EVALUACIóN FINAL CURSO DE MéTODOS EN BIOTECNOLOGíA

Evaluación Final Curso De Métodos En BiotecnologíaAlumno: Luis del Pozo YaunerTema: Métodos Electroforéticos

Introducción:El nivel de desarrollo científico de las sociedades humanas ha sido estimado, por lo general, por la cantidad de conocimientos que atesoran, y la velocidad con que pueden generarlos. El conocimiento puede dividirse en dos grandes grupos: el que es mera descripción de nuestro entorno, sea este tan lejano como los mismos límites de nuestro universo, o tan cercano como las moléculas que nos componen; y el conocimiento ÒherramientaÓ, aquel que en forma de métodos, procedimientos y tecnologías, nos permite describir y sobre todo, modificar el entorno. Se reconoce, pues, un vínculo de génesis entre todas las partículas de conocimiento. El descriptivo se va convirtiendo en ÒherramientasÓ, y estas ayudan a generar nuevo conocimiento. La historia de las ciencias, y por consiguiente del hombre, ha sido también la historia del desarrollo de los métodos. En el caso de la Biología, su desarrollo ha estado condicionado por el avance de los métodos para identificar, aislar y describir las propiedades de las moléculas que componen los sistemas vivos. Uno de los más poderosos y populares de estos métodos es la electroforesis, que dado su grado de diversificación actual, es propio hablar de Òmétodos electroforéticosÓ. Como concepto, ÒelectroforesisÓ se refiere al movimiento que experimentan partículas cargadas, ya sean átomos individuales o grandes complejos moleculares, cuando son sometidas al efecto del un campo eléctrico. En este trabajo nos ocuparemos brevemente de su desarrollo histórico, de la descripción de los conceptos y principios básicos, así como de los métodos fundamentales usados en la actualidad, señalando algunas aplicaciones específicas de estos. La historia se inicia con Tiselius:En 1937 el bioquímico sueco Arne Tiselius dio a conocer un nuevo método para separar las proteínas de una mezcla. Este consistía en colocar un cierto volumen de una solución proteica, conteniendo dos o más de estas sustancias, en un tubo de vidrio en forma de ÒUÓ, sin llenar completamente este. Luego el espacio

restante en cada extremo se ocupaba con el volumen necesario de una solución de electrolitos libre de proteína, que actuaba como amortiguador. Debía tenerse mucha precaución al adicionar la solución amortiguadora, evitando que se mezclara con la solución de proteínas a separar. Luego se sumergía un electrodo en la solución amortiguadora libre de proteína de cada extremo, se conectaban los electrodos a una fuente de electricidad y se sometía todo el sistema a una campo eléctrico. Los componentes de la mezcla de proteínas comienzan a moverse a diferente velocidad, en virtud del signo y el número global de cargas, hacia el electrodo de carga opuesta. Se formaban frentes de movimiento que reflejaban la movilidad electroforética de cada componente, o grupos de componentes de la mezcla, por lo que se le dio el nombre de Òelectroforesis de frentes en movimientoÓ o Òmoving boundary electrophoresisÓ a este sistema. Este fue uno de los pocos métodos analíticos poderosos disponibles para los investigadores en la época del desarrollo inicial de la química de proteínas; no obstante, presentaba importantes desventajas, como exigir gran cantidad de muestra, ser muy laborioso y en la práctica no permitir la separación total de los componentes de la mezcla. Las desventajas radicaban en su propia concepción; la separación electroforética ocurría en el medio fluido del amortiguador, sin un soporte que por su naturaleza limitara la mezcla convectiva de los frentes en movimiento, sobre todo, apenas se suprimía el campo eléctrico. Como consecuencia, el poder de resolución del método era escaso. El sistema de Tiselius dio paso a aquellos que usaban un soporte sólido, o gelatinoso, embebido en la solución amortiguadora de pH, y sobre el que ocurría la separación electroforética. En estos sistemas la muestra era constreñida a moverse en el seno del soporte, y dada las cantidades relativamente pequeñas de muestras que podían aplicarse, la separación total de los componentes en bandas o zonas discretas se convertía en algo realizable. Al desconectar el campo eléctrico, cada componente de la mezcla se mantenía por un tiempo suficiente para fines prácticos, en la zona del soporte que había alcanzado durante la electroforesis. Se les llamó genéricamente Òelectroforesis de zonaÓ. Como puede suponerse, el soporte es un factor muy importante para este tipo de electroforesis. Uno de los primeros soportes usados fue el papel de filtro, pero tenían inconvenientes como su fragilidad y la escasa resolución que se alcanzaba en la separación, debido fundamentalmente al gran

tamaño de sus poros moleculares. Luego, el papel fue sustituido por las membranas de acetato de celulosa. Este material posee menor tamaño de poro, lo que resulta en mayor poder de resolución, pero de igual manera es frágil, lo cual hace su manipulación laboriosa, y no permitía alta reproducibilidad de los resultados. Probablemente esto se debió a la dificultad de establecer adecuados controles de calidad en su manufactura. El acetato de celulosa, no obstante, tuvo amplia aplicación en los métodos electroforéticos usados en los estudios clínicos, y aún se usa en muchas instituciones de salud de varios países. Los geles han sido muy usados como soportes electroforéticos, y en la actualidad permanecen, en el caso de los geles de agarosa y de poliacrilamida, como los soportes de elección para la mayoría de las técnicas electroforéticas aplicadas en las investigaciones en Biología y ciencias afines. El gel de almidón, muy utilizado por la década de los años 50s-60s, incrementó la resolución para algunas aplicaciones, pero tiene inconvenientes, como rango limitado de poro, lo cual limita el tamaño molecular de las especies que pueden ser separadas, y además provoca intensa electroendoosmosis. Este fenómeno también fue una limitante en el uso de otro gel, el de agar. Hace relativamente poco tiempo, y gracias al desarrollo de metodologías de purificación, se hizo posible obtener agarosa de alta pureza a partir de ciertas algas productoras de agar. Con este polisacárido natural se confeccionan geles que, en dependencia de la concentración del reactivo usada, poseen rango de poros de variado tamaño, aunque la distribución práctica de poros es fundamentalmente hacia el rango de poro grande. La eliminación de los grupos cargados disminuyó considerablemente el fenómeno de electroendoosmosis. El gel de poliacrilamida, cuyo desarrollo se inició antes que el de agarosa, es, junto con este último, los soportes que poseen importancia práctica en la actualidad. Las ventajas fundamentales del gel de poliacrilamida son la posibilidad de seleccionar con suficiente exactitud el rango de poro, que en su conjunto es de menor talla que en el gel de agarosa; además, es un gel resistente, inerte y transparente; esta última propiedad es muy importante para su uso en ciertos sistemas ópticos de detección y estimación de la concentración de componentes separados por electroforesis en él. El desarrollo de los soportes electroforéticos hizo posible el diseño de nuevos procedimientos de separación de las biomoléculas y otras sustancias sobre la base a su carga eléctrica. Hoy son

varios las técnicas que se valen de este principio, y constituyen valiosas herramientas en el laboratorio bioquímico. Las más conocidas son la electroforesis en gel de agarosa, la electroforesis en gel de poliacrilamida en condiciones nativas o desnaturalizantes, el enfoque isoeléctrico, la electroforesis bidimencional, la electroforesis en gel de campo pulsante y la electroforesis capilar. A ellas haremos referencia, en algunos casos brevemente, en este trabajo.Referencia:

■ Ninfa, A. J., & Ballou, D. P. Fundamental Laboratory Approach For Biochemistry and Biotechnology. Pps 125-156. Fitzgerald Scienc4e Press. Inc., Bethesda, MD.

■ Voet, D., & Voet, J. (1995). Biochemistry. 2nd Edition, pps. 89-96. John Wiley & sons, Inc., New York.

Algunos Conceptos Básicos:De inicio trataremos los conceptos básicos que se aplican a la mayoría de los métodos electroforéticos.

■ Una molécula de carga q en un campo eléctrico E experimenta una fuerza (F):

F=qE

■ La movilidad electroforética (µ) de una molécula es:

µ = = Donde v es la velocidad de movimiento de la molécula, y f es su coeficiente friccional.

■ Ley de Ohm: V= P=VI=RI2

Donde V es voltaje, R es resistencia del conductor, I es intensidad de la corriente y P es potencia. Esta relación nos indica que si incrementamos la I de la corriente en el experimento electroforético, la P, y por consiguiente, el calor generado se incrementa al cuadrado.

■ Dada la relación establecida en la ecuación:

P= Cuando disminuimos la fuerza iónica del amortiguador, y por tanto su conductividad, aumenta la resistencia (R) al paso de la corriente. Como consecuencia disminuye el calor generado por el sistema (P), si V se mantiene constante. Por cada dos veces que disminuyamos la fuerza iónica del amortiguador, podemos incrementar 1.4 veces el V sin provocar incremento significativo del calor generado

■ La movilidad efectiva (µeff) de un electrolito débil es un promedio de la movilidad efectiva de su forma ionizada:

µeff=µxDonde x es la fracción de moléculas ionizadas a un pH determinado, y µ es la movilidad de la forma ionizada. Referencias:

■ Garfin, D. E. Electrophoretic Methods (1996). Reprinted from: Introduction to Biophysical Methods for Protein and Nucleic Acid Research. Edited by Jay A. Glasel and Murray P. Detscher. Academic Press.

■ Ninfa, A. J., & Ballou, D. P. (1998) Fundamental Laboratory Approach For Biochemistry and Biotechnology, pps 125-156. Fitzgerald Science Press. Inc., Bethesda, MD.


Electroforesis en gel de agarosa: La agarosa es un polisacárido lineal que se extrae de ciertas variedades de algas productoras de agar (FMC Bioproduct, 1988, Sambrook et al, 1989). La estructura básica que se repite a lo largo del polímero lineal es una unidad compuesta por una molécula de D-galactosa y una de 3,6 anhidrido galactosa, unidas por enlace glucosídico. Cuando la agarosa es hidratada y fundida a

temperaturas altas y luego dejada enfriar, forma un gel en el que las cadenas polisacarídicas se entrelazan, en asociación física estabilizada por innumerables puentes de H, formando ases interconectador, que finalmente resultan en una densa red o malla cuyos poros moleculares limitan los espacios internos de la trama. Los poros del gel de agarosa son mayores que los del gel de poliacrilamida. En el gel de agarosa solo pueden ser separadas con eficiencia proteínas de peso molecular mayor de 500kDa, y moléculas de DNA de alrededor de 2000 pb. En la medida que se incrementa la concentración de agarosa en el gel, el tamaño de los poros disminuye. En el mercado hay disponibles muchos tipos de celdas electroforéticas para realizar electroforesis en geles de agarosa. Igualmente, están disponibles muchos tipos de agarosas, que varían en sus propiedades químicas y físicas, como la intensidad de la electroendoosmosis que su composición química promueve, la resistencia mecánica del gel, la porosidad, y la temperatura de gelificación, que para algunas de grado especial es significativamente baja sin que se afecten otras características como la resistencia mecánica del gel. Estas propiedades influyen significativamente en los resultados del proceso de separación electroforética (Sambrook et al, 1989). Aunque se ha usado en técnicas para proteínas, el uso más frecuente de la agarosa es en los métodos electroforéticos de separación de ácidos nucleicos. Varios factores determinan la velocidad de migración de una molécula de DNA en el gel de agarosa. Estos factores son: Tamaño molecular del DNA: Debido a la resistencia friccional que ejerce el gel sobre las moléculas de DNA que se mueven en su seno en respuesta al campo eléctrico, aquellas de mayor talla se moverán a una menor velocidad que las más pequeñas. La velocidad de migración es inversamente proporcional al log10 del número de pares de bases, para el caso de una molécula de DNA de doble cadena y lineal. Conformación del DNA: La forma o conformación del DNA influye significativamente en su movilidad electroforética. El DNA súper enrollado (forma I), el DNA circular mellado (forma II) y el DNA lineal de igual talla molecular migran a través del gel de agarosa a diferente velocidad. Si bien la

concentración de agarosa del gel es el factor que primariamente determina la velocidad relativa de migración de estas formas, otros factores como la fuerza del campo eléctrico aplicado, la fuerza iónica de amortiguador y el grado de súper enrollamiento de la molécula en forma I, también influyen en el valor de este término (Johnson and Grossman, 1977). Bajo algunas condiciones el DNA en forma I migra más rápidamente que el III, pero en otras el orden se invierte.Concentración de agarosa en el gel: Un fragmento de DNA lineal de un tamaño molecular dado migrará a diferentes velocidades en geles que posean diferente concentración de agarosa. La relación entre la movilidad electroforética (µ) y la concentración del gel (τ) se expresa en la siguiente ecuación: logµ=logµ0-Kr τDonde µ0 es la movilidad electroforética libre del DNA, Kr es el coeficiente de retardo, una constante que se relaciona a las propiedades del gel y al tamaño y la forma de la molécula que migra. Esta relación nos dice que, modificando la concentración del gel, podemos resolver mezclas no totalmente separadas en un primer intento.El rango óptimo de separación de geles de agarosa en dependencia de la concentración se indica en la tabla siguiente:

Cantidad de agarosa en el gel, expresado en % (p/v)

Rango óptimo de separación de fragmentos lineales de

DNA (kb)0,3 5-600,6 1-200,7 0,8-100,9 0,5-71,2 0,4-61,5 0,2-32,0 0,1-2

Voltaje aplicado: A bajo voltaje la velocidad de migración del DNA es proporcional al voltaje aplicado. Pero en la medida que el campo eléctrico es intensificado, la movilidad de las moléculas de DNA de alto peso molecular comienza a no corresponder proporcionalmente al incremento del voltaje. Como

consecuencia, el rango efectivo de separación en el gel de agarosa disminuye con la elevación del voltaje. La máxima resolución en la separación de fragmentos de DNA de más de 2kb se logra cuando se realiza la corrida electroforética a 5 V/cm. La distancia a usar en este cálculo es la que separa en línea recta a los electrodos de la celda electroforética.Dirección del campo eléctrico: Los fragmentos lineales de DNA se alinean con relación a la dirección del campo eléctrico, de modo que al cambiar la dirección de este, estos fragmentos deben realinealse para iniciar el movimiento en la nueva dirección del campo. Los fragmentos de mayor talla consumen más tiempo en el proceso de realineamiento, y quedan rezagados con relación a los de talla menor. El cambio frecuente de la dirección del campo eléctrico es la base de la electroforesis en gel de campo pulsante. Composición de bases del DNA y la temperatura de corrida: La influencia de estos factores es menor en el gel de agarosa que en la electroforesis en gel de poliacrilamida, no obstante, en el caso de geles confeccionados con agarosa de bajo punto de fusión, o aquellos que contienen menos del 0,5% del polisacárido, es muy conveniente realizar el ensayo a 40C, para evitar la fusión del gel. Presencia de colorantes intercalantes: El colorante fluorescente intercalante Bromuro de Etidio reduce la movilidad electroforética del DNA lineal en un 15%. Este efecto se debe al intercalamiento de las moléculas del colorante entre las bases apiladas en la estructura del DNA, que torna rígidas y alargas a las moléculas lineales o circulares con mella.Composición del amortiguador: la movilidad electroforética del DNA es afectada por la composición y la fuerza iónica de amortiguador usado en el ensayo. En cuanto a la fuerza iónica, su efecto es expresión de la relación entre este parámetro y la corriente eléctrica que fluye a través del gel. En un medio de fuerza iónica cero, la conductancia es mínima y la movilidad electroforética del DNA es escaza, si es que migra. En cambio, si la fuerza iónica del amortiguador es muy alta, el flujo de corriente se incrementa, lo que resulta en mayor generación de calor, lo que puede incrementar la temperatura del sistema más allá del punto de fusión de la agarosa, y eventualmente desnaturalizar el DNA. Fragmentos lineales de DNA de doble cadena migran alrededor de un 10% más rápido en amortiguador TAE que en TBE o TPE, sin que esto signifique

incremento en la resolución. En cambio, la resolución de mezclas de DNA súper enrollado es mejor en TAE que en TBE. Varios aparatos y protocolos han sido desarrollados para satisfacer las diferentes exigencias del trabajo experimental relacionado con la separación electroforética de mezclas de DNA. Información abundante sobre este particular se encuentra disponible en el web y textos de métodos.Referencia:

■ FMC BioProducts (1988) ÒFMC BioProducts Sources Books,Ó Agarose Monog., 4th ed., pp. 51-106. FMC BioProducts, Rockland, Me.¤ Garfin, D. E. Electrophoretic Methods (1996). Reprinted from: Introduction to Biophysical Methods for Protein and Nucleic Acid Research. Edited by Jay A. Glasel and Murray P. Detscher. Academic Press.

■ Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). ÒMolecular Cloning: A Laboratory Manual,Ó 2nd ed. Cold Spring Harbor Lab., Cold Spring Harbor, NY.

Electroforesis en Gel de Poliacrilamida: La casi inigualable resolución y flexibilidad que ofrece la electroforesis en gel de poliacrilamida ha convertido a esta técnica en una de las más usadas para la separación de proteínas y ácidos nucleicos. Es posible variar la porosidad del gel y la composición del sistema de amortiguadores para satisfacer una amplio rango de requerimientos experimentales. La separación puede realizarse en virtud de las diferencias de carga, del tamaño molecular, o por una combinación de ambos. El gel de poliacrilamida se forma por la copolimerización de la acrilamida y un comonómero entrecruzante que es usualmente la N,NÕ-metilenebisacrilamida (bisacrilamida). El entrecruzante es un agente acrílico bifuncional que une covalentemente las cadenas lineales de poliacrilamida (Righetti et al, 1989). Además de la bisacrilamida se han usado otros entrecruzantes para propósito especiales, como el piperazinediacrilamide, o PDA, que permite obtener un gel con bajo fondo en la coloración con plata.

La reacción de polimerización, representada en la figura situada debajo, es de adición de grupos vinilos, iniciada por un sistema generador de radicales libres (Flory et al, 1953).

En las recetas más comunes de confección de geles de poliacrilamida, la polimerización es iniciada por la adición de persulfato de amonio, que es el agente iniciador, y de tetrametiletilenediamina (TEMED), que funciona como agente acelerador. El TEMED acelera la descomposición del persulfato de amonio en radicales sulfato libres, los que inician la reacción al convertir a los monómeros de acrilamida en radicales libres que reaccionan, a su vez, con los monómeros no activados, procediendo de esta forma la polimerización; para esta se necesita la presencia de la base libre del TEME. La estructura del agente entrecruzante N,NÕ-metilenebisacrilamida es la de dos moléculas de acrilamidas unidas a través de sus grupos amino por una unidad de metilene. Esta estructura le permite a una molécula de entrecruzante participar en la reacción de polimerización de dos cadenas lineales en crecimiento. Cuando esto ocurre, las cadenas lineales quedan conectadas covalentemente por un puente de Cs y Ns alternantes. Este puente posee dos enlaces C-N de tipo peptídico, lo cual probablemente aporta cierto grado de rigidez a la estructura. Cuanto mayor es la concentración relativa del agente entrecruzante, mayor es la probabilidad de que se entrecrucen las cadenas lineales de poliacrilamida, aunque esto parece ser verdadero hasta cierto valor de concentración, como veremos luego (Gelfi C et al 1981a). Muchos factores influyen profundamente el curso de la reacción de polimerización, y si no son adecuadamente controlados, se pueden comprometer las propiedades del gel y Consecuentemente la reproducibilidad del ensayo electroforético. Entre estos factores destacan:

q Pureza de los reactivos: La presencia de impurezas como el ácido acrílico,

iones metálicos y acrilamida lineal, tanto en la acrilamida como en la Bis, pueden inhibir la reacción de polimerización, y modificar las propiedades del gel.q La concentración de los comonómeros: Los geles de poliacrilamidan son caracerizados convencionalmente por dos términos: El %T y el %C. El primero indica la concentración total de monómeros (acrilamida + bisacrilamida) expresadas en g/100mL de solución, y el %C indica la proporción relativa del entrecruzante en la concentración total de monómeros ([Acril+Bis]/[Bis]x100). El tamaño de poro efectivo del gel es una función inversa del valor de %T, lo cual significa que, a mayor valor de %T, para un valor fijo bajo de %C, menor el tamaño de poro. Por otra parte, la función que expresa la relación entre el valor de %C y el tamaño de poro es bifásica; puede ser representada como una campana invertida en un eje de coordenadas, donde x es %C e y es tamaño de poro. En la medida que se incrementa %C a un %T constante, disminuye en tamaño de poro, siendo este mínimo al valor de 5%C. A valores mayores de 5%C el tamaño de poro se incrementa (Gelfi C et al 1981a).

El rango de fraccionamiento con relación al valor de %T se representa en la siguiente tabla:

Rango de fraccionamiento para geles de poliacrilamida%T Rango Mr óptimo5-12 20,000-150,00010-15 10,000-80,000> 15 <15,000

q El pH: Cuando el pH es menor de 6, la eficiencia de la reacción de polimerizaci˜n disminuye significativamente (Caglio & Righetti, 1993). Una forma de atenuar este efecto cuando se necesitan confeccionar geles a pH ácido, es usar riboflavina como fuente de radicales libres adicional, y la luz como activador en presencia de O2, de la liberación de estos radicales.q La presencia de O2: Puesto que la formación de poliacrilamida procede como una reacción de polimerización de radicales libres, la presencia de

agentes que reaccionan con estos y los atrapan, la inhibe. El oxígeno es ese tipo de inhibidor. Por eso se hace necesario, para ciertos requerimientos, eliminar el oxígeno disuelto en las soluciones de reactivos mediante degasado. En ocasiones también se hace necesario aislar del aire circundante a la solución donde está ocurriendo la reacción de polimerización, mediante una fina capa de agua previamente degasada. q Temperatura: En control de la temperatura es crítico para lograr reproducibilidad en la polimerización de la acrilamida. La temperatura tiene un efecto directo sobre la velocidad de la reacción de polimerización, pues esta es exotérmica. El calor liberado acelera, pues, a la reacción. La temperatura de polimerización también influye en las características finales del gel. El rango óptimo es de 230C a 250C. A temperaturas cercanas a cero se forman geles de aspecto turbio, porosos e inelásticos, y a temperaturas muy altas se forman cadenas poliméricas cortas y el gel también es inelástico (Gelfi, C. et al 1981b). q Aditivos: Aditivos como los detergentes SDS y Tritón X-100 pueden adicionarse a la mayoría de los sistemas de amortiguadores sin que se altere la polimerización. En cambio, la formamida y la urea provocan la formación de geles con tamaño de poro menor que el que acontecería si no estuvieran estos reactivos presentes. Además de los propios aditivos, los contaminantes que en ocasiones les acompañan pueden también inhibir la polimerización. q Tiempo: Si bien luego de transcurridos 15-20 min puede constatarse la formación del gel, la polimerización aún está procediendo. Se sugiere que cuando se emplee el sistema persulfato de amonio-TEMED, se permita que esta transcurra durante al menos 2 horas, para asegurar reproducibilidad en las características del gel. En cuanto al sistema riboflavina-luz, por lo general la reacción procede más lentamente, dependiendo de la intensidad de la luz.

La electroforesis en gel de poliacrilamida puede ser efectuada en condiciones no desnaturalizantes para las proteínas, y en ese caso se clasifica como Òelectroforesis en condiciones nativaÓ, o por el contrario, en Òcondiciones desnaturalizantesÓ, tales que aseguran total desnaturalización de estas. Esto es importante, pues si bien en ambos casos la fuerza que mueve a las moléculas cargadas a lo largo del gel es la atracción del polo con carga opuesta, la

separación de los componentes de la muestra no ocurre por el mismo principio en ambos casos.El sistema de amortiguadores tienen un efecto profundo en la corrida electroforética. El amortiguador determina las condiciones de potencia del sistema, y afecta la separación y resolusión de las bandas. Adicionalmente, cuando se desea realizar un experimento en condiciones nativas, es muy importante tener en cuenta el efecto de las características de fuerza iónica, composición de iones y pH del amortiguador sobre la actividad de las proteínas de la muestra. Los sistemas de amortiguadores para electroforesis en gel de poliacrilamida se clasifican en continuos y discontinuos.Sistema continuo: Es aquel en el que un mismo amortiguador, con el mismo pH, es usado en el gel y en los reservorios de los electrodos. Adicionalmente, las muestras son cargadas directamente en el gel donde ocurrirá la separación. Este sistema tiende a la difusión de las bandas, lo cual hace que tenga menor resolución que el sistema discontinuo. Cuando la concentración de los componentes de la muestra se acerca a 1mg/mL, este sistema puede ser usado.Sistema discontinuo: El primer sistema discontinuo fue desarrollado por Ornstein y Davis en 1964. Este fue un sistema para el análisis de proteínas del suero en estado nativo, que desde entonces ha sido aplicado a muchas otras proteínas. Este sistema consiste en cuatro partes interrelacionadas: (1) el gel concentrador (stacking gel), (2) el gel separador (running gel), (3) el amortiguador de electrodos, y (4) la muestra. El gel concentrador es de poros grandes (4%T), contiene amortiguador Tris-Cl 0.125M, pH 6.8 y se prepara en la parte superior del gel separador. Este último posee valores de 5-30%T, contiene Tris-Cl 0.375M, pH 8.8. El amortiguador de los electrodos es Tris 0.025 Glicina 0.192M, pH 8.3. La muestra se deposita sobre el gel concentrador, diluida en amortiguador de muestra Tris-Cl 0.0625 M, pH 6.8, y es cubierta por el amortiguador del cátodo (reservorio superior). Cuando es aplicado el campo eléctrico, los aniones Cl- y glicinato, así como las proteínas cargadas negativamente, comienzan a moverse hacia el ánodo. El Tris y otros cationes se mueven hacia el cátodo. Como la µeff del Cl- es mayor que la de la glicina a pH 8.3 (Dado que el pKa de la glicina es 9.8, solo 1/30 moléculas están disociadas a glicinato), se forma una

región de baja conductividad y alto campo eléctrico detrás del frente de movimiento del Cl-. Esto acelera el movimiento de las proteínas de la muestra y del frente de movimiento de la glicina a la misma velocidad del frente del Cl-. Se forma un arreglo de frentes en movimiento, con el Cl- al frente, la glicina en la retaguardia, y las proteínas en varios frentes comprimidos entre ambos. Esto ejerce un efecto de concentración de las proteínas que se ha planteado pueden llegar a alcanzar concentraciones del orden de 100mg/mL (Chen et al, 1978). El pH del gel concentrador asciende a 8.3 en la medida que la glicina se mueve a través de él. Dado que el gel concentrador posee poros grandes, no ejerce efecto de tamiz, y las proteínas avanzan solo dependiendo del efecto del campo eléctrico. Cuando los frentes atraviesan el gel concentrador y entran en el gel separador, la menor porosidad de este sí ejerce efecto de tamiz, resistiéndose al paso de las proteínas en dependencia de su tamaño molecular. Como este es un sistema en condiciones nativas, también la forma influye en la movilidad electroforética. A lo anterior debe añadirse que al penetrar las proteínas en el gel separador, experimentan un súbito cambio del pH del medio, de 8.3 a 8.8, el que luego asciende a 9.5 cuando el amortiguador Tris-CL es sustituido por Tris-glicina. Este es el valor de pH al ocurre la separación de los componentes de la muestra en bandas discretas.Laemmlin en 1970 modificó es sistema de Ornstein-Davis para usarlo en la estimación del peso molecular de las proteínas. él incorporó el detergente aniónico dodecil-sulfato de Na (SDS) al 0.1% en los amortiguadores del sistema de Ornstein-Davis, y adicionó un paso de desnaturalización al calor en presencia de 0.2% de SDS y 2-mercaptoetanol al 5%. En tales condiciones, los enlaces S-S de las proteínas se reducen, y las proteínas se desnaturalizan totalmente. El SDS se asocia a estas en una relación de 1.4g de SDS por gramo de proteína. Las propiedades del SDS se hacen dominantes sobre las de las proteínas, estas adoptan una conformación extendida, similar a bastoncillos, con dimensiones proporcionales a su masa molecular. En adición, la densidad de carga del complejo SDS-proteína es independiente del pH en el rango de 7 a 9. Por todo lo dicho, se hace claro que el fraccionamiento de los componentes de la muestra depende únicamente del efecto de tamiz del gel de poliacrilamida, de modo que es posible estimar con buena exactitud la masa molecular de una proteína, si se

posee varias de ellas de masa conocida (marcadores de peso molecular para SDS-PAGE), pues la relación entre la movilidad relativa de estas y el logaritmo de su peso molecular es lineal dentro de un rango relativamente amplio para cada sistema. Están disponibles en el comercio numerosos kits de proteínas de referencia de peso molecular, tanto para rango de peso bajo, alto, como rango amplio. En la tabla que sigue se muestra un ejemplo, comercializado por la compañía Bio-Rad, así como el gráfico donde se representa la relación lineal entre movilidad relativa relativa y log del peso molecular.

Aunque el formato más usado en la actualidad es el gel en forma de lámina, también se han desarrollado sistemas que emplean geles de forma tubular. En las electroforesis en gel de poliacrilamida, para la detección de las bandas se emplean colorantes como el negro amido, el azul brillante de Coommassie, pero también coloraciones con mayor sensibilidad para detectar proteínas. Uno de los métodos más sensibles es la coloración con plata (Rabilloud et al., 1990). También se emplea la coloración con Cu cuando se desea evitar el paso de

fijación en electroforesis de SDS-PAGE. La especificidad de reconocimiento de los anticuerpos permite identificar a componentes específicos de la muestra, separados durante la electroforesis, mediante el método de inmuno-fijación (immuno-blotting). Dado que los anticuerpos reconocen fundamentalmente epitopes conformacionales, su uso se limita a sistemas de electroforesis nativa.Como veremos en el caso de la 2D-SDS PAGE, la electroforesis en gel de poliacrilamida puede combinarse con otros métodos, como el PCR, para detectar cambios puntuales en genes que pueden ser causa de enfermedad, como es el caso de la Fenilcetonuria (Michiels, L., et al.)Referencias:

■ Caglio, S & Righetti, P. G. (1993) On the pH dependence of polymerization efficiency, as investigated by cappillary zone electrophoresis. Electrophoresis 14, 554-558.

■ Chen, D., Rodbard, D., Chrambach, A: (1978) Poliacrilamide gel electroforesis with optical scanning, using multiphasic buffer systems: The syack. Anal. Biochem. 89, 596-608.

■ Davis, B. J. (1964) Disc electrophoresis. II. Methods and application to human serum proteins. Ann. N. Y. Acad. Sci. 121, 404-427.

■ Flory, P. J. (1953) ÒPrinciples of Polimer Chemistry.Ó Cornell University Press, Ithaca, NY.¤ Garfin, D. E. Electrophoretic Methods (1996). Reprinted from: Introduction to Biophysical Methods for Protein and Nucleic Acid Research. Edited by Jay A. Glasel and Murray P. Detscher. Academic Press¤ Gelfi C and Righetti P. G. (1981a)Polymerization kinetics of polyacrylamide gels I. Effect of different cross-linkers, Electrophoresis 2, 213Ð219

■ Gelfi C and Righetti PG, Polymerization kinetics of polyacrylamide gels II. Effect of temperature, Electrophoresis 2, 220Ð228 (1981b)

■ Laemmli, U. K. (1970). Cleavage of structural proteinsdurind the assembly of the head og bacteriophage T-4. Nature, 227, 680

¤ Michiels, L., Fran•ois, F., Raus, J. (1999) Identification of Disease Causing Mutations in Phenylketonuria by Denaturing Gradient Gel Electrophoresis Using the DCodeª System. Bulletin 2377. Bio-Rad. www.bio-rad.com.


■ Ornstein, L. (1964) Disc electrophoresis. I. Background and theory. Ann. N. Y. Acad. Sci. 121, 321-349.

■ Rabilloud, T. (1990). Mechanism of protein silver staining in polyacrilamide gels: A 10-years synthesis. Electrophoresis 11: 785-794.

■ Righetti PG et al., (1981) Polymerization kinetics of polyacrylamide gels. III. Effect of catalysts, Electrophoresis 2, 291Ð295.


Enfoque isoeléctrico: Este es un poderoso método para separar las moléculas anfotéricas de una mezcla sobre la base del valor de su pI. Las proteínas son el caso mejor conocido de moléculas afotéricas. Ellas se cargan negativamente cuando el pH del medio es mayor que el valor de su pI, en cambio, cuando ocurre lo contrario, el pH<pI, entonces su carga neta es de signo positivo. En ambas condiciones las moléculas se moverán en un campo eléctrico, obviamente dirigiéndose hacia el electrodo de carga opuesta. Pero la esencia del método radica en la respuesta de las moléculas anfotéricas cuando el valor de pH del medio se iguala al valor de su pI. En esas condiciones las cargas negativas de la especie son compensadas por igual número de cargas positivas, resultando en carga neta igual cero. Cuando esto ocurre, la molécula deja de responder al campo eléctrico y se detiene en la región donde pH=pI. Para la realización del enfoque isoeléctrico es necesario establecer un gradiente lineal y estable de pH a lo largo del gel, cuyo rango de valores estará determinado por los pI de los componentes de la muestra que son de nuestro

http://www.bio-rad.com/

interés. Los soportes que se han usado para este método son, fundamentalmente, los geles de agarosa y el de poliacrilamida. Para el isoelectroenfoque analítico son deseables geles con tamaño de poro grande, como el de agarosa al 1% o el de poliacrilamida con 5% de T y 3% de C. Para establecer el gradiente se emplean moléculas pequeñas (alrededor de 300-1000 da) llamadas portadores sintéticos de anfolitos, los cuales son estructuralmente compuestos poliamino-policarboxilatos (Just et al, 1983). Estos compuestos son incluidos en la mezcla de monómeros precursores del polimero del gel, y posterior a la confección de este, son distribuidos a todo lo largo mediante la aplicación de un campo eléctrico. Ellos se distribuyen en un gradiente de pH que se incrementa monotónicamente del ánodo al cátodo (Laas et al, 1989). La pendiente del gradiente de pH es determinado por el intervalo de pH cubierto por la mezcla de anfolitos usada, y por la distancia entre los electrodos.Para lograr buenos resultados en el ensayo de enfoque isoelétrico, un elemento clave es la selección del rango de anfolitos a usar, el cual, idealmente, debe ser centrado por el valor de pI de la proteína de interés. Esto aseguraría que la proteína enfocara en la zona lineal del gradiente. El isoelectroenfoque es una técnica muy poderosa, que permite la separación de moléculas que difieren en su valor de pI por menos de 0,05 unidades de pH. Adicionalmente posee la característica de no provocar la desnaturalización de las proteína, lo cual a permitido desarrollar técnicas a nivel preparativo.Posterior al enfoque de los componentes de la muestra, se ejecuta la etapa de detección, la que se realiza mediante el uso de colorantes con afinidad para proteína (Ej. Azul de Comassie y Negro Amido) o específicos para glicoproteinas o lipoproteínas.Referencias:

¤ Garfin, D. E. Electrophoretic Methods (1996). Reprinted from: Introduction to Biophysical Methods for Protein and Nucleic Acid Research. Edited by Jay A. Glasel and Murray P. Detscher. Academic Press.

■ Just, W. W. (1983) Synthesis of carrier ampholytes for isoelectrric focusinf. In ÒMethods in EnzimologyÓ (C. Hirs and S. Timasheff, eds.), Vol. 91,

pp281-298. Academic Press, New York.■ Laas, T. (1989) Isoelectric focusing in gels. In ÒProtein purification.

Principles, Hight Resolution Methods, and ApplicationsÓ (J-C Janson and L. Ryden, eds), pp. 376-403. VCH, Weinheim.



Electroforesis Bidimensional:Dado que cada método electroforético hace uso de, fundamentalmente, una propiedad física diferente de las moléculas de la muestra para lograr su separación, la combinación de dos de estos puede incrementar la capacidad de resolución del ensayo en general. Esta es la lógica de la electroforesis bidimensional, en la que se combinan dos técnicas que separan las moléculas en base a características diferentes. La combinación más frecuentemente usada es la de enfoque isoeléctrico con la electroforesis en gel de poliacrilamida con SDS. A este ensayo de le ha denominado 2-D PAGE. Los dos principios de separación son ortogonales y complementarios. Primero las proteínas son separadas en bases a su pI en un ensayo de enfoque isoeléctrico, usando un solo carril de un gel en lámina, o un gel en forma cilíndrica. La separación ocurre en una sola dimensión. Luego la porción de gel que contiene las proteínas es cortada en una tira, en el caso de ser un gel en lámina, o el cilindro de gel es tomado, y en ambos casos se adosan al extremo superior de un gel para SDS-PAGE. El gel del enfoque isoeléctrico se orienta de modo tal que el eje de movimiento de las proteínas durante el enfoque queda perpendicular a la dirección de movimiento durante la SDS-PAGE. Luego se realiza la segunda corrida, de modo que ahora la separación ocurre en una segunda dimensión, perpendicular a la anterior. Una vez concluida la separación electroforética, se efectúa el paso de identificación mediante coloraciones estándar para proteína, o especiales, según el interés. El uso de la coloración de plata aporta mucha sensibilidad de detección al método. El resultado final va ha depender de las condiciones elegidas para cada ensayo;

como, por ejemplo, el rango del gradiente de pH usado en el ensayo de enfoque isoeléctrico.La combinación de ambos métodos incrementa significativamente la resolución, y hace posible resolver mezclas muy complejas de proteínas. Esta característica ha convertido a la 2D-PAGE en una herramienta muy poderosa para la identificación de nuevas proteínas, y el estudio de los patrones de expresión de estas en la célula, bajo condiciones específicas, lo cual es muy importante en la comprensión de nuestro proteoma. Referencias:

¤ 2-D Electrophoresis for proteomics. A Method and Products Manual. Bulletin 2651. Bio-Rad. www.bio-rad.com.

■ Hoving, S., Voshol, H., van Oostrum, J. (2001) High-Performance 2-D Gel Electropho resis using Narrow pH-Range ReadyStrip- IPG Strips. Bulletin 2587. Bio-Rad. www.bio-rad.com

■ Pandey, A., Mann, M. (2000). Proteomics to study genes and genomes. Nature, 405(15):837-846.

Electroforesis Capilar (EC):La EC es una familia de técnicas relacionadas que emplean capilares de pequeño diámetro interior (20-200µm) como celda donde estará contenido el sistema de amortiguadores donde ocurrirá la separación electroforética de los componentes de la muestra. Este formato permite acometer la separación, con muy alta eficiencia, de moléculas de estructura química muy diferentes, como proteínas y péptidos, DNA y carbohidratos. También es posible separar moléculas estructuralmente muy similares entre sí, como los enantiómeros, y pequeñas del tipo de los compuestos farmacéuticos. Las prestancias de este conjunto de técnicas se deben, en buena medida, al uso de alto voltaje, lo cual genera dentro del capilar tanto flujo electroforético de las especies iónicas de la muestra, como electroosmótico de la solución amortiguadora. Las altas propiedades de separación y los electroforegramas que resultan, recuerdan a un híbrido entre la SDS-PAGE y la HPLC. Las características generales y ventajas que supone el uso de la EC se pueden resumir en:



■ Emplea un tubo capilar donde ocurre la separación electroforética.■ Utiliza campos eléctricos de muy alta tensión, de hasta 500 V/cm.■ Usa metodología de detección muy moderna, que le dan al electroforegrama

semejanza a un cromatograma.■ Posee un grado de eficiencia del orden de la cromatografía de gas en capilar

, llegando a superarla en algunos casos.■ Requiere cantidades de muestra muy pequeñas, lo cual le convierte en un

procedimiento analítico muy útil para el trabajo con sustancias muy escasas y valiosas.

■ Su explotación es económica, pues consume pequeñas cantidades de reactivos.

■ Es aplicable a un amplio grupo de compuestos, superando en este rubro a muchos otros métodos electroforéticos.

La configuración instrumental básica para la EC es simple, y consta de los siguientes elementos:

a) Un tubo capilar de sílice fundida que posea una ventana óptica para el análisis del contenido.b) Una fuente de alto voltaje regulable.c) Un par de electrodos.d) Dos reservorios de solución amortiguadora.e) Un detector de UV.

Cada extremo del capilar es sumergido en uno de los reservorios de amortiguador, y la ventana óptica es alineada con el detector. Luego de llenar el capilar con solución amortiguadora, la muestra puede ser introducida en él sumergiendo el extremo opuesto al que posee la ventana óptica en la solución de muestra, y posteriormente elevando este extremo unos 30 cm por encima del nivel del reservorio del extremo del detector. Ya introducida la muestra, se aplica el campo eléctrico y se procede a la separación electroforética. De hecho, la mayoría de los trabajos realizados en EC antes de 1988 se valieron de sistema confeccionados por los mismos investigadores. Los sistemas comerciales modernos se caracterizan, sobre todo, por su alto grado de automatización, conferido por el control computarizado de todas las operaciones, como la

inyección electrocinética de la muestra, el control exacto de la presión y la temperatura del sistema, el desarrollo de los procedimientos y las estrategias de separación; poseen eficaces sistemas de disipación de calor y de recolección de fracciones. Es necesario precisar el uso de algunos términos en el contexto de la electroforesis capilar. Los fundamentales son:Tiempo de migración (tm): Es el tiempo que toma a un soluto en moverse desde el inicio del tubo capilar hasta la ventana óptica del detector.Movilidad electroforética (µep): Se expresa en cm2/Vs y es la velocidad con que se mueve un ion a lo largo del capilar en respuesta a la intensidad del campo eléctrico.Velocidad electroforética (Vep): Se expresa en cm/s, y es la distancia que recorre un soluto por unidad de tiempo.Fuerza del campo eléctrico (E): Se expresa en V/cm. La relación entre estos términos es la siguiente:

µep = = (1)La Vep se calcula dividiendo la longitud del capilar (Ld) entre el valor de tm. La µep se calcula dividiendo la Vep entre E. La Vep es independiente del voltaje y de la longitud del capilar, pero es altamente dependiente del tipo de amortiguador empleado, el pH y la temperatura.Las dos medidas del capilar que son importantes son Ld, que es la longitud del capilar desde su inicio hasta la ventana del detector, y la longitud total (Lt), que incluye el segmento de capilar que continúa luego de la ventana del detector, y que sirve para conectar al capilar con el reservorio de amortiguador. La separación mensurable ocurre en el segmento Ld, mientras que el cálculo de E se realiza dividiendo el voltaje total entre Lt. La ecuación 1 nos permite calcular la µep aparente, pero el cálculo de la real debe tener en cuenta el efecto del flujo electroosmótico. Este factor debe ser

cuidadosamente controlado durante los experimentos de separación, para lograr alta reproducibilidad.La electroosmosis: Es uno de los procesos determinantes de la separación de los componentes de la muestra en la EC. Este fenómeno es consecuencia de los grupos ionizables presentes en la superficie interna de las paredes del capilar. La sílica fundida posee grupos silanol que están en contacto con el amortiguador contenido en el capilar. El grado de ionización de estos grupos depende del pH del amortiguador empleado en el ensayo. La sílica posee un pI de 1.5. A pH por encima de 1.5 la sílica se carga negativamente, y atrae cationes componentes del amortiguador, lo cual crea una doble capa de iones. Cuando se aplica un campo eléctrico a lo largo del capilar, la tendencia es que los aniones se muevan hacia el ánodo, y los catones hacia el cátodo, pero los grupos silanol de carga negativa están fijados en la pared del capilar y non pueden moverse libremente, pero los cationes de la doble capa, que si pueden hacerlo, se mueven hacia en cátodo, y en su movimiento arrastran moléculas de agua. El resultado es un flujo neto de solución amortiguadora en dirección al electrodo negativo, que puede ser muy intenso. A pH igual a 9.0 en amortiguador borato 20 mM el flojo electroosmótico (FEO) es de alrededor de 2mm/s, lo que en un capilar de 50 µm de diámetro interno significa un flojo neto de 4 nL/s. A pH igual a 3 el FEO es mucho menor, de 0.5 nL/s, lo que está determinado por el menor grado de ionización de los grupos silanol de la pared capilar. El FEO (Veo) es definido por la siguiente relación:

Veo = Donde es la constante dieléctrica del medio, η es la viscosidad del amortiguador, y ζ es el potencial zeta medido en el plano de corte cerca de la interfase líquido-sólido. El ζ está relacionado al inverso de la carga por unidad de área de superficie, al número de electrones de valencia, y a la raíz cuadrada de la concentración de electrolito. Puesto que esta es una relación inversa, el incremento de la concentración de electrolitos resulta en la disminución del FEO.El FEO puede ser controlado, o incluso suprimido, para realizar ciertos modos de

ensayo de EC. Por otro lado, el FEO permite el análisis simultaneo de aniones, cationes y especies neutras en una misma corrida electroforética. A pH neutro o alcalino el FEO es suficientemente intenso como para obligar a todas las especies a moverse hacia el ánodo. El orden de movimiento es: Cationes, especies neutras y finalmente aniones.Para describir el efecto del pH sobre el FEO y la movilidad de las especies ionizables de la muestra, emplearemos el caso de péptidos. Los péptidos son zwitteriones; a pH muy alcalino se cargan negativamente, y se mueven hacia el electrodo positivo (ánodo) en respuesta al campo eléctrico. Pero a pH alcalino el FEO es tan intenso que sobrepasa el efecto de atracción que ejerce el ánodo sobre los péptidos, y estos se mueven junto al amortiguador hacia el cátodo. A pH ácido, los péptidos se cargan positivamente, y el FEO es débil. Como resultado. ambos, los péptidos y el amortiguador fluyen hacia el cátodo. Para asegurar el total control del sistema se hace necesario medir la intensidad del FEO. Un modo de calcularlo es inyectando una muestra que contenga una molécula neutra, como el metanol o la acetona, y midiendo el tiempo que le toma en llegar hasta el detector.Para realizar modalidades de EC que se basan en el enfoque isoeléctrico (EI-EC) y la isotacoforesis (ISF-EC), se necesita eliminar el efecto de FEO. Ello se logra recubriendo la superficie interna del capilar con una sustancia no cargada, o usando capilares de teflón, que no poseen grupos químicos ionizables. También se han usado aditivos como la metilcelulosa. Cinética del flujo:Una consecuencia importante del uso del campo eléctrico como generador del flujo de amortiguador (FEO) en la EC, en lugar de los sistemas hidrodinámicos como las bombas de presión que se usan en los sistemas de cromatografía líquida, es que, en este último caso, el flujo posee un perfil de tipo laminar o parabólico, como consecuencia de la caída de presión que provoca la fricción de las moléculas del fluido con la superficie interna de los empaques y las paredes de las tuberías. Se establece un gradiente de flujo en el que la velocidad es mayor en las capas centrales del fluido y se aproxima a cero en las capas cercanas a la interfase líquido-sólido. Este gradiente de velocidad provoca que la banda o zona donde se mueve cada componente de la muestra se haga más ancha. Por el contrario, en los

sistemas en los que el campo eléctrico es quien provoca el flujo, la fuerza impulsora del FEO se distribuye uniformemente a todo lo largo del capilar, y en consecuencia, no se producen caídas de presión. La velocidad de flujo es uniforme en toda la sección transversal del capilar, excepto en la región muy cercana a la interfase líquido-sólido, donde la velocidad también se aproxima a cero.Eficiencia del sistema:La velocidad de migración, (Vep), se calcula según la ecuación:

νep= µep E = µep Donde V es el voltaje y L es la longitud total del capilar. El tiempo de migración (t) es:

t = =

Durante la migración a través del capilar ocurre dispersión de los picos (σ2) como consecuencia de la difusión molecular. La σ2 se calcula a partir de la ecuación:

σ2= 2Dmt= Donde Dm es el coeficiente de difusión del soluto (cm2/s). El número de platos teóricos (N) está dado por la relación siguiente:

N= Sustituyendo la ecuación de dispersión en la ecuación de cálculo de los platos teóricos, tenemos la siguiente relación:

N = Sustituyamos los valores de la mioglobina de corazón de caballo en la ecuación para el cálculo del número de platos teóricos.MW=13,900, µep=0.65x10-4 cm2/Vs (en amortiguador bicina/TEA 20mM, pH=8.5)Dm= 10-6 cm2/s, V=30,000 Volts.El cálculo de N es igual a 975x103 platos teóricos.

La dispersión en un sistema simple como el que analizamos aquí, se asume como difusión dependiente del tiempo solamente. Esta ecuación indica que las moléculas grandes como las proteínas y los Ac. Nucleicos, con coeficientes de difusión (D) pequeños, generarán el número mayor de platos teóricos. Un modo de incrementar la eficiencia en el sistema de EC es incrementando el voltaje, lo cual disminuye el tiempo de separación. El voltaje límite en los sistemas actuales es de 30kV. No obstante la posibilidad de usar capilares muy cortos para generar campos muy fuertes, el límite práctico para el incremento de la fuerza del campo eléctrico es la generación de calor, que se denomina Òcalentamiento de JouleÓ. El calentamiento de Joule es consecuencia de la resistencia del amortiguador al flujo de la corriente eléctrica. Esto puede ser atenuado mediante el uso de sistemas apropiados de disipación de calor, aunque con límites reales. No obstante las limitaciones que establece la difusión, la EC es un método útil para la separación de moléculas pequeñas, pues, dado que el valor de µep es una función de la relación carga / masa, las moléculas pequeñas tienden a tener mayor valor de µep.La resolución (Rs):La resolución entre dos especies componentes de la muestra se calcula mediante la ecuación:

Rs =

Donde Æ ep es la diferencia de movilidad electroforética entre los dos componentes, µep es el promedio de la movilidad electroforética de ambas especies y N es el número de platos teóricos. Si sustituimos el conteo de platos teóricos en la ecuación de resolución, se obtiene:

Rs= (0.177) Esta relación demuestra que incrementando el voltaje es una medida limitada para incrementar la resolución, pues para lograr el doble de la resolución es necesario incrementar el voltaje al cuádruplo. La clave de cualquier acción para incrementar

la resolución radica en Æµep. La mejor forma de controlar la movilidad de los distintos componentes de la muestra es eligiendo la modalidad de EC apropiada unido a la selección del sistema amortiguador adecuado. La producción de calor durante los ensayos de EC es un fenómeno inevitable, dado el uso de campos eléctricos intensos. Los dos problemas fundamentales que provoca la generación de calor son:

¤ Gradientes de temperatura a través de la sección transversal del capilar.¤ Cambios de la temperatura del sistema en función del tiempo (consecuencia de la disipación ineficiente del calor generado).

Dado la siguiente relación:

= El calor generado es proporcional al cuadrado de la intensidad del campo eléctrico. Por consiguiente, disminuyendo la intensidad del campo o incrementando la longitud del capilar, se disminuye dramáticamente el calor generado.Una consecuencia de la disipación del calor producido es el gradiente de temperatura antes mencionado. Dado que el calor se disipa por difusión, es de esperar que la temperatura en la zona central del capilar sea mayor que en las paredes del capilar. Como la viscosidad disminuye con el aumento de la temperatura, tanto el FEO como la µep se incrementarán para las moléculas que están en las capas de fluido más centrales en el capilar. La consecuencia es un perfil de flujo similar al hidrodinámico, con incremento del ancho de zonas o bandas de migración de las especies. El uso de capilares de muy pequeña sección transversal mejora la situación, al disminuir en el cuadrado del radio del capilar la corriente que pasa a través de él, y al mejorar la disipación del calor generado. La relación entre gradiente térmico (ÆT), radio del capilar (r), la potencia disipada (W) y la conductividad térmica (K) se expresa en la siguiente ecuación:

ÆT= 0.24 El segundo problema, el incremento de la temperatura en función del tiempo, se debe a la diferencia entre la velocidad de disipación y de generación del calor. La

variación de la temperatura resultante modifica la viscosidad del fluido y por consiguiente los factores que determinan el tiempo de migración de las moléculas. La solución del problema depende tanto del uso de capilares estrechos como del acoplamiento de sistemas de enfriamiento, que pueden ser por aire o por líquido. La limitante en el uso de capilares de pequeño diámetro interno es que al disminuir este también disminuye el path length y consecuantemente el límite de detección. La EC, como antes se mencionó, comprende una familia de técnicas que poseen diferencias dramáticas en cuanto a operatividad y características de separación. Estas técnicas son:

a. Electroforesis capilar de zona.b. Enfoque isoeléctrico.c. Electroforesis capilar en gel.d. Isotacoforesis.e. Cromatografía capilar miscelar electrocinética.

Dado los objetivos de este trabajo, nos ocuparemos de la descripción breve de las cuatro primeras técnicas.Electroforesis capilar de zona: Es la forma más simple de EC, y se le conoce también como EC de solución libre. El mecanismo de separación es basado en las diferencias de la relación carga / masa de los componentes de la muestra. Dos factores, la homogeneidad de la solución amortiguadora y la constancia de la intensidad del campo eléctrico son claves para lograr buenos resultados con esta técnica. Luego de la inyección de la muestra y aplicado el campo eléctrico, los componentes de la mezcla se separan en zonas discretas. El parámetro fundamental, µep, puede ser aproximado a partir de la teoría de Debeye-Hunckel-Henry:

µep= donde q es la carga neta, R es el radio de Stokes, y η es la viscosidad del medio.La EC de zona es muy útil para el análisis de mezclas de proteínas, como las proteasas, y péptidos, pues es posible lograr la completa separación de, por ejemplo, proteínas que solo difieren en un residuo de aminoácido. Esto puede ser muy importante en el análisis de los mapas trípticos, haciendo posible la

http://www.ibt.unam.mx/computo/pdfs/met/Electroforesis/Electroforesis/CE/CZE/A1SLIDE5.MOV

detección de modificaciones post-traduccionales y mutaciones puntuales. La EC se ha usado como método micropreparativo para separar los productos de la digestión de proteínas que serán secuenciadas. La cuantificación de mRNA celular, el análisis de los productos de PCR y de mezclas de fragmentos de DNA de doble cadena son otras de las aplicaciones de esta técnica. También es posible monitorear la formación de puentes disulfuros en péptidos durante el plegamiento. Otras aplicaciones son la separación de cationes y aniones inorgánicos y de productos farmacéuticos, siempre que sean especies cargadas. Enfoque isoeléctrico (EI): Como para la técnica antes descrita en geles de poliacrilamida o agarosa, la premisa fundamental de esta técnica es que las moléculas migrarán en el campo eléctrico mientras posean carga. Apenas se haga una especie neutra, detendrá su migración. La separación en el enfoque isoeléctrico se realiza en un gradiente de pH generado por una serie de compuestos zwitterionicos conocidos como anfolitos. El gradiente de pH es menor en el ánodo y mayor en el cátodo. Cuando es aplicado el campo eléctrico, la mezcla de anfolitos se distribuye a lo largo del capilar, los de carga positiva se dirigen hacia el cátodo, y los negativos se mueven hacia el ánodo. El resultado es que el pH disminuye en la región del ánodo y se eleva en la región del cátodo. La migración de cada especie de anfolito cesa cuando alcanza la zona cuyo pH es igual a si pI. Los solutos componentes de la muestra migrarán en el campo eléctrico de igual forma, hasta encontrar la zona de pH igual a su pI, siendo este el mecanismo de separación. Como en el enfoque isoeléctrico para geles, las dos etapas iniciales del experimento son la carga de la muestra y el enfoque, pero en la técnica de formato EC es necesaria una fase adicional, que es la movilización. Luego del enfoque de los componentes de la mezcla, es necesario hacerlos pasar por la ventana óptica del detector, y esto se logra adicionando NaOH/NaCl en el reservorio del cátodo, lo cual provoca la movilización en dirección catódica; en cambio, si se adiciona NaCl en el reservorio del ánodo, se provocará la movilización en dirección anódica. La adición de sales altera el pH y esto provoca la movilización de las especies, ahora nuevamente cargadas, pero manteniendo el patrón de zonas de la etapa de enfoque.El poder de resolución del enfoque isoeléctrico (ÆpI) se describe en la ecuación siguiente:

http://www.ibt.unam.mx/computo/pdfs/met/Electroforesis/Electroforesis/CE/CZE/ProtSequence.pdf

http://www.ibt.unam.mx/computo/pdfs/met/Electroforesis/Electroforesis/CE/IEF/2.101.mov

ÆpI= Donde D es el coeficiente de difusión, E es la fuerza del campo eléctrico y µ es la movilidad electroforética de la proteína. Se ha calculado un poder de resolución teórico de hasta 0.02 unidades de pH. Dado su alto poder de resolución, el enfoque isoeléctrico es un poderoso método de análisis de proteínas, como anticuerpos monoclonales y factores de crecimiento; se ha usado en la identificación y cuantificación de hemoglobinas anormales, y además de para el cálculo del pI de péptidos y proteínas. Es muy útil para separar variantes de inmunoglobulinas, de Hb y para identificar modificaciones post-traduccionales de proteínas recombinantes.Electroforesis capilar en gel (ECG): El gel que se usa en esta técnica es el de poliacrilamida, y se aplican las mismas consideraciones antes mencionadas para la EGPA , incluyendo la EGPA-SDS para las proteínas. Se usan capilares de entre 50 a 100 µm de diámetro interno, y entre 10cm y 1 m, aunque el tiempo de corrida se extiende excesivamente en el caso de los capilares muy largos. La fuerza del campo eléctrico es limitada a valores de 500 V/cm, y la resolución puede incrementarse manipulando la composición del gel. Los geles de agarosa no se han podido usar en esta técnica por la temperatura relativamente elevada que puede alcanzar el sistema, dado el voltaje que se usa. Esta técnica puede usarse para la separación de diferentes muestras de DNA, siempre que se pueda lograr el tamaño de poro adecuado en el gel de poliacrilamida, según el MW de las especies de DNA presentes en la muestra.Isotacoforesis (ITF): Fue la técnica más ampliamente usada en el formato capilar antes de 1981, en la que se empleaban capilares de 250-500 µm, muy gruesos para los estándares actuales. Como para el enfoque isoeléctrico, se requiere eliminar el fenómeno de flujo electro-osmótico, y también usa un sistema heterogéneo de amortiguadores. En esta técnica, el capilar es llenado con una solución de electrolitos cuya movilidad electroforética en las condiciones de corrida es mayor que la de todos los componentes de la muestra, a ese electrolito se le nombra electrolito líder. Luego se inyecta la mezcla de componentes a separar, y finalmente se adiciona al reservorio opuesto una solución de

http://www.ibt.unam.mx/computo/pdfs/met/Electroforesis/Electroforesis/CE/IEF/A2A.mov

http://www.ibt.unam.mx/computo/pdfs/met/Electroforesis/Electroforesis/CE/IEF/A2A.mov

http://www.ibt.unam.mx/computo/pdfs/met/Electroforesis/Electroforesis/CE/GCE/a31.mov

electrolitos cuya movilidad electroforética en las condiciones de corrida es menor que la de todos los componentes de la muestra, a ese electrolito se le nombra electrolito terminador. Al establecerse el campo eléctrico, todas las especies cargadas se moverán en dependencia del signo de su carga (que para los fines de esta técnica ha de ser igual en todos los componentes) y de su movilidad electroforética. Por la diferencia en este último parámetro entre el electrolito líder y el electrolito terminador, se forma una brecha que es ocupada por los componentes de la muestra. La separación de los analitos ocurre en función de su movilidad electroforética, distribuyéndose estos en zonas o bandas estables bien definidas, lo que resulta en una muy eficiente separación. Dado que las bandas de movimiento de dos compuestos contiguos están en contacto (pues de separarse estas se formaría una zona de no-conductividad, y cesaría en campo eléctrico), se han usado compuestos con movilidad electroforética intermedia entre dos componentes que se desea separar totalmente. Estos componentes son nombrados espaciadores, y poseen como característica adicional no absorber en la longitud de onda donde se realizará la lectura de los componentes de interés. Referencias:

¤ Introduction to capillary electrophoresis. (1999) Beckman.¤ Lux, J. A.; Yin, H. F.; Schomburg, G., (1990) Construction, evaluation and analytical operation of a modular capillary electrophoresis instrument, Chromatographia 30, 7-15

■ Lux, J. A.; Yin, H. F.; Schomburg, G., J. (1990) A simple method for the production of gel-filled capillaries for capillary gel electrophoresis, High Resolut.Chromatogr. 13, 436-437

■ Paulus, A.; Ohms, J. I., (1990) Analysis of oligonucleotides by capillary gel electrophoresis, J. Chromatogr. 507, 113-123¤ McCormick, R. M., (1988) Capillary zone electrophoretic separation of peptides and proteins using low pH buffers in modified silica capillaries,., Anal. Chem.60, 2322-2328

■ Compton, S.W.; Brownlee, R. G (1988) Capillary electrophoresis,., BioTechniques 6, 432-440

■ Cohen, A. S.; Paulus, A.; Karger, B. L., (1987) High-performance capillary

electrophoresis using open tubes and gels, Chromatographia 24, 15-24 ■ Altria, K. D.; Simpson, C. F., (1988) The effect of electrolyte chain length

on electroendosmotic flow in high voltage capillary zone electrophoresis, Anal. Proc. (London) 25, 85.

Sub-Cell® GTAgarose Gel

ElectrophoresisSystems

Instruction Manual

Catalog Numbers170-4401 to 170-4406170-4481 to 170-4486

For Technical Service Call Your Local Bio-Rad Off ice or in the U.S. Call 1-800-4BIORAD (1-800-424-6723)

WarrantyBio-Rad Laboratories warrants the Sub-Cell GT, Wide Mini-Sub® cell GT, and Mini-Sub cell GT

electrophoresis systems against defects in materials and workmanship for 1 year. If any defects occurin the instrument during this warranty period, Bio-Rad Laboratories will repair or replace the defectiveparts free. The following defects, however, are specifically excluded:

1. Defects caused by improper operation.

2. Repair or modification done by anyone other than Bio-Rad Laboratories or an authorized agent.

3. Use of fittings or other spare parts supplied by anyone other than Bio-Rad Laboratories.

4. Damage caused by accident or misuse.

5. Damage caused by disaster.

6. Corrosion due to use of improper solvent or sample.

This warranty does not apply to parts listed below:

1. Platinum Electrode Wires

To insure the best performance from the Sub-Cell GT electrophoresis systems, become fullyacquainted with these operating instructions before use. Bio-Rad recommends that you first read theseinstructions carefully. Assemble and disassemble the unit completely without casting a gel. After thesepreliminary steps, you should be ready to cast and run a gel.

Bio-Rad also recommends that all Sub-Cell GT system components and accessories be inspected fordamage, cleaned as recommended in this manual, and rinsed thoroughly with distilled water before use.

Record the following for your records:

For any inquiry or request for repair service, contact Bio-Rad Laboratories after confirming themodel and serial number of your instrument.

Model

Catalog No.

Date of Delivery

Warranty Period

Serial No.

Invoice No.

Purchase Order No.

Table of ContentsPage

Section 1 General Information....................................................................................11.1 Introduction ................................................................................................................11.2 Safety ..........................................................................................................................11.3 System Components...................................................................................................21.4 Specifications .............................................................................................................4

Section 2 Operating Instructions ................................................................................42.1 DNA Gel Preparation.................................................................................................42.2 Casting Agarose Gels.................................................................................................62.3 Electrophoresis ...........................................................................................................82.4 Nucleic Acid Staining and Visualization...................................................................92.5 Note on Blotting .......................................................................................................10

Section 3 Gel and Electrophoresis Reagent Preparation .......................................10

Section 4 Care and Maintenance...............................................................................114.1 Cleaning Sub-Cell GT Components ........................................................................124.2 Compatible Cleaning Agents ...................................................................................124.3 Maintenance Schedule .............................................................................................124.4 Electrode Replacement.............................................................................................134.5 RNase Decontamination ..........................................................................................14

Section 5 Troubleshooting..........................................................................................14

Section 6 Product Information..................................................................................156.1 Sub-Cell GT Systems...............................................................................................156.2 Sub-Cell GT System Accessories............................................................................166.3 Related Bio-Rad Products........................................................................................18

Section 7 References ...................................................................................................21

1

Section 1General Information

1.1 IntroductionThe Sub-Cell GT instruments (basic Sub-Cell GT cell, Wide Mini-Sub® cell GT, and Mini-

Sub cell GT) comprise a comprehensive and flexible gel electrophoresis system that effective-ly separates nucleic acids using submerged agarose gels. Submarine agarose gels are easy tocast and readily dissipate heat. These gels allow sample underlaying and prevent electrical field dis-continuities caused by wicks or sample well dehydration. Agarose gels are ideal for the separation ofDNA restriction digestions, Polymerase Chain Reaction (PCR*)-amplified fragments, and genomicDNA and RNA prior to Southern or northern blotting. If operated correctly, agarose gel submarineelectrophoresis can effectively separate nucleic acids from 20 base pairs to 20 kilobase pairs in length.

The Sub-Cell GT systems are designed for years of reproducible and rigorous use. Theserugged systems incorporate many features that make casting and running agarose gels simpleand efficient. The gel caster provides tape-free gel casting in trays. Gels can also be cast in theGT bases using specially designed wedge gates. Replaceable electrode cassettes provide asimple way to replace electrode wires. A comprehensive assortment of base and tray sizes,including a variety of preparative, analytical, and multichannel pipet compatible combs, makesthese systems ideal for any agarose gel application.

Note:This manual contains instructions for the Sub-Cell GT electrophoresis systems only.Prior to the release of the Sub-Cell GT systems, Bio-Rad supplied similar agarose gel elec-trophoresis cells: the original Sub-Cell DNA electrophoresis cell, Wide Mini-Sub cell, and Mini-Sub cell systems. This manual does not provide information on these earlier versions. Contactyour local Bio-Rad representative for information concerning the original Sub-Cell systems.

Definition of Symbols

Caution, risk of electrical shock Caution (refer to accompanying documents)

1.2 SafetyThe Sub-Cell GT electrophoresis systems are designed for maximum user safety. The buffer

chambers are made of 3/16 inch (.476 cm) thick injection-molded acrylic to create a leak-freeelectrophoresis environment. The safety lids surround the buffer chamber to protect the user fromexposure to electrical currents. All Sub-Cell GT systems were designed for indoor use only.

Before use, inspect the GT base for cracks or chips, which may allow the buffer to leakfrom the base and cause a potential electrical hazard. Additionally, inspect all electrical cables,banana jacks, and plugs for loose connections, cracks, breaks, or corrosion. Do not use any partthat is cracked, charred, or corroded. These parts may also cause a potential electrical hazard.Contact your local Bio-Rad representative before using a part that may be considered hazardous.

During electrophoresis, inspect the base and workbench for any signs of buffer leakage.If leaking buffer is detected, disconnect the power to the cell immediately and contact yourlocal Bio-Rad representative.

Power to Sub-Cell GT units is supplied by an external DC voltage power supply. Thispower supply must be ground isolated in such a way that the DC voltage output floats with

!

!

respect to ground. All of Bio-Rad’s power supplies meet this important safety requirement.The recommended power supply for this apparatus is the PowerPac 300 power supply. ThePowerPac 300 power supply contains safety features such as no load, overload, rapid resis-tance change, and ground leak detection capabilities. The maximum specified operating param-eters for the Sub-Cell GT systems are given in Table 1.1.

Table 1.1 Sub-Cell GT systems operating parametersSub-Cell Wide Mini-Sub Mini-Sub GT Cell Cell GT Cell GT

Maximum voltage limit 200 VDC 150 VDC 150 VDCMaximum power limit 40 Watts 45 Watts 10 WattsMaximum Buffer temperature 40 ˚C 40 ˚C 40 ˚C

Current to the cell, provided from the external power supply, enters the unit through thelid assembly, providing a safety interlock. Current to the cell is broken when the lid is removed.Do not attempt to circumvent this safety interlock, and always turn the power supply offbefore removing the lid or when working with the cell.

Important: These Bio-Rad instruments are certified to meet IEC 1010-1** safety stan-dards. IEC-certified products are safe to use when operated in accordance with the instructionmanual. This instrument should not be modified in any way. Alteration of this instrument will:

• Void the manufacturer’s warranty• Void the IEC 1010-1 safety certification• Create a potential safety hazard

IEC 1010-1 certification applies to equipment designed to be safe between the operating temperatures of 4 °C and 40 °C and altitudes up to 2,000 meters. Instruments are also safe at a max-imum relative humidity of 80% for temperatures up to 31 °C decreasing linearly to 50 % at 40 °C. Bio-Rad is not responsible for any injury or damage caused by the use of this instru-ment for purposes other than those for which it is intended, or by modifications of the instrumentnot performed by Bio-Rad or an authorized agent. No user-serviceable parts are contained in thisapparatus. To insure electrical safety, do not attempt to service this apparatus.

1.3 System ComponentsEach of the Sub-Cell GT systems comes with the components listed in Table 1.2 (see

Figure 1.1 for part description). Check your instrument to be sure all items are present. Noteany damage to the unit which may have occurred during shipping. Notify Bio-RadLaboratories if any items are missing or damaged.

Table 1.2 Sub-Cell GT System ComponentsSub-Cell GT Wide Mini-Sub Cell Mini-Sub Cell

System GT System GT SystemItem Quantity Quantity QuantityGT Base (buffer chamber) 1 1 1Gel Casting Gates 2 2 2Safety Lid and Cables 1 1 1UVTP Gel Tray 1 1 1Fixed Position Comb 2 2 2

(15 well, 1.5 mm thick) (15 well, 1.5 mm thick) (8 well, 1.5 mm thick)(20 well, 1.5 mm thick) (20 well, 1.5 mm thick) (15 well, 1.5 mm thick)

Leveling Bubble 1 1 1Gel Caster (optional) 1 1 1Instruction Manual 1 1 1

2

Fig.1.1. Sub-Cell GT parts.

3

Safety lid

Electrical cables

Electrical leads

Gel casting gates

Fixed heightcomb

GT Base

Safety Lid removal peg

Leveling feet

Combslots

Banana plug/Electrode wireassembly

UVTP gel tray

Fluorescent ruler

Fixed height comb

Leveling feet

Leveling feet

UVTP gel tray

Gel caster

Cam lever

Fluorescent ruler

Leveling bubble

1.4 Specifications

Sub-Cell GT Wide Mini-Sub Cell Mini-Sub Cell System GT System GT System

GT base footprint (L x W x H) 42 x 19.5 x 10 cm 26 x 20 x 7.5 cm 26 x 12 x 6.5 cmGT base buffer volume✝ 1,500–2,000 ml 650–900 ml 265–320 mlGT base gel size 15 x 15 cm 15 x 7 cm 7 x 7 cmGel tray sizes 15 x 10 cm 15 x 7 cm 7 x 7 cm

15 x 15 cm 15 x 10 cm 7 x 10 cm15 x 20 cm15 x 25 cm

ConstructionGT base Molded clear plasticGel casting gates AluminumSafety lid Molded clear plasticBanana plug/electrode cassette Molded polycarbonate Banana plugs Gold-plated brass, 4.4 cm lengthElectrodes Platinum, 0.25 mm diameterElectrical cables Dual, 20 AWG, tinned copper wire cable

Flame-retardant polyurethane insulation jacketElectrical leads Nickel silverGel tray UV-transparent acrylic plastic (UVTP)Combs Molded plastic and machined acrylicGel casting device Polycarbonate

0.64 cm silicon foam

Section 2Operating Instructions

Note: See Section 3, Gel and Electrophoresis Reagent Preparation, for information onthe preparation of RNA gels. See References 1 and 2 for more information on DNA andRNA electrophoresis.

2.1 DNA Gel PreparationDNA agarose gels can be used to separate and visualize DNA of various sizes. Before cast-

ing an agarose gel, consult Table 2.1 to determine the appropriate percent agarose gel to use,based on the size of DNA to be separated.

Procedure

1. Determine the amount of agarose (grams) required to make the desired agarose gel con-centration and volume. Use Tables 2.1 and 2.2 as a guide for agarose concentration andgel volume requirements.

Example: For a 1% agarose gel, add 1 gram of agarose to 100 ml of 1x electrophoresis buffer.

4

✝ GT base buffer volumes will vary depending on the size and thickness of the gel used.

Table 2.1 Gel Concentration Required for DNA Separation 1-2

Gel Concentration (%) DNA Size (Kb)0.50 1–300.75 0.8–121.00 0.5–101.25 0.4–71.50 0.2–32–5* 0.01–0.5

* Sieving agarose such as AmpliSize® agarose

Table 2.2 Gel Volume RequirementsGel Size 0.25 cm thick 0.5 cm thick 0.75 cm thick 1.0 cm thickBase7 x 7 cm 10 ml 20 ml 30 ml 40 ml

15 x 7 cm 20 ml 40 ml 60 ml 80 ml15 x 15 cm 50 ml 100 ml 150 ml 200 mlTray7 x 7 cm 10 ml 20 ml 30 ml 40 ml7 x 10 cm 15 ml 30 ml 45 ml 60 ml

15 x 7 cm 20 ml 40 ml 60 ml 80 ml15 x 10 cm 30 ml 60 ml 90 ml 120 ml15 x 15 cm 50 ml 100 ml 150 ml 200 ml15 x 20 cm 70 ml 140 ml 210 ml 280 ml15 x 25 cm 90 ml 180 ml 270 ml 360 ml

2. Add the agarose to a suitable container (e.g., 250 ml Erlenmeyer flask, Wheaton bottle, etc.).Add the appropriate amount of 1x electrophoresis buffer (see Section 3, Gel andElectrophoresis Reagent Preparation, for electrophoresis buffer preparation) and swirl to sus-pend the agarose powder in the buffer. If using an Erlenmeyer flask, invert a 25 ml Erlenmeyerflask into the open end of the 250 ml Erlenmeyer flask containing the agarose. The smallflask acts as a reflux chamber, allowing long or vigorous boiling without much evaporation.

Note: A mark can be put on the lower flask at the same level as the liquid. If evaporationoccurs, water can be added to bring the liquid back to the original starting level.

3. The agarose can be melted by boiling on a magnetic hot plate (Step 4a) or in a microwaveoven (Step 4b).

Caution: Always wear protective gloves, goggles, and a lab coat while preparing and cast-ing agarose gels. The vessels containing hot agarose can cause severe burns if allowed tocontact skin. Additionally, molten agarose can boil over when swirled.

Magnetic Hot Plate Method

4a. Add a stir bar to the undissolved agarose solution. Heat the solution to boiling while stirringon a magnetic hot plate. Bubbles or foam should disrupt before rising to the neck of the flask.

Microwave Oven Method

4b. Place the gel solution into the microwave. Using a low to medium setting, set the timerfor a minimum of 5 minutes, stopping the microwave oven every 30 seconds and swirlingthe flask gently to suspend the undissolved agarose. This technique is the fastest andsafest way to dissolve agarose.

5. Boil and swirl the solution until all of the small translucent agarose particles are dissolved.With the small flask still in place, set aside to cool to 60 °C before pouring.

5

2.2 Casting Agarose Gel SlabsThere are several ways to cast agarose submarine gels using the Sub-Cell GT systems.

Gels may be cast with or without a UV-transparent plastic (UVTP) tray directly on the stageof the Sub-Cell GT bases using the gel casting gates. Gels may also be cast on the removableUVTP trays with the aid of the gel caster or with standard laboratory tape.

Casting gels on the base stages

1. Level the Sub-Cell base using the leveling bubble provided.

2. Slide the gel casting gates into the slots at opposite ends of the gel stage.

3. Place the comb(s) into the appropriate slot(s) of the base so that the sample wells are near thecathode (black). DNA samples will migrate toward the anode (red) during electrophoresis.

4. Prepare the desired concentration and amount of agarose in 1x electrophoresis buffer (seeSection 2.1). When the agarose solution has cooled to 50–60 ˚C pour the molten agarosebetween the gates.

Warning: Hot agarose (>60 ˚C) may cause the plastic in the cell to warp or craze and willdecrease the lifetime of the Sub-Cell base. Warping may also result in sample wells ofuneven depth.

5. Allow 20–40 minutes for the gel to solidify at room temperature.

6. Carefully remove the comb from the solidified gel. Remove the gel casting gates.

7. Submerge the gel beneath 2 to 6 mm of 1x electrophoresis buffer (see Section 3, Gel andElectrophoresis Reagent Preparation). Use greater depth overlay (more buffer) withincreasing voltages to avoid pH and heat effects.

Casting Gels on the Base Stage With UVTP Tray

1. Level the cell using the leveling bubble provided.

2. Place the UVTP tray on the cell base stage.

Note: The Mini-Sub cell GT requires the 7 x 7 cm UVTP tray for casting in the base.The Wide-Mini-Sub cell GT requires the 15 x 7 cm UVTP tray and the Sub-Cell GT sys-tem requires the 15 x 15 cm UVTP tray for casting in the base.

3. Slide the gel casting gates into the slots at opposite ends of the base stage. Insure the gatesare evenly seated in the slots and the gates uniformly contact all edges of the UVTP tray. Theweight of the gates provides a tight seal to prevent any leakage problems during gel casting.

4. Place the comb(s) into the appropriate slot(s) of the trays so that the sample wells are near thecathode (black). DNA samples will migrate toward the anode (red) during electrophoresis.

5. Prepare the desired concentration and amount of agarose in 1x electrophoresis buffer (seeSection 2.1). When the agarose solution has cooled to 50-60 °C, pour the molten agarosebetween the gates.

Warning: Hot agarose (>60 °C) may cause the tray to warp or craze and will decrease thelifetime of the tray. Warping may also result in sample wells of uneven depth.

6. Allow 20-40 minutes for the gel to solidify at room temperature.


8. Submerge the gel beneath 2 to 6 mm of 1x electrophoresis buffer (see Section 3, Gel andElectrophoresis Reagent Preparation). Use greater depth overlay (more buffer) withincreasing voltages to prevent pH and heat effects.

6

Removable tray (UVTP) gel casting using a Gel Caster or Mini-Gel Caster

1. Level the Gel Caster or Mini-Gel Caster using the leveling feet in the gel caster and the leveling bub-ble provided.

2. Disengage and slide the movable wall to the open end of the Gel Caster or Mini-Gel Casterby turning and lifting the cam peg upward.

Note: If casting more than one gel with the Gel Caster, add the removable gel castingwall to the gel caster. The removable wall will allow casting of two 15 x 10 cm trays,four 7 x10 cm trays or one 15 x 10 cm and one 15 x15 cm trays.

3. Place the open edge of the UVTP tray against the fixed wall of the Gel Caster or Mini-GelCaster.

4. Slide the movable wall against the edge of the UVTP tray (Figure 2.1).

5. To seal the open tray ends, engage the cam peg by turning and pressing downward simul-taneously.

6. When the cam peg has dropped into the appropriate slot, turn the peg in either directionuntil resistance is felt. This action seals the edges of the tray for casting.

7. Place the comb(s) into the appropriate slot(s) of the tray.

Fig. 2.1. Sealing the UVTP tray for gel casting.

8. Prepare the desired concentration and amount of agarose in 1x electrophoresis buffer (seeSection 2.1). When the agarose solution has cooled to 50–60 ˚C pour the molten agarosebetween the gates.

Warning: Hot agarose (>60 ˚C) may cause the tray to warp or craze and will decrease thelifetime of the tray. Warping may also result in sample wells of uneven depth.

7

Movable wall of gel caster

Fixed wall of gel caster

Engage and seal (press down and rotate)

Slide forward

Lift cam lever up


10. Carefully remove the comb from the solidified gel.

11. Disengage the cam peg by turning and lifting upward. Slide the movable wall away fromthe tray. Remove the tray from the Gel Caster or Mini-Gel Caster.

Note: While the gel is solidifying, a light seal is formed between the gasket and the gel(especially for low percentage agarose gels [<0.8%]). Before moving the wall away fromthe tray, carefully lift the tray on one side to release the seal or use a spatula to break theseal between the agarose and gasket.

12. Place the tray onto the leveled Sub-Cell base so that the sample wells are near the cathode(black). DNA samples will migrate toward the anode (red) during electrophoresis.


Removable tray (UVTP) gel casting using tape

1. Seal the ends of the UVTP gel tray securely with strips of standard laboratory tape. Pressthe tape firmly to the edges of the gel tray to form a fluid-tight seal.

2. Level the gel tray on a leveling table or workbench using the leveling bubble provided withthe instrument.

3. Prepare the desired concentration and amount of agarose in 1x electrophoresis buffer (seeSection 2.1). When the agarose solution has cooled to 50–60 ˚C pour the molten agaroseinto the gel tray.

Warning: Hot agarose (>60 ̊ C) may cause the tray to warp or craze and will decrease thelifetime of the tray. Warping may also result in sample wells of uneven depth.



6. Remove the tape from the edges of the gel tray.

7. Place the tray onto the leveled Sub-Cell base so that the sample wells are near the cath-ode (black). DNA samples will migrate toward the anode (red) during electrophoresis.

8. Submerge the gel under 2 to 6 mm buffer (see Section 3, Gel and Electrophoresis ReagentPreparation). Use greater depth overlay (more buffer) with increasing voltages to avoidpH and heat effects.

2.3 ElectrophoresisAfter the agarose gel has solidified, sample loading and electrophoresis can begin. Agarose

gels can be run in many different types of electrophoresis buffers. Nucleic acid agarose gelelectrophoresis is usually conducted with either Tris-Acetate-EDTA (TAE) buffer or Tris-Borate-EDTA (TBE) buffer. While TAE buffer provides faster electrophoretic migration oflinear DNA and better resolution of supercoiled DNA, TBE buffers have a stronger bufferingcapacity for longer or higher voltage electrophoresis runs. Bio-Rad offers premixed 50x TAEand 10x TBE buffers, as well as individual buffer reagents for use with the Sub-Cell GT systems.

1. Prepare samples for gel loading. The maximum sample loading volumes for Bio-Rad’scombs are listed in Section 6.2. Loading volume is dependent upon the type of comb used(i.e., well thickness and length) and thickness of the gel.

8

2. When loading volume is determined, add standard nucleic acid sample loading dye to a final1x concentration to make samples dense for underlaying into sample wells (see Section 3,Gel and Electrophoresis Reagent Preparation, for sample loading dye preparation).

3. Load the samples into the wells using standard pipets. Multichannel pipets can be used forloading samples only with the Bio-Rad MP combs (see Section 6.2).

Note: Sample wells are often difficult to see. Well visualization can be enhanced by placingblack paper or tape under the base or trays where comb placement or well formation is common.

4. Place the lid on the DNA cell carefully. Do not disturb the samples. The Sub-Cell GTsystem lids attach to the base in only one orientation. To attach the lid correctly, match thered and black banana jacks on the lid with the red and black banana plugs of the base.

5. Power requirements vary depending on gel thickness, length and concentration, and type ofelectrophoresis buffer used. Refer to Tables 2.3 and 2.4 for relative sample migration ratesand for DNA size migration with sample loading dyes for the different Sub-Cell GT systems.

Note: Buffer recirculation is not required for most standard DNA and RNA agarose gelelectrophoresis. If buffer recirculation is required, simply turn off the power supply,remove the safety lid, and mix the running buffer as desired. After the buffer has beenmixed, reconnect the safety lid and continue electrophoresis.

Table 2.3 Relative Sample Migration Rates *

Bromophenol Blue Cell Type Voltage Migration RateSub-Cell GT cell, 15 x 15 cm gel 75 V 3.0 cm/hrWide Mini-Sub cell GT, 15 x 10 cm gel 75 V 4.5 cm/hrMini-Sub cell GT, 7 x 10 cm gel 75 V 4.5 cm/hr* These sample migration rates were determined based on a 0.5 cm thick 1.0% agarose gel using Bio-Rad’s Molecular Biology

Certified Agarose in 1x TAE electrophoresis buffer (diluted from Bio-Rad’s Premixed 50x TAE Buffer). Migration rates willvary depending on the voltage, current, and type of agarose or buffer used.

Table 2.4 DNA size migration with sample loading dyesAgarose Concentration (%) Xylene Cyanol Bromophenol Blue

0.5–1.5 4–5 Kb 400–500 bp2.0–3.0** 750 bp 100 bp4.0–5.0** 125 bp 25 bp

** Sieving agarose such as AmpliSize agarose.

2.4 Nucleic Acid Staining and VisualizationGels can be removed from the base or gel tray for nucleic acid staining. The gel can also

remain on the UVTP gel tray for staining.

Ethidium Bromide Staining Procedure

1. Place the gel into the appropriate volume of 0.5 µg/ml ethidium bromide (EtBr) stain for15–30 minutes. Use enough staining solution to cover the entire gel.

Caution: Ethidium bromide is a suspected carcinogen and should be handled with extreme care.Always wear gloves, eye glasses, and a laboratory coat. Dispose of used EtBr solutions and gelsappropriately (Review EtBr Material Safety Data Sheet [MSDS] for proper disposal methods).

2. Destain the gel for 10–30 minutes in dH2O using the same volume used for staining.

Note: Ethidium Bromide can be removed from the DNA with extended destaining. Thiswill cause lower sensitivity of detection. However, insufficient destaining will createhigher background fluorescence.

9

3. Rinse the gel briefly with dH2O to remove any residual staining solution.

4. Place the gel on a UV transilluminator for nucleic acid visualization and analysis. DNA-Ethidium Bromide complexes may be illuminated with UV light of 254, 302, or 366 nm.Sensitivity decreases with illumination at higher wavelengths. However, nicking of DNAwill increase below 302 nm. Table 2.5 gives the percentage of transmittance of UV lightthrough 1/4” (.64 cm) UV-transparent plastic.

Note: Nucleic acids in the gel can be visualized through the UVTP trays. If a UVTP trayis not used, place household plastic wrap between the UV transilluminator and the gel toavoid contaminating the transilluminator with nucleic acids or EtBr.

Table 2.5 Percent UV Transmittance through 1/4” (.64 cm) UVTransparent Plastic (UVTP)

ApproximateWavelength (nm) % Transmittance

254 0302 80360 90

5. Photograph the gel using standard cameras and film (e.g., Bio-Rad’s Standard PolaroidGel Documentation System) or with CCD-based digitized image analysis systems (e.g.,Gel Doc™ 1000 UV fluorescent gel documentation system). Gels are generally pho-tographed with a yellow, orange, or red interference filter. Red filters generally give thecleanest background. Bio-Rad offers a full-line of standard photography and CCD-basedimaging systems for nucleic acid gel analysis.

2.5 Note on BlottingNucleic acids within the gel can be transferred to membranes using the techniques of

Southern and Northern blotting. It is beyond the scope of this instruction manual to includeblotting procedures. Consult references 1 and 2 for blotting techniques. Bio-Rad offers a fullline of nitrocellulose and positively charged nylon membranes, as well as vacuum and elec-trophoretic blotting apparatus for Southern and Northern blotting.

Section 3Gel and Electrophoresis Reagent Preparation

RNA Agarose Formaldehyde Gels

For 100 ml of a 1% agarose formaldehyde gel prepare as follows:62 ml of 1.6% melted agarose20 ml 5x MOPS electrophoresis buffer (1x final concentration)18 ml 12.3 M (37.5%) formaldehyde (2.2 M final concentration)

Caution: Formaldehyde solutions and formaldehyde vapors are toxic. When handling solu-tions or gels that contain formaldehyde use a chemical hood. Always wear gloves, eye glass-es, and a laboratory coat while using formaldehyde. See the MSDS for safety information.

Nucleic Acid Electrophoresis Buffers 1-2

DNA agarose gel electrophoresis is usually conducted with either Tris-Acetate-EDTA(TAE) or Tris-Boric Acid-EDTA (TBE). While TAE provides faster electrophoretic migra-tion of linear DNA and better resolution of supercoiled DNA, TBE buffers have a strongerbuffering capacity for longer or higher voltage electrophoresis runs. Bio-Rad offers premixed50x TAE and 10x TBE buffers for use with the Sub-Cell GT systems. RNA formaldehyde gelsrequire a MOPS [3-(N-morpholino)-propanesulfonic acid] electrophoresis buffer.

10

1x Tris-Acetate-EDTA (TAE)—40 mM tris (pH 7.6), 20 mM acetic acid, and 1 mM EDTA.

50x Stock (1 liter)—dissolve in 600 ml distilled water:242 g tris base (FW = 121)57.1 ml glacial acetic acid100 ml 0.5 M EDTA (pH 8.0).Fill to a final volume of 1 liter with distilled water.

1x Tris-Boric Acid-EDTA (TBE) —89 mM tris (pH 7.6), 89 mM boric acid, 2 mM EDTA

10x Stock (1 liter)—dissolve in 600 ml distilled water:108 g tris base (FW = 121)55 g boric acid (FW = 61.8)40 ml 0.5 M EDTA (pH 8.0)Fill to a final volume of 1 liter with distilled water.

1x MOPS Buffer (RNA Gels)—0.02 M MOPS [3-(N-morpholino)-propanesulfonicacid] (pH 7.0), 8 mM sodium acetate, 1 mM EDTA (pH 8.0)

5x Stock (1 liter)—dissolve in 600 ml DEPC-treated distilled water:20.6 g MOPS 13.3 ml 3 M sodium acetate (DEPC treated), pH 7.410 ml 0.5 M EDTA (DEPC-treated), pH 8.0Fill to a final volume of 1 liter with DEPC-treated distilled water.

Caution: DEPC is a suspected carcinogen. Always wear gloves, eye glasses, and a lab-oratory coat. Use caution when handling DEPC containing solutions. Consult the DEPCMSDS (Material Safety Data Sheet) for more information.

DNA and RNA Sample Loading Dye 1-2

A convenient 10x sample buffer stock consists of 50% glycerol, 0.25% bromophenol blue, and0.25% xylene cyanole FF in 1x TAE buffer. Only 1–10 ml of the 10x loading dye should be prepared.

RNA Sample Preparation 1-2

Prior to loading RNA onto an agarose formaldehyde gel prepare each RNA sample as follows:6 µl RNA in DEPC-treated water10 µl 5x MOPS buffer (final concentration 1.67x)9 µl 12.3 M formaldehyde (final concentration 3.7 M)25 µl formamide (final concentration 50% v/v)

Caution: Formamide is a teratogen. Always wear gloves, eye glasses, and a laboratory coat. Usecaution when handling formamide. Consult the formamide MSDS for more information.

Ethidium Bromide Solution

Add 10 mg of EtBr to 1 ml distilled water. Bio-Rad offers EtBr solutions (10 mg/ml).

Section 4Care and Maintenance

4.1 Cleaning Sub-Cell GT Components1. All Sub-Cell GT system parts should be washed with a mild detergent solution in warm water.

Note: Be careful not to snag or break the electrode wire in the GT base while cleaning.

2. Rinse all parts thoroughly with warm water or distilled water and air dry, if possible.

11

4.2 Compatible Cleaning AgentsChemically compatible cleaners must be used to insure long life of parts. These include:

• Aqueous solutions of soaps and mild detergents:Bio-Rad Cleaning Concentrate (catalog number 161-0722)Dishwashing liquid

• Organic Solvents:HexaneAliphatic hydrocarbons

Do not leave plastic parts to soak in detergents more than 30 minutes. A short detergentrinse typically is all that is required.

Caution: Do not use the following chemicals to clean Sub-Cell GT parts. Exposure tothese chemicals may cause the plastic parts to crack, craze, etch, or warp.

• Chlorinated HydrocarbonsCarbon tetrachlorideChloroform

• Aromatic HydrocarbonsBenzenePhenolTolueneMethyl ethyl ketoneAcetone

• AlcoholsMethanolEthanolIsopropyl alcohol

Do not use abrasive or highly alkaline cleaners on Sub-Cell GT parts.

Do not expose Sub-Cell GT parts to temperatures >60 ˚C. Do not sterilize Sub-Cell GTparts by autoclaving or dry heat.

4.3 Maintenance ScheduleItem Look For Frequency Action

All parts Dried salts, agarose, Each use Clean parts as described ingrease, and dirt Section 4.1

Electrical cables Breaks or fraying Each use Replace cables

Trays Chips or cracks Each use Replace tray

Electrode wires Breaks Each use See Section 4.4 (Electrode Cassette Replacement)

Cable connections Looseness Weekly Replace banana jacks or(banana jacks and banana plug holdersplugs)

GT base Crazing, cracks, Monthly Replace GT base or or leaks banana plug holder o-ring

12

4.4 Electrode ReplacementThe Sub-Cell GT systems allow easy replacement of broken electrode wires by remov-

ing the banana plug/electrode wire assembly and ordering a new assembly from Bio-Rad(Figure 4.1). Order the new assembly using the part description and catalog numbers listed inSection 6, Product Information.

1. Remove the thumb screw and rubber gasket from the banana plug chamber of the GTbase to release the banana plug/electrode wire assembly. Do not discard this thumb screwor rubber gasket (keep these parts with the GT base).

2. Remove the broken wire assembly by lifting upward on the banana plug. Discard the bro-ken assembly.

3. Insert the new assembly into the banana plug chamber of the GT base. Make sure theelectrode wire guard guides are properly seated into the electrode wire guard slots in thebottom of the GT base.

4. Replace and tighten the thumb screw and rubber gasket to secure the assembly in thebase and to form a leak-free seal in the banana plug holder chamber.

Note:Test for buffer leakage, by filling the base with water and checking for leakage of waterthrough the banana plug chamber of the base. If leakage occurs, tighten the thumb screw.

Fig. 4.1. Removal of banana plug/electrode wire assembly.

Banana plug

O-ring

Electrode wire

Banana plug/Electrode wireassembly

Banana plugchamber

Thumb screw

Rubber gasket

13

4.5 RNase DecontaminationSub-Cell GT parts can be cleaned with a mild detergent and treated for 10 minutes with

3% hydrogen peroxide (H2O2), and then rinsed with 0.1% DEPC- (diethyl pyrocarbonate)treated distilled water, to eliminate RNases prior to using the Sub-Cell GT systems for RNAgels.1-2 Consult references 1-2 for other suggestions regarding the use of DEPC in RNasedecontamination.

Caution: DEPC is a suspected carcinogen. Always wear gloves, eye glasses, and a lab-oratory coat. Use caution when handling DEPC-containing solutions. Consult the DEPCMSDS for more information.

Do not attempt to RNase decontaminate Sub-Cell GT parts using extreme dry heat.

Note: Several commercial products are available for eliminating RNase contamination.RNaseZAP™ (Ambion) is a safe, simple, and effective method that if used properly doesnot craze or fog the Sub-Cell GT parts. See manufacturer’s instructions for proper use.

Section 5Troubleshooting

Symptoms Cause SolutionsSlanted lanes (bands) Gel not fully solidified Let gel solidify for at least 30–45 minutes.

Comb warped or at an Check alignment of comb.angle

Curved line or distortion Bubbles in sample Remove bubbles prior to electrophoresis.of lanes (bands) wells

Differential relative Sample spilled out of Samples should have proper density. mobilities wells Apply carefully.

Unit not leveled Level unit. Place on steady work bench.

Curved bands, smiles Sample overload Reduce load.

Ragged bands Sample density incorrect See sample application instructions.

Sample well deformed Carefully remove comb, especially from soft gels. Be sure gel has solidified. Cooling soft gels aids in comb removal.

Excessive power or Reduce voltage. See electrophoresis heating instructions.

Band smearing and Agarose has improper Consult Bio-Rad about agarose.streaking endosmosis (mr)

Salt concentration in Reduce salt concentration to ≤ 0.1 M.sample too high

Excessive power and Reduce voltage. See electrophoresis heating instructions.

Sample spilled out of Apply sample carefully. Increase gel well thickness for large sample volumes.

Adjust comb height.

Incomplete digest, Heat sample. Check enzyme activity. nuclease contamination, Digest sample further.bad enzyme

14

Symptoms Cause SolutionsSample wells cast Comb should be placed 1 to 2 mm above through the gel. Sample the base of the running surface.leaks along bottom of running surface.

Sample overload Dilute sample.

Bands sharp but too Too high gel Lower gel percentage.few bands seen percentage

Incomplete digest Check enzyme activity, digest further.

High MW bands Gel percentage Increase gel percentage.sharp; Low MW too low Switch to polyacrylamide.bands smeared

Gels crack Too high voltage Reduce voltage. Run gel at lower gradient, especially temperature.with low melting temperature agarose orlow gel strength gels

Section 6Product Information

6.1 Sub-Cell GT Systems

Catalog Number Product Description

170-4401 Sub-Cell GT System, with 15 x 10 cm tray




170-4481 Sub-Cell GT System, with 15 x 10 cm tray and gel caster




170-4405 Wide Mini-Sub Cell GT System

170-4485 Wide Mini-Sub Cell GT System, with gel caster

170-4406 Mini-Sub Cell GT System

170-4486 Mini-Sub Cell GT System, with gel caster

Sub-Cell GT/PowerPac 300 Power Supply Systems *

165-4349 Sub-Cell GT/PowerPac 300 System, 100/120 V

165-4350 Sub-Cell GT/PowerPac 300 System, 220/240 V

165-4348 Wide Mini-Sub Cell GT/PowerPac 300 System, 100/120 V

165-4351 Wide Mini-Sub Cell GT/PowerPac 300 System, 220/240 V

165-4347 Mini-Sub Cell GT/PowerPac 300 System, 100/120 V

165-4352 Mini-Sub Cell GT/PowerPac 300 System, 220/240 V

* All Sub-Cell GT/PowerPac 300 systems come with 15 x 15 cm UVTP tray and gel caster.

15

6.2 Sub-Cell GT System Accessories

Catalog Number Product Description

Sub-Cell GT Systems

170-4410 Sub-Cell GT Base

170-4411 Sub-Cell GT Safety Lid with Cables

170-4412 Gel Caster

170-4413 Sub-Cell GT Electrode (Anode), red

170-4414 Sub-Cell GT Electrode (Cathode), black

170-4415 Sub-Cell GT Gel Casting Gates

170-4416 GT UVTP Gel Tray, 15 x 10 cm




Wide Mini-Sub Cell GT Systems

170-4420 Wide Mini-Sub Cell GT Base

170-4421 Wide Mini-Sub Cell GT Safety Lid with Cables

170-4422 Mini-Gel Caster

170-4423 Wide Mini-Sub Cell GT Electrode (Anode), red

170-4424 Wide Mini-Sub Cell GT Electrode (Cathode), black

170-4425 Wide Mini-Sub Cell GT Gel Casting Gates



Mini-Sub Cell GT Systems

170-4430 Mini-Sub Cell GT Base

170-4431 Mini-Sub Cell GT Safety Lid with Cables

170-4422 Mini-Gel Caster

170-4432 Mini-Sub Cell GT Electrode (Anode), Red

170-4433 Mini-Sub Cell GT Electrode (Cathode), Black

170-4434 Mini-Sub Cell GT Gel Casting Gates



16

Sub-Cell Systems CombsFixed Height Combs For Sub-Cell GT and Wide Mini-Sub Cell GT Systems

Catalog Well Thickness Well Width Well VolumeNumber Number (mm) (mm) Capacity* (µl)

170-4440 1 1.50 106.43 800.0170-4441 2 1.50 50.29 377.0170-4442 4 1.50 26.42 200.0170-4443 10 0.75 9.87 37.0170-4444 10 1.50 9.87 74.0170-4445 15 0.75 5.52 20.7170-4446 15 1.50 5.52 41.4170-4447 20 0.75 4.84 18.2170-4448 20 1.50 4.84 36.4170-4449 30 1.50 2.69 20.2

Multi-channel Pipet Compatible (MP) Fixed Height Combs For Sub-Cell GTand Wide Mini-Sub Cell GT Systems

Catalog Well Thickness Well Width Well VolumeNumber Number (mm) (mm) Capacity * (µl)

170-4450 10 0.75 5.82 21.8170-4451 10 1.50 5.82 43.6170-4452 14 0.75 5.82 21.8170-4453 14 1.50 5.82 43.6170-4454 18 0.75 2.91 10.9170-4455 18 1.50 2.91 21.8170-4456 26 0.75 2.91 10.9170-4457 26 1.50 2.91 21.8

Adjustable Height Combs For Sub-Cell GT and Wide Mini-Sub Cell GTSystems (Adjustable height combs require a comb holder [catalog 170-4320])


170-4328 1 1.50 106.43 800.0170-4345 2 1.50 50.29 377.0170-4325 10 0.75 9.87 37.0170-4326 10 1.50 9.87 74.0170-4323 15 0.75 5.52 20.7170-4324 15 1.50 5.52 41.4170-4321 20 0.75 4.84 18.2170-4322 20 1.50 4.84 36.4170-4344 30 1.50 2.69 20.2

17

Fixed Height Combs for Mini-Sub Cell GT


170-4460 1 1.50 43.43 325.7170-4461 2 1.50 20.32 152.4170-4462 8 0.75 5.54 20.8170-4463 8 1.50 5.54 41.6170-4464 15 0.75 2.59 9.7170-4465 15 1.50 2.59 19.4

Adjustable Height Combs for Mini-Sub Cell GT (Adjustable height combs requirea comb holder [catalog 170-4331])


170-4342 1 1.50 43.43 325.7170-4333 8 1.50 5.54 41.6170-4332 15 1.50 2.59 19.4* Well volume capacity was determined based on a well depth of 0.5 cm.

6.3 Related Bio-Rad Products

Power Supplies

165-5050 PowerPac 300 Power Supply, 100/120 V


Blotting Membranes

161-0153 Zeta-Probe ® Positively Charged Nylon Blotting Membrane,sheets, 9 x 12 cm, 15

161-0154 Zeta-Probe Positively Charged Nylon Blotting Membrane,sheets, 10 x 15 cm, 15





161-0159 Zeta-Probe Positively Charged Nylon Blotting Membrane,roll, 30 cm x 3.3 m, 1

161-0165 Zeta-Probe Positively Charged Nylon Blotting Membrane,roll, 20 cm x 3.3 m, 1

161-0190 Zeta-Probe GT (Genomic Tested) Positively Charged NylonBlotting Membrane, sheets, 9 x 12 cm, 15


18





161-0196 Zeta-Probe GT (Genomic Tested) Positively Charged NylonBlotting Membrane, roll, 30 cm x 3.3 m, 1

161-0197 Zeta-Probe GT (Genomic Tested) Positively Charged NylonBlotting Membrane, roll, 20 cm x 3.3 m, 1

161-0090 Supported Nitrocellulose Membrane, 0.45 micron, sheets, 7 x 8.4 cm, 10

161-0091 Supported Nitrocellulose Membrane, 0.45 micron, sheets, 10 x 15 cm, 10



161-0094 Supported Nitrocellulose Membrane, 0.45 micron, roll, 30 cm x 3 m, 1

161-0095 Supported Nitrocellulose Membrane, 0.20 micron, sheets, 7 x 8.4 cm, 10


161-0097 Supported Nitrocellulose Membrane, 0.20 micron, roll, 30 cm x 3 m, 1

Vacuum Blotting Apparatus

165-5000 Model 785 Vacuum Blotter

165-5001 Model 785 Vacuum Blotter System, 120 VAC

165-5002 Model 785 Vacuum Blotter System, 220/240 VAC

Semi-Dry Transfer Cells

170-3940 Trans-Blot ® SD Semi-Dry Electrophoresis Transfer Cell

170-3948 Trans-Blot SD System, 100/120 VAC


UV Crosslinking Chamber

165-5031 GS Gene Linker ® UV Chamber, 120 VAC

165-5032 GS Gene Linker UV Chamber, 220 VAC



19

Gel Reagents and Standards

162-0019 Low Melt Preparative Grade Agarose, 100 g

162-0133 Molecular Biology Certified Agarose, 500 g

162-0126 High Strength Analytical Grade Agarose, 500 g

170-8200 AmpliSize DNA Size Standard, 50-2,000 bp

170-8210 DNA Size Standard, 1-4.2 Kb ladder

170-8220 DNA Size Standard, 0.7-8.4 Kb

170-3470 DNA Size Standard, λ-Hind III

170-3465 DNA Size Standard, pBR322 AVa II/Eco RI

161-0404 Bromophenol Blue, 10 g

161-0423 Xylene Cyanole FF, 25 g

161-0433 Ethidium Bromide Solution, 10 ml, 10 mg/ml

Electrophoresis Buffers

161-0733 10x Tris/Boric Acid/EDTA (TBE), 1 l

161-0743 50x Tris/Acetic Acid/EDTA (TAE), 1 l

161-0719 Tris, 1 kg

161-0751 Boric Acid, 1 kg

161-0729 EDTA, 500 g

DNA Gel Image Analysis and Documentation Systems

170-3742 Standard Polaroid ® Documentation System, 120 VAC

170-3746 Standard Polaroid Documentation System, 100 VAC

170-3747 Standard Polaroid Documentation System, 220/240 VAC

170-7520 Gel Doc ™ 1000 UV Gel Documentation System-PC, 100 VAC

170-7521 Gel Doc 1000 UV Gel Documentation System-PC, 120 VAC

170-7522 Gel Doc 1000 UV Gel Documentation System-PC, 220/240 VAC

170-7525 Gel Doc 1000 UV Gel Documentation System-Mac, 100 VAC

170-7526 Gel Doc 1000 UV Gel Documentation System-Mac, 120 VAC

170-7527 Gel Doc 1000 UV Gel Documentation System-Mac, 220/240 VAC

20

Section 7References

1. Sambrook, Fritsch, and Maniatis, Molecular Cloning, A Laboratory Manual, Second Edition,Cold Spring Harbor Laboratory Press, 1989.

2. Current Protocols in Molecular Biology,Greene Publishing Associates and Wiley-Interscience,1989.

Additional Reading

3. Kopchick, J. J., Cullen, B. R. and Stacey, D. W., Anal. Biochem., 115, 419 (1981).

4. Southern, E., Methods in Enzymol., 68, 152 (1979).

5. The Bio-Rad Silver Stain - Bulletin 1089, Bio-Rad Laboratories, Hercules, CA.

6. Bittner, M., Kupferer, P. and Morris, C .F., Anal. Biochem., 102, 459 (1980).

7. Bio-Rad Trans-Blot Cell Operation Instructions, Bulletin 1082, Bio-Rad Laboratories, Hercules, CA.

8. Winberg, G. and Hammarskjold, M. L., Nucleic Acids Res., 8, 253 (1980).

9. Jytatekadze, T. V., Axelrod, V. D., Gorbulev, V. G., Belzhelarskaya, S. N. and Vartikyan, R. M.,Anal. Biochem., 100, 129 (1979).

10. Dretzen, G., Bellard, M., Sassone-Corsi, P. and Chambon, P., Anal. Biochem., 112, 295 (1981).

* The Polymerase Chain Reaction (PCR) process is covered by patents owned by Hoffmann-LaRoche. Use of the PCR processrequires a license.

** IEC 1010-1 is an internationally accepted electrical safety standard for laboratory instruments.

21

Australia, Bio-Rad Laboratories Pty Limited, Block Y Unit 1, Regents Park Industrial Estate, 391 Park Road, Regents Park, NSW 2143 • Phone 02-9414-2800 • Fax 02-9914-2888Austria, Bio-Rad Laboratories Ges.m.b.H., Auhofstrasse 78D, 1130 Wien • Phone (1) 877 89 01 • Fax (1) 876 56 29Belgium, Bio-Rad Laboratories S.A./N.V., Begoniastraat 5, 9810 Nazareth Eke • Phone 09-385 55 11 • Fax 09-385 65 54Canada, Bio-Rad Laboratories (Canada) Ltd., 5671 McAdam Road, Mississauga, Ontario L4Z 1N9 • Phone (905) 712-2771 • Fax (905) 712-2990China, Bio-Rad Laboratories, 14, Zhi Chun Road, Hai Dian District, Beijing 100088 • Phone (01) 2046622 • Fax (01) 2051876Denmark, Bio-Rad Laboratories, Symbion Science Park, Fruebjergvej 3, DK-2100 Copenhagen • Phone 39 17 9947 • Fax 39 27 1698Finland, Bio-Rad Laboratories, Business Center Länsikeskus, Pihatörmä 1A SF-02240, Espoo, • Phone 90 804 2200 • Fax 90 804 1100France, Bio-Rad S.A., 94/96 rue Victor Hugo, B.P. 220, 94 203 Ivry Sur Seine Cedex • Phone (1) 49 60 68 34 • Fax (1) 46 71 24 67Germany, Bio-Rad Laboratories GmbH, Heidemannstraße 164, D-80939 München/Postfach 450133, D-80901 München • Phone 089 31884-0 • Fax 089 31884-100India, Bio-Rad Laboratories, C-248 Defence Colony, New Delhi 110 024 • Phone 91-11-461-0103 • Fax 91-11-461-0765Italy, Bio-Rad Laboratories S.r.l.,Via Cellini, 18/A, 20090 Segrate Milano • Phone 02-21609 1 • Fax 02-21609-399Japan, Nippon Bio-Rad Laboratories, 7-18, Higashi-Nippori 5-Chome, Arakawa-ku, Tokyo 116 • Phone 03-5811-6270 • Fax 03-5811-6272The Netherlands, Bio-Rad Laboratories B. V., Fokkerstraat 10, 3905 KV Veenendaal • Phone 0318-540666 • Fax 0318-542216New Zealand, Bio-Rad Laboratories Pty Ltd., P. O. Box 100-051, North Shore Mail Centre, Auckland 10 • Phone 09-443 3099 • Fax 09-443 3097Pacific, Bio-Rad Laboratories, Unit 1111, 11/F., New Kowloon Plaza, 38, Tai Kok Tsui Road, Tai Kok Tsui, Kowloon, Hong Kong • Phone 7893300 • Fax 7891257Singapore, Bio-Rad Laboratories (Singapore) Ltd., 221 Henderson Rd #05-19, Henderson Building, Singapore 0315 • Phone (65) 272-9877 • Fax (65) 273-4835Spain, Bio-Rad Laboratories, S. A. Avda Valdelaparra 3, Pol. Ind. Alcobendas, E-28100 Alcobendas, Madrid • Phone (91) 661 70 85 • Fax (91) 661 96 98Sweden, Bio-Rad Laboratories AB, Gärdsvägen 7D, Box 1276, S-171 24 Solna • Phone 46-(0)8-735 83 00 • Fax 46-(0)8-735 54 60Switzerland, Bio-Rad Laboratories AG, Kanalstrasse 17, Postfach, CH-8152 Glattbrugg • Phone 01-809 55 55 • Fax 01-809 55 00United Kingdom, Bio-Rad Laboratories Ltd., Bio-Rad House, Maylands Avenue, Hemel Hempstead, Herts HP2 7TD • Free Phone 0800 181134 • Fax 01442 259118

Life Science Group

2000 Alfred Nobel DriveHercules, California 94547Telephone (510) 741-1000Fax: (510) 741-5800

SIG 020996 Printed in USA

M1704400 REV A

Bio-Rad Laboratories

Sub-Cell® Model 96 andModel 192

Agarose GelElectrophoresis Systems

Instruction Manual

Catalog Numbers170-4500 through 170-4511

For Technical Service Call Your Local Bio-Rad Off ice or in the U.S. Call 1-800-4BIORAD (1-800-424-6723)

Warranty

Bio-Rad Laboratories warrants the Sub-Cell Model 96 and Model 192 electrophoresis systems againstdefects in materials and workmanship for 1 year. If any defects occur in the instrument during this warranty period, Bio-Rad Laboratories will repair or replace the defective parts free. The followingdefects, however, are specifically excluded:

1. Defects caused by improper operation.

2. Repair or modification done by anyone other than Bio-Rad Laboratories or an authorizedagent.

3. Use of fittings or other spare parts supplied by anyone other than Bio-Rad Laboratories.

4. Damage caused by accident or misuse.

5. Damage caused by disaster.

6. Corrosion due to use of improper solvent or sample.

This warranty does not apply to parts listed below:

1. Platinum Electrode Wires

For any inquiry or request for repair service, contact Bio-Rad Laboratories after confirming themodel and serial number of your instrument.

To insure the best performance from the Sub-Cell electrophoresis systems, become fully acquaintedwith these operating instructions before use. Bio-Rad recommends that you first read these instructionscarefully. Assemble and disassemble the unit completely without casting a gel. After these prelimi-nary steps, you should be ready to cast and run a gel.

Bio-Rad also recommends that all Sub-Cell system components and accessories be inspected fordamage, cleaned as recommended in this manual and rinsed thoroughly with distilled water beforeuse. Record the following for your records:

Model

Catalog No.

Date of Delivery

Warranty Period

Serial No.

Invoice No.

Purchase Order No.

Table of ContentsPage

Warranty Information ..........................................................................Inside Front Cover

Section 1 General Information....................................................................................11.1 Introduction ................................................................................................................11.2 Safety ..........................................................................................................................11.3 List of System Parts....................................................................................................31.4 Specifications .............................................................................................................5

Section 2 Operating Instructions ................................................................................62.1 DNA Gel Preparation.................................................................................................62.2 Comb Set-up...............................................................................................................72.3 Casting Agarose Gels.................................................................................................82.4 Recirculation Ports ...................................................................................................112.5 Electrophoresis .........................................................................................................122.6 Nucleic Acid Staining and Visualization.................................................................132.7 Note on Blotting .......................................................................................................14

Section 3 Gel and Electrophoresis Reagents Preparation......................................143.1 Electrophoresis Buffer Preparation..........................................................................143.2 DNA and RNA Gel Preparation ..............................................................................153.3 RNA Sample Preparation.........................................................................................153.4 DNA and RNA Sample Loading Dye .....................................................................163.5 Gel Staining Solution ...............................................................................................16

Section 4 Care and Maintenance ..............................................................................164.1 Cleaning Sub-Cell System Components..................................................................164.2 Compatible Cleaning Agents ...................................................................................164.3 Maintenance Schedule .............................................................................................174.4 Electrode Replacement.............................................................................................174.5 RNase Decontamination ..........................................................................................18

Section 5 Troubleshooting..........................................................................................19

Section 6 Ordering Information ...............................................................................206.1 Sub-Cell Model 96 and 192 Systems.......................................................................206.2 Sub-Cell Model 96 and 192 System accessories.....................................................206.3 Related Bio-Rad Products........................................................................................21

Section 7 References ...................................................................................................25

i

Section 1General Information

1.1 IntroductionThe Sub-Cell instruments comprise a comprehensive and versatile gel electrophoresis

system that effectively separates nucleic acids using submerged agarose gels. Submarineagarose gels are easy to cast and readily dissipate heat. These gels allow sample underlayingand also prevent electrical field discontinuities caused by wicks or sample well dehydration.Agarose gels are ideal for the separation of DNA restriction digestions, Polymerase ChainReaction (PCR)* amplified fragments, and genomic DNA and RNA prior to Southern orNorthern blotting. If operated correctly, agarose gels can effectively separate nucleic acidsfrom 20 base pairs to 20 kilobase pairs in length.

The Sub-Cell Model 96 and 192 electrophoresis cells have been created specifically formultiple sample analysis. The width of each cell and the analytical combs were designedbased on the fixed spacing of multichannel pipets used to transfer samples from standardmicroplates. Forty eight nucleic acid samples (plus three DNA size standards) can be visualizedin one row using the 51-well comb. If two combs are used, samples from all 96 wells of amicroplate can be analyzed on the Model 96. Four or more combs can be used on the Model 192 for even higher throughput. The Model 96 can run gels 10 or 15 cm in length,whereas the Model 192 can run gels 10, 15, 20 or 25 cm in length for the analysis of more sam-ples or applications such as genomic DNA separations for Southern blotting.

The Sub-Cell systems give years of rigorous use. These rugged systems incorporate manyfeatures that make casting and running agarose gels simple and efficient. The gel caster pro-vides tape-free gel casting in trays and gels can be cast in the Sub-Cell bases using the gelcasting gates. Replaceable electrode cassettes provide a simple way to replace electrode wires.A comprehensive assortment of tray sizes and multichannel pipet-compatible combs makethese systems ideal for most high throughput agarose gel applications. Recirculation portsare provided to prevent heat and pH effects during high voltage or extended run elec-trophoresis.

Note: This manual contains instructions for the Sub-Cell Model 96 and Model 192 electrophoresis systems only. Bio-Rad supplies similar but smaller agarose gel electrophoresis cells: the original Sub-Cell, Wide Mini-Sub Cell, and Mini-Sub Cell systemsand the Sub-Cell GT, Wide Mini-Sub Cell GT and Mini-Sub Cell GT systems. This manual does not provide information concerning these smaller versions. Contact your localBio-Rad representative for information concerning the original Sub-Cell and Sub-Cell GT systems.

* The Polymerase Chain Reaction (PCR) process is covered by patents owned by Hoffmann-LaRoche. Use of the PCR processrequires a license.

1.2 Safety The Sub-Cell electrophoresis systems are designed for maximum user safety. The buffer

chambers are made of 1/4-inch (.635 cm) thick cast acrylic to create a leak-free electrophoresis environment. The safety lids surround the entire buffer chamber to protect theuser from exposure to electrical currents. Sub-Cell systems were designed for indoor use only.

1

Before every use, inspect the base for cracks or chips. Cracks or chips may cause thebuffer to leak from the base and cause a potential electrical hazard. Additionally, inspect allelectrical cables, banana jacks, recirculation port fittings, tubing, and plugs for loose connections, cracks, breaks or corrosion. Do not use any part that is cracked, charred or corroded. These parts may also cause a potential electrical shock. Contact your local Bio-Rad representative before using a part that may be considered hazardous.

During electrophoresis inspect the base and workbench for any signs of buffer leakage.If leaking buffer is detected disconnect the power to the cell immediately and contact your Bio-Rad representative.

Power to Sub-Cell units is to be supplied by an external DC-voltage power supply. Thispower supply must be ground isolated in such a way that the DC voltage output floats withrespect to ground. All Bio-Rad power supplies meet this important safety requirement. Therecommended power supply for this apparatus is the PowerPac 300. The PowerPac 300 contains several safety features such as no load, overload, rapid resistance change, and groundleak detection capabilities. The maximum specified operating parameters* for the Sub-Cell Model 96 and Model 192 systems are:

200 VDC Maximum voltage limit

70 Watts Maximum power limit

50 °C Maximum buffer temperature

4 °C – 40 °C Ambient temperature limits

* IEC 1010-1 certification applies to equipment designed to be safe at the operating parameters listed above. Additionally, bothSub-Cell Model 96 and Model 192 have a maximum operating relative humidity of 80% for temperatures up to 31 °C decreasinglinearly to 50% relative humidity at 40 °C. Certification is valid when systems are operated at altitudes up to 2000 meters.

Current to the cell, provided from the external power supply, enters the unit through thelid assembly, providing a safety interlock to the user. Current to the cell is broken when thelid is removed. Do not attempt to circumvent this safety interlock, and always turn the powersupply off before removing the lid or when working with the cell in any way.

Important: These Bio-Rad instruments are designed and certified to meet IEC 1010-1*safety standards. IEC-certified products are safe to use when operated in accordance withthis instruction manual. This instrument should not be modified in any way. Alteration ofthis instrument will:

• Void the manufacturer’s warranty

• Void the IEC 1010-1 safety certification

• Create a potential safety hazard

Bio-Rad is not responsible for any injury or damage caused either by the use of this instru-ment for purposes other than for which it is intended or by modifications of the instrument notperformed by Bio-Rad or any authorized agent. No user-serviceable parts are contained inthis apparatus. To ensure electrical safety, do not attempt to service this apparatus.

* IEC 1010-1 is an internationally accepted electrical safety standard for laboratory instruments.

Definition of Symbols

Caution, risk of electrical shock Caution (refer to accompanying documents)

2

1.3 List of System PartsEach Sub-Cell system comes with the components listed in Table 1.1. Check your

instrument to insure all items are present. Note any damage to the unit which may haveoccurred during shipping. Notify Bio-Rad Laboratories if any items are missing or damaged(see Figure 1.1 for part descriptions, on the following page).

Table 1.1 Sub Cell System Components

Item Quantity

Base (buffer chamber) 1

Gel Casting Gates 2

Safety Lid and Cables 1

UVTP Gel Tray 1

Comb (51 well, 1.5 mm thick) 1

Comb Holder 1

Leveling Bubble 1

Gel Caster (optional) 1

Instruction Manual 1

3

Figure 1.1 Sub-Cell Model 96 and Model 192 Parts

4

Safety Lid ElectricalCables

Recirculation Port Plugs

Comb Holder

CombElectrical

Leads

UVTP Gel Tray

Comb Holder Slot

FluorescentRuler

Banana Plug/ElectrodeWire Assembly

Gel CastingGates

Base(Buffer Chamber)

LevelingFeet

Figure 1.2 Sub-Cell Model 96 and Model 192 Gel Caster Parts

1.4 Specifications

Sub-Cell Model 96

Base Footprint (L x W x H) 29 .5 cm x 29.0 cm x 9.0 cm

Base Buffer Volume* 2.0 L

Base Gel Size 25 x 10 cm

Gel Tray Sizes 25 x 10 cm25 x 15 cm

Sub-Cell Model 192

Base Footprint (L x W x H) 39.5 cm x 29.0 cm x 9.0 cm

Base Buffer Volume* 3.0 L

Base Gel Size 25 x 15 cm

Gel Tray Sizes 25 x 10 cm25 x 15 cm25 x 20 cm25 x 25 cm

5

Comb andComb Holder

UVTP Gel Tray

Gel Caster LevelingBubble

Leveling Feet

Movable Wall

Fixed Wall

Contruction

Base Cast Acrylic

Gel Casting Gates Anodized Aluminum

Safety Cover Cast Acrylic

Banana Plug/Electrode Cassette Polycarbonate

Banana Plugs Gold-Plated Brass, 4.4 cm Length

Electrodes Platinum, 0.25 mm Diameter

Electrical Cables Dual, 20 AWG, Tinned Copper Wire CableFlame-Retardant Polyurethane Insulation jacket

Electrical Leads Nickel Silver

Gel Tray UV-Transparent Acrylic Plastic (UVTP)

Combs Machined Acrylic

Comb Holder Polycarbonate

Gel Casting Device Polycarbonate0.64 cm Silicon Foam

* Base buffer volumes will vary depending on the size and thickness of gel used.

Section 2Operating Instructions

Note: Refer to Section 3 for information on preparation of RNA gels. See References 1 and 2 for more information on DNA and RNA electrophoresis.

2.1 DNA Gel PreparationDNA agarose gels can be used to separate and visualize DNA of various sizes. Before

casting an agarose gel, consult Table 2.1 to determine the appropriate percent agarose gel touse based on the size of DNA to be separated.

Procedure

1. Determine the amount of agarose (grams) and volume needed. Use Tables 2.1 and 2.2as a guide for agarose concentration and gel volume requirements.

Example: For a 1% agarose gel, add 1 gram of agarose to 100 ml of electrophoresis buffer.

Table 2.1 Gel Concentration Required for DNA Separation 1-2

Gel Concentration % DNA Size (Kbp)

0.50 1 – 30

0.75 0.8 – 12

1.00 0.5 – 10

1.25 0.4 – 7

1.50 0.2 – 3

2-5* 0.01 – 0.5

* Sieving agarose such as Bio-Rad AmpliSize® agarose.

6

Table 2.2 Gel Volume Requirements

Gel Size 0.5 cm thick 0.75 cm thick 1.0 cm thick

Bases

25 x 10 cm (Model 96) 125 ml 185 ml 250 ml

25 x 15 cm (Model 192) 185 ml 280 ml 375 ml

Trays

25 x 10 cm 125 ml 185 ml 250 ml

25 x 15 cm 185 ml 280 ml 375 ml

25 x 20 cm 250 ml 375 ml 500 ml

25 x 25 cm 310 ml 465 ml 625 ml

2. Add the agarose to a suitable container (e.g., 500-ml Erlenmeyer flask, Wheaton bottle, etc.).Add the appropriate amount of electrophoresis buffer (see Section 3) and swirl to sus-pend the agarose powder in the buffer. If using an Erlenmeyer flask, invert a 50 mlErlenmeyer flask into the open end of the 500-ml Erlenmeyer flask containing the agarose.The small flask acts as a reflux chamber, thus allowing long or vigorous boiling withoutmuch evaporation.

Note: Place a mark on the flask at the liquid level. If evaporation occurs, water can beadded to bring the volume back to the original liquid level.

3. The agarose can be melted by boiling on a magnetic hot plate or in a microwave oven.

Caution: Always wear protective gloves, goggles, and a lab coat while preparing andcasting agarose gels. Boiling molten agarose or the vessels containing hot agarose cancause severe burns if allowed to contact skin. Molten agarose can become super-heatedand boil over vessels when swirled which can also cause severe burns.

Magnetic Hot Plate Method

4a. Add a stir bar to the undissolved agarose solution. Heat the solution to boiling while stirring on a magnetic hot plate. Use the appropriate size container to allow bubbles orfoam to disrupt before rising to the neck of the container.

5a. Boil the solution until all of the small translucent agarose particles are dissolved. Set asideto cool to 50-60 °C before pouring.

Microwave Oven Method

4b. Place the gel solution into the microwave. Using a low to medium setting, set the timer fora minimum of 5 minutes, stopping the microwave oven every 30 seconds and swirling thecontainer gently to suspend the undissolved agarose. This technique is the fastest andsafest way to dissolve agarose.

5b. Boil and swirl the solution until all of the small translucent agarose particles are dissolved.Set aside to cool to 50-60 °C before pouring.

2.2 Comb Set-up

Comb and Comb Holder Set-up

The comb holder used for the Model 96 and 192 was designed to incorporate all the necessary features required for any agarose gel application. The comb holder allows foradjustable comb height and can be adjusted so that the comb can be placed anywhere on the

7

base stage or UVTP tray. The following instructions describe how to manipulate the comb andcomb holder for obtaining comb height and comb holder position on a UVTP tray or base stage.

Adjusting and Setting Comb Height

1. Loosen the five thumbscrews from the front plate of the comb holder.

2. Align the slots of the well-forming comb with the thumbscrews on the comb holder. Insertthe slots over the shaft (threaded portion) of the thumbscrews and tighten until the flat head(shoulder) of the screws come in contact with the comb.

3. Place the comb holder assembly on the cell base or UVTP tray and adjust the height ofthe comb to the desired distance from the surface of the base stage or tray (typically 1-2 mm).Tighten all five screws once the full-length of the comb is at a uniform distance from thebase stage or tray.

Adjusting and Setting Comb Position on UVTP Tray or Base Stage

1a. Turn the two thumbscrews clockwise on the sides of the comb holder until resistance isfelt. With the screws in this position, the comb holder can be placed into the comb slotsof the base and UVTP tray for gel casting.

OR

1b. Turn the two thumbscrews counterclockwise on the sides of the comb holder until theshaft (threaded portion) of the thumbscrews can no longer be seen in the comb holdernotches. With the screws in this position, this will allow the comb holder assembly to beplaced anywhere on the base or UVTP tray. The comb can be secured to the tray or baseby turning the thumbscrews clockwise until resistance is felt.

2.3 Procedures for Casting Agarose Gel SlabsThere are several ways to cast agarose submarine gels for the Model 96 and Model 192.

Gels may be cast with or without UV-transparent plastic (UVTP) trays directly on the stageof the Sub-Cell bases using the gel casting gates. Gels may also be cast on UVTP trays withthe aid of the gel caster or with standard laboratory tape.

Casting Gels on the Base Stages

1. Level the Sub-Cell base using the leveling bubble provided.

2. Slide the gel casting gates into the slots at opposite ends of the gel stage.

3. Place the comb(s) into the appropriate slot(s) of the base so that the sample wells are nearthe cathode (black) (refer to Section 2.2 for comb adjustments). DNA samples will migratetowards the anode (red) during electrophoresis.

4. When the solution of agarose has cooled to 60 °C (Section 2.1), pour the molten agarosebetween the gates.

Warning: Hot agarose (>60 °C) may cause the cell to warp or craze and will decrease thelifetime of the cell. Warping may also result in sample wells of uneven depth.

5. Allow 30 – 60 minutes for the gel to solidify at room temperature.

6. Carefully remove the comb and then remove the gel casting gates from the gate slots ofthe base.

7. Submerge the gel beneath 4 to 6 mm of electrophoresis buffer (Section 3.1).

8

Removable Tray (UVTP) Gel Casting

Casting gels on the base stage with UVTP tray

1. Level the cell using the leveling bubble provided.

2. Place the UVTP tray on the cell base stage.

Note: The Sub-Cell Model 96 system requires the 25 x 10 cm UVTP tray for casting inthe base. Sub-Cell Model 192 system requires the 25 x 15 cm UVTP tray for casting inthe base.

3. Slide the gel casting gates into the slots at opposite ends of the base stage. Ensure thegates are evenly seated in the slots and the gates uniformly contact all edges of the UVTPtray. The weight of the gates provide a tight seal to avoid any leakage problems during gelcasting.

4. Place the comb(s) into the appropriate slot(s) of the trays so that the sample wells arenear the cathode (black). DNA samples will migrate towards the anode (red) during elec-trophoresis.

5. Prepare the desired concentration and amount of agarose in 1x electrophoresis buffer (see section 2.1). When the agarose solution has cooled to 50-60˚ C pour the moltenagarose between the gates.

Warning: Hot agarose (>60 ˚C) may cause the tray to warp or craze and will decrease thelifetime of the tray. Warping may also result in sample wells of uneven depth.




Gel Caster Method

1. Level the gel caster on the lab bench using the leveling bubble provided.

2. Disengage and slide the movable wall to the open end of the gel caster by turning andlifting the cam peg upward.

3. Place the UVTP tray against the fixed wall of the gel caster.

4. Slide the movable wall against the edge of the UVTP tray (Figure 2.1).

5. To seal the open tray ends, engage the cam peg by turning and pressing downward simul-taneously.

6. Once the cam peg has dropped down into the appropriate slot, turn the peg in either direction until resistance is felt. This action seals the ends of the tray for casting.

9

Figure 2.1 Sealing the UVTP tray for gel casting.

7. Place the comb(s) into the appropriate slot(s) of the tray (refer to Section 2.2 for combadjustments).

8. When the solution of agarose has cooled to 60 °C (Section 2.1), pour the molten agaroseonto the tray.




11. Disengage the cam peg by turning and lifting upward. Slide the movable wall away fromthe tray. Remove the tray from the gel caster.

Note: While the gel is solidifying, a light seal is formed between the gasket and the gel(especially for low percentage agarose gels (<0.8%). Carefully lift the tray on one side torelease the seal.

12. Place the tray onto the leveled Sub-Cell base so that the sample wells are near the cathode(black). DNA samples will migrate towards the anode (red) during electrophoresis.

13. Submerge the gel beneath 4 to 6 mm of electrophoresis buffer (Section 3.1).

10

Movable WellLift cam lever up

UVTP Gel Tray

Fixed wall

Leveling Feet

Gel Caster

Leveling Bubble

Engage and seal(press down and rotate)

Slide Forward

Tape Method

1. Seal the ends of the UVTP gel tray securely with strips of standard laboratory tape. Pressthe tape firmly to the edges of the gel tray to form a fluid-tight seal.

2. Level the gel tray on a leveling table or workbench using the leveling bubble provided withthe instrument.

3. Place the comb(s) into the appropriate slot(s) of the tray (refer to Section 2.2 for combadjustments).

4. When the solution of agarose has cooled to 60 °C (Section 2.1), pour the molten agaroseonto the tray.


5. Allow the gel to solidify at room temperature for 30 – 60 minutes.


7. Remove the tape from the edges of the gel tray. Be careful when removing tape so the geldoes not slide off the tray.

8. Place the tray onto the leveled Sub-Cell base so that the sample wells are near the cathode(black). DNA samples will migrate towards the anode (red) during electrophoresis.

9. Submerge the gel under 4 to 6 mm of electrophoresis buffer.

2.4 Recirculation PortsBuffer recirculation is not required for most run conditions on the Sub-Cell systems. We

recommend buffer recirculation for extended run times (over 2 hours) or for high voltage run conditions (150-200 volts). This will prevent lane distortion that can arise from uneven heating or buffer pH gradients. If recirculation is desired, the buffer recirculation kit (Bio-Rad catalog number 170-4537) contains the adapters required to connect the pump tubing to the Sub-Cell lid.

1. Carefully remove the port plugs from the safety lid.

2. Turn clockwise and tighten the elbow-shaped recirculation port fitting into the threadedport holes (Figure 2.2).

Note: There should be at least three threads extending below the bottom surface of thesafety lid.

Figure 2.2. Connecting the recirculation ports and tubing.

11

Elbow-shapedFitting

Safety Lid

Recirculation Port

Straight Fitting

Tubing

3. Attach and tighten (10 lb.-in. torque) the straight fitting to the elbow-shaped fitting.

4. Connect tubing to the elbow-shaped fittings on the safety lid. Connect the other end of thetubing to a suitable buffer recirculation pump (Section 6.3). Attach the tubing clips at alltubing/fitting connections to insure that tubing does not disengage during electrophoresis.

5. Recirculate the buffer at a rate of 300-500 ml/min. Pumping at a higher rate will cause thegel to float or slide off the tray causing variable sample migration rates during electrophoresis. Check for any leaking in the fitting, tubing, and pump connections beforeturning on the power supply and starting electrophoresis.

Note: If recirculation port fittings are to be removed, always cover the port holes byreplacing the port plugs (use 5 lb.-in. torque to tighten).

2.5 ElectrophoresisOnce the agarose gel has solidified, sample loading and electrophoresis can begin. Agarose

gels can be run in many different types of electrophoresis buffers. Nucleic acid agarose gelelectrophoresis is usually conducted with either Tris-Acetate-EDTA (TAE) buffer or Tris-Borate-EDTA (TBE) buffer. While TAE buffer provides faster electrophoretic migration oflinear DNA and better resolution of supercoiled DNA, TBE buffers have a stronger bufferingcapacity and are less conductive than TAE buffers and therefore are used for longer or highervoltage electrophoresis runs.

Note: Because of the higher voltages and resulting higher currents often used with theModel 96 and Model 192, it is strongly recommended that only TBE buffers be used forelectrophoresis. TBE buffers have a stronger buffering capacity and are less conductive.Thus, pH or temperature gradient formation during extended electrophoresis will bereduced. If pH or temperature gradients cause uneven sample migration reduce the voltage,add more buffer or recirculate the buffer during electrophoresis to eliminate these effects(Section 2.4). Bio-Rad offers premixed 10x TBE buffers as well as individual bufferreagents for use with the Sub-Cell systems (Section 6.3).

1. When placing the gel tray into the base, make sure that the sample wells are at the cathode(black). DNA samples will migrate towards the anode (red) during electrophoresis.

2. Prepare the desired concentration of electrophoresis buffer (the electrophoresis bufferused should be identical to the type used for gel preparation).

3. Submerge the gel under 4 to 6 mm of electrophoresis buffer. Do not fill buffer above themax. buffer mark on the Sub-Cell base.

4. Prepare samples for gel loading. The maximum sample loading volume for Bio-Radcombs is listed in Section 6.2. Loading volume is dependent upon the type of comb used(i.e., well thickness and length) and thickness of the gel.

5. Once loading volume is determined, samples are made dense for underlaying into samplewells by using standard nucleic acid sample loading dyes (refer to Section 3.4 for sampleloading dye preparation). Add loading dye to a final 1x concentration.

6. Load the samples into the wells using standard pipets or multichannel pipets.

Note: Sample wells are often difficult to see. Well visualization can be enhanced by placingblack paper or tape under the base or tray where comb placement or well formation iscommon.

7. Place the lid on the DNA cell carefully. Do not disturb the samples. The Sub-Cell systemslid attaches to the base in one orientation only. To attach the lid correctly, match the redand black banana jacks on the lid with the red and black banana plugs of the base.

12

8. Power requirements vary depending on gel thickness, length and concentration, and typeof electrophoresis buffer used. Refer to Table 2.3 for relative sample migration rate for theSub-Cell Model 96 and Model 192 systems. Also, review Table 2.4 for DNA size migrationwith sample loading dyes.

Note: Buffer recirculation is not required for most standard DNA and RNA agarose gelelectrophoresis. For most electrophoresis, TBE buffer is recommended. If buffer recirculation is required, use the recirculating ports (Section 2.4).

Table 2.3 Relative Sample Migration Rates*

Bromophenol BlueCell Type Voltage migration rate

Sub-Cell Model 96 200 V 5.15 cm/hr

Sub Cell Model 192 200 V 6.20 cm/hr

* Note: These sample migration rates were determined based on a 0.5 cm thick 1.0% agarose gel using Bio-Rad Molecular BiologyCertified Agarose in 1x TBE buffer diluted from Bio-Rad Premixed 10x TBE Buffer). Migration rates will vary depending on thevoltage, current, and type of agarose or buffer used.

Table 2.4 DNA Size Migration with Sample Loading Dyes

Agarose Concentration (%) Xylene Cyanol Bromophenol Blue

0.5 – 1.5 4-5 Kbp 400-500 bp

2.0 – 3.0 * 750 bp 100 bp

4.0 – 5.0* 125 bp 25 bp

* Sieving agarose such as Bio-Rad AmpliSize agarose.

9. With the desired power requirements, begin electrophoresis. If using buffer recirculation,electrophorese for 15 minutes before turning the pump ON.

Note: Buffer recirculation is optional for gels that require short run times. Gels run athigher voltages (200 volts) may require recirculation to prevent heat or pH effects.Recirculate the buffer at a rate of 300-500 ml/min. Do not pump at a higher rate, it willcause the gel to float or slide off the tray causing variable sample migration rates duringelectrophoresis.

10. After electrophoresis is complete, turn off the power. If using buffer recirculation, do notturn the pump OFF and do not disconnect the tubing from the safety lid. Lift the safety lidwith the pump still ON and empty the buffer contained in the tubing and pump into thebase buffer chamber. When the tubing is empty, turn the pump OFF and disconnect thetubing if desired.

2.6 Nucleic Acid Staining and VisualizationGels can be removed from the base or gel tray for nucleic acid staining. The gel can also

remain on the UVTP gel tray for staining.

Ethidium Bromide Staining Procedure

1. Place the gel into the appropriate volume of 0.5 µg/ml ethidium bromide (EtBr) and stainfor 15–30 minutes. Use enough staining solution to cover the entire gel.

Caution: Ethidium bromide is a suspected carcinogen and should be handled with extremecare. Always wear gloves, eye glasses and a laboratory coat. Dispose of used EtBr solutions

13

and gels appropriately (Review EtBr Material Safety Data Sheet [MSDS] for properdisposal methods).

2. Destain the gel for 10-30 minutes in dH2O using the same volume used for staining.

Note: Ethidium Bromide can be removed from the DNA with extended destaining. Thiswill cause lower sensitivity of detection. However, insufficient destaining will createhigher background fluorescence.

3. Rinse the gel briefly with dH2O once to remove any residual staining solution.

4. Place the gel on a UV transilluminator for nucleic acid visualization and analysis. DNA-Ethidium Bromide complexes may be illuminated with UV light of 254, 302, or 366 nm.Sensitivity decreases with illumination at higher wavelength. However, nicking of DNAwill increase below 302 nm. Table 2.5 indicates the percentage of transmittance of UVlight through 1/4″ (.635 cm) UV-transparent plastic.

Note: Nucleic acids in the gel can be visualized through the UVTP trays. If a UVTP trayis not used, place household plastic wrap between the UV transilluminator and the gel toavoid contaminating the transilluminator with nucleic acids or EtBr.

Table 2.5 Percent UV Transmittance through 1/4” (.635 cm) UV Transparent Plastic

Approximate % Wavelength (nm) Transmittance

254 0

302 80

360 90

5. Photograph the gel using standard cameras and film (e.g., Bio-Rad Standard Polaroid GelDocumentation System) or with CCD-based digitized image analysis systems (e.g., Bio-RadGel Doc™ 1000). Gels are generally photographed with a yellow, orange, or red inter-ference filter. Red filters generally give the cleanest background. Bio-Rad offers a full-lineof standard photography and CCD-based imaging systems for nucleic acid gel analysis.

2.7 Note on BlottingNucleic acids within the gel can be transferred to membranes using the techniques of

Southern and Northern blotting. It is beyond the scope of this instruction manual to includeblotting procedures. Consult References 1 and 2 for blotting techniques. Bio-Rad offers afull-line of nitrocellulose and positively-charged nylon membranes, as well as vacuum andelectrophoretic blotting apparatus for Southern and Northern blotting (Section 6.3).

Section 3Gel and Electrophoresis Reagents Preparation

3.1 Electrophoresis Buffer PreparationDNA agarose gel electrophoresis is usually conducted with either Tris-Acetate-EDTA

(TAE) or Tris-Boric Acid-EDTA (TBE). While TAE provides faster electrophoretic migrationof linear DNA and better resolution of supercoiled DNA, TBE buffers have a stronger bufferingcapacity for longer- or higher-voltage electrophoresis runs. Bio-Rad offers premixed 50x TAE and 10x TBE buffers for use with the Sub-Cell systems. RNA formaldehyde gelsrequire a MOPS [3-(N-morpholino)-propanesulfonic acid] electrophoresis buffer.

14

1x Tris-Acetate-EDTA (TAE) — 40 mM Tris (pH 7.6), 20 mM Acetic Acid, and 1 mM EDTA.50x Stock (1 liter): dissolve in 600 ml distilled water:

Tris Base (FW = 121) 242.0 gGlacial acetic acid 57.1 ml0.5 M EDTA (pH 8.0) 100.0 ml

Fill to a final volume of one liter with distilled water.

1x Tris-Boric Acid-EDTA (TBE) — 89 mM Tris (pH 7.6), 89 mM Boric Acid, 2 mM EDTA10x Stock (1 liter): dissolve in 600 ml distilled water:

Tris Base (FW = 121) 108 gBoric Acid (FW = 61.8) 55 g0.5 M EDTA (pH 8.0) 40 ml

Fill to a final volume of one liter with distilled water.

1x MOPS Buffer (RNA Gels) — 0.02 M MOPS [3-(N-morpholino)-propanesulfonic acid] (pH 7.0), 8 mM Sodium Acetate, 1 mM EDTA (pH 8.0)5x Stock (1 liter): dissolve in 600 ml DEPC-treated distilled water:

MOPS 20.6 g 3 M Sodium Acetate (DEPC treated) pH 7.4 13.3 ml0.5 M EDTA (DEPC-treated) pH 8.0 10.0 ml

Fill to a final volume of one liter with DEPC-treated distilled water.

Caution: DEPC is a suspected carcinogen. Always wear gloves, eye glasses and a laboratory coat. Use caution when handling DEPC containing solutions. Consult theDEPC MSDS (Material Safety Data Sheet) for more information.

3.2 DNA and RNA Gel Preparation

DNA Agarose Gels

(See Section 2.1)

RNA Agarose Formaldehyde Gels 1-2

For 100 ml of a 1% agarose formaldehyde gel prepare as follows:1.6% melted agarose 62 ml5x MOPS Electrophoresis Buffer (1x final concentration) 20 ml12.3 M (37.5%) Formaldehyde (2.2 M final concentration) 18 ml

Caution: Formaldehyde solutions and formaldehyde vapors are toxic. When handlingsolutions or gels that contain formaldehyde use a chemical hood. Always wear gloves, eyeglasses and a laboratory coat while using formaldehyde. See the formaldehyde MSDSfor more safety information.

3.3 RNA Sample Preparation 1-2

Prior to loading RNA onto an agarose formaldehyde gel prepare each RNA sample asfollows:

3.0 µl RNA in DEPC-treated water5.0 µl 5x MOPS Buffer (final concentration 1.67x)4.5 µl 12.3 M Formaldehyde (final concentration 3.7 M)12.5 µl Formamide (final concentration 50% v/v)Caution: Formamide is a teratogen. Always wear gloves, eye glasses and a laboratorycoat. Use caution when handling formamide. Consult the formamide MSDS for moreinformation.

15

3.4 DNA and RNA Sample Loading Dye 1-2

A convenient 10x sample buffer stock consists of 50% glycerol, 0.25% bromophenolblue, and 0.25% xylene cyanole FF in 1x electrophoresis buffer. Prepare only 1-10 ml of the10x loading dye.

3.5 Gel Staining SolutionAdd 10 mg of ethidium bromide to 1 ml distilled water. Bio-Rad offers pre-mixed

EtBr solutions (10 mg/ml). Store reagent in the dark.

Section 4Care and Maintenance

4.1 Cleaning Sub-Cell System Components1. All Sub-Cell systems parts should be washed with a mild detergent solution in warm

water. If necessary, use a soft-bristled brush or sponge to remove dried buffer salts oragarose.

Note: Be careful not to snag or break the electrode wire in the base while cleaning.

2. Rinse all parts thoroughly with warm water or distilled water and air dry, if possible.

3. To clean recirculation ports and tubing, simply pump distilled water into the tubing torinse. Thoroughly empty tubing of liquid before use.

4.2 Compatible Cleaning AgentsChemically compatible cleaners must be used to ensure long life of parts. These include:

• Aqueous solutions of soaps and mild detergents:Bio-Rad Cleaning Concentrate (catalog number 161-0722)Dishwashing Liquid

• Organic Solvents:HexaneAliphatic Hydrocarbons

Do not leave plastic parts to soak in detergents more than 30 minutes. A short detergentrinse typically is all that is required.

Caution: Do not use the following chemicals to clean Sub-Cell parts. Exposure to thesechemicals may cause the plastic parts to crack, craze, etch or warp.

• Chlorinated HydrocarbonsCarbon TetrachlorideChloroform

• Aromatic HydrocarbonsBenzenePhenolTolueneMethyl Ethyl KetoneAcetone

• AlcoholsMethanolEthanolIsopropyl Alcohol

16

Do not use abrasive or highly alkaline cleaners on Sub-Cell parts. Do not expose Sub-Cellparts to temperatures >60 °C. Do not sterilize Sub-Cell parts by autoclaving or dry heat.

4.3 Maintenance ScheduleItem Look For Frequency Action

All Parts Dried salts, agarose, Each Use Clean parts as grease, and dirt described in Section 4.1

Electrical cables Breaks or fraying Each Use Replace Cables

Trays Chips or cracks Each Use Replace Tray

Electrode Wires Breaks Each Use See Section 4.4(Electrode CassetteReplacement)

Cable Connections Looseness Weekly Replace Banana Jacks(Banana Jacks or Banana Plug Holdersand Plugs)

Base Crazing, cracks, Monthly Replace Baseor leaks

Recirculation Looseness or cracks Each Use Tighten or ReplaceTubing

4.4 Electrode ReplacementThe Sub-Cell systems allow easy, hassle-free replacement of broken electrode wires by simply

removing the banana plug/electrode wire assembly and ordering a new assembly from Bio-Rad(Figure 4.1). See Ordering Information (Section 6.2) for catalog numbers and part descriptions.

Instructions

1. Remove the thumb screw from the side wall of the base to release the banana plug/electrodewire assembly. Do not discard this thumb screw (keep this screw with the base).

2. Remove the broken wire assembly from the base and discard the broken assembly.

3. Insert the new electrode assembly ensuring the electrode wire guard is properly seated intothe electrode wire guard slot in the bottom of the base.

4. Replace and tighten the thumb screw to secure the assembly in the base.

Figure 4.1 Replacement of banana plug/electrode wire assembly.

17

4.5 RNase DecontaminationSub-Cell parts can be cleaned with a mild detergent and treated for 10 minutes with 3%

hydrogen peroxide (H2O

2) and then rinsed with 0.1% DEPC- (diethyl pyrocarbonate) treated

distilled water to eliminate RNases prior to using the Sub-Cell systems for RNA gels1-2. Donot soak Sub-Cell parts in DEPC water. Consult references 1-2 for other suggestions regardingthe use of DEPC in RNase decontamination.

Caution: DEPC is a suspected carcinogen. Always wear gloves, eye glasses and a laboratory coat. Use caution when handling DEPC-containing solutions. Consult theDEPC MSDS for more information.

Do not attempt to RNase decontaminate Sub-Cell parts using extreme dry heat.

Note: Several commercial products are also available for eliminating RNase contamination.RNaseZAP™ (Ambion) or RNase AWAY™ (Molecular Bio-Products) are safe, simple andeffective methods that if used properly do not craze or fog the Sub-Cell parts. See manufacturer instructions for proper use.

18

Section 5Troubleshooting

Symptoms Probable Causes Solutions

Slanted lanes (bands)

Curved line or distortion oflanes (bands)

Differential relative mobilities

Curved bands, smiles

Ragged bands

Band smearing and streaking

Bands sharp but too few bandsseem

High MW bands sharp/Low MWbands smeared

Gels crack

• Gel not fully solidified.

• Comb warped or at an angle.

• Bubbles in sample wells.

• Sample spilled out of wells.

• Unit not leveled.

• Gel floated or slid off tray.

• Sample overload.

• Temperature or pH buffer gradients

• Sample density incorrect.

• Sample well deformed.

• Excessive power or heating.

• Agarose has improper endos-mosis (-m

r).

• Salt concentration in sampletoo high.

• Excessive power and heating.

• Sample spilled out of well.

• Incomplete digest, nucleasecontamination, bad enzyme.

• Sample wells cast through thegel. Sample leaks along bot-tom of running surface.

• Sample overload.

• Too high gel percentage.

• Incomplete digest.

• Gel percentage too low.

• Too high voltage gradientespecially with low meltingtemperature agarose or low gelstrength gels.

• Let gel solidify for at least 30-60 minutes.

• Check alignment of comb.

• Remove bubbles prior to electrophoresis.

• Samples should have proper density. Apply carefully.

• Level unit. Place on steadywork bench.

• Recirculate at a rate of 300-500 ml/min.

• Reduce the amount of sampleloaded.

• Reduce load.• Add more buffer.• Recirculate buffer.

• See sample application instructions.

• Carefully remove comb, espe-cially from soft gels. Be suregel has solidified. Cooling softgels aids in comb removal.Add buffer to help lubricateremoval of the comb.

• Reduce voltage. See elec-trophoresis instructions.

• Consult Bio-Rad aboutagarose.

• Reduce salt concentration to ≤ 0.1 M.

• Reduce voltage. See elec-trophoresis instructions

• Take care in applying sample.Increase gel thickness for largesample volumes.

• Heat sample. Check enzymeactivity. Digest sample further.

• Comb should be placed 1 to 2 mm above the base of the run-ning surface. Add buffer to helplubricate removal of the comb.

• Dilute sample.

• Lower gel percentage.

• Check enzyme activity, digestfurther.

• Increase gel percentage.• Switch to polyacrylamide.

• Reduce voltage. Run gel atlower temperature.

19

Section 6Ordering Information

6.1 Sub-Cell Model 96 and Model 192 SystemsCatalogNumber Product Description

Sub-Cell Model 96 Systems170-4540 Sub-Cell Model 96/PowerPac 300 System, 100/120 V

170-4542 Sub-Cell Model 96/PowerPac 300 System, 220/240 V

170-4500 Sub-Cell Model 96, with 25 x 10 cm tray and Gel Caster


170-4502 Sub-Cell Model 96, with 25 x 10 cm tray


Sub-Cell Model 192 Systems170-4541 Sub-Cell Model 192/PowerPac 300 System, 100/120 V

170-4543 Sub-Cell Model 192/PowerPac 300 System , 220/240 V








170-4511 Sub-Cell Model 192 , with 25 x 25 cm tray

6.2 Sub-Cell Model 96 and Model 192 Systems AccessoriesCatalogNumber Product Description

Sub-Cell Model 96 Accessories170-4512 Sub-Cell Model 96 Base

170-4513 Sub-Cell Model 96 Safety Lid , with cables

170-4514 Model 96 Gel Caster

170-4518 Electrode Assembly (Anode) – Red

170-4519 Electrode Assembly (Cathode) – Black

170-4520 Gel Casting Gates

170-4521 UV Transparent Tray, 25 x 10 cm


170-4525 Comb Holder

170-4537 Buffer Recirculation Kit

20

CatalogNumber Product Description

Sub-Cell Model 192 Accessories170-4515 Sub-Cell Model 192 Base

170-4516 Sub-Cell Model 192 Safety Lid , with cables

170-4517 Model 192 Gel Caster

170-4518 Electrode Assembly (Anode) – Red

170-4519 Electrode Assembly (Cathode) – Black

170-4520 Gel Casting Gates





170-4525 Comb Holder

170-4537 Buffer Recirculation Kit

Sub-Cell Model 96 and 192 Comb SpecificationsCatalog Well Thickness Well Width Well VolumeNumber Number (mm) (mm) Capacity* (µl)

170-4526 26 0.75 6.0 22.50

170-4527 26 1.5 6.0 45.00

170-4528 51 0.75 3.0 11.25

170-4529 51 1.5 3.0 22.50

170-4530 2 0.75 97 364.0

170-4530 4 0.75 46 172.5

170-4531 2 1.5 97 727.5

170-4531 4 1.5 46 345.0

All Sub-cell Model 96 and Model 192 combs require a comb holder (170-4525)

* Well volume capacity determined based on 0.5 cm thick gel

6.3 Related Bio-Rad ProductsContact your local Bio-Rad representative concerning the following products for nucleic

acid electrophoresis and blotting.


Sub-Cell GT Systems170-4400 Sub-Cell GT System




21


Sub-Cell GT Systems (continued)

170-4404 Sub-Cell GT System, 15 x 25 cm tray

170-4405 Wide Mini-Sub Cell GT System

170-4406 Mini-Sub Cell GT System

170-4481 Sub-Cell GT System, with 15 x 10 cm tray and Gel Caster




170-4485 Wide Mini-Sub Cell GT System and Gel Caster

170-4486 Mini-Sub Cell GT System and Gel Caster

Power Supplies165-5050 PowerPac 300 Power Supply, 100/120 V


Buffer Recirculation Pump Systems170-2929 Buffer Recirculating Pump, 120/100 V

170-2930 Buffer Recirculating Pump, 220/240 V

Zeta-Probe® Positively-Charged Nylon Blotting Membranes161-0153 Sheets, 9 x 12 cm, 15

161-0154 Sheets, 10 x 15 cm, 15

161-0155 Sheets, 15 x 15 cm, 15

161-0156 Sheets, 15 x 20 cm, 15

161-0157 Sheets, 20 x 20 cm, 15

161-0158 Sheets, 20 x 25 cm, 3

161-0159 Roll, 30 cm x 3.3 m, 1

161-0165 Roll, 20 cm x 3.3 m, 1

Zeta-Probe GT (Genomic Tested) Positively Charged NylonBlotting Membranes161-0190 Sheets, 9 x 12 cm, 15

161-0191 Sheets, 10 x 15 cm, 15

161-0192 Sheets, 15 x 15 cm, 15

161-0193 Sheets, 15 x 20 cm, 15

161-0194 Sheets, 20 x 20 cm, 15

161-0195 Sheets, 20 x 25 cm, 3

161-0196 Roll, 30 cm x 3.3 m, 1

161-0197 Roll, 20 cm x 3.3 m, 1

22


Supported Nitrocellulose Membrane (0.45 micron)161-0090 Sheets, 7 x 8.4 cm, 10

161-0091 Sheets, 10 x 15 cm, 10

161-0092 Sheets, 15 x 15 cm, 10

161-0093 Sheets, 20 x 20 cm, 10

161-0094 Roll, 30 cm x 3 m, 1

Supported Nitrocellulose Membrane (0.20 micron)161-0095 Sheets, 7 x 8.4 cm, 10

161-0096 Sheets, 15 x 15 cm, 10

161-0097 Roll, 30 cm x 3 m, 1

Vacuum Blotting Apparatus165-5000 Model 785 Vacuum Blotter

165-5001 Model 785 Vacuum Blotter System, 120 VAC

165-5002 Model 785 Vacuum Blotter System, 220/240 VAC

Semi-Dry Transfer Cells170-3940 Trans-Blot ® SD Semi-Dry Electrophoresis Transfer Cell



UV Crosslinking Chamber165-5031 GS Gene Linker ® UV Chamber, 120 VAC




Gel Reagents162-0019 Low Melt Preparative Grade Agarose, 100 g

162-0133 Molecular Biology Certified Agarose, 500 g

162-0126 High Strength Analytical Grade Agarose, 500 g

162-0144 Amplisize Agarose, 50 g

170-8200 AmpliSize ® DNA Size Standard, 50-2,000 bp

170-8210 DNA Size Standard, 1-4.2 Kb ladder

170-8220 DNA Size Standard, 0.7-8.4 Kb

161-0404 Bromophenol Blue, 10 g

161-0423 Xylene Cyanole FF, 25 g

161-0433 Ethidium Bromide Solution, 10 ml, 10 mg/ml

23


Electrophoresis Buffers161-0733 10x Tris/Boric Acid/EDTA (TBE), 1 L

161-0743 50x Tris/Acetic Acid/EDTA (TAE), 1 L

161-0719 Tris, 1 kg

161-0751 Boric Acid, 1 kg

161-0729 EDTA, 500 g

DNA Gel Image Analysis and Documentation Systems170-3742 Standard Polaroid ® Documentation System, 120 VAC

170-3746 Standard Polaroid Documentation System, 100 VAC

170-3747 Standard Polaroid Documentation System, 220/240 VAC

170-7520 Gel Doc ® 1000 UV Gel Documentaion System-PC, 100 VAC

170-7521 Gel Doc 1000 UV Gel Documentaion System-PC, 120 VAC

170-7522 Gel Doc 1000 UV Gel Documentaion System-PC, 220/240 VAC

170-7525 Gel Doc 1000 UV Gel Documentaion System-Mac, 100 VAC

170-7522 Gel Doc 1000 UV Gel Documentaion System-Mac, 120 VAC

170-7522 Gel Doc 1000 UV Gel Documentaion System-Mac, 220/240 VAC

24

Section 7References

1. Sambrook, Fritsch, and Maniatis, Molecular Cloning, A Laboratory Manual, Second Edition, ColdSpring Harbor Laboratory Press, 1989

2. Current Protocols in Molecular Biology, Greene Publishing Associates and Wiley-Interscience,1989

Additional Reading1. Kopchick, J.J., Cullen, B.R. and Stacey, D.W., Anal. Biochem, 115, 419 (1981).

2. Southern, E., Methods in Enzymol., Academic Press, N.Y., 68, 152 (1979).

3. The Bio-Rad Silver Stain – Bulletin 1089.

4. Bittner, M., Kupferer, P. and Morris, C.F., Anal. Biochem., 102, 459 (1980).

5. Bio-Rad Trans-Blot Cell Operation Instructions – Bulletin 1082.

6. Winberg, G. and Hammarskjold, M.L., Nucleic Acids Res., 8, 253 (1980).

7. Jytatekadze, T.V., Axelrod, V.D., Gorbulev, V.G., Belzhelarskaya, S.N. and Vartikyan, R.M., Anal. Biochem., 100, 129 (1979).

8. Dretzen, G., Bellard, M., Sassone-Corsi, P. and Chambon, P., Anal. Biochem., 112, 295 (1981).

25

Australia, Bio-Rad Laboratories Pty Limited, Block Y Unit 1, Regents Park Industrial Estate, 391 Park Road, Regents Park, NSW 2143 • Phone 02-805-5000 • Fax 02-805-1920Austria, Bio-Rad Laboratories Ges.m.b.H., Auhofstrasse 78D, 1130 Wien • Phone (1) 877 89 01 • Fax (1) 876 56 29Belgium, Bio-Rad Laboratories S.A./N.V., Begoniastraat 5, 9810 Nazareth Eke • Phone 09-385 55 11 • Fax 09-385 65 54Canada, Bio-Rad Laboratories (Canada) Ltd., 5671 McAdam Road, Mississauga, Ontario L4Z 1N9 • Phone (905) 712-2771 • Fax (905) 712-2990China, Bio-Rad Laboratories, 14, Zhi Chun Road, Hai Dian District, Beijing 100088 • Phone (01) 2046622 • Fax (01) 2051876Denmark, Bio-Rad Laboratories, Symbion Science Park, Fruebjergvej 3, DK-2100 Copenhagen • Phone 39 17 9947 • Fax 39 27 1698Finland, Bio-Rad Laboratories, Business Center Länsikeskus, Pihatörmä 1A SF-02240, Espoo, • Phone 90 804 2200 • Fax 90 804 1100France, Bio-Rad S.A., 94/96 rue Victor Hugo, B.P. 220, 94 203 Ivry Sur Seine Cedex • Phone (1) 49 60 68 34 • Fax (1) 46 71 24 67Germany, Bio-Rad Laboratories GmbH, Heidemannstraße 164, D-80939 München/Postfach 450133, D-80901 München • Phone 089 31884-0 • Fax 089 31884-100India, Bio-Rad Laboratories, C-248 Defence Colony, New Delhi 110 024 • Phone 91-11-461-0103 • Fax 91-11-461-0765Italy, Bio-Rad Laboratories S.r.l.,Via Cellini, 18/A, 20090 Segrate Milano • Phone 02-21609 1 • Fax 02-21609-399Japan, Nippon Bio-Rad Laboratories, 7-18, Higashi-Nippori 5-Chome, Arakawa-ku, Tokyo 116 • Phone 03-5811-6270 • Fax 03-5811-6272The Netherlands, Bio-Rad Laboratories B. V., Fokkerstraat 10, 3905 KV Veenendaal • Phone 0318-540666 • Fax 0318-542216New Zealand, Bio-Rad Laboratories Pty Ltd., P. O. Box 100-051, North Shore Mail Centre, Auckland 10 • Phone 09-443 3099 • Fax 09-443 3097Pacific, Bio-Rad Laboratories, Unit 1111, 11/F., New Kowloon Plaza, 38, Tai Kok Tsui Road, Tai Kok Tsui, Kowloon, Hong Kong • Phone 7893300 • Fax 7891257Singapore, Bio-Rad Laboratories (Singapore) Ltd., 221 Henderson Rd #05-19, Henderson Building, Singapore 0315 • Phone (65) 272-9877 • Fax (65) 273-4835Spain, Bio-Rad Laboratories, S. A. Avda Valdelaparra 3, Pol. Ind. Alcobendas, E-28100 Alcobendas, Madrid • Phone (91) 661 70 85 • Fax (91) 661 96 98Sweden, Bio-Rad Laboratories AB, Gärdsvägen 7D, Box 1276, S-171 24 Solna • Phone 46-(0)8-735 83 00 • Fax 46-(0)8-735 54 60Switzerland, Bio-Rad Laboratories AG, Kanalstrasse 17, Postfach, CH-8152 Glattbrugg • Phone 01-809 55 55 • Fax 01-809 55 00United Kingdom, Bio-Rad Laboratories Ltd., Bio-Rad House, Maylands Avenue, Hemel Hempstead, Herts HP2 7TD • Free Phone 0800 181134 • Fax 01442 259118

Molecular Bioscience Group

2000 Alfred Nobel DriveHercules, California 94547Telephone (510) 741-1000Fax: (510) 741-5800

SIG 101295 Printed in USA M1704498 Rev A


electrophoresis tech note 1156

Paul Menter, Bio-Rad Laboratories, 2000 Alfred Nobel Drive, Hercules, CA 94547 USA

IntroductionThe unparalleled resolution and flexibility possible with polyacrylamide gel electrophoresis (PAGE) has led to its widespread use for the separation of proteins and nucleicacids. Gel porosity can be varied over a wide range to meetspecific separation requirements. Electrophoresis gels andbuffers can be chosen to provide separation on the basis ofcharge, size, or a combination of charge and size.

The key to mastering this powerful technique lies in the polymerization process itself. By understanding the importantparameters, and following a few simple guidelines, the novicecan become proficient and the experienced user can optimizeseparations even further.

This bulletin takes a practical approach to the preparation ofpolyacrylamide gels. Its purpose is to provide the informationrequired to achieve reproducible, controllable polymerization.For those users interested only in the “bare essentials,” thePolymerization Protocols can be used as a laboratory guide.

Mechanism of PolymerizationPolyacrylamide gels are formed by copolymerization of acrylamide and bis-acrylamide (“bis,” N,N'-methylene-bis-acrylamide). The reaction is a vinyl addition polymerization initiated by a free radical-generating system (Chrambach 1985).Polymerization is initiated by ammonium persulfate and TEMED(tetramethylethylenediamine): TEMED accelerates the rate offormation of free radicals from persulfate and these in turncatalyze polymerization. The persulfate free radicals convertacrylamide monomers to free radicals which react with unactivated monomers to begin the polymerization chainreaction (Shi and Jackowski 1998). The elongating polymerchains are randomly crosslinked by bis, resulting in a gel witha characteristic porosity which depends on the polymerizationconditions and monomer concentrations.

Riboflavin (or riboflavin-5'-phosphate) may also be used as asource of free radicals, often in combination with TEMED andammonium persulfate. In the presence of light and oxygen,riboflavin is converted to its leuco form, which is active in initiating polymerization. This is usually referred to as photochemical polymerization.

Acrylamide Polymerization — A Practical Approach

Polyacrylamide Gel Polymerization

Purity of Gel-Forming ReagentsAcrylamide

Gel-forming reagents include the monomers, acrylamide and bis,as well as the initiators, usually ammonium persulfate andTEMED or, occasionally, riboflavin and TEMED. On a molarbasis, acrylamide is by far the most abundant component in themonomer solution. As a result, acrylamide may be the primarysource of interfering contaminants (Dirksen and Chrambach1972). Poor-quality acrylamide contains significant amounts ofthe following contaminants:

1. Acrylic acid — Acrylic acid is the deamidation product of acrylamide. Acrylic acid will copolymerize with acrylamide and bis, thereby conferring ion exchange properties on theresulting gel. This can lead to local pH changes in the gel andcause artifacts such as aberrant relative mobility, precipitation ofsome proteins and nucleic acids, streaking or smearing ofbands, and run-to-run irreproducibility. In acrylamide, acrylic acidshould be below 0.001% (w/w). This is determined by directtitration, and supported by both conductivity and pH measurement.

2. Linear polyacrylamide — Contaminants with catalytic propertiesmay cause what appears to be autopolymerization during theproduction, processing, or storage of marginally pure acrylamide.This results in the presence of linear polyacrylamide in the drymonomer. Linear polyacrylamide will affect polymerization, since itserves as a nucleus for polymerization. The most important effectis the loss of reproducibility in gel porosity and relative mobilitiesof proteins and nucleic acids. Linear polyacrylamide is detectedas water or alcohol insolubles and should be <0.005% (w/w).

CH2 CH

C O

NH2

CH2 CH

C O

NH

Acrylamide PolyacrylamideBis

+

CH2

NH

CHCH2

CH2 CH

C O

NH2

CH2 CH

C O

NH

CH2

NH

C O

CH

C O

CH2 CH

C O

NH2

CH2 CH

C O

NH2

CH2CH2 CH

C O

NH2

3. Ionic contaminants — Ionic contaminants can include bothinhibitors and accelerators of polymerization. Aside fromacrylic acid, the most notable ionic contaminants are metalssuch as copper, which can inhibit gel polymerization. Metals can also poison enzymes, alter the relative mobility of metal binding proteins such as calmodulin, and inhibitdigestion of electrophoretically purified nucleic acids byrestriction and modification enzymes. Ionic contaminants are detected indirectly by their effects on chemical and photochemical polymerization, and by the conductivity ofmonomer solutions.

bis-Acrylamide

Bis is present in much smaller quantities than acrylamide in monomer solutions. However, improperly purified bis contains some of the same contaminants as acrylamide.These include products of autopolymerization and ionic contaminants, which have the same deleterious effects, andcan be detected in the same ways, as the correspondingacrylamide contaminants.

Initiators

Chemical polymerization is initiated by ammonium persulfate,while photochemical polymerization is initiated by riboflavin (or riboflavin-5'-phosphate), or by a combination of riboflavinand ammonium persulfate. Initiation and polymerization are catalyzed by TEMED. Because polymerization is initiated bythe generation of free radicals from persulfate or riboflavin, it is not surprising that these compounds are reactive, andprone to oxidation or decomposition. The contaminants ofthe initiators tend to be the products of their own breakdownas well as other contaminating compounds.

TEMED is subject to oxidation, which causes the gradual loss of catalytic activity. This process is greatly accelerated by contaminating oxidizing agents. TEMED that containsoxidation products is characterized by a yellow color. Thepractical consequences of the oxidative process are therequirement for greater amounts of TEMED to achieve adequate polymerization, and a gradual loss of TEMED reactivity with time. TEMED is also very hygroscopic and willgradually accumulate water, which will accelerate oxidativedecomposition. TEMED with maximum activity and shelf life is obtained by redistillation immediately prior to bottling, resulting in a product that is clear, water free, and greaterthan 99% pure (14.4 M).

Ammonium persulfate is also very hygroscopic. This propertyis particularly important, since ammonium persulfate beginsto break down almost immediately when dissolved in water.Therefore, the accumulation of water in ammonium persulfateresults in a rapid loss of reactivity. This is why ammoniumpersulfate solutions should be prepared fresh daily. Persulfateis consumed in the polymerization reaction. Excess persulfatecan cause oxidation of proteins and nucleic acids. This oxidation problem can be avoided if inhibitor-free gel-formingreagents are used, and ammonium persulfate is used at therecommended levels.

Contaminants in Buffers

Contaminants in buffer reagents (Tris, borate, acetate, glycine, etc.), gel additives (SDS, urea, etc.), and laboratorywater can have a profound effect on polymerization. Themost common contaminants of these reagents are metals,non-buffer ions, and breakdown products. The most frequenteffect of these contaminants is to inhibit polymerization. Whenpolymerization is partially inhibited, the resulting gel will havegreater porosity than intended, and molecules will havegreater mobilities. Furthermore, control over polymerizationreproducibility is compromised.

Initiator Type and ConcentrationInitiators are the effectors of polymerization. Of course, the rateof polymerization depends on the concentration of initiators,but more importantly, the properties of the resulting gel alsodepend on the concentration of initiators. Increasing the concentration of initiators (e.g., ammonium persulfate andTEMED) results in a decrease in the average polymer chainlength, an increase in gel turbidity, and a decrease in gel elasticity. In extreme cases, excess initiator can produce a gelsolution that does not appear to polymerize at all. This is dueto the formation of polymer chains so short that visible gelation does not take place and the polymer stays in solution.The only indication that a reaction has taken place is anincrease in viscosity.

Excess ammonium persulfate and TEMED have other effects, including oxidation of sample proteins (especiallysulfhydryl-containing compounds) and changes in buffer pH.Excess TEMED can increase buffer pH, react with proteins(Dirksen and Chrambach 1972; Chrambach et al. 1976), and alter the banding pattern (Gelfi and Righetti 1981a).Ammonium persulfate acts as a buffer between pH 8 and 9.Potassium persulfate is recommended instead of ammoniumpersulfate in weakly buffered basic systems (~pH 9). Excess riboflavin may cause the oxidation of some compounds, especially sulfhydryl-containing compounds(Dirksen and Chrambach 1972), and can denature proteins(Righetti et al. 1981).

Reducing the concentration of initiators results in longer polymer chain lengths, lower turbidity, and greater elasticity.These are desirable properties. However, lower initiator concentrations also mean slower polymerization. If polymeri-zation is too slow, oxygen will begin to enter the monomersolution and inhibit polymerization, resulting in gels which aretoo porous and mechanically weak. Inhibition will be especiallypronounced at surfaces exposed to air, or at the surfaces of combs and spacers, which appear to trap air at their surfaces. The remaining unpolymerized monomer can reactwith alpha amino, sulfhydryl, and phenolic hydroxyl groups ofproteins (Allison et al. 1974; Chrambach et al. 1976; Dirksenand Chrambach 1972).

For discontinuous systems which employ a stacking gel (e.g.,Laemmli system), optimal polymerization of the overlaid lowergel (resolving gel) is achieved when visible gelation takesplace 15–20 min after the addition of the initiators ammoniumpersulfate and TEMED (note that polymerization continueslong after visible gelation takes place; see Figure 1). Forstacking gels and continuous systems (which do not containstacking gels) — i.e., any gel which is not overlaid — optimalpolymerization results when visible gelation takes place in 8–10 min. Higher initiator concentrations and faster polymerization are required in these cases because of theinhibitory effect of atmospheric oxygen associated with thecomb. In any case, conversion of monomer to polymershould be greater than 95%. If gelation takes longer than20 min, the inhibitory effects of atmospheric oxygen will begin to appear.

As a general rule, use the lowest catalyst concentrations thatwill allow polymerization in the optimal period of time. In thecase of ammonium persulfate/TEMED-catalyzed reactions, for example, approximately equimolar concentrations of bothcatalysts in the range of 1 to 10 mM are recommended.

Riboflavin is often used as an initiator along with TEMED, orwith TEMED and ammonium persulfate. The major advantageof riboflavin is that it is active in very low concentrations (~5–10 µg/ml). Thus, when riboflavin is used with TEMED andammonium persulfate, the total amount of initiator required(sum of the three initiators) is less. Given the possible effectsof initiators on buffer pH, riboflavin-based initiator systems areuseful for poorly buffered systems such as electrofocusinggels, in which the only buffering components are ampholytes.

Visible gelation takes longer in riboflavin-based initiator systems, usually 30–60 min. Oxygen does not have the dramatic inhibitory effect on riboflavin-based initiator systemsthat it has on TEMED/ammonium persulfate systems. This ispresumably due to the oxygen-scavenging property ofriboflavin. As a result, longer gelation time can be tolerated.

In chemical polymerization, visible gelation occurs in 15–20 min and polymerization is essentially complete in 90 min. In photochemical polymerization, however, visiblegelation takes 30–60 min and complete polymerizationrequires up to 8 hr (Righetti et al. 1981). Shorter times lead to more porous and elastic gels, increased risk of proteinmodification, and pore size irreproducibility.

TemperatureTemperature control is critical for reproducibility of acrylamidepolymerization. Temperature has a direct effect on the rate of gel polymerization; the polymerization reaction is alsoexothermic. Consequently, the generated heat drives thereaction more quickly. Thus, gelation usually occurs very rapidly once polymerization begins.

Temperature also affects the properties of the gel (Chen andChrambach 1979). For example, polymerization at 0–4°Cresults in turbid, porous, inelastic gels, and reproducibility isdifficult to achieve. These properties may be due to increasedhydrogen bonding of monomer at low temperatures. Gelspolymerized at 25°C are more transparent, less porous, andmore elastic. However, if the polymerization temperature istoo high, short polymer chains are formed and the gels are inelastic. This is thought to be due to increased polymerchain termination at higher temperatures.

A temperature of 23–25°C is optimal (as well as most convenient) for polymerization. It is important that the monomersolution and the gel mold (e.g., glass plates or tubes) be atthe optimal temperature when the gel is poured. Furthermore,reproducibility is dependent on using the same temperatureeach time gels are poured.

Since monomer solutions are usually stored at 4°C along withbuffer concentrates, it is important to allow the monomer gelsolution, once prepared, to equilibrate to room temperaturebefore being evacuated (if cold solutions are placed under vacuum they tend to stay cold).

OxygenThe formation of polyacrylamide gels proceeds via free radicalpolymerization. The reaction is therefore inhibited by any element or compound that serves as a free radical trap(Chrambach 1985). Oxygen is such an inhibitor. Oxygen,present in the air, dissolved in gel solutions, or adsorbed tothe surfaces of plastic, rubber, etc., will inhibit, and in extremecases prevent, acrylamide polymerization. Proper degassingis critical for reproducibility. Therefore, one of the most important steps in the preparation of polyacrylamide gels isthe evacuation, or “degassing” of gel solutions immediatelyprior to pouring the gel. This is done by placing the flask ofgel solution in a vacuum chamber or under a strong aspirator.In some cases, a vacuum pump may be required.

Buffer stock solutions and monomer stock solutions are usuallystored at 4°C. Cold solutions have a greater capacity for dissolved oxygen. The process of degassing is faster andmore complete if the gel solution is brought to room temperature (23–25°C)‚ before degassing begins.Furthermore, if a cold gel solution is placed under vacuum,the process of evacuation tends to keep the solution cold.Pouring a gel with a cold solution will have a substantial negative effect on the rate of polymerization and on the quality of the resulting gel.

Polymerization in which riboflavin is used as one of the initiators calls for degassing. The conversion of riboflavin fromthe flavo to the leuco form (the species active in initiation)actually requires a small amount of oxygen (Gordon 1973).

Gel AdditivesThe most common gel additives include SDS (sodium dodecyl sulfate), Triton* X-100 detergent, and chaotropicagents such as urea and formamide. Detergents can beadded to most common buffer systems without significantlyaffecting polymerization. Agents such as urea and formamide,however, cause the formation of smaller pore-size gels thanwould be formed in their absence (urea is often a componentof gel systems used to separate small proteins and peptides).This may be due to the disruption of hydrogen bondsbetween monomer molecules during polymerization. Smallerpore size may also be achieved at higher polymerization temperatures, an effect also attributed to hydrogen bond disruption. Contaminants of gel additives can affect polymerization. Nonionic additives such as urea, formamide,and Triton X-100 can be deionized with a mixed-bed ionexchange resin. Use 10 gm Bio-Rad Ag 5O1 X-8 resin per100 ml additive solution and let sit overnight. However,removal of nonionic contaminants from nonionic reagents is not practical. Therefore, all additives should be quality-assured for electrophoresis.

TimeAlthough visible gelation occurs in 15–20 min for chemical polymerization and 30–60 min for photochemicalpolymerization, polymerization continues much longer (seeFigure 1). Ammonium persulfate/TEMED-initiated reactionsshould be allowed to proceed for 2 hr to ensure maximumreproducibility in gel pore size. Photochemical polymerization(riboflavin-based initiator system) usually proceeds more slowly than chemical polymerization, and is also dependenton light intensity (Shi and Jackowski 1998). However,riboflavin is usually used for polymerization of electrofocusinggels in which separation is based on charge, and for whichgel porosity is of secondary importance. Thus these gels canbe used shortly after visible gelation without being affected byslight variations in porosity.

Monomer ConcentrationThe practical range for monomer concentration is between3%T and 30%T, where %T refers to % (w/v) of total monomer(acrylamide + bis) in solution. A higher concentration ofmonomer results in faster polymerization. Therefore, changingfrom 5% gels to 30% gels will probably allow a reduction of20–50% in the concentration of initiators.

Polymerization ProtocolsThere are 2 major initiator formulations for acrylamide polymerization. The first, for chemical polymerization, is usedfor SDS-PAGE and DNA sequencing. Chemical polymerizationemploys ammonium persulfate and TEMED as initiators. The second, for photochemical polymerization, is used primarily for horizontal electrofocusing gels. Photochemicalpolymerization calls for riboflavin as well as ammonium persulfate and TEMED. Riboflavin phosphate can be substituted for riboflavin. Riboflavin phosphate is often preferred for its greater solubility.

This explains why polymerization initiated primarily by riboflavincan be completely blocked by exhaustive degassing. However,oxygen in excess of that needed to convert riboflavin to theactive form will inhibit polymer chain elongation, as it does inreactions initiated only by ammonium persulfate and TEMED.Thus, while degassing is still important for limiting inhibition, it must not be so extensive that it prevents conversion ofriboflavin to the active form. For polymerization initiated byriboflavin/TEMED, or riboflavin/TEMED/ammonium persulfatesystems, degassing should not exceed 5 min.

A consequence of the interaction of riboflavin with oxygen isthat riboflavin seems to act as an oxygen scavenger. This issupported by the observation that the addition of riboflavin (5 µg/ml) to stacking gel solutions containing ammonium persulfate/TEMED initiators results in cleaner, more uniformpolymerization at gel surfaces exposed to oxygen (such ascombs). The same effect could likely be achieved by morethorough degassing of solutions without riboflavin.

Whether using chemical polymerization (ammonium persulfate/TEMED) or photochemical polymerization(riboflavin/TEMED or riboflavin/TEMED/ammonium persulfateinitiators), reproducible gel quality and separation characteristics require careful attention to gel solution temperature before degassing, and to degassing time, temperature, and vacuum. These parameters should be kept constant every time gels are prepared.

pHThe majority of electrophoresis systems are buffered at neutral or basic pH, at which the common initiators, ammonium persulfate, TEMED, and riboflavin, are effective. Riboflavin is the better choice for polymerization at low pH(Shi and Jackowski 1998); however, at low pH, TEMED maybecome protonated. This can result in slower polymerization,since the free base form of TEMED is required for initiation.For acidic buffer systems, alternative initiator systems aresometimes used (Andrews 1990).

Alternative CrosslinkersPDA (piperazine di-acrylamide), a crosslinking agent that canbe substituted for bis in polyacrylamide gels, offers severaladvantages for electrophoresis. These include reduced background for silver staining, increased gel strength, andhigher-resolution gels. PDA can be substituted for bis on aweight basis without changing polymerization protocols.

Crosslinkers other than bis and PDA may be used for specialized purposes, the most common of which is gel solubilization during post-electrophoresis recovery of proteinsor nucleic acids. These crosslinkers include DATD (diallyl-tartardiamide), DHEBA (dihydroxyethylene-bis-acrylamide),and BAC (bis-acrylylcystamine). Alternative crosslinkers maybe more or less reactive in polymerization than bis. Therefore,some adjustment in the concentration of initiators may benecessary to achieve optimal polymerization. For a discussionof alternative crosslinkers, see Gelfi and Righetti (1981b).

Preparation for Polymerization1. Prepare 10% ammonium persulfate shortly prior to use

(prepare fresh daily). TEMED is used undiluted. Prepare 0.1% riboflavin (or riboflavin phosphate, which is more soluble)if photopolymerization will be performed.

2. Combine buffer stock solution, monomer stock solution, and water in the appropriate proportions in an Erlenmeyer flask.Since stock solutions are usually stored at 4°C, the gel solutionshould be allowed to warm to room temperature beforedegassing.

3. Prepare the gel casting mold, i.e., plates, spacers, and clampsfor gel casting. Be sure they are neither hot nor cold.

4. Once the gel solution is prepared and brought to room temperature (23–25°C), degas the solution under a vacuum of125 torr or better for 15 min at room temperature (for systemsin which constant agitation is used during degassing, 10 min is sufficient). Longer periods of degassing are generally notdeleterious, although long degassing will result in somewhatfaster polymerization.

Chemical Polymerization in Discontinuous Systems — Lower (Resolving) GelIn a discontinuous system, such as that of Laemmli, the resolving gel is polymerized first. Then, the stacking gel is caston top of the resolving gel. Use the following protocol to prepare resolving gels for all discontinuous systems (see the next section for preparation of stacking gels).

Initiator volume per Initiator final10 ml gel solution concentration

Ammonium persulfate (10% w/v) 50 µl 0.05%

TEMED (undiluted) 5 µl 0.05%

Swirl the solution gently but thoroughly. Holding the flask by the neck with one hand, swirl it 8 to 10 cycles. This mixes the initiators completely without introducing too much oxygen.Swirling too little can result in uneven polymerization.

Cast the gel by introducing the monomer solution into the gelmold in a steady stream to minimize the introduction of oxygen.Overlay the monomer solution using water, isoamyl alcohol, or water-saturated isobutyl alcohol to exclude oxygen from the surface.

Allow polymerization to occur at room temperature at least 90 min prior to use (see Figure 1).

Chemical Polymerization in Continuous Systems andStacking GelsContinuous systems consist of a single gel. Continuous systems are used for some types of protein electrophoresis, and for DNA sequencing. Stacking gels are part of discontinuoussystems. These gels have in common contact with the well-forming comb and greater exposure to molecular oxygen atthe surface. Use the following levels of initiators for continuoussystems and stacking gels.


Ammonium persulfate (10% w/v) 50 µl 0.05%


Swirl the solution gently but thoroughly.

Cast the gel and insert the well-forming comb without trappingair under the teeth.

Allow polymerization to occur at room temperature at least 90 min prior to use (see Figure 1).

Photochemical PolymerizationThis protocol is recommended for isoelectric focusing(IEF) gels. Since molecules remain in the IEF gel during electrophoresis, excess ions from initiators can cause distortion ofbands. Photochemical initiation is recommended for IEF gelsbecause it is effective at low initiator concentrations.

1. When the gel solution is prepared and brought to room temperature (23–25°C), degas the solution under a vacuum of125 torr or better for 2 min at room temperature (for systems inwhich constant agitation is used for degassing, 1 min is sufficient).

2. Add initiators as follows:


Riboflavin (0.1% w/v) 50 µl 0.0005%(5 µg/ml)

Ammonium persulfae (10% w/v) 15 µl 0.015%


3. Swirl the solution gently but thoroughly.

4. Cast the gel and allow polymerization to occur for at least 2 hr for isoelectric focusing gels. If separation is to be based on size, allow photochemically initiated gels, to polymerize for 8 hr under light from a nearby fluorescent lamp, (Righetti et al. 1981).

Polymerization AnalysisThere are several ways to assess the extent and reproducibilityof polymerization. One of the easiest methods is to routinelymonitor the time required for visible gelation. There are severalfactors which affect the polymerization rate. A significant changein the time required for visible gelation indicates that one of theparameters has changed.

The polymerized gel should be inspected for evidence of inhibition or nonuniform polymerization. A swirled or “schlieren”pattern, for example, indicates that polymerization was too fastor that the polymerization initiators were not mixed thoroughly with the monomer solution prior to casting the gel.

Polymerization ProfileAs acrylamide polymerizes, UV-absorbing double bonds areeliminated. The progress of a reaction can therefore be followedby monitoring absorbance at 260 nm. As the reaction proceeds,the UV absorbance drops. Absorbance increases with theamount of unreacted monomer.

Fig. 1. Polymerization profile for 12% monomer (12%T, 2.6%C*) containing 0.375 M Tris-HCl, pH 8.8 (Laemmli), polymerized in a quartz cuvette at room temperature with ammonium persulfate and TEMED at final concentrations of0.05% each. Polymerization was monitored at 260 nm. A, “enzyme grade” acrylamide. B, Bio-Rad’s electrophoresis-purity acrylamide, control # 25281.

Figure 1 shows a UV profile of chemical polymerization for 2 samples of acrylamide polymerized under identical conditionsin a quartz cuvette. Sample A was an “enzyme grade” acrylamide with a conductivity (50% w/w) of 3.75 µS. Sample Bwas Bio-Rad’s electrophoresis-purity acrylamide with a conductivity of 0.56 µS. As the figure shows, polymerization islargely complete after about 90 min, even though the reaction proceeds to a small extend beyond that time. While sample A began to polymerize faster, sample B polymerized more completely, as indicated by the lower final UV absorbance.

20 40 60 80 100 120 140Time, min

0.6

1.0

1.5

2.0

2.5

3.0

A26

0

BA

Contaminants in acrylamide may be accelerators or inhibitors ofpolymerization. Therefore, initiation of polymerization, as indicated by reduced absorbance, may be faster with crudeacrylamide than with highly refined acrylamide. However, themost important consideration is the completeness of polymerization. Polymerization of highly refined acrylamide may be initiated more slowly, but conversion of monomer topolymer, as indicated by the low final absorbance, is more complete. Therefore, less residual monomer remains. Completepolymerization is critical for reproducibility in gel porosity.

Gel Exclusion Limit DeterminationEstimation of protein molecular weight by SDS-PAGE is a widelyemployed procedure. The relative mobility of a protein in anSDS-PAGE gel is related to its molecular weight. A standardcurve is constructed with proteins of known molecular weightby plotting the logarithms of their molecular weights versus therelative mobilities of the proteins. The relative mobility of a protein of unknown molecular weight is then fitted to the curveto determine its molecular weight.

A standard curve can be extrapolated to give the y-intercept,which represents the molecular weight exclusion limit of thatparticular gel. That is, proteins with a molecular weight greaterthan the y-intercept value will show zero mobility and will beexcluded from the gel matrix.

Poorly polymerized gels have greater porosity due to incompletechain elongation and crosslinking. As a result, the exclusion limitwill be greater than for a well-polymerized gel of the same percent acrylamide. Furthermore, when polymerization is incomplete, exclusion limits are irreproducible. Use of highlypurified gel-forming reagents and proper polymerization technique will result in the lowest and most reproducible exclusion limits for a given percent total monomer.

Figure 2 shows a typical curve obtained by plotting log molecular weight versus relative mobility following SDS-PAGEfor a group of standard proteins. The antilog of the y-interceptvalue of this plot is 115,000 as determined by linear regressionanalysis. The approximate molecular weight exclusion limit ofthe gel is thus 115,000. The y-intercept value should be considered approximate because it depends upon the relative mobility of the proteins used as standards.

Although the y-intercept value will be different for every gel acrylamide percentage, and slightly different for every set ofstandards, the value should be highly reproducible from gel togel if the same acrylamide percentage and standards are used.Thus, monitoring the y-intercept of the log molecular weight vs.relative mobility plot is an excellent assessment of reproducibilityin polymerization technique.

* %C = (grams crosslinker x 100)/(grams monomer + grams crosslinker)

Fig. 2. A representative calibration curve for molecular mass (Mr) estimation. In the run that is plotted here (solid line), Bio-Rad SDS-PAGE standards with Mrof 200, 116.2, 97.4, 66.2, 45, 31, 21.5, 14.4, and 6.5 kD (closed circles, left toright) were separated on a 15%T SDS-PAGE gel. The plot shows the inherentnonlinearity of such curves. The straight-line segment in the middle of the plot is the most accurate range for Mr estimation. Larger polypeptides experiencegreater sieving than do those in the middle range, so the corresponding upperpart of the curve has a different slope than in the middle. The curve in thesmaller polypeptide range also deviates from a straight line because of lesssieving. Because the scale is logarithmic, an estimate of Mr from a “best fit”straight line (dashed) is acceptable for many purposes.

Handling of AcrylamideUse good laboratory practices, work in a well ventilated areaand wear proper personnel proctective equipment. Refer tothe MSDS for further information.

Reagent Storage and Shelf LifeAcrylamide and bis-acrylamide — Electrophoresis-purity acrylamide and bis can be stored dry at room temperature(23–25°C) for at least 1 year.

Ammonium persulfate and potassium persulfate — These initiators can be stored tightly sealed at room temperature forat least 1 year. Solutions should be made fresh daily, sincepersulfate in solution decomposes rapidly. Persulfate is astrong oxidizing agent. Disposal should be in accordance with local regulations.

TEMED — This initiator can be stored tightly closed either at4°C or at room temperature for at least 6 months.

After 10 to 12 months, a significant reduction in reactivityrequires an increase in the concentration required for properpolymerization. This loss of reactivity is probably due, at leastin part, to the gradual accumulation of water.

Riboflavin and riboflavin-5'-phosphate — These photoinitiators can be stored dry at room temperature for atleast 1 year. In aqueous solution, they are stable for at least 1 month if kept in the dark at 4°C. Riboflavin phosphate isusually preferred because of its greater solubility.

0.9

1.1

1.3

1.5

1.7

1.9

2.1

2.3

2.5

0.5

0.7

0 0.1 0.2 3.0 0.4 0.5 0.6 0.7 0.8 0.9 1

Relative mobility

log

(Mr x

10- 3

)

Life ScienceGroup

Web site www.bio-rad.com USA (800) 4BIORAD Australia 02 9914 2800 Austria (01)-877 89 01 Belgium 09-385 55 11 Brazil 55 21 507 6191Canada (905) 712-2771 China 86-10-8201-1366/68 Denmark 45 44 52-1000 Finland 358 (0)9 804 2200 France 01 47 95 69 65 Germany 089 318 84-177 Hong Kong 852-2789-3300 India (91-124) 6398112/113/114 Israel 03 951 4124 Italy 34 91 590 5200 Japan 03-5811-6270 Korea 82-2-3473-4460 Latin America 305-894-5950 Mexico 52 5 534 2552 to 54 The Netherlands 0318-540666 New Zealand 64-9-4152280 Norway 47-23-38-41-30Russia 7 095 979 98 00 Singapore 65-2729877 Spain 34-91-590-5200 Sweden 46 (0)8-55 51 27 00 Switzerland 061-717-9555 United Kingdom 0800-181134

00-305 0401 Sig 0101Bulletin 1156 US/EG Rev E

Bio-RadLaboratories, Inc.

ReferencesAllison JH et al., Effect of N,N,N’,N’-tetramethylethylenediamine on the migration of proteins in SDS polyacrylamide gels, Anal Biochem 58, 592–601 (1974)

Andrews AT, Acid-urea detergent gels, pp 141–143 in Electrophoresis Theory,Techniques, and Biochemical and Clinical Applications, 2nd ed, Oxford SciencePublications, Oxford (1990)

Chen B and Chrambach A, Estimation of polymerization efficiency in the formation of polyacrylamide gel, using continuous optical scanning duringpolymerization, J Biochem Biophys Methods 1, 105–116 (1979)

Chrambach A et al., Analytical and preparative polyacrylamide gel electrophoresis. An objectively defined fractionation route, apparatus, and procedures, Methods Protein Sep 2, 27–144 (1976)

Chrambach A, The Practice of Quantitative Gel Electrophoresis, VCH, Deerfield Beach (1985)

Dirksen ML and Chrambach A, Studies on the redox state in poly acrylamidegels, Sep Sci 7, 747–772 (1972)

Gelfi C and Righetti PG, Polymerization kinetics of polyacrylamide gels I. Effectof different cross-linkers, Electrophoresis 2, 213–219 (1981a)

Gelfi C and Righetti PG, Polymerization kinetics of polyacrylamide gels II. Effectof temperature, Electrophoresis 2, 220–228 (1981b)

Gordon AH, Electrophoresis of Proteins in Polyacrylamide and Starch Gels, 2nded, Elsevier/North-Holland Biomedical Press, Amsterdam (1975)

Righetti PG et al., Polymerization kinetics of polyacrylamide gels. III. Effect ofcatalysts, Electrophoresis 2, 291–295 (1981)

Shi Q and Jackowski G, One-dimensional polyacrylamide gel electrophoresis,pp 1–52 in Hames BD (ed) Gel Electrophoresis of Proteins: A PracticalApproach, 3rd edn, Oxford University Press, Oxford (1998)

* Triton is a trademark of Union Carbide Chemicals and Plastics Technology Corp.

SDS gel doesn’tpolymerize

Swirls in gel

IEF gel doesn’tpolymerize

Long polymerizationtime, incompletecatalysis

• Too little or too muchAPS or TEMED

• Failure to degas• Temperature too low

• Poor-qualityacrylamide or bis

• APS not freshly made

• Excessive catalysis —gel polymerized in <10 min

• Gel inhibition — poly-merization time >1 hr

• Basic gradient; gelpolymerization problem

• Gel has no structure• Riboflavin-catalyzed;

gel was degassed too long

• APS alone doesn’tpolymerize

• Didn’t degas forchemical polymeri-zation. Temperaturetoo low.

• Poor-quality acryl-amide or bis

• APS not freshly made• Failure to degas• Too little APS or

TEMED• Temperature too low

• Poor-quality acryl-amide or bis

• APS not freshly made• TEMED old

• Requires 0.05% APSand 0.05% TEMED

• Degas 10–15 min• Cast at room

temperature, warmingglass plates if necessary

• Use electrophoresis-purity reagents

• Make solution fresh daily

• Reduce APS andTEMED by 25% each

• Increase APS andTEMED by 50%; degas

• Requires 0.015% APS, 1 µl/ml TEMED and 0.0005% riboflavin

• Degas 1–2 min• O2 required for this

reaction to initiate —don’t degas

• Requires 0.015% APS, 1 µl/ml TEMED

• Degas 10–15 min atroom temperature

• Use electrophoresis-purity acrylamide and bis

• Make fresh APS• Degas 10–15 min• Increase both to 0.05%

• Cast at room temperature


• Make APS fresh daily• Use new TEMED

Problem Cause Remedy

Gel feels soft

Gel turns white

Gel brittle

Diffuse or broadbands

Inconsistent relative mobilities

High silver-staining background

Severe cathodicdrift in IEF, molecules diffuseor don’t reachproper pl

• Low % T• Poor-quality acryl-

amide or bis• Too little crosslinker

• bis concentration too high

• Crosslinker too high

• Poor-qualityacrylamide or bis;incomplete catalysis

• Sample notequilibrated

• Excessive TEMED orAPS

• SDS or sample buffertoo old

• Gel temperature high

• Incomplete catalysis;excessive TEMED or APS

• Did not degas

• Acrylic acid contamination inacrylamide and bis

• Acrylic acid contamination inacrylamide and bis

• Gel not aged longenough after photopolymerization

• Poor-quality urea

• No remedy• Use electrophoresis-

purity reagents• Make sure of proper %C

• Recheck solutions or weights

• Recheck %C


• Equilibrate sample torunning conditions

• Reduce initiator concentrations by 25%

• Prepare fresh solutions

• Cool during run or runmore slowly

• TEMED and APS shouldbe 0.05%

• Degas 10–15 min



• Allow polymerization for8 hr before using the gel

• Deionize urea; use electrophoresis-purityreagents

Problem Cause Remedy

Polymerization Artifacts and TroubleshootingComplete and reproducible polymerization is dependent on proper technique and pure gel-forming reagents. The table belowlists commonly observed polymerization artifacts and problems, along with their probable cause and solution.

SpecificationsContentsApproximately 400 µg of each pro-

tein blended to give bands of equal intensity on SDS polyacrylamide gels run according to Laemmli1and stained with Coomassie Blue R-250

Storage 50% glycerol, 300 mM NaCl, buffer10 mM Tris, 2 mM EDTA,

3 mM NaN3

Volume200 µl concentrated solution

Storage-20 °C

Shipping Room temperature,conditions

Shelf life1 year at -20 °C

Applications 400 with Coomassie R-250per vial

RecommendedLow range12.5%gel percentage*High range7.5%

Broad range4-20 % gradient gels

*Note:These standards can be run on other percentagegels, but all proteins may not be visible. Lower percent-age gels may cause the low molecular weight proteins tomigrate with or in front of the dye front. Higher percent-age gels may prevent the high molecular weight proteinsfrom separating.

Protein Molecular Weights (daltons)ProteinMolecularBroadLowHigh

WeightRangeRangeRange

Myosin200,000XXß-galactosidase116,250XXPhosphorylase b97,400XXXSerum albumin66,200XXXOvalbumin45,000XXXCarbonic anhydrase31,000XXTrypsin inhibitor21,500XXLysozyme14,400XXAprotinin6,500X

Reference1. Laemmli, U. K., Nature, 227, 680 (1970).

2. Hames, B. D. and Rickwood, D., Gel Electrophoresis ofProteins: A Practical Approach, Second Edition, p. 17,Oxford University Press, New York (1990).

Ordering InformationCatalogNumber Product Description

Molecular Weight Standards161-0303 SDS-PAGE Standards, High, 200 µl161-0304 SDS-PAGE Standards, Low, 200 µl161-0317 SDS-PAGE Standards, Broad, 200 µl161-0314 Silver Stain SDS-PAGE Standards, Low, 200 µl161-0315 Silver Stain SDS-PAGE Standards, High, 200 µl161-0306 Biotinylated SDS-PAGE Standards, Low, 250 µl161-0311 Biotinylated SDS-PAGE Standards, High, 250 µl161-0319 Biotinylated SDS-PAGE Standards, Broad, 250 µl161-0320 2-D SDS-PAGE Standards161-0326 Polypeptide SDS-PAGE Standards, 200 µlPrestained Standards161-0305 Prestained SDS-PAGE Standards, Low, 500 µl161-0309 Prestained SDS-PAGE Standards, High, 500 µl161-0318 Prestained SDS-PAGE Standards, Broad, 500 µl161-0324 Kaleidoscope Prestained Standards, 500 µl161-0325 Kaleidoscope Polypeptide Standards, 500 µlIEF Standards161-0310 IEF Standards, pI range 4.45-9.6, 250 µlBio-Rad Laboratories, 2000 Alfred Nobel Drive, Hercules CA 94547

4006035 Rev C

SDS-PAGE Molecular Weight Standards,

Broad Range

Catalog Number161-0317

Product shipped at room temperature.Store at -20 °C upon arrival.

12

4006035C 8/28/98 03:01 PM Page 1

SpecificationsContents Approximately 400 µg of each pro-

tein blended to give bands of equal intensity on SDS polyacrylamide gels run according to Laemmli1 and stained with Coomassie Blue R-250

Storage 50% glycerol, 300 mM NaCl, buffer 10 mM Tris, 2 mM EDTA,

3 mM NaN3

Volume 200 µl concentrated solution

Storage -20 °C

Shipping Room temperature,conditions

Shelf life 1 year at -20 °C


Recommended Low range 12.5%gel percentage* High range 7.5%

Broad range 4-20 % gradient gels

*Note: These standards can be run on other percentagegels, but all proteins may not be visible. Lower percent-age gels may cause the low molecular weight proteins tomigrate with or in front of the dye front. Higher percent-age gels may prevent the high molecular weight proteinsfrom separating.

Protein Molecular Weights (daltons)Protein Molecular Broad Low High

Weight Range Range Range

Myosin 200,000 X Xß-galactosidase 116,250 X XPhosphorylase b 97,400 X X XSerum albumin 66,200 X X XOvalbumin 45,000 X X XCarbonic anhydrase 31,000 X XTrypsin inhibitor 21,500 X XLysozyme 14,400 X XAprotinin 6,500 X

Reference1.Laemmli, U. K., Nature, 227, 680 (1970).

2.Hames, B. D. and Rickwood, D., Gel Electrophoresis ofProteins: A Practical Approach, Second Edition, p. 17,Oxford University Press, New York (1990).

Ordering InformationCatalogNumberProduct Description

Molecular Weight Standards161-0303SDS-PAGE Standards, High, 200 µl161-0304SDS-PAGE Standards, Low, 200 µl161-0317SDS-PAGE Standards, Broad, 200 µl161-0314Silver Stain SDS-PAGE Standards, Low, 200 µl161-0315Silver Stain SDS-PAGE Standards, High, 200 µl161-0306Biotinylated SDS-PAGE Standards, Low, 250 µl161-0311Biotinylated SDS-PAGE Standards, High, 250 µl161-0319Biotinylated SDS-PAGE Standards, Broad, 250 µl161-03202-D SDS-PAGE Standards161-0326Polypeptide SDS-PAGE Standards, 200 µlPrestained Standards161-0305Prestained SDS-PAGE Standards,Low, 500 µl161-0309Prestained SDS-PAGE Standards,High, 500 µl161-0318Prestained SDS-PAGE Standards,Broad, 500 µl161-0324Kaleidoscope Prestained Standards,500 µl161-0325Kaleidoscope Polypeptide Standards,500 µlIEF Standards161-0310IEF Standards,pI range 4.45-9.6, 250 µlBio-Rad Laboratories, 2000 Alfred Nobel Drive, Hercules CA 94547

4006035 Rev C


Broad Range



1 2

4006035C 8/28/98 03:01 PM Page 1

5

Myosin

ß-galactosidasePhosphorylase b

Bovine serum albumin

Ovalbumin

Carbonic anhydrase

Soybean trypsininhibitor

LysozymeAprotinin

Fig. 1. SDS polyacrylamide gels run in theMini-PROTEAN® II cell according to the method ofLaemmli.1 Broad molecular weight standards run on a 4-20% gradient gel, stained with Coomassie R-250.

3 4

6

ProtocolDilute standards 1:20 in SDS Reducing SampleBuffer.* Heat for 5 minutes at 95 °C. Cool andload 10 µl/well for full length gels (16-20 cm) or 5 µl/well for mini gels.

* SDS Reducing Sample Buffer (prepare immediatelybefore use)

ß-mercaptoethanol 25 µlStock Sample Buffer 475 µl

500 µl

Stock Sample Buffer (store at room temperature)

Distilled water 4.8 ml0.5M Tris-HCl pH 6.8 1.2 mlGlycerol 1.0 ml10% (w/v) SDS 2.0 ml0.1% (w/v) Bromophenol blue 0.5 ml

9.5 ml

Use of Sample Buffer with insufficient or old ß-mercaptoethanol may result in doublets at thesoybean trypsin inhibitor and ovalbumin bands.

Lo

g M

W

Rf

3

4

5

6

1.00.80.60.40.2

Fig. 3. Curve generated by plotting the log of themolecular weight of the broad range standards vs.the relative mobility (Rf).

Rf = distance migrated by proteindistance migrated by dye

The curve can be used to determine molecular weightsof unknown proteins.2

Protein References

Protein Reference

Rabbit skeletal Woods, E. F., Himmelfarb, S. andmuscle myosin Harrington, W. F., J. Biol. Chem.,

238, 2374 (1963).

E. coli ß-galac- Fowler, A. V. and Zabin, I., Proc. tosidase Natl. Acad. Sci. USA, 74, 1507 (1977).

Rabbit muscle Titani, K., et al., Proc. Natl. phosphorylase b Acad. Sci. USA, Vol. 74, 4762 (1977).

Bovine serum Brown, J. R., Fed. Proc., 34, 591 (1975).albumin (BSA)

Hen egg white Warner, R. C., “Egg Proteins,” ovalbumin in: The Proteins, Vol. IIA, p. 435

(Neurath, H. and Bailey, K., eds.),Academic Press, New York (1954).

Bovine carbonic Davis, R. P., “Carbonic Anhydrase,” in:anhydrase The Enzymes, Vol V, p. 545, (Boyer, P.

D., ed.) Academic Press, New York (1971)

Soybean trypsin Wu, Y. V. and Scherage, H. A., inhibitor Biochemistry, 1, 698 (1962).

Hen egg white Jolles, P., Angew. Chem Intl. Edit., 8, 227lysozyme (1969).

Bovine pancreatic Kassell,B. and Laskowski, M., Biochem.trypsin inhibitor Biophys. Res. Comm., 20, 463 (1965).(Aprotinin)

4006035C 8/28/98 03:01 PM Page 2

SpecificationsContentsApproximately 400 µg of each pro-

tein blended to give bands of equal intensity on SDS polyacrylamide gels run according to Laemmli1and stained with Coomassie Blue R-250

Storage 50% glycerol, 300 mM NaCl, buffer10 mM Tris, 2 mM EDTA,

3 mM NaN3

Volume200 µl concentrated solution

Storage-20 °C

Shipping Room temperatureconditions

Shelf life1 year at -20 °C


RecommendedLow range12.5%gel percentage*High range7.5%

Broad range4-20 % gradient gels

*Note:These standards can be run on other percentagegels, but all proteins may not be visible. Lower percent-age gels may cause the low molecular weight proteins tomigrate with or in front of the dye front. Higher percent-age gels may prevent the high molecular weight proteinsfrom separating.

Protein Molecular Weights (daltons)ProteinMolecularBroadLowHigh

WeightRangeRangeRange

Myosin200,000XXß-galactosidase116,250XXPhosphorylase b97,400XXXSerum albumin66,200XXXOvalbumin45,000XXXCarbonic anhydrase31,000XXTrypsin inhibitor21,500XXLysozyme14,400XXAprotinin6,500X

Reference1. Laemmli, U. K., Nature, 227, 680 (1970).

2. Hames, B. D. and Rickwood, D., Gel Electrophoresis ofProteins: A Practical Approach, Second Edition, p. 17,Oxford University Press, New York (1990).

Ordering InformationCatalogNumber Product Description

Molecular Weight Standards161-0303 SDS-PAGE Standards, High, 200 µl161-0304 SDS-PAGE Standards, Low, 200 µl161-0317 SDS-PAGE Standards, Broad, 200 µl161-0314 Silver Stain SDS-PAGE Standards, Low, 200 µl161-0315 Silver Stain SDS-PAGE Standards, High, 200 µl161-0306 Biotinylated SDS-PAGE Standards, Low, 250 µl161-0311 Biotinylated SDS-PAGE Standards, High, 250 µl161-0319 Biotinylated SDS-PAGE Standards, Broad, 250 µl161-0320 2-D SDS-PAGE Standards161-0326 Polypeptide SDS-PAGE Standards, 200 µlPrestained Standards161-0305 Prestained SDS-PAGE Standards, Low, 500 µl161-0309 Prestained SDS-PAGE Standards, High, 500 µl161-0318 Prestained SDS-PAGE Standards, Broad, 500 µl161-0324 Kaleidoscope Prestained Standards, 500 µl161-0325 Kaleidoscope Polypeptide Standards, 500 µlIEF Standards161-0310 IEF Standards, pI range 4.45-9.6, 250 µlBio-Rad Laboratories, 2000 Alfred Nobel Drive, Hercules CA 94547

4006033 Rev C


Low Range



12

4006033C 8/28/98 02:53 PM Page 1

SpecificationsContents Approximately 400 µg of each pro-

tein blended to give bands of equal intensity on SDS polyacrylamide gels run according to Laemmli1 and stained with Coomassie Blue R-250

Storage 50% glycerol, 300 mM NaCl, buffer 10 mM Tris, 2 mM EDTA,

3 mM NaN3

Volume 200 µl concentrated solution

Storage -20 °C

Shipping Room temperatureconditions

Shelf life 1 year at -20 °C


Recommended Low range 12.5%gel percentage* High range 7.5%

Broad range 4-20 % gradient gels

*Note: These standards can be run on other percentagegels, but all proteins may not be visible. Lower percent-age gels may cause the low molecular weight proteins tomigrate with or in front of the dye front. Higher percent-age gels may prevent the high molecular weight proteinsfrom separating.

Protein Molecular Weights (daltons)Protein Molecular Broad Low High

Weight Range Range Range

Myosin 200,000 X Xß-galactosidase 116,250 X XPhosphorylase b 97,400 X X XSerum albumin 66,200 X X XOvalbumin 45,000 X X XCarbonic anhydrase 31,000 X XTrypsin inhibitor 21,500 X XLysozyme 14,400 X XAprotinin 6,500 X

Reference1.Laemmli, U. K., Nature, 227, 680 (1970).

2.Hames, B. D. and Rickwood, D., Gel Electrophoresis ofProteins: A Practical Approach, Second Edition, p. 17,Oxford University Press, New York (1990).

Ordering InformationCatalogNumberProduct Description

Molecular Weight Standards161-0303SDS-PAGE Standards, High, 200 µl161-0304SDS-PAGE Standards, Low, 200 µl161-0317SDS-PAGE Standards, Broad, 200 µl161-0314Silver Stain SDS-PAGE Standards, Low, 200 µl161-0315Silver Stain SDS-PAGE Standards, High, 200 µl161-0306Biotinylated SDS-PAGE Standards, Low, 250 µl161-0311Biotinylated SDS-PAGE Standards, High, 250 µl161-0319Biotinylated SDS-PAGE Standards, Broad, 250 µl161-03202-D SDS-PAGE Standards161-0326Polypeptide SDS-PAGE Standards, 200 µlPrestained Standards161-0305Prestained SDS-PAGE Standards,Low, 500 µl161-0309Prestained SDS-PAGE Standards,High, 500 µl161-0318Prestained SDS-PAGE Standards,Broad, 500 µl161-0324Kaleidoscope Prestained Standards,500 µl161-0325Kaleidoscope Polypeptide Standards,500 µlIEF Standards161-0310IEF Standards,pI range 4.45-9.6, 250 µlBio-Rad Laboratories, 2000 Alfred Nobel Drive, Hercules CA 94547

4006033 Rev C


Low Range



1 2

4006033C 8/28/98 02:53 PM Page 1

5

Phosphorylase b

Bovine serumalbumin

Ovalbumin

Carbonicanhydrase

Soybean trypsininhibitor

Lysozyme

Fig. 1. SDS polyacrylamide gels run in theMini-PROTEAN® II cell according to the method ofLaemmli.1 Low molecular weight standards run on a 12% gel, stained with Coomassie R-250.

3 4

6

ProtocolDilute standards 1:20 in SDS Reducing SampleBuffer.* Heat for 5 minutes at 95 °C. Cool andload 10 µl/well for full length gels (16-20 cm) or 5 µl/well for mini gels.

* SDS Reducing Sample Buffer (prepare immediatelybefore use)

ß-mercaptoethanol 25 µlStock Sample Buffer 475 µl

500 µl

Stock Sample Buffer (store at room temperature)

Distilled water 4.8 ml0.5M Tris-HCl pH 6.8 1.2 mlGlycerol 1.0 ml10% (w/v) SDS 2.0 ml0.1% (w/v) Bromophenol blue 0.5 ml

9.5 ml

Use of Sample Buffer with insufficient or old ß-mercaptoethanol may result in doublets at thesoybean trypsin inhibitor and ovalbumin bands.

Lo

g M

W

Rf

3

4

5

6

1.00.80.60.40.2

Fig. 3. Curve generated by plotting the log of themolecular weight of the broad range standards vs.the relative mobility (Rf).

Rf = distance migrated by proteindistance migrated by dye

The curve can be used to determine molecular weightsof unknown proteins.2

Protein References

Protein Reference

Rabbit skeletal Woods, E. F., Himmelfarb, S. andmuscle myosin Harrington, W. F., J. Biol. Chem.,

238, 2374 (1963).

E. coli ß-galac- Fowler, A. V. and Zabin, I., Proc. tosidase Natl. Acad. Sci. USA, 74, 1507 (1977).

Rabbit muscle Titani, K., et al., Proc. Natl. phosphorylase b Acad. Sci. USA, Vol. 74, 4762 (1977).

Bovine serum Brown, J. R., Fed. Proc., 34, 591 (1975).albumin (BSA)

Hen egg white Warner, R. C., “Egg Proteins,” ovalbumin in: The Proteins, Vol. IIA, p. 435

(Neurath, H. and Bailey, K., eds.),Academic Press, New York (1954).

Bovine carbonic Davis, R. P., “Carbonic Anhydrase,” in:anhydrase The Enzymes, Vol V, p. 545, (Boyer, P.

D., ed.) Academic Press, New York (1971)

Soybean trypsin Wu, Y. V. and Scherage, H. A., inhibitor Biochemistry, 1, 698 (1962).

Hen egg white Jolles, P., Angew. Chem Intl. Edit., 8, 227lysozyme (1969).

Bovine pancreatic Kassell,B. and Laskowski, M., Biochem.trypsin inhibitor Biophys. Res. Comm., 20, 463 (1965).(Aprotinin)

4006033C 8/28/98 02:54 PM Page 2

mutation analysis tech note 2377

Luc Michiels1, Baudouin François1,2, and Jef Raus1, 2, 1 Dr. L. Willems-Instituutand 2 Limburgs Universitair Centrum, Universitaire Campus, B-3590Diepenbeek, Belgium

IntroductionPhenylketonuria (PKU) is a hereditary disease that gives riseto elevated blood levels of phenylalanine. The disease mostfrequently is caused by mutations in the phenylalaninehydroxylase gene. This gene is situated at the humanchromosomal locus 12q22-q24.1 and encodes a hepaticenzyme that hydroxylates phenylalanine (PAH; phenylalanine4-monooxygenase, EC 1.14.16.1). The large variety ofmutations described in this gene results in a broad range ofplasma phenylalanine concentrations associated withphenotypic differences ranging from the severest form of PKUto the mildest hyperphenylalaninaemia. PKU is autosomalrecessively inherited, carriers therefore are phenotypicallynormal, with no pronounced elevated phenylalanine ortyrosine blood levels. The need for a reliable method toidentify persons carrying PAH mutations is evident. To identifymutations in each of the 13 exons of the PAH gene, Guldbergand coworkers (1993) reported a method based on DGGE.1

For the rapid screening of PKU patients and their relatives formutations in the PAH exons and intron/exon boundaries wefurther refined this method by using multiplex PCR*amplifications combined with Multiplex Denaturing GradientGel electrophoresis analysis (Michiels et al., 1996).3

MethodsMULTIPLEX PCR AMPLIFICATION OF PAH DNA

Genomic DNA of PKU patients was prepared from dried bloodspots on Guthrie cards using the Chelex® 100 extractionprocedure (Walsh et al., 1991).5 Alternatively 0.5 ml freshEDTA-treated blood sample was treated with 1 ml haemolysisbuffer (HB) at 4 °C for 10 minutes. White blood cells werepelleted (5 minutes at 2,500 rpm in a microfuge) and washedwith HB until no red cell debris was left. After the removal ofthe supernatant, 500 µl of 5% Chelex 100 (Bio-Rad, California)suspension in water was added and mixed thoroughly.

This step was followed by incubations at 56 °C for 30 minutesand subsequently at 100 °C for 5 minutes. After centrifuging (3 minutes at 2,500 rpm), the supernatant was stored at -20 °C. Five µl of these DNA preparations was used in a 20 µlmultiplex amplification reaction of the PAH exons and theirintron-exon boundaries. GC-clamped primers for DGGEanalysis of the PAH gene have been described by Guldberg et al., 1993.1 To achieve comparable PCR amplificationefficiencies and non-overlapping, but high mutation resolutionDGGE patterns, combinations of these primer sets wereevaluated (Michiels et al. 1996).3 Amplifications were carriedout in 20 µl reaction mixtures, containing Perkin Elmer PCRreaction buffer with 1.5 mM MgCl2, 200 nM each dNTP, 800 nM each primer (MP 1: exon 1 + exon 4 + exon 13, MP2:exon 2 + exon 6+ exon 8, MP3 : exon 3 + exon 5 + exon 9,MP7 : exon 7 + exon 11, MP10: exon 10 + exon 12.) and 1.5 U of Taq DNA polymerase (Perkin Elmer). The Perkin Elmer9600 thermal cycler was set at 5 minutes denaturation at 95 °C followed by 40 cycles of 10 seconds denaturation at 95 °C, 20 seconds annealing at 56 °C and 40 secondselongation at 72 °C. After cycling, final elongation was done at72 °C for 5 minutes. Finally, to generate heteroduplexesbetween heterozygous PAH gene fragments, the PCRfragments were incubated for 5 minutes at 95 °C, 60 minutesat 65 °C and 60 minutes at 37 °C. Complete amplificationmixes were used for DGGE analysis.

MP-DGGE was performed under two different electrophoresisconditions. A 6% polyacrylamide gel containing a 20–70%(for MP2, MP3 and MP10) or a 30–80% (for MP1 and MP7)gradient (Model 475 Gradient Delivery system, Bio-Rad) ofurea and formamide (100% is 7 M urea and 40% formamide)was run at 60 °C in TAE (40 mM Tris-acetate, 1 mM EDTA,pH 8.0). Electrophoresis conditions were 130 V for 6 hourson the DCode apparatus (Bio-Rad). Gels were stained in 1 µg/ml ethidium bromide and visualized by UV transillumination.

Identification of Disease Causing Mutations in Phenylketonuria byDenaturing Gradient Gel Electrophoresis Using the DCode™ System

Bulletin 2377 2/3/99 10:51 AM Page 1

The point mutations were confirmed by sequencing the PCRamplified fragments with the non-GC clamped primer of eachexon, using the Applied Biosystems Model 373A DNASequencing System and the PRISM™ Ready Reaction DyeDeoxy™ Terminator Cycle Sequencing Kit (Applied Biosystems).

ResultsDenaturing gradient gel electrophoresis allows detection ofmore than 95% of the PKU causing mutations. In the Belgianpopulation, we observed about 100 different mutations usingDGGE technology. Among these, seven new mutations wereidentified (Michiels et al.,1998).4 The high mutation resolvingpower of DGGE is illustrated in Figure 1. This figure shows allthe mutations or mutation combinations we located in exon 7of Belgian PKU patients. The screening of PKU patients andtheir relatives is rather laborious when exon specificamplification and DGGE conditions are used. Therefore, toobtain compatible co-amplifications with comparableefficiencies, different exon PCR amplifications were combined.Moreover, to make sure that the resolving power to identify

the exon specific mutations remains intact and that the DGGEpatterns of the combined exons do not interfere with eachother, the multiplex PCR amplifications have to be resolvedsubsequently on denaturing gradient gels. This resulted in amultiplex DGGE (MP-DGGE) of the PAH gene (Michiels et al.1996).3 Examples of such an MP-DGGE are shown in Figure 2.The 13 different PAH exons are analyzed in 5 distinct multiplexPCR and DGGE patterns. Each exon shows distinct mutations,demonstrating that the resolving power of the DGGE gelremains intact. In addition, the DGGE patterns for the differentexons do not overlap. Cycle sequencing of the DGGE DNAbands excised from the gel as described in the methods allowsthe identification of new mutations.

Several DNA preparation protocols have been tested: homemade (see Methods) or commercially available kits such asSplit Second™ from Boehringer Mannheim, Germany andDNA Easy-Prep from Lifecodes Corporation, USA, and can allbe used in the MP-DGGE assay. All of them gave comparableresults in the multiplex PCR/DGGE experiments.

Fig. 1. Mutations detected in exon 7 of the PAH gene using DGGE analysis. Lane 1: WT/WT; lane 2: R243Q/WT; lane 3: R243X/WT; lane 4: V245V/WT; lane 5: R252W/WT; lane 6: R261Q/WT; lane7: : R261P/WT; lane 8: : R261X/WT; lane 9: G272X/WT; lane 10: Y277D/WT; lane 11: E280K/WT; lane 12: P281L/WT; lane 13: V245V/V245V; lane 14: V245V/V245V +R243X/WT; lane 15: G272X/WT + E280K/WT; lane 16: V245V/ WT + G272X/WT; lane 17: V245V/V245V +IVS7nt1/WT; lane 18: V245V/WT + P281L/WT; lane 19: V245V/WT + Y277D/WT; lane 20: V245V/WT + R243X/WT; lane 21: V245V/WT + IVS7nt1/WT + R261Q/WT;lane 22: V245V/WT + R243Q/WT.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

Bulletin 2377 2/3/99 10:51 AM Page 2

Fig. 2. Examples of multiplex DGGE analysis. Heterozygous mutations detected are Q20X in exon 1 (MP1, E1, lane 2), IVS3nt-22c/t in intron 3 (MP1, E4, lanes 2,5),IVS12nt-35c/t in intron 12 (MP1, E13, lane 1), Q232Q in exon 6 (MP2, E6, lane 1), L48S in exon 2 (MP2, E2, lane 1), IVS12nt19t/c in intron 2 (MP2, E2, lane 5),R158Q in exon 5 (MP3, E5, lane 1), L311P in exon 9 (MP3, E9, lane 2), I65T in exon 3 (MP3, E3, lane 3), L385L in exon 11 (MP7, E11, lane 1), combination of V245Vand R261Q in exon 7 (MP7, E7, both homozygous in lane 2 and both heterozygous in lane 3), V245V in exon 7 (MP7, E7, lane 4), IVS10nt-11g/a in intron 10 (MP7,E11, lane 5), combination of D415N and IVS12nt1g/a in exon 12 and at the splice junction of intron 12 (MP10, E12, lane 1), R408W in exon 12 (MP10, E12, lanes2,3), K341T in exon 10 (MP10, E10, lane 3), S349P in exon 10 (MP10, E10, lanes 4,5) and IVS12nt1g/a in intron 12 (MP10, E12, lane 5).

1 2 3 4 5

MP1 MP2 MP3 MP7 MP10

1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

ConclusionIn this report we demonstrate the use of powerful multiplexPCR amplification combined with DGGE analysis to rapidlyidentify mutations causing PKU. This MP-DGGE allows analysisof the complete PAH gene for three different individuals on onegel, whereas previously such analysis would take three gelruns. Moreover, the multiplex DGGE analysis is compatible withall the tested rapid DNA isolation methods commerciallyavailable. With this procedure DNA extracted from dried bloodspots on Guthrie cards can be successfully analyzed.

AcknowledgementsThis work was carried out within the framework of Biomed 1GENE-CT93-0081. Further support was obtained from theFonds ter Bevordering van het Wetenschappelijk Onderzoekin het DWI (FWI).

References1 Guldberg P., Henriksen K.F. and Güttler F. Molecular Analysis of

Phenylketonuria in Denmark: 99 % of the Mutations Detected by DenaturingGradient Gel Electrophoresis. Genomics, 17, 141–146 (1993).

2 Güttler F., Hyperphenylalaninemia: Diagnosis and classification of the varioustypes of phenylalanine hydroxylase deficiency in childhood. Acta Pediatr.Scand. 280 (Suppl): 1–80 (1989).

3 Michiels L., François B., Raus J. and Vandevyver C. Rapid identification ofPKU-associated mutations by multiplex DGGE analysis of the PAH gene. J. of Inherited Metabolic Disease, 19, 735–738, (1996).

4 Michiels L., François B., Raus J. and Vandevyver C. The identification of 7new mutations in the phenylalanine hydroxylase gene associated withhyperphenylalanemia in the Belgian population. Human Mutation,Supplement 1, S123–124, (1998).

5 Walsh P.S., Metzger D.A. and Higuchi R. Chelex 100 as a medium forsimple extraction of DNA for PCR-based typing from forensic material.BioTechniques 10, 506–513 (1991).

E13

E4

E1

E3

E5

E9

E11

E7

E10

E12

E2

E6

E8

The Polymerase Chain Reaction (PCR) process is covered by patents ownedby Hoffmann-LaRoche. Use of the PCR process requires a license.

Bulletin 2377 2/3/99 10:51 AM Page 3

Life ScienceGroup

Website www.bio-rad.com U.S. (800) 4BIORAD Australia 02 9914 2800 Austria (01)-877 89 01 Belgium 09-385 55 11 Canada (905) 712-2771 China 86-10-62051850/51 Denmark 45 39 17 99 47 Finland 358 (0)9 804 2200 France 01 43 90 46 90 Germany 089 318 84-0 Hong Kong 852-2789-3300 India (91-11) 461-0103 Israel 03 951 4127 Italy 02-21609.1 Japan 03-5811-6270 Korea 82-2-3473-4460 The Netherlands 0318-540666 New Zealand 64-9-4152280 Russia 7-95-4585822 Singapore 65-2729877 Spain (91) 661 70 85 Sweden 46 (0)8 627 50 00 Switzerland 01-809 55 55 United Kingdom 0800-181134

98-373 0998 Sig 072498Bulletin 2377 US/EG Rev A

Bio-RadLaboratories

Bulletin 2377 2/3/99 10:51 AM Page 4

e l e c t ro p h o re s i s tech note 2587

Sjouke Hoving, Hans Voshol, and Jan van Oostrum, Novartis Pharma AG,Functional Genomics Area/Protein Sciences Unit, CH-4002 Basel, Switzerland

IntroductionP roteomics is now generally accepted as an unbiased methodto analyze protein expression; e.g., to elucidate cellularp rocesses at the molecular level or to identify surrogate markersin treatment models. Two-dimensional polyacrylamide gele l e c t ro p h o resis (2-D PAGE) is the heart of pro t e o m et e c h n o l o g i e s , as it is the only method currently capable ofsimultaneously separating thousands of proteins (Rabilloud2 0 0 0 ) . The technique involves the separation of proteins byisoelectric focusing (IEF) in the first dimension. Proteins areseparated in a pH gradient until they reach a stationary positionw h e re their net charge is zero. The pH at which a protein hasz e ro net charge is called its isoelectric point (pI). In the seconddimension, proteins are separated according to their relativemolecular weight by SDS-PAGE (Figure 1). The 2-D PAGEtechnique that is used today originated from the work ofO ’ F a rrell (1975), who used denaturing IEF in the first dimension.

The main drawbacks of the original procedure for 2-D PAGEmethodology were poor reproducibility due to instability of thecarrier ampholyte pH gradient in the first dimension and thelack of sufficient sample loading capacity to visualize low-abundance proteins, which are probably more interesting.The introduction of immobilized pH gradients (IPGs) largelyovercame these problems (Bjellqvist et al. 1982). The IPGsare formed by copolymerization of the pH gradient and thepolyacrylamide support. Since the IPGs are commerciallyavailable (as ReadyStrip pH 3–6, pH 5–8, and pH 7–10 IPGstrips from Bio-Rad), this has dramatically improved thereproducibility of 2-D PAGE and boosted its application inproteomics studies. It has become the method of choice forthe separation of complex mixtures of proteins from tissuesand cells because of its enormous high resolution and thefact that parameters affecting separation are controlledindependently in the 2 dimensions.

H i g h - P e rf o rmance 2-D Gel Electro p h o resis using Narrow pH-RangeR e a d y S t r ipT M IPG Strips

Fig. 1. Principle of 2-D electro p h o resis. A, pre-B lymphoma cell extract (1 mg)was separated by IEF on a ReadyStrip pH 5–8 IPG strip, and stained withB i o - S a feT MCoomassie stain. B, Equilibrated strip was run in the second dimensionby SDS-PAGE (12% acrylamide). The gel was stained with Coomassie® B l u e .

kD

8

200

85

A

B

6pH 3

10pH 3 (nonlinear)

pH 5 pH 7

A

8 10

B C

D

ReadyStrip pH 3–6 and pH 5–8 IPG strips (17 cm; Bio-Rad)were rehydrated overnight in 300 ml buffer containing 1–2 mgprotein (Sanchez et al.1997). Passive rehydration is preferredwhen loading semi-preparative amounts. After rehydration,paper wicks were added to absorb salts. Focusing conditionson the Bio-Rad PROTEAN® IEF cell were: 2 hr at 300 V toremove excess salts, voltage ramping for 4 hr to 10,000 V(current limit per strip was set at 50 A), followed by focusing for5.5 hr at 10,000 V (11.5 hr; 76 kV.hr). The focusing temperaturewas set at 15°C.

It was necessary to rehydrate the ReadyStrip pH 7–10 I P Gstrip in buffer without the protein. After rehydration, pro t e i nsamples were applied through sample cups. In this case aconventional flatbed IEF system was used (Gorg et al. 1988).On this system, focusing conditions are adapted to ensurethat the sample enters the gel with minimal precipitation at theinterface with the cup. The following focusing conditions were

Fig. 2. High-performance 2-D electrophoresisusing narrow pH-range ReadyStrip IPG strips.P roteins were separated by 2-D SDS-PAGE usingIPG strips and were stained with Coomassie Blue.A, 2 mg protein on pH 3–6 IPG; B, 2 mg pro t e i non pH 5–8 IPG; C, 1 mg protein on pH 7–10 IPG;D, 0.8 mg protein on a nonlinear pH 3–10 IPG.With the n a rrow-range strips (A–C), more samplewas loaded, and consequently, more spots wereresolved and detected after staining compare dto the wide-range IPG strip (D).

Detailed Protocol for 2-D PA G ETwo-dimensional electrophoresis was essentially performedaccording to standard protocols (Gorg et al. 1988) with minormodifications as described here. Protein samples wereprepared from the 697 pre-B lymphoma cell line, of which107 cells corresponded to about 1 mg protein. Cells were dire c t l ydissolved in the sample buffer described by Rabilloud (1998).The sample buffer, which contains 7 M urea, 2 M thiourea,4% CHAPS, 1% DTT, 2% Pharmalyte® carrier ampholytespH 3–10 (Pharmacia) for IEF, 0.01% Bromophenol Blue and1 tablet of a cocktail of protease inhibitor (complete; RocheBioscience), was aliquoted, stored at -80°C and thawed onlyonce. Before use, DNase and RNase were added. Proteinsamples were incubated for 15 min at room temperatureand centrifuged for 10 min at 15,000 x g before applying themto the strips.

a c t i n

t u b u l i n

a c t i n e n o l a s ee n o l a s e

g l y c e r a l d e h y d e - 3 - p h o s p h a t ed e h y d ro g e n a s e

s u p e roxide dismutase

s t a t h m i n - P

a c t i n e n o l a s e

s t a t h m i n

HSP 60t u b u l i n

therefore used: 3 hr at 300 V to ensure all proteins enteredthe strip, voltage ramping for 5 hr to 3,500 V, followed byfocusing for 20 hr at 3,500 V (28 hr; 81 kV.h r). The focusingt e m p e r a t u re was set at 20°C.

After focusing, the strips were equilibrated in a solution containing6 M urea, 50 mM Tris.HCl, pH 8.5, 30% glycerol, 2% SDS for12 min at room temperature with 2% DTT and for 6 min with5% iodoacetamide. In the second dimension, strips wereapplied to 12% T, 2.6% C polyacrylamide gels. Gels wereoverlaid with 0.5% agarose solution to ensure good contactbetween the strip and the gel.

Gels were stained with colloidal Coomassie Blue (Anderson 1991)or with silver (Blum et al. 1987). The gels were scanned in a laserdensitometer and image analysis was perf o rmed using Melanies o f t w a re (Appel et al. 1997a, b).

Results and DiscussionOn a standard Coomassie-stained 2-D gel with a wide pH 3-10nonlinear IPG as the first dimension, about 1,500 spots canbe detected with a protein load of about 1 mg. (Figure 2D).However, to improve resolution and sensitivity, narrow-rangeIPG strips should be used. Such strips have higher proteinloading capacity than standard wide-range IPGs (Righetti1990), and consequently, an increased number of protein spotscan be visualized at the same level of detection sensitivity.

F i g u re 2 shows the pH range of 3–10 expanded over 3 part i a l l yoverlapping narrow pH-range ReadyStrip IPGs. The ReadyStrippH 3–6 and pH 5–8 IPG strips were loaded with 2 mg ofprotein, whereas the ReadyStrip pH 7–10 strip was l o a d e dwith only 1 mg protein (by cup-loading at the anodic site).D e p e n d i n g on the pH range, much higher amounts of p ro t e i ncan be loaded, although resolution might be diminished, especially at the m o re alkaline pH range. Our image analysiss o f t w a re d e t e c t e d 1,000 spots (for the IPG 3–6), 2,000 spots(IPG 5–8 ) , and 900 spots (IPG 7–10) on these gels. While theo v e r l a p p i n g pH ranges by definition have data repeatedbetween gels, in total, more spots are easily resolved anddetected on the narrow pH-range ReadyStrip IPGs.

Obviously, in these experiments the possibilities of sensitiveprotein detection have not been exhausted. With silverstaining the detection limit can be lowered, albeit at the costof simplicity and reproducibility. In preliminary experiments,the post-separation fluorescent dye SYPRO® Ruby (availablefrom Bio-Rad) seems to have the optimal combination of easeof use and sensitivity to make it the proteomics dye of thenear future.

ReferencesAnderson L, Two-Dimensional Electro p h o resis: Operation of the ISO-DALT®

System, 2nd edn, Large Scale Biology Press, Washington DC, pp 128–129 (1991)

Appel RD et al., Melanie I I—A third-generation software package for analysis oftwo-dimensional electrophoresis images: I. Features and user interface,Electrophoresis, 18, 2724–2734 (1997a)

Appel RD et al., Melanie II—A third-generation software package for analysis oftwo-dimensional electro p h o resis images: II. Algorithms, Electro p h o resis, 18,2735–2748 (1997b)

Bjellqvist B et al., Isoelectric focusing in immobilized pH gradients: principle,methodology and some applications, J Biochem Biophys Methods, 6,317–339 (1982)

Blum H et al., Improved silver staining of plant proteins, RNA and DNA inpolyacrylamide gels, Electrophoresis, 8, 93–99 (1987)

Gorg A et al., The cur rent state of two-dimensional electrophoresis withimmobilized pH gradients, Electrophoresis, 9, 531-546 (1988)

O’Farrell PH, High resolution two-dimensional electrophoresis of proteins,J Biol Chem, 250, 4007–4021 (1975)

Rabilloud T, Use of thiourea to increase the solubility of membrane proteins intwo-dimensional electrophoresis, Electrophoresis, 19, 758–760 (1998)

Rabillloud T (ed), Proteome Research: Two-Dimensional Gel Electrophoresisand Identification Methods, Springer-Verlag, Berlin (2000)

Righetti PG, Immobilized pH Gradients: Theory and Methodology, ElsevierBiomedical Press, Amsterdam, pp. 5–116 (1990)

Sanchez JC et al., Improved and simplified in-gel sample application usingreswelling of dry immobilized pH gradients, Electro p h o resis, 18, 324–327 (1997)

Coomassie is a trademark of Imperial Chemical Industries PLC. ISO-DALT is atrademark of Large Scale Biology Corporation. SYPROis a trademark ofMolecular Probes, Inc. Bio-Rad is licensed by Molecular Probes to sellSYPROdyes for research use only, under USpatent 5,616,502.

2-DElectrophoresisfor Proteomics

A Methods and Product Manual

2-D electrophoresis for proteomics

sample preparation discussion

the second dimension: SDS-PAGE

detection of proteins in gels

image acquisition and analysis

sample solubilization and preparation methods

first-dimension separation methods

the first dimension: isoelectric focusing (IEF)

detection of proteins on western blots

identification and characterization of 2-D protein spots

references and related Bio-Rad literature

second-dimension separation methods

methods for protein detection in gels

troubleshooting guide

ordering information

ProteomeWorksSystem.com

About This Manual

This manual is a guide to experimental approaches and

methods in proteomics. As a reference tool, it provides

a background on technologies and approaches that

are general to all proteomics studies as well as sample

protocols and tools to use as a starting point for most

proteomics studies. It emphasizes how experimental

conditions can be varied and interpreted to optimize your

results, and also provides an extensive set of specialized

references that you can consult for more information.

Because each sample, experimental approach, and

objective is different, this manual provides overall guidance

on how to develop customized protocols suitable for the

analysis of your samples.

The ProteomeWorks System

Two-dimensional (2-D) electrophoresis is an integral

component of any proteomics program and is the core

separation technology of the ProteomeWorks system.

Because of its ability to separate and resolve complex

mixtures of thousands of proteins in a single gel, 2-D

electrophoresis has become the standard proteomics

separation technique. Proteomics combines 2-D gel

electrophoresis with high-throughput tools for image

analysis, automated protein excision and digestion, and

mass spectrometry identification. The ProteomeWorks

system is a completely integrated system for protein

discovery and proteomics analysis.

The ProteomeWorks system is the globalalliance between Bio-Rad Laboratories, Inc.(USA) and Micromass, Ltd. (UK), dedicatedto furthering proteomics research.

2-D Electrophoresis for Proteomics

A Methods andProduct Manual
















Two-Dimensional Electrophoresis for ProteomicsTable of Contents

Part I. Discussion

Chapter 1 — 2-D Electrophoresis for Proteomics 1The context of proteomics 1Overview of experimental design 1

Chapter 2 — Sample Preparation Discussion 3Solublization 3

Chaotropic agents 3Detergents 4Carrier ampholytes 4Reducing agents 5

Prefractionation 5Sequential extraction 6

Removal of DNA 6Protein load 7Preventing keratin contamination 7Further resources 7

Chapter 3 — The First Dimension: Isoelectric Focusing (IEF) 8Isoelectric point (pI) 8IEF 8

IPGs versus carrier ampholytes 8IPG strips 9Choice of pH gradient ranges 9IPG strip (2-D array) size 10Estimation of pI 10Sample application 10

Sample application during rehydration 11Sample application by cup loading 11

Power conditions and resolution in IEF 11

Chapter 4 — The Second Dimension: SDS-PAGE 13Protein separation by molecular weight (MW) 13Gel composition 13

Single percentage gels 13Gradient gels 13

Precast gels 14Choosing a size format 14 Transition from first to second dimension 14Second dimension and high throughput 15MW estimation 15

Chapter 5 — Detection of Proteins in Gels 16Guidelines for detection of proteins in gels 16

Coomassie Blue staining 16SYPRO Ruby fluorescent staining 16Silver staining 17

Chapter 6 — Detection of Proteins on Western Blots 18Apparatus for blotting 18Membranes and buffers for immunoblotting 18Immunoblotting 19On-membrane detection of glycoproteins after 2-D electrophoresis 20Total protein detection on blots 20

Chapter 7 — Image Acquisition and Analysis 21Image acquisition instruments 21Densitometry 21Storage phosphor and fluorescence scanners 21Computer-assisted image analysis of 2-D

electrophoretic gels 22Spot detection and spot quantitation 22Gel comparison 22Data analysis 22

Chapter 8— Identification and Characterization of 2-D Protein Spots 23Sequence data from 2-D gels 23Integration of image analysis with automated spot cutting 23Automated protein digestion 23Rapid, high-throughput protein identification by MALDI-TOF-MS 24Advanced protein characterization with

ESI-LC-MS-MS and MALDI-MS-MS 25

Part II. Methods

Chapter 9 — Sample Solubilization and Preparation Methods 27Standard sample solubilization solution 27

Urea stock solutions 27Ampholytes for sample solutions 28

Enhanced solubilizing solutions 28Multiple chaotropic agent solution 28Multiple surfactant solution 28

Sequential extraction of proteins 28Sequential extraction protocol 28Endonuclease treatment 29Protein determination for IEF samples 30

Modified Bio-Rad protein assay 30RC DC™ protein assay 31

Chapter 10 — First-Dimension Separation Methods 32Protein load for 2-D gels 32IPG strip rehydration 32

Passive rehydration with sample 32Active rehydration 33

Performing IEF 33Positioning strips and use of wicks 33Focusing conditions for IPG strips on the PROTEAN® IEF cell 34Voltage ramping modes 34

Storage of IPG strips after IEF 34

Chapter 11 — Second-Dimension Separation Methods 35Using precast gels 35Casting SDS-PAGE gels using the multi-casting chambers 35Running multiple gels 35IPG equilibration for the second dimension 36Placement and agarose embedding of IPG strips 36Running the second dimension 37Applying MW standards 37

Chapter 12 — Methods for Protein Detection in Gels 38Coomassie Brilliant Blue R-250 stain 38Bio-Safe™ colloidal Coomassie stain (G-250) 38SYPRO Ruby protein stains 38Bio-Rad silver stain (Merril) 39Silver Stain Plus™ stain 39Gel drying 39Storage of gels in plastic bags 39

Part III. Troubleshooting Guide 41

Part IV. Ordering Information 45

Part V. References and Related Bio-Rad Literature 51

Part I

Discussion









One of the greatest challenges of proteome analysis is thereproducible fractionation of these complex protein mixtureswhile retaining the qualitative and quantitative relationships.Currently, two-dimensional polyacrylamide gel electrophoresis (2-D PAGE) is the only method that can handle this task(Cutler et al. 1999, Fegatella et al. 1999, Gorg et al. 2000),and hence has gained special importance. Since 2-D PAGE iscapable of resolving over 1,800 proteins in a single gel (Choeand Lee 2000), it is important as the primary tool of proteomicsresearch where multiple proteins must be separated for parallelanalysis. It allows hundreds to thousands of gene products tobe analyzed simultaneously. In combination with computer-assisted image evaluation systems for comprehensive qualitative and quantitative examination of proteomes, this electrophoresis technique allows cataloging and comparison of data among groups of researchers.

Overview of Experimental DesignThe general workflow (Figure 1.2) in a 2-D gel based proteomics experiment and some of the factors affecting theway the experiment is performed are outlined below.

Sample Preparation

The method of sample preparation depends on the aim of the research and is key to the success of the experiment. Factors such as the solubility, size, charge, and isoelectricpoint (pI) of the proteins of interest enter into sample preparation. Sample preparation is also important in reducingthe complexity of a protein mixture. The protein fraction to beloaded on a 2-D PAGE gel must be in a low ionic strengthdenaturing buffer that maintains the native charges of proteinsand keeps them soluble. Chapter 2 (pages 3–7) discussessample preparation.

First-Dimension Separation

Proteins are first separated on the basis of their pI, the pH atwhich a protein carries no net charge and will not migrate in anelectrical field. The technique is called isoelectric focusing (IEF).For 2-D PAGE, IEF is best performed in an immobilized pHgradient (IPG). Chapter 3 (pages 8–12) discusses IEF.

Equilibration

A conditioning step is applied to proteins separated by IEFprior to the second-dimension run. This process reducesdisulfide bonds and alkylates the resultant sulfhydryl groupsof the cysteine residues. Concurrently, proteins are coatedwith SDS for separation on the basis of molecular weight(MW). Equilibration is discussed on page 14.

Second-Dimension Separation

The choice for the SDS-PAGE second-dimension gel dependson the protein MW range to be separated, as for 1-D PAGE.The ability to run many gels at the same time and under thesame conditions is important for the purpose of gel-to-gelcomparison. Discussion of second-dimension gels is found inChapter 4 (pages 13–15).

The Context of ProteomicsProteome analysis is a direct measurement of proteins in termsof their presence and relative abundance (Wilkins et al. 1996).The overall aim of a proteomic study is characterization of thecomplex network of cell regulation. Neither the genomic DNAcode of an organism nor the amount of mRNA that isexpressed for each gene product (protein) yields an accuratepicture of the state of a living cell (Lubec et al. 1999), whichcan be altered by many conditions (Figure 1.1). Proteomeanalysis is required to determine which proteins have beenconditionally expressed, how strongly, and whether anyposttranslational modifications are affected. Two or moredifferent states of a cell or an organism (e.g., healthy anddiseased tissue) can be compared and an attempt made toidentify specific qualitative and quantitative protein changes.

Chapter 1 — Two-Dimensional Electrophoresis for Proteomics

Fig. 1.1. Environmental or experimental perturbations can greatly changethe proteins expressed in a cell, even when changes to the DNA code areminor or absent.

Chemicals or drugs

Radiation

1

2

Staining

In order to visualize proteins in gels, they must be stained insome manner. The exceptions are metabolically labeled proteins or iodinated proteins. The choice of staining method isdetermined by several factors, including desired sensitivity, linear range, ease of use, expense, and the type of imagingequipment available. At present there is no ideal universalstain. Sometimes proteins are detected after transfer to amembrane support by western blotting. These topics are discussed in Chapter 5 (pages 16–17) and Chapter 6 (pages 18–20).

Imaging

The ability to collect data in digital form is one of the majorfactors that enables 2-D gels to be a practical means of collecting proteome information. It allows unprejudiced comparison of gels, transfer of information among researchgroups, and cataloging of immense amounts of data. Many types of imaging devices interface with software designedspecifically to collect, interpret, and compare proteomicsdata. Imaging equipment is briefly discussed in Chapter 7(pages 21–22).

Image Analysis

Bio-Rad’s PDQuest™ software and similar image analysis software packages compare gel images, annotate proteinspots, and catalog data. PDQuest also drives Bio-Rad imaging instruments and the ProteomeWorks™ spot cutter. These software packages truly enable proteomic experimentsby making comparison of large sets of data possible.Discussion of image analysis is found on pages 21–22.

Protein Identification

Once interesting proteins are selected by differential analysisor other criteria, the proteins can be excised from gels andidentified. The ability to precisely determine MW by massspectrometry and to search databases for peptide massmatches have made high-throughput protein identificationpossible. One workflow procedure utilizes the ProteomeWorksfamily of products. The ProteomeWorks spot cutter automatically cuts resolved protein spots from gels with highprecision and deposits them in the wells of microplates. Thespot cutter can be operated independently or programmed torun from PDQuest software. The MassPREP station roboticallydestains and digests excised proteins in preparation for massspectrometry. Automatic peptide mass fingerprinting is donewith Micromass’ M@LDI spectrometer. Proteins not identifiedby MALDI can be identified by sequence tagging or de novosequencing using the Q-Tof Ultima electrospray LC-MS-MSworkstation. Alternative procedures can be found in severaltexts (Wilkins et al. 1997, Link 1999, Rabilloud 2000,Pennington and Dunn 2001).

Fig. 1.2. Schematic of workflow involvedin a typical proteomics experiment.

Table 2.1. Reagents for sample preparation by function.

Chapter 2 — Sample Preparation DiscussionEfficient and reproducible sample preparation methods arekey to successful 2-D electrophoresis (Link 1999, Rabilloud1999, Macri et al. 2000, Molloy 2000). Sample preparationmethods range from extraction with simple solubilization solutions to complex mixtures of chaotropic agents, detergents, and reducing agents. Sample preparation caninclude enrichment strategies for separating protein mixturesinto reproducible fractions.

An effective sample preparation procedure will:

1. Reproducibly solubilize proteins of all classes, includinghydrophobic proteins

2. Prevent protein aggregation and loss of solubility during focusing

3. Prevent postextraction chemical modification, including enzymatic or chemical degradation of the protein sample

4. Remove or thoroughly digest nucleic acids and other interfering molecules

5. Yield proteins of interest at detectable levels, which mayrequire the removal of interfering abundant proteins or nonrelevant classes of proteins

Most protein mixtures will require some experimentation todetermine optimum conditions for 2-D PAGE. Variations in theconcentrations of chaotropic agents, detergents, ampholytes,and reducing agents can dramatically affect the 2-D pattern.Some examples follow.

SolubilizationSolubilization of proteins is achieved by the use of chaotropicagents, detergents, reducing agents, buffers, and ampholytes.These are chosen from a small list of compounds that meet therequirements, both chemically and electrically, for compatibilitywith the technique of IEF in IPG strips. The compounds chosen must not increase the ionic strength of the solution, toallow high voltages to be applied during focusing without producing high currents. The various components of samplebuffers for IPG strips are discussed in the following paragraphs.See Chapter 9 for sample preparation procedures and solutions. See Table 2.1 for relevant products from Bio-Rad.Thorough discussion of solubilization methods, including newvariations, can be found in several books (Link 1999,Pennington and Dunn 2000, Rabilloud 2000).

Chaotropic Agents

Urea is the most commonly used chaotropic agent in samplepreparation for 2-D PAGE. Thiourea can be used to help solubilize many otherwise intractable proteins. Urea andthiourea disrupt hydrogen bonds and are used when hydrogenbonding causes unwanted aggregation or formation of secondary structures that affect protein mobility. Urea is typically used at 8 M. Thiourea is weakly soluble in water, butis more soluble in high concentrations of urea, so a mixture of2 M thiourea and 5–8 M urea is used when strongly chaotropicconditions are required (Rabilloud 1998).

3For methods related to this section see page 27

Sample Preparation Products Chaotropic Agent Detergent Reducing Agent Buffer Ampholyte

Individual Components 161-0731 Urea, 1 kg X - - - -

Thiourea X - - - -161-0460 CHAPS, 1 g - X - - -161-0465 CHAPSO, 1 g - X - - -

SB 3-10 - X - - -161-0407 Triton X-100, 500 ml - X - - -161-0611 Dithiothreitol, 5 g - - X - -163-2101 Tributylphosphine, 200 mM, 0.6 ml - - X - -161-0716 Tris, 500 g - - - X163-1112 Bio-Lyte® 3/10 Ampholyte, 40%, 10 ml - - - - X163-1132 Bio-Lyte 3/5 Ampholyte, 20%, 10 ml - - - - X163-1142 Bio-Lyte 4/6 Ampholyte, 40%, 10 ml - - - - X163-1152 Bio-Lyte 5/7 Ampholyte, 40%, 10 ml - - - - X163-1192 Bio-Lyte 5/8 Ampholyte, 40%, 10 ml - - - - X163-1162 Bio-Lyte 6/8 Ampholyte, 40%, 10 ml - - - - X163-1172 Bio-Lyte 7/9 Ampholyte, 40%, 10 ml - - - - X163-1182 Bio-Lyte 8/10 Ampholyte, 20%, 10 ml - - - - X

Bio-Lyte IEF Buffers*163-2093 100x ReadyStrip 7–10 Buffer, 1 ml - - - - X163-2094 100x Bio-Lyte 3/10 Ampholyte, 1 ml - - - - X163-2095 100x ReadyStrip 6.3–8.3 Buffer, 1 ml - - - - X163-2096 100x ReadyStrip 5.5–6.7 Buffer, 1 ml - - - - X163-2097 100x ReadyStrip 4.7–5.9 Buffer, 1 ml - - - - X163-2098 100x ReadyStrip 3.9–5.1 Buffer, 1 ml - - - - X

Solutions or Kits163-2100 ReadyPrep™ Sequential Extraction Kit X X X X X163-2102 ReadyPrep Reagent 1, 1 vial - - - X -163-2103 ReadyPrep Reagent 2, 1 vial X X X X X163-2104 ReadyPrep Reagent 3, 1 vial X X X X X163-2105 ReadyPrep 2-D Starter Kit X X X X X163-2106 ReadyPrep 2-D Starter Kit X X X X X

Rehydration/Sample Buffer

* Dilute ReadyStrip buffers to 1x final in each sample to equal 0.2% Bio-Lyte ampholyte.

Detergents

Detergents are added to disrupt hydrophobic interactions and increase solubility of proteins at their pI. Detergents mustbe nonionic or zwitterionic to allow proteins to migrateaccording to their own charges. Some proteins, especiallymembrane proteins, require detergents for solubilization during isolation and to maintain solubility during focusing.Ionic detergents such as SDS are not compatible with IEF, butcan be used with concentrated samples in situations where theSDS can be unbound from the proteins by IEF-compatibledetergents that compete for binding sites. Nonionic detergentssuch as octylglucoside, and zwitterionic detergents such asCHAPS and its hydroxyl analog CHAPSO, can be used.CHAPS, CHAPSO, or octylglucoside concentrations of 1–2%are recommended (Rabilloud 1999). New detergents areemerging that have great potential in proteomics, including SB 3-10 and ASB-14 (Chevallet et al. 1998). Some proteinsmay require detergent concentrations as high as 4% forsolubility (Hermann et al. 2000).

Carrier Ampholytes

A fundamental challenge with IEF is that some proteins tend to precipitate at their pI. Even in the presence of detergents,certain samples may have stringent salt requirements tomaintain the solubility of some proteins. Salt should be presentin a sample only if it is an absolute requirement, and then onlyat a total concentration less than 40 mM. This is problematicsince any salt included will be removed during the initial high-current stage of focusing. Salt limits the voltage that canbe achieved without producing high current, increasing thetime required for focusing. Proteins that require salt forsolubility are subject to precipitation once the salt is removed.Carrier ampholytes sometimes help to counteract insufficientsalt in a sample. They are usually included at a concentrationof ≤0.2% (w/v) in sample solutions for IPG strips. Highconcentrations of carrier ampholytes will slow down IEF untilthey are focused at their pI, since they carry current andhence limit voltage.

Some researchers have increased resolution by varying theampholyte composition. An example is shown in Figure 2.1,where the resolution in the first dimension is greatly increasedby using a mixture of ampholytes. See Table 2.1 for relevantproducts from Bio-Rad.

pI 5.25 pI 7.75

carrier ampholytes, pH 3–10, 80,000 V-hr

55.5 kD

31.0 kD

55.5 kD

31.0 kD

carrier ampholytes, pH 5–8/8–10 (2:1), 80,000 V-hr

Fig. 2.1. Effect of ampholytes on resolution. Matching sections of 2-Dimages are shown. In both A and B, 110 µg of a cytosolic extract of a human lymphoblastoid cell line was passively loaded into a 17 cm pH 5–8ReadyStrip™ IPG strip. Second-dimension separation was in 10–24% gradientgels with PDA crosslinker in PROTEAN® II XL format. In A, pH 3–10 carrierampholytes were used. In B, pH 5–8 carrier ampholytes were mixed with pH8–10 carrier ampholytes at a 2:1 ratio. The use of the ampholyte mixture greatlyimproved focusing. Data kindly provided by R Joubert-Caron, Laboratoire deBiochimie des Proteines et Proteomique.

pH 5–8 strippI 5.25 pI 7.75

For ordering information related to this section see page 45

A

B

4

For a more thorough discussion of the effects of detergents,denaturing agents, and reducing agents on protein solubility,consult the following papers: Rabilloud (1998, 1999), Herbertet al. (1998), Molloy (2000), and Taylor et al. (2000).

PrefractionationReducing the complexity of the sample loaded on a 2-D gelcan increase the visibility of minor proteins. Techniques suchas differential extraction (Molloy et al. 1998), subcellular fractionation (Taylor et al. 2000, Morel et al. 2000), chromatography (Fountoulakis et al. 1999), or prefocusing in a preparative IEF device such as the Rotofor® system (Masuoka et al. 1998, Nilsson et al. 2000) have been used to reduce the complexity of samples.

Reducing Agents

Reducing agents such as dithiothreitol (DTT) or tributyl-phosphine (TBP) are used to disrupt disulfide bonds. Bond disruption is important for analyzing proteins as single subunits. DTT is a thiol reducing agent added in excess toforce equilibrium toward reduced cysteines. At 50 mM it iseffective in reducing most cystines, but some proteins are notcompletely reduced by this treatment. If the concentration ofDTT is too high it can affect the pH gradient since its pKa isaround 8. Figure 2.2 shows the effect of DTT concentrationon samples of soluble E. coli proteins. The result will be different for samples from different sources.

TBP is a much more effective reducing agent than DTT. It reacts to reduce cystines stoichiometrically at low millimolar concentrations (Herbert et al. 1998). It is chemically more difficult to handle than DTT, but Bio-Rad has solved this problem by supplying it in a form safe for shipping and labuse. See Table 2.1 for these reducing agents from Bio-Rad.

Fig. 2.2. Effect of DTT concentration on 2-D protein spot pattern. A, 200 µg of E. coli extract was suspended in rehydration buffer containing 10 mM DTTand subjected to 2-D gel electrophoresis (first dimension in 11 cm, pH 4–7 IPG for 40,000 V-hr, second dimension in 8–16% SDS-PAGE gel). In B, C, and D, rehydration buffer included 10 mM, 50 mM, and 100 mM DTT, respectively. The gels were stained with Bio-Safe™ Coomassie Blue stain. The images show theacidic, low MW regions of each gel. Notice that as the DTT concentration was increased, the number of spots resolved in this region also increased, indicating that 10 mM DTT is insufficient to completely reduce the disulfides present in the protein mixture. Data kindly provided by William Strong of Bio-Rad Laboratories.

10 mM DTT

50 mM DTT

25 mM DTT

100 mM DTT

For methods related to this section see page 28

A

C

B

D

5

Removal of DNAThe presence of nucleic acids, especially DNA, interferes withseparation of proteins by IEF. Under denaturing conditions,DNA complexes are dissociated and markedly increase theviscosity of the solution, which inhibits protein entry and slowsmigration in the IPG. In addition, DNA binds to proteins in thesample and causes artifactual migration and streaking.

The simplest method for removal of DNA is enzymatic digestion.Adding endonuclease to the sample after solubilization athigh pH (40 mM Tris) allows efficient digestion of nucleic acidswhile minimizing the action of contaminating proteases. The advantage of the endonuclease method is that sample preparation can be achieved in a single step, by the additionof the enzyme prior to loading the first-dimension IPG. See page 29 for the experimental protocol.

Sequential Extraction

One method for reducing sample complexity is the basis ofthe ReadyPrep sequential extraction kit. This protocol takesadvantage of solubility as a third independent means of proteinseparation. Proteins are sequentially extracted in increasingly powerful solubilizing solutions. More protein spots are resolvedby applying each solubility class to a separate gel, therebyenriching for particular proteins while simplifying the 2-D patterns in each gel. An increase in the total number of proteins is detected using this approach (Molloy et al. 1998).

The reagents may be prepared by the protocols provided inChapter 9, or purchased from Bio-Rad as the ReadyPrepsequential extraction kit (see Table 2.1). Each of the 3 reagentssolubilizes an overlapping set of proteins, as illustrated by theflowchart in Figure 2.3 and by the results in Figure 2.4.Reagent 1 extracts soluble proteins, such as cytosolic proteins. Reagent 2 is used to extract proteins of intermediatesolubility, while reagent 3 extracts proteins insoluble inreagents 1 and 2. See page 28 for the protocol.

Fig. 2.4. Increased protein display in sequentially extracted E. colicell pellet. E. coli strain W3110 was collected by centrifugation. The cell pelletwas suspended in reagent 1 and the cells were lysed by sonication. One portionof the sonicated cell suspension, containing 200 µg of protein, was diluted inReadyPrep extraction reagent 3. The proteins soluble in reagent 3 were termeda whole cell extract. Another portion of the sonicated cell suspension was sequentially extracted. First-dimension separation was by IEF from pH 4–7 in anIPG gel. The second-dimension separation was by SDS-PAGE in an 8–16%Tpolyacrylamide gradient gel. A, 2-D PAGE of whole cell extract; B, 2-D PAGE of200 µg of protein solubilized with reagent 1; C, 2-D PAGE of 200 µg of proteinsolubilized with reagent 2; D, 2-D PAGE of 200 µg of protein solubilized withreagent 3.

A B

C D

Fig. 2.3. Flowchart for sequential extraction.

Cell sample

Supernatant 1

Insoluble pellet from extraction 1



Supernatant 2

Supernatant 3

Reagent 1 Reagent 2

Gel 1

Gel 2

Gel 3

Reagent 2

Reagent 3


Protein LoadThe amount of protein applied to an IPG strip (see page 8)can range from several micrograms to 1 mg or more(Bjellqvist et al. 1993b; see page 32). Some of the factorsaffecting the decision of how much protein to load are:

• Subsequent analysis — enough of the protein of interest mustbe loaded for it to be analyzed. With the Ready Gel® minisystem (7 cm IPG), detection of moderately abundant proteinsin complex mixtures with Coomassie Brilliant Blue R-250 dyerequires on the order of 100 µg total protein. With the sameload, many low-abundance proteins can be detected withmore sensitive stains such as silver or SYPRO Ruby proteingel stain. See Table 5.1 on page 16 for sensitivity of stains.

• The purpose of the gel. If the gel is being run solely for thesake of getting a good image of well-resolved proteins forcomparative studies or for publication, the protein load wouldbe the minimum amount that is stainable.

• The abundance of the proteins of interest. If the purpose is tostudy low copy number proteins, a large mass of a protein mixture might be loaded (Wilkins et al. 1998).

• The complexity of the sample. A highly complex sample containing many proteins of widely varying concentrationsmight requre a compromise load so that high-abundanceproteins don’t obscure low-abundance proteins.

By enriching a sample for specific types of proteins using prefractionation techniques, each individual protein will be at a higher relative concentration, which means that enoughmaterial can be loaded for detection of low-abundanceconstitutents. Examples are a fraction obtained by differentialsolubility, a chromatography fraction, a Rotofor® fraction, or any subcellular organelle fraction.

• pH range of IPG strip. In general, larger amounts of total protein can be loaded on a narrow-range IPG strip. Only theproteins with a pI within the strip pH range will be representedwithin the second-dimension gel.

Preventing Keratin ContaminationCareful sample handling is important when sensitive detectionmethods are employed. Silver-stained SDS-PAGE gels sometimes show artifacts in the 50 to 70 kD region and irregular but distinctive vertical streaking parallel to the direction of migration. This has been attributed to the reductionof skin keratin, a contaminant inadvertently introduced into thesamples (keratin in the sample solution usually is focused nearpH 5). Skin keratin is also a common contaminant seen inmass spectra. The best remedy for the keratin artifact is toavoid introducing it into the sample in the first place.Monomer solutions, stock sample buffers, gel buffers, andelectrode buffers should be filtered through nitrocellulose andstored in well-cleaned containers. It also helps to clean thegel apparatus thoroughly with detergent and to wear gloveswhile assembling the equipment.

Further ResourcesSamples can be prepared for 2-D electrophoresis usingmany other techniques. The scope of such a discussion ismuch too broad for this booklet. The resources cited on thefollowing topics should be consulted for further information:

• Cell disruption (Deutscher 1990, Bollag et al. 1996, Link 1999)

• Immunoprecipitation (Harlow and Lane 1988)

• Plant cell sample preparation (Link 1999)

• Subcellular organelles (Celis 1998)

• Microdissection (Celis et al. 1999)

For methods related to this section see page 32 7

IEFWhen a protein is placed in a medium with a pH gradient andsubjected to an electric field, it will initially move toward theelectrode with the opposite charge. During migration throughthe pH gradient, the protein will either pick up or lose protons.As it migrates, its net charge and mobility will decrease andthe protein will slow down. Eventually, the protein will arrive at the point in the pH gradient equal to its pI. There, beinguncharged, it will stop migrating (Figure 3.1). If this proteinshould happen to diffuse to a region of lower pH, it willbecome protonated and be forced back toward the cathodeby the electric field. If, on the other hand, it diffuses into aregion of pH greater than its pI, the protein will become negatively charged and will be driven toward the anode. In thisway, proteins condense, or are focused, into sharp bands inthe pH gradient at their individual characteristic pI values.

Focusing is a steady-state mechanism with regard to pH.Proteins approach their respective pI values at differing ratesbut remain relatively fixed at those pH values for extendedperiods. By contrast, proteins in conventional electrophoresiscontinue to move through the medium until the electric field isremoved. Moreover, in IEF, proteins migrate to their steady-state positions from anywhere in the system.

IPGs versus Carrier Ampholytes

IEF, either using IPG strips or using carrier ampholytes andtube gels, may be used to resolve proteins in the first dimension (Garfin 2000). The IPG method has numerousadvantages (summarized in Table 3.1) over the older tube gelmethod (Görg 1989, Görg 1991). Most proteomics labs arechoosing IPG technology, and for this reason, only this technique will be discussed here.

Isoelectric Point (pI)Differences in proteins’ pI are the basis of separation by IEF.The pI is defined as the pH at which a protein will not migratein an electric field and is determined by the number and typesof charged groups in a protein. Proteins are amphoteric molecules. As such, they can carry positive, negative, or zeronet charge depending on the pH of their local environment. For every protein there is a specific pH at which its net chargeis zero; this is its pI. Proteins show considerable variation in pI,although pI values usually fall in the range of pH 3–12, with themajority falling between pH 4 and pH 7. A protein is positivelycharged in solution at pH values below its pI and negativelycharged at pH values above its pI.

3.9

7.1

8.4

7.18.8

7.18.4 8.8 8.88.4

3 10

+

Before focusing

After focusing

–

3.93.93.73.7

5.35.3 7.17.17.18.88.88.8

8.48.48.4

5.3 5.33.7 3.73.9

Fig. 3.1. A mixture of proteins is resolved on a pH 3–10 IPG strip according toeach protein’s pI and independently of size, as described in the IEF section.

Chapter 3 — The First Dimension: Isoelectric Focusing (IEF)

Table 3.1. Advantages of IPG strips over tube gels for first-dimension IEF.

ReadyStrip IPG Strips Tube Gels with Carrier Ampholytes

ReproducibilitySupplied ready to use Poured by the user

Computer-controlled gradient formation Ampholyte-based self-forming internal gradients

pH gradient is covalently incorporated into acrylamide matrix and immobilized pH gradient may drift under conditions of:• Protein overload• Extended run length

Tightly controlled pH gradient and gel length (± 2 mm) for consistent Variable gel lengths because of difficulty in casting gel tubesfirst-dimension separations

Ease of UsePlastic backing supports acrylamide matrix Fragile low-percentage acrylamide matrix

Stored in freezer until ready to use Tube gels cast prior to experiment

Self-centering strips are applied easily to the top of standard second-dimension gels Gels must be extruded from tubes onto the top of second-dimension gels

Large protein loads can be applied Limited amount of sample can be applied

ThroughputUp to twelve 11 cm or 17 cm strips or twenty-four 7 cm strips in each run Sample loaded at beginning of run

Rehydration/equilibration trays allow for rehydration of one sample while Very difficult and fragile application of tube gel onto second-dimension gelanother is being focused


IPG Strips A stable, linear, and reproducible pH gradient is crucial to successful IEF. IPG strips offer the advantage of gradient stability over extended focusing runs (Bjellqvist et al. 1982).IPG strips are much more difficult to cast than carrierampholyte gels (Righetti 1983); however, IPG strips are commercially available, for example as ReadyStrip™ IPG strips.(See Table 3.2 for gradients and sizes currently available.)

pH gradients for IPG strips are created with sets of acrylamidobuffers, which are derivatives of acrylamide containing bothreactive double bonds and buffering groups. The general structure is CH2=CH–CO–NH–R, where R contains either acarboxyl [–COOH] or a tertiary amino group (e.g., –N(CH3)2).These acrylamide derivatives are covalently incorporated intopolyacrylamide gels at the time of casting and can formalmost any conceivable pH gradient (Righetti 1990).

Choice of pH Gradient RangesUse of broad-range strips (pH 3–10) allows the display of mostproteins in a single gel. With narrow-range and micro-rangeoverlapping gradient strips, resolution is increased by expanding a small pH range across the entire width of a gel.Since many proteins are focused in the middle of the pH range3–10, some researchers use nonlinear (NL) gradients to betterresolve proteins in the middle of the pH range and to compressthe width of the extreme pH ranges at the ends of the gradients. However, overlapping narrow-range and micro-range linear IPG strips can outperform a nonlinear gradient and display more spots per sample (see Figure 3.2).This result is due to the extra resolving power from use of anarrower pI range per gel. Use of overlapping gradients alsoallows the ability to create “cyber” or composite gels bymatching spots from the overlapping regions using imagingsoftware. Because proteins outside of the pH range of the

6pH 3 pH 5

pH 3

pH 7

A

8

10

10

B C

D

Fig. 3.2. Narrow overlapping IPG stripsoutperform nonlinear IPG strips. Proteins were separated by 2-D PAGE using 17cm IPGstrips, then stained with Coomassie Blue. A, 2 mg protein on pH 3–6 ReadyStrip IPG; B, 2 mg protein on pH 5–8 ReadyStrip IPG; C, 1 mg protein on pH 7–10 ReadyStrip IPG; D, 0.8 mg protein on a nonlinear pH 3–10 IPG.With the narrow-range strips (A–C), more sample was loaded, and consequently, morespots were resolved and detected after stainingcompared to the wide-range IPG strip (D).Kindly provided by Sjouke Hoving, Hans Voshol,and Jan van Oostrom of Novartis Pharma AG,Functional Genomics Area, Basel, Switzerland.

actin

tubulin

actin enolaseenolase

glyceraldehyde-3-phosphatedehydrogenase

superoxide dismutase

stathmin-P

actin enolase

stathmin

HSP 60tubulin

(nonlinear)

For ordering information related to this section see pages 45–46 9

7 cm ReadyStrip IPG Strips

strip are excluded, more total protein mass can be loaded perstrip, allowing more proteins to be detectable. Figure 3.2demonstrates the resolution achieved using 3 overlappinggradients with 17 cm ReadyStrip IPG strips and PROTEAN®

XL gels.

IPG Strip (2-D Array) SizeThe 17 cm IPG strips and large-format gels have a large area to resolve protein spots; however, they take a long time torun. Using a mini system instead of, or as a complement to, a large gel format can provide significant time savings. A minisystem is perfect for rapid optimization of sample preparationmethods. Switching to a large format then allows thoroughassessment of a complex sample and identification of proteins of interest. In many cases, a mini system consisting of narrow-range IPG strips can then be used to focus in on the proteins of interest.

Throughput of the 2-D process is a consideration in choosinggel size. Table 4.2 (page 15) compares size formats by experimental run times and equipment available. Table 3.2shows the appropriate ReadyStrip IPG strip sizes for each ofBio-Rad’s second-dimension gel formats. The ability to castor run 12 gels at a time in any of 3 size formats is very useful in gathering proteomic results. In some cases, mini systems (7 cm ReadyStrip IPG strips with Mini-PROTEAN® 3 formatgels, or 11 cm ReadyStrip IPG strips with Criterion™ precastgels) can completely replace large 2-D systems, providingspeed, convenience, and ease in handling. The availability of

narrow and micro overlapping pH-range ReadyStrip IPGstrips can increase the effective width of pI resolution morethan 5-fold after accounting for overlapping regions. When 3narrow-range overlapping ReadyStrip IPG strips are used withthe Criterion system, the resolution in the first dimension isincreased from 11 to 26 cm. When micro-range strips areused, the resolution in the first dimension is expanded from11 to 44 cm.

Estimation of pIThe pI of a protein can be estimated by comparing the positionof the protein spot of interest to the position of known proteinsor standards separated across the same pH gradient(Bjellqvist et al. 1993a, Garfin 2000). ReadyStrip IPG strips contain linear gradients, so the pI of an unknown protein canbe estimated by linear interpolation relative to proteins ofknown pI.

Sample ApplicationCommercial IPG strips are dehydrated and must be rehydratedto their original gel thickness (0.5 mm) before use. This allowsflexibility in applying sample to the strips. There are 3 methodsfor sample loading: passive in-gel rehydration with sample,active in-gel rehydration with sample, or cup loading of sample after IPG rehydration. Introducing the sample whilerehydrating the strips is the easiest, and in most cases themost efficient, way to apply sample. In some specificinstances, it is best to rehydrate the strips and then applysample through sample cups while current is applied. Each method is discussed in the following sections.

Table 3.2. ReadyStrip IPG strips available from Bio-Rad.

For methods related to this section see pages 32–3310

11 cm ReadyStrip IPG Strips

17 cm ReadyStrip IPG Strips Catalog # Second-Dimension Gel

3 10 163-20073 10 (NL) 163-20094 7 163-20083 6 163-20105 8 163-2011 PROTEAN XL, PROTEAN Plus, or7 10 163-2012 PROTEAN Ready Gel precast gels3.9 5.1 163-20204.7 5.9 163-20215.5 6.7 163-20226.3 8.3 163-2023

3 10 163-20143 10 (NL) 163-20164 7 163-20153 6 163-20175 8 163-2018 Criterion precast gels7 10 163-2019 Gels hand cast in empty Criterion cassettes3.9 5.1 163-20244.7 5.9 163-20255.5 6.7 163-20266.3 8.3 163-2027

3 10 163-20003 10 (NL) 163-20024 7 163-20013 6 163-20035 8 163-2004 Mini-PROTEAN 3 gels7 10 163-2005 Ready Gel precast gels3.9 5.1 163-20284.7 5.9 163-20295.5 6.7 163-20306.3 8.3 163-2031

Sample Application During Rehydration

For both active and passive rehydration methods, the sampleis introduced to the IPG strip at the time of rehydration. As the strips hydrate, proteins in the sample are absorbedand distributed over the entire length of the strip (Sanchez et al. 1997).

In the case of active rehydration, a very low voltage is appliedduring rehydration of the strips. Proteins enter the gel matrixunder current as well as by absorption. The PROTEAN IEFcell has preprogrammed methods designed to accommodateactive rehydration (see page 33 for the protocol). Active rehydration is thought to help large proteins enter the strip byapplying electrical “pull”. Because the voltage is applied beforeall the solution and proteins are absorbed into the gel, the pHof a protein’s environment will be the pH of the rehydrationbuffer, and the protein will move according to its mass-to-charge ratio in that environment. Thus, small proteins with ahigher mobility have a higher risk of being lost from the strip.

With passive rehydration, proteins enter the gel by absorptiononly (see page 30 for the protocol). This method allows efficient use of equipment since strips can be rehydrated insample rehydration trays while other samples are beingfocused in the IEF cell.

Whether the strips are hydrated actively or passively, it is veryimportant that they be incubated with sample for at least 11 hrprior to focusing. This allows the high molecular weight proteinstime to enter the gel after the gel has become fully hydratedand the pores have attained full size. These sample applicationmethods work because IEF is a steady-state technique, soproteins migrate to their pI independent of their initial positions.

The advantages of this approach are:

• Sample application is simple (Görg et al. 1999)

• Sample application during rehydration avoids the problem of sample precipitation, which often occurs with cup loading(Rabilloud 1999)

• Shorter focusing times can be used because the sampleproteins are in the IPG strip prior to IEF

• Very large amounts of protein can be loaded using this method

Sample Application by Cup Loading

Cup loading can be beneficial in the following cases (Cordwellet al. 1997, Görg et al. 2000):

• When samples contain high levels of DNA, RNA, or otherlarge molecules, such as cellulose

• For analytical serum samples that have not been treated toremove albumin

• When running basic IPG strips; e.g., pH 7–10

• For samples that contain high concentrations of glycoproteins

Because of its relative difficulty and tendency towardartifacts, cup loading should be avoided if possible. When loading the protein sample from a cup, the IPGstrips must be rehydrated prior to sample application. The IPG strips can be rehydrated in a variety of ways. We recommend the rehydration tray, although IPG strips are often rehydrated in 1 or 2 ml pipets that have beensealed at both ends with Parafilm. Sample volumes of upto 100 µl can be loaded later onto each gel strip using asample cup.

Power Conditions and Resolution in IEFDuring an IEF run, the electrical conductivity of the gelchanges with time, especially during the early phase. When an electrical field is applied to an IPG at the beginning of an IEF run, the current will be relatively highbecause of the large number of charge carriers present. As the proteins and ampholytes move toward their pIs, thecurrent will gradually decrease due to the decrease in thecharge on individual proteins and carrier ampholytes.

The pH gradient, strip length, and the applied electricalfield determine the resolution of an IEF run. According toboth theory and experiment, the difference in pI between 2 adjacent IEF-resolved protein bands is directly proportionalto the square root of the pH gradient and inversely proportional to the square root of the voltage gradient atthe position of the bands (Garfin 2000).

Thus, narrow pH ranges and high applied voltages yieldhigh resolution in IEF. The highest resolution can beachieved using micro-range IPG strips and an electro-phoretic cell, such as the PROTEAN IEF cell, capable ofapplying high voltages. IEF runs should always be carriedout at the highest voltage compatible with the IPG stripsand electrophoretic cell. However, high voltages inelectrophoresis are accompanied by large amounts of generated heat. The magnitude of the electric field that canbe applied and the ionic strength of the solutions that canbe used in IEF are limited. Thin gels are better able to dissipate heat than thick ones and are therefore capable ofwithstanding the high voltage that leads to higher resolution.Also, at the completion of focusing, the current drops tonearly zero since the carriers of the current have stoppedmoving. The PROTEAN IEF cell is designed to provide precise cooling, allowing the highest possible voltages to beapplied. (A default current limit of 50 µA per strip is intendedto minimize protein carbamylation reactions in urea samplebuffers. This limit can be increased to 99 µA per strip.)


Consistent and reproducible focusing requires that the time integral of voltage (volt-hours) be kept consistent. It is usually necessary to program IEF runs to reach finalfocusing voltages in stages. This approach clears ionic constituents in the sample from the IPG strips while limitingelectrical heating of the strips. The PROTEAN IEF cell allowsfor multistep runs at durations set by time or volt-hours. Some suggested starting electrical conditions and voltageramping options are discussed on page 32.

The number of volt-hours required to complete a run must bedetermined empirically. A simple E. coli extract can yieldequally acceptable 2-D results when 7 cm strips are run foreither 8,000 V-hr or 22,000 V-hr, when 11 cm strips are run foreither 20,000 V-hr or 41,500 V-hr, or when 17 cm strips are

run for either 50,000 V-hr or 85,600 V-hr. This result is sample-and buffer-dependent. A more complex sample in terms ofnumber of proteins or even a different sample buffer mightrequire increased volt-hours.

The time needed to achieve the programmed volt-hoursdepends on the pH range of the IPG strip used as well assample and buffer characteristics. Table 3.3 shows the variabilityin time required to run the same sample and rehydration solution on different strip sizes and pH gradients. These datasupport running similar strips and samples in batches. If different strips are run at the same time, the electrical conditions experienced by individual strips will be different,perhaps exposing some strips to more current than desired,since the total current limit is averaged over all strips in a tray.

Table 3.3. The time to reach programmed volt-hours varies with the pH gradient.

Strip Size; V-hr Programmed pH 3–10 pH 3–6 pH 4–7 pH 5–8 pH 7–10

7 cm, 8,000 V-hr 2 hr 30 min 3 hr 45 min 2 hr 1 hr 30 min 2 hr

11 cm, 20,000 V-hr 4 hr 50 min 6 hr 30 min 3 hr 45 min 3 hr 45 min 3 hr 45 min

17 cm, 50,000 V-hr 7 hr 20 min 9 hr 6 hr 30 min 6 hr 6 hr 30 min

Average amount of time for the PROTEAN IEF cell to achieve the programmed volt-hours. The time required varied by up to 20% between trials. Each strip containedthe same E. coli sample in the same rehydration solution. The current was limited to 50 µA per strip using a rapid ramp. The time is extended when the voltage is limitedby high current. In most cases, the voltage never reaches the maximum voltage set.


Protein Separation by Molecular Weight (MW)Second-dimension separation is by protein mass, or MW,using SDS-PAGE. The proteins resolved in IPG strips in thefirst dimension are applied to second-dimension gels and separated by MW perpendicularly to the first dimension.

The pores of the second-dimension gel sieve proteinsaccording to size because dodecyl sulfate coats all proteinsessentially in proportion to their mass. The net effect is thatproteins migrate as ellipsoids with a uniform negative charge-to-mass ratio, with mobility related logarithmically to mass(Garfin 1995). See Figure 4.1.

Fig. 4.1. Schematic diagram showing separation of proteins by SDS-PAGEafter separation by IEF.

Gel CompositionHomogeneous (single-percentage acrylamide) gels generallygive excellent resolution of sample proteins that fall within anarrow MW range. Gradient gels have 2 advantages: theyallow proteins with a wide range of MW to be analyzed simultaneously, and the decreasing pore size along the gradient functions to sharpen the spots.

Single-Percentage Gels

The percentage of acrylamide, often referred to as %T (total percentage of acrylamide plus crosslinker) determinesthe pore size of a gel. Most protein separations use 37.5 parts acrylamide to 1 part bis-acrylamide (bis). Some researchers substitute piperazine diacrylamide (PDA)for bis, which can reduce silver staining background and give higher gel strength. The higher the %T, the smaller thepore size. A suitable %T can be estimated from charts ofmobility for proteins of different MW. Table 4.1 shows the MWranges resolved on gels of different acrylamide percentages.

Gradient Gels

Gradient gels are cast with acrylamide concentrations thatincrease from top to bottom so that the pore size decreasesas proteins migrate further into the gels. As proteins movethrough gradient gels from regions of relatively large pores toregions of relatively small pores, their migration rates slow.Small proteins remain in gradient gels much longer than theydo in single-percentage gels that have the same average %T,so both large and small molecules may be resolved in thesame gel. This makes gradient gels popular for analysis ofcomplex mixtures that span wide MW ranges. A gradient gel,however, cannot match the resolution obtainable with a properly chosen single concentration of acrylamide. A goodapproach is to use gradient gels for estimates of the complexities of mixtures. A proteomics experiment might start out with an 8–16%T gradient for global comparison. After interesting regions of the 2-D array have been identified,a new set of single-percentage gels may be run to study a particular size range of proteins.

It is simplest and often most cost and labor effective to purchase commercially available precast gradient gels. Bio-Rad offers a full line of gels for 2-D PAGE (Table 4.1) aswell as devices to cast gels of different sizes in multi-castingchambers (Table 4.2).

Chapter 4 — The Second Dimension: SDS-PAGE

8.88.88.8

pH 3

MW

10

IEF-focused proteins

SDS-charged proteins in IPG strip

SDS-charged proteins resolved according to size in SDS-PAGE gel

7.17.17.1

7.1

3.93.9

3.9

8.48.48.4

8.4

3.73.7

3.7

5.35.3

5.3

8.8

3.93.93.73.7

5.35.3 7.17.17.18.88.88.8

8.48.48.4

Equilibrate in SDS and reducing agent to giveuniform protein shape, single subunits,

uniform negative charge/mass ratio

Apply to SDS-PAGE gel

For ordering information related to this section see page 46 13

Transition from First to Second DimensionThe transition from first-dimension to second-dimension gelelectrophoresis involves 2 steps: equilibration of the resolvedIPG strips in SDS reducing buffer, and embedding of the stripon the top of the second-dimension gel. Proper equilibrationsimultaneously ensures that proteins are coated with dodecylsulfate and that cysteines are reduced and alkylated. A method for IPG strip equilibration is discussed on page 34.The equilibrated IPG strips are placed on top of the gel andfixed with molten agarose solution to ensure good contactbetween the gel and the strip.

Table 4.1. Second-dimension precast gels from Bio-Rad.

Gel Format Catalog # Composition Protein Mass Range Resolved

PROTEAN II Ready Gel Precast Gels 161-1450 10% 30–150 kDTris-HCl, IPG well 161-1451 12% 20–100 kD(For use with 17 cm ReadyStrip IPG strips; 161-1452 10–20% 10–100 kDgel size 18.3 x 19.3 cm) 161-1453 8–16% 10–100 kD

345-0013 10% resolving, 4% stacking gel 30–150 kDCriterion Precast Gels 345-0018 12.5% resolving, 4% stacking gel 20–100 kDTris-HCl, IPG well 345-0031 4–15% 20–250 kD(For use with 11 cm ReadyStrip IPG strips; 345-0036 4–20% 10–200 kDgel size 13.3 x 8.7 cm) 345-0041 8–16%, 4% stacking gel 10–100 kD

345-0046 10–20%, 4% stacking gel 10–100 kD

161-1390 10% resolving, 4% stacking gel 30–150 kDReady Gel Precast Gels 161-1391 12% resolving, 4% stacking gel 20–100 kDTris-HCl, IPG well 161-1392 4–15% 20–250 kD(For use with 7 cm ReadyStrip IPG strips; 161-1393 4–20% 10–200 kDgel size 7.4 x 6.8 cm) 161-1394 8–16% resolving, 4% stacking gel 10–100 kD

161-1395 10–20% resolving, 4% stacking gel 10–100 kD

Precast GelsHigh-quality precast gels are preferred for high-throughputapplications. They provide savings in time and labor, and theprecision-poured gradients result in reproducibility amongruns. Bio-Rad offers precast gels with IPG wells in 3 size formats to hold 3 lengths of ReadyStrip™ IPG strips. A full listof the gels currently available with IPG wells can be found inTable 4.1.

Precast gels differ from handcast gels in that they are castwith a single buffer throughout and without SDS. During storage, different buffers in the stacking and resolving gelswould mingle without elaborate means to keep them separate,and thus have no practical value. Also, because the sample contains SDS, and the dodecyl sulfate ion in the cathodebuffer moves faster than the proteins in the gel, keeping them saturated with the detergent, precast gels are madewithout SDS.

Choosing a Size Format

Bio-Rad offers complete second-dimension systems for 3 gelsizes as detailed in Table 4.2. Mini-PROTEAN® 3, Criterion™,and PROTEAN® XL formats have precast gels available. A fourth PROTEAN Plus™ size is available to run in the PROTEAN Plus Dodeca™ cell and can be cast as 20 x 20.5cm gels or as 25 x 20.5 cm gels. The large gel can hold one17 cm or two 11 cm ReadyStrip IPG strips.

The large-format PROTEAN Ready Gel® precast gels may be run in either the PROTEAN XL cell (2 gels per run), the PROTEAN XL multi-cell (6 gels per run), or the PROTEANPlus Dodeca cell (12 gels per run). The Criterion precast gelscan be run in the Criterion cell (2 gels per run) or the CriterionDodeca cell (up to 12 gels per run). Mini-PROTEAN 3 orReady Gel precast gels can be run in the Mini-PROTEAN 3 cell (2 gels per run) or the Mini-PROTEAN 3 Dodeca cell (up to 12 gels per run).

Fig. 4.2. The Dodeca cells. Clockwise from upper left: Mini-PROTEAN 3 Dodecacell, Criterion Dodeca cell, and PROTEAN Plus Dodeca cell.


Second Dimension and High ThroughputSince the first dimension can be run in batches of 12–24strips at a time, it is desirable to run the same number ofsamples in the second dimension. Precast gels ensure highreproducibility among samples and help reduce the workinvolved in running large numbers of samples. Alternatively, gelscan be hand cast 12 at a time under identical conditions withmulti-casting chambers. The Dodeca cells save time, space,and effort, and help to ensure that gels are run under the sameelectrical conditions for highest throughput and reproducibility.

MW EstimationThe migration rate of a polypeptide in SDS-PAGE is inverselyproportional to the logarithm of its MW. The larger the poly-peptide, the more slowly it migrates in a gel. MW is determinedin SDS-PAGE by comparing the migration of protein spots tothe migration of standards. Plots of log MW versus the migration distance are reasonably linear. Gradient SDS-PAGEgels can also be used to estimate MW. In this case, log MWis proportional to log (%T). With linear gradients, %T is proportional to distance migrated, so the data can be plottedas log MW vs. log (migration distance).

Standard curves are actually sigmoid. The apparent linearityof a standard curve may not cover the full MW range for agiven protein mixture in a particular gel. However, log MWvaries sufficiently slowly to allow fairly accurate MW estimatesto be made by interpolation, and even extrapolation, over relatively wide ranges (Garfin 1995).

Mixtures of standard proteins with known MW are availablefrom Bio-Rad in several formats for calibrating the migrationof proteins in electrophoretic gels. Standards are availableunstained, prestained, or with tags for development with various secondary reagents (useful when blotting). Standards can be run in a reference well, attached to the end of a focused IPG strip by filter paper, or directly embedded in agarose onto the second-dimension gel (see method on page 37).

Table 4.2. Size formats for second-dimension electrophoresis.

PROTEAN PlusMini-PROTEAN 3 System Criterion System PROTEAN II XL System Dodeca Cell System

ReadyStrip IPG strip 7 cm 11 cm 17 cm 17cm or 2 x11 cm or 3 x 7cm

Precast second-dimension gel Ready Gel Criterion PROTEAN II Ready Gel PROTEAN II Ready Gel(W x L) precast gels precast gels precast gels precast gels

7.4 x 6.8 cm 13.3 x 8.7 cm 18.3 x 19.3 cm 18.3 x 19.3 cm

Handcast second-dimension gel Mini-PROTEAN 3 Criterion PROTEAN II XL PROTEAN Plus(W x L) handcast gels empty cassettes handcast gels handcast gels

7.4 x 6.8 cm 13.3 x 8.7 cm 18.3 x 19.3 cm 20 x 20.5, 25 x 20.5 cm

Cell formats available Mini-PROTEAN 3 cell Criterion cell PROTEAN II XL cell PROTEAN Plusruns 2 gels runs 2 gels runs 2 gels Dodeca cell

Mini-PROTEAN 3 Dodeca cell Criterion Dodeca cell PROTEAN II XL multi-cell runs 12 gelsruns 12 gels runs 12 gels runs 6 gels

second-dimension run time 35 min 1 hr 5–6 hr 5–6 hr

Total electrophoresis run time* 5.5–7.5 hr 5.75–8 hr 13–15 hr 13–15 hr

* Does not include 12 hr rehydration or 30 min equilibration time for strips, or staining time for gels.


Guidelines for Detection of Proteins in GelsGels are run for either analytical or preparative purposes. The intended use of the gel determines the amount of proteinto load and the means of detection. It is most common tomake proteins in gels visible by staining them with dyes ormetals. Each type of protein stain has its own characteristicsand limitations with regard to the sensitivity of detection andthe types of proteins that stain best (see Table 5.1).Sometimes proteins are transferred to membranes by westernblotting to be detected by immunoblotting, glycoprotein analysis, or total protein stain (see Chapter 6).

If the purpose of gel electrophoresis is to identify low-abundanceproteins (e.g., low-copy-number proteins in a cell extract, orcontaminants in a purification scheme), then a high proteinload (0.1–1 mg/ml) and a high-sensitivity stain, such as silveror a fluorescent stain, should be used (Corthals et al. 2000).When the intention is to obtain enough protein for use as an antigen or for sequence analysis, then a high protein loadshould be applied to the gel and the proteins visualized with a staining procedure that does not fix proteins in the gel.Quantitative comparisons require the use of stains with broad linear ranges of detection.

The sensitivity that is achievable in staining is determined by: 1), the amount of stain that binds to the proteins; 2), theintensity of the coloration; 3), the difference in colorationbetween stained proteins and the residual background in thebody of the gel (the signal-to-noise ratio). Unbound stain molecules can be washed out of the gels without removingmuch stain from the proteins.

All stains interact differently with different proteins (Carroll etal. 2000). No stain will universally stain all proteins in a gel in proportion to their mass. The only observation that seemsto hold for most stains is that they interact best with basicamino acids. For critical analysis, replicate gels should bestained with 2 or more different stains. Of all stains available,colloidal Coomassie Blue (Bio-Safe™ Coomassie) appears tostain the broadest spectrum of proteins. It is instructive, especially with 2-D PAGE gels, to stain a colloidal CoomassieBlue-stained gel with silver or to stain a fluorescently stainedgel with colloidal Coomassie Blue or silver. Very often, this

double staining procedure will show a few differencesbetween the protein patterns. It is most common to stain gelsfirst with Coomassie Blue or a fluorescent stain, then restainwith silver. However, the order in which the stains are useddoes not seem to be important, as long as the gels arewashed well with high-purity water between stains.

Coomassie Blue Staining

Coomassie Brilliant Blue R-250 is the most common stain for protein detection in polyacrylamide gels. Coomassie Brilliant Blue R-250 and G-250 are wool dyes thathave been adapted to stain proteins in gels. The “R” and “G”designations indicate red and green hues, respectively.

Coomassie R-250 requires on the order of 100 ng of proteinper spot for detection. Absolute sensitivity and staining linearitydepend on the proteins being stained. The staining solutionalso fixes most proteins in gels.

Bio-Safe Coomassie stain is made with Coomassie BrilliantBlue G-250. Bio-Safe Coomassie stain is a ready-to-use, single-reagent protein stain. Sensitivity can be down to 10 ng,and greater contrast is achieved by washing the gel in waterafter staining. Used stain can be disposed of as nonhazardouswaste and the procedure does not fix proteins in the gel.

SYPRO Ruby Fluorescent Staining

SYPRO Ruby protein gel stain has desirable features thatmake it popular in high-throughput laboratories. It is anendpoint stain with little background staining (high signal-to-noise characteristics) and it is sensitive and easy to use.SYPRO Ruby protein stain does not detect nucleic acids.

SYPRO Ruby protein stain is sensitive to 1–10 ng and can belinear over 3 orders of magnitude. It is compatible with high-throughput protocols and downstream analysis, including massspectrometry and Edman sequencing (Patton 2000). It alsoallows detection of glycoproteins, lipoproteins, low MW proteins, and metalloproteins that are not stained well byother stains. This fluorescent stain is easily visualized withsimple UV or blue-light transilluminators, as well as by theMolecular Imager FX Pro Plus™ multiimager and VersaDoc™

imaging system (see pages 21–22).

Table 5.1. Characteristics of protein stains.

Gel Stain Sensitivity Process Time/# Steps Advantages

SYPRO Ruby protein gel stain 1 ng 3 hr/2 steps Mass spectrometry compatible; linear over 3 orders of magnitude;allows protein analysis in fluorescent imagers

Bio-Safe Coomassie stain 10 ng 2.5 hr/3 steps Mass spectrometry compatible; easily visualized; nonhazardousSilver Stain Plus™ stain 1 ng 1.5 hr/3 steps Mass spectrometry compatible; high sensitivity; low backgroundBio-Rad silver stain 1 ng 2 hr/7 steps High sensitivity; detects some highly glycosylated and other

difficult-to-stain proteinsCoomassie Blue R-250 40 ng 2.5 hr/2 steps Oldest and least expensive method

Chapter 5 — Detection of Proteins in Gels


Silver Staining

Two popular methods for silver staining are recommended for2-D analysis. They are based on slightly different chemistriesbut have similar sensitivities for protein. Bio-Rad’s silver stainkit, based on the method of Merril et al. (1981), can be asmuch as 100 times more sensitive than Coomassie Blue R-250dye staining and allows visualization of heavily glycosylatedproteins in gels. Protein spots containing 10–100 ng of proteincan be easily seen. Proteins in gels are fixed with alcohol andacetic acid, then oxidized in a solution of potassium dichromatein dilute nitric acid, washed with water, and treated with silvernitrate solution. Silver ions bind to the oxidized proteins andare subsequently reduced to metallic silver by treatment with alkaline formaldehyde. Color development is stoppedwith acetic acid when the desired staining intensity has beenachieved. This method is not compatible with massspectroscopic analysis since the oxidative step changesprotein mass. See page 39 for more information on the method.

The Silver Stain Plus stain from Bio-Rad requires only onesimultaneous staining and development step and is based on the method developed by Gottlieb and Chavko (1987).Proteins are fixed with a solution containing methanol, aceticacid, and glycerol, and washed extensively with water. The gels are then soaked in a solution containing a silver-ammine complex bound to colloidal tungstosilicic acid. Silver ions transfer from the tungstosilicic acid to the proteinsin the gel by means of an ion exchange or electrophilicprocess. Formaldehyde in the alkaline solution reduces the silver ions to metallic silver to produce the images of proteinspots. The reaction is stopped with acetic acid when thedesired intensity has been achieved. Because silver ions donot accumulate in the bodies of gels, background staining islight. Since this method lacks an oxidizing step, visualizationof heavily glycosylated proteins and lipoproteins can be lesssensitive than with the Merril stain. This method is better foruse in proteomics when the end goal is identification by mass spectrometric analysis. See page 37 for more information onthe method.


Certain synthetic membranes bind proteins tightly enough thatthey can be used as supports for solid-phase immunoassays,staining, or other solid-phase analysis. These membranes arecollectively known as western blots (Bers and Garfin 1985,Garfin and Bers 1989, Ledue and Garfin 1997). Bound proteinsretain their antigenicity and are accessible to probes. Several techniques have been developed to probe proteins bound tosynthetic membranes.

Apparatus for BlottingA protein is electrotransferred from a gel to an adjoiningmembrane by directing an electric field across the gel. There are 2 types of apparatus for electrotransfer, buffer-filledtanks and semi-dry transfer devices.

Transfer tanks are made of plastic with 2 electrodes mountednear opposing tank walls. A nonconducting cassette holds the membrane in close contact with the gel. The cassetteassembly is placed vertically into the tank, parallel to the electrodes, and submerged in electrophoresis buffer. A largevolume of buffer in the tank dissipates the heat generated during transfer. Table 6.1 lists specifications for tank blotters.

In semi-dry blotting, the gel and membrane are sandwichedhorizontally between 2 stacks of buffer-wetted filter paper in

direct contact with 2 closely spaced solid plate electrodes.The close spacing of the semi-dry apparatus provides highfield strengths. The term “semi-dry” refers to the limitedamount of buffer, which is confined to the stacks of filterpaper. See Table 6.2 for Bio-Rad’s Trans-Blot SD semi-dryblotter specifications.

With tanks, transfers are somewhat more efficient than withsemi-dry devices. Under semi-dry electrotransfer conditions,some low MW proteins are driven through the membranes,and because low buffer capacity limits run times, some highMW proteins are poorly transferred. Conversely, the liquid intank blotters can be efficiently cooled, allowing slower transfers without heat buildup. Slower transfer conditions can allow the time needed for large proteins to move throughthe gel matrix, but the lower intensity allows small proteins to remain attached to the membrane after transfer.

Membranes and Buffers for ImmunoblottingThe 2 membranes most used for protein immunoblotting work are nitrocellulose and polyvinylidene fluoride (PVDF). Both bind proteins at about 100–200 µg/cm2. In addition,PVDF can be used when proteins are to be sequenced. It can withstand the harsh chemicals of protein sequencers,whereas nitrocellulose cannot.

Table 6.2. Bio-Rad’s Trans-Blot SD semi-dry blotter.

Blotter Number of Gels/Run Electrode Material Highlights(single layer, more can be run in stacks) and Surface Area

Trans-Blot SD 1 PROTEAN II XL Platinum-coated titanium anode Durable electrode pair with optimal electrochemical properties6 Mini-PROTEAN 3 gels Stainless-steel cathode Spring-loaded electrodes give uniform pressure for consistent results3 Criterion gels (18 x 25 cm) Easy to assemble and use

Voltage limit of 25 V

Table 6.1. Bio-Rad’s tank blotters.

Mini Trans-Blot® Cell Criterion™ Blotter Trans-Blot® Cell

Number of gels per run • 2 Mini-PROTEAN 3 gels • 2 Criterion gels Low intensity• 4 Mini-PROTEAN 3 gels • 3 PROTEAN II XL gels

• 6 Criterion gels• 12 Mini-PROTEAN 3 gelsHigh intensity• 1 PROTEAN II XL gel• 2 Criterion gels• 4 Mini-PROTEAN 3 gels

Buffer volume 450 ml 1.3 L 2.5 LElectrode material Wire Wire or plate Wire or plateElectrode distance 4 cm 4.3 cm 8 cm or 4 cmTime 1 hr 30 min to overnight 30 min to overnightTemperature control Bio-Ice™ block Sealed ice block or cooling coil Cooling coil

Chapter 6 — Detection of Proteins on Western Blots

For ordering information related to this section see pages 49–5018

Fig. 6.1 A. General color detection system. 1. Antigen-specific primary antibody binds to protein of interest2. Enzyme-conjugated secondary antibody or binding

protein binds to primary antibody3. Enzyme (E) converts added substrate (S) to colored product (P),

which precipitates onto the membrane surface

Fig. 6.1 B. Immun-Star™ chemiluminescent detection kit.1. Alkaline phosphatase (AP)-conjugated secondary antibody binds to

primary antibody2. Chemiluminescent substrate reacts with AP to emit light3. Film or phosphor screen is exposed to emitted light

Fig. 6.1 C. Amplified AP Immun-Blot® kit.1. Biotinylated secondary antibody binds to primary antibody2. Complex of streptavidin and biotinylated AP binds to biotin of secondary antibody3. Multiple AP molecules are now available to convert substrate (S) to colored

product (P), which precipitates


ImmunoblottingDevelopment of immunoblots will range in complexity, depending on the method. Figure 6.1 illustrates 3 methods of development; each is available as a kit from Bio-Rad. The general procedure is summarized below:

1. Proteins are transferred from a 2-D gel to a membrane surface. The transferred proteins become immobilized on the surface of the membrane in a pattern that is an exactreplica of the gel.

2. Unoccupied protein-binding sites on the membrane are saturated to prevent nonspecific binding of antibodies. This step is called either blocking or quenching.

3. The blot is probed for the protein of interest with a specificprimary antibody.

4. The blot is probed a second time. The second probe is anantibody that is specific for the primary antibody type and is conjugated to a detectable enzyme. The site of the protein of interest is thus tagged with an enzyme through the specificities of the primary and secondary antibodies.

5. Enzyme substrates that are converted into detectable products are incubated with the blot. The enzyme productsindicate the positions of the proteins of interest.

19

O O

O

O

O

O

O

Terminal nonreducing monosaccharide

Oxidize with periodate

Label with biotin-hydrazide

Incubate with streptavidin-AP

Incubate with NBT/BCIP

Color development

Colordevelopment

Biotin Biotin

Biotin-SAP Biotin-SAP

Biotin-SAP Biotin-SAP

Fig. 6.2. Schematic illustration of glycoprotein detection chemistryusing Bio-Rad’s Immun-Blot glycoprotein detection kit.

On-Membrane Detection of Glycoproteins after 2-D ElectrophoresisEukaryotic proteins often appear on 2-D PAGE gels as “trains”of spots that differ in apparent pI, MW, or both. These areusually isoforms of the same protein and result from a varietyof posttranslational modifications, including glycosylation.

The initial step, once a blot has been prepared, is to identifywhich spots are glycoproteins so that they can be furthercharacterized. Various methods have been developed for thedetection of glycoproteins on 2-D gels and blots by color andlectin analysis, and these can be carried out at the analyticallevel. The actual level of detection of course depends on theextent of glycosylation of the protein, since the reagents reactonly with the carbohydrate moiety.

The Immun-Blot glycoprotein detection kit is based on the initial oxidation of the carbohydrate with periodic acid. Periodicacid oxidizes vicinal diols on terminal monosaccharides todialdehydes. Biotin hydrazide is coupled to these aldehydegroups. The biotinylated glycoproteins are then detected bycoupling to streptavidin-alkaline phosphatase (AP) followed by reaction with a color-development substrate system (Figure 6.2). The limit of detection by this method is about 0.5 µg of glycoprotein containing a single N-linked oligosaccharide.

Total Protein Detection on BlotsFor proper identification of the proteins of interest in a blot,immunodetected proteins or glycoproteins must be comparedto the total protein pattern of the gel. This requires the indiscriminate staining of all the proteins in the blot. Colloidal gold stain is a very sensitive reagent for total proteinstaining. It is a stabilized solution of colloidal gold particles.The gold particles bind to proteins on the surfaces of membranes. Detection limits are in the 100 pg range and canbe enhanced an order of magnitude by subsequent treatmentwith silver (Dunn 1999).

Coomassie Blue R-250 and Bio-Safe Coomassie stains areother popular total protein stains. Researchers blotting 2-D PAGE gels particularly favor them since they are compatible with mass spectrometry. Stained blots are well suited for the archiving of 2-D PAGE separations. SYPRO Rubyprotein blot stain is a very sensitive total protein stain that isformulated for blots.


Image Acquisition InstrumentsBefore 2-D gels can be analyzed with an image evaluationsystem, they must be digitized. The most commonly useddevices are camera systems, densitometers, phosphorimagers, and fluorescence scanners. All of Bio-Rad’s imagingsystems (Table 7.1) are seamlessly integrated with PDQuest™

software, and they can export and import images to and from other software via TIFF files.

DensitometryDensitometers compare the intensity of a light beam beforeand after attenuation by a sample.

The GS-800™ calibrated imaging densitometer (Figure 7.1) hasbeen customized for analysis of gels, autoradiograms, andblots. The transmittance and true reflectance capabilities allowaccurate scans of samples that are either transparent (gelsand film) or opaque (blots). It provides high-quality imaging toresolve close spots and a variable resolution feature to previewand crop images. Wet 2-D gels may be scanned with red,green, and blue color CCD technology on the watertight platen.

Storage Phosphor and Fluorescence ScannersDigitization of 2-D gels stained with fluorescent dyes orradioactive compounds requires specific imaging systems(Patton 2000).

The Molecular Imager FX Pro Plus™ system (Figure 7.2) is flexible and expandable. 2-D gels of radiolabled proteins canbe imaged using a Kodak phosphor screen more rapidly andaccurately than with film. Popular proteomic fluorescent stains,including SYPRO Ruby protein gel and blot stains and SYPROOrange protein gel stain can be imaged with single-color andmulticolor fluorescence via direct laser excitation. This systempermits detection of almost any fluorophore that is excited in the visible spectrum. The internal laser and external laseroptions allow optimal excitation of single-color or multicolorfluorescent samples. Computer-controlled, user-accessible filter wheels have 8 filter slots, allowing detection of manymulticolor combinations of dyes (e.g., Gingrich et al. 2000).

Table 7.1 Bio-Rad imaging systems.

Instrument Detection Detects

Molecular Imager FX Pro Plus multiimager Isotopic, • SYPRO Ruby protein stains (for gels or blots)fluorescent • SYPRO Orange protein gel stain

• Radioisotopes

Molecular Imager FX Pro™ fluorescent imager Fluorescent • SYPRO Ruby protein stains (for gels or blots)• SYPRO Orange protein gel stain

VersaDoc™ Model 5000 imaging system Fluorescent, • SYPRO Ruby protein stains (for gels or blots)chemiluminescent, • SYPRO Orange protein gel stain

colorimetric • Autoradiographs• Silver, Coomassie R-250, or Bio-Safe™ Coomassie stain

VersaDoc Model 3000 imaging system Fluorescent, • SYPRO Ruby protein stains (for gels or blots)chemiluminescent, • SYPRO Orange protein gel stain

colorimetric • Autoradiographs• Silver, Coomassie R-250, or Bio-Safe Coomassie stain

GS-800 calibrated imaging densitometer Colorimetric, • Coomassie R-250 or Bio-Safe Coomassie stainisotopic (on film) • Silver stain

• Autoradiographs

Chapter 7 — Image Acquisition and Analysis

Fig. 7.2. Molecular Imager FX Pro Plus multiimager system.

Fig. 7.1. GS-800 calibrated imaging densitometer.


The Molecular Imager FX Pro™ system has all the featues ofthe FX Pro Plus for fluorescent dectection, but without thestorage phosphor option. The VersaDoc model 1000, model3000, and model 5000 imaging systems are able to produceexceptionally high-quality images from single-color or multicolorfluorescent sources, including samples stained with SYPRORuby protein gel and blot stains and SYPRO Orange proteingel stain. Chemiluminescent western blots can be detectedwithout the use of film. Gels stained with Coomassie dyes orsilver stains as well as autoradiographs can be imaged bydensitometry or with a multiimaging system.

Computer-Assisted Image Analysis of 2-D Electrophoretic GelsComputer-assisted image analysis software is an indispensabletool for the evaluation of complex 2-D gels. It allows:

• Storage and structuring of large amounts of collected experimental image data

• Rapid and sophisticated analysis of experimental information

• Supplementation and distribution of data among labs

• Establishment of 2-D-protein data banks

Image analysis systems deliver error-fee comprehensive qualitative and quantitative data from a large number of 2-D gels (Miller 1989).

PDQuest software from Bio-Rad is a popular analysis tool. Gel analysis of digitized gel images includes spot detection,spot quantitation, gel comparison, and statistical analysis.PDQuest software has the further advantage of seamlessintegration with any of Bio-Rad’s image acquisition instruments,as well as the ability to control the ProteomeWorks™ spot cutterdescribed in Chapter 8 (see Figures 8.1 and 8.2). Theadvanced annotation feature can be used to label spots withtext, URL links, document links, or mass spectrometry data.

Spot Detection and Spot Quantitation Before the software automatically detects the protein spots of a 2-D gel, the raw image data are corrected and the gelbackground is subtracted. The process is executed with simple menus and “wizards”.

PDQuest software models protein spots mathematically as 3-D Gaussian distributions and uses the models to determineabsorption maxima. This enables automatic detection and resolution of merged spots. Following this procedure, spotintensities are obtained by integration of the Gaussian function. The mathematical description of the spots is usedboth for data reduction and for increasing evaluation speed,since reevaluation of data after an image change takes onlyfractions of a second.

The hit rate of automatic spot detection is highly dependenton the quality of the 2-D gels. Correction capabilities ofPDQuest software can be used to add undetected spots tothe list of spots or to delete spots that arise from gel artifacts.

Gel ComparisonThe next step in 2-D gel evaluation is the identification of proteins that are present in all gels of a series. This task ismade difficult primarily because of inherent irreproducibility ingels, which affects the positions of spots within a gel series.Gel analysis software must detect minor shifts in individualspot position within the gel series. Many software packages forautomatic gel comparison are created with the assumption thatthe relative positions of spots are altered only slightly relative toeach other, and allocate the spots on this basis. Prior to automatic gel comparison, PDQuest software selects thebest 2-D gel of a gel series as a reference or standard gel andcompares all other 2-D gels to this gel. Proteins in a gel seriesthat are not present in the reference gel are added manuallyso that the reference gel will include all proteins of a gel series.

Before the software can detect and document matching ofdifferent spots, a number of landmarks, or identical spots inthe gel series, must be manually identified. The landmarkingtool speeds the process by making “best guess” assignmentsof landmark spots to images in the gel series. With PDQuestsoftware, it is possible to simultaneously display up to 100enlarged details of 2-D gels on the screen. This simultaneousdisplay of all 2-D gels of a test series enables rapid and error-free determination of the fixed points.

Using the landmarks, the image analysis software first attemptsto compare all spots lying very near these fixed points andthen uses the matched spots as starting points for furthercomparisons. Thus, the entire gel surface is systematicallyinvestigated for the presence or absence of matching spots ina gel series. The results of the automatic gel comparisonrequire verification, as does automatic spot detection. Two tools assist this verification process in PDQuest: Either identical protein spots are labeled with matching lettersand allocated section by section, or the deviations in the spot positions of a particular 2-D gel can be displayed as lines thatshow spot shifts in comparison to the reference gel.

Data AnalysisWith PDQuest software, all gels of an experiment are viewed as a unit. To compare gels from different experiments, the reference images are compared. In such comparisons, eachspot is automatically assigned a number so that identicalspots have identical numbers. Experimental data can also beanalyzed statistically — both parametric and nonparametrictests are available.


Sequence Data from 2-D Gels2-D electrophoresis has the virtually unique capability ofsimultaneously displaying several hundred gene products. 2-D gels are an ideal starting point for protein chemical identification and characterization. Peptide mass fingerprint orsequence data can be derived following 2-D electrophoresiswith mass spectrometry or amino acid sequence analysis(Eckerskorn et al. 1988, Ducret et al. 1996). The sensitivity ofcurrently available instruments makes 2-D electrophoresis anefficient “preparative” analytical method. Most current proteinidentification depends on mass spectrometry of proteinsexcised from gels or blots.

Integration of Image Analysis with Automated Spot CuttingImage analysis software obtains quantitative and qualitativeinformation about the proteins in sample, and stores the information in files, which may also contain additional annotations (Figure 8.1).

The ProteomeWorks™ spot cutter (Figure 8.2) expands thecapabilities of proteome labs by integrating PDQuest™ imageanalysis software. The image analysis files acquired byPDQuest direct automated spot cutting. Excised proteinspots are deposited into microtiter plates ready for furtherautomated processing. PDQuest software tracks the proteinspots through spot cutting and protein identification.Downstream protein spot identifications are generallyobtained from peptide mass fingerprint analysis using massspectrometry.

The ProteomeWorks spot cutter is a precision instrument with a small benchtop footprint. It is fully automatic to increasethroughput and minimize the amount of hands-on time spentexcising protein spots. The spot cutter individually exciseseven overlapping spots for unique identification.

Automated Protein DigestionThe ProteomeWorks spot cutter eliminates the first of twobottlenecks for excision and enzymatic digestion of proteinspots. Driven by PDQuest software, it enables automatedspot excision and deposition of cut gel spots into microtiterplate wells. Isolated proteins from the gel pieces are thendigested to release peptides for detailed sequence analysisby mass spectrometry, leading to protein identification.

Excised gel spots can be robotically destained, chemicallymodified (reduction and alkylation), and digested in preparationfor either MALDI-TOF-MS or ESI-MS-MS mass spectrometrywith the Micromass MassPREP station. Each process is executed under fully automated software control with a rangeof standard protocols enabling high throughput and flexibility.

Manual protein digestion is a tedious, time-consuming processthat is subject to variability and keratin contamination. Automationof this process with the MassPREP station eliminates a significantbottleneck for high-throughput protein identification.

Chapter 8 — Identification and Characterization of 2-D Spots

Fig. 8.1. PDQuest software’s increased flexibility includes simultaneous display of multiple annotation categories.

Fig. 8.2. ProteomeWorks spot cutter.


Operational features of the MassPREP station include a variable temperature control for optimized reduction, alkylation, and digestion of proteins and on-board coolingcapabilities for reagents and peptide digests to ensure reproducible digestion results. The station employs a varietyof sample cleaning technologies (MassPREP targets andMillipore ZipTip pipet tips) to prepare peptide digests prior to automated deposition of samples onto a M@LDI orMassPREP target plate. Contamination of peptide samples is also minimized with the MassPREP clean air enclosure. For more information visit the Micromass web site atwww.micromass.co.uk

Rapid, High-Throughput Protein Identification byMALDI-TOF-MSPeptide mass fingerprinting of protein digest products usingmatrix assisted laser desorption ionization time of flight massspectrometry (MALDI-TOF-MS) provides an ideal method for protein identification when samples have been separatedby 2-D PAGE. The M@LDI HT is one of a new generation of networked “2-D gel-MS” analyzers for high-throughputprotein identification. M@LDI HT is the primary MS dataacquisition device of the ProteomeWorks system, and featuresa fully automated target plate auto-changer for increasedthroughput. Networking enables distribution of data capture,protein assignment, and result presentation functions ofProteinLynx Global SERVER software within a secure client-server architecture, maximizing computing power to quicklyidentify proteins.

The M@LDI HT enables automated acquisition of optimizedmass spectra and the derivation of monoisotopic peptidemass fingerprint information. Interrogation of multiple FASTAdatabases using Global SERVER software following captureof MS results provides rapid identification of proteins that fitthe samples’ peptide mass fingerprint, along with aconfidence score indicating the validity of the identifications. Following MS identification, peptide mass fingerprint spectra and all of the identification results are availablethrough electronic reports. In addition, protein identificationresults are seamlessly integrated with the gel image inPDQuest software. Figure 8.5 shows the data for a MS identification and how this information is accessible inPDQuest as an annotation to a specific gel spot by simplyclicking on the spot.

24

Fig. 8.3. MassPREP station: the in-gel digestion and sample preparation robotof the ProteomeWorks system.

Fig. 8.4. M@LDI HT, the “2-D gel-MS” analyzer of the ProteomeWorksSystem. Q-Tof micro™ with capillary HPLC optimized for high-sensitivity LC-MS-MS in post genomic studies.

Using this system, the working time to process data fromspot cutting to protein digestion to MS analysis and imageannotation is reduced by over 50% compared to manual processing of gel samples, with a corresponding reduction in error. All of the instrumentation and software in this process is part of the integrated ProteomeWorks system, a set ofpowerful tools for proteomic analysis.

Advanced Protein Characterization with ESI-LC-MS-MS and MALDI-MS-MSMALDI-TOF MS provides an ideal high-throughput solution for protein identification; however, where protein identity isambiguous, known databases must be searched with a higher degree of sequence information. The Micromass Q-Tof family of MS-MS instruments incorporates quadrupole/orthogonal acceleration time-of-flight (Q/oa-TOF) technology,enabling exact mass measurement and acquisition of thehighest level peptide sequence information for de novosequencing and BLAST analyses. The Q-Tof Ultima family ofMS-MS instruments provides a flexible research platform foroptimal results with either API LC-MS-MS or MALDI MS-MS,or with a combined platform for both API and MALDI MS-MS.

Protein digest samples in microtiter plates, prepared with the MassPREP station, can be transferred directly to theMicromass CapLC (capillary HPLC) system for automatedinjection into the Q-Tof micro for integrated LC-MS-MS underMassLynx software control. The capability for MS to MS-MSswitching “on the fly” with the Q-Tof family of instruments maximizes the amount of amino acid sequence informationthat can be generated with these instruments. MassSeq software also provides the capacity for automated de novoamino acid sequencing based on the MS results. For more information, contact www.micromass.co.uk

25

Fig. 8.6. The Q-Tof™ Ultima family of electrospray LC-MS-MS systems provides the tools of choice for proteomics in the postgenomic era.

Fig. 8.5. The integration of ProteinLynx Global SERVER, the Micromass proteomics application manager and web-enabled database search engine,with PDQuest software enables annotation of gel images with protein identityand retrieval of MS data and database search results for all spots of interest.

Part II

Methods





27

Protein solubilization is sample dependent. Several solubilizationsolutions and protocols are detailed in this chapter. See pages 3–5 for background information on sample solubilization.

Standard Sample Solubilization SolutionThe sample solubilization solution described in Table 9.1 is commonly used as a general extraction solution and providesa good starting point for sample preparation. This solution isalso used for IPG strip rehydration, for sample application byin-gel rehydration or cup loading, and for sequential extractionof more complex samples. It is available from Bio-Rad asreagent 2 in the ReadyPrep™ sequential extraction kit (see page 28). A similar IPG rehydration/sample buffer ideal for E. coli samples can be ordered as ReadyPrep 2-D starter kitrehydration/sample buffer. Each vial in the starter kit reconstitutes to 10 ml of 8 M urea, 2% CHAPS, 40 mM DTT,0.2% Bio-Lyte™ 3/10 ampholyte.

Table 9.1. Sample solubilization solution (ReadyPrep reagent 2).

Dissolve the urea in about 25 ml of water with stirring. Add the

CHAPS, ampholytes, and Bromophenol Blue and adjust the solution to

a final volume of 50 ml. Tris may be added at 10–40 mM if pH control

is important. Tris will increase the conductivity and extend the time

required to focus the IPG strips. Add DTT or TBP immediately before

use. Use carrier ampholytes that span the pH range of the IPG strip

according to Table 9.4. Urea stock solutions should be used soon

after they are made, or treated with a mixed-bed ion exchange resin

to avoid protein carbamylation by cyanate, which forms in old urea.

Table 9.2 and the section on urea stock solutions describe the

preparation and storage of urea stock solutions. Store sample

solubilization solution in aliquots at -20ºC. Thaw only the required

number of aliquots and discard leftover solution. Add sample solution

to the protein sample so the final concentration of urea is ≥6.5 M.

Solid urea may be added as necessary. Proteins may be directly

extracted in sample solubilization solution using at least 9 ml of

solution for each 1 ml of packed cell pellet. Use sample solubilization

solution to rehydrate IPG strips.

Table 9.2. Urea stock solution, 8.5 M.

Dissolve the urea in about 600 ml of water with gentle heating

(<30°C) and vigorous stirring with a heavy stirbar. Remove from the

heat source as soon as the urea dissolves. Add 5 g of Bio-Rad

deionizing resin (Bio-Rex® 501-X8) and stir for 10 min. If the resin

de-colors, add an additional 5 g and repeat until the resin no

longer loses color. Filter the solution through Whatman No.1 paper

using a Buchner funnel.

Store convenient aliquots of this urea solution at -20°C until

required. This deionized 8.5 M stock can be used to make up all

urea-containing solutions. Do not store urea solutions at room

temperature (or 4°C) any longer than necessary. Urea in solution

exists in equilibrium with ammonium cyanate, which can cause

irreversible protein modification and interfere with IEF.

Urea Stock Solutions

Urea is a chaotropic agent comonly used in IEF samplepreparation (see discussion on page 3). To prepare an 8.5 Murea stock solution, see Table 9.2. High-purity urea should beused for IEF. Charged species can be removed by addition ofa mixed-bed ion exchange resin. The resin is then removedby filtration. Urea should not be heated above 30°C because carbamylation of the sample may occur, which gives rise tocharged artifacts detected in the second-dimension gel.

For some applications, it is convenient to prepare a saturatedurea solution (9.8 M) containing 4% CHAPS (Table 9.3). By diluting samples with the 9.8 M urea solution, the final urea andCHAPS concentrations remain high even when large volumes ofaqueous protein sample must be denatured. The solutionshould be stored frozen in aliquots. Thaw enough for use whenneeded. Add reducing agent and ampholytes immediatelybefore use. Discard unused reagent.

Table 9.3. Urea/CHAPS stock solution.

Dissolve the urea in about 25 ml of water with stirring. Add the CHAPS

and adjust the final volume to 50 ml. Store in aliquots at -20°C.

Urea Stock Solution Components Amount

8.5 M Urea 510 g

Water Adjust to 1 L

Components Amount

9.8 M Urea 29.4 g

4% CHAPS 2 g

Water Adjust to 50 ml

Chapter 9 — Sample Solubilization and Preparation Methods

For background information related to this section see pages 3–7

Components Amount

8 M Urea 47 ml of 8.5 M stock (Table 9.2) or 24 g dry urea dissolved in 25 ml of H2O

50 mM DTT or 385 mg or 2 mM TBP 500 µl of 200 mM TBP stock

4% CHAPS 2 g

0.2% Carrier ampholytes See Table 9.4

0.0002% Bromophenol Blue* 100 µl of 0.1% stock


* Bromophenol Blue is included in trace amounts in rehydration solutions both toview the rehydration of the strip with the solution and to observe the earlystages of electrophoresis. It is not required for solubilization.

Ampholytes for Sample Solutions

Ampholytes are added to all IPG rehydration buffers andsample solubilization solutions to help keep proteins soluble.The choice of ampholytes is dependent on the pH range ofthe IPG strip. Table 9.4 provides a starting point. See page 4for further discussion.

Table 9.4. Suggested Bio-Lyte® ampholyte composition for IPG use.

Enhanced Solubilizing SolutionsRecent improvements to protein solubilization include additionof multiple chaotropic agents, multiple surfactants, or both tothe solubilizing solution. For some samples, solutions thathave strong chaotropic properties extract more proteins thanthe standard sample solution (see page 3). In other instances,proteins are better solubilized by multiple surfactants (seepage 4). Each protein sample is different, and the most efficient solubilizing solution can only be determined by trial.

Multiple Chaotropic Agent Solution

This solution can be made according to the directions inTable 9.5. Tris may be added at 10–40 mM if pH control isimportant for a particular sample. Tris will increase the conductivity and extend the time required to focus the IPGstrips. The reducing agent TBP should be added immediatelyprior to use.

Table 9.5. Multiple chaotropic agent solution preparation.

Multiple Surfactant Solution

Reagent 3 in the Bio-Rad ReadyPrep sequential extraction kit is a multiple surfactant-containing solution and can be purchased from Bio-Rad (see Table 2.1). Alternatively, it canbe made according to the directions in Table 9.6.

The Tris in this solution inactivates some proteases. It may beomitted, in which case the conductivity will be reduced andthe focusing time will be shortened. The reducing agent TBPshould be added immediately prior to use.

Table 9.6. Multiple surfactant solution (ReadyPrep reagent 3).

Sequential Extraction of ProteinsSee page 6 for background information and a flowchart related to sequential extraction.

Sequential Extraction Protocol

Sequential extraction reagents are available premade in Bio-Rad's ReadyPrep sequential extraction kit. Alternatively, thecomposition of reagents is described in Table 9.7. The goal of the first step in sequential extraction is to lyse the cells of interest directly in sequential extraction reagent 1 using a physical lysis procedure. The method of lysis will varydepending on cell type. Follow standard procedures for cellor tissue growth and harvesting, and for physical cell lysis.For bacteria, growth medium can be washed away with reagent 1 before cells are lysed by sonication or French press.Animal tissues can be homogenized or sonicated in reagent 1.Plant tissues can be ground in liquid nitrogen and the resultingpowder suspended in reagent 1. The 2-D results can beimproved greatly by a nuclease treatment at this step.Thorough lysis is important for the best segregation of the different solubility classes of proteins. It may be necessary toseparate organelles or to ensure their lysis; these approachesprevent organelle lysis during subsequent solubilization stepsfrom affecting the protein classes obtained.

Components Amount

5 M Urea 1.5 g dry urea and 3 ml of water to dissolve, or 2.9 ml of 8.5 M stock as prepared in Table 9.2

2 M Thiourea 760 mg

2 mM TBP 50 µl of 200 mM TBP stock

2% CHAPS 100 mg

2% SB 3-10 100 mg


40 mM Tris 24.2 mg

0.0002% Bromophenol Blue 10 µl of 0.1% stock


Components Amount

7 M Urea 2.1 g dry urea and 3 ml of water to dissolve, or 4.1 ml of 8.5 M stock as prepared in Table 9..2

2 M Thiourea 760 mg

2 mM TBP 50 µl of 200 mM TBP stock

4% CHAPS 200 mg


0.0002% Bromophenol Blue* 10 µl of 0.1% stock


* Bromophenol Blue is included in trace amounts in rehydration solutions both toview the rehydration of the strip with the solution and to observe the earlystages of electrophoresis. It is not required for solubilization.

IPG Bio-Lyte Ampholyte (Stock) Sample Solution VolumepH Range Range Conc. (w/v) per 5 ml per 50 ml

3–10 3/10 40% 25 µl 250 µl

4–7 4/6 40% 12.5 µl 125 µl

5/7 40% 12.5 µl 125 µl

3–6 3/5 20% 25 µl 250 µl

4/6 40% 12.5 µl 125 µl

5–8 5/8 40% 25 µl 250 µl

7–10 7/9 40% 12.5 µl 125 µl

8/10 20% 25 µl 250 µl

28 For ordering information related to this section see page 45

Extraction 3

12. Prepare reagent 3 by adding 10 µl of TBP per 1 ml of reagent 3

from the ReadyPrep extraction kit or as the multiple surfactant

solution described on page 26.

13. Optional step: Wash the pellet twice with reagent 2 and discard the

supernatants.

14. Add the same volume of reagent 3 to the pellet from step 13 as

the volume of solution 2 added in step 8.

15. Vortex and centrifuge as described for the second extraction.

16. Centrifuge and recover the supernatant and label it as extract 3.

Determine the protein concentration.

17. Freeze the supernatant until ready for use in aliquots appropriate

for a typical experiment (in duplicate).

18. The pellet can be further extracted with SDS to dissolve highly

insoluble proteins for protein analysis in 1-D SDS-PAGE.

Endonuclease TreatmentAdd 150–300 U of endonuclease to each 1 ml of sample dissolved in the selected solubilization solution. Incubate thesample at room temperature for at least 20 min. The amountsand times necessary will vary depending on the sample.

For a discussion on the removal of nucleic acids, see page 6.

* Optional: Add 150 U of endonuclease and mix. Incubate for 20 min at room temperature.

29For background information related to this section see page 6

Table 9.7. Sequential extraction solutions.

Extraction 1

1. Place the appropriate amount of sample to yield 50 mg of

protein for each 1 ml of solution 1 into a suitable lysis vessel.

The amount of material needed will vary according to the type of

sample and must be determined empirically.

2. Lyse the cells according to standard protocols.*

3. Centrifuge to yield a firm pellet. A microcentrifuge spin at

12,000 x g for 8 min at room temperature is usually sufficient.

4. Collect the supernatant, label it as extract 1, and reserve the

pellet for further extraction (in step 8). Determine the

concentration of the supernatant as described on page 28.

5. Freeze the supernatant until ready for use in aliquots appropriate

for a typical experiment (in duplicate).

6. When ready to load IPG strips, dilute the supernatant in reagent

2 (made according to step 7 below). Use an appropriate volume

for the passive rehydration of IPG strips in the sample.

Extraction 2

7. Prepare reagent 2 by adding 10 µl of 200 mM TBP to each 1 ml

of reagent 2 from the ReadyPrep sequential extraction kit, or

prepare the sample solubilization solution described on page 27.

8. Wash the pellet from step 4 twice in reagent 1: vortex and

centrifuge, then discard each supernatant. Add 1/2 volume

of reagent 2 (compared to the volume of reagent 1 used in

step 1) to the washed pellet. (To track extraction efficiencies and

protein yields, determine the protein concentration of all washes

before discarding them.)

9. Vortex the mixture for 5 min; some samples may also require

sonication, or aspiration through a fine-gauge needle.*

10. Centrifuge the mixture to yield a firm pellet and a clear supernatant.

A microcentrifuge spin at 12,000 x g for 8 min at room

temperature is usually sufficient.

11. Collect the supernatant, label it as extract 2, and determine its

protein concentration. Reserve the pellet for further extraction

(in step 13). Freeze the supernatant until ready for use in aliquots

appropriate for a typical experiment (in duplicate).

Composition Amount

Reagent 140 mM Tris 96.8 mg Tris in 20 ml water

Reagent 28 M urea, 4% CHAPS, 2 mM TBP, See Table 9.1, Sample 40 mM Tris, 0.2% ampholytes solubilization solution

Reagent 35 M urea, 2 M thiourea, 2% CHAPS, See Table 9.6, Multiple 2% SB 3-10, 2 mM TBP, 40 mM Tris, surfactant solution0.2% ampholytes

Protein Determination for IEF SamplesProtein assays are generally sensitive to detergents or reducingagents used at the concentrations found in typical samplesolutions. The 2 assays described below are adaptations to the Bradford (modified Bio-Rad protein assay) or Lowry(RC DC™ protein assay) protein assays that compensate forthe interfering substances. The modified Bio-Rad proteinassay procedure is simple to perform, but the mechanism forthis protein assay, dye binding, will give variable results withdifferent proteins. However, it will give consistent results withina protein mixture, and is useful to standardize protein loads.The RC DC protein assay involves more steps, but is lessvariable for different proteins because the reaction is with thepeptide bond, not the amino acid side chains.

Modified Bio-Rad Protein Assay

This modification to the standard Bio-Rad protein assay(Figure 9.1) is recommended to determine the protein contentin typical sample solutions used to load IPG strips.

Materials:

• Bio-Rad protein assay kit I: contains 450 ml dye reagent concentrate and a bovine γ-globulin standard

• Deionized water

• Whatman #1 filter paper or equivalent (easiest to use if purchased fluted for funnel filtration)

• Concentrated HCl to make a 0.12 N stock

• 20 µl of each sample (40 µl for duplicates)

• IPG sample buffer, protein-free

Procedure:

1. Prepare the bovine γ-globulin standard at 14 mg/ml by

reconstituting the lyophilized protein in 1 ml of water. This is 10x

the concentration that is recommended in the kit instructions.

2. Prepare a 1:4 dilution of the dye reagent concentrate by mixing

1 part of dye with 3 parts of water, and filter the dye through

Whatman #1 filter paper. (Each assay point requires 3.5 ml of

diluted dye reagent.)

3. Prepare 0.12 N HCl (nominal) by diluting concentrated HCl 1:100.

4. Prepare a standard curve covering the range of 0.1–14 µg

protein/µl by diluting the 14 mg/ml standard in the same buffer

as the sample.

5. Mix 20 µl of each standard or sample with 80 µl of 0.12 M HCl

in separate assay tubes. It is a good idea to make duplicates for

each sample or standard.

6. Add 3.5 ml of diluted dye reagent to each tube. Vortex gently.

7. After 5 min, measure the absorbance of each sample or standard

at 595 nm.

8. Plot the absorbance values versus the amount of protein (in µg)

for the standard curve. Expect a nonlinear relationship.

9. Compare the absorbance readings for the samples to the

standard curve.

Fig. 9.1. The Bio-Rad protein assay can be performed in tubes, microtubes,or microtiter plates.

30 For ordering information related to this section see page 45

RC DC Protein Assay

The 2-D sample buffer components listed in Table 9.8 arecompatible with the RC DC protein assay. The presence ofone or more of these substances may change the response ofthe protein to the assay reagents. Thus, the protein standardshould always be prepared in the same buffer as the proteinsample (see Figure 9.2).

Materials:

• Bio-Rad RC DC protein assay kit 1, contains RC reagentspackage (RC reagent I, 250 ml; RC reagent II, 250 ml), DC™ protein assay components (DC reagent A, 250 ml; DCreagent B, 2 L; DC reagent S, 5 ml), and bovine γ-globulinstandard sufficient for 500 assays

• Deionized water

• 25 µl of each sample (50 µl for duplicates)

• IPG sample buffer

Microcentrifuge Tube Assay Protocol (1.5 ml)

1. Add 5 µl of DC reagent S to each 250 µl of DC reagent A that

is needed for the assay. This solution is referred to as reagent A´.

Each standard or sample assayed will require 127 µl of reagent A´.

(Reagent A´ is stable for 1 week even though a precipitate

may form after 1 day. If a precipitate forms, warm the solution and

vortex it. Do not pipet the undissolved precipitate as it will likely

plug the tip of the pipet and change the volume of reagent A´

added to the sample.

2. Prepare 3–5 dilutions of a protein standard ranging from

0.2–1.5 mg/ml protein. A standard curve should be prepared

each time the assay is performed.

(For best results, the standards should always be prepared in

the same buffer as the sample.)

3. Pipet 25 µl of each standard and sample into a clean, dry

microcentrifuge tube.

4. Add 125 µl RC reagent I into each tube and vortex. Incubate the

tubes for 1 min at room temperature.

5. Add 125 µl RC reagent II into each tube and vortex. Centrifuge

the tubes at 15,000 x g for 3–5 min.

6. Discard the supernatant by aspiration, then invert the tubes on

clean absorbent tissue paper. Allow the liquid to drain completely

from the tubes.

7. Add 127 µl of reagent A´ to each microcentrifuge tube and vortex.

Incubate the tubes at room temperature for 5 min or until the

precipitate is completely dissolved. Vortex before proceeding to

the next step.

8. Add 1 ml of DC reagent B to each tube and vortex immediately.

Incubate at room temperature for 15 min.

9. After the 15 min incubation, read the absorbance at 750 nm.

The absorbance should be read within 1 hr.

Fig. 9.2. Effect of reducing agent on standard curve. Gray line, carbonic anhydrase standards without reducing agent; black line, with 350 mM DTT.

0.5

0.4

0.3

0.2

A75

0

0.2 0.4 0.6 0.8 1.0 1.2 1.4

Amount of standard, µg

DTT

Compatible 2-D Sample Buffer Components After 1 Wash After 2 Washes (Optional)

Dithiothreitol (DTT) 100 mM 350 mM

Tributylphosphine (TBP) 2 mM -

β-Mercaptoethanol 5% 10%

ReadyPrep sequential extraction reagent 2* Not compatible Full strength†

ReadyPrep sequential extraction reagent 3** Not compatible Full strength†

Laemmli buffer (with 5% β-mercaptoethanol) Full strength -

CHAPS 2% -

Tween 20 2% -

Triton X-100 2% -

* Contains 8 M urea, 4% (w/v) CHAPS, 40 mM Tris, 0.2% (w/v) Bio-Lyte 3/10 ampholyte, 2 mM TBP ** Contains 5 M urea, 2 M thiourea, 2% (w/v) CHAPS, 2% (w/v) SB 3-10, 40 mM Tris, 0.2% (w/v) Bio-Lyte 3/10 ampholyte, 2 mM TBP

† The presence of these substances changes the response of protein to the assay reagents. Protein standards should be prepared in the same buffer as the protein samples.

Table 9.8. Reagent compatibility with the RC DC protein assay.

31For ordering information related to this section see page 45

Master 2-D techniques before proceeding to separate yourown samples with the ReadyPrep™ 2-D starter kit. Premixedreagents, a standardized sample, and a detailed optimizedprotocol allow you to get familiar with 2-D techniques and tovalidate your 2-D system. See page 45 for ordering information.

Protein Load for 2-D GelsTable 10.1 shows generally recommended protein loads for2-D gels. Because of sample-to-sample variation, theamounts are a guide only. For narrower pH range IPG strips,more protein can be loaded, because proteins outside therange of pI resolution will not remain on the strip to enter the2-D gel. For single-pH-unit IPG strips, the amount that canbe loaded can be as much as 4–5 times more, which allowsbetter detection of low-abundance proteins. For further discussion of factors related to protein load, see page 7.

Table 10.1. Approximate protein loads for IPG strips.

IPG Strip RehydrationFor a discussion of IPG rehydration and sample application,see pages 10–11.

Solutions used to rehydrate IPG strips prior to loading a sample are the same as those used to solubilize or dilute samples for in-gel rehydration (see page 27). Methods forrehydration of strips in buffers (with or without sample) aredescribed in the following sections.

Passive Rehydration with Sample

Passive sample application during rehydration is performed by placing the IPG strip gel side down in the channel of afocusing or rehydration tray that contains the sample in anappropriate rehydration solution. Use the sample volumesgiven in Table 10.2. This procedure will result in rehydration ofthe strips to their original thickness of 0.5 mm. Larger orsmaller volumes can be used and the strips will swell to accommodate more liquid up to a point (Görg et al. 2000). A minimum of 11 hr total rehydration time is recommended. It is important that the strips be left in the well for the entiretime, even if it appears that all of the liquid has been absorbed.High MW proteins cannot enter the gel until the pores arelarge enough to accept them, which only occurs when thepores have swelled to their maximum size.

Table 10.2. Approximate volumes to hydrate ReadyStrip™

IPG strips.

IPG Strip Length Analytical Load Preparative Load(Silver or SYPRO Ruby staining) (Coomassie staining)

7 cm 10–100 µg protein 200–500 µg protein

11 cm 50–200 µg protein 250–1,000 µg protein

17 cm 100–300 µg protein 1–3 mg protein

ReadyStrip IPG VolumeStrip Length

7 cm 125 µl

11 cm 185 µl

17 cm 300 µl

If too much solution remains outside the gel in the focusing tray during electrophoresis, a parallel current path along thesurface of the strip can form in which the proteins will not befocused. This can result in protein loss and streaking. To minimize the possibility of a parallel current path, rehydratethe strips in a disposable rehydration tray, then transfer themto the focusing tray. During transfer, carefully blot excess liquidfrom the strip with moist filter paper prior to beginning the run.

Remove the IPG strip from the protective cover using glovedhands and forceps. Carefully place the IPG strip in the rehydration buffer, gel side down, making sure the entire stripis wetted. There is no “best way” to place dry IPG strips incontact with solution in the trays. Any of the methods illustratedin Figure 10.1 are suitable.

It is helpful to add a trace of Bromophenol Blue to the samplesolution to observe the hydration process. Allow the liquid todistribute for about 1 hr before covering the strips with mineraloil. The IPG strips must be covered to prevent evaporation,which will cause the urea to precipitate as it becomes moreconcentrated. As a precaution against evaporation, mineral oil should be gently layered on top of each channel until itcompletely covers each strip.

Fig. 10.1 A and B. Strip rehydration method 1.Prop up one of the long edges of the tray at an angle to the lab bench. Pipetthe rehydration solution along the entire length of the lower corner of eachchannel (A); place the strip, edge first, into the liquid (B). Then place the tray flaton the benchtop.

Fig. 10.1 C and D. Strip rehydration method 2.Pipet the rehydration solution into the middle of each tray channel (C); bend thestrip into a “U” shape and lower it into the liquid from the center out to theedges (D).

A B

C D

32 For background information related to this section see pages 8–12

Chapter 10 — First-Dimension Separation Methods

Fig. 10.2. Placement of wicks on the electrodes in each channel that will beused. IPG strips will be placed on top of the wicks.

Alternatively, if strips are rehydrated in the focusing trays(either actively or passively), the ends of each strip can be liftedwith forceps and wet wicks inserted between the strip and the electrodes (Figure 10.3). Wicks should be wetted but notsoaked. Blot wetted wicks before placing them in the tray.

Cover the strip with mineral oil before starting the focusingrun to prevent evaporation and carbon dioxide absorptionduring focusing. Channels should be filled nearly to the topbut should not be overflowing. The PROTEAN IEF focusingtray has rounded corners at both ends of the individual channels that prevent mineral oil movement into the adjacentchannels. The rounded corners also reduce salt buildup dueto inadequate cleaning between IEF runs. It is important to clean the focusing trays properly between runs. Channel-to-channel leakage is common when salts accumulate in the channels.

Fig. 10.1 E. Strip rehydration method 3.Pipet the rehydration solution into one end of each tray channel. Butt the stripup to the same end of the channel and lower it into the liquid toward the opposite end (E).

Active Rehydration

For active rehydration of IPG strips with sample in a focusingtray, run the IEF cell under low voltage (50 V). Ensure that theliquid extends past the electrode wires at each end so that theentire strip rehydrates and no dry area creates a discontinuityin the current path. It might be necessary to lift the ends of theIPG strip slightly to get the liquid to flow to the ends of thestrip. After the sample has been in contact with the strips for1 hr, add mineral oil to cover each strip. The PROTEAN® IEFcell can be programmed for active rehydration and to transitionautomatically into a focusing run. Alternatively, a pause may beincorporated to allow the operator to insert a wick under eachend of the strip (see the section below on performing IEF).

If this method of sample application causes a disproportionateratio of large proteins to small proteins, try passive rehydration.

Performing IEF The PROTEAN IEF cell with integrated power supply andPeltier cooling is recommended for IEF protocols in this manual. It can simultaneously run up to twelve 11 or 17 cmIPG strips or up to twenty-four 7 cm strips. Running conditionscan be better controlled by running the same type of sample,buffer, and IPG strip pH range together. See Chapter 3 for ageneral discussion of IEF and page 11 for a discussion of running conditions that affect results.

Positioning Strips and Use of WicksAfter the strips have rehydrated, move them to the IEF focusing tray if they were rehydrated in other trays. Carefullyblot excess liquid from the strip with moist filter paper.

Wicks are highly recommended because they collect saltsand other contaminants in the sample. Without wicks, saltscollect at the anode and cathode, producing high conductivitythat can alter the gradient, cause discontinuities in the gel,and cause “hot spots” or burns. Place a dry wick on eachelectrode that is used (Figure 10.2). Position the wicks withinthe indentations of the channels. Pipet 5–8 µl of water on eachwick before positioning the IPG strips.

E

Fig. 10.3. Insertion of wicks under both ends of an IPG strip that has been rehydrated in a focusing tray.

33For ordering information related to this section see pages 45–46

• Linear ramping mode

In linear ramping mode, the voltage increases linearly withinthe programmed time frame, starting with the final voltage ofthe previous step and ending with the maximum voltage programmed. The resistance of the sample/rehydration buffersystem will determine whether the maximum set voltage canbe reached in the programmed time. This mode is used forsamples of intermediate resistance.

• Slow ramping mode

In this mode, the voltage is increased quadratically:

V= B + (N2 x (E - B)/T2)

where B = starting voltage, E = ending voltage, N = elapsedtime, and T = total time. The run will continue below or at thecurrent limit. This mode is used for high-resistance sample/rehydration buffer systems to minimize high power input initially while achieving high voltage as quickly as possible.

NOTE: The default current limit in the PROTEAN IEF cell is 50 µA per strip. A higher current limit, up to 99 µA per strip,can be programmed into a method. All preset methods havea fixed current limit of 50 µA per strip. In the rapid rampingmode, the system runs at the set current limit and adjusts thevoltage until the maximum voltage is reached. In the linear orslow ramping modes, the system follows a specific algorithmand does not always run at the current limit. The factor thatdetermines the time needed to reach maximum voltage is thecomposition of the sample solution. Systems with high saltconcentration and high sample loads require a long time toreach steady state. It is not always possible to reach themaximum set voltage within the programmed time. Highampholyte concentrations and high protein load also limit thefinal attainable voltage.

Storage of IPG Strips after IEFBecause the pH gradient is fixed in the IPG strip gel, focusedproteins are more stable at their pI than in conventional IEFgels. Focused IPG strips can be stored at -20˚C indefinitelywithout affecting the final 2-D pattern. IPG strips are bound toa plastic sheet, so gel cracking, which results from expansionand contraction during freezing and thawing, is avoided andthe IPG strips retain their original dimensions after thawing. It is convenient to store IPG strips in rehydration trays orscrewcap plastic tubes, which can then be used to equilibrate the strips for the second dimension (see page 36).

Focusing Conditions for IPG Strips on the PROTEAN IEF Cell

See pages 11–12 for further discussion of running conditionsfor IPG strip focusing. Table 10.3 gives suggested total volt-hours for IPG strip runs. These conditions are intended as aguide; individual samples may require more or less time.

Table 10.3. Suggested focusing conditions.

Voltage Ramping Modes

Voltage ramping can replace traditional stepwise voltage programming with continuous voltage changes. The PROTEAN IEF cell (Figure 10.4) includes 3 voltage ramping modes: rapid, linear, and slow. Each ramping mode is appropriate for the resistance of particular samples. The combined resistance of the IPG strips, the rehydrationbuffer, and the sample determines which ramping modeshould be used. During the focusing process, charged contaminants move to the electrodes and proteins move tothe pH equal to their pI. While the proteins are being focused,the resistance of the IPG strip gradually increases until itreaches a maximum.

Each voltage ramping mode controls the rate of voltagechange as follows:

• Rapid ramping mode

In rapid ramping mode, salts and other ionic contaminantsare driven from the IPG strips as rapidly as possible. The limiting factor in reaching the maximum set voltage is the current limit per strip. The maximum voltage can be reachedin ≤2 hr for high-resistance (low ionic strength) samples, or in >6 hr for low-resistance samples. In both cases the powersupply will run at the set current limit until a steady state isreached. This is the mode of choice for many samples, and isparticularly useful to minimize low-resistance sample run time.

ReadyStrip IPG Maximum Typical Volt-HoursStrip Length Voltage

7 cm 4,000 V 7,000–10,000 V-hr

11 cm 8,000 V 20,000–40,000 V-hr

17 cm 10,000 V 30,000–60,000 V-hr

Fig. 10.4. The PROTEAN IEF cell and accessories.


Fig. 11.1. Apparatus for casting multiple gels. Multi-casting chambers for 12 PROTEAN Plus™ 3 gels or for 12 Mini-PROTEAN® gels allow uniform castingof gradient gels. Gradient makers are available for both size formats.

Using Precast GelsFor discussion of precast gels and for a full list of the 3 sizesof Bio-Rad precast gels with IPG wells, see page 14.

To embed IPG strips onto precast gels, remove the gels fromtheir protective wrapping and remove any tape that seals thebottom of the gels. Follow the instructions for IPG equilibrationand agarose embedding (page 36).

Casting SDS-PAGE Gels Using Multi-Casting ChambersIn general, proteomics work requires running several IPGstrips and second-dimension gels per experiment. It is important that gels have a very similar composition. The bestway to ensure that handcast gels have the same compositionis to cast them at the same time in a multi-casting chamber.This is especially important when casting gradient gels. Details of the assembly and use of multi-casting chambersare available in their accompanying instruction manuals. Tips that generally apply to all multi-casting systems are:

• Before assembling the casting chamber, glass plates shouldbe carefully cleaned with Bio-Rad cleaning concentrate andthoroughly rinsed with deionized water

• Each pair of glass plates (2 per gel) should be separated fromthe next by a spacer sheet; the spacer sheet allows easierseparation of the cassettes after gel polymerization

• The volume of gel solution should be determined by measuringthe volume of water needed to fill the assembled glass platesto the desired level in the multi-casting stand

• Allow overnight polymerization to compensate for the lowconcentrations of catalysts (recommended to ensure thatpolymerization does not start while the gradient gels arebeing cast)

Running Multiple GelsBio-Rad's Dodeca™ cells allow electrophoresis of 12 gelssimultaneously under identical conditions, providing highthroughput and reproducible results. Three size formats areavailable (see pages 14–15, Figure 4.2, and Table 4.2). Referto the instruction manuals for specific directions on assemblyand use.

Chapter 11 — Second-Dimension Separation Methods


IPG Equilibration for the Second DimensionTo solubilize focused proteins and to allow SDS binding inpreparation for the second dimension, it is necessary to equilibrate focused IPG strips in SDS-containing buffers. This step is analogous to boiling samples in SDS buffer priorto 1-D SDS-PAGE. See page 14 for further discussion of equilibration. Rehydration/equilibration trays sized for eachsize strip can be used for equilibration.

Equilibration protocol:

Place one strip, gel side up, in each channel and fill the channels

successively with the equilibration buffers derived from the base

buffer (Table 11.1). First incubate with gentle agitation in DTT

equilibration buffer 1 for 10 min, then decant. Refill the channels

with iodoacetamide equilibration buffer 2 and incubate again for

10 min. This method requires 2.5 ml of each solution per strip for

7 cm strips, 4 ml for 11 cm strips, and 6 ml for 17 cm strips. After

equilibration, remove the IPG strip and embed it onto the prepared

second-dimension gel as described in the following section.

* This buffer may be frozen in aliquots. Lyophilized equilibration base buffer canbe ordered as ReadyPrep™ equilibration buffer II, which reconstitutes to 20 ml.

DTT Equilibration Buffer 1

This buffer reduces sulfhydryl groups. To prepare it, add DTT to

equilibration base buffer (Table 11.1) to 2% (200 mg/10 ml)

immediately prior to use.

Iodoacetamide Equilibration Buffer 2

This buffer alkylates sulfhydryl groups. While the strip is incubating in

equilibration buffer 1, add dry iodoacetamide to equilibration base

buffer (Table 11.1) to 2.5% (250 mg/10 ml).

Reagents Amount (Final Concentration)

Urea 36 g urea (6 M)

20% SDS 10 ml (2%)

1.5 M Tris/HCl, pH 8.8 gel buffer 3.3 ml (0.05 M)

50% Glycerol 40 ml (20%)


Table 11.1. Equilibration base buffer.*

Placement and Agarose Embedding of IPG StripsPosition the second-dimension gel cassette so that it is leaning slightly backwards (approximately 30° from vertical).Place the IPG strip onto the long plate with the plastic backingagainst the plate. Slide the strip, face down, between theplates using a spatula to push against the plastic backing. Becareful not to damage the gel with the spatula. Make sure theIPG strip is positioned directly on top of the second-dimensiongel. To secure the strip in place, overlay it with 0.5–1.0%molten agarose prepared in SDS-PAGE running buffer (asmall amount of Bromophenol Blue can be added to theagarose overlay in order to track the ion front during the run).Use warm molten agarose; hot agarose may acceleratedecomposition of the urea in the equilibration buffer. Bubblesmay form under or behind the strip when adding the agaroseoverlay. These bubbles may disturb protein migration andmust be removed. Immediately after overlaying, use the spatula to dislodge bubbles by tapping the plastic backing ontop of the strip. Stand the gel upright and allow the agaroseto set prior to loading the gel into the electrophoresis cell (see Figure 11.2).

Fig. 11.2 C and D. In C, an 11 cm ReadyStrip IPG strip is applied to the topof a Criterion™ precast gel held at an angle. The strip is aligned so the plasticbacking is against the back plate and the IPG strip is touching the top of the gel.Molten agarose may be added before the strip is placed in the well and thestrip positioned within the liquid agarose, or the agarose may be added afterthe strip is in position (D). The gel is moved to an upright position while theagarose is setting.

C D

Fig. 11.2 A and B. In A, a 17 cm ReadyStrip™ IPG strip (held in forceps) isapplied to the top of a PROTEAN® XL Ready Gel® precast gel. The gel is held at an angle with the short plate forward. In B, the strip is pushed into direct contact with the top of the gel using a spatula.

A B


Running the Second DimensionSecond-dimension gels can be run in any cell appropriate tothe size of the gel as shown in Table 4.2 (page 15). Bio-RadDodeca cells, which accommodate 12 gels per run, are wellsuited to high-throughput proteomics experiments. Figure11.3 shows the use of precast gels in Dodeca cells.

When using PROTEAN Ready Gel precast gels with the PROTEAN Plus Dodeca cell, apply a thick bead of Vaselinepetroleum jelly or other sealant between the glass platesalong the downward facing edge of each cassette. Thesealant prevents current leak and associated inward skewingof protein migration. PROTEAN Plus sandwiches are sealedalong the downward facing edge and do not require sealant.

Insert the gels into the appropriate electrophoresis cell andrun them according to the instruction manual provided withthe cell.

Applying MW StandardsSDS-PAGE standards can be applied to gels that have noreference lane using this protocol:

1. Trim a PROTEAN IEF cell wick from 4 x 20 mm to 4 x 5 mm.

2. Pipet 10 µl of the SDS-PAGE standards onto the wick. Unstained standards can serve as a control for the staining procedure.

3. Slip the wick into the slot in the gel sandwich next to or overlapping an end of the IPG strip.

4. Seal the wick and the IPG strip with molten agarose.

Fig. 11.3 A and B. In A, the seal is being removed from the bottom of a ReadyGel precast gel. Criterion precast gels and PROTEAN XL precast gels also mustbe unsealed. This can be done before the IPG strip is applied or after the agaroseis set. In B, Ready Gel precast gels are being placed in the Mini-PROTEAN 3Dodeca cell, which can run up to 12 gels at the same time.

Fig. 11.3 C. A gel is being inserted into the PROTEAN Plus Dodeca cell. The top of the gel is aligned with the agarose-embedded IPG strip at the cathode (negative electrode) side of the tank. This cell accommodates PROTEAN XL Ready Gel precast gels (shown), or PROTEAN Plus handcast gels up to 25 cm wide. A uniform electrical field is applied via metal plates. The pump visible behind the tank circulates the buffer to maintain a uniformtemperature during the run.

A B

C


Methods for dye staining, fluorescent staining, and silverstaining of gels are provided in this chapter. The different protein detection methods are discussed on pages 16–17.Stain gels at room temperature with gentle agitation (e.g., onan orbital shaker). Use any convenient container that is appropriate to the method chosen. Silver staining should bedone in glass, while SYPRO Ruby protein gel staining cannotbe done in glass. Casserole dishes, photography trays, orany flat dishes large enough to allow the gels to lie flat areappropriate. Wear gloves and use good laboratory safety precautions when using staining chemicals.

Coomassie Brilliant Blue R-250 StainUse the following protocol for Coomassie Blue R-250 staining:

1. Stain the gel with gentle agitation in an ample volume of Coomassie

Brilliant Blue R-250 staining solution (0.1% Coomassie Blue R-250

(w/v) in 40% methanol (v/v), 10% acetic acid (v/v)) for 20 min to 1 hr.

This solution should not be reused.

2. Wash the gel in an ample volume of Coomassie Brilliant Blue R-250

destaining solution (40% methanol, 10% acetic acid). Lab tissues

(e.g., Kimwipes) can be dropped in the destaining solution to help

capture the dye and aid in destaining. Agitate gently, and change the

solution until background staining has been removed.

Bio-Safe™ Colloidal Coomassie Blue G-250 StainBio-Safe colloidal Coomassie stain is a preformulated stainingsolution with sensitivity between that of Coomassie Blue R-250 and silver stains.

Use the following protocol for Bio-Safe Coomassie staining:

1. Rinse gels twice for 10 min in deionized water with agitation to

remove SDS.

2. Add sufficient Bio-Safe Coomassie stain to cover the gels. Incubate

with agitation for 1 hr to overnight. Color will continue to develop

after 1 hr.

3. Rinse the gel with deionized water with agitation until desired

contrast is achieved.

Fig. 12.1. Bio-Safe Coomassie stain is available in 1 L bottles or 5 L cubes.

Chapter 12 — Methods for Protein Detection in GelsSYPRO Ruby Protein Gel StainSYPRO Ruby protein gel staining is a simple procedure.Polymethylpentene dishes are ideal containers for stainingbecause the high-density plastic adsorbs a minimal amountof the dye. Glass dishes are not recommended.

Follow this protocol to stain and visualize SYPRO Ruby-stained gels:

1. First fix gels for 30 min to 1 hr in a mixture of 10% methanol and

7% acetic acid.

2. Add SYPRO Ruby protein stain to the gel. Do not dilute the stain.

(The minimum staining volumes for typical gel sizes are: 50 ml for

8 x 10 cm gels, 150 ml for 13.5 x 10 cm gels, 330 ml for 16 x 20 cm

gels, and 500 ml for 20 x 20 cm gels; using too little stain will lower

the sensitivity.)

SYPRO Ruby protein gel stain is an end-point stain. Some staining

can be seen in as little as 30 min. However, a minimum of 3 hr of

staining is required for maximum sensitivity. For convenience, gels

may be left in the dye solution overnight or longer without overstaining.

3. Prior to imaging the gel, to further decrease background fluorescence

and to reduce speckling, the stained gel should be washed for at

least 30 min in a mixture of 10% methanol and 7% acetic acid. The

gel may be monitored by UV epi-illumination to determine the level

of background fluorescence. Gels do not over-destain, although the

fluorescent intensity of gels left in destaining solution or water for

weeks will be reduced. Store stained gels in water.

4. Gels can be viewed on a “blue light” or UV box, or scanned on a

Bio-Rad Molecular Imager FX Pro Plus™ or VersaDoc™

imaging system (see Chapter 7).

Fig. 12.2. A 2-D gel stained with SYPRO Ruby protein gel stain.


Gel DryingThe GelAir™ drying system is perfect for drying polyacrylamidegels up to 20 x 20 cm (Figure 12.4). After drying, the gels arecompletely clear and enclosed in cellophane. The dried gelsare ideal for densitometry, photodocumentation, autoradio-graphy, or long-term storage.

The GelAir dryer is a heated drying chamber that dries minigels in 45 min or 20 x 20 cm gels in 60 min. The dryer holdsup to 4 drying frames at once.

Fig. 12.4. Drying gels with the GelAir drying system. Step 1, place aframe on the assembly table and cover it with wet cellophane; 2, lay a gel ontop of the cellophane; 3, place another sheet of wet cellophane over the gel; 4, clamp the drying frames together; 5, slide the drying frame into the dryer.

Gels can also be dried with a vacuum gel dryer. Bio-Rad’sModel 583 gel dryer can dry multiple gels on its 35 x 45 cmsurface. The system includes a heated lid and a gasket toensure uniform application of vacuum. Gels are dried on asheet of filter paper or a cellophane membrane support. The quiet water-cooled HydroTech™ pump provides a self-contained, constant vacuum.

Storage of Gels in Plastic BagsAs an alternative to drying gels, they can also be sealed inzip-top plastic bags. Gels are usually sealed in either water or,for long-term storage, water with 0.005% sodium azide. Fill the bag with water, then insert the gel, expel the water,and seal the bag.

Bio-Rad Silver Stain (Merril)Silver staining is a highly sensitive method for detection ofproteins and nucleic acids in polyacrylamide gels. The Bio-Radsilver stain (Figure 12.3), based on the method of Merril et al.(1981), is sensitive to 0.1 ng protein per mm2 of gel, which is10- to 50-fold more sensitive than Coomassie Brilliant Blue R-250 for proteins. This method is not recommended if massspectrometry will be performed on spots excised from thegels. Use Silver Stain Plus™ stain, described in the followingsection, for mass spectrometry applications.

Follow the detailed instructions included in the kit. For largegels, the procedure involves 13 steps and may take up to 13 hr to perform.

Silver Stain Plus StainProteins can be visualized in 1 hr with very little hands-ontime by using a carrier-complex silver-staining chemistry similar to that developed by Gottlieb and Chavko (1987) fordetection of DNA in agarose gels. Silver Stain Plus stain is 30- to 50-fold more sensitive than Coomassie Blue R-250dye and will detect nanogram amounts of protein. For optimalstaining results, follow the instruction manual for the kit.

Fig. 12.3. Bio-Rad Silver Stain Plus kit.

1 2

3 4 5

39For ordering information related to this section see page 49

Part III

TroubleshootingGuide


41

Troubleshooting GuideProblems with 2-D Gel Results Page

No spots or fewer than expected spots detected on the 2-D gel 421. Not enough sample loaded 422. Insufficient sample entered the IPG strip 423. pH gradient oriented incorrectly during focusing 424. Detection method not sensitive enough 425. Failure of detection reagents 42

Streaking and smearing on 2-D gel 42• Horizontal streaks 421. Sample preparation problems 422. Too much protein loaded 423. Nucleic acids bound to protein 424. Focusing time not optimized 42

• Vertical streaks 431. Pinpoint vertical streaks in the background 432. Vertical streaks, usually broad, connected to a spot 433. Twin vertical spots or vertical doublets throughout the gel 43

Blank stripes in the vertical dimension 43

Known proteins showing up as multiple spots or at the wrong position 43

Spots absent from one side of the second-dimension gel 43

No spots or fewer than expected spots in the high molecular weight regions 43

Problems during runs 441. Focusing time too long or final voltage not reached 442. Arcing or burned IPG strips 443. Mineral oil migrating between wells in PROTEAN® IEF focusing trays 444. Interruption due to power failure 445. Second-dimension run too long 446. Second-dimension run too fast 44

42

1. Not enough sample loaded. Check the concentration of starting material by protein assay procedures (see page 30). Check that your protein

assay is functioning properly and not responding to interfering substances in your sample. Consider what percentage of the sample loaded might

be represented in each spot. For example, if you want to detect 1,000 spots, and your detection method can detect 10 ng per spot as a lower limit

(SYPRO Ruby, Bio-Safe™ Coomassie, or Silver Stain Plus™ stains), you should load at least 100,000 ng (100 µg) of protein per gel if you expect

most proteins to be 10x the abundance of your lower limit. If your detection method is less sensitive, or your protein of interest is of lower abundance,

you must load more sample. See pages 7 and 32 for further information on protein load. A guideline for new samples is to adjust the total protein

concentration in the rehydration solution to 1 mg/ml.

2. Insufficient sample entered the IPG strip. Insoluble proteins will not enter the IPG strip pores. Proteins can be solubilized by adjusting

various components in the sample extraction solution, for example, detergent, reducing agent, and ampholyte concentration, as well as pH and

ionic strength. See the sample preparation discussion (Chapter 2) for more thorough discussion of factors affecting solubility of proteins applied

to IPG strips. Sample application methodology can also affect solubilization; a discussion on sample application is found on page 10. See page 28

for enhanced solubilization solutions.

3. The pH gradient was oriented incorrectly during focusing. Check to make sure that the end of the IPG strip marked with a “+” is oriented

toward the positive electrode.

4. Detection method not sensitive enough. Make sure that the detection method is sensitive enough to detect the amount of protein loaded.

A titration may be helpful: Load strips with increasing concentrations of sample, focus, and stain the strips with Bio-Safe Coomassie Blue or IEF

gel staining solutions. Select the concentration of subsequent loads from the load that has the most detectable proteins without distortion or

severe overloading.

5. Failure of detection reagents. Loading a lane of unstained protein standards as described on page 37 can help to diagnose this problem.

If the standards are not detected, you should check the expiration dates and the formulations of all detection reagents.

Horizontal streaksHorizontal streaks indicate a problem with the IEF run. They can be caused by a number of factors, including liquid in excess of the amount

absorbed by the strip, high viscosity, overloading, sample solubility problems, incomplete focusing, or nonproteinaceous material (especially nucleic

acids) adhering to proteins.

1. Sample preparation problems. This is the most common cause of horizontal streaking. In this case, significant amounts of protein may

also remain at the origin of the second-dimension gel. A new sample preparation method must be worked out for each type of sample.

The concentations of urea, detergents, ampholytes, and reducing agents may be critical; see the section on solubilization on page 3. A literature

search on preparation of proteins from the system of interest is a good place to start. If the sample is a membrane protein sample or a cell lysate,

then high-speed centrifugation may help. The sample should be centrifuged at 100,000 x g after addition of IEF sample buffer (ampholytes, urea,

detergent, reducing agent) even if it appears clear.

An excellent reference is Molloy (2000).

2. Too much protein loaded. Try using 1/10th the amount of sample, or dilute the sample by at least one half. If your protein is not abundant, you

may want to prefractionate your sample using the ReadyPrep™ sequential extraction kit, Rotofor® system, chromatography, or other procedure to

enrich the protein of interest and lower the amounts of abundant proteins. See pages 7 and 32 for information on the amount of protein to load.

3. Nucleic acids bound to protein. Treatment with an endonuclease as discussed on pages 6 and 29 can reduce viscosity and improve protein

absorption into the IPG strip. Make sure that the nuclease is active and that digestion is adequate. A very viscous sample implies that nuclease

treatment has failed.

4. Focusing time not optimized. If focusing is not complete, proteins will not focus as tight spots. See pages 11 and 34.

Problem: Streaking and smearing on 2-D gel

Problem: No spots or fewer than expected spots

43

Problem: Streaking and smearing on 2-D gelVertical streaksVertical streaks are related to second-dimension electrophoresis and are caused by various factors, including loss of solubility of a protein

at its pI, dust contaminants, improper reducing agent, or incorrect placement of the IPG strips. These streaks can be analyzed like any other

SDS-PAGE problem. For example, if SDS-PAGE standards have been applied next to the IEF strip on the second-dimension gel, and the

standards do not have streaks, then the problem lies with the sample, rather than with the electrophoresis run.

1. Pinpoint vertical streaks in the background. These are most likely due to dust or particles in the water. The water purifier and storage

containers should be checked for contamination. All gel solutions should be filtered through 0.45 µm nitrocellulose into a dust-free container.

2. Vertical streaks, usually broad, connected to a spot. Some streaks are above the spots, others are below the spots. If spots are streaky,

some protein aggregation is indicated. Isoelectric point precipitation is difficult to predict and control. Some streaking is caused by incomplete

reduction/alkylation. In such cases, an increased amount of reducing agent or changing the reducing agent to TBP will help. In a small number

of cases, increasing the SDS concentration during equilibration or increasing the time of equilibration will decrease streaking. Protein overloading

can also cause vertical streaking.

3. Twin vertical spots or vertical doublets throughout the gel. This effect is sometimes seen when thiourea is included in the sample

solution. Its cause is unclear.

Problem: Blank stripes in the vertical dimension1. An air bubble may have been trapped in the agarose that joins the strip to the top of the second-dimension gel.

2. A region of the IPG strip may not have been sufficiently rehydrated or may have torn during handling, resulting in the absence of focused

protein in that region.

3. Blank stripes near pH 7 are often caused by excessive DTT (above 50 mM) in the IPG sample buffer.

4. Blank stripes at the electrodes, especially at the cathode, can be caused by a buildup of salt.

Problem: Known proteins showing up as multiple spots or at the wrong position

1. Proteins may be carbamylated if prepared too far ahead of time in urea, if exposed to high pH while in urea, or if allowed to exceed 30°C while in

urea solutions.

2. Proteins may have been oxidized if the concentration of DTT is not sufficient. Many researchers prefer TBP (tributylphosphine) for this

reason; see page 5 for further discussion.

3. Proteolysis can be a problem if cells are harvested under slow or warm conditions, especially in physiological buffers (lacking chaotropic agents).

Use protease inhibitors, perform manipulations as quickly as possible, and keep solutions as cold as possible.

Problem: Spots absent on one side of the second-dimension gel1. The IPG gradient range could be incorrect for your sample. For example, if all proteins in your sample have basic pI, and you are using an

acidic gradient strip, then most of the proteins will focus to the basic side of the gel.

2. If you apply the sample after rehydration of the strip, then the movement of the proteins can be restricted if the gel is unevenly rehydrated.

3. Basic IPG strips (pH 7–10 and higher) generally give better results when cup loading is used.

4. Reverse endosmosis limits focusing to pH 11 in high-pH IPG strips.

Problem: No spots or fewer than expected spots in the high molecular weight regions

1. Sample may have been subject to proteolysis prior to focusing. Include appropriate protease inhibitors and keep the sample on ice

or in a cold room.

2. Equilibration steps between the first and second dimension were not long enough. Make sure to incubate strips in sufficient volumes of each

equilibration buffer for at least 10 min with mild agitation.

3. Poor entry of high molecular weight proteins during rehydration. The pore size of the acrylamide in the IPG strip is very small during the early

stages of rehydration. Active sample loading in the focusing tray may help the entry of large proteins. See page 10 for an in-depth discussion

of sample application.

44

1. Focusing time too long or final voltage not reached. One of the most important factors in IEF is to control the total ionic strength of the

sample so that the current is kept low. With IPG strips, a low ionic strength sample is especially important. If the ionic strength is too high, the

final voltage will not be reached. The optimum salt concentration is about 10 mM, although 40 mM can be tolerated. High conductivity can result

from buffer salts in the sample, or from an inherently high conductivity of the sample (e.g., serum). We suggest you either desalt or dilute the sample

so that the total conductivity is as low as possible. Ampholytes can often substitute for salts in the sample solution. We recommend a final

concentration of 0.2% (w/v) ampholytes. Note that other commercially available IPG buffers may contribute conductivity to the sample.

Many cell lysates must be desalted before they can be run on ReadyStrip™ IPG strips. A total ionic strength greater than 10 mM salt in the

sample will increase your focusing time and could prevent the voltage from reaching the maximum setting. Salt will eventually electrophorese out

of the strips. At 50 µA, it takes 4 hr for 40 mM salt to clear from a 17 cm strip. Use of wicks helps. It is also acceptable to increase the run time,

giving as much time as necessary for the run to reach high voltage.

If the sample has a high ionic strength, we recommend you remove the ions with Micro Bio-Spin™ 6 or Micro Bio-Spin 30 spin columns. These

columns contain 10 mM Tris and they will remove all small molecules and ions from a protein sample with excellent protein recovery and no

dilution of the sample.

The simplest approach to focusing highly conductive samples is to be patient. Use electrode wicks and give the system sufficient time to reach

high voltage.

2. Arcing or burned IPG strips during focusing run. Wicks should be used to absorb ionic contaminants that migrate to the ends of IPG strips

during the run. Otherwise, the ions will collect at the electrodes and create regions of low resistance. Regions of low resistance create high current

and high heat, leading to water evaporation, which may result in arcing or burning of the strip. The variable resistance error message will appear if

you are using the PROTEAN® IEF cell.

Whenever possible, we recommend no more than 10 mM total ionic strength in the sample. We suggest desalting with a column such as the

Econo-Pac™ 10DG desalting column or the Micro Bio-Spin column. The 10DG column is useful for samples between 1 and 3 ml. The Micro

Bio-Spin column is useful for samples of ≤75 µl.

3. Mineral oil migrating between wells in the PROTEAN IEF focusing trays. The sample does not migrate along with the oil. It is an

aqueous solution and is not miscible with the oil. The oil will not migrate between wells unless the wells are overfilled or dirty. Note that proper

cleaning of the trays eliminates the wicking problem (mineral oil migrating to next lane). We recommend cleaning in hot water with detergent and

thorough rinsing (pay special attention to the corners). Note that any residual detergent left in the tray can also lead to wicking.

The use of mineral oil is needed during IEF for 2 reasons:

• To prevent the gel from dehydrating during a long focusing run, which would cause urea to precipitate and possibly cause the strip to burn.

• To prevent CO2 in the air from entering the gel and disrupting the pH gradient.

However, oil does not necessarily need to be used in the rehydration trays. When oil is omitted, use the tray lid and seal the tray with plastic

wrap or Parafilm.

4. Interruption due to power failure. If your laboratory is susceptible to power failures, we suggest an uninterruptible power supply for the

PROTEAN IEF cell and other electronic instruments.

5. Second-dimension run too long. Long run times usually indicate a gel buffer that is too concentrated (conductivity is too high).

The distribution of the applied voltage between the leading ions in the gel buffer and the trailing ions from the electrode buffer can slow protein

migration and distort the gel pattern compared to gels run in properly made gel buffers.

6. Second-dimension run too fast. Short run times indicate that the gel buffer is too dilute (conductivity is too low). In this case too, run time and

gel pattern are a function of the distribution of the applied voltage between the gel and electrode buffers.

Problem: Problems during runs

Part IV

Ordering Information


Ordering InformationChapters 2 and 9 — Sample Preparation and Protein AssayCatalog # Description Quantity Page

163-2105 ReadyPrep™ 2-D Starter Kit, includes protein sample and reagents sufficient to rehydrate, focus, and 1 kit 32transfer to the second-dimension gel six 17 cm, ten 11 cm, or sixteen 7 cm ReadyStrip™ IPG strips

163-2103 ReadyPrep Sequential Extraction Kit Reagent 2, reconstitutes to 10 ml 1 vial 28–29163-2106 ReadyPrep 2-D Starter Kit Rehydration/Sample Buffer, reconstitutes to 10 ml 1 vial 27–28163-2110 E. coli Protein Sample, lyophilized 2.7 mg 27161-0730 Urea 250 g 27–28161-0731 Urea 1 kg 27–28161-0611 Dithiothreitol (DTT) 5 g 27163-2101 Tributylphosphine (TBP), 200 mM 0.6 ml 27–29161-0460 CHAPS 1 g 27–29161-0407 Triton X-100 500 ml 3163-1112 Bio-Lyte® 3/10 Ampholyte, 40% (w/v) 10 ml 3, 27–28163-1132 Bio-Lyte 3/5 Ampholyte, 20% (w/v) 10 ml 3, 27–28163-1142 Bio-Lyte 4/6 Ampholyte, 40% (w/v) 10 ml 3, 27–28163-1152 Bio-Lyte 5/7 Ampholyte, 40% (w/v) 10 ml 3, 27–28163-1172 Bio-Lyte 7/9 Ampholyte, 40% (w/v) 10 ml 3, 27–28163-1182 Bio-Lyte 8/10 Ampholyte, 20% (w/v) 10 ml 3, 27–28163-1192 Bio-Lyte 5/8 Ampholyte, 40% (w/v) 10 ml 3, 27–28163-2093 100x ReadyStrip 7–10 Buffer 1 ml 3, 27–28163-2094 100x Bio-Lyte 3/10 Ampholyte 1 ml 3, 27–28163-2095 100x ReadyStrip 6.3–8.3 Buffer 1 ml 3, 27–28163-2096 100x ReadyStrip 5.5–6.7 Buffer 1 ml 3, 27–28163-2097 100x ReadyStrip 4.7–5.9 Buffer 1 ml 3, 27–28163-2098 100x ReadyStrip 3.9–5.1 Buffer 1 ml 3, 27–28161-0404 Bromophenol Blue 10 g 27–28161-0716 Tris 500 g 27–28161-0719 Tris 1 kg 27–28142-6425 AG 501-X8 (D) Mixed Bed Resin 500 g 27163-2104 ReadyPrep Sequential Extraction Kit Reagent 3, reconstitutes to 10 ml 1 vial 28163-2100 ReadyPrep Sequential Extraction Kit, includes 1 vial of reagent 1 (to make 50 ml), 1 kit 28

2 vials of reagent 2 (to make 10 ml each), 2 vials of reagent 3 (to make 10 ml each), 1 vial containing 0.6 ml of 200 mM TBP

500-0001 Bio-Rad Protein Assay Kit I, contains 450 ml dye reagent concentrate and a bovine γ-globulin 1 kit 30standard, based on method of Bradford

500-0121 RC DC™ Protein Assay Kit I, includes RC reagents package, DC™ protein assay reagents package, 1 kit 31bovine γ-globulin standard, 500 standard assays

Sigma E8263 Endonuclease, recombinant, from Serratia marcescens 29Sigma T8656 Thiourea (ACS Reagent Grade) 28–29

Catalog # Description Quantity Page

PROTEAN® IEF System165-4000 PROTEAN IEF System, complete, includes basic unit, 17, 11, and 7 cm focusing trays with lid, 1 system 32–34

1 pack each of 17, 11, and 7 cm rehydration/equilibration trays with lid, 2 pairs of forceps, pack of electrode wicks, mineral oil, cleaning brush

165-4001 PROTEAN IEF Cell, 90–240 VAC, basic unit, includes cell, instructions 1 unit 33–34165-4050 Cup Loading Tray, includes 1 pair moveable electrodes, 1 pack each of large and small replacement cups 1 unit165-4051 Large Replacement Cups 1 pack165-4052 Small Replacement Cups 1 pack165-4071 Electrode Wicks, precut 500 33165-4035 Disposable Rehydration/Equilibration Tray with Lid, 7 cm 25 pack 32165-4015 Disposable Rehydration/Equilibration Tray with Lid, 17 cm 25 pack 32165-4025 Disposable Rehydration/Equilibration Tray with Lid, 11 cm 25 pack 32165-4041 Disposable Rehydration/Equilibration Tray with Lid, 18 cm 25 pack 32165-4043 Disposable Rehydration/Equilibration Tray with Lid, 24 cm 25 pack 32165-4030 7 cm Focusing Tray with Lid 1 33165-4020 11 cm Focusing Tray with Lid 1 33165-4010 17 cm Focusing Tray with Lid 1 33165-4040 18 cm Focusing Tray with Lid 1 33165-4042 24 cm Focusing Tray with Lid 1 33165-4070 Forceps 1 pair 32–33163-2129 Mineral Oil 500 ml 32–33165-4080 Thermal Printer, 100 V, includes cable and power adaptor165-4082 Thermal Printer, 120 V165-4085 Thermal Printer, 220 V

Chapters 3 and10 — Isoelectric Focusing

45

46


ReadyStrip™ IPG Strips163-2000 ReadyStrip IPG Strip, 7 cm, pH 3–10 12 32–33163-2002 ReadyStrip IPG Strip, 7 cm, pH 3–10 (NL) 12 32–33163-2001 ReadyStrip IPG Strip, 7 cm, pH 4–7 12 32–33163-2003 ReadyStrip IPG Strip, 7 cm, pH 3–6 12 32–33163-2004 ReadyStrip IPG Strip, 7 cm, pH 5–8 12 32–33163-2005 ReadyStrip IPG Strip, 7 cm, pH 7–10 12 32–33163-2028 ReadyStrip IPG Strip, 7 cm, pH 3.9–5.1 12 32–33163-2029 ReadyStrip IPG Strip, 7 cm, pH 4.7–5.9 12 32–33163-2030 ReadyStrip IPG Strip, 7 cm, pH 5.5–6.7 12 32–33163-2031 ReadyStrip IPG Strip, 7 cm, pH 6.3–8.3 12 32–33163-2014 ReadyStrip IPG Strip, 11 cm, pH 3–10 12 32–33163-2016 ReadyStrip IPG Strip, 11 cm, pH 3–10 (NL) 12 32–33163-2015 ReadyStrip IPG Strip, 11 cm, pH 4–7 12 32–33163-2017 ReadyStrip IPG Strip, 11 cm, pH 3–6 12 32–33163-2018 ReadyStrip IPG Strip, 11 cm, pH 5–8 12 32–33163-2019 ReadyStrip IPG Strip, 11 cm, pH 7–10 12 32–33163-2024 ReadyStrip IPG Strip, 11 cm, pH 3.9–5.1 12 32–33163-2025 ReadyStrip IPG Strip, 11 cm, pH 4.7–5.9 12 32–33163-2026 ReadyStrip IPG Strip, 11 cm, pH 5.5–6.7 12 32–33163-2027 ReadyStrip IPG Strip, 11 cm, pH 6.3–8.3 12 32–33163-2007 ReadyStrip IPG Strip, 17 cm, pH 3–10 12 32–33163-2009 ReadyStrip IPG Strip, 17 cm, pH 3–10 (NL) 12 32–33163-2008 ReadyStrip IPG Strip, 17 cm, pH 4–7 12 32–33163-2010 ReadyStrip IPG Strip, 17 cm, pH 3–6 12 32–33163-2011 ReadyStrip IPG Strip, 17 cm, pH 5–8 12 32–33163-2012 ReadyStrip IPG Strip, 17 cm, pH 7–10 12 32–33163-2020 ReadyStrip IPG Strip, 17 cm, pH 3.9–5.1 12 32–33163-2021 ReadyStrip IPG Strip, 17 cm, pH 4.7–5.9 12 32–33163-2022 ReadyStrip IPG Strip, 17 cm, pH 5.5–6.7 12 32–33163-2023 ReadyStrip IPG Strip, 17 cm, pH 6.3–8.3 12 32–33

Chapters 3 and10 — Isoelectric Focusing (cont.)


PROTEAN II Ready Gel® Precast Gels, IPG Well, for Use with 17 cm ReadyStrip IPG Strips161-1450 Ready Gel Tris-HCl Gel for PROTEAN II Cell, 10%, IPG well, 1.0 mm thick, 18.3 x 19.3 cm each 14, 35161-1451 Ready Gel Tris-HCl Gel for PROTEAN II Cell, 12%, IPG well, 1.0 mm thick, 18.3 x 19.3 cm each 14, 35161-1452 Ready Gel Tris-HCl Gel for PROTEAN II Cell, 10–20%, IPG well, 1.0 mm thick, 18.3 x 19.3 cm each 14, 35161-1453 Ready Gel Tris-HCl Gel for PROTEAN II Cell, 8–16%, IPG well, 1.0 mm thick, 18.3 x 19.3 cm each 14, 35

Criterion™ Precast Gels, IPG Well, for Use with 11 cm ReadyStrip IPG Strips345-0013 Criterion Tris-HCl Gel, 10% resolving, 4% stacking gel each 14, 35345-0018 Criterion Tris-HCl Gel, 12.5% resolving, 4% stacking gel each 14, 35345-0031 Criterion Tris-HCl Gel, 4–15% each 14, 35345-0036 Criterion Tris-HCl Gel, 4–20% each 14, 35345-0041 Criterion Tris-HCl Gel, 8–16% resolving, 4% stacking gel each 14, 35345-0046 Criterion Tris-HCl Gel, 10–20% resolving, 4% stacking gel each 14, 35

Ready Gel Precast Gels, IPG Well, for Use with 7 cm ReadyStrip IPG Strips161-1390 Ready Gel Tris-HCl Gel, 10% resolving, 4% stacking gel each 14, 35161-1391 Ready Gel Tris-HCl Gel, 12% resolving, 4% stacking gel each 14, 35161-1392 Ready Gel Tris-HCl Gel, 4–15% each 14, 35161-1393 Ready Gel Tris-HCl Gel, 4–20% each 14, 35161-1394 Ready Gel Tris-HCl Gel, 8–16% resolving, 4% stacking gel each 14, 35161-1395 Ready Gel Tris-HCl Gel, 10–20% resolving, 4% stacking gel each 14, 35

Chapters 4 and11— Second Dimension — Precast Gels


161-0722 Cleaning Concentrate 1 kg 7, 35, 44161-0798 Resolving Gel Buffer, 1.5 M Tris-HCl, pH 8.8 1 L 35161-0799 Stacking Gel Buffer, 0.5 M Tris-HCl, pH 6.8 1 L 35161-0158 30% Acrylamide/Bis Solution, 37.5:1 500 ml 35161-0148 40% Acrylamide/Bis Solution, 37.5:1 500 ml 35161-0149 40% Acrylamide/Bis Solution, 37.5:1 2 x 500 ml 35161-0122 Acrylamide/Bis Powder, 37.5:1 30 g 14, 35161-0125 Acrylamide/Bis Powder, 37.5:1 150 g 14, 35161-0141 40% Acrylamide Solution 2 x 500 ml 14, 35161-0142 2% Bis Solution 500 ml 14, 35161-0202 Piperazine Diacrylamide (PDA; alternative crosslinker to bis) 10 g 14, 35161-0800 TEMED 5 ml 35161-0801 TEMED 50 ml 35161-0700 Ammonium Persulfate (APS) 10 g 35

Chapters 4 and11— Casting Multiple Gels — Reagents

47


Casting Multiple PROTEAN Plus™ Gels165-4160 PROTEAN Plus Multi-Casting Chamber, includes casting chamber, sealing plate, silicone 1 35

gasket, tapered luer connector, leveling bubble, acrylic blocks, separation sheets (PROTEAN Plus plates and combs must be ordered separately)

165-4121 Model 495 Gradient Former, includes body with valve stem and tubing connection kit 1 35165-4170 PROTEAN Plus Hinged Spacer Plate, 20 x 20.5 (W x L) cm, 1.0 mm 1 35165-4171 PROTEAN Plus Hinged Spacer Plate, 20 x 20.5 (W x L) cm, 1.5 mm 1 35165-4172 PROTEAN Plus Hinged Spacer Plate, 20 x 20.5 (W x L) cm, 2.0 mm 1 35165-4173 PROTEAN Plus Hinged Spacer Plate, 25 x 20.5 (W x L) cm, 1.0 mm 1 35165-4174 PROTEAN Plus Hinged Spacer Plate, 25 x 20.5 (W x L) cm, 1.5 mm 1 35165-4175 PROTEAN Plus Hinged Spacer Plate, 25 x 20.5 (W x L) cm, 2.0 mm 1 35165-4176 PROTEAN Plus Comb, 2-D (1 reference well), 20 cm, 1.0 mm 1 35165-4177 PROTEAN Plus Comb, 2-D (1 reference well), 20 cm, 1.5 mm 1 35165-4178 PROTEAN Plus Comb, 2-D (1 reference well), 20 cm, 2.0 mm 1 35165-4179 PROTEAN Plus Comb, 2-D (1 reference well), 25 cm, 1.0 mm 1 35165-4180 PROTEAN Plus Comb, 2-D (1 reference well), 25 cm, 1.5 mm 1 35165-4181 PROTEAN Plus Comb, 2-D (1 reference well), 25 cm, 2.0 mm 1 35

Casting Criterion Cassettes345-9905 Criterion Empty Cassette, 1.0 mm thick with IPG comb 10 sets 35

Casting Multiple Mini-PROTEAN® 3 Gels165-4110 Mini-PROTEAN 3 Multi-Casting Chamber, includes 8 acrylic blocks, 15 separation sheets, each 35

tapered luer fitting, stopcock valve 165-4120 Model 485 Gradient Former, includes body with valve stem and tubing connection kit each 35165-3308 Short Plates 5 35165-3311 Spacer Plates with 1.0 mm Integrated Spacers 5 35165-3312 Spacer Plates with 1.5 mm Integrated Spacers 5 35165-3362 Mini-PROTEAN 3 Comb, IPG well, 1.0 mm 2 35165-3368 Mini-PROTEAN 3 Comb, IPG well, 1.5 mm 2 35

AnyGel™ Stands165-4131 AnyGel Stand, single-row, holds 1 PROTEAN gel, 2 Criterion gels, or 3 mini gels; includes instructions 1 36165-5131 AnyGel Stand, 6-row, holds 6 PROTEAN gels, 12 Criterion gels, or 18 mini gels; includes instructions 1 36165-3221 Mini-PROTEAN 3 Cell and Single-Row AnyGel Stand, includes 165-3301 and 165-4131 1 36165-6020 Criterion Cell and Single-Row AnyGel Stand, includes 165-6001 and 165-4131 1 36

Chapters 4 and11— Casting Multiple Gels — Equipment

48


Reagents and Premixed Solutions163-2111 ReadyPrep Overlay Agarose 50 ml 36161-3111 Certified™ Low-Melt Agarose 25 g 36161-0404 Bromophenol Blue 10 g 36161-0732 10x Tris/Glycine/SDS 1 L 36–37161-0772 10x Tris/Glycine/SDS 5 L 36–37

Running up to 12 PROTEAN Plus Gels or PROTEAN II Ready Gel Precast Gels165-4150 PROTEAN Plus Dodeca™ Cell, 100/120V, includes tank and lid, 14, 15, 37

buffer recirculation pump with tubing 165-4151 PROTEAN Plus Dodeca Cell, 220/240V, includes tank and lid, 14, 15, 37

buffer recirculation pump with tubing

Running Dual PROTEAN XL Gels or PROTEAN II Ready Gel Precast Gels165-3188 PROTEAN II XL Cell, wide format, 1.0 mm 1 system* 14–15, 36–37165-3189 PROTEAN II XL Cell, wide format, 1.5 mm 1 system* 14–15, 36–37165-3190 PROTEAN II XL Cell, wide format, 2.0 mm 1 system* 14–15, 36–37

*Each system includes PROTEAN XL basic unit with casting stand, upper and lower buffer chambers, cooling core, lid with cables, plates, spacers

Running up to 12 Criterion Gels165-4130 Criterion Dodeca Cell, includes tank and lid 14–15, 36–37

Running Dual Criterion Gels165-6001 Criterion Cell, includes tank, lid with power cables 1 cell 14–15, 36–37

Running up to 12 Mini-PROTEAN 3 Gels or Ready Gel Precast Gels 165-4100 Mini-PROTEAN 3 Dodeca Cell, includes 6 clamp assemblies, 2 buffer dams, 14–15, 36–37

drain line, 2 gel releasers165-4101 Mini-PROTEAN 3 Dodeca Cell, with multi-casting chamber 14–15, 35–37

Running Dual Mini-PROTEAN 3 Gels or Ready Gel Precast Gels165-3302 Mini-PROTEAN 3 Electrophoresis Module, for Ready Gel precast gel applications 36–37165-3337 Mini-PROTEAN 3 Casting Module, 1.0 mm, IPG comb 36–37165-3343 Mini-PROTEAN 3 Casting Module, 1.5 mm, IPG comb 36–37

Chapters 4 and11— Running the Second Dimension


165-4035 Disposable Rehydration/Equilibration Tray with Lid, 7 cm 25 pack 36165-4025 Disposable Rehydration/Equilibration Tray with Lid, 11 cm 25 pack 36165-4015 Disposable Rehydration/Equilibration Tray with Lid, 17 cm 25 pack 36163-2108 ReadyPrep 2-D Starter Kit Equilibration Buffer II, without DTT or iodoacetamide, lyophilized 20 ml 36163-2107 ReadyPrep 2-D Starter Kit Equilibration Buffer I, with DTT, lyophilized 20 ml 36161-0730 Urea 250 g 36161-0731 Urea 1 kg 36161-0300 SDS 25 g 13, 36161-0301 SDS 100 g 13, 36161-0302 SDS 1 kg 13, 36161-0418 SDS Solution, 20% 1 L 13, 36161-0798 Resolving Gel Buffer, 1.5 M Tris-HCl, pH 8.8 1 L 13, 36161-0611 Dithiothreitol (DTT) 5 g 36163-2109 Iodoacetamide 30 g 36Sigma Glycerol 36

Chapters 4 and11— Equilibration of Focused Strips for the Second Dimension


Catalog # Detection in Gels161-0436 Coomassie Brilliant Blue R-250 Staining Solution 1 L 16, 38161-0437 Coomassie Brilliant Blue R-250 Staining Solution 4 x 1 L 16, 38161-0439 Coomassie Brilliant Blue R-250 Destaining Solution 4 x 1 L 16, 38161-0786 Bio-Safe™ Colloidal Coomassie Blue G-250 Stain 1 L 16, 38161-0787 Bio-Safe Colloidal Coomassie Blue G-250 Stain 5 L 16, 38170-3125 SYPRO Ruby Protein Gel Stain 1 L 16, 38170-3138 SYPRO Ruby Protein Gel Stain 5 L 16, 38161-0443 Bio-Rad Silver Stain Kit 1 kit 17, 39161-0449 Silver Stain Plus™ Kit 1 kit 17, 39

Methanol, reagent grade 38Acetic Acid, reagent grade 38

165-1771 GelAir™ Drying System, 115 V, 60 Hz, includes dryer, 2 drying frames, 16 clamps, assembly table, 1 system 3950 precut sheets of cellophane support, gel drying solution

165-1772 GelAir Drying System, 230 V, 50 Hz 1 system 39165-1775 GelAir Drying Frames, includes 2 frames and 16 clamps 2 39165-1779 GelAir Cellophane Support, precut sheets 50 sheets 39161-0752 Gel Drying Solution 1 L 39165-1789 HydroTech™ Gel Drying System, 100/120 V, includes Model 583 gel dryer, HydroTech vacuum pump 1 system 39165-1790 HydroTech Gel Drying System, 220/240 V 1 system 39165-0962 Filter Paper Backing, 35 x 45 cm 25 sheets 39165-0963 Cellophane Membrane Backing, 35 x 45 cm 50 sheets 39

Chapters 5 and12 — Protein Detection Methods — Gels


Apparatus and Accessories170-3940 Trans-Blot® SD Semi-dry Electrophoretic Transfer Cell 18170-3939 Trans-Blot Cell with Plate Electrodes and Super Cooling Coil, includes 2 gel holder cassettes, 18

cell with lid and power cables, fiber pads, blot absorbent filter paper170-4070 Criterion Blotter with Plate Electrodes, includes cell assembled with plate electrodes, lid with cables, 18

2 Criterion gel holder cassettes, filter paper pack, fiber pad pack, gel blot assembly tray, roller, sealed ice cooling unit, instructions

165-5052 PowerPac™ 200 Power Supply, 100/120 V 18165-5053 PowerPac 200 Power Supply, 220/240 V 18161-0734 10x Tris/Glycine, to make 25 mM Tris, 192 mM glycine, pH 8.3 (to make Towbin buffer, add methanol) 1 L 18161-0771 10x Tris/Glycine, to make 25 mM Tris, 192 mM glycine, pH 8.3 (to make Towbin buffer, add methanol) 5 L 18161-0778 10x Tris/CAPS Buffer 1 L 18161-0418 SDS Solution, 20% 1 L 18

Methanol, reagent grade 18

Mini-Blot Membranes and Papers170-3966 Extra Thick Blot Absorbent Filter Paper, for semi-dry blotting of mini gels 60 sheets 18162-0186 Sequi-Blot™ PVDF Membrane, 7 x 8.4 cm 10 sheets 18162-0216 Sequi-Blot PVDF/Filter Paper Sandwich, 7 x 8.5 cm (tank blotting) 20 pack 18162-0217 Sequi-Blot PVDF/Filter Paper Sandwich, 7 x 8.5 cm (tank blotting) 50 pack 18

Criterion Blot Membranes and Papers170-3967 Extra Thick Blot Absorbent Filter Paper, for semi-dry blotting of Criterion gels 60 sheets 18162-0180 Sequi-Blot PVDF Membrane, 10 x 15 cm 10 sheets 18162-0236 Sequi-Blot PVDF/Filter Paper Sandwich, 8.5 x 13.5 cm (tank blotting) 20 pack 18162-0237 Sequi-Blot PVDF/Filter Paper Sandwich, 8.5 x 13.5 cm (tank blotting) 50 pack 18

Large Blot Membranes and Papers170-3969 Extra Thick Blot Absorbent Filter Paper, for semi-dry blotting of PROTEAN XL-size gels 30 sheets 18162-0182 Sequi-Blot PVDF Membrane, 20 x 20 cm 10 sheets 18162-0184 Sequi-Blot PVDF Membrane, 24 cm x 3.3 m 1 roll 18

Blot Detection Reagents170-6527 Colloidal Gold Total Protein Stain 500 ml 20161-0400 Coomassie Brilliant Blue R-250 Stain 10 g 20161-0786 Bio-Safe Coomassie G-250 Stain 1 L 20170-3127 SYPRO Ruby Protein Blot Stain 200 ml 20170-6490 Immun-Blot® Kit for Glycoprotein Detection 1 kit 20

Chapters 6 and12 — Protein Detection Methods — Blots

49

50


Image Acquisition Systems170-7850 Molecular Imager FX Pro Plus™ MultiImager System, PC, 100–240 V 21170-7851 Molecular Imager FX Pro Plus MultiImager System, Mac, 100–240 V 21170-7856 Molecular Imager FX Pro™ Fluorescent Imaging System, PC, 100–240 V 21170-7857 Molecular Imager FX Pro Fluorescent Imaging System, Mac, 100–240 V 21170-8010 VersaDoc™ Model 1000 Imaging System, PC, 100/240 V 21173-8011 VersaDoc Model 1000 Imaging System, Mac, 100/240 V 21173-8030 VersaDoc Model 3000 Imaging System, PC, 100/240 V 21173-8031 VersaDoc Model 3000 Imaging System, Mac, 100/240 V 21173-8050 VersaDoc Model 5000 Imaging System, PC, 100/240 V 21173-8051 VersaDoc Model 5000 Imaging System, Mac, 100/240 V 21170-7980 GS-800™ Calibrated Imaging Densitometer, PC 21170-7981 GS-800 Calibrated Imaging Densitometer, Mac 21

Image Analysis Software 170-8603 PDQuest™ 2-D Image Analysis Software, PC 22170-8611 PDQuest 2-D Image Analysis Software, Mac 22170-8637 PDQuest Extended License Contract, entitles licensed software user to request upgrade to update 22

software released during one year period of the contract; network floating licenses also available

Chapter 7 — Image Acquisition and Analysis


165-7042 ProteomeWorks™ Spot Cutter with Fluorescent Enclosure, Windows, includes viewer mode PDQuestsoftware for basic excision; not included, licensed version PDQuest 2-D software 1 23

165-7043 ProteomeWorks Spot Cutter with Fluorescent Enclosure, Mac, includes viewer mode PDQuestsoftware for basic excision; not included, licensed version PDQuest 2-D software 1 23

165-7009 ProteomeWorks Spot Cutter, Windows, includes viewer mode PDQuest software for basic excision; not included, licensed version PDQuest 2-D software 1 23

165-7039 ProteomeWorks Spot Cutter, Mac, includes viewer mode PDQuest software for basic excision; not included, licensed version PDQuest 2-D software 1 23

Chapter 8 — Protein Spot Excision


General Immunodetection Color Detection Systems170-6460 Goat Anti-Rabbit IgG (H + L) Alkaline Phosphatase and BCIP/NBT 1 kit/200 assays 19170-6461 Goat Anti-Mouse IgG (H + L) Alkaline Phosphatase and BCIP/NBT 1 kit/200 assays 19170-6462 Goat Anti-Human IgG (H + L) Alkaline Phosphatase and BCIP/NBT 1 kit/200 assays 19170-6463 Goat Anti-Rabbit IgG (H + L) HRP and 4CN 1 kit/200 assays 19170-6464 Goat Anti-Mouse IgG (H + L) HRP and 4CN 1 kit/200 assays 19170-6465 Goat Anti-Human IgG (H + L) HRP and 4CN 1 kit/200 assays 19

Immun-Star™ Chemiluminescent Detection170-5010 Goat Anti-Mouse Immun-Star Detection Kit 50 miniblots 19170-5011 Goat Anti-Rabbit Immun-Star Detection Kit 50 miniblots 19170-5012 Immun-Star Substrate Pack 50 miniblots 19170-5013 Goat Anti-Mouse Immun-Star Intro Kit 8 miniblots 19170-5014 Goat Anti-Rabbit Immun-Star Intro Kit 8 miniblots 19

Amplified Detection Systems170-6412 Amplified Alkaline Phosphatase Immun-Blot Assay Kit, goat anti-rabbit IgG (H + L)-biotin 50 miniblots 19170-8235 Opti-4CN™ Substrate Kit 50 miniblots 19170-8237 Goat Anti-Mouse Opti-4CN Detection Kit 50 miniblots 19170-8236 Goat Anti-Rabbit Opti-4CN Detection Kit 50 miniblots 19170-8238 Amplified Opti-4CN Substrate Kit 50 miniblots 19170-8240 Goat Anti-Mouse Amplified Opti-4CN Detection Kit 50 miniblots 19170-8239 Goat Anti-Rabbit Amplified Opti-4CN Detection Kit 50 miniblots 19

Chapters 6 and 12 — Protein Detection Methods — Blots (cont.)

Part V

References andRelated Literature


Bers G and Garfin D, Protein and nucleic acid blotting and immunochemicaldetection, Biotechniques 3, 276–288 (1985)

Bjellqvist B, Ek K, Righetti PG, Gianazza E, Gorg A, Westermeier R and PostelW, Isoelectric focusing in immobilized pH gradients: principle, methodology andsome applications, J Biochem Biophys Methods 6, 317–339 (1982)

Bjellqvist B, Hughes GJ, Pasquali C, Paquet N, Ravier F, Sanchez JC, Frutiger Sand Hochstrasser D, The focusing positions of polypeptides in immobilized pHgradients can be predicted from their amino acid sequences, Electrophoresis14, 1023–1031 (1993a).

Bjellqvist B, Sanchez JC, Pasquali C, Ravier F, Paquet N, Frutiger S, HughesGJ and Hochstrasser D, Micropreparative two-dimensional electrophoresisallowing the separation of samples containing milligram amounts of proteins,Electrophoresis 14, 1375–1378 (1993b)

Bollag DM, Rozycki MD and Edelstein SJ, Protein Methods, 2nd ed. Wiley-Liss, Inc., New York (1996)

Carroll K, Ray K, Helm B and Carey E, Two-dimensional electrophoresis revealsdifferential protein expression in high- and low-secreting variants of the ratbasophilic leukaemia cell line, Electrophoresis 21, 2476–2486 (2000)

Celis JE (ed) Cell Biology. A Laboratory Handbook, Vol. 2, Academic Press,New York (1998)

Celis JE, Celis P, Ostergaard M, Basse B, Lauridsen JB, Ratz G, RasmussenHH, Orntoft TF, Hein B, Wolf H and Celis A, Proteomics and immunohisto-chemistry define some of the steps involved in the squamous differentiation ofthe bladder transitional epithelium: a novel strategy for identifying metaplasticlesions, Cancer Res 59, 3003–3009 (1999)

Chevallet M, Santoni V, Poinas A, Rouquie D, Fuchs A, Kieffer S, Rossignol M,Lunardi J, Garin J and Rabilloud T, New zwitterionic detergents improve theanalysis of membrane proteins by two-dimensional electrophoresis,Electrophoresis 19, 1901–1909 (1998)

Choe LH and Lee KH, A comparison of three commercially available isoelectricfocusing units for proteome analysis: The multiphor, the IPGphor and the protean IEF cell, Electrophoresis 21, 993–1000 (2000)

Cordwell SJ, Basseal DJ, Bjellqvist B, Shaw DC and Humphery-Smith I,Characterisation of basic proteins from Spiroplasma melliferum using novelimmobilised pH gradients, Electrophoresis 18, 1393–1398 (1997)

Corthals GL, Wasinger VC, Hochstrasser DF and Sanchez JC, The dynamicrange of protein expression: a challenge for proteomic research, Electrophoresis21, 1104–1115 (2000)

Cutler P, Bell DJ, Birrell HC, Connelly JC, Connor SC, Holmes E, Mitchell BC,Monte SY, Neville BA, Pickford R, Polley S, Schneider K and Skehel JM, An integrated proteomic approach to studying glomerular nephrotoxicity,Electrophoresis 20, 3647–3658 (1999)

Ducret A, Bruun CF, Bures EJ, Marhaug G, Husby G and Aebersold R,Characterization of human serum amyloid A protein isoforms separated by two-dimensional electrophoresis by liquid chromatography/electrospray ionization tandem mass spectrometry, Electrophoresis 17, 866–876 (1996)

Deutscher MP (ed) Guide to Protein Purification, Methods Enzymol 182 (1990)

Dunn MJ, Detection of total proteins on western blots of 2-D polyacrylamidegels, Methods Mol Biol 112, 319–329 (1999)

Eckerskorn C, Jungblut P, Mewes W, Klose J and Lottspeich F, Identification of mouse brain proteins after two-dimensional electrophoresis and electro-blotting by microsequence analysis and amino acid composition analysis,Electrophoresis 9, 830–838 (1988)

Fegatella F, Ostrowski M and Cavicchioli R, An assessment of protein profilesfrom the marine oligotrophic ultramicrobacterium, Sphingomonas sp. strainRB2256, Electrophoresis 20, 2094–2098 (1999)

Fountoulakis M, Takacs MF, Berndt P, Langen H and Takacs B, Enrichment oflow abundance proteins of Escherichia coli by hydroxyapatite chromatography,Electrophoresis 20, 2181–2195 (1999)

Garfin DE, Electrophoretic methods, pp 53–109 in Glasel JA and Duetscher MP(eds) Introduction to Biophysical Methods for Protein and Nucleic AcidResearch, Academic Press, San Diego (1995)

Garfin DE, Isoelectric focusing, pp 263–298 in Ahuja S (ed) Separation Scienceand Technology, Vol 2, Academic Press, San Diego (2000)

Garfin DE and Bers G, Basic aspects of protein blotting, pp 5–42 in Baldo BAand Tovey ER (eds) Protein Blotting, Karger, Basel (1989)

Gingrich JC, Davis DR and Nguyen Q, Multiplex detection of quantification ofproteins on western blots using fluorescent probes, Biotechniques 29, 636–642(2000)

Görg A, Two-dimensional electrophoresis, Nature 349, 545–546 (1991)

Görg A, Obermaier C, Boguth G, Harder A, Scheibe B, Wildgruber R andWeiss W, The current state of two-dimensional electrophoresis with immobilizedpH gradients, Electrophoresis 21, 1037–1053 (2000)

Görg A, Obermaier C, Boguth G and Weiss W, Recent developments in two-dimensional gel electrophoresis with immobilized pH gradients: wide pH gradients up to pH 12, longer separation distances and simplified procedures,Electrophoresis 20, 712–717 (1999)

Görg A, Postel W, Domscheit A and Günther S, Methodology of two-dimensional electrophoresis with immobilized pH gradients for the analysis of cell lysates and tissue proteins, in Endler AT and Hanash S (eds) Two-Dimensional Electrophoresis. Proceedings of the International Two-DimensionalElectrophoresis Conference, Vienna, Nov. 1988, VCH, Weinheim FRG (1989)

Gottlieb M and Chavko M, Silver staining of native and denatured eucaryoticDNA in agarose gels, Anal Biochem 165, 33–37 (1987)

Harlow E and Lane D (eds), Antibodies: A Laboratory Manual, Cold SpringHarbor Laboratory, Cold Spring Harbor, NY (1988)

Herbert BR, Molloy MP, Gooley AA, Walsh BJ, Bryson WG and Williams KL,Improved protein solubility in two-dimensional electrophoresis using tributylphosphine as reducing agent, Electrophoresis 19, 845–851 (1998)

Hermann T, Finkemeier M, Pfefferle W, Wersch G, Kramer R and Burkovski A,Two-dimensional electrophoretic analysis of Corynebacterium glutamicummembrane fraction and surface proteins, Electrophoresis 21, 654–659 (2000)

Ledue TB and Garfin D, Immunofixation and immunoblotting, pp 54–64 in Rose NR, Conway de Macario E, Folds JD, Lane HC and Nakamura RM (eds)Manual of Clinical Laboratory Microbiology, 5th ed, American Society forMicrobiology, Washington, D.C. (1997)

Link AJ (ed) 2-D Proteome Analysis Protocols, Methods Mol Biol 112 (1999)

Lubec G, Nonaka M, Krapfenbauer K, Gratzer M, Cairns N and FountoulakisM, Expression of the dihydropyrimidinase related protein 2 (DRP-2) in Downsyndrome and Alzheimer’s disease brain is downregulated at the mRNA anddysregulated at the protein level, J Neural Transm Suppl 57, 161–177 (1999)

Macri J, McGee B, Thomas JN, Du P, Stevenson TI, Kilby GW and RapundaloST, Cardiac sarcoplasmic reticulum and sarcolemmal proteins separated bytwo-dimensional electrophoresis: surfactant effects on membrane solubilization,Electrophoresis 21, 1685–1693 (2000)

Masuoka J, Glee PM and Hazen KC, Preparative isoelectric focusing andpreparative electrophoresis of hydrophobic Candida albicans cell wall proteinswith in-line transfer to polyvinylidene difluoride membranes for sequencing,Electrophoresis 19, 675–678 (1998)

Merril CR, Goldman D, Sedman SA and Ebert MH, Ultrasensitive stain for proteins in polyacrylamide gels shows regional variation in cerebrospinal fluidproteins, Science 211, 1437–1438 (1981)

Miller MJ, Computer-assisted analysis of two-dimensional gel electrophero-grams, pp 182–217 in Chrambach A, Dunn MJ and Radola BJ (eds) Advancesin Electrophoresis, Vol. 3, Verlag Chemie, Weinheim (1989)

Molloy MP, Two-dimensional electrophoresis of membrane proteins using immobilized pH gradients, Anal Biochem 280, 1–10 (2000)

Molloy MP, Herbert BR, Walsh BJ, Tyler MI, Traini M, Sanchez JC, HochstrasserDF, Williams KL and Gooley AA, Extraction of membrane proteins by differentialsolubilization for separation using two-dimensional gel electrophoresis,Electrophoresis 19, 837–844 (1998)

Morel V, Poschet R, Traverso V and Deretic D, Towards the proteome of therhodopsin-bearing post-Golgi compartment of retinal photoreceptor cells,Electrophoresis 21, 3460–3469 (2000)

References

51

52

Nilsson CL, Larsson T, Gustafsson E, Karlsson KA and Davidsson P,Identification of protein vaccine candidates from Helicobacter pylori using apreparative two-dimensional electrophoretic procedure and mass spectrometry, Anal Chem 72, 2148–2153 (2000)

Patton WF, A thousand points of light: the application of fluorescence detection technologies to two-dimensional gel electrophoresis and proteomics, Electrophoresis 21, 1123–1144 (2000)

Pennington SR and Dunn MJ, Proteomics. From Protein Sequence toFunction, Springer/Bios, New York (2001)

Rabilloud T, Use of thiourea to increase the solubility of membrane proteins intwo-dimensional electrophoresis, Electrophoresis 19, 758–760 (1998)

Rabilloud T, Solubilization of proteins in 2-D electrophoresis. An outline,Methods Mol Biol 112, 9–19 (1999)

Rabilloud T (ed) Proteome Research: Two Dimensional Gel Electrophoresisand Identification Tools, Springer, Berlin (2000)

Righetti PG, Isoelectric Focusing: Theory, Methodology and Applications,Elsevier, Amsterdam (1983)

Righetti PG, Recent developments in electrophoretic methods, J Chromatogr516, 3–22 (1990)

Sanchez JC, Rouge V, Pisteur M, Ravier F, Tonella L, Moosmayer M, WilkinsMR and Hochstrasser DF, Improved and simplified in-gel sample applicationusing reswelling of dry immobilized pH gradients, Electrophoresis 18,324–327 (1997)

Taylor RS, Wu CC, Hays LG, Eng JK, Yates JR 3rd and Howell KE,Proteomics of rat liver Golgi complex: minor proteins are identified throughsequential fractionation, Electrophoresis 21, 3441–3459 (2000)

Wilkins MR, Gasteiger E, Sanchez JC, Bairoch A and Hochstrasser DF, Two-dimensional gel electrophoresis for proteome projects: the effects of protein hydrophobicity and copy number, Electrophoresis 19, 1501–1505(1998)

Wilkins MR, Sanchez JC, Gooley AA, Appel RD, Humphery-Smith I,Hochstrasser DF and Williams KL, Progress with proteome projects: Why allproteins expressed by a genome should be identified and how to do it,Biotechnol Genet Eng Rev 13, 19–50 (1996)

Wilkins MR, Williams KL, Appel RD and Hochstrasser DF (eds) ProteomeResearch: New Frontiers in Functional Genomics, Springer, Berlin (1997)

Related Bio-Rad LiteratureGeneral2619 The ProteomeWorks system brochure

2563 2-D in a day brochure

2670 Separation and comparison of proteins from virulent and nonvirulent strains of the fish pathogen Flavobacterium psychrophilum, using a 2-D electrophoretic approach

RP0005 Analytical and micropreparative two-dimensional electrophoresisof proteins (reprinted from Methods: A Companion to Methodsin Enzymology, vol 3, pp 98–108, 1991)

RP0014 Isoelectric focusing nonporous RP HPLC: A two-dimensional liquid-phase separation method for mapping of cellular proteins with identification using MALDI-TOF mass spectrometry (reprinted from Anal Chem 72, 1099–1111, 2000)

RP0015 Identification of protein vaccine candidates from Helicobacterpylori using a preparative 2-D electrophoretic procedure and mass spectrometry (reprinted from Anal Chem 72,2148–2153, 2000)

RP0016 Isoelectric focusing (reprinted from Garfin 2000)

RP0017 Electrophoretic methods (reprinted from Garfin 1995)

Sample Preparation 2495 ReadyPrep™ sequential extraction kit

2404 SmartSpec™ 3000 spectrophotometer brochure

2610 RC DC™ protein assay

RP0008 Extraction of membrane proteins by differential solubilization forseparation using two-dimensional gel electrophoresis (reprintedfrom Electrophoresis 19, 837–844, 1998)

2634 Quantitation of proteins by the Bradford, Lowry, and BCA protein assays using the SmartSpec 3000 spectrophotometer

First-Dimension Electrophoresis2642 ReadyPrep 2-D starter kit

2442 ReadyStrip™ IPG strips

2587 High-performance 2-D gel electrophoresis using narrow pH-range ReadyStrip IPG strips

2426 PROTEAN®

IEF system brochure

RP0013 A comparison of three commercially available isoelectric focusing units for proteome analysis (reprinted fromElectrophoresis 21, 993–1000, 2000)

Second-Dimension Electrophoresis2571 Mini-PROTEAN

®3 Dodeca cell brochure

2710 Criterion electrophoresis system brochure

2622 Criterion Dodeca cell brochure

2621 PROTEAN Plus Dodeca cell brochure

2444 Large precast gels for 2-D electrophoresis

2317 Electrophoresis buffers and solutions brochure

2702 AnyGel stands flier

Immunoblotting 2033 Western blotting products brochure

2558 The Criterion blotter flier

1956 Immun-Blot®

kit for glycoprotein detection

Protein Staining in Gels2423 Bio-Safe Coomassie stain flier

2533 SYPRO Ruby protein gel stain flier

Image Acquisition and Analysis2596 GS-800 calibrated densitometer brochure

2586 Molecular Imager FX Pro Plus system brochure

2597 Three-color imaging with the Moleculer Imager FX Pro Plus system and Quantity One

®software

2566 Normalizing between 2-D gels with PDQuest software

2629 PDQuest software citations

2553 PDQuest tutorial CD

2443 PDQuest software brochure

Protein Spot Excision2655 ProteomeWorks spot cutter fluorescent enclosure flier

2425 ProteomeWorks spot cutter brochure

2503 Performance of the 2-D spot cutter system

Protein IdentificationBR30/DAM MassPREP station — the robotic protein handling system

BR29/DAM M@LDI 2-D gel-MS analyzer

Bioinformatics2637 WorksBase software for proteomics

53

Bio-Rad TrademarksAnyGel™

Bio-Ice™

Bio-Lyte®

Bio-Rad®

Bio-Rex®

Bio-Safe™

Certified™

Criterion™

DC™

Dodeca™

Econo-Pac®

GelAir™

GS-800™

HydroTech™

Immun-Blot®

Immun-Star™

Micro Bio-Spin™

Mini-PROTEAN®

Mini Trans-Blot®

Molecular Imager®

Molecular Imager FX Pro™

Molecular Imager FX Pro Plus™

PDQuest™

PowerPac™

PROTEAN®

PROTEAN Plus™

ProteomeWorks™

RC DC™

Ready Gel®

ReadyPrep™

ReadyStrip™

Rotofor®

Sequi-Blot™

Silver Stain Plus™

SmartSpec™

Trans-Blot®

VersaDoc™

WorksBase™

TrademarksThe following trademarks are the property of their respective owners:

Trademark OwnerCapLC Micromass, Ltd.Coomassie Imperial Chemical Industries PLCKimwipes Kimberly-ClarkM@LDI HT Micromass, Ltd.Mac Apple ComputerMassLynx Micromass, Ltd.MassPREP Micromass, Ltd.Parafilm American National Can Co.ProteinLynx Micromass, Ltd.Q-Tof micro Micromass, Ltd.Q-Tof Ultima Micromass, Ltd.SYPRO Molecular Probes, Inc.Triton Union Carbide Chemicals and Plastics Technology Corp.Tween ICI Americas, Inc.Vaseline Chesebrough Ponds, Inc.Windows Microsoft Corp.ZipTip Millipore

The ProteomeWorks system is the global alliance between Bio-Rad Laboratories, Inc. (USA) and Micromass, Ltd. (UK), dedicated to furthering proteomics research.

Bio-Rad is licensed to sell SYPRO products for research use only, under US patent 5,616,502.

AcknowledgementsEditors:

Dave Garfin

Lauri Heerdt

Contributors:


Linda Castle

Emily Dale

Adriana Harbers

Bruce Sadownick

William Strong

Christina Whitman

Mingde Zhu

Special thanks:

Dr A Posch

GPC, Munich

S Cordwell, C Vockler, B Walsh

Australian Proteome Analysis Facility

D Bladier, OM Caron, R Joubert-Caron, N Imam,

F Montandon, and F Poirier

Laboratoire de Biochimie des Proteines et Proteomique

Universite Paris 13

S Hoving, J van Oostrom, and H Voshol

Novartis Pharma AG

Functional Genomics Area

Basel, Switzerland

Life ScienceGroup

01-769 0202 Sig 1001Bulletin 2651 US/EG Rev B

Bio-Rad Laboratories, Inc.

Web site www.bio-rad.com USA (800) 4BIORAD Australia 02 9914 2800 Austria (01)-877 89 01 Belgium 09-385 55 11 Brazil 55 21 507 6191Canada (905) 712-2771 China (86-21) 63052255 Denmark 45 44 52-1000 Finland 358 (0)9 804 2200 France 01 47 95 69 65 Germany 089 318 84-177 Hong Kong 852-2789-3300 India (91-124) 6398112/113/114, 6450092/93 Israel 03 951 4127 Italy 39 02 216091 Japan 03-5811-6270 Korea 82-2-3473-4460 Latin America 305-894-5950 Mexico 52 5 534 2552 to 54 The Netherlands 0318-540666 New Zealand 64-9-4152280 Norway 47-23-38-41-30 Portugal 351-21-472-7700 Russia 7 095 721 1404 Singapore 65-2729877 South Africa 00 27 11 4428508 Spain 590 5200 Sweden 46 (0)8-55 51 27 00 Switzerland 061 717-9555 United Kingdom 0800-181134

Contents

About this handbook..................................................................................... ii

Acronyms and symbols used ....................................................................... iii

Capillary electrophoresis ...............................................................................1

Electrophoresis terminology ..........................................................................3

Electroosmosis ...............................................................................................4

Flow dynamics, efficiency, and resolution ....................................................6

Capillary diameter and Joule heating ............................................................9

Effects of voltage and temperature ..............................................................11

Modes of capillary electrophoresis ..............................................................12

Capillary zone electrophoresis ..........................................................12

Isoelectric focusing ...........................................................................18

Capillary gel electrophoresis ............................................................21

Isotachophoresis ...............................................................................26

Micellar electrokinetic capillary chromatography ............................28

Selecting the mode of electrophoresis .........................................................36

Approaches to methods development by CZE and MECC .........................37

Suggested reading ........................................................................................40

ii

About this handbook

This handbook, the first of a series on modern high performance capillaryelectrophoresis (CE), is intended for scientists who are contemplating use ofor have recently started using this rapidly evolving family of techniques.The goals of this book are: to introduce you to CE; to help you understandthe mechanisms of the various modes of CE; to guide you in methodselection; and to provide a set of approaches towards methods developmentfor both large and small molecules.

iii

Acronyms and symbols used

The following acronyms and symbols are used throughout this handbook.

BSA bovine serum albuminCE capillary electrophoresisCTAB cetyltrimethylammonium bromideCGE capillary gel electrophoresisCMC critical micelle concentrationCZE capillary zone electrophoresisDMF dimethylformamideDMSO dimethyl sulfoxideE electric field strengthEDTA ethylenediaminetetraacetic acidEOF electroosmotic flowEPF electrophoretic flowHPLC high performance liquid chromatographyIEF isoelectric focusingITP isotachophoresisLC liquid chromatographyLd length of capillary to the detectorLt total capillary lengthMECC micellar electrokinetic capillary chromatographyµep electrophoretic mobilityPAGE polyacrylamide gel electrophoresisPCR polymerase chain reactionpI isoelectric pointSDS sodium dodecyl sulfateTHF tetrahydrofuranUV ultravioletV voltV voltageveo electroosmotic flow velocityvep electrophoretic velocity

1

Capillary electrophoresis

Capillary electrophoresis (CE) is a family of related techniques that employnarrow-bore (20-200 µm i.d.) capillaries to perform high efficiency separa-tions of both large and small molecules. These separations are facilitated bythe use of high voltages, which may generate electroosmotic and electro-phoretic flow of buffer solutions and ionic species, respectively, within thecapillary. The properties of the separation and the ensuing electropherogramhave characteristics resembling a cross between traditional polyacrylamidegel electrophoresis (PAGE) and modern high performance liquid chroma-tography (HPLC).

CE offers a novel format for liquid chromatography and electro-phoresis that:

• employs capillary tubing within which the electrophoreticseparation occurs;

• utilizes very high electric field strengths, often higher than500 V/cm;

• uses modern detector technology such that the electrophero-gram often resembles a chromatogram;

• has efficiencies on the order of capillary gas chromatographyor even greater;

• requires minute amounts of sample;

• is easily automated for precise quantitative analysis and easeof use;

• consumes limited quantities of reagents;

• is applicable to a wider selection of analytes compared toother analytical separation techniques.

The basic instrumental configuration for CE is relatively simple.All that is required is a fused-silica capillary with an optical viewingwindow, a controllable high voltage power supply, two electrode assem-blies, two buffer reservoirs, and an ultraviolet (UV) detector. The ends ofthe capillary are placed in the buffer reservoirs and the optical viewingwindow is aligned with the detector. After filling the capillary with buffer,the sample can be introduced by dipping the end of the capillary into thesample solution and elevating the immersed capillary a foot or so above thedetector-side buffer reservoir. Virtually all of the pre-1988 work in CE was

2

carried out on homemade devices following this basic configuration. Whilerelatively easy to use for experimentation, these early systems were incon-venient for routine analysis and too imprecise for quantitative analysis.

A diagram of a modern instrument, the P/ACE™ 2000 Series, isillustrated in Figure 1. Compared to the early developmental instruments,this fully automated instrument offers computer control of all operations,pressure and electrokinetic injection, an autosampler and fraction collector,automated methods development, precise temperature control, and anadvanced heat dissipation system. Automation is critical to CE sincerepeatable operation is required for precise quantitative analysis.

ElectrolyteBuffer

ElectrolyteBuffer

HV

ReservoirReservoir

CapillaryInlet

CapillaryOutlet

Detector

DataAcquisition

Figure 1. Basic Configuration of the P/ACE Capillary ElectrophoresisSystem

3

Electrophoresis terminology

There are a few significant differences between the nomenclature ofchromatography and capillary electrophoresis. For example, a fundamentalterm in chromatography is retention time. In electrophoresis, under idealconditions, nothing is retained, so the analogous term becomes migrationtime. The migration time (tm) is the time it takes a solute to move from thebeginning of the capillary to the detector window.

Other fundamental terms are defined below. These include theelectrophoretic mobility, µep (cm2/Vs), the electrophoretic velocity,vep (cm/s), and the electric field strength, E (V/cm). The relationshipsbetween these factors are shown in Equation 1.

µ epvE

L tm

L t=

ep=

d

V (1)

Several important features can be seen from this equation:

1) Velocities are measured terms. They are calculated bydividing the migration time by the length of the capillary tothe detector, Ld.

2) Mobilities are determined by dividing the velocity by the fieldstrength. The mobility is independent of voltage and capillarylength but is highly dependent on the buffer type and pH aswell as temperature.

3) Two capillary lengths are important: the length to the detector,Ld, and the total length, Lt. While the measurable separationoccurs in the capillary segment, Ld, the field strength iscalculated by dividing the voltage by the length of the entirecapillary, Lt. The excess capillary length, Lt - Ld, is required tomake the connection to the buffer reservoir. For the P/ACEsystem, this length is 7 cm. By reversing the configuration ofthe system, this 7-cm length of capillary can be used toperform very rapid separations.

Equation 1 is only useful for determining the apparent mobility. Tocalculate the actual mobility, the phenomenon of electroosmotic flow mustbe accounted for. To perform reproducible electrophoresis, the electroos-motic flow must be carefully controlled.

4

Electroosmosis

One of the fundamental processes that drive CE is electroosmosis. Thisphenomenon is a consequence of the surface charge on the wall of thecapillary. The fused silica capillaries that are typically used for separationshave ionizable silanol groups in contact with the buffer contained within thecapillary. The pI of fused silica is about 1.5. The degree of ionization iscontrolled mainly by the pH of the buffer.

The electroosmotic flow (EOF) is defined by

v Eeo=

∈ζ4πη (2)

where ∈ is the dielectric constant, η is the viscosity of the buffer, and ζ isthe zeta potential measured at the plane of shear close to the liquid-solidinterface.

The negatively-charged wall attracts positively-charged ions from thebuffer, creating an electrical double layer. When a voltage is applied acrossthe capillary, cations in the diffuse portion of the double layer migrate in thedirection of the cathode, carrying water with them. The result is a net flowof buffer solution in the direction of the negative electrode. This electroos-motic flow can be quite robust, with a linear velocity around 2 mm/s at pH 9in 20 mM borate. For a 50µm i.d. capillary, this translates into a volumeflow of about 4 nL/s. At pH 3 the EOF is much lower, about 0.5 nL/s.

The zeta potential is related to the inverse of the charge per unit surfacearea, the number of valence electrons, and the square root concentration ofthe electrolyte. Since this is an inverse relationship, increasing the concen-tration of the electrolyte decreases the EOF.

As we will see later on, the electroosmotic flow must be controlled oreven suppressed to run certain modes of CE. On the other hand, the EOFmakes possible the simultaneous analysis of cations, anions, and neutralspecies in a single analysis. At neutral to alkaline pH, the EOF is suffi-ciently stronger than the electrophoretic migration such that all species areswept towards the negative electrode. The order of migration is cations,neutrals, and anions.

5

The effect of pH on EOF is illustrated in Figure 2. Imagine thata zwitterion such as a peptide is being separated under each of the twoconditions described in the figure. At high pH, EOF is large and the peptideis negatively charged. Despite the peptide’s electrophoretic migrationtowards the positive electrode (anode), the EOF is overwhelming, and thepeptide migrates towards the negative electrode (cathode). At low pH, thepeptide is positively charged and EOF is very small. Thus, peptide electro-phoretic migration and EOF are towards the negative electrode. In untreatedfused silica capillaries most solutes migrate towards the negative elec-trode regardless of charge when the buffer pH is above 7.0. At acidicbuffer pH, most zwitterions and cations will also migrate towards thenegative electrode.

O++

+

High pH

Low pH

+

O O O O++ ++ ++ ++

OH O OH O OH+ + + +

Electroosmotic Flow

–

–

Figure 2. Effect of pH on theElectroosmotic Flow

To ensure that a system is properly controlled, it is often necessary tomeasure the EOF. This is accomplished by injecting a neutral solute andmeasuring the time it takes to reach the detector. Solutes such as methanol,acetone, and mesityl oxide are frequently employed. In the micellar electro-kinetic capillary chromatography (MECC) technique to be discussed later,a further requirement that the marker solute not partition into the micelle isalso imposed.

To perform techniques such as isoelectric focusing (IEF) or isotacho-phoresis (ITP), EOF must be suppressed. This is possible if an uncharged,e.g., Teflon,1 or a suitably coated capillary is used. Additives such asmethylcellulose are also effective in suppressing EOF. EOF suppressionwill be discussed later.

1 Teflon is a trademark of E.I. Du Pont de Nemours & Co.

6

Flow dynamics, efficiency, and resolution

When employing a pressure-driven system such as a liquid chromatograph,the frictional forces at the liquid-solid interfaces, such as the packing andthe walls of the tubing, result in substantial pressure drops. Even in an opentube, the frictional forces are severe enough at low flow rates to result inlaminar or parabolic flow profiles. As a consequence of parabolic flow, across-sectional flow gradient, shown in Figure 3, occurs in the tube,resulting in a flow velocity that is highest in the middle of the tube andapproaches zero at the tubing wall. This velocity gradient results in substan-tial bandbroadening.

Cross-Sectional Flow ProfileDue to Electroosmotic Flow

Cross-Sectional Flow ProfileDue to Hydrodynamic Flow

Figure 3. Capillary Flow Profiles

In electrically driven systems, the driving force of the EOF is uniformlydistributed along the entire length of the capillary. As a result, there is nopressure drop and the flow velocity is uniform across the entire tubingdiameter except very close to the wall where the velocity again approacheszero.

7

The efficiency of a system can be derived from fundamental principles.The migration velocity, vep, is simply

Vµ epv E L=ep = µ ep (3)

The migration time, t, is defined as

v µ ep VLep

=t2

=L

(4)

During migration through the capillary, molecular diffusion occursleading to peak dispersion, σ2, calculated as

µ ep VL= t2 2Dσ m =

2Dm2

(5)

where Dm = the solute’s diffusion coefficient cm2/s. The number of theoreti-cal plates is given as

L2σ

2=N (6)

Substituting the dispersion equation into the plate count equation yields

=N 2Dm

µ epV(7)

The dispersion, σ2, in this simple system is assumed to be time-relateddiffusion only. The equation indicates that macromolecules such as proteinsand DNA, which have small diffusion coefficients, D, will generate thehighest number of theoretical plates. In addition, the use of high voltageswill also provide for the greatest efficiency by decreasing the separationtime. The practical voltage limit with today’s technology is about 30 kV.The practical limit of field strength (one could use very short capillaries togenerate high field strength) is Joule heating. Joule heating is a conse-quence of the resistance of the buffer to the flow of current. The problemsof heat generation/dissipation will be covered shortly.

Substituting some numbers into the plate count equation using theprotein horse heart myoglobin (MW 13,900) as an illustration, whereµep = 0.65× 10-4 cm2/Vs (20 mM bicine/TEA buffer, pH 8.5) andDm = 1 × 10-6 cm2/s at 30,000 V, gives a plate count of 975,000 theoreticalplates.

8

In spite of the diffusional limitation, CE is still useful for small-molecule separations because µep is a function of the charge-to-mass ratio.Small molecules tend to be more mobile. For example, the mobility ofquinine sulfate is 4 × 10-4 cm2/Vs. Despite the higher diffusion coefficientof 0.7 × 10-5 cm2/s, the equation solves for N = 857,000 theoretical plateswhen V = 15,000 volts.

The resolution, Rs, between two species is given by the expression

=Rsµ ep

N∆µ ep

1

4(8)

where ∆µep is the difference in electrophoretic mobility between the twospecies, µ ep is the average electrophoretic mobility of the two species andN is the number of theoretical plates. If we substitute the plate countequation, we get

=Rsµ ep Vµ ep Dm

(0.177) (9)

This expression indicates that increasing the voltage is a limited meansof improving resolution. To double the resolution, the voltage must bequadrupled. The key to high resolution is to increase ∆µep. The control ofmobility is best accomplished through selection of the proper mode ofcapillary electrophoresis coupled with selection of the appropriate buffers.Both of these areas will be covered later in this book.

9

Capillary diameter and Joule heating

The production of heat in CE is the inevitable result of the application ofhigh field strengths. Two major problems arise from heat production:temperature gradients across the capillary and temperature changes withtime due to ineffective heat dissipation.

The rate of heat generation in a capillary can be approximated asfollows

=dH iVdt LA

(10)

where L is the capillary length and A, the cross-sectional area. Since i = V/Rand R = L/kA where k is the conductivity, then

=dH kVdt L 2

2

(11)

The amount of heat generated is proportional to the square of the fieldstrength. Either decreasing the voltage or increasing the length of thecapillary has a dramatic effect on the heat generation. The use of low-conductivity buffers is also helpful in this regard although sample loading isadversely affected.

Temperature gradients across the capillary are a consequence of heatdissipation. Since heat is dissipated by diffusion, it follows that the tempera-ture at the center of the capillary should be greater than at the capillarywalls.

Cross-Sectional Temperature Gradientand Electrophoretic Velocity Profile

vep ∆TR

Figure 4. Cross-Sectional Thermal Gradient andthe Electrophoretic Velocity Profile

10

Since viscosity is lower at higher temperatures, it follows that both the EOFand electrophoretic mobility (EPM) will increase as well. Mobility for mostions increases by 2% per degree kelvin. The result is a flow profile thatresembles hydrodynamic flow, and bandbroadening occurs. Operating withnarrow-diameter capillaries improves the situation for two reasons: thecurrent passed through the capillary is reduced by the square of the capillaryradius, and the heat is more readily dissipated across the narrower radialpath. The resulting thermal gradient is proportional to the square of thediameter of the capillary, which can be approximated from the followingequation

Wr4K

T∆ 0.242

= (12)

where W is the power, r is the capillary radius, and K, the thermal conduc-tivity.

The second problem is ineffective heat dissipation. If heat is notremoved at a rate equal to its production, a gradual but progressive tempera-ture rise will occur until equilibrium is reached. Depending on the specificexperimental conditions, imprecision in migration time will result due tovariance in both EOF and electrophoretic velocity. Narrow-diametercapillaries help heat dissipation, but effective cooling systems are requiredto ensure heat removal. Liquid cooling is the most effective means of heatremoval and capillary temperature control.

Capillary inner diameters range from 20-200µm. From the standpointof resolution, the smaller the capillary i.d., the better the separation.However, smaller-bore capillaries yield poorer limits of detection due toreduced detector path length and sample loadability. Narrow capillaries arealso more prone to clogging. As long as buffers are filtered through<0.5-µm filters, clogging is seldom a problem in the above mentionedsize range. Since it may be impractical to filter samples, high speed cen-trifugation is usually sufficient to settle suspended particles.

11

Effects of voltage and temperature

Both the electroosmotic and electrophoretic velocities are directly propor-tional to the field strength, so the use of the highest voltages possible willresult in the shortest times for the separation. Theory predicts that shortseparation times will give the highest efficiencies since diffusion is the mostimportant feature contributing to bandbroadening. The limiting factor hereis Joule heating. Experimentally, the optimal voltage is determined byperforming runs at increasing voltages until deterioration in resolution isnoted.

The electrophoretic mobility (Eq. 13) and the electroosmotic flow(Eq. 2) expressions both contain a viscosity term in the denominator.Viscosity is a function of temperature; therefore, precise temperaturecontrol is important. As the temperature increases, the viscosity decreases;thus, the electrophoretic mobility increases as well. Some buffers such asTris are known to be pH-sensitive with temperature. For complex separa-tions such as peptide maps, even small pH shifts can alter the selectivity.

Most separations are performed at 25°C (i.e., near room temperature).With liquid cooling of the capillary it is possible to maintain excellenttemperature control, even with high-concentration buffers and large-borecapillaries. Whenever temperature control starts to become a problem, theusual strategy is to use a smaller-bore capillary (less current reduces theheat produced) or a longer capillary (more surface area dissipates the heatgenerated). An alternative is to reduce the buffer concentration, but thisalso reduces peak efficiency by decreasing the focusing effect. Inadequatetemperature control is the main reason for using low-concentration(e.g., 20 mM) buffers or operating at elevated temperatures.

12

Modes of capillary electrophoresis

Capillary electrophoresis comprises a family of techniques that havedramatically different operative and separative characteristics. The tech-niques are:

• Capillary zone electrophoresis

• Isoelectric focusing

• Capillary gel electrophoresis

• Isotachophoresis

• Micellar electrokinetic capillary chromatography

Each of these modes of CE will be covered in the following sections.Other less mature modes of electrophoresis such as electroosmotic chroma-tography will not be covered here.

Capillary zone electrophoresis

Capillary zone electrophoresis (CZE), also known as free solution capillaryelectrophoresis, is the simplest form of CE. The separation mechanism isbased on differences in the charge-to-mass ratio. Fundamental to CZE arehomogeneity of the buffer solution and constant field strength through-out the length of the capillary.

Following injection and application of voltage, the components of asample mixture separate into discrete zones as shown in Figure 5. Thefundamental parameter, electrophoretic mobility, µep, can be approximatedfrom Debeye-Huckel-Henry theory

=µ epq

6πηR (13)

where q is the net charge, R is the Stokes radius, and η is the viscosity. Inpractice, analytical chemists infrequently calculate electrophoretic mobili-ties, although some understanding of the parameters describing the phenom-enon is useful.

13

+ —

+ —

Figure 5. Capillary Zone Electrophoresis

The net charge is usually pH dependent. For example, within the pHrange of 4-10, the net charge on sodium is constant as is its mobility. Otherspecies such as acetate or glutamate are negatively charged within that pHrange and thus have negative mobilities (they migrate towards the positiveelectrode). At alkaline pH, their net migration will still be towards thenegative electrode because of the EOF. Zwitterions such as amino acids,proteins, and peptides exhibit charge reversal at their pI’s and, likewise,shifts in the direction of electrophoretic mobility.

Separations of both large and small molecules can be accomplished byCZE. Even small molecules, where the charge-to-mass ratio differencesmay not be great, may still be separable.

Capillaries. Capillaries with an internal diameter of 25-75 µm are usuallyemployed. Fused silica is the material of choice due to its UV transparency,durability (when polyimide coated), and zeta potential. Functionalized andgel-filled capillaries are becoming available and will be covered in othersections.

A new capillary must be conditioned before it can be used. Pretreat thecapillary for 10 min with 0.1 M sodium hydroxide, 5 min with water, and10 min with run buffer. This conditioning procedure is important to ensurethat the surface of the capillary is fully and uniformly charged. For somemethods it is necessary to regenerate this surface between runs with0.1 M sodium hydroxide, and in extreme cases, 1 M sodium hydroxide. Theregeneration procedure is frequently necessary if migration times change ona run-to-run basis. This is most common when using buffers in the pH 2-6

14

region. Regeneration is seldom necessary when working above pH 9 exceptas a daily startup procedure unless sample or matrix components areadhering to the capillary walls.

Not all attempts to store a fused silica capillary are successful. Thesmaller the capillary i.d., the more prone it is to clogging. While this is nottoo serious for bare silica, capillary damage can be costly when usingchemically modified capillaries. The following procedure will maximizeyour chances of successfully storing a capillary.

1) Rinse the capillary with 0.1 M NaOH for a few minutes.(Do not do this with chemically-modified capillaries.)

2) Rinse for 5 min with distilled water.

3) Place an empty, uncapped vial at the outlet end and blowN2through the capillary for 5 min.

4) Remove the capillary cartridge from the instrument.

Effect of pH. At a high pH where EOF is substantial, the order of migrationwill be cations, neutrals, and anions. None of the neutral molecules will beseparated since the net charge is zero. The anions will still migrate towardthe cathode because the EOF is greater than the electrophoretic migration.

At lower pH where the EOF is greatly reduced, both cations and anionscan still be measured, although not in a single run. To measure anions, theanode must be beyond the detector window. Likewise, to measure cations,the cathode must reside beyond the detector window. The proper electricalconfiguration is achieved by simply reversing the polarity of the electrodes.

The impact of pH on the analyte can also be substantial, particularlyfor complex zwitterionic compounds such as peptides. The charge on thesecompounds is pH-dependent, and the selectivity of separation is affectedsubstantially by pH. As a rule of thumb, select a pH that is at least two unitsabove or below the pKa of the analyte to ensure complete ionization. Athighly alkaline pH, the EOF may be so rapid that incomplete separationsmay occur.

Certain protein separations can be performed at acidic pH. Underthese conditions, the capillary wall is uncharged. The proteins under theseconditions will be positively charged and will not electrostatically interactwith the wall, although hydrophobic interaction may still occur. Thelimitation of this technique is protein precipitation. At the pI of the protein,band symmetry tends to be poor.

15

Buffers. A wide variety of buffers (Table 1) can be employed in CZE. Abuffer is most effective within one or two pH units of its pI. For example,phosphate is used around pH 2.5 and pH 7, and borate around pH 9. Thetypical buffer concentration is 50-100 mM.

Zwitterionic buffers such as bicine, tricine, CAPS, MES, and Tris arealso common, particularly for protein and peptide separations. The advan-tage of the zwitterionic buffer is its low conductivity when the buffer isemployed around the pI. The advantage therein is the low current draw andthus reduced Joule heating. In certain buffer preparations, particularly thosedirected at protein separations, salts such as chloride, phosphate, and sulfateare added to the buffer medium. These added salts affect the conformationof the protein, which in turn can impact the separation. The salt concentra-tion also impacts the EOF due in part to disruption of the charged doublelayer at the walls of the capillary. The major limitation on the amount of saltthat can be added to a buffer preparation is Joule heating.

Table 1. Buffers for Capillary Electrophoresis

Buffer Useful pH Range

Phosphate 1.14 - 3.14Acetate 3.76 - 5.76Phosphate 6.20 - 8.20Borate 8.14 - 10.14

Zwitterionic Buffers

MES 5.15 - 7.15PIPES 5.80 - 7.80HEPES 6.55 - 8.55Tricine 7.15 - 9.15Tris 7.30 - 9.30

Various buffer additives (Table 2) can be employed to change theselectivity of the separation. Buffer additives can alter, among other things,electrophoretic mobilities. In other words, two compounds that haveidentical mobilities in a simple buffer system may be differentiated with anadditive. Other additives, such as surfactants or cyclodextrins, form aheterogeneous environment that defines new classes of CE that will becovered later.

All buffers should be filtered through 0.45-µm filters prior to use.

16

Table 2. Buffer Additives for CZE

Additive Function

Inorganic salts Protein conformational changesOrganic solvents Solubilizer, modify electroosmotic flowUrea Solubilize proteins and denature

oligonucleotidesSulfonic acids Ion pairing agents, hydrophobic

interaction agentsCationic surfactants Charge reversal of capillary wallCellulose derivatives Reduce electroosmotic flow, provide a

sieving mediumAmines Cover free silanol groups

Capillary wall binding. One problem with CZE is electrostatic binding ofcationic substances to the walls of the tubing. This effect is observed withproteins when operating in a buffer that has a pH below the pKa of theanalyte. This problem can be overcome by operating at least two pH unitsabove the pKa of the protein, but in some cases, the higher pH may not beoptimal for the separation. In spite of these problems, proteins, especiallythose of similar size, can be best separated in free solution. The use oftreated capillaries is one of several ways that can serve to reduce wallbinding.

Another recently developed procedure involves the use of high-concentration phosphate buffer to screen the charge on the inner capillarywall and short (20 cm), small-bore (25 µm) capillaries operated at highelectric field strengths which allow for very rapid separations, therebyminimizing the residence time of the proteins in the capillary. This approachdoes require an excellent capillary temperature control system in order toremove the heat generated by the high voltages and currents.

Capillary coating. The reduction of or elimination of EOF can be useful toenable direct electrophoretic separations to be performed. More compellingis the ability to eliminate solute adsorption. A variety of coatings is possibleincluding some phases used for capillary gas chromatography. The use ofhydrophilic coatings can be useful in suppressing adsorption of hydrophobiccompounds. Electrostatic binding can also be suppressed. Hydrophobiccoatings, in conjunction with nonionic surfactants as buffer additives,appear promising as well. Many companies are beginning to introducebonded-phase capillaries. These capillaries should further extend the rangeof compounds applicable to separation by CZE.

17

Applications. CZE is very useful for the separation of proteins and peptidessince complete resolution can often be obtained for analytes differing byonly one amino acid substituent. This is particularly important in trypticmapping where mutations and post-translational modifications must bedetected. Figure 6 illustrates the separation of a tryptic digest of reduced,denatured and alkylated bovine serum albumin (BSA). The separation is runat low pH with 1.5 M urea as an additive. The urea induces peptide unfold-ing, thereby exposing the internal structural elements.

.030

.022

.015

.0075

.00

0 2 4 6 8 10 12 14 16 18 20

Time (min)

Abs

orba

nce

(214

nm

)

Figure 6. CZE Separation of a Tryptic Digest of Reduced, Denatured andAlkylated Bovine Serum Albumin. Buffer, 21 mM sodium phosphate(monobasic), 1.5 M urea, pH 2.5; capillary, 59 cm. Courtesy of R. Rush,A. Cohen, and B. Karger, Northeastern University.

Other applications where CZE may be useful include inorganic anionsand cations such as those typically separated by ion chromatography. Smallmolecules such as pharmaceuticals can often be separated provided they arecharged. In most cases, the technique of micellar electrokinetic capillarychromatography gives superior results for charged as well as neutral smallmolecules. This mode of CE will be covered later.

A summary of applications including the buffer recipe and reference isgiven in Table 3.

18

Table 3. Applications of CZE

Analyte Buffer Reference

Dipeptides/proteins 150 mM H3PO4, pH 1.5 Anal. Chem. 60, 2322 (1988)Angiotensin II octapeptides 150 mM NaH2PO4, pH 3.0 Anal. Chem. 60, 2322 (1988)Tryptic digest 12.5 mM phosphate, pH 6.86 J. Chromatogr. 352, 337

(1986)ß-Lactoglobulin 50 mM borate, 50 mM KCl J. Chromatogr. 447, 117

(1988)Carbonic anhydraseWhale skeletal myoglobin

Myoglobins 10 mM tricine, 40 mM KCl, pH 8.24Anal. Chem. 58, 743a (1986)Carbonic anhydrasesß-LactoglobulinsHorse heart myoglobin 10 mM borate, 40 mM KCl, pH 9.5J. Chromatogr. 480, 157

(1989)Endorphins 20 mM citrate, pH 2.5 Anal. Chem. 61, 1186 (1989)Ribonucleases 20 mM CAPS, pH 11.0 Anal. Chem. 61, 1186 (1989)Immune complexes 0.1 M tricine, pH 8.0 Anal. Chem. 61, 1186 (1989)Tobacco mosaic virus 2 mM potassium borate Anal. Chem. 62, 1592 (1990)Adenosine-5'[α-32Ρ]triphosphate200 mM borate, pH 8.1 J. Chromatogr. 480, 259

(1989)Lysozyme, α-chymotrypsin 0.1 M CHES, 0.25 M K2SO4, pH 9 J. Chromatogr. 480, 301

(1989)Nucleoside phosphates 40 mM glutamic acid/GABA J. Chromatogr. 480, 321

(1989)Collagens 2.5 mM sodium tetraborate J. Chromatogr. 480, 371

(1989)Membrane proteins 0.2 M borate, 7 M urea, pH 9.2J. Chromatogr. 516, 89 (1990)

Human growth hormone 100 mM phosphate, pH 2.6, pH 8J. Chromatogr. 480, 379(1989)

Anions 25 mM salicylate, pH 4 J. Chromatogr. 480, 169(1989)

Isoelectric focusing

The fundamental premise of isoelectric focusing (IEF) is that a moleculewill migrate so long as it is charged. Should it become neutral, it will stopmigrating in the electric field. IEF is run in a pH gradient where the pH islow at the anode and high at the cathode (Figure 7). The pH gradient isgenerated with a series of zwitterionic chemicals known as carrierampholytes. When a voltage is applied, the ampholyte mixture separates inthe capillary. Ampholytes that are positively charged will migrate towardsthe cathode while those negatively charged migrate towards the anode. The

19

pH then will decrease at the anodic section and increase at the cathodicsection. Finally, the ampholyte migration will cease when each ampholytereaches its isoelectric point and is no longer charged. Initially, a solute witha net negative charge will migrate towards the anode where it encountersbuffer of decreasing pH. Finally, the solute encounters a pH where its netcharge becomes zero, the isoelectric point (pI), and migration halts. Thegreater the number of ampholytes in solution, the smoother the pH gradient.

pI

High pHLow pH

+ –+

–

Figure 7. Isoelectric Focusing

The pH of the anodic buffer must be lower than the pI of the mostacidic ampholyte to prevent migration into the analyte. Likewise, thecatholyte must have a higher pH than the most basic ampholyte.

It is apparent that the EOF and other convective forces must be sup-pressed if IEF is to be effective. The capillary walls can be coated withmethylcellulose or polyacrylamide to suppress EOF. The coating tends tosuppress protein adsorption as well. IEF is generally used for high resolu-tion separations of proteins and polypeptides but could be used for anyamphoteric substance, provided a series of ampholytes that cover the entirepI range is used.

Resolving power. The resolving power, ∆pI, of IEF is described by theequation

=∆ 3D(dpH/dx)pIE(dµ/dpH)

(14)

where D is the diffusion coefficient, E is the electric field strength, and µ isthe electrophoretic mobility of the protein. A resolving power of 0.02 pHunits has been calculated.

The three basic steps of IEF are loading, focusing, and mobilization.In traditional slab-gel techniques, the mobilization technique is unnecessary.

20

Once focusing is complete, the gel is stained using traditional methods. Incapillary IEF, the bands must migrate past the detector, so the mobilizationstep becomes necessary. In contrast to CZE, the buffer medium is discon-tinuous, i.e., a pH gradient is formed along the capillary. Commercialampholytes are available from several suppliers covering many pH ranges.Broad-range buffers are used to estimate the pI. Then a narrower pH rangecan be employed to improve precision. A series of calibrants of known pI’sare employed to correlate the migration time with the isoelectric point.

Loading. The sample is mixed with the appropriate ampholytes (availablefrom Bio-Rad or Pharmacia) to a final concentration of 1-2% ampholytes.The mixture is loaded into the capillary either by pressure or vacuumaspiration.

Focusing. The buffer reservoirs are filled with sodium hydroxide (cathode)and phosphoric acid (anode). Field strengths on the order of 500-700 V/cmare employed. As the focusing proceeds, the current drops to less than 1 µA.Overfocusing can result in precipitation due to protein aggregation at highlocalized concentrations. Dispersants such as nonionic surfactants (TritonX-100 or Brij-35), or organic modifiers such as glycerol or ethylene glycolmay minimize aggregation. These agents are mild and usually do notdenature the protein. Urea could also be used, but the protein will becomedenatured. Because of precipitation problems, very hydrophobic proteinsare not usually separated by IEF. Gel-filled capillaries are sometimes usefulfor separating troublesome proteins.

Mobilization. Mobilization can be accomplished in either the cathodic oranodic direction. For cathodic mobilization, the cathode reservoir is filledwith sodium hydroxide/sodium chloride solution. In anodic mobilization,the sodium chloride is added to the anode reservoir. The addition of saltalters the pH in the capillary when the voltage is applied since the anions/cations compete with hydroxyl/hydronium ion migration. As the pH ischanged, both ampholytes and proteins are mobilized in the direction of thereservoir with added salt. As mobilization proceeds, the current rises as thesaline ions migrate into the capillary. Detection is performed at 280 nm forproteins since the ampholytes absorb strongly in the low UV range.

Applications. In addition to performing high resolution separations, IEFis useful for determining the pI of a protein. IEF is particularly useful forseparating immunoglobulins, hemoglobin variants and post-translationalmodifications of recombinant proteins. A separation of a protein mixtureis shown in Figure 8.

21

0.05

0.045

0.04

0.035

0.03

0.025

0.02

0.015

0.01

0.005

0

0 10 20 30 40

1

2

3

4

5

Time (min)

Abs

orba

nce

(280

nm

)

Figure 8. Separation of a Protein Mixture by IEF. Ampholyte pH range,3.5-10; catholyte, 50 mM sodium hydroxide; analyte, 150 mM phosphoricacid; voltage, 25 kV; mobilization (catholyte), 50 mM sodium chloride,50 mM sodium hydroxide. Courtesy of R. Nelson and B. Karger,Northeastern University.

Capillary gel electrophoresis

Traditional gel electrophoresis is conducted in an anticonvective mediumsuch as polyacrylamide or agarose. The composition of the media can alsoserve as a molecular sieve to perform size separations (Figure 9). Further-more, the gel suppresses the EOF. Because of the long history of thistechnique, the adaptation to CE is very desirable. This is particularlyvaluable for DNA separations since no other technique to date has providedsuch dramatic separations. Commercial capillary gel columns are nowbeginning to be introduced to the marketplace from numerous sources.Polyacrylamide gel-filled capillaries are usually employed, although newpolymer formulations with greater stability to the applied electric field arelikely to be introduced shortly. Agarose gels are unable to withstand theheating produced by the high voltages used in capillary gel electrophoresis(CGE).

22

��

Figure 9. Capillary Gel Electrophoresis

There are two fundamental classes of gels that can be employed inCGE. These are illustrated in Figure 10. The physical gel obtains its porousstructure by entanglement of polymers and is quite rugged to changes in theenvironment. Hydroxypropylmethylcellulose and similar polymers can beused to form physical gels. Chemical gels use covalent attachment to formthe porous structure. These gels are less rugged, and it is difficult to changethe running buffers once the gels are formed. In CZE and other forms of“open-tubular” CE, the capillary is filled with buffer by pressurization. Forgel-filled capillaries, this technique would result in extrusion of the gel.Urea and other buffer agents such as Tris-borate-EDTA are added prior topolymerization. Cross-linked polyacrylamide is usually selected as the gel-forming agent.

Physical Gel Chemical Gel

Figure 10. Physical and Chemical Gels

CGE is typically performed in 50- to 100-µm capillaries in lengths ofabout 10 cm to 1 m. Better resolution is found for the longer capillaries, butthe run times are excessive. The capillary gel composition is better manipu-lated to optimize the resolution. For example, increasing the gel concentra-tion improves resolution but decreases the molecular weight range acces-sible within the run. The voltage is somewhat limiting since field strengthsabove 500 V/cm may cause capillary heating and, ultimately, voids.

Proteins. Proteins denatured with 2-mercaptoethanol are usually run withan SDS-PAGE system. Under these conditions, all proteins have the samecharge-to-mass ratio since the native charge is obscured by SDS binding.

23

Indeed, a constant amount of SDS, 1.4 g, is adsorbed onto each gram ofprotein. SDS is anionic; therefore, all proteins become negatively chargedand migrate towards the anode. Proteins unfold (provided disulfide linkagesare broken) and become rod-like in structure allowing uniform molecularsieving for size separation. A typical buffer is 90 mM Tris-phosphate,pH 8.6, 0.1% SDS. A calibration plot of mobility versus molecular weightpermits size assignments of the various fragments. Electrokinetic injectionsare used in CGE since pressure- mediated injection would result in extru-sion of the gel. Short 10-20 cm capillaries are employed with field strengthsapproaching 400 V/cm. Under these conditions, there is a linear correlationbetween mobility and molecular weight. The log of the mobility versus thepercent monomer composition (%T) of the gel is also linear. Proteins areusually denatured in 1% SDS and 2% β-mercaptoethanol for 30 min at90°C.

As in conventional slab-gel electrophoresis, the migration timesdecrease as the pore size increases (lower %T). Separation times can also bereduced by using higher voltages, provided heat dissipation is adequate.Short capillaries can also be used at the same field strength to further speedthe separation.

DNA. Separation of oligonucleotides and DNA sequence products havebeen accomplished in polyacrylamide gels. For restriction fragments andlarger oligos, gels with little or no crosslinker seem most effective due tothe larger pore size of the gel. Separation of deoxyoligonucleotides such aspoly(dA)40-60 is readily accomplished in an 8% T gel with a buffer consist-ing of 100 mM Tris-borate, pH 8.3 with 2 mM EDTA and 7 M urea, inunder 35 min with unit base resolution. Determining the purity of syntheticoligos is an important application of CGE. Figure 11 shows the separationof a synthetic 50-mer homopolymer of thymidine with modulo 5 enhance-ment. The failure sequences are well separated with unit base resolution.

24

Abs

orba

nce

1520

25

30 35 40 45

0 5 10 15 20 25 30

Time (min)

Capillary, 27 cm x 75 µm25 mM Tris, 25 mM boric acid, 7 M ureaPAG 7.5% T, 3.3% C10-18 kV 25°C

Figure 11. CGE of Thymidine Synthetic Homopolymer with Modulo 5Enhancement. Buffer, 25 mM Tris, 25 mM boric acid, 7 M urea; gel,polyacrylamide, 7.5% T, 3.3% C. Courtesy of A. Paulus and J. Ohms,Beckman Instruments, Inc.

Double-stranded DNA can be separated with physical gels. Surfacemodified capillaries are best employed since the electroosmotic flow istotally suppressed. Under these conditions, the fragment migration time isdirectly related to the number of bases present. Figure 12 shows theseparation of a Hae III restriction digest of øX174 DNA. These separationsare very efficient as the peak representing 118 base pairs yields about2,000,000 theoretical plates per meter. The use of ethidium bromide as abuffer additive enhanced the separation, permitting good resolution of thepeaks representing 261 and 271 base pairs. The resolution of this system is3 base pairs. Among the other useful applications for this system areseparations of PCR-amplified DNA as well as restriction digests.

25

0.018

0.016

0.014

0.012

0.010

0.008

0.006

0.004

0.002

0

-0.002

16 19.6 23.2 26.8 30.4 34

Time (min)

Abs

orba

nce

(260

nm

)

Time (min)18.1 18.2

72118

194234

271281

310

603

872

1078

1353

Figure 12. Separation of Hae III RestrictionDigest of øX174 DNA by CGE with a PhysicalGel. Buffer, 89 mM Tris-borate, 2 mM EDTA,pH 8.5, 0.5% hydroxypropylmethylcellulose,10 µM ethidium bromide; capillary, DB-17coated, 100 µm × 50 cm (length to detector);field strength, 175 V/cm; detection, 260 nm;sample concentration, 10 µg/mL.

Capillary gel electrophoresis is still in its infancy. While most academiclabs are manufacturing their own capillaries either by covalent bonding ofpolyacrylamide to the capillary walls or packing the capillaries with thegels, it is expected that most users will purchase capillaries from commer-cial vendors. Likewise, optimized buffer solutions for separating DNA inthe physical gel format are also being provided by manufacturers. One of

26

the many driving forces for the development of gel-filled capillaries is DNAsequencing. The hope is that a multiplex CE instrument with gels will formthe basis of a genomic scale instrument.

Isotachophoresis

Prior to 1981, isotachophoresis (ITP) was the most widely used instrumen-tal capillary electrophoretic technique, although the capillaries were quitewide (250-500 µm) by today’s standards. A commercial instrument, theLKB Tachophor, was introduced in the mid-1970s.

Like IEF, ITP relies on zero electroosmotic flow, and the buffer systemis heterogeneous. The capillary is filled with a leading electrolyte that has ahigher mobility than any of the sample components to be determined. Thenthe sample is injected. A terminating electrolyte occupies the oppositereservoir, and the ionic mobility of that electrolyte is lower than any of thesample components. Separation will occur in the gap between the leadingand terminating electrolytes based on the individual mobilities of theanalytes. As shown in Figure 13, stable zone boundaries form betweenindividual components resulting in highly efficient separations. Both anionsand cations could be determined, though not in the same run.

+_

+_

Sample

LeadingElectrolyte

TerminatingElectrolyte

Direction of Electrophoretic Flow

Figure 13. Anionic Isotachophoresis

In the early instrumentation, detection was by conductivity or differen-tial conductivity. Conductivity detection gave a stair-step pattern as eachindividual ion passed the electrodes. Differential conductivity could restorethe isotachopherogram to a series of conventional peaks. Direct UV detec-

27

tion also gives a more familiar looking electropherogram in the presence ofspacers. A spacer is a nonabsorbing solute with a mobility value that fallsin between the mobilities of two peaks that need to be resolved. The disad-vantage of ITP is that unless spacers are employed, adjacent bands are incontact with each other. A second problem is that, compared to CZE, theselection and optimization of the buffer are less straight forward. For ex-ample, to determine cations, the leading cathodic electrolyte might containhighly mobile acid (H+) while the terminating anodic electrolyte mightcontain a weaker acid such as propionic acid. Most of these points areillustrated in Figure 14 for ITP of some anions. The lower plot is the directconductivity (plotted as resistance) tracing while the upper plot representsdifferential conductivity.

60

40

R(mV)

T

1211

109

87

65

43 2

1 L

10

30 sec

t

Figure 14. ITP of a Mixture of Anions withConductivity Detection. Capillary, 105 µm i.d.fluorinated ethylene-propylene copolymer;leader, 10 mM hydrochloric acid titrated to pH6.0 with histidine, 2 mM hydroxyethylcellulose;terminator, 5 mM MES; driving current, 10 µA.Key: R, signal on conductivity meter (increasingresistance); L, chloride; 1, sulfate; 2, chlorate;3, chromate; 4, malonate; 5, adipate; 6, benzo-ate; 7, impurity; 8, acetate; 9, ß-bromopropion-ate; 10, naphthalene-2-sulfonate; 11, glutamate;12, enantate; T, MES. Reproduced in part fromJ. Chromatogr., 267, 67 (1983) (Fig. 3).

28

Isotachophoresis has two characteristics, the combination of which isunique to electrophoretic methods: all bands move at the same velocity, andthe bands are focused. For example, highly mobile bands have high conduc-tivity, and as a result, have a lower voltage drop across the band. Since themobility is the product of the conductivity and the voltage drop, andconductivity and voltage drop are inversely proportional, the individualband velocities are self-normalizing. Focusing is also a consequence ofvelocity normalization. For example, if a band diffuses into a neighboringzone, it will either speed up or slow down based on the field strengthencountered and rejoin the original band.

Operating on modern instrumentation has been successful, both withtreated and untreated fused silica capillaries. EOF can be suppressed with0.25% hydroxypropylmethylcellulose. A good leading electrolyte is 5 mMphosphoric acid. Valine (100 mM, adjusted to an appropriate pH with aprimary amine) is a useful terminating electrolyte. At the start of a separa-tion, the current may be quite high since the highly mobile electrolytecompletely fills the capillary. As the separation progresses, the currentalways declines as the less mobile terminator enters the capillary.

There have been some recent publications that employ ITP as a pre-concentration step for CZE. Because ITP is a focusing technique, the use oflarge-diameter capillaries does not cause resolution to deteriorate the way itdoes for CZE.

Micellar electrokinetic capillary chromatography

Perhaps the most intriguing mode of CE for the determination of smallmolecules is MECC. The use of micelle-forming surfactant solutions cangive rise to separations that resemble reverse-phase LC with the benefits ofCE. Unlike IEF or ITP, MECC relies on a robust and controllable EOF.

Micelles. Micelles are amphiphilic aggregates of molecules known assurfactants. They are long chain molecules (10-50 carbon units) and arecharacterized as possessing a long hydrophobic tail and a hydrophilic headgroup. Normal micelles are formed in aqueous solution with the hydropho-bic tails pointing inward and the hydrophilic heads pointing outward intothe aqueous solution. A schematic is shown in Figure 15. Micelles form as aconsequence of the hydrophobic effect, that is, they form to reduce the freeenergy of the system. The hydrophobic tail of the surfactant cannot besolvated in aqueous solution. Above a surfactant concentration known asthe critical micelle concentration (CMC), the aggregate is fully formed.Physical changes such as surface tension, viscosity, and the ability to scatter

29

light accompany micelle formation. Reverse micelles, which form in or-ganic solvents, have not been studied in MECC.

veovep

O

OO

O

OO

O

OO

O

OO

O

OO+ –

-

-

--

-

-

--

Figure 15. Micellar Electrokinetic Capillary Chroma-tography

There are four major classes of surfactants: anionic, cationic, zwitteri-onic, and nonionic, examples of which are given in Table 4. Of these four,the first two are most useful in MECC. Both synthetic and naturallyoccurring compounds have been employed for MECC. Synthetic varietiesinclude anionic SDS and cationic cetyltrimethylammonium bromide(CTAB). Naturally occurring compounds such as bile salts (sodiumtaurocholate, etc.) are also useful.

Table 4. Surfactant Classes and Properties

Surfactant Type CMC Aggregation #

SDS Anionic 8.1 × 10-3 62CTAB Cationic 9.2 × 10-4 170Brij-35 Nonionic 1.0 × 10-4 40Sulfobetaine Zwitterionic 3.3 × 10-3 55

SDS = sodium dodecyl sulfate; CTAB = cetyltrimethylammonium bromide;Brij-35 = polyoxyethylene-23-lauryl ether; sulfobetaine = N-dodecyl-N,N-dimethylammonium-3-propane-1-sulfonic acid.

30

SDS is the most widely used surfactant in MECC. It is available inhighly purified forms and is inexpensive. Its molecular weight is 288 andthe CMC is 8 mM. The aggregation number (number of molecules/micelle)is 62.

Micelles have the ability to organize analytes at the molecular levelbased on hydrophobic and electrostatic interactions. Even neutral moleculescan bind to micelles since the hydrophobic core has very strong solubilizingpower. Surfactant solutions have been employed in spectroscopy andchromatography to take advantage of these unique micellar properties.For example, room temperature phosphorescence is readily observable inmicellar media since the micellar environment prevents many of the normalquenching mechanisms from operating. More significantly with regard toMECC, these same surfactant solutions can serve as chromatographicmobile phase modifiers. Micellar chromatography can mimic reverse-phaseLC in that increasing the surfactant concentration increases the elutingpower of the mobile phase. The analyte can partition between the micelleand the bulk phase, the micelle and the stationary phase, or the bulk andstationary phase. Thus, “pseudophase” or micellar LC has more complexequilibria than conventional LC. This three-phase equilibrium can belikened to ion-pair chromatography in many instances. In certain aspects,the mechanism of MECC, because of only a two-phase equilibrium, issimpler than micellar chromatography. The complicating factor in MECC isthat analyte electrophoretic mobility often contributes to the overallseparation.

Separation mechanism. At neutral to alkaline pH, a strong EOF moves inthe direction of the cathode. If SDS is employed as the surfactant, theelectrophoretic migration of the anionic micelle is in the direction of theanode (Figure 15). As a result, the overall micellar migration velocity isslowed compared to the bulk flow of solvent.

Since analytes can partition into and out of the micelle, the require-ments for a separation process are at hand. When an analyte is associatedwith a micelle, its overall migration velocity is slowed. When an unchargedanalyte resides in the bulk phase, its migration velocity is that of the EOF.Therefore, analytes that have greater affinity for the micelle have slowermigration velocities compared to analytes that spend most of their time inthe bulk phase.

31

On examination of the fundamental equations governing MECC (Eqs.15-16), note that both k' (capacity ratio) and Rs (resolution) at infinitemicelle concentration simplify to the general LC equations. MECC is achromatographic or pseudo-chromatographic process.

=t

k'- t

mcRto (1 - t /t )oR

(15)

=k'R mc- o1 - t /t 2

sN4 ( )α 1

α ( k'21+ )(1+ (t /t k'mco 1) (16)

Both cationic and anionic surfactants can be employed in MECC.When using a cationic surfactant, the EOF is reversed; therefore, theelectrode polarity must also be reversed to detect the analyte.

Migration order. With SDS micelles, the general migration order will beanions, neutrals, and cations. Anions spend more of their time in the bulkphase due to electrostatic repulsions from the micelle. The greater theanionic charge, the more rapid the elution. Neutral molecules are separatedexclusively based on hydrophobicity. Cations elute last due to strongelectrostatic attraction (e.g., ion-pairing with the micelle). While this is auseful generalization, strong hydrophobic interaction can overcomeelectrostatic repulsions and attractions. Likewise, the electrophoreticmigration of the analytes can also affect the elution order.

The separation of neutral species by CE is a compelling example of thegeneral applicability of the technique. Figure 16 shows the separation ofsome corticosteroids by MECC. Rather than use SDS, the surfactant mostfrequently employed, this separation uses sodium cholate, a naturallyoccurring bile salt, as the micelle-forming surfactant. Since only hydropho-bic mechanisms can influence the order of migration, the less polar steroidesters are expected to show longer migration times. This is proven in Figure16 with hydrocortisone (peak 2) and hydrocortisone acetate (peak 4).

32

0.02

0.01

0 4 8 12 16 20

Time (min)

Abs

orba

nce

1

2

3

4

5

6

7

8

Figure 16. Separation of Corticosteroids by MECC.Buffer, 100 mM sodium cholate, 100 mM borate,pH 8.45. 1, triamcinolone; 2, hydrocortisone;3, betamethasone; 4, hydrocortisone acetate; 5, dexam-ethasone acetate; 6, triamcinolone acetonide;7, fluocinolone acetanide; 8, fluocinolone.

Use of organic modifiers. While organic modifiers have been used in free-solution separations to overcome solubility problems, their use in MECC ismuch more profound. Because the organic modifier reduces EOF, the over-all peak capacity of the separation is increased due to the greater gap be-tween to and tmc. A more important role of the modifier is the impact on thepartition coefficient of a solute between the micelle and the bulk solution.Clearly, the modifier makes the bulk solution more hospitable for hydro-phobic analytes. Without the modifier, hydrophobic solutes will elute at ornear tmc. The addition of the modifier generally increases migration velocityof hydrophobic species since they now spend more of their time in the bulkphase.

33

Many organic modifiers are useful in MECC. Methanol and acetonitrileare most commonly employed at concentrations from 5-25%. Methanol andother alcohols tend to slow EOF more so than acetonitrile. For certainseparations, aprotic solvents such as tetrahydrofuran (THF), dimethylsulfoxide (DMSO) and dimethylformamide (DMF) might be useful. DMF isa good solvent because its boiling point is high, thereby minimizingoutgassing, a common problem if elevated temperatures are employed. Thepercentage of organic modifier that is effective is limited by the effect of thesolvent on the aggregation number and the micellar ionization number.

Chiral Recognition. Chiral recognition is dependent on the formation ofdiastereomers either through covalent or electrostatic interactions. There areseveral approaches for performing chiral separations by CE. Additives suchas optically active bile salts and cyclodextrins permit chiral resolution bystereoselective interaction with the solute. With cyclodextrins, this interac-tion occurs within the molecular cavity by formation of an inclusioncomplex. The L-complexes tend to be more stable and have longer migra-tion times. With bile salts, the interaction probably occurs at the surface ofthe micelle. The mechanism of separation is similar, if not identical, toconventional MECC. When an analyte is complexed with the micellar orcyclodextrin additive, its migration velocity is slowed relative to the bulkphase. The enantiomer that forms the more stable complex will alwaysshow a longer migration time because of this effect. The main disadvantageof this approach towards chiral recognition is that it is difficult to predictwhich analytes will optically resolve with a particular additive.

Another approach to chiral selectivity is precapillary derivatization.The analyte is derivatized with an optically active reagent to form cova-lently bound diastereomers. The diastereomers are usually easily separatedby MECC. There are several advantages and disadvantages with thisapproach. The advantages include: enhanced sensitivity since the tag canbe a good chromophore or fluorophore, and predictable results, particularlywhen the analyte’s chiral center and reactive site are relatively close to eachother. The major disadvantage is that complex assay validation as a checkfor completeness of derivatization, derivative stability, and freedom fromracemization must be performed. The separation of chiral amino acidsderivatized with Marfey’s reagent (1-fluoro-2,4-dinitrophenyl-5-L-alanine)is shown in Figure 17.

34

6.95

8.65

10.34

12.04

13.74

15.43

17.13

18.83

20.52

22.22

23.92

25.61

27.31

-0.0020 0.0040 0.0100 0.0160 0.0220 0.0280

7.38 Unknown7.17 L-Ala

7.77 L-Val

8.55 D-Ala

9.75 L-Leu11.07 L-Phe

12.06 L-Trp

12.40 D-Val

18.00 D-Leu

19.75 Unknown

21.10 D-Phe

22.55 D-Trp

Absorbance (340 nm)

Tim

e (m

in)

Figure 17. Separation of Chiral Amino Acids Derivatizedwith Marfey’s Reagent. Buffer, 100 mM sodium borate,200 mM SDS, 5% acetonitrile, pH 8.5.

Applications. A broad base of small-molecule applications has alreadyappeared in the literature. Table 5 contains a summary of some of theseapplications along with buffer recipes.

35

Tab

le 5

. App

licat

ions

of M

EC

C

Ana

lyte

Buf

fer

Ref

eren

ce

Mod

ified

nuc

leic

aci

ds75

mM

SD

S, 1

0 m

M b

orat

e, 1

0 m

M P

O4,

pH

8.4

J. C

hrom

atog

r. 4

8, 193

(19

87)

Ang

iote

nsin

s10

mM

Tris

-PO 4, 5

0 m

M D

TA

B, p

H 7

.05

J. C

hrom

atog

r. 5

19, 1

89(1

990)

Pen

icill

ins

20 m

M P

O 4, 1

00 m

M S

DS

, pH

8.5

J. C

hrom

atog

r. 5

15, 2

45(1

990)

OP

A-a

min

o ac

ids

50 m

M b

orat

e, 1

5% M

eOH

, 2%

TH

F, 5

0 m

M S

DS

, pH

9.5

J. C

hrom

atog

r. 4

86, 5

5 (1

988)

Urin

ary

porp

hyrin

s85

mM

SD

S, 1

7 m

M C

AP

S, 1

5% M

eOH

, pH

11

J. C

hrom

atog

r. 5

16, 271

(199

0)A

spiri

n/ca

ffein

e50

mM

SD

S, 2

0 P

O 4, p

H 1

1A

nal.

Che

m. 5

9, 277

3 (1

987)

Col

d pr

ep (

14 d

rugs

)50

mM

sod

ium

deo

xych

olat

e or

100

mM

sod

ium

taur

ocho

late

, 20

mM

PO

4, p

H 9

J. C

hrom

atog

r. 4

98, 3

13(1

990)

Wat

er-s

olub

le v

itam

ins

50 m

M S

DS

, 20

mM

PO

4-bo

rate

, pH

9J.

Chr

omat

ogr.

465

, 331

(198

9)W

ater

-sol

uble

vita

min

s50

mM

SD

S, 2

0 m

M P

O4,

pH

9J.

Chr

omat

ogr.

447

, 133

(198

8)ß

-Lac

tam

ant

ibio

tics

150

mM

SD

S, 2

0 m

M P

O4-

bora

te, p

H 9

J. C

hrom

atog

r. 4

77, 259

(198

9)P

heno

ls50

mM

SD

S, 2

5 m

M b

orat

e, 5

0 m

M P

O4,

pH

7A

nal.

Che

m. 5

6, 111

(19

84)

DN

S-m

ethy

lam

ine/

DN

S-m

ethy

l-d 3-a

min

e25

mM

SD

S, 2

5% M

eOH

, 25

mM

PO

4, 6

2.5

mM

bor

ate,

4.2

8 m

M N

aHC

O3,

pH

8A

nal.

Che

m. 6

1, 491

(19

89)

Chi

ral d

rugs

50 m

M ta

urod

eoxy

chol

ate,

20

mM

bor

ate-

PO

4, p

H 7

J. M

icro

col.

Sep

. 1, 234

(19

89)

Cat

echo

lam

ines

10 m

M S

DS

, 10

mM

PO

4, 6

mM

bor

ate,

pH

7J.

Chr

omat

ogr.

411,

299

(198

8)C

hlor

inat

ed p

heno

ls70

mM

SD

S, 5

0 m

M P

O4,

25

mM

bor

ate

J. C

hrom

atog

r. 3

48, 39

(198

5)S

mal

l olig

onuc

leot

ides

50 m

M S

DS

, 5 m

M T

ris-b

orat

e, 3

mM

Zn(

II) 2

M u

rea,

pH

7C

hrom

atog

raph

ia 3

0, 15

(199

0)S

ulfo

nam

ides

50 m

M S

DS

, 20

mM

PO

4, p

H 7

Chr

omat

ogra

phia

30, 7

(19

90)

36

Selecting the mode of capillary electrophoresis

Table 6 can be used to help select the most advantageous mode of electro-phoresis as a starting point in methods development. The uppermost listingin each category of the chart is likely to yield acceptable results in theshortest time frame.

Table 6. Selecting the Mode of Capillary Electrophoresis

Small SmallPeptides Proteins Oligonucleotides DNA

Ions Molecules

CZE MECC CZE CZE CGE CGEITP CZE MECC CGE MECC

ITP IEF IEFCGE ITPITP

37

Approaches to methods developmentby CZE and MECC

Before you attempt a new separation, some information gathering will bevery useful. Is the compound soluble in water at concentrations up to1 mg/mL? Is it soluble at all pH’s? If aqueous solubility is a problem, willsmall amounts (up to 25%) of methanol or acetonitrile solubilize it? Willsmall molecules solubilize using 100 mM SDS? For proteins, will 7 M ureaor a dispersant such as ethylene glycol help? Is the analyte unstable atcertain pH’s? Is the compound thermally labile? What is the wavelength ofmaximum UV absorption? How many components are expected in themixture? What is the concentration expected of each component?

Developing a method by CZE

A great number of options and tools for methods development are availablefor CZE. In this exercise, the processes for separating a new protein will bereviewed. For starting conditions, use a 75-cm capillary run at 25°C at20 kV with the detector set at 214 nm. Make a 1-s injection of a 1 mg/mLsolution. Use a 100 mM buffer at the appropriate pH.

1) Acid stable—use a buffer pH below the pI;Acid labile—select a pH at least 1 unit above the pI.

2) Solubility problem—add a modifier such as urea or ethyleneglycol.

3) Adsorption problem—use an additive such as a sulfonic acid,a salt, or switch to a treated capillary.

4) Good efficiency, poor separation—adjust the pH.

5) Poor efficiency—increase ionic strength of buffer, add a saltin which the protein is stable.

In most cases, you will be able to get a good separation in a relativelyshort time frame. Some samples may be quite difficult and you may have tospend considerable time carefully selecting buffers and buffer additives.

38

Developing a method by MECC

MECC is a good separation mechanism for small molecules. The uppermolecular weight has not yet been established. Proteins are not wellseparated by this technique.

Good starting conditions are 100 mM SDS in pH 7, 50 mM phosphate-borate buffer, after which adjustments in SDS concentration, pH, andorganic modifier may be necessary. Some guidelines are:

1) Long separation times, good resolution—increase pH,decrease SDS.

2) Long separation times, poor resolution—use organic modifier.

3) Short separation times, poor resolution—increase SDS.

4) Short separation times, moderate resolution—decrease pH,increase SDS.

The use of the organic modifier is especially powerful in MECC.Acetonitrile is the solvent of first choice since it has little impact on theEOF. Alcohols may also be useful, but the separation times can becomelengthy. Under the proper conditions, the resolving power and peak capacityfar exceed HPLC. It takes no more than a few days to develop mostseparations.

Automated methods development

The P/ACE™ Systems 2050 and 2100 have the capability of performingmultiple separations with a variety of buffer solutions. This feature offersthe possibility of automated, unattended methods development since freshbuffers in both the anode and cathode reservoirs can be used for each run.In this fashion, parameters such as pH, buffer concentration, additive type,and surfactant concentration, among other factors, can be optimized in alogical and systematic manner.

There are a few guidelines that will prove useful to ensure the efficientgeneration of applications information:

1) Perform a few preliminary separations to gain a generalunderstanding of the problem.

2) Set the run time for the maximum expected separation time.For example, with MECC, this would correspond to thehighest surfactant concentration employed.

39

3) Some buffers such as phosphate permanently alter the wallchemistry; use these buffers last.

4) When switching from CZE to MECC, allow sufficient time(at least 0.5 h) for the capillary to equilibrate with the surfac-tant solution.

5) Program a 0.1 M sodium hydroxide wash between each run.

6) When determining the optimum temperature, allow sufficienttime for thermal equilibration.

7) Modify only one experimental variable at a time.

40

Suggested reading

At the present time, no textbooks are available on capillary electrophoresis.The following bibliography summarizes a number of articles that are usefulfor developing a basic understanding of modern CE. For starters, readJournal of Chromatography volumes 480 and 516. These are the conferenceproceedings of HPCE ’89 and ’90, respectively. A tremendous amount ofinformation is available in these volumes and the soon-to-be-publishedproceedings from HPCE ’91. Other useful papers are listed below in reversechronological order.

Benefits of automation in the separation of biomolecules by high perfor-mance capillary electrophoresis, McLaughlin, G.; Palmieri, R.; Anderson,K., Techniques in Protein Chemistry II, Academic Press, 3-19 (1991)

Indirect detection methods for capillary separations, Yeung, E. S.; Kuhr,W. G., Anal. Chem. 63, 275A-282A (1991)

Fluorescence detection in capillary electrophoresis: evaluation ofderivatizing reagents and techniques, Albin, M.; Weinberger, R.; Sapp, E.;Moring, S., Anal. Chem. 63, 417-422 (1991)

Production of polyacrylamide gel filled capillaries for capillary gel electro-phoresis (CGE): Influence of capillary surface pretreatment on performanceand stability, Yin, H. F.; Juergen, A.; Schomburg, G., J. High Resolut.Chromatogr. 13, 624-627 (1990)

Influence of buffer concentration, capillary internal diameter and forcedconvection on resolution in capillary zone electrophoresis, Rasmussen,H. T.; McNair, H. M., J. Chromatogr. 516, 223-231 (1990)

Isotachophoresis in open-tubular fused-silica capillaries: impact of elec-troosmosis on zone formation and displacement, Thormann, W., J.Chromatogr. 516, 211-217 (1990)

High speed DNA sequencing by capillary electrophoresis, Luckey, J. A.;Drossman, H.; Kostichka, A.J.; Mead, D. A.; D’Cunha, J.; Norris, T. B.;Smith, L. M., Nucleic Acids Res. 18, 4417-4421 (1990)

Method optimization in capillary zone electrophoretic analysis of hGHtryptic digest fragments, Nielsen, R. G.; Rickard, E. C., J. Chromatogr. 516,99-114 (1990)

41

Construction, evaluation and analytical operation of a modular capillaryelectrophoresis instrument, Lux, J. A.; Yin, H. F.; Schomburg, G.,Chromatographia 30, 7-15 (1990)

Atmospheric pressure ionization mass spectrometry. Detection for theseparation sciences, Huang, E. C.; Wachs, T.; Conboy, J. J.; Henion, J. D.,Anal. Chem. 62, 713A-722A, 724A-725A (1990)

A simple method for the production of gel-filled capillaries for capillary gelelectrophoresis, Lux, J. A.; Yin, H. F.; Schomburg, G., J. High Resolut.Chromatogr. 13, 436-437 (1990)

Quantitation of ribonucleotides from base-hydrolyzed RNA using capillaryzone electrophoresis, Huang, X.; Shear, J. B.; Zare, R.N., Anal. Chem. 62,2049-2051 (1990)

Analysis of oligonucleotides by capillary gel electrophoresis, Paulus, A.;Ohms, J. I., J. Chromatogr. 507, 113-123 (1990)

Effect of surfactant structures on the separation of cold medicine ingredientsby micellar electrokinetic chromatography, Nishi, H.; Fukuyama, T.;Matsuo, M.; Terabe, S., J. Pharm. Sci. 79, 519-523 (1990)

Separation and determination of the ingredients of a cold medicine bymicellar electrokinetic chromatography with bile salts, Nishi, H.;Fukuyama, T.; Matsuo, M.; Terabe, S., J. Chromatogr. 498, 313-323 (1990)

Enantioselective hydrophobic entanglement of enantiomeric solutes withchiral functionalized micelles by electrokinetic chromatography, Dobashi,A.; Ono, T.; Hara, S.; Yamaguchi, J., J. Chromatogr. 480, 413-20 (1989)

High-speed micellar electrokinetic capillary chromatography of the com-mon phosphorylated nucleosides, Liu, J.; Banks, J. F., Jr.; Novotny, M.,J. Microcolumn Sep. 11, 36-41 (1989)

Chiral separation by electrokinetic chromatography with bile salt micelles,Terabe, S.; Shibata, M.; Miyashita, Y., J. Chromatogr. 480, 403-411 (1989)

Separation of β-lactam antibiotics by micellar electrokinetic chromatogra-phy, Nishi, H.; Tsumagari, N.; Kakimoto, T.; Terabe, S., J. Chromatogr.477, 259-270 (1989)

Capillary zone electrophoretic separation of peptides and proteins using lowpH buffers in modified silica capillaries, McCormick, R. M., Anal. Chem.60, 2322-2328 (1988)

42

Improved electrospray ionization interface for capillary zone electrophore-sis-mass spectrometry, Smith, R. D.; Barinaga, C. J.; Udseth, H. R., Anal.Chem. 60, 1948-1952 (1988)

Capillary electrophoresis, Compton, S.W.; Brownlee, R. G., BioTechniques6, 432-440 (1988)

High-performance capillary electrophoresis using open tubes and gels,Cohen, A. S.; Paulus, A.; Karger, B. L., Chromatographia 24, 15-24 (1987)

The effect of electrolyte chain length on electroendosmotic flow in highvoltage capillary zone electrophoresis, Altria, K. D.; Simpson, C. F., Anal.Proc. (London) 25, 85 (1988)

Miniaturization in pressure and electroendosmotically driven liquid chroma-tography: some theoretical considerations, Knox, J. H.; Grant, I. H.,Chromatographia 24, 135-143 (1987)

High voltage capillary zone electrophoresis: operating parameters effects onelectroendosmotic flows and electrophoretic mobilities, Altria, K. D.;Simpson, C. F., Chromatographia 24, 527-532 (1987)

Modification of electroosmotic flow with cetyltrimethylammonium bromidein capillary zone electrophoresis, Tsuda, T., HRC & CC, J. High Resolut.Chromatogr. Chromatogr. Commun. 10, 622-624 (1987)

Bias in quantitative capillary zone electrophoresis caused by electrokineticsample injection, Huang, X.; Gordon, M. J.; Zare, R. N., Anal. Chem. 60,375-377 (1988)

High-performance sodium dodecyl sulfate polyacrylamide gel capillaryelectrophoresis of peptides and proteins, Cohen, A. S.; Karger, B. L.,J. Chromatogr. 397, 409-417 (1987)

Measurement of electroendosmotic flows in high-voltage capillary zoneelectrophoresis, Altria, K. D.; Simpson, C. F., Anal. Proc. (London) 23,453-454 (1986)

Control of electroosmosis in coated quartz capillaries, Herren, B. J.; Shafer,S. G.; Van Alstine, J.; Harris, J. M. Snyder, R. S., J. Colloid Interface Sci.115, 46-55 (1987)

Electrokinetic resolution of amino acid enantiomers with copper(II)-aspartame support electrolyte, Gozel, P.; Gassmann, E.; Michelsen, H.;Zare, R. N., Anal. Chem. 59, 44-49 (1987)

43

Capillary zone electrophoresis of proteins in untreated fused silica tubing,Lauer, H. H.; McManigill, D., Anal. Chem. 58, 166-170 (1986)

Electrokinetic separations with micellar solutions and open-tubular capillar-ies, Terabe, S.; Otsuka, K.; Ichikawa, K.; Tsuchiya, A.; Ando, T., Anal.Chem. 56, 111-113 (1984)

High-resolution separations based on electrophoresis and electroosmosis,Jorgenson, J. W.; Lukacs, K. D., J. Chromatogr. 218, 209-216 (1981)

Free-zone electrophoresis in glass capillaries, Jorgenson, J. W.; Lukacs, K.D., Clin. Chem. 27, 1551-1553 (1981)

Zone electrophoresis in open-tubular glass capillaries, Jorgenson, J. W.;Lukacs, K. D., Anal. Chem. 53, 1298-1302 (1981)

BE

CK

MA

NS

eparation of Proteins and P

eptides by Capillary E

lectrophoresisV

olume V

BECKMAN

Beckman Instruments, Inc. • 2500 Harbor Boulevard, Box 3100 • Fullerton, California 92634-3100

Sales: 1-800-742-2345 • Service: 1-800-551-1150 • TWX: 910-592-1260 • Telex: 678413 • Fax: 1-800-643-

4366

Worldwide Offices: Africa, Middle East, Eastern Europe (Switzerland) (22) 994 07 07. Australia (61) 02

816-5288. Austria (2243) 85656-0. Canada (800) 387-6799. China (861) 5051241-2. France (33) 1 43 01 70

00. Germany (49) 89-38871. Hong Kong (852) 814 7431. Italy (39) 2-953921. Japan 3-3221-5831. Mexico 525

575 5200, 525 575 3511. Netherlands 02979-85651. Poland 408822, 408833. Singapore (65) 339 3633. South

Africa (27) 11-805-2014/5. Spain (1) 358-0051. Sweden (8) 98-5320. Switzerland (22) 994 07 07. Taiwan (886)

02 378-3456. U.K. (0494) 441181. U.S.A. 1-800-742-2345.

727484 $14.95 Printed in U.S.A. ©Copyright 1994 Beckman Instruments, Inc.

BECKMAN

Separation of Proteins and Peptidesby Capillary Electrophoresis:

Application toAnalytical Biotechnology

Separation of Proteinsand Peptides

by Capillary Electrophoresis:Application to Analytical

Biotechnology

Herb Schwartz

Palomar Analytical Services Redwood City, CA

and

Tom Pritchett

Beckman Instruments, Inc. Fullerton, CA

iii

Table of Contents

About the Authors...................................................................................vAcknowledgments................................................................................. viFront Cover........................................................................................... viAcronyms and Symbols Used............................................................... viiPreface.................................................................................................. ix

Part 1—CE Techniques Applied to Proteins and Peptides1.1 Prediction of Electrophoretic Mobility......................................... 1-31.2 Buffer Selection and pH Control in CZE...................................... 1-9

1.2.1 Ohm’s Law Plot.................................................................. 1-91.2.2 Buffer Composition, pH Control........................................1-11

1.3 Separation of Peptides in Free Solution......................................1-141.3.1 Types of Capillaries..........................................................1-151.3.2 Use of Buffer Additives to Optimize

Peptide Separations...........................................................1-151.4 Separation of Proteins in Free Solution.......................................1-20

1.4.1 Adsorption........................................................................1-201.4.2 Strategies to Prevent Protein Adsorption...........................1-221.4.3 Protein Modification: Use of Ionic Surfactants and Urea .. 1-37

1.5 Detectability Enhancements: Matrix Effects, Sample Stacking,and ITP Preconcentration...........................................................1-411.5.1 Laser-Induced Fluorescence Detection (LIF)....................1-421.5.2 Effect of Sample Matrix; Stacking of Sample

Components.....................................................................1-421.5.3 ITP Preconcentration........................................................1-47

1.6 SDS Capillary Gel Electrophoresis (SDS-CGE).........................1-521.7 Capillary Isoelectric Focusing (CIEF)........................................1-55

1.7.1 Principle of IEF................................................................1-551.7.2 Classical IEF....................................................................1-561.7.3 Capillary Isoelectric Focusing (CIEF)...............................1-56

1.8 Micropreparative CE..................................................................1-601.8.1 Collection from a Single Run with Standard Capillaries.... 1-611.8.2 Collection by Performing Multiple Runs..........................1-611.8.3 Collection from Large-Diameter Capillaries.....................1-61

1.9 Affinity Capillary Electrophoresis (ACE)...................................1-631.9.1 Receptor-Ligand Studies...................................................1-641.9.2 Antibody-Antigen Interactions..........................................1-66

iv

Part 2—Protein/Peptide Applications of CEto Analytical Biotechnology

2.1 Introduction.................................................................................. 2-32.1.1 The Importance of Analytical Chemistry to

Biotechnology.................................................................... 2-32.1.2 The “Eight Points” Model of Analytical Development........ 2-3

2.2 Identity......................................................................................... 2-42.2.1 Peptide Mapping: Utility of CZE........................................ 2-52.2.2 Molecular Weight Estimation: Utility of SDS-CGE............ 2-72.2.3 Identity of Monoclonal Antibodies (MAbs): Utility of

Capillary Isoelectric Focusing (CIEF)................................. 2-92.2.4 Confirmation of Peak Identity by

CE-Mass Spectrometry (CE-MS).....................................2-132.3 Quantity.....................................................................................2-16

2.3.1 Accuracy and Precision in Quantitative CE Analysis........2-172.3.2 Quantitation of Dosage Forms by CIEF............................2-182.3.3 Quantitation with CZE and SDS-CGE..............................2-212.3.4 Immunoassays Using Affinity CE (ACE).........................2-23

2.4 Purity.........................................................................................2-262.4.1 Purity of Proteins..............................................................2-262.4.2 Screening the Purity of Peptides.......................................2-30

2.5 Heterogeneity.............................................................................2-302.5.1 Monoclonal Antibodies.....................................................2-312.5.2 Glycoforms of Recombinant Proteins...............................2-312.5.3 Heterogeneity in Proteins Relevant to

Clinical/Diagnostic Applications.......................................2-332.6 Stability......................................................................................2-342.7 Process Consistency...................................................................2-382.8 Method Validation......................................................................2-39

References............................................................................Ref-1

v

About the Authors

Herbert E. Schwartz is a native of the Netherlands, and holds a Ph.D. in Ana-lytical Chemistry from Northeastern University, Boston, and an M.S. degreefrom the Free University, Amsterdam. He has been working in the analyticalinstrumentation industry for over ten years, and has authored fifty publicationsin the field of separation science. Before starting his own consulting firm,Palomar Analytical Services in Redwood City, CA, Herb managed the CEapplications group at Beckman Instruments, Inc., Palo Alto, CA, and wasinvolved in the development of the first fully automated, commercial CE in-strument at Microphoretic Systems. Prior to that, he was employed as a re-search chemist at Applied Biosystems and Brownlee Labs. He edited theprevious primers, Volumes I–IV, on capillary electrophoresis for Beckman.

Thomas J. Pritchett is a Principal Scientist with Beckman Instruments,Inc., specializing in development and validation of CE and HPLC methods forthe analysis of (glyco)proteins. He received a Ph.D. in Biological Chemistryfrom the University of California, Los Angeles, and did post-doctoral work atthe Scripps Research Institute, La Jolla, CA. He was formerly manager of QCand Analytical Development at Cytel Corporation, an early-stage biopharma-ceutical company in San Diego, CA. Previous to Cytel, Tom was associatedwith the Genetics Institute, Andover, MA, and Baxter Diagnostics, Inc., Mi-ami, FL, where he was the founding member and manager of the AnalyticalChemistry/Method Development group.

vi

Acknowledgments

We would like to thank Richard Palmieri, Judy Nolan, Jeff Allen, JeffChapman (Canada), and Ron Brown (UK) at Beckman Instruments for themany suggestions and contributions to this book; Bert Rietveld at Beckman forinitiating this project and fruitful discussions; Andras Guttman and RonBiehler at Beckman for reviewing drafts of the manuscript; Gale Leach andAnnette Hurst at WordsWorth (Pacifica, CA) for the desktop publishing; DonGregory at Molecular Simulations (Burlington, MA) for the computer-gener-ated front cover.

Front Cover

The illustration on the cover depicts a computer-generated glycoprotein mol-ecule. Color code: white, oligosaccharide; red, Asn residues; yellow and blue,other amino acids of ribonuclease B. Courtesy of Don Gregory, MolecularSimulations, Burlington, MA.

Other Beckman primers (Volumes I, II, III, and IV) oncapillary electrophoresis:

Title BeckmanPart Number

Introduction to Capillary Electrophoresis 360643

Introduction to Capillary Electrophoresisof Proteins and Peptides 266923

Micellar Electrokinetic Chromatography 266924

Introduction to the Theory and Applicationsof Chiral Capillary Electrophoresis 726388

All trademarks and registered trademarks are the property of their respectiveowners.

vii

Acronyms and Symbols Used

The following acronyms and symbols may be found throughout this book.

ACE affinity capillary electrophoresisACTH adrenocorticotropic hormoneBGE background electrolyteCE capillary electrophoresisCEA carcinoembryonic antigenCGE capillary gel electrophoresisCGMP current good manufacturing proceduresCHO Chinese hamster ovaryCIEF capillary isoelectric focusingCMC chemistry, manufacturing, and controlCTAB cetyltrimethylammonium bromideCTAC cetyltrimethylammonium chlorideCZE capillary zone electrophoresisDAB diaminobutaneDAP diaminopentaneDMF dimethylformamideEDTA ethylenediaminetetraacetic acidELISA enzyme-linked immunoassayEOF electroosmotic flowG-CSF granulocyte colony stimulating factorGM-CSF granulocyte macrophage colony stimulating factorHGH human growth hormoneHPCE high-performance capillary electrophoresisHPIEC high-performance ion exchange chromatographyHPLC high-performance liquid chromatographyHPMC hydroxypropylmethylcelluloseHSA hexanesulfonic acidIEF isoelectric focusingIFN-α recombinant human interferon-αIgG immunoglobulin GIND investigative new drug applicationITP isotachophoresisLE leading electrolyteLIF laser-induced fluorescenceµ mobility

viii

m/z mass to chargeMAbs monoclonal antibodiesMECC micellar electrokinetic capillary chromatographyMHC major histocompatibility complexMS mass spectrometryMW molecular weightNDA new drug applicationPA polyacrylamidePAGE polyacrylamide gel electrophoresisPBS phosphate-buffered salinepI isoelectric pointPLA product license agreementrbst recombinant bovine somatotropinrhIL recombinant human interleukinrHuEPO recombinant human erythropoietinRIA radiolabeled immunoassayRMT relative migration timeRP reversed phaserpst recombinant porcine somatotropinRSD relative standard deviationrtPA recombinant human tissue plasminogen activatorSDS sodium dodecyl sulfateTE terminating electrolyteTFA trifluoroacetic acidTHF tetrahydrofuranTMR tetramethylrhodamineTNF tumor necrosis factorUV ultravioletVis visible

ix

Preface

The emergence of a new biotechnological industry utilizing hybridoma, pep-tide synthesis, and recombinant DNA techniques for the production of highlyspecialized biomolecules has increased the demand for sophisticated analyticalinstrumentation and methodologies. In pharmaceutical biotechnology, theseanalytical instruments and methods are used for the identification of chemicalsand structures, control of purity, and assays of potency. Quantity, stability,heterogeneity and process consistency are important issues. In biologicalsamples, proteins, nucleic acids, and polysaccharides are often present in verysmall quantities and sample sizes are often limited, requiring highly sensitiveand selective separation techniques. The same is true for the analysis of com-plex samples of biological origin, i.e., in clinical and diagnostic applications.Miniaturized, microcolumn separation techniques, such as micro-HPLC, capil-lary gas chromatography, or capillary electrophoresis (CE), are particularlysuitable for the detection of analytes in very small, (sub)microliter-size vol-umes.

Because many samples of biological origin are quite complex, two ormore inherently different yet complementary techniques are often used toperform qualitative and quantitative assays. The use of such complementarymethods provides more confidence in the analytical results. HPLC and CE(or, in a broader sense, chromatography and electrophoresis) fulfill the aboverequirements. These techniques provide fully automated, computer-controlled,quantitative assays as well as high-resolution separations with fast analysistimes. For example, whereas in (reversed-phase) HPLC, species are typicallyseparated on the basis of hydrophobicity, in free-solution CE, the charge-to-mass ratio often plays a key role. This difference in separation mechanism ishelpful in the characterization or elucidation of complex samples of biologicalorigin. In addition, clean-up or purification of biomolecules is often done bypreparative HPLC, thus requiring a second technique for purity control. CE,therefore, is a welcome addition to the toolbox of the bioanalytical chemist forsolving problems related to proteins and peptides.

HPLC has undergone a tremendous development in the past two decades.This growth has been possible because of the advent of high-performancepacking materials, suitable instrumentation, and a thorough theoretical under-standing of separations. We are seeing a similar development in the field ofCE. The development in CE of capillary pretreatment procedures and perma-nent coatings is similar to developments in chromatography decades ago.Research in CE—in its present form with fused-silica columns and on-columndetection—is more than a decade old (Jorgenson and Lukacs, 1981, 1983), and

x

commercial CE instruments have been available since 1988. The field is mov-ing rapidly, illustrated by the fact that already six annual symposia on CE havebeen held and a number of comprehensive textbooks (see, for example, thetextbooks edited by Camillieri, 1993; Landers, 1994; and Guzman, 1993)recently have been published. The reviews by Karger (1993) and Landers et al.(1993) capture the state of the art of CE. A sampling of the type of protein-related applications for which CE has been utilized in analytical biochemistryand clinical/diagnostic assays is given below.

Applications of CE for Proteins and Peptides

Purity Assays• QC of a manufacturing process• Quantitation of contaminants or excipients• Screening of samples prior to protein sequencing

Structural Studies• Peptide mapping• Two-dimensional methods, i.e., HPLC-CE, IEC-CE• Microheterogeneity of complex proteins e.g., glycoproteins,

monoclonal antibodies, histones, transferrins• Oligomerization

Binding Studies• Calcium, zinc-binding proteins• Antigen-antibody immune complexes

• Drug–protein complexes• Protein–DNA complexes

Process Analysis• Quantitation of the final product• Enzymatic digestion monitoring• Derivatization (labeling) monitoring• Deamidation, disulfide bridge reduction• Conformational changes

Stability Studies• pH, environment, temperature effects

Mobility Measurements• (Semi-)empirical models for prediction of separation

Micropreparative PurificationSee applications listed in Table 1-4

xi

Clinical/Diagnostic• Serum protein assays• Hemoglobin variants; globin chains• Neuropeptide isolation from brain tissue• Apolipoproteins• Cytokines, transferrins, metallothioneins• Proteins in single cells• Isoenzymes

As in HPLC, various separation modes are applicable to the analysis ofproteins and peptides: capillary zone electrophoresis (CZE), micellar electroki-netic capillary chromatography (MECC), capillary isoelectric focusing(CIEF), capillary gel electrophoresis (CGE), and displacement electrophoresisor isotachophoresis (ITP). Whereas in CE the separation of small peptides(in the CZE mode) often is relatively straightforward and well understood, itappears that no single strategy is applicable for large peptides and proteins.As might be expected, this is due largely to the wide diversity and complexityassociated with these biomolecules. Thus, different strategies often work fordifferent protein separation problems, hence requiring different CE separationmodes. An introduction to the general principles of CE has appeared in anotherBeckman primer (part number 360643) and should serve as a quick referenceguide and introduction to the present booklet on protein and peptide separa-tions. Please contact your local Beckman office if you wish to order a copy.

This edition on the CE of proteins and peptides is a completely revisededition of the first printing of 1992. It is divided into two parts. Part I is anupdated version of the first booklet, with added new material on coated capil-laries, detectability enhancements, CIEF, and Affinity CE. It gives generalbackground information on various strategies devised for protein separations.Part II discusses the specific application of the CE techniques to analyticalbiotechnology, while some clinical/diagnostic applications are also discussed.The objective is to show where and why CE can be used as a solution to acritical analytical need. Specifically, the role of capillary electrophoresis inproduct identity, quantity, purity, heterogeneity, stability, and process consis-tency is addressed.

In reading this primer, some familiarity with the basics of CE and proteinstructure and function is assumed. It is not meant as an all-inclusive, comprehen-sive review, but rather as a non-theoretical, practical guide which can be helpfulas a reference or in designing separations. Relevant examples from our laborato-ries at Beckman Instruments in Fullerton, California, and from the literature areselected to serve as illustrations. For easy reference, the articles and bookscited in the text are listed alphabetically at the back of this primer.

1

I: Introduction

Electrokinetic chromatography (EKC) is a family of electrophoresis tech-niques named after electrokinetic phenomena, which include electroosmo-sis, electrophoresis, and chromatography. Micellar electrokineticchromatography (MEKC) is a mode of EKC in which surfactants (mi-celles) are added to the buffer solution. Surfactants are molecules whichexhibit both hydrophobic and hydrophilic character. They have polar“head” groups that can be cationic, anionic, neutral, or zwitterionic andthey have nonpolar, hydrocarbon tails. The formation of micelles or“micellization” is a direct consequence of the “hydrophobic effect.” Thesurfactant molecules can self-aggregate if the surfactant concentrationexceeds a certain critical micelle concentration (CMC). The hydrocarbontails will then be oriented toward the center of the aggregated molecules,whereas the polar head groups point outward. Micellar solutions may solu-bilize hydrophobic compounds which otherwise would be insoluble inwater. The front cover picture shows an aggregated SDS molecule. In thecenter of the aggregate, p-fluorotoluene is situated depicting the partition-ing of a neutral, hydrophobic solute into the micelle. Every surfactant has acharacteristic CMC and aggregation number, i.e., the number of surfactantmolecules making up a micelle (typically in the range of 50-100). (See alsoTable 1 and the discussion on page 10). The size of the micelles is in therange of 3 to 6 nm in diameter; therefore, micellar solutions exhibit proper-ties of homogeneous solutions. Micellar solutions have been employed in avariety of separation and spectroscopic techniques. In 1980, Armstrongand Henry pioneered the use of micellar solutions as mobile phases forreversed-phased liquid chromatography (RPLC).

In the literature, MEKC is also often referred to as MECC (micellarelectrokinetic capillary chromatography) since the separations are mostoften performed in a capillary tube. Other modes of EKC are cyclodextrinEKC (CDEKC), ion-exchange EKC (IXEKC), and microemulsion EKC(MEEKC). Cyclodextrin derivatives, polymer ions, and microemulsionsare used in CDEKC, IXEKC, and MEEKC, respectively, instead of themicelles used in MEKC. The references listed on page 3 provide furtherdetail on the differences between the various kinds of EKC techniques. Inthe following chapters, relevant references are listed in reverse chronologi-cal order after each chapter. All EKC techniques are based on the same

2

separation principle: the differential partitioning of an analyte between atwo-phase system (i.e., a mobile/aqueous phase and a stationary phase).

The same instrument that is used for capillary zone electrophoresis(CZE) is also used for MEKC. Both MEKC and CZE are modes of capil-lary electrophoresis (CE), as are capillary gel electrophoresis, capillaryisoelectric focusing, and capillary isotachophoresis (for an introduction toCE, see the Beckman Primer Introduction to Capillary Electrophoresis,part number 360643). MEKC is different in that it uses an ionic micellarsolution instead of the simple buffer salt solution used in CZE. The micel-lar solution generally has a higher conductivity and hence causes a highercurrent than the simple buffer does in CZE. MEKC can separate both ionicand neutral substances while CZE typically separates only ionic sub-stances. Thus MEKC has a great advantage over CZE for the separation ofmixtures containing both ionic and neutral compounds. However, inMEKC the size of the sample molecules is limited to molecular weights ofless than 5000, whereas CZE has virtually no limitation in molecular size.The separation principle of MEKC is based on the differential partition ofthe solute between the micelle and water; CZE is based on the differentialelectrophoretic mobility.

3

Further Reading(in reverse chronological order)

Janini, G. M., Isaaq, H. J. Micellar electrokinetic capillary chromatogra-phy: basic considerations and current trends. J. Liq. Chromatogr. 15, 927-960 (1992)

Kuhr, W. G., Monnig, C. A. Capillary Electrophoresis. Anal. Chem. 64,389R-407R (1992)

Watarai, H. Microemulsion capillary electrophoresis. Chem. Lett., 391-394(1991)

Nishi, H., Terabe, S. Application of micellar electrokinetic chromatogra-phy to pharmaceutical analysis. Electrophoresis 11, 691-701 (1990)

Terabe, S., Isemura, T. Ion-exchange electrokinetic chromatography withpolymer ions for the separation of isomeric ions having identical electro-phoretic mobilities. Anal. Chem. 62, 650-652 (1990)

Terabe, S. Electrokinetic chromatography: an interface between electro-phoresis and chromatography. Trends Anal. Chem. 8, 129-134 (1989)

Khaledi, M. G. Micellar reversed phase liquid chromatography.Biochromatography 3, 20-35 (1988)

Burton, D. E., Sepaniak, M. J. Analysis of B6 vitamers by micellar electro-kinetic capillary chromatography with laser-excited fluorescence detection.J. Chromatogr. Sci. 24, 347-351 (1986)

Terabe, S., Ozaki, H., Otsuka, K., Ando, T. Electrokinetic chromatographywith 2-O-carboxymethyl-β-cyclodextrin as a moving "stationary" phase.J. Chromatogr. 332, 211-217 (1985)

Terabe, S., Otsuka, K., Ando, T. Electrokinetic chromatography with mi-cellar solution and open-tubular capillary. Anal. Chem. 57, 834-841 (1985)

Terabe, S., Otsuka, K., Ichikawa, K., Tsuchiya, A., Ando, T. Electrokineticseparations with micellar solution and open-tubular capillaries. Anal Chem.56, 111-113 (1984)

Armstrong, D. W., Henry, S. J. Use of an aqueous micellar mobile phasefor separation of phenols and polynuclear aromatic hydrocarbons viaHPLC. J. Liq. Chromatogr. 3, 657-662 (1980)

4

II: Separation Principle/Fundamentals

Figure 1 shows a schematic representation of the separation principle ofMEKC. When an anionic surfactant such as sodium dodecyl sulfate (SDS)is employed, the micelle migrates toward the positive electrode by electro-phoresis. The electroosmotic flow transports the bulk solution toward thenegative electrode due to the negative charge on the surface of fused silica.The electroosmotic flow (EOF) is usually stronger than the electrophoreticmigration of the micelle under neutral or alkaline conditions and, therefore,the anionic micelle also travels toward the negative electrode at a retardedvelocity.

= Surfactant(negative charge)

= Solute

= Electroosmotic Flow

= Electrophoresis

Figure 1. Schematic of the separation principle of MEKC. The detectorwindow is assumed to be positioned near the negative electrode.

When a neutral analyte is injected into the micellar solution, a fractionof it is incorporated into the micelle and it migrates at the velocity of themicelle. The remaining fraction of the analyte remains free from the mi-celle and migrates at the electroosmotic velocity. The migration velocity ofthe analyte thus depends on the distribution coefficient between the micel-lar and the non-micellar (aqueous) phase. The greater the percentage ofanalyte that is distributed into the micelle, the slower it migrates. Theanalyte must migrate at a velocity between the electroosmotic velocity andthe velocity of the micelle (see Figure 2A), provided the analyte is electri-cally neutral. In other words, the migration time of the analyte, tR, is lim-ited between the migration time of the bulk solution, t0, and that of the

5

micelle, tmc (see Figure 2B). This is often referred to in the literature as themigration time window in MEKC.

Water Solute Micelle

Micelle Solute Water

inj. column det.

Time0 t0 tR tmc

(A)

(B)

Figure 2. Schematic of the zone separation in MEKC (A) and chromato-gram (B). Reproduced with permission from Terabe, et al., Anal. Chem.57, 834 (1985).

Capacity Factor

We can define the capacity factor, k', similarly to that of chromatographyas

k' = nmc

naq(1)

where nmc and naq are the amount of the analyte incorporated into the mi-celle and that in the aqueous phase, respectively. We can obtain the rela-tionship between the capacity factor and the migration times as

tR = 1 + k'1+ (t0 / tmc )k'

t0 (2)

The migration time of the analyte is equal to t0 when k' = 0, or whenthe analyte does not interact with the micelle at all; the migration timebecomes tmc when k' is infinity or the analyte is totally incorporated intothe micelle. Thus, the migration time window is limited between t0 andtmc.

6

When t0 is infinity (electroosmosis is completely suppressed), equa-tion (2) becomes

tR = (1 + 1 / k' )tmc (3)

In this case, the bulk solution remains stationary in the capillary andthe micelle migrates only by electrophoresis. If we define the capacityfactor as the reciprocal of equation (1), equation (3) becomes identical withthe relationship between tR, t0, and k' in conventional chromatography.

Figure 3 shows a typical example of MEKC separation. Eight electri-cally neutral compounds were successfully resolved in 17 min. The capac-ity factor scale is inserted in the figure to indicate the relationship betweenthe migration time and the capacity factor. The capacity factor of infinitymeans that analyte has the same migration time as the micelle. Theoreticalplate numbers calculated from the peak widths range from 200,000 to250,000 which is typical for MEKC separations.

0.004 AU

0 1 2 6 10 20 50∞

0 5 10 15Time (min)

Capacity Factor

1

2

3

4

5

6

7

8

Figure 3. Micellar electrokinetic chromatogram of a test mixture:1 = methanol; 2 = resorcinol; 3 = phenol; 4 =p-nitroaniline;5 = nitrobenzene; 6 = toluene; 7 = 2-naphthol; 8 = Sudan III. Conditions:capillary, 50µm i.d. × 65 cm (effective length 50 cm); run buffer, 30 mMSDS in 50 mM phosphate/100 mM borate (pH 7.0); applied voltage, 15 kV;current, 33µA; detection, UV absorbance at 210 nm; temperature, 35°C.Reproduced with permission from Terabe, Trends Anal. Chem. 8, 129 (1989).

Part 1

CE TechniquesApplied to Proteins

and Peptides

1-3

1.1 Prediction of ElectrophoreticMobility

The ability to predict the migration behavior of peptides and proteins underdifferent experimental conditions (e.g., changes in pH) would be extremelydesirable as it would allow the optimization of separation conditions. For ex-ample, investigators could choose conditions such that a native protein couldbe distinguished from closely related variants or degradation products. Predic-tive models also might aid in the characterization of unknown species and inpurity assays. This optimization approach could be applicable to the study ofmicroheterogeneity of proteins such as glycoproteins, immunoglobulins, trans-ferrins, and histone proteins.

In general, small peptides1 consisting of just a few amino acids are“well behaved” in CZE and their electrophoretic migration (mobility) can bepredicted based on their mass (size) and charge characteristics. The charge of asmall peptide can be estimated from the pKa values of the individual aminoacids. With this information, the migration of a small peptide in a particular CEbuffer can be easily calculated (for example, by a computer program—seereviews by McCormick, 1994; Palmieri and Nolan, 1994). For larger peptidesand proteins, calculation of the charge based on ionization constants is nottrivial and cannot be calculated easily from the pKas of the free amino acids.Aside from the mass-to-charge ratio, other factors which may affect mobilityare hydrophobicity, primary sequence, conformational differences, and thechirality of amino acids.

To illustrate this point, let us examine the case of a number of nonapep-tides with identical amino acid composition but different primary sequences.The sequences of the peptides are shown in Table 1-1. The question is whetheror not these peptides can be separated by CE.

1 In this text, and in accordance with other textbooks, peptides containing ten or feweramino acids are called oligopeptides (or small peptides). Peptides with a molecularmass of approximately 5000–7000 Daltons (approximately 50–70 amino acids) lieon the borderline between polypeptides and proteins.

1-4

Table 1-1. Six Nonapeptide Isomers and Their Sequences

Peptide Amino Acid Sequence

5 NH2-ALDYALAHR-COOH6 NH2-ALDYARLAH-COOH7 NH2-ALDYHALAR-COOH8 NH2-HALDYARLA-COOH9 NH2-ALDYHARLA-COOH

10 NH2-DHAYLLAAR-COOH

In Figure 1-1, the zone velocity of the peptides is plotted versus the pH.The plot indicates that, at different pH values, these peptides should be sepa-rable. Apparently the proteins do not have the same mass-to-charge ratio and,with some fine-tuning (i.e., pH manipulation), they can be resolved from eachother. The net charge of a peptide can be modified by a small change in the pHof the buffer, particularly near the pKas of the amino terminal, carboxy termi-nal, or side groups. Since all these peptides contain the same amino acids, theamino acid sequence can influence the pKas of the ionizable side groups and,therefore, the migration behavior of the peptides. Under low-pH conditions(as are present in the CE runs of Figure 1-1), the Asp side group, as well as thecarboxy terminal group, is titrated. The charge on peptide #10, which is uniquein that it has an amino terminal group near the Asp and Arg near the carboxylterminal, is evidently altered more strongly than the other peptides. This de-creases the positive charge on peptide #10 and causes a reduced zone velocity.

1-5

3.5

2.5

1.52.25 3.25 4.25

Peptide #5

Peptide #6

Peptide #7

Peptide #8

Peptide #9

Peptide #10

Zon

e V

eloc

ity (

cm/m

in)

pH

Figure 1-1. Zone velocity of peptides plotted against pH. Linear sequences ofthe peptides are listed in Table 1-2. From Field et al., Beckman ApplicationData Sheet DS-791 (1991).

Semi-empirical and theoretical models have been described to predictmobility in CE buffer systems. In 1966, Offord proposed the following equa-tion relating the mobility (µ) of peptides with their valence (Z) and molecularmass (M):

µ = k.Z.M-2/3

where k is a constant. Thus a plot of mobility versus Z.M-2/3 should yield astraight line. This relationship was later confirmed by several researchers inthe field of CE. The equation states that the frictional forces opposing theelectrophoretic migration are proportional to the surface area of the species(assuming a spherical shape of the molecule with a radius proportional to thecube root of M). Researchers at Eli Lilly have extensively studied the mobilitycharacteristics of peptides and proteins. An example from their data, derivedfrom peptide fragments of an hGH digest, is shown in Figure 1-2A.

1-6

4.00e-4

3.00e-4

2.00e-4

1.00e-4

0.00e-4

0.00 0.01 0.02 0.03 0.04 0.05 0.06

q / MW2/3

Ele

ctro

phor

etic

Mob

ility

A

4.00e-4

3.00e-4

2.00e-4

1.00e-4

0.00e-4

0.00 0.01 0.02 0.03 0.04 0.05

q / MW2/3 from amino acid pKa

Ele

ctro

phor

etic

Mob

ility

B

Figure 1-2. Fit of electrophoretic mobility (cm2/V·s) versus the charge-to-sizeparameter for hGH digest separated in 0.1 M glycine, pH 2.35. (A) Data frompKa values calculated at Eli Lilly. Correlation coefficient, 0.989. (B) Databased on pKa values of the isolated amino acids. Correlation coefficient,0.956. Data reproduced with permission from Rickard et al., Anal. Biochem.197, 197 (1991).

1-7

Note that the charges in Figure 1-2A were calculated from ionizationconstants (pKas consistent with those of typical peptides and proteins; for adetailed discussion and a list of ionization constants, see Rickard et al. (1991).A significant lack of fit was observed when the charges were calculated simplybased on the pKa values of the isolated amino acids. As a comparison, the plotbased on these data is shown in Figure 1-2B.

Because mobility is defined as velocity per unit field strength, the aboveequation can also be re-written in the form of a relative migration time (relativeto a standard reference peptide) as:

t/t0 = K0.Z-1.M2/3

where t/t0 is the relative migration time and K0 is a constant pertaining to thereference peptide. Therefore, a plot of relative migration time versus M2/3/Zshould also give a straight line as was found and published by several research-ers (Deyl, 1989; Hjerten, 1989).

For closely related peptides, the above “Offord” equation can be used toestimate the optimum pH of the CE run buffer. This was shown by Bongers,et al. (1992) for synthetic peptides derived from human growth hormone re-leasing factor, a 44-residue peptide. The 11-residue peptides differed only bysubstitution of Asn, Asp, or β-Asp at a single residue. Titration curves (chargevs. pH) were constructed based on the pK data from Skoog and Wichman(1986) as illustrated in Figure 1-3. The curves reveal that it can be predictedthat the optimum pH to separate the peptides should be between 3 and 5, ashere the charge differences are maximized. Since the MWs of the peptides areapproximately the same, their mass-to-charge ratios and, consequently, theirmobilities differ the most in this pH range. The electropherogram shown at thebottom of Figure 1-3 shows that, indeed, baseline separation occurs at pH 4.3.

As shown above, the “Offord” charge/size parameter, Z.M-2/3 is related tothe peptide mobility in CZE, and thus can be used to predict the relative migra-tion order of peptides. This was verified by McCormick (1994) who usedpublished data from Strickland and Strickland (1990) to calculate the charge-to-size parameter for 14 peptides with known sequence. The peptides variedwidely in size and composition. A commercially available computer program(IntelliGenetics, Mountain View, CA) was used for the calculation of chargeand Z.M-2/3. Exactly as predicted, the CE migration time order of the 14 pep-tides correlated well (inversely) with the order of calculated Z.M-2/3 values.Thus, the use of such procedures would be a valuable tool in the determinationof peak identity for peptide separations in CZE.

1-8

2

1 1

0

-1

-2

-3 0 1 2

2

3

3

4 5 6 7 8 9 10 11 12 13 pH

Cha

rge

A

B pH 2.3 pH 2.6 pH 3.4 pH 4.3

1

147 19

11,21,2,32

2 33

3

A21

0

t (min)12

t (min)7 12

t (min)6 11

t (min)

Figure 1-3. (A) Calculated charge versus pH profiles and (B) electrophero-grams at varying pH for the synthetic model peptides (Leu27, Asn28) GRF(22-32)-OH,2, and (Leu27,β−Asp28)-GRF(22-32)-OH,3. The pH of the 25-mMNa2HPO4 /H3PO4 running buffer is shown in the upper left corner of eachpanel. Reprinted with permission from Bongers et al., J. Liq. Chromatogr. 15,1115 (1992).

Deyl (1989) demonstrated that, for a large number of collagen proteinswhich varied considerably in molecular weight, the relative migration time islinearly related to their pIs. This relationship was studied in untreated fused-silica capillaries within the pH range of 6.9 to 10.5 and included data replottedfrom other researchers. Other investigators (Compton, 1991) have modeled themobility of proteins in CE and applied it to protein microheterogeneity analy-sis. In this study, mobility was found to be a continuous function of M-1/3 toM-2/3, depending on the magnitude of M and the ionic strength of the buffer.Research on the modeling of mobility of peptides and proteins in CE is ongo-ing and various approaches are being investigated at this time.

1-9

1.2 Buffer Selection and pH Controlin CZE

As pointed out in Section 1.1, the choice of a run buffer is very important inCE because it determines the charge on the analyte molecule and, therefore,its migration velocity. Briefly, the characteristics of a useful CE buffer are:

1. Good pH control/buffer capacity2. Low conductivity3. Good UV transparency (especially in the low UV region)

To obtain high-performance separations and special selectivities, theaddition of certain compounds to the buffer is often required. These additiveswill be discussed in detail in Sections 1.3 and 1.4 dealing with peptides andproteins, respectively. The type of buffer, its ionic strength, and its pH can bevaried and optimized for a particular separation problem. With untreatedfused-silica capillaries, the use of high-pH buffers generally produces fastseparations because the electroosmotic flow (EOF) is high. At a low pH, pep-tides migrate primarily on the basis of their charge-to-mass characteristics andthe EOF is diminished. Buffers based on sodium phosphate, citrate, acetate, orcombinations thereof are frequently used in CE. Typical buffer concentrationsare 20 to 200 mM. Borate (pH range 7.5 to 10), in particular, is a popularbuffer in CE for a wide variety of applications. This buffer has an inherentlylow conductivity (its use in CE has been reported with concentrations of 500mM). Furthermore, borate is known to complex with diol groups which facili-tate analysis of sugars and glycoproteins (Landers, 1993). Often, the conduc-tivities associated with phosphate, citrate, and acetate buffers are relativelyhigh, necessitating the selection of smaller i.d. capillaries or the adjustment ofelectric field strength conditions. Such limitations can be greatly minimized byemploying efficient capillary cooling designs such as circulating liquids usedin Beckman’s P/ACE™ system. In certain cases, low conductivities can beachieved by using zwitterionic buffers, i.e., the so-called “Good’s” buffers(see Sigma catalog, 1993, page 1556). Several examples of the utility of thesebuffers in CE are discussed below.

1.2.1 Ohm’s Law PlotTo estimate the maximum voltage that can be applied during a CE run, an“Ohm’s Law” plot can be constructed (Nelson et al., 1989). It is also a usefulprocedure to compare run buffers (or even CE cooling systems) in terms ofcurrent and Joule heat generation. By recording the current at each appliedvoltage (e.g., at 1-minute intervals), an ideal Ohm’s Law plot should yield a

1-10

straight line. This is the case when the heat generated inside the capillary isadequately dissipated (aided by the cooling system of the CE instrument).Deviations from linearity are indicative of inadequate Joule heat dissipation.At the voltage where linearity is lost, the heat dissipation capacity of the sys-tem has been exceeded. As a rough rule of thumb, the heat generated with aCE run buffer should not exceed ≈ 5 W/m. Figure 1-4 shows Ohm’s Law plotsfor three buffers, i.e., 100 mM phosphate, pH 2.5; 100 mM borate, pH 8.3; and100 mM CAPS, pH 11.0. It can be seen that, while the borate buffer yields astraight line over 0 to 30 kV, deviation from linearity occurs much earlier,i.e.,at a lower voltage, with CAPS and phosphate. At an applied voltage of20 kV, the power generated with CAPS and phosphate is 5.88 and 10.07 W/m,respectively. Therefore, it is advisable to use lower voltages for these buffersduring CE runs. On the other hand, with borate, only 0.58 W/m is generated.At 20 kV, the recorded current was only 10 mA for borate, while 100 and150 mA for CAPS and phosphate, respectively.

A B

0 10

10

20

200

150

100

50

0

100

00 10 20 30 2.5 5 7.5 10 12.5 15 17.5 20 22.5 25 27.5 30

20 30

Borate PhosphateCAPSBorate

AcceptableSeparation

Voltage

ExcessiveJoule

Heating

Applied Voltage (kV)

Cur

rent

(µA

)

Figure 1-4. Ohm’s Law Plot. (A) Plot of observed current vs. applied voltagefor each of three buffers. The voltage was incremented at 2.5 kV/min. Bufferswere 100 mM phosphate, pH 2.5, made by dilution of phosphoric acid andtitration with NaOH; 100 mM borate, pH 8.3, made by titrating 25 mM sodiumtetraborate with 100 boric acid; and 100 mM CAPS, pH 11.0, made by titra-tion of the appropriate concentration dissolved in water with NaOH. The insetshows the borate data plotted on an expanded scale. (B) Direct plot of currentvs. applied voltage for 100 mM CAPS, pH 11.0. Voltage is incremented by2.5 kV/min. A straight line drawn through the front edge of the plateau illus-trates the ability of the cooling system to dissipate the heat generated by thepassage of current. The departure from linearity indicates the excessive in-crease in current at the applied voltage, and is a reflection of the increase incapillary temperature. Reprinted with permission from Oda and Landers,Handbook of Capillary Electrophoresis, Landers (Ed.), Boca Raton: CRCPress, 1994.

1-11

While it is relatively easy to make up CE run buffers, it is now possibleto buy ready-to-go CE buffers from various sources (e.g., Fluka, ISCO, Scien-tific Resources; Beckman sells a number of useful CE buffers designed for usewith coated and untreated capillaries and in conjunction with application-specific kits).

1.2.2 Buffer Composition, pH Control

Camilleri and co-workers (1991) have suggested the use of deuterium oxide(D2O) instead of water when making a run buffer. The increased viscosity of aD2O buffer is favorable with respect to resolution of closely spaced peaks.To illustrate this point, Figure 1-5 shows the separation of a tryptic digest inboth water and D2O buffers. While the cost of D2O (≈ $70.00 per 100 g) is rela-tively high compared to water, CE only requires 1 to 2 mL of buffer per experi-ment (≈ 5 runs). Thus, expensive buffers can be used in CE rather economically.

0

Abs

orba

nce

Time (min)

0.06

0.04

0.02

0.0

105 14.5

pH = 7.93

A

0

Abs

orba

nce

Time (min)

0.06

0.04

0.02

0.0

105 15

pD = 7.93

B0.08

Figure 1-5. Comparison of CE separations of tryptic digests of salmon calcito-nin in (A) water (pH 7.93) and (B) D2O (pD 7.93). Reproduced with permis-sion from Camilleri et al., Anal. Biochem. 198, 36 (1991).

1-12

In practice, manipulation of the pH of a buffer is a powerful tool used tocontrol the charge on a peptide (or protein) and, hence, the resolution betweenneighboring peaks in an electropherogram. This point is illustrated in Figure 1-6.

1.00

Time (min)

0.0

Buffer 1pH 2.2

Buffer 3pH 4.4

Buffer 5pH 7.5

Buffer 2pH 3.8

Buffer 4 pH 6.2

Buffer 6pH 8.3

A

C

E

B

D

F

Abs

orba

nce

UV

0.62

0.26

-0.12

-0.502.5 5.0 7.5 10.0 12.5

x 10

-1x

10-2

0.30

0

0.21

0.13

0.04

-0.062 4 6 8

x 10

-1

1.00

0

0.62

0.26

-0.12

-0.502 4 6

0 2 4 6 8

0.20

0.14

0.07

0.01

-0.06x

10-1

0.06

-0.00

-0.06

-0.10

-0.15

x 10

-1

0.20

0.14

0.07

0.01

-0.06

x 10

-1

0 2 4 6

12

3

5

4

6

12

35

4

6

123

5

4

7

1

2

3

5

4

6

12

3

5

46

1

2

3

5 4

6

0.0 2.5 5.0 7.5 10.0 12.5

Figure 1-6. CE of peptides separated with the buffers listed in Table 1-2.UV detection at 214 nm. Peak identification of the ACTH fragments: (A) 4-9;(B) 5-9; (C) 6-9; (D) 7-9; (E) 8-9; (F) 4-6. Reproduced with permission fromvan de Goor et al., J. Chromatogr. 545, 379 (1991).

In this example from Van der Goor et al. (1991), a mixture of six peptideswas separated using different buffers ranging in pH from 2.2 to 8.3. The buff-ers and some of their properties are listed in Table 1-2. In this experiment, thepeptide samples were dissolved in water prior to injection into the CE instru-ment.

1-13

Table 1-2. Operational Buffer Systems

UVBuffer pH Concentration* Conductivity EOF .10-5 AbsorbanceNo. (mS /cm) (cm2/V· s) (AU at

214 nm)

1 2.2 25 mM phosphate + KOH 4.03 < 3.2 0.0002 3.8 20 mM formate + alanine 0.95 16.4 0.00023 4.4 20 mM α−aminocaproate 0.84 29.7 0.007

+ acetic acid4 6.2 20 mM histidine + MES 0.39 52.8 0.0595 7.5 40 mM imidazole + MOPS 0.73 57.6 0.0576 8.3 100 mM borate + KOH 1.90 67.9 0.000

* MES = 2-(N-morpholino)ethanesulphonic acid;MOPS = 3-(N-morpholino)propanesulphonic acid.

The peptides were synthetic fragments derived from the hormone ACTHand ranged in size from two to six amino acids. Several points can be noted:

1. Whereas all six panels show fast separations, the choice of a buffer withrelatively high conductivity (25 mM phosphate—see panel A) results insomewhat broader peaks. This may be due to inadequate zone focusingusing the 25 mM phosphate concentration (100 mM would result insharper peaks—see the later section on “Detectability Enhancements”).Also, Joule heating in the capillary (relative to the other buffers) may playa role. In contrast, panels B and C show sharp, highly efficient peaks(theoretical plate counts of 200,000 to 300,000 are obtained).

2. Going from low to high pH, the electroosmotic flow (EOF) increases.Note: the EOF can be observed as the negative peak in the electrophero-grams; this negative peak is the result of water which is not associatedwith buffer ions and, therefore, has a decreased absorbance, migrating asa neutral species.

3. Selectivity is greatly affected by pH (e.g., note the position of peak 4relative to the other peaks).

4. The use of high-UV-absorbing buffers decreases the linear dynamic rangeof the detector and therefore leads to smaller peak heights. Note thesmaller peak heights resulting from the choice of MES/histidine andMOPS/imidazole buffers.

1-14

In another paper by Langenhuizen and Janssen (1993), the effectiveness ofdifferent buffers was studied for the separation of pharmaceutical peptides.The selected buffers were:

• 25 mM phosphoric acid, adjusted to pH 2.20 with 1 M NaOH• 20 mM formic acid, adjusted to pH 3.80 with β-alanine• 20 mM L-histidine, adjusted to pH 6.20 with MES• 50 mM Tris, adjusted to pH 7.50 with acetic acid• 100 mM boric acid, adjusted to pH 8.30 with 1 M NaOH

Adrenocorticotropic hormone (ACTH), endorphins, cholecystokinin andfragments thereof were chosen as model compounds. It was found that theoptimum buffer pH depends strongly on the pI values of the peptides. Forpeptides with an acidic character, the neutral pH region was preferred foroptimum separation. Basic and neutral peptides were best separated in thelow pH region.

1.3 Separation of Peptides in FreeSolution

In physiology and medicine, peptides play important roles as hormones orneurotransmitters, and epitopes for such receptors as major histocompatibilitycomplex (MHC) molecules. Many microorganisms produce peptides, oftenwith antibiotic activity. Some peptides are very toxic (e.g., phalloidin, a cyclicpeptide originating from a mushroom). Examples of peptides with growth-promoting activity are the streptogenins. Other peptides act as enzyme inhibi-tors (e.g., pancreatic trypsin inhibitor). A wide variety of biologically activepeptides has been isolated from plants. The analysis of synthetically madepeptide analogs is important to the biopharmaceutical industry in the develop-ment of new therapeutic agents. The analysis of peptides may also be part ofextensive protein characterization schemes. For example, this is the case inpeptide mapping applications where proteins are enzymatically or chemicallycleaved into smaller subunits and subsequently analyzed by several methodssuch as slab gel electrophoresis, HPLC, thin-layer chromatography, or CE.Peptide mapping is important in a quality control environment.

During recent years, CE in free solution, as opposed to CE in a gel matrix(see Section 1.6), has been successfully applied to peptide separations. As wewill see, many of the separation strategies used for peptides also work for

1-15

proteins. However, being more complex, proteins often require special pre-treated capillaries and conditions. This is due in part to the adsorption prob-lems encountered with proteins. MECC conditions—also considered afree-solution technique—generally work better for peptides than for proteins,although protein applications have been reported. The CGE and IEF tech-niques, on the other hand, are almost exclusively used with proteins.

1.3.1 Types of CapillariesIn the vast majority of peptide separations by CE, uncoated, fused-silica capil-laries may be employed. For peptides, coated capillaries are used less than inconjunction with proteins but may be useful in some cases. Coated capillariessuitable for protein and peptide separations (e.g., the eCAP™ Neutral andAmine capillaries) will be discussed in more detail in Section 1.4. Generally,it is a good idea to pretreat the uncoated capillary with sodium hydroxide(≈ 0.1 M) followed by rinsing the capillary with buffer and/or water before theactual run and repeat this rinsing procedure after each or several runs. Therinsing of the capillary with a strong base removes adsorbed contaminants.These contaminants may influence the magnitude of the EOF—and, therefore,run-to-run reproducibility— during subsequent runs. On the capillary wallsurface, siloxane bonds are hydrolyzed to free silanol groups, the number ofwhich determines the charge on the capillary wall and the magnitude of EOFat the existing buffer pH. Good run-to-run reproducibility is generally ob-tained by rinsing the capillary with a consistent rinse protocol (either after eachrun or after a number of runs); this ensures that—if changed—the EOF alwaysreturns to a certain, constant value at the start of a new run.

1.3.2 Use of Buffer Additives to OptimizePeptide Separations

In addition to pH control and buffer selection, one can use a number of bufferadditives to optimize selectivity and fine-tune a separation. The reviews bySchwartz et al. (1993), Palmieri and Nolan (1994), and McCormick (1994)provide additional information on this subject.

1.3.2.1 Ionic Surfactants (e.g., SDS or CTAB)

An effective way to achieve better selectivity—especially for neutral solutes—is the addition of micelle-forming reagents to the buffer. These can be eitheranionic (e.g., SDS) or cationic (e.g., CTAB). The resulting separations, knownas micellar electrokinetic capillary chromatography (MECC—see Terabe,1989), resemble reversed-phase HPLC in that the analytes partition between a

1-16

mobile phase (i.e., the background electrolyte) and a pseudo-stationary phase(the “micellar phase,” e.g., SDS, which migrates against the EOF). In MECC,the detergent is added to the buffer above its critical micelle concentration(e.g., with SDS, generally a concentration of 50 to 100 mM is used.) Note:with cationic surfactants such as CTAB, the polarity of the power supply mustbe reversed because the EOF changes direction. This occurs because thesesurfactants bind to the negatively charged silanol sites on the fused-silica sur-face and effectively changes the charge on the wall from negative to positive.Another Beckman CE primer authored by Terabe (P/N 266924) describes theMECC technique in more detail.

1.3.2.2 Non-Ionic Surfactants

This strategy is often useful when the above-described MECC approach fails(e.g., when peptides with subtle differences in hydrophobicity are being sepa-rated). The non-ionic surfactant may provide a better balance between theelectrostatic and hydrophobic forces controlling the separation. For example,in work done in our lab on the separation of analogs of growth hormone releas-ing peptide (all consisting of six amino acids), two analogs which had the samemass-to-charge ratio where separated by adding 20 mM polyoxyethylene-10(Sigma Co.) to the buffer.

1.3.2.3 Ion-Pairing Reagents

Short-chain, ion-pairing reagents (e.g., hexanesulfonic acid (HSA)) have beenused in HPLC for protein and peptide separations. Research in our lab indi-cates that this reagent is also particularly effective in CE for hydrophobic pep-tides that are difficult to separate. For example, a 30 mM sodium phosphatebuffer, pH 2.5, with 100 mM HSA buffer was used for the separation of twoproprietary synthetic peptides (Figure 1-7). The mechanism by which resolu-tion enhancement occurs is by a hydrophobic pairing between the short alkylchain of the sulfonate and hydrophobic surfaces on the peptide at HSA concen-trations below the critical micelle concentration. This results in an increase innegative charge on the peptide surface and a corresponding increase in migra-tion time. Improvement in resolution is seen as the peptide-associated HSAinduces a repulsion between peptides that would otherwise be attracted byhydrophobic forces. Yet another mechanism may be at work, in that the HSAcan also ion-pair as in the traditional HPLC application, decreasing the surfacepositive charge on the peptides and any wall interaction with the negativelycharged capillary surface. Both hydrophobic forces and ion pairing may be atwork, since either association would demonstrate the experimentally observedresult of increased resolution and migration times.

1-17

0.4

2

0.3

0.2

0.1

06 8 10 12

0.4

2

0.3

0.2

0.1

06 8 10 12

Time (min)

Abs

orba

nce

(200

nm

)

Time (min)A

bsor

banc

e (2

00 n

m)

with HSA

without HSA

with HSA

without HSA

A B

Figure 1-7. Comparison of separations of two proprietary synthetic peptides(A) with and (B) without 100 mM hexanesulfonic acid added to the buffer.A 30 mM sodium phosphate, pH 2.5 buffer was used. Reproduced with permis-sion from McLaughlin et al., J. Liq. Chromatogr. 15, 961 (1992).

1.3.2.4 Cyclodextrins

It has been mentioned (Novotny et al., 1990) that addition of cyclodextrinssignificantly enhances sensitivity of the fluorescence detection of derivatizedamino acids and peptides. These additives improve resolution for peptideseparations—presumably due to host–guest interactions with the cyclodextrincavities). The use of these additives was first explored in HPLC (and later inCE by Terabe, 1989) for the separation of chiral substances such as pharma-ceuticals.

1.3.2.5 Organic Solvents

Small amounts of organic solvents (approximately 0 to 30% methanol or ac-etonitrile, 1 to 2% THF) can be added to the buffer. This is a well-known prac-tice in MECC when dealing with small molecules and is often used to increaseanalyte solubility in the buffer. Adding organic solvents to the buffer causes adecrease in the EOF due to a decreased zeta potential, resulting in a lowercurrent and less Joule heat generation. In Figure 1-8, the MECC separation ofa mixture of enkephalins is shown with and without 5% acetonitrile. A muchbetter separation is obtained using acetonitrile as a buffer additive.

1-18

0

Abs

orba

nce

(200

nm

)

Time (min)

15

A

5 10 3020 25 35 40

1

2

3

54

6

0

Abs

orba

nce

(200

nm

)

Time (min)

15

B

5 10 3020 25 35 40

1 2354 6 7

Figure 1-8. MECC of enkephalin analogs (A) without and (B) with the addi-tion of 5% acetonitrile to the buffer. Capillary, 65 cm × 75 µm, 100 mM so-dium borate, 100 mM SDS, pH 8.5, 15 kV, 25°C, 200 nm. Peak identification:(1) metsulfoxide enkephalin; (2) methionine enkephalin; (3) [ala2] methionineenkephalin; (4) leucine enkephalin; (5) leucine enkephalin amide; (6) leucineenkephalin-arg; (7) proenkephalin.

1-19

1.3.2.6 Divalent Amines

While adsorption on the capillary wall is usually not a problem with smallpeptides, larger peptides and proteins can be irreversibly adsorbed on the capil-lary wall. This is one of the major problems encountered with CE of proteins(see also the following Section 1.4 on proteins). Efforts to reduce adsorptionhave been aimed at modifying the interfacial double layer at the wall (i.e., tomodify its zeta potential). This can be done by adding small amounts (0 to5 mM) of cationic, divalent amines to the buffer. 1,4-diaminobutane (DAB)(see Lauer and McManigill, 1986; Stover et al., 1989), 1,5-diaminopentane(DAP), and morpholine (Nielsen et al., 1989) have been used for this purpose.Much larger amounts of amine (30 to 60 mM) are useful for protein separa-tions (Bullock and Yuan, 1991). Figure 1-9 shows the effect of adding 5 mMDAP to the buffer, resulting in better selectivity and sharper peptide peaks.

0

Abs

orba

nce

Time (min)

A

10 20-0.002

0.018

No DAP

2B

0

Abs

orba

nce

Time (min)

10 20-0.002

0.01810

8

1

4

With 5 mM DAP

Figure 1-9. Separation of a mixture of five synthetic nonapeptides. (A) WithoutDAP added to the buffer, four peaks can be discerned. (B) With 5 mM DAPadded to the buffer, all five peaks are resolved. Peak identification, see Fieldet al., Beckman Application Data Sheet DS-791.

1-20

1.3.2.7 Complexing Reagents

Cohen et al. (1987) found that metal ions enhance the resolution of nucleicacids and amino acids in MECC buffers. In a similar fashion, 0 to 30 mM zincperchlorate has been used as a complexing reagent for histidine-containingpeptides (Stover et al., 1989). Metal ions originating from zinc or copper saltscan interact with certain N, O, or S atoms on proteins or peptides. This causesthe mobility of these species to decrease relative to species in which no com-plexation takes place. On the other hand, trace amounts of EDTA are some-times used to complex undesirable trace metals during electrophoresis (Ludiet al., 1988).

1.4 Separation of Proteins in FreeSolution

1.4.1 AdsorptionProteins have the unfortunate property of sticking to many different surfaces,including metals, plastics, and glass. In chromatography, concerns regardingthe adsorption of proteins on packing materials or ancillary equipment have ledto the development of “biocompatible” instrumentation. One of the main prob-lems associated with protein separations using CE on untreated fused-silicacapillaries is adsorption by the charged sites of proteins on fixed, negativelycharged sites (silanol groups) on the capillary wall. This process leads to bandbroadening and results in far lower actual plate numbers (a measure of theefficiency of the separation) than would be expected on the basis of theory.The electrostatic interaction of proteins with the wall is schematically depictedin Figure 1-10.

1-21

migration

capillary wall capillary wall

PROTEIN ka kd

Figure 1-10. Schematic of electrostatic interactions of proteins with the nega-tively charged capillary wall. The adsorption-desorption process affects peakwidth as migration occurs. Reproduced with permission from Novotny et al.,Electrophoresis 11, 735 (1990).

The adsorption and desorption processes have specific rate constants(ka and kd, respectively) which ultimately affect the band width of the analytezone as migration progresses. Protein adsorption may also occur throughhydrophobic forces if the stability of the protein in solution is affected bychanges in its environment (e.g., by changes in pH or temperature). In solution,proteins are held together by H-bonding and hydrophobic forces. Whenunfolding occurs, hydrophobic areas may be exposed, resulting in aggregationor hydrophobic adsorption. Reagents such as urea or guanidine HCl are oftenused by protein chemists to force this process of denaturation. β-mercaptoethanolis used to break the disulfide bonds which hold the polypeptide chain together,as illustrated for ribonuclease in Figure 1-11. The following sections deal withstrategies to prevent undesirable protein adsorption on the capillary wallsurface when working with CE in the free-solution mode.

��

��

��

��

��

��

��

��

��

8 M urea andβ-mercaptoethanol

SHHS

1

26

4058

65

72

84

95

110

SH

HSHS

HS HS

HS126

40 58

65

72

84

95 110

Denatured reduced ribonucleaseNative ribonuclease

��

Figure 1-11. Denaturing of ribonuclease by treatment with β-mercaptoethanolin 8 M urea.

1-22

1.4.2 Strategies to Prevent Protein Adsorption

1.4.2.1 High- and Low-pH CE Buffer Conditions

Several investigators (Lauer and McManigill, 1986; McCormick, 1988) havetried to overcome the electrostatic adsorption of proteins on the capillary wallsimply by working at extremes of pH. Under these conditions, the silanolgroups on the wall surface are either negatively charged or neutral. In solution,proteins exhibit acid–base behavior and their net charge is dependent on thepH of the buffer. The net charge is zero at the characteristic isoelectric point(where the number of positive charges is balanced precisely by the numberof negative charges). This is illustrated in Figure 1-12 for a protein with apI of 7.0.

Increasing Positive

Charge

Increasing Negative

Charge

pI = 7.0

No Net Charge

pH0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

+ -

+

+

+ - -

-+

+

+

+

+ +-

- -

-

--

Figure 1-12. Schematic of the effect of buffer pH on the net protein charge.At pH 7.0 (in this case, the pI of the model protein), there is no net charge onthe protein. Increasing net positive and net negative charge is obtained bydecreasing and increasing the pH of the buffer, respectively.

At high pH (e.g., pH approximately 10), the pI of the protein typically isless than the pH of the buffer (an exception being a very basic protein). There-fore, both the protein and the capillary wall are negatively charged, and theadsorption process is minimized as a result of a charge repulsion effect. Notethat, at high pH, a relatively high EOF is generated. Theoretically, resolutionis improved when the EOF is suppressed and balanced against the electro-phoretic migration (e.g., by means of buffer additives—see Section 1.4.2.3).

This is also the case when working at the other extreme with low pHconditions (e.g., pH approximately 2). In this situation, the capillary wall is

1-23

protonated while the proteins are positively charged, again minimizing electro-static interactions. However, while these approaches are useful in some cases,working at the pH extremes is not always feasible for the following reasons:

• It precludes the study of many native molecular conformation interactions.For example, at the pH extremes, unfolding and/or aggregation of proteinsmay occur which could lead to the undesirable appearance of multipleand/or broad peaks. In addition, at high pH, deamidation or peptide bondscission may occur.

• The biological activity of the protein may be very different under extremepH conditions, precluding enzymatic assays.

• At very high pH (greater than 11), dissolution of the silica (wall material)becomes an issue.

• The electrophoretic mobility of proteins tends to be very sensitive to cer-tain pH regions. The largest number of titratable groups occurs in theregion of pH < 6 or pH > 9.

As a result of the enormous diversity in proteins, a wide range of runconditions at varying pH values needs to be available.

Two basic strategies have been applied in CE to address the above issues:(1) a “static” approach through permanent modification of the capillary wallby suitable coatings, and (2) a “dynamic” approach using certain buffer addi-tives. In the latter, the capillary wall is dynamically modified each run bymasking the charged sites.

1.4.2.2 Permanent Capillary Coatings

As CE is maturing, coated capillaries specifically designed for protein separa-tions are becoming more available. A recent survey by Majors (1994) de-scribes the latest introductions of companies at the 1994 PittsburghConference. Wehr (1993), Swerdberg (1994), and Guzman’s book (1993,several chapters) have reviewed column technology for CE. Some of the majorsuppliers of CE instrumentation, including Beckman, also sell coated capillar-ies. However, some of these capillaries are designed to fit only in certain in-struments, so it advisable to check first if these capillaries will fit directly in aP/ACE system. Specialty companies such as J & W Scientific, Supelco, Scien-tific Glass Engineering, Scientific Resources, and MetaChem Technologiesalso provide coated capillaries. Beckman has introduced several coated capil-laries which are available either by themselves or as part of a method develop-ment kit. In general, the strategy is to modify the capillary wall by attachinghydrophilic polymers to the Si-OH sites on the silica surface, either directly orthrough suitable spacers. Hydrophobic coatings, such as those frequently used

1-24

in GC, are less suitable for protein CE separations. A sampling of coatingprocedures is summarized in Table 1-3. In most cases, the capillary is firstdeactivated by means of a silanization reagent and is then coated with a func-tional group.

Table 1-3. Some Capillary Coatings for Protein Separations

Coating Type Investigator Reference

Methylcellulose Hjerten J. Chromatogr. 347, 191 (1985)

Polyacrylamide (through Si-O-Si-C bonds) Hjerten J. Chromatogr. 347, 191 (1985)

Polyacrylamide (through Si-C bonds) Novotny Anal. Chem. 62, 2478(1990)

3-glycidoxypropyltrimethoxysilane Jorgenson Science 222, 266(1983)

Epoxy-diol, maltose Poppe J. Chromatogr. 480, 339 (1989)

Polyethyleneglycol Poppe J. Chromatogr. 471, 429 (1988)

Polyvinylpyrrolidone McCormick Anal. Chem. 60, 2322(1988)

Aryl-pentafluoro(aminopropyl- Swerdberg Anal. Biochem. 185, trimethoxy)silane 51 (1990)

α-lactalbumin Swerdberg J. High Res.Chromatogr. 14, 65(1991)

Polyether El Rassi J. Chromatogr. 559,367 (1991)

Hydrophilic, C1, C8, C18 Dougherty Supelco Reporter,Vol. X, No. 3 (1991)

Ion exchangers, polyacrylamide EngelhardtJ. Microcol. Sep. 3,491 (1991)

Neutral, hydrophilic Karger J. Chromatogr. 652,149 (1993)

Hydrophilic & hydrophilic polymers Lee Anal. Chem. 65, 2747(1993)

1-25

It is interesting to mention that, as early as 1967, Hjerten demonstrated theutility of polyacrylamide or methylcellulose polymer coatings for glass tubesin free-solution electrophoresis. In 1985, Hjerten and Zhu described a capillarysurface modification which has been frequently cited in the CE literature.Before the advent of commercial CE coatings, the “Hjerten” coating was themost widely used among practitioners of CE. A bifunctional silane was used tofirst derivatize the surface, after which a hydrophilic polymer (polyacrylamide)was covalently attached. This type of coating minimizes solute adsorption andstrongly suppresses the EOF (the negative charge on the wall is “neutralized”).However, an often-heard complaint was that these home-made capillaries werenot stable in the long run and, therefore, would not yield reproducible migra-tion times.

Beckman recently introduced a similar type of neutral coating (eCAPNeutral Capillary, P/N 477441) designed for protein separations in the pH 3 to8 range (use of solutions with pHs outside this range may damage the capil-lary). This capillary also should prove useful for certain peptide separations.A Neutral Capillary Method Development Kit (P/N 477445) includes threebuffers designed for use in the normal- or reversed-polarity mode. In applica-tions requiring normal polarity, the citrate, pH 3.0 buffer, and the citrate/MES,pH 6.0 buffer are most useful for proteins with pIs > 4.0 and 6.7, respectively.The tricine, pH 8.0 buffer is designed for applications in the reversed-polaritymode. Note that, with this type of coating, migration is mainly by electro-phoretic flow as the EOF is minimized (≈ 5% of the EOF determined for un-treated fused silica). The excellent stability of these neutral capillaries isdemonstrated in Figures 1-13 and 1-14 for acidic and basic proteins, respec-tively. The Neutral Capillary is also suitable for CIEF applications (see Sec-tions 1.7, 2.2, and 2.3).

1-26

0.030

0.025

0.020

0.015

0.010

0.005

0.000

-0.005

Abs

orba

nce

Run 1

α-La

c

β-La

c A

Car

boni

c A

nhyd

rase

pI 5

.4

Car

boni

c A

nhyd

rase

II p

I 5.9

0.025

0.020

0.015

0.010

0.005

0.000

-0.005

Abs

orba

nce

Run 120

0 2 4 6 8 10Minutes

Figure 1-13. e-CAP Neutral Capillary stability. 1st and 120th run of acidicproteins (% RSD for absolute migration time). β-lactoglobulin (2.99), α-lacto-globulin (0.45), carbonic anhydrase II, pI 5.4 (1.00) and pI 5.9 (1.30).

1-27

1st injection0.025

0.020

0.015

0.010

0.005

0.000

-0.005

Abs

orba

nce

(214

nm

)

240th injection

0 20Minutes

1

2

3

4

1

2

3

4

5 10 15

0.030

0.025

0.020

0.015

0.010

0.005

0.000

-0.005

Abs

orba

nce

(214

nm

)

1 = Reference Marker2 = Lysozyme3 = Cytochrome C4 = Ribonuclease A

Figure 1-14. e-CAP Neutral Capillary stability. 1st and 120th run of basicproteins (% RSD for absolute migration time). Lysozyme (0.47), cytochrome C(0.51), myoglobin (1.05), ribonuclease A (0.96).

A second type of coating (eCAP Amine Capillary, P/N 477431) availablefrom Beckman utilizes a polyamine-modified surface. A strong cationic chargeis created on the capillary wall and, consequently, the EOF is reversed. Basicanalytes should be repelled from the wall surface; hence, adsorption is mini-mized for cationic peptides, proteins, and other positively charged analytes.This capillary must be operated in the reversed-polarity mode as, otherwise,the EOF is in the direction away from the detector. Figure 1-15 shows thereversed order of elution when the amine capillary is used instead of the un-treated capillary. In addition, the analysis time on the coated capillary is muchshorter than with the untreated capillary. In the Beckman Amine Capillary Kit(P/N 477430), three buffers are supplied which will yield relatively fast runtimes: acetate, pH 4.5; MES, pH 6.0; and Tris pH 8.0. Also included are twophosphate buffers (pH 2.5 and 7.0) which will reduce the positive surfacecharge on the capillary wall through ion pairing, and consequently, the EOF.

1-28

The effect of the phosphate concentration (varied from 0 to 300 mM) on themobility of a neutral marker (benzyl alcohol) is demonstrated in Figure 1-16.Typically with the phosphate buffers, peak efficiency and resolution are supe-rior to those obtainable with non-phosphate buffers. The reduced surfacecharge on the wall is especially advantageous for acidic analytes, as electro-static interaction will be minimized.

0.030

0.025

0.020

0.015

0.010

0.005

0.000

0 2 4 6 8 10 12

Time (min)

Abs

orba

nce

2

13

A

0.025

0.020

0.015

0.010

0.005

0.000

0 1

Time (min)

Abs

orba

nce

2 3 4 5

2

1

3

4

B

Figure 1-15. (A) Separation of (1) lysozyme, (2) cytochrome C, and (3) ribo-nuclease A on fused-silica capillary. (B) the eCAP Amine Capillary improvesthe resolution, speed of separation, and reverses the migration order. Benzylalcohol (4) is a marker.

1-29

700

600

500

400

300

200

100

0

0 100 200 300

(Phosphate), mM

Mob

ility

Figure 1-16. Mobility vs. phosphate buffer concentration

Whereas coated capillaries may dramatically reduce protein adsorption,it still may be necessary to clean the capillary in between runs. This is neces-sary when severe peak broadening occurs and/or poor precision is observed.The extent of capillary cleanup is dependent on the type and number ofsamples exposed to the capillary. With the Neutral capillary, a recommendedregeneration procedure involves rinsing with 0.1 N HCl (30 to 60 seconds)followed by a 1.5-minute rinse with run buffer. For the Amine capillary, rins-ing with 1 N NaOH is recommended in between runs. However, with “dirty”samples, a rinse procedure consisting of 1 N HCl, 1 N NaOH, and AmineRegenerator solution (5 minutes each) may be required.

1.4.2.3 Dynamic Capillary Coating: Buffer Additives

The dynamic approach for capillary coating involves the use of buffer addi-tives. A number of ways have been suggested to reduce protein adsorption onthe capillary wall, conceivably by interfering with the protein–wall ion-ex-change mechanism.

1.4.2.3.1 High-Salt or High-Ionic-Strength Buffers

Lauer and McManigill found that a relatively large amount of salt (i.e., 0.25 MK2SO4) led to improved separation efficiency. Under these conditions, saltcompetes with protein for adsorption sites. An issue of concern with this ap-

1-30

proach is the absorptivity (purity) of high salt buffers in the low UV region.A variation of this method was applied by Chen (Beckman, Fullerton, CA)who used 0.5 M sodium phosphate buffers, pH 5 to 10, for the separation ofprotein standards in the pI 5.2 to 10.5 range. Short, small-i.d. (25-µm) capillar-ies were used, resulting in fast, efficient separations. With standard-size capil-laries (75 to 100 µm i.d.), the use of high-ionic-strength buffers may result inintolerable Joule heat generation. (In Figure 1-9, for example, the 75µm i.d.capillary could not be used with buffer concentrations exceeding 0.125 M.)However, 25 µm i.d. capillaries dissipate heat much more efficiently and,therefore, are preferable with high-ionic-strength buffers. The 0.5 M sodiumphosphate buffers are UV transparent and can be used with low-UV detection.An example of this separation system applied to milk analysis is shown inFigure 1-17. Urea was added to the buffer to prevent aggregation of thecaseins.

0.030

0.020

0.010

0.0001 2 3 4

Time (min)

Abs

orba

nce

(200

nm

)

DMF

1

2

3

4

Figure 1-17. CE of nonfat milk on a 23 cm × 21 µm capillary. Buffer,0.5 M sodium phosphate, 4 M urea, pH 7.0. Peak identification: (1) β-casein;(2) α-lactalbumin; (3) α-casein and β-lactoglobulin B; (4) β-lactoglobulin A.Dimethylformamide (DMF) was added as an EOF marker. From Chen, Beck-man Application Data Sheet DS-818 (1991).

1-31

Phosphate buffers dynamically modify the capillary wall by convertingresidual silanols on the capillary surface to phosphate complexes that are moreeasily protonated (McCormick, 1988). The resulting dynamically coated sur-face exhibits a different selectivity than the untreated fused-silica capillary.

1.4.2.3.2 Zwitterionic Salts

Instead of using ionic salts, Bushey and Jorgensen (1989) proposed the use ofzwitterionic salts as a buffer additive. Zwitterions such as betaine, sarcosine,and triglycine do not contribute to the conductivity and, consequently, can beused at relatively high concentrations (up to 2 M). A similar approach wastaken by workers at Waters Associates, who used n-propyl(trimethyl)ammo-nium sulfate in concentrations up to 1 M. The quaternary ammonium function-ality of the zwitterions interacts with the negatively charged silanol groups onthe surface; they contain sulfate and carboxyl groups as the anionic compo-nents. Zwitterionic salts are effective over a wide pH range and are limited bythe pKas of the titratable groups. In many cases, however, the initial concentra-tion must be in excess of 1 M to be effective. Alternatively, primary amino-phosphoryl reagents exert their effect at lower concentrations and provideevidence of increased ion-pair stability when compared to phosphate buffersalone (see Chen, F. T. et al., 1992). This may be due to the increased hydrogenbonding potential for these additives compared to the quaternary ammoniumcompounds.

In Figure 1-18, 250-mM O-phosphorylethanolamine was used as thezwitterionic additive to separate five protein standards. This reagent is widelyavailable, low in cost, and UV transparent. It can be seen that, with the addi-tive, considerably better peak shape is obtained, particularly for β-lactoglobu-lin A and myoglobin. It is apparent from the two pKas (i.e., 5.8 and 9.4) thatthe effective working range is limited to pHs between 6 and 9.

1-32

Cha

nnel

A: A

bsor

banc

e

0.000

0.000

0.000

0.000

5 10 15 20 25 30

Cha

nnel

B: A

bsor

banc

e

0.000

0.000

0.000

0.000

Time (min)

1

2

34

5

6

1

2

4

3

5 6

Abs

orba

nce

(200

nm

)

A

B

Figure 1-18. Effect of zwitterionic additive (O-phosphorylethanolamine) onseparation efficiency. (A) run buffer, 250 mM sodium phosphate, pH 6.0;(B) 100 mM sodium phosphate, plus phosphorylethanolamine, pH 5.8. Peakidentification: (1) lysozyme; (2) cytochrome C; (3) EOF masker; (4) myoglo-bin; (5) carbonic anhydrase; (6) β-lactoglobulin A. Reproduction with permis-sion from Chen et al., J. Liq. Chromatogr. 15, 1143 (1992).

1.4.2.3.3 Divalent, Cationic Amines

As mentioned earlier regarding peptide separations (Figure 1-9), divalentamines suppress solute–capillary wall interactions. For protein separations,1,4-diaminobutane, 1,5-diaminopentane, and 1,3-diaminopropane have beenused for this purpose. The latter additive (in relatively large quantities of 30 to60 mM) appeared particularly suitable in conjunction with moderate amountsof salt. Figure 1-19 shows the effect of varying amounts of 1,3-diaminopro-pane on the resolution of basic proteins in a run buffer of pH 7.0. Increasingthe amine concentration in the buffer results in longer separations as the EOFbecomes increasingly suppressed.

This method is well suited for basic proteins. Acidic proteins (pI 5.4to 7.4) showed considerable band broadening, however.

1-33

100 20 30 40

A 12, 3

4, 5

6

100 20 30 40

B1

2

4, 56

100 20 30 40

C 1

2 4

6

100 20 30 40

D1

2

4

6

3

3

5

5

3

10 mM

20 mM

30 mM

50 mM

25 m

V25

mV

16 m

V13

mV

Figure 1-19. Effect of adding 1,3-diaminopropane on peak shape and resolu-tion in a pH 7.0 phosphate buffer. Peak identification: (1) lysozyme; (2) cyto-chrome C; (3) ribonuclease A; (4) trypsinogen; (5) α-chymotrypsinogen A;(6) rhuIL-4. Reproduced with permission from Bullock and Yuan, J. Microcol.Sep. 3, 241 (1991).

1-34

1.4.2.3.4 Non-Ionic Surfactants

Hydrophobic proteins are often difficult to analyze by CE due to their interac-tion with the (coated) capillary wall. Towns and Regnier (1991) showed thatnon-ionic surfactants such as Brij35 and Tween20 can be used to dynamicallycoat deactivated (bonded-phase) capillaries, thus allowing high-efficiencyseparations for hydrophobic proteins. At Supelco, Dougherty et al. used asimilar approach with their hydrophobically bonded “CElect” capillaries andthe non-ionic surfactant Brij35. The detergent coating makes the capillary wallsurface less hydrophobic and, therefore, more suitable for hydrophobic proteinseparations.

Typically, the procedure of preparing the capillary involves rinsing thebonded-phase capillary (by pressure) for approximately 2 hours with aqueousbuffer containing 5% of the desired detergent (e.g., Brij35); 0.001% detergentis added to the run buffer during the CE runs. The latter helps to maintain aconstant detergent concentration on the capillary wall surface. Figure 1-20shows the separation of three hydrophobic proteins—lysozyme, cytochrome C,and ribonuclease A—analyzed on a CElect C18 capillary dynamically coatedwith Brij35.

0.0400

0.0300

0.0200

0.0100

0.0000

0 20 40 60

Time (min)

Abs

orba

nce

1

2 3

Figure 1-20. Separation of basic proteins using a CElect-H2 (C18) capillarycoated with Brij35. Run buffer: 10 mM sodium phosphate, pH 7.0, 0.001%Brij35. Field strength: 300 V/cm. Peak identification: (1) lysozyme; (2) cyto-chrome C; (3) ribonuclease A. Figure reproduced with permission from AnnDougherty, Supelco, Bellefonte, PA.

1-35

1.4.2.3.5 Charge Reversal Reagents

Another way to reduce adsorption of cationic proteins is to add to the buffercertain surfactants or polymers which reverse the charge on the wall fromnegative to positive (note: this is also the case with the Amine capillary fromBeckman except here a permanent coating is created). Consequently, the EOFis also reversed which necessitates reversal of the power supply. Figure 1-21from Emmer et al. (1991) depicts this situation for a fluorocarbon surfactant,FC-134 (3M Company), and is thought to involve a bilayer formation of hy-drophobic chains at the wall.

A

B

C

Electroosmotic Flow

No Flow

Electroosmotic Flow

+-

+-

+-

Figure 1-21. Schematic of the charge reversal process at the capillary wall.(A) No surfactant added. (B) Electrostatic interaction of the positively chargedsurfactant headgroup to the negatively charged silanol sites of the capillarywall. (C) Admicellar bilayer formation by hydrophobic interaction between thenonpolar chains, resulting in a reversal of the electroosmotic flow. Repro-duced with permission from Emmer et al., J. Chromatogr. 547, 544 (1991).

Due to the hydrophobic behavior of the fluorocarbon chain, interactionwith proteins is minimized. High-efficiency separations were obtained in arelatively low-ionic-strength buffer (10 mM phosphate, pH 7) as illustrated inFigure 1-22. With this approach, high-efficiency separations at a neutral pHare feasible without creating high current conditions which may adverselyaffect separation performance.

1-36

A similar approach, involving charge reversal through a cationic, poly-meric surfactant was investigated by Wiktorowicz and Colburn (1990) and iscommercially available as MicroCoat from Applied Biosystems. For dynamiccoating of the capillary wall, the additive solution is rinsed through the capil-lary prior to analysis. Subsequently, excess surfactant is removed by rinsingwith run buffer (periodic replenishment is required).

1

23

4

10 12 14 16 18 20

Time (min)

Figure 1-22. Electropherogram of four model proteins using the charge rever-sal system depicted in Figure 1-21. Buffer: 0.01 M phosphate, pH 7.0 with50µg/mL FC-134 added. Peak identification: (1) myoglobin (run separately);(2) ribonuclease A; (3) cytochrome C; (4) lysozyme. Reproduced with permis-sion from Emmer et al., J. Chromatogr. 547, 544 (1991).

1.4.2.3.6 EOF and/or Adsorption Suppressors

Small amounts of ethylene glycol or cellulose derivatives, e.g., hydroxypro-pylmethylcellulose (HPMC), have been used by several research groups tosignificantly increase the resolution of protein separations. From previousstudies in isotachophoresis (Everaerts et al., 1976), it is known that low con-centrations (0.01 to 0.03%) of these additives decrease the zeta potential at thefused-silica wall and may assist (by shielding) in preventing undesirable ad-sorption. In one CE report (Gordon et al., 1991), ethylene glycol was effectiveby adding it to the sample (as opposed to adding it to the buffer). Higher con-

1-37

centrations of these additives (e.g., 0.5% methylcellulose or 10 to 30% ethyl-ene glycol) substantially increase the viscosity of the buffer and cause a mo-lecular sieving effect through the formation of an entangled polymer network(see Section 1.6). Figure 1-23 from Lindner et al. (1992) shows the utility ofapplying 0.03% HPMC to the separation of rat liver core histone proteins.Histones are a class of complex, basic proteins (pI 11 to 12) which show con-siderable microheterogeneity.

0.025

0.020

0.015

0.010

0.005

0.000

13.0 13.5 14.0 14.5 15.0

H3.3

H3.2+

H3.3

H3.2+

H2A.1

H2A.1

H2A.1

H2B

H4

H4

H3.3

H3.2+

Time (min)

Abs

orba

nce

(200

nm

)

Figure 1-23. Separation of rat liver core histones. Buffer: 110 mM phosphate,pH 2.0, with 0.03% HPMC added. Histone peaks are labeled as indicated inthe figure. Reproduced with permission from Lindner et al., J. Biochem. 283,467-471 (1992).

1.4.3 Protein Modification: Use of Ionic Surfactantsand Urea

1.4.3.1 Ionic Surfactants—MECC Conditions

The anionic surfactant SDS can bind to proteins in a variety of modes. SDSbinds to proteins in a constant ratio of 1.4 g detergent per gram of protein,yielding a rod-shaped, SDS–protein micelle complex. All molecules, intheory, possess the same charge-to-mass ratios and can be separated according

1-38

to relative molecular weight by SDS-polyacrylamide gel electrophoresis(see Section 1.6). However, differences in SDS saturation levels may occur asa consequence of disulfide bond formation, amino acid composition, and gly-cosylation. Nolan and Palmieri (1994) exploited this phenomenon in CE forthe separation of a glycosylated protein (avidin) and a nonglycosylated protein(streptavidin). In this example (Figure 1-24), SDS was added to both thesample (1%) and the run buffer (0.1%). By contrast, this protein pair could notbe resolved in a simple borate buffer. Note that differences in mobility resultfrom different levels of SDS saturation.

Abs

orba

nce

0.030

0.025

0.020

0.015

0.010

0.005

0

-0.005 2 4 6 8 10 12 14

Time (min)

Abs

orba

nce

0.030

0.025

0.020

0.015

0.010

0.005

0

2

2

1

4 6 8 10 12 13 14-0.005 1 3 5 7 9 11

Time (min)

A

B

Figure 1-24. SDS-containing run buffer for the separation of glycosylatedand nonglycosylated proteins. Capillary, 50-µm i.d. × 27-cm length; wave-length, 200 nm; proteins, streptavidin (peak 1) and avidin (peak 2) stocksmade 1 mg/mL each in 25% phosphate-buffered saline (PBS). (A) Run buffer,500 mM sodium borate, pH 8.5; sample preparation, diluted 1:3 in waterfor analysis; voltage, 0- to 10-kV linear ramp over 20 min, normal polarity(toward cathode). (B) Run buffer, 200 mM sodium phosphate, pH 2.0, 0.1%SDS; sample preparation, stocks diluted 1:3 in 1% SDS, 5% 2-mercapto-ethanol, 2.5 mM Tris-glycine (pH 8.9), 10% glycerol, and boiled for 5 min;voltage, 10 kV, constant, reverse polarity (toward anode). Reprinted withpermission from Palmieri and Nolan, Handbook of Capillary Electrophoresis,Landers (Ed.), Boca Raton: CRC Press, 1994.

1-39

SDS can also be used to break up protein–carbohydrate–lipid complexes.This is shown in Figure 1-25 where 50% methanol extracts of wheat grainswere separated with 1% SDS (approximately 35 mM) in the borate buffer.In this application, the electrophoretic profile obtained from an extract or com-plex sample is more important than knowing precisely the identity of all spe-cies in that sample. The electropherogram shows that three distinctly differentpatterns are obtained for the three varieties of wheat. The run buffer employed,rich in acetonitrile (20%), was most effective in the resolution of the manyprotein peaks resulting from the methanolic extraction of the wheat kernels.

It is interesting to mention that, strictly speaking, the separation mecha-nism with protein–SDS free-solution electrophoresis is not MECC, eventhough the SDS concentration in the buffer may exceed the critical micelleconcentration (8 mM for SDS). The large SDS–protein complexes do notpartition in the micelles as do smaller molecules. However, in the literature,protein separations with SDS buffers are often classified as MECC separations(Vinther et al., 1992; Arentoft et al., 1993).

Alternatively, SDS may interact with proteins in limited amounts. SDSmay bind and ion-pair, bind and modify the net charge, or bind and modify thetertiary structure (or activity). In most cases, these interactions tend to be spe-cific and, depending on the experimental conditions, may or may not alter theexpected mobility. Since SDS binds in a reversible manner, it is important toadjust the concentration of this additive in order to maintain equilibrium duringseparation. Typically, concentrations of 0.5 to 10 mM SDS are used and theoptimum level is determined by titration. For example, Harrington et al. (1991)used 0.5 mM SDS in the borate run buffer to characterize enzyme–antibodyconjugates. Plasma apolipoproteins (HDL, LDL) were characterized by Tadeyand Purdy (1993) using a pH 8.3 borate buffer to which 0.1% SDS was added.Strege and Lagu (1993) separated model proteins with coated capillaries in thepresence of anionic (SDS, 0.3%) or cationic (CTAC, 0.1%) surfactants.Thus, as shown in these examples, differential SDS binding may be exploitedin CE to separate proteins by increasing differences in the charge-to-massratio.

1-40

0.040

0.030

0.020

0.010

0.000

Abs

orba

nce

0.100

0.050

0.000

Abs

orba

nce

Abs

orba

nce

0.100

0.000

Time (min)

4 5 10 15 16

A

B

C

Figure 1-25. Profiling of wheat varieties. Capillary: 57 (50) cm × 50 µm;30°C; 200 nm; buffer: 0.06 M sodium borate, pH 9.0, 1% SDS, 20% acetoni-trile added. Three wheat varieties are shown: (A) Galahead; (B) Avalone;(C) Mercia.

1-41

1.4.3.2 Denaturants (Urea)

In many cases, urea has been found to be an effective denaturant for electro-phoresis. Urea is included in the sample and/or run buffers to reduce con-former formation and prevent aggregation. In relatively high concentration(4 to 8 M), urea prevents aggregation while protein structure is randomized sothat a single, symmetrical peak is observed on the electropherogram. Becauseurea has a significant absorbance in the low UV range, it is recommended tokeep the concentration of urea in the run buffer as low as possible to preservethe linearity of the dynamic range of the UV detector. However, it is still pos-sible to add higher concentrations of urea to the sample buffer when the EOF islow because urea is uncharged at run conditions and will not migrate with thecomponent of interest. An example of the use of urea is the work of Josic et al.(1990) on hydrophobic membrane proteins. This class of proteins tends toprecipitate in aqueous solution. Urea (7 M) was effective in preventing precipi-tation which led to increased precision in migration times and peak heights.In other work, Strege and Lagu (1993) used a urea-containing buffer to studyprotein folding.

1.5 Detectability Enhancements:Matrix Effects, Sample Stacking,and ITP Preconcentration

Almost all commercially available CE instruments are equipped with aUV-visible absorbance detector. For most protein and peptide assays, theUV-Vis detector provides more than adequate sensitivity. However, becauseof limited sample loadability and, furthermore, a short detector pathlength(typically 20-100 µm), concentration sensitivity with CE is limited in traceanalysis (≈ 10−6 Μ). Research has been carried out to improve the detectionlimits in CE. To achieve this goal, basically two strategies have been applied.The first involves a detection scheme other than UV absorbance, i.e., laser-induced fluorescence (LIF). The second strategy makes use of samplepreconcentration techniques through suitable chemistries and will bediscussed next.

1-42

1.5.1 Laser-Induced Fluorescence Detection (LIF)Unless the natural fluorescence of a compound can be exploited, it is necessaryto derivatize the analyte with a suitable fluorescent label. The majority of CE-LIF applications involve pre-capillary labeling with reagents similar to thoseused with fluorescence detection in HPLC. However, whereas it is relativelyeasy to derivatize and analyze at high amino acid concentrations, it becomesincreasingly difficult at low levels (< 10-7 M), as either the labeling efficiencydecreases or the background signal increases. Proteins may present additionalcomplications as multiple sites can be tagged; the resulting labeled species areusually electrophoretically separable. An interesting new development is theemployment of fluorescently labeled immunochemicals. These probes act ashighly selective tagging reagents and may circumvent the problems encoun-tered with labeling of protein analytes at low levels (see the discussion ofFigures 2-7 and 2-14 in Part 2.) Full discussion of LIF detection is beyond thescope of this book, but readers are referred to a recent review by Schwartzet al. (1994) and chapters in the textbooks by Guzman (1993) and Landers(1994).

1.5.2 Effect of Sample Matrix; Stacking of SampleComponents

Samples from biological origin frequently contain significant amounts of saltsor buffer ions. When injected into the CE instrument, these “matrix” ions candramatically influence resolution and detectability. It is not unusual to find thatseparations performed on standards in relatively clean sample matrices aresuperior to the separations in more complex environments, i.e., “real” samples.Thus, it appears that the composition of the sample is important with respect tothe peak efficiency of the analyte under investigation.

The above phenomenon is related to the relative conductivities of thesample zone and the run buffer. Figure 1-26 shows the effect of the ionicstrength of the run buffer on the migration and peak height of a standardmixture of bioactive peptides. The sample components were dissolved in0.03% TFA; the run buffer was sodium phosphate. In changing from 0.025 Mto 0.125 M, an increase in peak efficiency and peak height can be seen.

1-43

4 6 8 10 12 14 16

Time (min)

0.125 M

0.100 M

0.075 M

0.050 M

0.025 M

0.05 AU

12

34

5 6

7 8

9

Figure 1-26. Effect of the buffer ionic strength on peak shape and migrationtime of peptides. Buffer: sodium phosphate, pH 2.44; 30 kV; 20°C; 200 nm;57 (50) cm × 75 µm capillary. Peak identification: (1) dynorphin; (2) bradyki-nin; (3) angiotensin II; (4) TRH; (5) LHRH; (6) bombesin; (7) leu-enkephalin;(8) met-enkephalin; (9) oxytocin. From McLaughlin et al., Beckman TechnicalInformation Bulletin TIBC-106 (1991).

The greater the difference in conductivity between the sample zone andthe run buffer, the greater the focusing (also referred to as sample stacking).Under typical conditions, a 5- to 10-fold increase in detection sensitivity can beobtained by using sample stacking. In this case, the electric field in the samplezone is relatively high, causing the analytes to migrate rapidly until they reachthe interface between the sample buffer and the run buffer (see Figure 1-27).This causes the sample to be “stacked” at that interface. Thus, the sampleshould be applied in a medium of relatively low conductivity. If the opposite isthe case, uneven migration and zone spreading will result. Thus, for samples

1-44

which often contain salts, it is desirable to select a buffer with a relatively highionic strength. Note: the use of small-i.d. capillaries, i.e., 25 to 50µm, is favor-able in this respect as they permit better heat dissipation; an efficient capillarycooling system is also beneficial. High-ionic-strength buffers are also used inprotein separations (see Section 1.4.2.3.1). In addition to zone-focusing effectsbased on differences in conductivity, differences in pH between the zones mayfurther contribute to the sharpening of the peaks. Figure 1-28 provides a sche-matic example of a sample solution which has a pH that is greater than the pIof the peptides injected into a low-pH buffer. The diagram shows thatdeprotonated, negatively charged peptides stack up against the sample–bufferinterface.

+++

+

++

+

++

++

+

+

Interface

++++

++

+ +

+ ++

+

+

++++

++

+

(+)

(+)

(-)

(-)

SampleIntroduction

ApplyVoltage

SampleStacking

Sample Plug(low conductivity)

Sample Buffer(high conductivity)

Figure 1-27. The sample “stacking” mechanism. Sample ions have anenhanced electrophoretic mobility in a lower conductivity environment(i.e.,elevated local field strength). When a voltage is applied, sample ions inthe sample plug instantaneously accelerate toward the adjacent separationbuffer zone where, on crossing the boundary, a higher-conductivity environ-ment (lower field strength) causes a decrease in electrophoretic velocity and“stacking” of the sample ions into a buffer zone smaller than the originalsample plug. Reprinted with permission from Oda and Landers, Handbook ofCapillary Electrophoresis, Landers (Ed.), Boca Raton: CRC Press, 1994.

1-45

B

Sample Introduction

A

C

(+) (-)

(+) (-)

Concentration

Dissipation/Separation

Low pH High pH Low pH

Low pH High pH Low pH

Figure 1-28. Electrophoretic stacking of peptides. (A) Sample is applied to thecapillary in a solution with a pH higher than the pI of the peptides, resulting innegatively charged peptides. (B) Electric field is applied, resulting in stackingof peptides at the buffer/sample zone interface. (C) After dissipation of the pHstep gradient, peptides move toward the cathode. Reproduced with permissionfrom Abersold et al., J. Chromatogr. 516, 79 (1990).

As discussed earlier, sharper peaks also result in better detectability.For example, significantly lower detection limits were obtained by exploitingzone focusing effects. By dissolving peptides in water or 0.03% TFA andusing a pH 2.5 phosphate buffer, a 5- to 10-fold increase in signal can be ob-tained compared to a sample matrix consisting of run buffer. For the highestsensitivity, the sample injection volume should be maximized (as a rule ofthumb, not to exceed 2% of the total capillary volume). The reviews by Chienand Burgi (1992) and Albin et al. (1993) give further information on sampleconcentration techniques.

The effect of salts dissolved in the sample matrix on peak shape is demon-strated in Figure 1-29. This situation is commonly encountered when samplesare prepared by ion-exchange or reversed-phase HPLC. In the example inFigure 1-28, three peptides at a concentration of 100 µg/mL were dissolvedin the presence of 30 mM and 100 mM NaCl. A 100 mM borate run buffer,pH 9.2, was used. As can be seen, inefficient, broad peaks were obtained with100 mM salt in the sample. Doubling the injection time from 10 seconds to 20seconds only marginally increased the peak height. Similar, salt-related, peak-broadening effects also occur under acidic conditions (data not shown here).

1-46

AA

bsor

banc

e

Time (min)

Abs

orba

nce

Abs

orba

nce

Abs

orba

nce

Time (min)

Time (min)

Time (min)

C

B

D

0.110

-0.005

0.150

-0.005

0.110

-0.005

0.150

-0.005

Figure 1-29. Effect of salts in the sample zone on peak height and peak shapeof bioactive peptides. (A) 30 mM sodium chloride, 10 s injection. (B) 30 mMsodium chloride, 20 s injection. (C) 100 mM sodium chloride, 10 s injection.(D) 100 mM sodium chloride, 20 s injection. Reproduced with permission fromSatow et al., HRC, J. High Resolut. Chromatogr. 14, 276 (1991).

1-47

Because of the effects described above, desalting of the sample may berequired. When CE protocols require electrokinetic injection (with gel-filledcolumns, for instance), desalting is often required as only a relatively smallamount of sample is introduced into the capillary otherwise. This occurs be-cause salt ions compete with the analyte ions during the electrophoreticsample introduction process. Protein and nucleic acid samples are frequentlydesalted using procedures based on ultrafiltration, dialysis, or centrifugation.A micro-concentrating/desalting device, consisting of a small disc with re-versed-phase packing material suitable for small peptides, is commerciallyavailable. Another device, based on ultrafiltration and centrifugation, is ideallysuited for the small sample volumes typically encountered in CE (Amicon,Beverly, MA). Filters with different molecular mass cut-offs are available fromthe manufacturer.

Finally, it should be mentioned that, under carefully selected sample/runbuffer conditions, the addition of salt to the sample matrix may actually benefitzone sharpening. This case (e.g., Beckers and Everaerts, 1990) involves thecreation of a temporary isotachophoretic (ITP) zone during the injection pro-cess prior to the electrophoretic migration of the species. The use of ITP as apreconcentration technique will be discussed next.

1.5.3 ITP PreconcentrationIsotachophoresis (ITP) or displacement electrophoresis is an electrophoretictechnique which can be used as a concentration method for dilute samples.The concentration of the sample is adapted to that of the leading zone accord-ing to the Kohlrausch regulating function (dating back to 1897!). In contrast toCZE, where dilution of the sample takes place due to dispersion, ITP is inher-ently a focusing technique, as is isoelectric focusing. In the literature, differentapproaches have been described to couple ITP to CZE (e.g., Wanders andEveraerts, 1994; Foret et al., 1992 and 1993; Schwer and Lottspeich, 1992).The preconcentration step can be performed either in a dual- or single-columnmode. The dual-column mode involves ITP in a pre-capillary with subsequenttransfer of the concentrated zone to an analytical capillary where the separationof the individual zones takes place in a CZE mode. While it has been shownthat several milliliters of sample can be introduced with this approach, thedual-column system cannot be easily implemented in commercial CE systems.In the single-column configuration, a discontinuous buffer system is employed.ITP takes place only at the beginning of the experiment, and, after a change inconditions, separation in the CZE mode takes over. With this approach, theinjection volume, which can be focused and analyzed with on-column ITPpreconcentration, is typically 10 to 100 times higher than in normal-mode

1-48

CZE: 100 to 1000 nL can be injected into a 75-µm-i.d. capillary. The increasein sample loadability translates directly to an increase in sample detectability.The use of preconcentration appears to be particularly useful when the samplecontains not just proteins but also other ions; the latter often is the case withreal samples. The following section discusses two methods—one suited forbasic proteins, the other for acidic proteins—in which ITP preconcentration isused in P/ACE instrument (Foret et al., 1993). The schemes of the two basicelectrolyte arrangements for the on-column ITP sample preconcentration usedin this study are depicted in Figures 1-30 and 1-32 as Methods A and B.

1.5.3.1 Method A: For Basic Proteins

In this case, the sample is injected into the column filled with a leading electro-lyte (LE) as shown in step I of Figure 1-30. Next, the injection end of the capil-lary is placed in an electrode reservoir containing the terminating electrolyte(TE). Then the voltage is applied, and the sample components having mobili-ties intermediate to those of the LE and TE stack into sharp ITP zones asshown in step II. In practice, the final concentrations in focused ITP zones areclose to the concentration of the leading electrolyte, irrespective of the originalsample concentration.

Figure 1-30. Schematic illustration of transient ITP sample preconcentrationin CE. Method A: replacement of the terminating electrolyte. From: Foretet al., Beckman Technical Information Bulletin A-1740 (1993).

1-49

After achieving the ITP steady state, the buffer reservoir containing theterminating electrolyte is replaced by a reservoir containing the leading elec-trolyte, resulting in destacking of ITP zones, shown in step III. At this point,individual species begin to move in a zone-electrophoretic mode. The separa-tion of basic proteins according to method A is shown in Figure 1-31. Thesample volume was approximately 450 nL. After 4 minutes of ITP migration,the anode electrode reservoir (on the injection side of the capillary) was re-placed by a buffer reservoir containing leading electrolyte. As can be observedin Figure 1-31, excellent peak shape and resolution of proteins is obtainedwithin a short run time.

Time (min)

-0.002

0.002

0.006

0.010

0.014

0.018

0.022

0.026

0.030

Abs

orba

nce

0 2 4 6 8 10 12 14 16 18

1

2

3

4

56

7

8

Figure 1-31. Cationic CZE separation of basic proteins with ITP sample pre-concentration. BGE = leading electrolyte: 0.02 M triethylamine + acetic acid,pH 4.3; Terminating electrolyte: 0.01 M HAc, applied for 4 min. Capillary:50 cm × 75 mm, coated. Constant current mode: 15 mA, 17-28 kV. Sample:1) lysozyme, 6.3 × 10-7 M; 2) cytochrome C, 3.1 × 10-7 M; 3) ribonuclease A,4.7 × 10-8 M, 4) myoglobin 5.3 × 10-7 M; 5) α-chymotrypsinogen,1.6× 10-7 M; 6) β-lactoglobulin A, 3.1 × 10-7 M; 7) β-lactoglobulin B,3.1× 10-7 M; 8) carbonic anhydrase, 1.6× 10-7 M. Analytes dissolved in BGE.Injected volume: 450 nL. From Foret et al., Beckman Technical InformationBulletin A-1740 (1993).

1-50

1.5.3.2 Method B: For Acidic Proteins

Method B consists of a single electrolyte (the “background” electrolyte—BGE) used for both ITP preconcentration and CZE separation (Figure 1-32).The BGE acts as a terminating electrolyte when the sample itself contains ionswith high electrophoretic mobility which can serve as a leading zone duringtransient ITP migration. Biological samples usually contain such ions (e.g.,sodium, potassium, or ammonium ions in the form of chlorides, sulfates, orphosphates); however, partial desalting may still be desirable in cases when thesalt concentration is high. The best results will be obtained with the salt con-centration in the range 5 × 10-3 to 5× 10-2 M. At higher salt concentrations,longer capillaries must be used for completion of the separation. After sampleinjection and application of the electric current, the leading ions from thesample (having higher mobility) form a zone with an asymmetric leading andsharp rear boundary, resulting in a non-uniform electric field; consequently,ITP stacking of the sample ions is achieved. During the migration, the leadingzone broadens due to electromigration dispersion and its concentration de-creases. At a certain concentration of the leading zone, the sample bandsdestack and move with independent velocities in the zone-electrophoreticmode.

Figure 1-32. Schematic illustration of transient ITP sample preconcentration inCE. Method B: sample contains leading ions. LE = leading electrolyte,TE = terminating electrolyte. From Foret et al., Beckman Technical InformationBulletin A-1740 (1993).

1-51

Since zones with different conductivities may develop during the ITPstep, it is convenient to work at constant current. This will keep electric fieldstrength in the BGE constant and, thus, also maintain constant velocity ofprotein zones during the detection, assuring good reproducibility and quantita-tion. The use of transient ITP for preconcentration of acidic proteins accordingto method B is shown in Figure 1-33. In this case, 1 µL of the sample wasinjected. The protein zones preconcentrated behind the zone of chloride, thusforming a transient leading electrolyte during the early stages of the migration.

0 2 4 6 8 10 12-0.005

0.005

0

0.010

0.015

0.020

0.0251

2

3 4

5

Time (min)

Abs

orba

nce

Figure 1-33. Anionic CZE separation of 1 mL of the sample of acidic proteinsdissolved in 5 mM Tris-HCl. Sample: 1) glucose-6-phosphate dehydrogenase,1.25 × 10-8 M; 2) trypsin inhibitor, 1.6 × 10-7 M; 3) β-lactoglobulin B,1.2× 10-7 M, 4) L-asparaginase, 1.9 × 10-8 M; 5) α-lactalbumin, 4.4 × 10-8 M.BGE: 0.02 M TAPS-TRIS, pH 8.3. Capillary: 40 cm × 75 µmm, coated. Con-stant current mode: 7 mA, 22 kV. From Foret et al., Beckman ApplicationInformation Bulletin A-1740 (1993).

The above examples have shown that transient ITP preconcentration caneasily be adopted with the P/ACE instrument. The methods are reproduciblewith typical RSDs of migration times ≈ 0.5 to 1%. Sample volumes up to 1 µLcan be effectively preconcentrated in a 75-µm-i.d. coated capillary. Conse-quently, detection limits of proteins can be substantially improved (detectionlimits below 10-8 M or 0.1µg/mL are typical).

1-52

1.6 SDS Capillary Gel Electrophoresis(SDS-CGE)

Gels such as those made of agarose and cross-linked polyacrylamide (PA)were originally used as an anticonvective medium in slab gel electrophoresis.The gel structure creates a molecular sieving effect, allowing separations to beperformed based on the size of the molecules. The technique of SDS-PAGE isthe most popular among all electrophoretic techniques. SDS binds to polypep-tide chains such that, in theory, similar charge densities and constant mass-to-charge ratios of different proteins are obtained. As a result of the presence ofthe gel, separation then takes place based solely on the basis of the size of theproteins. SDS-PAGE is frequently used to analyze the purity of peptides andproteins and also to determine apparent molecular weight (MW) by means ofa comparison with calibration standards. Earlier approaches of SDS-PAGE incapillaries were described by Cohen and Karger (1987), Widhalm et al.(1991), and Tsuji (1991).

Beckman recently developed a new system for the SDS-CGE of proteins.A proprietary, linear polymer formulation (UV transparent at 214 nm) is usedin conjunction with 100-µm-i.d. coated capillaries. In CE, UV-transparentpolymers such as dextran or polyethylene glycol are preferred over polyacryla-mide as a sieving medium (Ganzler et al., 1992; Guttman et al., 1992, 1993;Lausch et al., 1993). Samples can be introduced by means of pressure injectionwhile, after completion of the run, the gel matrix can be replaced by using thepressure-rinsing feature of the P/ACE instrument. With an “in-between-runs”rinsing procedure, excellent reproducibility can be achieved for proteins of upto ≈ 200,000 Daltons (Tsuji, 1993).

Figure 1-34 shows the separation of seven protein standards (size rangefrom 14,200 to 205,000 Daltons) in approximately 15 minutes on a 27-cm-length capillary. A semilog plot of the MW of the standards versus the inverseof the relative migration time (RMT) exhibits excellent linearity (Figure 1-35).With the use of Gold™ Molecular Weight software, estimation of the MW ofunknown proteins is easily obtained.

1-53

5.00 10.00 15.0016.010.00

0.0

000

0.0

100

0.0

200

0.0

300

0.0

400

0.0

000

0.0

100

0.0

200

0.0

300

0.0

400

Absorb

ance

Absorb

ance

1

2

3

4

5

6

7

8

27 cm Total Length Capillary1 = Orange G 2 = α-lactalbumin (14.2 KDa)3 = carbonic anhydrase (29 KDa)4 = ovalbumin (45 KDa)5 = bovine serum albumin (66 KDa)6 = phosphorylase B (97.4 KDa)7 = β-galactosidase (116 KDa)8 = myosin (205 KDa)

Figure 1-34. Standard test mix separation on an eCAP-SDS 14-200 capillary

.4 .45 .5 .55 .6 .65 .7 .75 .84

4.2

4.4

4.6

4.8

5

5.2

5.4

5.6

5.8y = -3.377x + 6.833, R-squared: .995

1/RMT

log m

ol. w

t.

Figure 1-35. Molecular weight standard curve

1-54

For proteins which do not bind to SDS in the 1 to 1.4 ratio (e.g., glycopro-teins), generally inaccurate estimates are obtained. In this case, a Ferguson plotcan be constructed to account for this nonideal behavior. First, the RMTs ofstandards at different gel concentrations are measured. A log (1/RMT) vs. thegel buffer concentration plot yields a retardation coefficient, KR, for each ofthe standard proteins (Figure 1-36). Next, a plot of log MW vs. the square rootof KR can be used for estimation of MWs (Figure 1-37). The Ferguson methodis a rather time-consuming and labor-intensive method when used with SDS-PAGE. However, using CE, this method is easy to perform in an automatedfashion (for further details see the Beckman eCAP SDS 14-200 Kit,P/N 477420, and a recent publication by Werner et al. (1993).

40 45 50 55 60 65 70 75 80 85 90 95 100 105-.4

-.35

-.3

-.25

-.2

-.15

-.1

-.05

0

Gel Buffer Concentration

Log

(1/R

MT

)

α-LactalbuminPhosphorylase

Carbonic anhydraseGalactosidase

OvalbuminMyosin

Bovine serum albumin

Figure 1-36. Ferguson plot of seven standard proteins

.0275 .03 .0325 .035 .0375 .04 .0425 .045 .0475 .05 .05254

4.2

4.4

4.6

4.8

5

5.2

5.4

Sqrt KR

Log M

ol. W

t.

Figure 1-37. KR plot used to determine molecular weights of proteins

1-55

While SDS-CGE is in many respects very similar to SDS-PAGE on slabgels (e.g., sample preparation), a number of distinct advantages can be noted:

• CE analyses are fast, reproducible and quantitative.• In SDS-CGE, the gel can be replaced at least 100 times.• Direct, on-line UV detection is feasible.• There is no need to stain or de-stain the proteins. It is known that, with

slab gels, the staining reagent Coumassie Blue does not bind to proteins ina stoichiometric ratio; therefore quantitation with slab gels is not alwaysreliable.

1.7 Capillary Isoelectric Focusing (CIEF)

1.7.1 Principle of IEFIn isoelectric focusing (IEF), amphoteric molecules such as proteins are sepa-rated by electrophoresis in a pH gradient generated between the cathode andanode. The technique takes advantage of the fact that each protein has a differ-ent pH at which it is electrically neutral; its isoelectric point (pI). Briefly, theprinciple of IEF separation is as follows. Under the influence of an electricfield, charged species will start to migrate through the electrophoresis medium(gel or solution). If the sample component has a net negative charge, migrationis toward the anode. During the migration, the sample encounters progres-sively lower pH, thus picking up more positive charge. Eventually, a zone isreached where the net charge is zero. At this point (at the pI), migration stopsand the sample component is focused in a tight zone. Likewise, if a componenthas a positive charge, migration will be toward the cathode. Thus, each samplecomponent migrates to its own isoelectric point. In classical IEF techniques,the separated zones are usually visualized by means of staining. IEF is a trueelectrophoretic focusing technique, as is isotachophoresis, i.e., the separatedzones are self-sharpened during the electrophoretic separation. Protein mol-ecules diffusing out of a focused zone will acquire a charge and are pulledback into the center of the zone where the net charge is zero. IEF is most oftenapplied to proteins (including enzyme isoforms, polyclonal, and monoclonalantibodies, hemoglobin variants, and r-DNA-made proteins), although pep-tides, whole cells and subcellular particles, viruses, and bacteria also have beenstudied by IEF. The technique is particularly useful to estimate the pI of anunknown protein through calibration with known protein standards. Manyapplications of IEF in the biomedical/clinical fields have been reported(Righetti, 1983).

1-56

1.7.2 Classical IEFIn IEF, the pH gradient, which is key to the success of the technique, is pro-vided by molecules called “carrier ampholytes.” In the mid 1960’s, the Swed-ish researchers Svensson and Vesterberg were able to synthesize moleculeswith the exact properties needed for IEF (e.g., good conductivity at their iso-electric points and good buffering capacity). The ampholytes consist ofpolyamino-polycarboxylic acids with slightly different pI values; their averagemolecular mass is approximately 750 Daltons. Ampholyte mixtures in wideand narrow pH ranges can be purchased from various commercial sources.Typically, polyacrylamide or agarose gels are used as anticonvective media,similar to types of slab gel electrophoresis. It is also possible to immobilizepH gradients on a suitable matrix such as polyacrylamide. In this case, thebuffering groups of the pH gradient are acrylamide derivatives which are co-polymerized into the gel matrix. This IEF technique with immobilized pHgradients was introduced in 1982 by Bjellqvist, Righetti, and co-workers. IEFyields typically higher resolution than do other modes of electrophoresis. Withimmobilized pH gradients, proteins differing by 0.001 pI unit have been sepa-rated. Routinely, IEF provides resolution of 0.1 to 0.01 pI units. For evenhigher resolution, IEF can also be combined with other modes of electrophore-sis, e.g.,with SDS-PAGE or immunoelectrophoresis (“two-dimensional electro-phoresis”).

1.7.3 Capillary Isoelectric Focusing (CIEF)CIEF offers the potential to combine the high resolving power of conventionalgel IEF with the advantages of modern CE instrumentation. Small-diametercapillaries with efficient dissipation of Joule heat permit the use of relativelyhigh field strengths for rapid separations. Thus far, all CIEF separations havebeen carried out in free solution (i.e., without a gel, as in classical IEF). InCIEF, as in classical IEF, proteins are separated in a pH gradient created byampholytes under the influence of an electric field. As will be discussed below,the key to high performance is to effectively displace the protein zones out ofthe capillary without introducing band broadening. Direct, on-line UV detec-tion is feasible, without the requirement for staining of the focused proteinzones (most often, detection is at 280 nm as the ampholytes absorb in the lowUV region). With the P/ACE system, CIEF can be performed automatically,allowing unattended analysis of multiple samples.

1-57

1.7.3.1 Two-Step CIEF Methods: Focusing and Mobilization

Hjerten and Zhu (1985) were the first to develop IEF in capillaries. A two-stepprocess was performed: first, proteins were focused in the capillary. Aftercompletion of this process (monitored by a drop-off in current), displacement(“mobilization”) of the zones out of the capillary took place by means ofchanging the chemical composition of anolyte or catholyte solution (i.e., byadding acid, base, or salt). In this method, termed “chemical mobilization,”the change in anolyte or catholyte causes a shift in the pH gradient, resultingin migration of the zones past the detection point in the capillary. CIEF withchemical mobilization was further optimized and applied to a variety of sub-stances including monoclonal antibodies (Wehr et al., 1990), hemoglobinvariants (Zhu et al., 1992), human serum transferrin (Kilar and Hjerten, 1989),and glycoforms of recombinant proteins (Yim, 1991). The method involves theuse of coated capillaries to eliminate the electroosmotic flow (EOF) and toprevent undesirable adsorption of proteins to active sites on the capillary wall.

Instead of the above-described chemical mobilization, focused zones canalso be moved past the detection point by hydrodynamic means. In their earlyexperiments, Hjerten and Zhu (1985) used a low-flow pump to displace thecapillary contents while leaving the voltage on to maintain resolution of thefocused protein zones. Recently, the principle of this method was implementedin a modern CE instrument by Chen and Wiktorowicz (1992) and Nolan(1993), using vacuum and pressure mobilization, respectively. The formerworkers demonstrated excellent linearity of a pI vs. mobility plot in the pHrange of 2.75 to 9.5. Figure 1-38 shows the principle of the method employedby Nolan. A plug of sample is introduced into a 27 cm × 50 mm capillary.The catholyte containing 20 mM sodium hydroxide is backflushed just past thedetection point. The base acts as a “blocking agent,” i.e., focusing of basicproteins will not extend beyond the detection window which would preventtheir detection. After the focusing step (completed in approximately 3 min-utes), a low-pressure rinse is applied, moving the focused zones past the detec-tion point (Figure 1-38B). During this step, the electric field is turned on tomaintain the sharpness of the bands. This CIEF system yielded sharply focusedprotein peaks and was found to be reproducible (RSDs for migration times< 0.6%). Calculated pI values of proteins correlated well with reported pIs.

1-58

Catholyte

Anolyte Catholyte

+Anolyte

Focused Protein Zones

Detection WindowPressure

1.5% Ampholytein 0.4% MC

Sample Detection Window

A

B

Anolyte

Figure 1-38. Principle of CIEF by simultaneous pressure/voltage mobiliza-tion. A) Catholyte is backflushed past the detection point and a sample plug isintroduced into the coated capillary (no high voltage). B) Focusing of sampleis complete and the sample components are driven toward the detector by alow-pressure rinse. High voltage is applied during this step.

1.7.3.2 One-Step CIEF Methods:Simultaneous Focusing and Mobilization

The need for a simple, reproducible method has led to a one-step CIEF methodwhich involves utilizing the EOF for mobilization of the focused zones(Mazzeo et al., 1992; Thormann et al., 1992; Yao and Regnier, 1993; Pritchett,1994). In one variant, Thormann et al. (1992) introduced a large plug ofsample (approximately 6% of the capillary volume) dissolved in an ampholytesolution into a 90-cm-length, untreated capillary which was filled withcatholyte (20 mM sodium hydroxide, 0.3% hydroxypropylmethylcellulose).After placing the capillary between the catholyte and the anolyte (10 mMphosphoric acid), the electric field was turned on, causing the formation of apH gradient and focusing of the protein zones; simultaneously, the sample wasswept towards the detection point by the EOF.

The method of Mazzeo and Krull (1992) is different from the above-mentioned method of Thormann in that, initially, the entire capillary is filledwith sample and ampholytes. Good results were achieved with commercially

1-59

available (Supelco) C8 coated capillaries. Yowell, et al. (1993) used this one-step CIEF method for the QC analysis of recombinant protein formulations.Pritchett (1994) applied the methods to the qualitative and quantitative analysisof monoclonal antibodies. The principle of their method is schematicallyshown in Figure 1-39. After filling the capillary with the sample mixture(which contained the basic “blocking agent,” TEMED), the capillary wasplaced between the inlet (catholyte, 20 mM sodium hydroxide) and outlet(anolyte, 10 mM phosphoric acid). Subsequently, the voltage was turned on,focusing the protein zones in the short (7 cm) end of the capillary (note: theTEMED blocks the long (40 cm) end of the capillary). Simultaneously, theEOF drives the zones past the detection point in the reverse direction.

AnolyteCatholyte

Catholyte+

Anolyte

Focused Protein Zones

Detection WindowEOF

TEMED

0.2% HPMC, 0,75% TEMEDSample in 2% Ampholyte,

Detection Window

A

B

Figure 1-39. Principle of the one-step CIEF method with EOF mobilization.(A) Coated capillary is filled with sample. (B) Under high voltage, proteins arefocused in the short end of the capillary and mobilized past the detection pointby the EOF.

CIEF has the potential to replace many of the established, classical IEFprocedures. CIEF has advantages over classical IEF with slab gels with regardto time savings, sample preparation, quantitation, and reproducibility. In addi-tion, no staining or destaining of the bands needs to be performed as the detec-tion takes place directly on the capillary, in real time.

1-60

1.8 Micropreparative CEBecause CE involves separations in small-i.d. capillaries, the technique hasfound limited use as a micropreparative tool. In comparison with HPLC, CE isa priori more limited with respect to the maximum sample load (i.e., theamount of sample which can be loaded onto the capillary without causingexcessive zone broadening). Consequently, the amounts of material collectedfrom capillaries are relatively small, which constrains further manipulation andanalysis. However, a growing number of publications have recently indicatedthat, under proper conditions, CE can be a useful tool for protein chemists. In1988, Cohen et al. have shown the utility of micropreparative CE for the puri-fication of oligonucleotides (800 ng of a primer was collected and subse-quently used as a probe) and peptides. While gel-filled capillaries were used inthis work, CE in the free-solution mode can also be utilized for microprepara-tive work. Table 1-4 lists a number of applications in which CE has been usedin the micropreparative mode.

Table 1-4. Micropreparative CE Applications

• Peptide purification from digests prior to microsequencing• Amino acid compositional analysis after hydrolysis• Slab gel electrophoresis (radioactivity labeling)• Identification of fractions by mass spectrometry• Enzymatic activity analysis of fractions collected by HPLC or CZE

Fraction collection with CE is different from HPLC in that the capillarymust stay in contact with a solution containing water or buffer and the elec-trode (i.e., the electric field, which drives the separation, must be maintained).This is probably the most commonly used way to collect fractions from CEruns. Alternatively, the fraction to be collected can be mobilized by utilizingthe low-pressure (0.5 psi) rinsing capability of the P/ACE system (Gagnon,1991). Yet another possibility is to connect the capillary to a syringe pumpfor a “dynamic” elution after switching off the field (as was done by Camilleriet al., 1991). The latter approach was applied successfully for sequencingpeptides at the low-picomole level.

Basically, three strategies are conceivable for micropreparative CE: col-lection from a single run with standard capillaries, collection by performingmultiple runs, and collection from large-diameter capillaries.

1-61

1.8.1 Collection from a Single Run with StandardCapillaries

When enough analyte is present in the sample or when the sample can be pre-concentrated, a one-time collection can be performed on a standard capillary of50 to 75 µm i.d. Typically, the initial concentration of peptide required is from1 to 10 mg/mL to allow further analysis by spectroscopic or chemical means.Banke et al. (1991) have shown that, when enzymatic amplification is applied,even minor components of a fermentation broth can be analyzed by means ofcollection from 75-µm-i.d. capillaries. In this application, quantities as small as3 ng could be identified as an alkaline protease of the subtilisin family.

1.8.2 Collection by Performing Multiple RunsTo increase the amount of collectable material, multiple runs can be performedwith the same collection vial. The concentration of the targeted analyte shouldincrease in proportion to the number of runs. A prerequisite is that adequaterun-to-run reproducibility has been obtained. More detailed information onprogramming the CE instrument for micropreparative work can be found inBeckman Technical Information Bulletin TIBC-105 (Biehler and Schwartz,1991).

1.8.3 Collection from Large-Diameter CapillariesThe loadability on a capillary is roughly proportional to the square of its diam-eter. Thus, for example, seven times the sample load can be collected using a200-µm-i.d. capillary than when using a 75-µm-i.d. capillary. However, thepenalty paid for increasing the diameter is increased Joule heat generation.Ultimately, this is the limiting factor in the separation performance. A goodcapillary cooling system for heat dissipation decreases this problem.

The feasibility of this approach (Smith and Ohms, 1992) is illustrated inFigure 1-40 with the collection of peptides from a tryptic digest (β-lactoglobu-lin). The sample load on the micropreparative capillary was approximately100 pmol of digest and a 50 mM sodium phosphate run buffer, pH 2.7, wasused. It can be seen that the integrity of the individual peaks is largely main-tained when scaling up from the analytical (75 µm i.d.) to the micropreparativecapillary (200 µm i.d.). However, the analysis time is much longer because alower voltage was applied to prevent excessive Joule heating. After collection,various fractions were subjected to microsequence analysis. The results of thesequence analysis are shown in Table 1-5. Recoveries of the peptides rangedfrom 10 to 75%.

1-62

0.012

0.010

0.005

0.000

0 10 20 30

A

Time (min)

Abs

orba

nce

(214

nm

)

0.03

0.02

0.01

0.00

0 40 80 120

B

Time (min)

Abs

orba

nce

(214

nm

)

Figure 1-40. Comparison of analytical and micropreparative CE of aβ-lactoglobulin tryptic digest. (A) analytical CE with a 75-µm-i.d capillary.(B) micropreparative CE with a 200-µm-i.d. capillary. See text for details.From Smith and Ohms, Techniques in Protein Chemistry III (1992).

1-63

Table 1-5. Recovery of Peptide fromβ-Lactoglobulin Tryptic Digest*

Frac Amino Acid Sequence Yield Recovery PurityNo. (pmol) % %

8 TK 70 70 > 9011 ALK 35 35 50

TMK 35 35 5013 LQK 50 50 8015 FPK 40 40 > 9018 VYEELKPTPEGDDLEILLQK 25 25 > 9021 TPEVDDEALEK 40 40 7528 SLAMAASDISLLD(AQSAPLR) 35 35 65

VAGT(W)Y* Data from Smith and Ohms, Techniques in Protein Chemistry III, 1992.

In a later paper, Kenny et al. (1993) used 150-µm-i.d. capillaries for frac-tion collection prior to microsequencing at the low picomole level. A compari-son with HPLC was made. The results indicate that, for the peptides selected,the recovery with CE was superior to that with HPLC at load levels below50 picomoles. With CE, recoveries in the 60 to 70% range at the 5- to 10-pico-mole level were routinely obtained, whereas with HPLC yields were less than10% at the 50-picomole level.

1.9 Affinity Capillary Electrophoresis(ACE)

Affinity electrophoresis is a technique which has been used for the character-ization of biomolecules and for the analysis of specific interactions with ofbiomolecules with affinity ligands, including antibody–antigen interactions(for a review, see Takeo, 1987). When CE is used to study receptor–ligandinteractions, the technique is referred to as affinity capillary electrophoresis(ACE). While still in its infancy, it is clear that ACE is rapidly becoming animportant tool for the bioanalytical/clinical researcher. At the HPCE 1994 meet-ing in San Diego, a complete session was devoted to applications in this field.

1-64

1.9.1 Receptor-Ligand StudiesThe binding of a ligand (protein, peptide, small molecule) to a receptor (typi-cally a protein or peptide) is schematically shown in Figure 1-41.

Kb

Receptor Ligand Receptor/Ligand Complex

Figure 1-41. Binding of a ligand to a receptor

The protein–ligand complex has different charge-to-mass characteristicsthan the unbound protein which permits its separation by CE. By measuringthe migration time as a function of the concentration of the ligand present inthe CE run buffer, the equilibrium constant Kb can be measured (Chu et al.,1992). Traditionally, such affinity measurements have been performed withtechniques such as equilibrium dialysis or spectroscopy. However, these meth-ods are often labor intensive and require labeling or depend on secondaryreagents for quantitation. ACE has several potential advantages:

• Only a small amount of material is required.• The receptor need not be highly purified and its concentration does not

need to be known.• ACE allows, in principle, simultaneous determination of several ligands.• ACE has short run times and is precise and quantitative.

Interactions between receptors and ligands have been studied by a numberof workers and various CE-based reaction schemes have been devised.For example, Kajiwara (1991) used ACE for the analysis of conformationand interaction of metal-binding proteins such as calmodulin and parvalbumin(calcium-binding) and carbonic anhydrase (zinc-binding). Using EGTA as aCa2+-chelating agent in the run buffer, the binding shift of soluble calciumbinding proteins was investigated by Huch and D’Haese (1993). In the caseof calmodulin, binding of Ca2+ resulted in a change in protein conformationand subsequent interaction with various target enzymes. The electrophoreticchange upon adding the chelating agent for parvalbumin is shown in Fig-ure 1-42. Whereas in the presence of Ca2+ a single protein peak is obtained(panel A), adding EGTA to the buffer results in a double peak and shift towardlonger migration times (panel B). Potential applications of this assay includethe diagnosis of pathological alterations of skeletal muscle.

1-65

† (min.)2 4 6 8

0.08

0.04

A214

A214

† (min.)2 4 6 8

0.04

0.06

0.02

B

A

Figure 1-42. CE profiles of parvalbumin from white skeletal muscle of the chub(Leuciscus cephalus) in the presence of 0.1 mmol L-1 Ca++ (A) or 10 mmol L-1

EGTA (B) Voltage: 20 kV (Current 1: 10 µA). Temperature: 30°C. Injectedamount of protein: 4 ng in 2 nL (≈ 0.3 pmol). From Huch and D’Haese, Dis-covery, Spring 1993, Beckman Instruments, Inc.

In other recent applications of ACE, Honda et al. (1992) determined theassociation constant of monovalent-mode protein–sugar interaction. Heegaardand Robey (1992) used CE to evaluate the binding of anionic sugars to syn-thetic peptides. Chu et al. (1992), as well as Biehler and Jacobs (1993), appliedACE to study molecular recognition with low-molecular-weight receptors.Kraak et al. (1992) tested three different reaction schemes for determination ofbinding constants resulting from protein–drug interactions. Frontal analysisgave the best results for a bovine serum albumin–warfarin model system.Finally, Kuhn et al. (1994) used ACE for determination of lectin–sugar inter-actions.

1-66

1.9.2 Antibody-Antigen InteractionsImmune complexes can also be studied by ACE. Nielsen et al. (1991) investi-gated MAb-antigen complexation with hGH as the antigen. Immunoassaysbased on ACE and laser-induced fluorescence (LIF) detection have also beendescribed recently. An example of this approach—digoxin assay in serum—will be discussed later in Part 2. LIF detection, in contrast to UV detection,permits the (ultra-) low detection of labeled analytes required for immunoas-says. P/ACE can be coupled to various laser light sources, permitting optimumselectivity and sensitivity. [A recent review (Schwartz et al., 1994) describesvarious applications of the P/ACE-LIF detector, including those of proteins,peptides and amino acids.] Immunoassays are based on the specific chemicalreaction between an antibody and its corresponding antigen. Quantitationinvolves the separation and detection of antibody–bound antigen from the freeantigen or antibody, depending upon the analytical scheme employed. Theantigen–antibody reaction-equilibrium has a slow off-rate, resulting in theformation of an observable—by CE—complex (this is different in some of theexamples mentioned in Section 1.9.1). In a competitive binding immunoassay,a labeled antigen and the antigen of interest compete for an appropriateamount of antibody. Differentiation between the (labeled) complex and thefree (labeled) antigen can be accomplished by CE-LIF. In a noncompetitiveassay, a known amount of fluorescently labeled antigen is added to the samplewhich forms a complex specifically with the antibody. The feasibility of bothtypes of immunoassay with CE-LIF was demonstrated by Schultz andKennedy (1993). A He-Cd laser emitting at 442 nm was used to determineFab fragments of a monoclonal antibody and human insulin. Shimura andKarger (1994) used a fluorescently labeled antibody fragment to tag anantigen (methionine–recombinant human growth hormone). In this competi-tive immunoassay, a 488-nm Ar-ion laser was used with detection limitsdown to 5 × 10-12 M.

Part 2

Protein/Peptide Applications of CEto Analytical Biotechnology

2-3

2.1 Introduction

2.1.1 The Importance of Analytical Chemistry toBiotechnology

Biotechnology products, especially protein and peptide biopharmaceuticals,must be analyzed and characterized as rigorously and completely as currenttechnology allows. Such extensive analysis is driven by the complexity of themolecules, and is needed to meet the demands of good science, ethical busi-ness practice, and regulatory requirements. As we will discuss in this part, theseparation power and versatility of CE make it an ideal solution for many ofthe analytical challenges provided by proteins and peptides.

An important goal of analytical biotechnology is to have adequate charac-terization, specification, and control assays for the Chemistry, Manufacturing,and Control (CMC) section of a regulatory filing such as an IND (investigativenew drug application), PLA (product license agreement), or NDA (new drugapplication). The importance of analytical methods can be seen in the fact that,according to biotechnology consultant and ex-Food and Drug Administration(FDA) official Peter Hoyle, CMC deficiencies can lead to rejection of regula-tory submissions. A company invests millions of dollars to bring a drug toPhase 1 clinical trials, and hundreds of millions to conduct the trials throughPhase 3 and achieve licensing. This is all money wasted if applications arerejected or clinical trials fail because of inadequate or inappropriate assays.

2.1.2 The “Eight Points” Model of AnalyticalDevelopment

The U.S. FDA current Good Manufacturing Procedures for drugs (CGMP,Title 21 of the Code of Federal Regulations, Parts 210 and 211) states that allpharmaceutical products must be analyzed for “identity, strength, quality, andpurity.” Many analytical development scientists in biotechnology considerthese precepts too vague to serve as effective guidelines for biopharmaceuticalanalysis. When one of the authors (TJP) was working in the biopharmaceuticalindustry, he developed the following “eight points” model to guide analyticaldevelopment for any new macromolecule he and his staff encountered. Severalsuccessful regulatory filings attest to the adequacy of the model.

2-4

According to the model, assays should be developed and validated toaddress each of the following points:

1) Identity 5) Heterogeneity

2) Quantity 6) Stability

3) Purity 7) Process Consistency

4) Activity 8) Safety

CE can be used directly for all the points except, perhaps, numbers 4and 8, which are usually addressed using biological and microbiological assays(supplemental CE activity assays may, however, sometimes be validatedagainst biological assays). In the following pages, details of the use of CE toassess the identity, quantity, purity, heterogeneity, stability, and process consis-tency of proteins and peptides will be discussed.

While the emphasis is on the role of CE in the analysis of biopharmaceuti-cal proteins, many of the examples also pertain to the analysis of proteins inacademic and clinical research laboratories. For more information on the roleof analytical techniques in biotechnology with regard to quality control testingand FDA concerns, the reviews by Garnick et al., (1988) and Avallone (1986)are a good starting point.

2.2 IdentityThe purpose of an identity assay is to provide scientific proof that the contentsof a container correspond, qualitatively, to what is claimed on that container’slabel. Until identity is established, all other analytical concerns are secondary.Because of this central role, proof of identity is usually approached by a sum-mation of evidence from several assays. In addition, using multiple structuraldetermination methods ensures that products are thoroughly characterized aswell as adequately identified. For proteins, commonly used identity assaysinclude specific activity, amino acid composition and sequence, and assess-ment of such physicochemical parameters as molecular weight and isoelectricpoint. A list of assays commonly used to establish the identity of a biopharma-ceutical protein is presented in Table 2-1.

2-5

Table 2-1. Methods Commonly Used forCharacterization of Proteins: Identity

• Specific activity assays• Immunological methods

- Immunoblots- ELISAs- RIAs

• Electrophoretic profile• Chromatographic profile• Peptide mapping• Amino acid analysis• Carbohydrate analysis• N-terminal sequencing• Mass spectrometry• Colorimetric assay• Optical rotatory dispersion• Circular dichroism

Several CE techniques readily lend themselves to providing evidence of aprotein’s identity. These include:

• Peptide mapping by free-solution CE (CZE)

• SDS-capillary gel electrophoresis (SDS-CGE) for evaluation of relativemolecular weight

• Capillary Isoelectric Focusing (CIEF) for evaluation of a protein’s charac-teristic isoelectric point

• CE-Mass Spectrometry (CE-MS) for direct assignment of molecular weight

In the following section, examples of work in which these CE techniquesare used for identity assays are discussed.

2.2.1 Peptide Mapping: Utility of CZEIn peptide mapping, a protein is enzymatically or chemically cleaved intospecific peptide fragments which are then separated and detected. Basedmainly on its primary structure and the specificity of the cleavage reagent, agiven protein will exhibit a characteristic pattern of peptides called its “peptidemap.” Digestion of the protein into smaller peptide fragments allows subtlestructural features of the protein to be detected. A peptide map thus provides a“fingerprint” of a protein, and evidence of its correct identity.

2-6

Additional applications of peptide mapping include primary structuredetermination (i.e., peaks are collected and sequenced), detection of posttrans-lational amino acid modifications, identification of genetic variants, and thedetermination of glycosylation and/or disulfide sites. For these reasons, peptidemapping has found widespread use in the quality control (QC) and character-ization of recombinant DNA derived products.

Traditionally, analysis of peptide fragments has been achieved using high-resolution techniques such as reversed-phase HPLC (RP-HPLC). Slab gelelectrophoresis (SDS-PAGE) is sometimes used, although less frequently thanHPLC. CZE peptide mapping provides not only an excellent high-resolutionseparation tool, but is eminently suitable to complement RP-HPLC in identifi-cation or proof of structure cases. This is because the separation mechanismsof RP-HPLC and CZE are different, being based on hydrophobicity and mass-to-charge ratio, respectively. Hence, whereas certain peptide fragments cannotbe resolved by RP-HPLC (because of similar hydrophobic character), CZEreadily provides baseline resolution; likewise, RP-HPLC may resolve peaksnot separable by CZE (see Grossman et al., 1989; Bullock, 1993).

Figure 2-1, from Rush et al. (1993), illustrates the power of CE peptidemapping in the analysis of recombinant human erythropoietin (rHuEPO).Here, the enzyme trypsin, which specifically cleaves at argenine and lysineresidues, was used as the cleavage reagent. A pH-2.5, phosphate CE run bufferwas used, and an ion pairing reagent, 100 mM heptanesulfonic acid, was in-cluded, similar to the example of Figure 1-7, Section 1.3.2.3. The tryptic mapsof rHuEPO derived from two different expression systems are shown. Theupper trace shows the map of rHuEPO expressed in E. coli, a prokaryoticsystem, while the lower trace shows the map of rHuEPO expressed in Chinesehamster ovary (CHO) cells, a eukaryotic system. The former (E. coli) proteinis not glycosylated while the latter (CHO) originally contained carbohydrategroups typical of mammalian systems, but was treated with N-glycanase whichcleaves the N-linked oligosaccharides, leaving an aspartic acid residue in placeof the original asparagine residue. While 16 peaks have identical migrationtimes in the two electropherograms, the peaks labeled a through f are onlypresent in the E. coli-expressed rHuEPO material and can be attributed tostructural differences between the two preparations (Rush et al., 1993). Thepeptide maps also provide information on the heterogeneity associated with thethree rHuEPO glycopeptides (see Section 2.5). The combination of CE withmass spectrometry (CE-MS) would, in principle, allow the unknown peaks inFigure 2-1 to be unequivocally identified (see Section 2.2.4).

2-7

Det

ecto

r R

espo

nse

(mvo

lts)

1 2 34

56

7

8 910

1214

16

11

0123456789

10

12

14

16

18

12 13 14 15

11

13

15

16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34Migration Time (min)

a

b

c

d

e

f

Figure 2-1. Comparison of N-glycanase-treated CHO cell-expressed (lowertrace) versus E. coli expressed (upper trace) rHuEPO tryptic maps by HPCE.Reproduced with permission from Rush et al., Anal. Chem. 65, 1834 (1993).

2.2.2 Molecular Weight Estimation: Utility of SDS-CGE

While accurate molecular weight (MW) determination of proteins usuallyinvolves ultracentrifugation or mass spectrometry, estimation of relative MWis typically performed using size-exclusion HPLC and SDS-PAGE. The capil-lary analog of SDS-PAGE (SDS-capillary gel electrophoresis) was describedin Part 1 (Section 1.6) and involves the use of SDS-containing polymer solu-tions (“replaceable gels”) which can be pumped in and out of the capillary. Thecapillary format offers fast and reproducible analyses, accurate quantitation,and direct, on-line UV detection of proteins without the need for staining ordestaining. Beckman’s Gold software allows rapid calibration and MW esti-mation of proteins.

Examples of SDS-capillary gel electrophoresis (SDS-CGE) can be seen inFigures 2-2 and 2-3 which are representative of the separation of relatively lowand high MW proteins, respectively. The IgG monoclonal antibody (Fig-ure 2-3) was also run under reduced conditions—the sample was treated with

2-8

2-mercaptoethanol—resulting in the appearance of light and heavy chains(MW 25,000 and 50,000).

(Orange G)

0.0

00

5.0

0

8.0

0

12

.00

16

.00

20

.00

0.0

10

0.0

20

0.0

30

0.0

40

(IL-3)

(IL-5)M

W=

15

,80

0

MW

=2

2,3

00

Tim

eC

hannel A

. A

bsorb

ance

Figure 2-2. SDS-CGE Separation of interleukins 3 and 6. Orange G was usedas a reference standard.

00

Ora

nge G

Lig

ht C

hain

(M

W=

25,0

00)

Heavy C

hain

(M

W=

50,0

00)

IgG

(M

W=

155,0

00)

4 6 8 10 12 14 16 18

2012

40

60

80

100

%

Figure 2-3. Purity check of IgG monoclonal antibody by SDS-CGE. Trace 2was run under reduced conditions.

2-9

SDS-CGE, as well as CZE, can also be very useful in monitoring dimer–oligomerization processes (Palmieri and Nolan, 1993; Tsuji, 1993). Figure 2-4shows the SDS-CGE electropherogram of a dimer-enriched recombinant bo-vine somatotropin (rbSt) preparation. Baseline resolution of the monomer,dimer, trimer, and tetramer peaks was obtained, demonstrating the resolvingpower of the method.

0 5 10 15 20

DC

A

B

Hei

ght

Minutes

Figure 2-4. SDS gel-filled capillary electropherogram of a dimer-enrichedrbSt sample indicating a baseline resolution of monomer, dimer, trimer, andtetramer peaks. Conditions, 300 V/cm (24 µA); detector: 214 nm; columntemperature: 20°C; migration distance: 40 cm; coated capillary: 100 µm i.d.;Peaks: A = monomer; B = dimer; C = trimer; D = tetramer. Reproduced withpermission from Tsuji, J. Chromatogr. 652, 139 (1993).

2.2.3 Identity of Monoclonal Antibodies (MAbs):Utility of Capillary Isoelectric Focusing (CIEF)

Monoclonal antibodies (MAbs) are increasingly used in a variety of diagnosticand therapeutic applications. According to a 1993 survey by the Pharmaceuti-cal Manufacturers Association, 35% of biotechnology medicines in clinicaltrials in the United States were MAbs. Monoclonal antibodies are also used ashighly selective agents for the affinity purification of recombinant proteins,allowing the recovery of the protein antigen from crude mixtures and the re-moval of contaminants from a preparation. According to FDA regulations,MAbs used in the purification process must have quality control proceduresessentially equivalent to those for the drugs they are being used to purify.

2-10

As is the case for most of the proteins used in biotechnology, the composi-tion of MAbs is often heterogeneous (that is the product consists of a popula-tion of related molecules rather than a singular molecular species). For MAbs,this heterogeneity is due mainly to differing post-translational modifications.Such heterogeneity makes high-resolution analytical techniques particularlyimportant. Traditionally, analytical biotechnologists have turned to HPLC, inparticular RP-HPLC, to provide high-resolution analysis. Developing high-resolution chromatographic techniques for MAbs, however, has proven ex-tremely difficult. For example, with RP-HPLC, poor peak shapes andextremely low recoveries are often obtained due to adsorption of the MAb onthe reversed-phase packing material.

CE—in the free solution CZE, gel, or IEF modes—provides the analystwith high resolution tools to study MAbs. CE offers all of the advantages ofclassical, HPLC techniques in terms of automation, analysis time, direct, on-line detection and computerized data storage. CE, however, offers a significantadvantage over RP-HPLC. Newly developed coated capillaries such asBeckman’s Neutral Capillary minimize MAb interactions to the capillary wall,resulting in high-resolution separations. Finally, with the new coated capillar-ies methods development time is greatly reduced, as there is no longer a needto try numerous additives and dynamic coatings to prevent with minimal meth-ods development time. Following is an example of high-resolution MAbanalysis using capillary isoelectric focusing in a coated capillary.

One important application of IEF is the determination of a protein’s iso-electric point (pI). As in classical IEF, the CIEF method involves calibrationwith standard proteins whose pIs are known. A plot of the mobilization timevs. the pI should exhibit linearity within a defined pH range. In Figure 2-5(Pritchett, 1994), such a calibration plot is shown for the determination of thepIs of components of anti-carcinoembryonic antigen (CEA) MAb. The CIEFmethod described in Section 1.7 (Figure 1-39) was used along with a coatedcapillary (eCAP Neutral, P/N 477441). As shown in the electropherogram, thesample was spiked with internal standards and, from the calibration plot, it wasestimated that the isoelectric points of the anti-CEA MAb varied from 6.73 to7.46. Figure 2-6 shows a separate run of anti-CEA MAb. The difference in themigration times between the two runs is due to differences in sample concen-trations and matrix compositions, and highlights the importance of runninginternal standards for pI determinations.

2-11

Figure 2-5. CIEF Analysis of anti-CEA MAb. The method of Figure 1-39 wasused with an eCAP Neutral coated capillary. The pI was calculated by com-parison with standards.

Figure 2-6. CIEF analysis of anti-CEA MAb without internal standards. Theestimated pIs of the individual zones are indicated.

2-12

2.2.3.1 Immunotoxins: Utility of MECC with P/ACE-LIFDetection

Immunotoxins are biotechnology-derived drugs that consist of cytotoxinsconjugated to monoclonal antibodies. They deliver cytotoxins (e.g., doxorubi-cin) to the site of cancerous cells at relatively high local drug concentrations,while allowing the total systemic dose to be relatively low. The specificity ofthe antibody avoids the toxicity to normal cells (particularly those of the bonemarrow and gastrointestinal tract) characteristic of conventional anticancerdrug therapy.

Hughes and Richberg (1993, 1994) have shown that CE is an excellenttechnique to examine doxorubucin-linked chimeric antibodies. While UVdetection can be used, P/ACE with laser-induced fluorescence (LIF) detectionprovides both high sensitivity and selectivity for the analysis of antibody-conjugated anthracyclines: only those protein species that are conjugated withthe fluorescent drug are detected. Detectability with LIF is typically severalorders of magnitude better than that with UV absorbance (for a review of CE-LIF applications, see Schwartz et al., 1994). With drug-labeled antibodies,different drug–antibody ratios and conformations are possible. In an SDS-containing run buffer (i.e., MECC conditions were used), the doxorubicin-conjugated antibody was separated into three species, as shown in theelectropherogram of Figure 2-7 (peaks 1, 2, 3). These peaks are well separatedfrom the unconjugated drug (peak 6) which separates on the basis of micellarpartitioning (see Sections 1.3.2.1 and 1.4.3.1 for MECC principles). Also ap-pearing in the electropherogram are the conjugated light and heavy chains(peaks 4, 5). Thus CE-LIF permits sensitive detection of doxorubicin and itsconjugated antibody species. Examination of the antibody conjugate at pico-molar levels would be difficult, if not impossible, by other separation/detectiontechniques.

2-13

0.00

0.000

0.500

1.000

1.200

2.00 4.00 6.00 8.00Time (min)

Cha

nnel

A: F

luor

esce

nce

1

2

3

4

56

Figure 2-7. CE-LIF electropherogram of a 420-pmol solution of IgG antibodyconjugated with doxorubicin. Peak identification: 1, 2, 3 = antibody conju-gate; 4, 5 = conjugated heavy and light chains; 6 = unconjugated doxorubicin.From Hughes and Richberg, Beckman Application Information BulletinA-1763.

2.2.4 Confirmation of Peak Identity byCE-Mass Spectrometry (CE-MS)

In principle, the combination of CE with a mass spectrometer would permitdirect identification of unknown peaks (e.g., peaks a to f in Figure 2-1). WhileCE-MS is not routine yet, significant progress has been made lately in proteinand peptide applications (see reviews by Smith et al., 1994; Pleasance andThibault, 1993).

Electrospray MS has revolutionized the way in which large—and some-times labile—biomolecules are analyzed by MS. Electrospray is a “soft” ion-ization technique which produces an abundance of molecular ions relative to

2-14

ion fragments. Multiply charged protein or peptide molecular ions permit MWmeasurements on single quadropole MS instruments. A CE-MS electrosprayinterface is commercially available from Beckman (P/N’s 727616 and 727617for interfacing with the Finnigan and PE Sciex mass spectrometers, respec-tively). In contrast to HPLC-MS, with CE-MS the effluent flow from the capil-lary is not sufficient to generate an effective electrospray. The interface has acoaxial flow design in which a sheath liquid flow (“make-up buffer”) is used toeffect the spray of capillary effluent into the ion source. Figure 2-8 shows aschematic drawing of the P/ACE CE-MS interface. The capillary extends tothe very tip of the needle assembly where the electrospray is produced with theaid of a dry nitrogen gas. This type of coaxial flow design produces minimalzone broadening; hence, the integrity of the electrophoretic profile is com-pletely maintained.

Figure 2-8. Schematic of the CE-MS interface. Reproduced with permissionfrom Tomlinson et al., Electrophoresis 15, 62 (1994).

As new and even more sensitive MS methods are being developed, it isexpected that CE-MS will become an extremely powerful tool for proteincharacterization. With the electrospray MS, proteins with MWs exceeding100,000 Daltons have been analyzed with excellent accuracy. The added di-mension of CE-MS requires some restrictions in the choice of (make-up) buff-ers and capillaries (Smith et al., 1994) as volatile buffers such as ammoniumacetate must be used.

An example of the utility of CE-MS for MW determination is shownin Figure 2-9 from Tsuji et al. (1992). Recombinant porcine somatotropin

2-15

(rpSt) was analyzed using P/ACE coupled to a Vestec electrospray MS.A characteristic pattern of multiply charged ion clusters can be seen rangingfrom m/z1363.2 (16 charges) to m/z1982.5 (11 charges). Software deconvolu-tion of the mass spectrum resulted in an average MW of 21798.3 ± 3.6, closeto the theoretical value of 21797.9. The CE-MS method allowed identificationof both mono- and dioxidized homologs in samples of rbSt and rpSt.

2000190018001700160015001400130012001100m/z

1363.2

1454.5

1558.1

1678.2

1817.3

1982.5

0

10

20

30

40

50

60

Rel

ativ

e A

bund

ance

Figure 2-9. Electrospray CE-MS of the major rpSt peak showing multiplycharged ion clusters (m/z). Conditions: electrolyte buffer, 20% acetonitrile in20 mM ammonium acetate, pH 9.0; makeup solution, 5% acetic acid in 50%methanol. Reproduced with permission from Tsuji, Anal. Chem. 64, 1868 (1992).

Fragmentation patterns of molecular ions, needed for sequence informa-tion, can be obtained with triple quadropole MS units. For example, using on-line CE-MS, Tomlinson et al. (1994) have shown preliminary results on thesequencing of peptides which were shown to bind class I major histocompat-ibility complex molecules. The CE method involved isotachophoresis (ITP)preconcentration of sample components to improve peptide detectability, asdiscussed in Section 1.5.

ITP preconcentration of proteins was also utilized by Thompson et al.(1993); full scan CE-MS data of proteins were obtained at the 10-7 M level.Finally, Kelly et al. (1993) demonstrated the utility of CE-MS for the identifi-cation of protein glycoforms using a P/ACE system combined with a PerkinElmer-Sciex triple quadropole mass spectrometer. Contour plots (intensity vs.m/z vs. time) were used to identify closely related glycoforms present at agiven attachment site. It was possible to identify unambiguously oligosaccha-rides in digests of complex-type glycoproteins (e.g., horseradish peroxidase).

2-16

2.3 QuantityProviding information on the mass or concentration of a given substance in avariety of matrices is the goal of quantity assays. Knowledge of concentrationis essential for every aspect of drug development and manufacturing, includingprocess control (e.g., calculation of yield), pre-clinical studies (e.g., animaltoxicity), and clinical trials (e.g., for dose escalation studies and to monitorserum concentration). The crucial importance of quantitative assays, even earlyin the development cycle, is underscored in the FDA’s Guideline for Submit-ting Documentation for the Manufacture of and Controls for Drug Products,Section F.1, which states, “The product tests and specifications appropriate toinvestigational drug products are, understandably, not as well developed aswhen an NDA is submitted . . .. Information should also be submitted to sup-port the specificity, linearity, precision, and accuracy applicable to specificquantitative methods used to test the dosage form.”

Table 2-2 lists commonly used, traditional methods for protein quantita-tion. As detailed in the following sections, several modes of CE can providesolutions to protein quantitation problems: CZE, MECC, CIEF, and SDS-CGE. Affinity capillary electrophoresis (ACE) appears very promising for thestudy of receptor–ligand reaction kinetics and to quantitate protein binding(e.g., with drugs, DNA, antigens), e.g., in immunoassays.

Table 2-2. Methods Commonly Used forCharacterization of Proteins: Quantity

• Amino acid analysis• Spectroscopy

- UV extinction coefficient- Tyrosine titration difference spectroscopy- Colorimetric methods

• HPLC• Electrophoresis (SDS-PAGE)• Immunoassays

- ELISA- RIA

2-17

2.3.1 Accuracy and Precision in Quantitative CEAnalysis

As pointed out in Section 1.4, CE protein analysis, when performed in baresilica capillaries, often requires special capillary pretreatment procedures andadditives to prevent protein interactions with and adsorption to the capillarywall. In addition, special rinsing protocols are often necessary between runs toremove adsorbed protein. These steps are needed because proteins fouling thecapillary wall can cause severe decreases in quantitative accuracy and preci-sion. Further evidence for the importance of certain buffer additives on theCZE analysis of MAb preparations in uncoated capillaries was presented byGuzman et al. (1992). They found that improper separation conditions maylead to significant errors in the quantitation of proteins in pharmaceuticalpreparations. Addition of alkylamines and/or zwitterions (see also Section1.4.2.3) was recommended to improve the performance of separation and toenhance the resolution and reproducibility of the protein analytes.

The new Beckman coated capillaries for CZE and SDS-CGE are designedto minimize protein–capillary wall interactions. They often give superior pro-tein separations compared to untreated capillaries, even when the bare silicacapillaries are used in conjunction with buffer additives (see examples in fol-lowing sections). In addition, the time formerly spent finding the right pretreat-ments, additives, and rinses is eliminated, greatly reducing methoddevelopment time.

The complexity and/or variability of the sample matrix also plays an im-portant role with regard to quantitative precision and accuracy. Many first-timeusers of CE equipment have the experience that results with “clean” samplesare often superior to those obtained with “dirty” or complex samples. Manysuch problems are the result of adsorption of proteins to the capillary wall andcan be avoided by using the aforementioned coated capillaries, or goingthrough the process of finding the right additives.

The quality of the separation also affects the analytical accuracy and pre-cision. With well-defined, baseline-separated peaks, precision is generallyexcellent, i.e., typically < 0.5% RSD for migration times and ≈ 0.5 to 3.0%RSD for peak areas.

When low-level impurities are quantitated, precision decreases as not onlyresolution but also peak detectability (signal-to-noise ratio) becomes an issue.For example, Bullock (1993) quantified with CZE low-level impurities inrecombinant human interleukin-4 (rhIL-4) preparations with RSDs varying

2-18

from 5 to 26% (Table 2-3). However, the precision in the overall purity deter-mination of rhIL-4 was excellent: 0.843% (n = 8). The use of on-capillarysample concentration techniques, discussed in Section 1.5, provides a possibleway to improve detectability (with concomitant improvement of precision) oflow-level impurities.

Table 2-3. Precision Data for the Purity Determination of rhIL-4 by CZE

rhIL-4 Impurity 1 Impurity 2 Impurity 3 Impurity 4

Mean 79.75% 1.16% 1.72% 11.95% 5.42%(n = 8) (peak area%)S.D. 0.6724 0.132 0.457 0.651 0.815(peak area%)RSD. (%) 0.843 11.4 26.6 5.45 15.0

From: Bullock, J. Chromatogr. 633, 235 (1993)

Finally, automated data systems with suitable peak-defining algorithmsfor CE such as Beckman’s Gold are imperative to achieve optimum quantita-tive precision and accuracy from CE runs. Gold software is designed for quan-titation with both HPLC and CE. CE and HPLC systems can be simultaneouslyoperated and controlled from one computer. Gold allows for internal and exter-nal standardization and the detector response can be tracked using single- ormultilevel concentration calibrations (see, for example, the insets in Figures2-10, 2-12, 2-13). In electropherograms, up to three reference peaks may beselected for precise alignment and comparison of data. Examples of the utilityof the various CE modes for quantitation purposes are described next. Theyshow that CE can provide rapid, accurate, and precise protein assays in oftencomplex media.

2.3.2 Quantitation of Dosage Forms by CIEFPharmaceutical and diagnostic products typically do not consist of pure drugsubstances, but rather of formulated mixtures. Excipients are added for severalreasons (for example to enhance stability during storage, or to provide protec-tion during the manufacturing process). Many of the excipients used in biop-harmaceutical products are themselves proteins (for instance, human serumalbumin), which adds significantly to the challenges of developing a goodquantitative method. In addition to its speed and high resolution, a decidedadvantage of CIEF for dosage form analysis is that, since detection is accom-plished at 280 nm, many small molecule excipients are transparent and thus donot interfere with the analysis.

2-19

MAbs were quantitated in dosage forms by Pritchett et al. (1994) usingthe recently introduced eCAP Neutral coated capillary. Figure 2-10 shows theCIEF analysis (method of Figure 1-39, Section 1.7.3.2) of anti-tumor necrosisfactor (TNF) MAb. For these types of assays, the method should have suffi-cient sensitivity and range to measure low and high levels of the drug in vari-ous investigational and licensed dosage forms. In addition, resolution shouldbe adequate to separate the active drug from the excipients. The electrophero-gram shows two protein peaks, the MAb, and human serum albumin (HSA)which was included in this experimental formulation. The CIEF method’sdynamic range was from ≈ 5 to 250 mg/mL, as shown in the insert of Figure 2-10. Table 2-4 lists the results of the CIEF assay as percent of label along withthe precision of three measurements. It is our experience that the peak areaprecision with the above CIEF method is quite good (≈ 0.5 to 2% RSD) andeven slightly better than with other CE methods. This may be due to the factthat, with the CIEF method, the entire capillary is filled with sample. In con-trast, only a small fraction of the capillary is filled with sample in the CZE orSDS-CGE methods.

Figures 2-10. Quantitative CIEF analysis of anti-TNF dosage form. Condi-tions as in Figure 2-5.

2-20

Table 2-4. Quantitative CIEF Analysis of Anti-TNF MAb: Assay Percent ofTwo Dosage Forms

Dosage Form % of Label

0.5 mg vial 101.0101.8102.2

Mean 101.7RSD (%) 0.6

1.0 mg vial 99.299.298.6

Mean 99.0RSD (%) 0.35

In a similar application (involving the same CIEF method and also per-formed on a P/ACE instrument), Yowell et al. (1993) determined granulocytemacrophage colony stimulating factor (GM-CSF) in the presence of HSA(Figure 2-11). The results of three different dosage forms are shown inTable 2-5 along with results obtained by CZE and reversed-phase HPLC.Dosage forms of GM-CSF showed 5 to 6 major components in the electro-pherogram which were well separated from GM-CSF. The figure also shows arun of recombinant granulocyte colony stimulating factor (G-CSF), trade nameNeupogen, which has a different pI. CIEF has excellent potential for routineQC testing. Compared to HPLC, it lacks the problems associated with cloggedcheck valves, leaking seals, large solvent consumption (and waste production),and relatively long analysis times.

Table 2-5. Comparison of FSCE, CIEF, and HPLC Analysesfor GM-CSF Dosage Forms

Sample RP-HPLC FSCE CIEF

1-SFG-303 98.45 105.37 102.440-SFG-303 98.73 99.80 104.320-SFG-304 104.63 102.22 105.21

From Yowell, et al., J. Chromatogr. 652, 215 (1993).

2-21

0.025

0.020

0.010

0.000

-0.0050.50 4.00 8.00 12.00 16.00

0.023

0.020

0.010

0.000

-0.003

GCSF (Neupogen)

GM-CSF

HSA

Time (min)

Cha

nnel

A (

Abs

orba

nce)

Cha

nnel

B (

Abs

orba

nce)

Figure 2-11. A CIEF electropherogram of GM-CSF and G-CSF (Neupogen)showing different pI values as a method for product identification. FromYowell et al., Beckman Application Information Bulletin A-1744.

2.3.3 Quantitation with CZE and SDS-CGEFigure 2-12 shows the CZE assay of a MAb, anti-CEA, in serum-free tissueculture medium (Pritchett et al., 1994). A 20-mM citrate, pH 3, run buffer andeCAP Neutral capillary were used. The method had a linear dynamic range of5 to 1000 mg/mL and the analysis time was less than 10 minutes. In our expe-rience, and that of others (Yowell et al., 1993), the linear dynamic range withCZE is typically higher (5 to 10 times) than with CIEF; since the latter methodis more prone to protein precipitation at higher concentrations. The peak areaprecision from a variety of studies with sample sizes ranging from 3 to 9 wasbetween 2.5% and 3% RSD. To determine the recovery, purified MAb wasspiked into Hybrimax (Sigma) medium at the levels of 25 and 50 µg/mL. Thepercent recovery found at these levels was 97.2 and 98.6%, respectively.

2-22

Figure 2-12. CZE analysis of anti-CEA MAb in tissue culture. Capillary,eCAP Neutral coated, 37 (30) cm × 50 µm; run buffer, 20 mM citrate pH 3;run temperature, 20°C; field strength, 400 V/cm; detection, 214 nm.

Evidence that SDS-CGE can also be used as a quantitative method waspresented by Pritchett et al. (1994) and Tsuji (1993). In the former, the sameMAb as that of Figure 2-12, anti-CEA, was analyzed by in a serum-containingtissue culture medium (10% newborn calf serum in QBSF-52 [Sigma]) usingthe Beckman eCAP SDS 14-200 kit (see Section 1.6 for more information onthis method development kit). Figure 2-13 shows the electropherograms ofanti-CEA MAb in the serum-containing medium along with the blank run.At the 25 mg/mL level, the recovery was 98% (RSD 1.4%). Using the samecoated capillary and gel buffer, Tsuji (1993) determined the percent monomer,dimer, trimer, and tetramer in recombinant bovine somatotropin (rbSt)samples. The SDS-CGE electropherogram of a dimer-enriched rbSt samplewas shown earlier in Figure 2-4. Composition of monomer and dimer inrbSt was 96 and 4% (tri- and tetramer < 1%) with RSDs of 0.2 and 4%,respectively (n = 7).

2-23

Figure 2-13. Quantitative analysis of anti-CEA in serum-containing medium.Capillary, eCAP SDS 14-200, 27 (20) cm × 100 µm; field strength, 250 V/cm;run temperature, 20°C; detection, 200 nm.

2.3.4 Immunoassays Using Affinity CE (ACE)The principles of ACE were outlined in Section 1.9. Immunoassays are widelyused for quantitative studies of proteins, particularly in clinical and diagnosticapplications. However, the UV absorbance detection used in conjunction withCE is generally not sensitive enough for detection in biological media at verylow protein levels, i.e., concentrations < 10-6 M. Laser-induced fluorescencedetection (LIF) provides sensitivity several orders of magnitude higher thanUV absorbance and is now commercially available (P/N 477125). The P/ACELIF detector permits coupling to various laser sources and can be used in awide variety of bioapplications (Schwartz et al., 1994).

Chen and Sternberg (1994) described a competitive immunoassay fordetermination of digoxin in human serum using the green He-Ne laser emittingat 543.5 nm. The antigen was conjugated with a dye-labeled oligonucleotide,(dT)10. The laser dye used was tetramethylrhodamine (TMR). The oligonucle-otide serves as a charge modifier so that the species of interest are well re-solved with CE. The competitive immunoassay can be described by theequilibrium

Ag + Ag* + Ab Ag-Ab + Ag*-Ab + Ag

2-24

where Ag = digoxin, Ag* = TMR-(dT)10-digoxigenin, Ab = Fab fragment, andAg-Ab and Ag*-Ab are the antigen–antibody complexes. Digoxin calibrationwith CE-LIF is demonstrated in Figure 2-14 with serum-based digoxin calibra-tors of 0.42, 2.72, and 5.13 ng/mL, respectively. As the digoxin concentrationis increased in the sample, the amount of free Ag* also increases whileAg*-Ab decreases. Using the ratio of the areas of free Ag* and the total fluo-rescence signal in each electropherogram, a digoxin calibration curve can beconstructed. In this immunoassay, both the free Ag* and the antibody-com-plexed Ag* can be measured simultaneously. This approach to immunoassayusing ACE with LIF detection should be promising for a variety of other clini-cal and diagnostic applications.

2-25

Rel

ativ

e F

luor

esce

nce

Inte

nsity

Time (min)

20

15

10

5

02 3 4 5 6 7 8

2

2

2

A

B

C

1

1

1

Per

cent

age

of F

ree

Labe

l

Digoxin, ng/mL

70

0.4210

20

30

40

50

60

2.72 5.21

Figure 2-14. (Top) Digoxin immunoassay with serum calibrator at (A) 0.42,(B) 2.72 and (C) 5.21 ng/mL, respectively. Peaks: 1, digoxigenin-3'-(dT)10-TMR; 2, antigen–antibody complex of the Fab. (Bottom) Calibration curve.Percentage of free digoxigenin-3'-(dT)10-TMR vs. digoxin calibrates. Repro-duced with permission from Chen and Sternberg, Electrophoresis 15,13 (1994).

2-26

2.4 PurityOnce it can be proven what is in the container (identity), and how much isthere (quantity), the next question often asked is: How pure is the active com-ponent? In pharmaceutical biotechnology, the requirements are often such thatminor amounts of impurities and degradation products must be determined inthe presence of much larger quantities of the main component. In addition,these components may be structurally very similar to the main component.Here, the resolving power, sensitivity, quantitation capability, and speed of CEoffer solutions to these analytical problems. Methods commonly used forpurity determination are listed in Table 2-6. CE can be added to this list, eitheras a stand-alone technique or used in conjunction with these methods. Thefollowing examples illustrate the utility of CE for purity assays.

Table 2-6. Methods Commonly Used for Characterization of Proteins: Purity

• Chromatography• Electrophoresis• Residual DNA testing• Microbiological testing• Tests for process additives• Immunological methods

2.4.1 Purity of ProteinsDuring the production of MAbs and rDNA proteins, process monitoring isrequired in order to assure that, among other things, the required purity levelsare achieved and maintained. The recovery and purification of the productfrom a fermentation broth typically involve various procedures, includingfiltration, centrifugation, and chromatography. After each purification step, andespecially after the final step, the intermediate or bulk product must be checkedto see if the desired purity level has been achieved. In addition to its controlfunction, this purity information is frequently used to further optimize thepurification process. The CZE, CIEF, and SDS-CGE modes of CE are all wellsuited for such purity checks.

2.4.1.1 Purity Check by CZE and CIEF

Figure 2-15 shows an example of a purity check by CZE for recombinanthirudin, an inhibitor of α-thrombin. The top panel (A) shows the electrophero-gram of a sample extracted from a fermentation broth, showing more than

2-27

30 constituents. Hirudin-65 is a 65-amino-acid recombinant protein. Twoprocess by-products, hirudin-64 and hirudin-63, are present in this sample andelute just after the major peak (migration time, 20.09 min). Even though thesecompounds are structurally similar to hirudin-65, they can be baseline sepa-rated from each other by CZE, thus showing its resolving power (Ludi et al.1988). The bottom panel (B) of Figure 2-15 shows a purity check at a certainstage in the purification process. The two minor components eluting at 20.27and 20.44 minutes are hirudin-64 and hirudin-63, respectively. Other, similarexamples of the use of CZE and CIEF in the purity of recombinant proteinscan be found in publications by Palmieri (1989), Arcelloni et al. (1993), Frenzet al. (1989), and Grossman et al. (1989)—dealing with the analysis of insulinand hGH by CZE and CIEF; Bullock (1993)—Analysis of rhIL-4 by CZE;Yowell et al. (1993)—Analysis of GM-CSF by CIEF.

0 10 20 30Time (min)

Abs

orba

nce

(214

nm

)

0.000

0.010

0.020

0.030

0.040 20.0

920

.28

20.4

7

A

0 10 20 30Time (min)

Abs

orba

nce

(214

nm

)

0.000

0.010

0.020

0.030

0.040

20.1

120

.27

20.4

4

B

Figure 2-15. (A) Separation by CZE of an r-hirudin-containing sample from afermentation broth. (B) Purified r-hirudin sample. Run buffer: 20 mM tricine,10 mM sodium borate, 1 mM EDTA, 0.2 mM diaminobutane, pH 8.4. FromPaulus and Gassman, Beckman Application Data Sheet DS-752.

2-28

2.4.1.2 Purity Check by SDS-CGE

A biotechnology company, Hemosol (Etobicoke, Ontario, Canada), used SDS-CGE to check the purity of a potential blood substitute product, Hemolink,during the manufacturing process.

Outside the red blood cell, hemoglobin does not readily release its oxygento body tissue. It can break down and damage the kidneys during excretion. Tobe an effective, efficient, and safe oxygen carrier, hemoglobin must be stabi-lized to prevent its breakdown and have an oxygen-carrying capacity similar tothat of whole red blood cells. Hemosol’s proprietary crosslinked product,Hemolink, involves the stabilization of a highly purified hemoglobin throughcrosslinking with a reagent prepared from raffinose. The hemoglobin is modi-fied to stabilize it against fragmentation into hemoglobin half-molecules, andto have oxygen binding and release properties similar to that of whole redblood cells. Figure 2-16 shows the electropherogram of a Hemolink sample.

8.00 10.00 15.00 20.00 25.00

0.0200

0.0400

0.0600

Abs

orba

nce

0.0800

10.8

1 O

rang

e G

12.4

2

13.0

3

Figure 2-16. SDS-CGE of Hemolink, a crosslinked and polymerized purifiedhemoglobin-based oxygen carrier. The standard procedure outlined in theeCAP SDS 200 kit was followed. Sample injection, 60 s, pressure. Orange Gwas used as a reference standard. Courtesy of Dr. David Wicks, Hemosol,Inc., Etobicoke, Ontario, Canada.

2-29

At Biomira, Edmonton (Canada) SDS-CGE is used to monitor the purityof raw monoclonal antibody (IgG1). The latter is used in the manufacturing ofimmunoscintography kits for adenocarcinomas. The MAb is processed by apatented photoactivation method into a product, Tru-Scint AD, which is radio-labeled with Technetium (99m) at the clinic.

The MAb product is monitored for its consistency and purity by the Beck-man SDS-200 kit on a routine basis (Figure 2-17). The percent of light andheavy chains generated during the manufacturing process, as well as the mono-meric antibody content, can be quantified.

8.00 10.00 15.00 20.00 25.00

0.0100

0.0200

0.0300

0.0400

Abs

orba

nce

Time (min)

11.6

7 R

EF

-1

14.9

9 15.0

8

16.4

1 P

eak-

2

17.7

1 P

eak-

3

19.9

4 P

eak-

420

.77

Pea

k-5

13.5

4

8.00 10.00 15.00 20.00 25.00

0.0100

0.0200

0.0300

0.0400

Abs

orba

nce

Time (min)

15.0

1 lig

ht 17.5

9 he

avy

11.5

6 R

EF

-1

Figure 2-17. Quality control of antibodies by SDS-CGE. Top: product;Bottom: reduced antibody. Courtesy of Dr. A. Abdul-Wajid, Biomira, Inc.,Edmonton, Canada.

2-30

2.4.2 Screening the Purity of PeptidesPurity checks of peptide products are needed in the biopharmaceutical industryfor process control and QC of bulk peptides and final peptide products. Thesame criteria as discussed previously for protein products (Section 2.4.1) alsoapply to peptides. On a laboratory bench scale, purity checks of newly synthe-sized compounds generated on a peptide synthesizer are usually required.

In protein characterization studies involving N-terminal sequencing andpeptide mapping, it is highly desirable to screen the purity of peptide fractionscollected from HPLC by another (i.e., non-HPLC), complementary method.Peptides often co-elute on HPLC columns, especially when their hydrophobic-ity is similar (Strickland, 1991; Grossman et al., 1989). Thus, a quick puritycheck with CE would save valuable sequencer run time (and sample) as im-pure fractions cannot be easily sequenced and interpreted.

CZE is ideally suited to quickly check the purity of collected peptidefractions for two reasons: 1) very little sample (nL) is consumed so that almostall of the sample can be used for subsequent sequence analysis; 2) the separa-tion mechanism is different from reversed-phase HPLC so that peptides similarin hydrophobicity can be resolved based on differences in charge-to-massratio. Thus, peaks appearing pure after HPLC analysis are often resolved inmultiple peaks with a CE check (Strickland, 1991). In that case, a further puri-fication step of the fraction might be considered. For further information onpeptide separation conditions using CZE, see Section 1.3, and the papers byMcCormick (1994), Langenhuizen and Janssen (1993), and Grossman et al.(1989).

2.5 HeterogeneityMost purified proteins, even in their native state, consist not of a single mo-lecular species, but rather are composed of a population of related molecules.One source of this heterogeneity is differences in amino acid sequence, due tomutant amino acids and modification of amino acid side chains (e.g., deamida-tion of Asp or Gln; oxidation of Met). Differential, post-translational modifica-tions provide additional heterogeneity. Common, post-translationally inducedheterogeneities include variable polypeptide chain length resulting from differ-ential cleavage by maturation enzymes (e.g., some of the pro-protein speciesmay be present in the population) and differential glycosylation of glycopro-

2-31

teins. The latter can exist at several levels of carbohydrate structure: 1) theremay be different numbers of oligosaccharide chains; 2) the oligosaccharidechains may have different monosaccharide composition (e.g., variable sialyla-tion and fucosylation); 3) the oligosaccharides may have different monosac-charide sequences; 4) the oligosaccharides may have differing anomericlinkages among their monosaccharides.

CE offers valuable solutions to the analytical challenge of protein hetero-geneity. Especially useful are CZE (e.g., peptide mapping), MECC, and CIEFmodes, while CE-MS provides unique capabilities for identification of iso-forms.

2.5.1 Monoclonal AntibodiesPreparations of monoclonal antibodies often show considerable heterogeneityeven though the amino acid sequence of the various forms observed is thesame. This may be due to post-translational modifications involving glycosyla-tion. It is important to monitor changes in glycosylation patterns because suchparameters as serum half-life and bioactivity of the product may be adverselyaffected. The earlier-discussed example of Figure 2-6 (anti-CEA MAb usingCIEF) shows a typical heterogenous distribution of glycosylated antibodyisoforms with pIs ranging from 6.7 to 7.5.

2.5.2 Glycoforms of Recombinant Proteins.Recombinant human tissue plasminogen activator (rtPA) is a fibrin-specificplasminogen activator used in the treatment of heart disease. The protein con-sists of 527 amino acids (MW ≈ 60,000) with 4 possible glycosylation sites.Yim (1991) examined glycoforms of rtPA with both CZE and CIEF. Twoglycosylation variants of rtPA, named Type I and Type 2, were examined.Even though these variants contained the same amino acid composition, morethan 20 peaks were visible in the electropherograms, reflecting the heterogene-ity present in the two preparations. The number of potential glycoforms inproteins increases with the number of glycosylation sites. In addition, onlysome changes in glycosylation effect the protein’s pI. Each peak may, there-fore, represent a combination of glycoforms with the same pI. Resolutionobtained with Yim’s CIEF method was better than that with CZE. Treatmentof the samples with neuramidase (which removes the sialic acids) dramaticallysimplified the electropherogram patterns. This would indicate that differentlevels of sialylation are largely responsible for the heterogeneity.

2-32

The heterogeneity of recombinant human erythropoietin (rHuEPO) hasbeen characterized by various researchers (Watson and Yao, 1993 and 1994;Rush et al., 1993; Tran et al., 1991) using CZE and CE-based tryptic mapping.This 165-amino-acid glycoprotein has a MW of 34,000 to 38,000 while 40%of its weight consists of carbohydrate. Figure 2-18 shows the separation ofrHuEPO glycoforms with the Beckman eCAP Amine coated capillary. Themigration of the glycoforms is in the order of decreasing number of sialic acids(Watson et al., 1994). The CZE method, while yielding slightly less resolutionthen slab gel IEF, offers the potential for quantitation in much shorter analysistimes than with IEF.

0

.002

.004

.006

.008

0 10 20 30

1

2

3

4

5

6

Time (min)

Abs

orba

nce

(200

nm

)

Figure 2-18. CZE of r-HuEPO into six glycoforms on a eCAP Amine coatedcapillary. Voltage, 15 kV; sample concentration, 0.3 mg/mL; run buffer,200 mM sodium phosphate, pH 4.0. Reproduced with permission from Watsonet al. (1994).

The separation of recombinant human interferon-γ (IFN-γ) under MECCconditions was described by James et al. (1994). With this glycoprotein,N-linked oligosaccharides may be attached to the 17-kDa protein backbone atone or two glycosylation sites, or may be absent. The electropherogram shownin Figure 2-19 reveals the difference in glycoform variants at these sites. Itshould be noted that, prior to the MECC analysis, the protein was denaturedwith SDS to prevent dimer formation. Further work by this group demon-

2-33

strated that CE can be used for real-time monitoring of glycosidase digestion.Other publications in which CZE is used for glycoprotein microheterogeneityanalysis are Rudd et al. (1992) and Grossman et al. (1989)—ribonucleases;and Landers et al. (1992)—ovalbumin. Kelly et al. (1993) used the powerfulcombination of CE and electrospray MS to characterize ribonuclease, ovalbu-min, and horseradish peroxidase (see Section 2.2.4).

Figure 2-19. Separation of IFN-λ glycoforms by MECC. RecombinantHuIFN-λ (1 mg/mL in 50 mM borate, 50 mM SDS, pH 8.5) was injected for5 s prior to electrophoresis at 22 kV (120 µA) in 400 mM borate, 100 mMSDS, pH 8.5. Peak groups are designated as: 2N, two glycosylation sites occu-pied; 1N, one glycosylation site occupied: 0N, no glycosylation sites occupied.From James, et al., Beckman Application Information Bulletin A-1761.

2.5.3 Heterogeneity in Proteins Relevant toClinical/Diagnostic Applications

The utility of CE in clinical/diagnostic research is reviewed in several papers(see, for example, the special issue (Volume 15) of Electrophoresis, January1994) and the book chapter by Klein and Jolliff (1994). Full discussion of thistopic is beyond the scope of this book. The following papers deal with proteinheterogeneity analysis by CE:

• Human transferrins. The heterogeneity results from differences in ironcontent, genetic polymorphism, and differences in glycosylation. Kilarand Hjerten (1989) used CIEF and CZE to study the glycosylation pat-terns in transferrin samples. Detection at 280 nm (for protein) and 460 nm(for iron) was used to identify the different transferrin isoforms (iron-free,monoferric, and diferric complexes).

2-34

• Hemoglobin variants. Zhu et al. (1992,1993) and Molteni et al. (1994)describe the utility of CZE and CIEF.

• Isoenzyme analysis. Klein and Jolliff (1994) proposed a new concept forlactate dehydrogenase isoenzyme analysis.

• Metallothionein isoforms. Beattie et al. (1993, 1994) used CZE andMECC for this type of analysis.

• Ferritin analysis. Zhao et al. (1994) used CZE for study of the apo-, holo-,recombinant-, and subunit-dissociated ferritins.

2.6 StabilityMethods commonly used for control and characterization of the stability ofprotein products are listed in Table 2-7. Physical and chemical instabilities ofproteins can be related to such parameters as:

1) Chemical reactions of amino acids. For example, deamidation of aspar-agine and glutamine residues is common, as is oxidation of methionineand isoaspartate formation. The high resolving power of CE can be usedto monitor the stability of products during storage. For example, Palmieri(1989), Frenz et al., (1989), and Grossman et al. (1989) demonstrated thatCZE readily resolves the deamidated forms of hGH from nondegradedhGH. Figure 2-20 shows the difference between preparations of natural,pituitary-derived hGH (A) and recombinant hGH (B). The former shows alarger percent of mono- and dideamidated by-products. The desamidoderivatives have an increased net negative charge due to the replacementof the side chain amide group with an ionized carboxylate. CIEF is also avery useful tool for these types of studies, as was demonstrated by Frenzet al. (1989) for hGH characterization. Peptide mapping (either by HPLCor CE) is another sensitive technique to detect subtle changes in aminoacid composition. It has been shown that single amino acid deletions in apolypeptide chains can be detected by peptide mapping (Garnick et al.1988).

2-35

0 10 20 30Time (min)

-0.005

0.020

Abs

orba

nce

(206

nm

) A

0 10 20 30Time (min)

-0.005

0.010

Abs

orba

nce

(206

nm

)

B

Figure 2-20. ( A) Separation of recombinant human growth hormone by CZE.(B) Separation of natural human growth hormone; run buffer, 100 mM boratepH 8.5; 7.5 kV; 20°C. From Palmieri, Beckman Application Data SheetDS-749.

2) Deglycosylation of glycoproteins. Glycoproteins from different manufac-turing lots may differ in glycosylation pattern. Additionally, process-induced degradation may occur. Sialic acid groups, which have a bulkycarboxyl group near the glycosidic bond, are particularly labile and proneto loss. CE can monitor deglycosylation, as shown in Figure 2-21. Theenzymatically induced deglycosylation was followed over a 20-hour pe-riod. First an increase in singly glycosylated forms at the expense of dou-

2-36

bly glycosylated forms was observed. After 20 h, all of the N-linked oli-gosaccharides had been removed from the protein backbone.

Figure 2-21. Digestion of IFN-λ by PNGase F. Samples of the digest wereinjected for 4 s at the specified intervals and resolved by the electrophoreticconditions specified in Figure 2-19. From James et al., Beckman ApplicationInformation Bulletin A-1761.

3) Peptide bond cleavage. In general, these types of instabilities are easilydetected by chromatographic or electrophoretic techniques, as the molecu-lar weight and also the hydrophobicity and/or the mass-to-charge ratio ofthe protein changes.

4) Disulfide bond formation or reduction and scrambling. Disulfidebonds are often required for biological activity. In cases where disulfidebond formation is undesirable, gentle oxidation can be carried out withoutdimerization which would occur under stronger conditions. CE can beutilized to monitor disulfide bond formation, as demonstrated by Landerset al., (1993). Figure 2-22 shows the formation of homo- and het-erodimers in the co-oxidation of synthetic peptides (termed Ntc and Ctc).

2-37

Time (minutes)

Abs

orba

nce

(200

nm

)

1 5 10 15 1 5 10 15

A Ntc

Ctc

C

DB

Ntc:Ctc

Ctc:Ctc

Ctc:CtcCtc:Ctc

Ntc:Ntc

Figure 2-22. Co-oxidization of the Ctc and Ntc peptides. A mixture of thepurified Ntc and Ctc monomers was incubated at 27°C and analyzed at 0 min(Panel A), 2 h 46 min (Panel B), 5 h 31 min (Panel C), and 11 h 1 min(Panel D). Separation was carried out in 20 mM citrate buffer, pH 2.5. Barrepresents 0.005 AU. From Oda et al., Beckman Application InformationBulletin A-1757.

5) Conformational changes (denaturation, aggregation). CE can monitor,for example, conformational changes with temperature variations, asshown by Rush et al., (1991), protein folding (Strege and Lagu, 1993),and oligomerization processes (Tsuji, 1993). In Figure 2-23, the peakshape of α-lactalbumin (pI 4.2 to 4.5) is changing when the temperaturecontrol of the capillary is changed from 2 to 50°C. The broadened ormultiple peaks (at 35 to 40°C) do not indicate that the protein sample isimpure, but are a consequence of the fact that both the folded and confor-mationally altered species exist under the electrophoretic run conditions.Other temperature-related effects were observed with myoglobin (on-capillary reduction of the heme group).

2-38

Table 2-7. Methods Commonly Used for Characterizationof Proteins: Stability

• Electrophoresis- SDS-PAGE- CE- IEF

• Chromatography- Size exclusion- Reversed phase

• Specific activity• Differential scanning colorimetry• Spectroscopy• Mass spectrometry• Analytical ultracentrifugation

0.120

0.100

0.080

0.060

0.040

0.020

0.000

10 32 4 5 6

Time (min)

Abs

orba

nce

(214

nm

) 50°C

40°C

35°C

30°C

20°C

45°C

Figure 2-23. Effect of temperature on the electrophoretic behavior of α-lactalbu-min. Reproduced with permission from Rush et al., Anal. Chem. 63, 1346 (1991).

2.7 Process ConsistencyWhen results from the analysis of multiple lots are compared, all of the abovemethods give information about process consistency. An additional examplecan be found the discussion of the CE analysis of Eminase in Section 2.8,following. From this information, control limits are established. Such informa-tion is also very valuable in pointing out areas in need of process optimization.

2-39

2.8 Method ValidationAs required by the demands of good laboratory practice, as well as regulatoryagencies, analytical procedures are typically validated for accuracy, precision,specificity, detection limit, quantitation limit, linearity, and dynamic range.This is particularly important in pharmaceutical biotechnology when themethod is to be used for process or quality control purposes. With the adventof commercial instrumentation (including software allowing for referencepeaks, internal and external standardization) and several years experience inhundreds of laboratories, it has become clear that optimized CE methods canbe validated for process and quality control applications in biotechnology.Such CE methods are just as precise and accurate as methods based on chro-matography. Support for this statement can be found in many recent papers onthe CE of rDNA proteins and MAbs (see for example: Guzman et al., 1992—CZE of MAbs and rDNA proteins; Arcelloni et al., 1993—CZE of hGH;Pascual et al., 1992—CZE of glutathione peroxidase; Harrington et al.,1991—CZE of enzyme labeled IgG; Pande et al., 1992—CZE of bovine serumalbumin; Phillips, 1993—CZE of tissue cytokines; Moring, 1992, andGrossman et al., 1989—quantitative aspects of CE; Bullock, 1993—analysisof rhIL-4; Tsuji, 1993—SDS-CGE of rbSt; Yowell et al., 1993—CIEF andCZE of GM-CSF; Strege and Lagu, 1993—MECC of rDNA proteins in fer-mentation broths.

In many cases in the above papers, the CE method was validated by com-parison with an already existing non-CE method. A case in point is the in-process assay for Eminase, a freeze-dried SmithKline-Beecham product usedin the treatment of myocardial infarction. It is composed of an activator com-plex (lys-plasminogen:streptokinase) and human serum albumin which isadded as a stabilizer. For the manufacturer of eminase, it was important todevelop and validate a rapid, in-process assay which could result in consider-able time savings in production runs (Figure 2-24). Eminase dissociates into itsconstituents under acidic conditions. Note that lys-plasminogen reveals twopeaks, corresponding to the two glycosylated forms.

2-40

0.060

0.040

0.020

0.000

0 2 4 6 8 9

1

2

3

5

4

Time (min)

Abs

orba

nce

(200

nm

)

Figure 2-24. In-process assay for Eminase. Buffer: 40 mM sodium phosphate,pH 2.5, with 0.01% HPMC added. Peak identification: (1) Eminase buffercomponent; (2) human serum albumin; (3) lys-plasminogen; (4) acylatingagent; (5) streptokinase. From Birell and Brightwell, SmithKline-BeechamPharmaceuticals, UK (1990).

Four in-process samples, together with 9 different batches of freeze-driedeminase, were assayed for lys-plasminogen content relative to an in-housereference preparation of known potency.

The results of the CE method and the existing activity method are shownin Table 2-8. The CE method was also shown to be linear over the range tested(0 to 250% of the expected assay result) whereas precision with CE (2 to 3%RSD) was better than with the existing activity method (3 to 4% RSD).

2-41

Table 2-8. Comparison of CE with an Existing Activity Assay

In-process Activity CE CE: Eminase Activity CE CE:Samples Assay (mg/mL) Activity Assay (mg/vial) Activity

(mg/mL) (mg/vial)

1 9.32 9.54 0.982 1 29.0 29.4 1.012 6.14 6.07 0.989 2 31.7 29.1 0.313 5.38 5.78 0.983 3 31.3 32.3 1.034 6.66 6.58 0.988 4 31.5 30.9 0.98

5 30.2 30.6 1.016 32.2 31.0 0.967 31.5 28.7 0.918 31.5 29.1 0.929 29.7 28.3 0.95

Four in-process samples together with 9 different batches of freeze-dried Eminaseassayed for lys-plasminogen content relative to an in-house reference preparation ofknown potency. From Birell and Brightwell, SmithKline-Beecham Pharmaceuticals,UK.

Ref-1

References

Abersold R., Morrison, H. D. J. Chromatogr. 516, 79 (1990)

Advis, J. P., Hernandez, L., Guzman, N. A. Peptide Research 2, 389(1989)

Albin, M., Grossman, P. D., Moring, S. E. Anal. Chem. 65, 489A (1993)

Arcelloni, C., Fermo, I., Banfi, G., Pontiroli, A. E., Paroni, R. Anal.Biochem. 212, 160 (1993)

Arentoft, A. M., Frokiaer, H., Michaelsen, S, Sorensen, H., Sorensen, S.J. Chromatogr. 652, 189 (1993)

Avallone, H. L. J. Parenteral Science and Technology 40, 178 (1986)

Banke, N., Hansen, K., Diers, I. J. J. Chromatogr. 559, 325 (1991)

Beattie, J. H., Richards, M. P. J. Chromatogr. 664, 129 (1994)

Beattie, J. H., Richards, M. P., Self, R. J. Chromatogr. 632, 127 (1993)

Beckers, J. L., Everaerts, F. M. J. Chromatogr. 508, 19 (1990)

Bergman, T., Agerberth, B., Jornvall, H. FEBS Letters 283, 100 (1991)

Biehler, R., Jacobs, A. Application Information Bulletin A-1727, BeckmanInstruments, Inc. (1993)

Biehler, R., Schwartz, H. E. Technical Information Bulletin TIBC-105,Beckman Instruments, Inc.(1991)

Bongers, J., Lambros, T., Felix, A. M., Heimer, E. P. J. Liq. Chromatogr.15, 1115 (1992)

Bullock, J. J. Chromatogr. 633, 235 (1993)

Bullock, J. A., Yuan, L.-C. J. Microcol. Sep. 3, 241 (1991)

Bushey, M. M., Jorgenson, J. W. J. Chromatogr. 480, 301 (1989)

Camilleri, P. (Ed.). Capillary Electrophoresis: Theory and Practice. BocaRaton: CRC Press, 1993.

Camilleri, P., Okafo, G. N. J. Chem. Soc. Chem. Commun. 3, 196 (1991)

Camilleri, P., Okafo, G. N., Southan, C. Anal. Biochem. 196, 178 (1991)

Chen, F.-T. J. Chromatogr. 559, 445 (1991)

Chen, F.-T. A. Technical Information Data Sheet-DS-818, Beckman In-struments, Inc. (1991)

Chen, F.-T. A., Kelly, L., Palmieri, R., Biehler, R., Schwartz, H. E. J. Liq.Chromatogr. 15, 1143 (1992)

Ref-2

Chen, F.-T. A., Liu, C.-M., Hsieh, Y.-Z., Sternberg, J. C. Clin. Chem. 37,14 (1991)

Chen, F.-T. A., Pentoney, S. L., Jr. J. Chromatgr. (in press)

Chen, F.-T. A., Sternberg, J. C. Electrophoresis 15, 13 (1994)

Chen, S. M., Wiktorowicz, J. E. Anal. Biochem. 206, 84 (1992)

Chien, R. L., Burgi, D. S. Anal. Chem. 64, 489A (1989)

Chu, Y.-H., Whitesides, G. M. J. Med. Chem. 35, 2915 (1992)

Cohen, A. S., Karger, B. L. J. Chromatogr. 397, 409 (1987)

Cohen, A. S., Najarian, D. R., Paulus, A., Guttman, A., Smith, J. A.,Karger, B. L. Proc. Natl. Acad. Sci. USA 85, 9660 (1988)

Cohen, A. S., Terabe, S., Smith, J. A., Karger, B. L. Anal. Chem. 59, 1021(1987)

Compton, B. J. J. Chromatogr. 559, 357 (1991)

Deyl, Z., Rohlicek, V., Struzinsky, R. J. Liq. Chromatogr. 12, 2515 (1989)

Deyl, Z., Struzinski, R. J. Chromatogr. 569, 63 (1991)

Emmer, A., Jansson, M., Roeraade, J. J. Chromatogr. 547, 544 (1991)

Everaerts, F. M., Beckers, J. L., Verheggen, T. P. E. M. Isotachophoresis:Theory, Instrumentation and Applications. Amsterdam: Elsevier, 1976.

Field, M., O’Conner, J., Palmieri, R. Application Data Sheet DS-791,Beckman Instruments, Inc.(1991)

Foret, F., Szoko, E., Karger, B. L. Application Information Bulletin A-1740, Beckman Instruments, Inc. (1993)

Foret, F., Szoko, E., Karger, B. L. J. Chromatogr. 608, 3 (1992)

Frenz, J., Wu, S.-L., Hancock, W. S. J. Chromatogr. 480, 379 (1989)

Gagnon, J., Goeltz, P. Application Data Sheet DS-801, Beckman Instru-ments, Inc. (1991)

Ganzler, K., Greve, K. S., Cohen, A. S., Karger, B. L., Guttman, A.,Cooke, N. C. Anal. Chem. 64, 2665 (1992)

Garnick, R. L., Solli, N. J., Papa, P. A. Anal.Chem. 60, 2546 (1988)

Gordon, M. J., Lee, K.-J., Zare, R. N. Anal. Chem. 63, 69 (1991)

Grossman, P. D., Colburn, J. C., Lauer, H. H. Anal. Biochem. 179, 28(1989)

Grossman, P. D., Colburn, J. C., Lauer, H. H., Nielsen, R. G., Riggin,R. M., Sittampalam, G. S., Rickard, E. C. Anal. Chem. 61, 1186 (1989)

Guttman, A., Nolan, J. A., Cooke, N. J. Chromatogr. 632, 171 (1993)

Ref-3

Guttman, A., Shieh, P., Cooke, N. Technical Information Data Sheet DS-827, Beckman Instruments, Inc. (1992)

Guzman, N. A. (Ed.). Capillary Electrophoresis Technology. New York:Marcel Dekker, 1993.

Guzman, N. A., Hernandez, L., Advis, J. P. Current Research in ProteinChemistry, p. 203. San Diego: Academic Press, 1990.

Guzman, N. A., Moschera, J., Iqbal, K., Malinck, A. W. J. Chromatogr.608, 197 (1992)

Harrington, S. J., Varro, R., Li, T. M. J. Chromatogr. 559, 385 (1991)

Heegaard, N. H. H., Robey, F. A. Anal. Chem. 64, 2479 (1992)

Hjerten, S. Chromatogr. Rev. 9, 122 (1967)

Hjerten, S., Vatcheva, L., Elenbring, K., Eaker, D. J. Liq. Chromatogr. 12,2471 (1989)

Hjerten, S., Zhu, M. J. Chromatogr. 346, 265 (1985)

Honda, S., Taga, A., Suzuki, S., Kahehi, K. J. Chromatogr. 597, 377(1992)

Huch, R., D’Haese, J. Discovery, Spring, Beckman Instruments, Inc.(1993)

Hughes, D. E. , Richberg, P. Application Information Bulletin A-1763,Beckman Instruments, Inc. (1994)

Hughes, D. E. , Richberg, P. J. Chromatogr. 635, 313-318 (1993)

James, D. C., Freedman, R. B., Hoare, M., Jenkins, N. Application Infor-mation Bulletin A-1761, Beckman Instruments, Inc. (1994)

Jorgenson, J. W., Lukacs, K. D. Anal. Chem. 53, 1298 (1981)

Jorgenson, J. W., Lukacs, K. D. Science 222, 266 (1983)

Josic, D., Zeilinger, K., Reutter, W., Bottcher, A., Scmitz, G. J.Chromatogr. 516, 89 (1990)

Kajiwara, H. J. Chromatogr. 559, 345 (1991)

Karger, B. L. Am. Lab. 25, 23 (1993)

Karger, B. L., Cohen, A. S., Guttman, A. J. Chromatogr. 492, 585 (1989)

Kelly, J. F., Locke, S. J., Thilbault, P. Discovery, Fall, Beckman Instru-ments, Inc. (1993)

Kenndler, E., Schmidt-Beiwl, K. J. Chromatogr. 545, 397 (1991)

Kenny, J. W., Ohms, J. I., Smith, A. J. Techniques in Protein ChemistryIV, Angeletti, R. (Ed.). San Diego: Academic Press, 1993.

Kilar, F., Hjerten, S. Electrophoresis 10, 23 (1989)

Ref-4

Kilar, F., Hjerten, S., J. Chromatogr. 480, 351 (1989)

Klein, G. L., Jolliff, C. R. Handbook of Capillary Electrophoresis, p. 419,Landers, J.P. (Ed.). Boca Raton: CRC Press, 1994.

Kuhn, R., Frei, R., Christen, M. Anal. Biochem. 218, 131 (1994)

Landers, J. P. (Ed.). Handbook of Capillary Electrophoresis. Boca Raton:CRC Press, 1994.

Landers, J. P., Oda, R. P., Liebenow, J. A., Spelsberg, T. C.J. Chromatogr. 652, 109 (1993)

Landers, J. P., Oda, R. P., Madden, B. J., Spelsberg, T. C. Anal. Biochem.205, 115 (1992)

Landers, J. P., Oda, R. P., Spelsberg, T. C., Nolan, J. A., Ulfelder, K.J. Biotechniques 14, 98 (1993)

Langenhuizen, M. H. J. M., Janssen, P. S. L. J. Chromatogr. 638, 311(1993)

Lauer, H. H., McManigill, D. Anal. Chem. 58, 166 (1986)

Lausch, R., Scheper, T., Reif, O. W., Schlosser, J., Fleischer, J., Freitag, R.J. Chromatogr. 654, 190 (1993)

Lindner, H., Helliger, W., Dirschlmayer, A., Jacquemar, M., Puschendorf,B. J. Biochem. 283, 467 (1992)

Liu, J., Cobb, K. A., Novotny, M. J. Chromatogr. 468, 55 (1988)

Ludi, H., Gassman, E., Grossenbacher, H., Marki, W. Anal. Chim. Acta213, 215 (1988)

Majors, R. E. LCGC 12, 278 (1994)

Mazzeo, J. R., Martineau, J. A., Krull, I. S. Methods 4, 205 (1992)

McCormick, R. M. Anal. Chem. 60, 2322 (1988)

McCormick, R. M. Handbook of Capillary Electrophoresis, p. 287,Landers, J. P. (Ed.). Boca Raton: CRC Press, 1994.

McLaughlin, G. M., Nolan, J. A., Lindahl, J. L., Palmieri, R. H., Anderson,K. W., Morris, S. C., Morrison, J. A., Bronzert, T. J. J. Liq. Chromatogr.15, 961 (1992)

McLaughlin, G., Palmieri, R., Anderson, K. Techniques in Protein Chem-istry II, pp. 3-19. San Diego: Academic Press, 1991.

Molteni, S., Frishknecht, H., Thormann, W. Electrophoresis 15, 22 (1994)

Moring, S. E. Capillary Electrophoresis: Theory and Practice, p. 87,Grossman, P. D. and Colburn (Eds.). Orlando: Academic Press, 1992.

Mosher, R. A. Electrophoresis 11, 765 (1990)

Ref-5

Nelson, R. J., Paulus, A., Cohen, A. S., Guttman, A., Karger, B. L.J. Chromatogr. 480, 111 (1989)

Nielsen, R. G., Rickard, E. C., Santa, P. F., Sharknas, D. A., Sittampalam,G. S. J. Chromatogr. 539, 177 (1991)

Nielsen, R. G., Riggin, R. M., Rickard, E. C. J. Chromatogr. 480, 393(1989)

Nolan, J., Application Information Bulletin A-1750, Beckman Instruments,Inc. (1993)

Novotny, M., Cobb, K. A., Liu, J. Electrophoresis 11, 735 (1990)

Nyberg, F., Zhu, M. D., Liao, J. L., Hjerten, S. Electrophoresis ’88, p.141, C. Shaefer-Nielsen (Ed.). New York: VCH Publishers, 1988.

Offord, R. E. Nature 211, 591 (1966)

Palmieri, R. Application Data Sheet DS-749, Beckman Instruments, Inc.(1989)

Palmieri, R. P., Nolan, J. A. Handbook of Capillary Electrophoresis,p.325, Landers, J. P. (Ed.). Boca Raton: CRC Press, 1994.

Pande, P. G., Nellore, R. V., Bhagat, H. R. Anal. Biochem. 204, 103(1992)

Pascual, P., Martinez-Lara, E., Barcena, J. A., Lopez-Barea, J., Toribio, F.J. Chromatogr. 581, 49 (1992)

Phillips, T. M. LC GC 11, 36 (1993)

Pleasance, S., Thilbault, P. Capillary Electrophoresis: Theory and Prac-tice, p. 311, Camilleri, P. (Ed.). Boca Raton: CRC Press, 1993.

Pritchett, T. Application Information Bulletin A-1769. Beckman Instru-ments, Inc. (1994)

Pritchett, T. Poster at HPCE ’94, San Diego, CA, 1994.

Pritchett, T., Gouesclou, L., Rietveld, B. Poster at AAPS Western Re-gional Meeting, San Francisco, CA, 1994.

Regnier, F. E. Science 238, 319 (1987)

Rickard, E. C., Strohl, M. M., Nielsen, R. G. Anal. Biochem. 197, 197(1991)

Righetti, P. G. Isoelectric Focusing: Theory, Methodology and Applica-tions. Amsterdam: Elsevier Biomedical Press, 1983.

Rudd, P. M., Scragg, I. G., Coghill, E. C., Dwek, R. A. GlycoconjugateJournal 9, 86 (1992)

Rush, R. S., Cohen, A. S., Mayer, B. L. Anal. Chem. 63, 1346 (1991)

Ref-6

Rush, R. S., Derby, P. L., Strickland, T. W., Rhode, M. F. Anal. Chem. 65,1834 (1993)

Satow, T., Machida, A., Funakushi, K., Palmieri, R. HRC, J. High Resolut.Chromatogr. 14, 276 (1991)

Schultz, N. M., Kennedy, R. T. Anal. Chem. 65, 3161 (1993)

Schwartz, H. E., Palmieri, R. P., Brown, R. Capillary Electrophoresis:Theory and Practice, p.201, Camilleri, P. (Ed.). Boca Raton: CRC Press,1993.

Schwartz, H. E., Ulfelder, K. J., Chen, F.-T. A., Pentoney, S. L., J. Capil-lary Electrophoresis, (in press, 1994)

Schwer, C., Lottspeich, F. J. Chromatogr. 623, 345 (1992)

Shimura, K., Karger, B. L. Anal. Chem. 66, 9 (1994)

Skoog, B., Wichman, A. Trends Anal. Chem. 5, 82 (1986)

Smith, A., Ohms, J. I. Techniques in Protein Chemistry III, R. Hoque-Angeletti (Ed.). San Diego: Academic Press, 1992.

Smith, R. D., Goodlett, D. R., Wahl, J. H. Handbook of Capillary Electro-phoresis, p.185, Landers, J. P. (Ed.). Boca Raton: CRC Press, 1994.

Stover, F. S., Haymore, B. L., McBeath, R. J. J. Chromatogr. 470, 241(1989)

Strege, M. A., Lagu, A. L. Anal. Biochem. 210, 402 (1993)

Strege, M. A., Lagu, A. L. J. Chromatogr. 652, 179 (1993)

Strickland, M. Am. Lab., March (1991)

Swerdberg, S. A. Handbook of Capillary Electrophoresis, p.513, Landers,J. P. (Ed.). Boca Raton: CRC Press, 1994.

Tadey, T., Purdy, W. C. J. Chromatogr. 652, 131 (1993)

Takeo, K. Advances in Electrophoresis, p. 229. New York: VCH Publish-ers, 1987.

Terabe, S. Trends Anal. Chem. 8, 129 (1989)

Thompson, T. J., Foret, F., Karger, B. L. Anal. Chem. 65, 900 (1993)

Thormann, W., Caslavska, J., Molteni, S., Chmelík, J. J. Chromatogr.589, 321 (1992)

Tomlinson, A. J., Benson, L. M., Scanlan, G. F., Nevala, W. K.,Strausbach M., Wettstein, P. J., Naylor, S. Poster at HPCE ’94, San Diego,CA, 1994.

Towns, J. K., Regnier, F. E. Anal Chem. 63, 1126 (1991)

Tran, A. D., Park, S., Lisi, P. J., Huynh, O. T., Ryall, R. R., Lane, P. A.J. Chromatogr. 542, 459 (1991)

Ref-7

Tsuji, K. J. Chromatogr. 550, 823 (1991)

Tsuji, K. J. Chromatogr. 652, 139 (1993)

Tsuji, K., Baczynskyj, L, Bronson, G. E. Anal. Chem. 64, 1864 (1992)

Van de Goor, T. A. A. M., Janssen, P. S. L., Van Nispen, J. W., VanZeeland, M. J. M., Everaerts, F. M. J. Chromatogr. 545, 379 (1991)

Vinther, A., Petersen, J., Soeberg, H. J. Chromatogr. 608, 205 (1992)

Wanders, B. J., Everaerts, F. M. Handbook of Capillary Electrophoresis,p. 111, Landers, J. P. (Ed.). Boca Raton: CRC Press,1994.

Watson, E., Yao, F. Anal. Biochem. 210, 389 (1993)

Watson, E., Yao, F., Barber, W. Poster at HPCE ’94, San Diego, CA,1994.

Wehr, T. LCGC 11, 14 (1993)

Wehr, T., Zhu, M., Rodriguez, R., Burke, D., Duncan, K. Am. Biotech.Lab. 8, 22 (1990)

Werner, W. E., Demorest, D. M., Wiktorowicz, J. E. Electrophoresis 14,759 (1993)

Widhalm, A., Schwer, C., Blass, D., Kenndler, E. J. Chromatogr. 549, 446(1991)

Wiktorowicz, J. E., Colburn, J. C. Electrophoresis 11, 769 (1990)

Yao, X.-W., Regnier, F. E. J. Chromatogr. 632, 185 (1993)

Yim, K. W. J. Chromatogr. 559, 401 (1991)

Yowell, G. G., Fazio, S. D., Vivilecchia, R. V. Application InformationBulletin A-1744, Beckman Instruments, Inc. (1993)

Yowell, G. G., Fazio, S. D., Vivilecchia, R. V. J. Chromatogr., 652, 215(1993)

Zhao, Z., Malik, A., Lee, M. L., Watt, G. D. Anal. Biochem. 218, 47(1994)

Zhu, M., Rodriguez, R., Wehr, T., Siebert, C. J. Chromatogr. 608, 225(1992)

Zhu, M., Wehr, T., Levi, V., Rodriguez, R., Shiffer, K., Cao, Z. A.,J. Chromatogr. 652, 119 (1993)

BE

CK

MA

NS

eparation of DN

A by C

apillary Electrophoresis

Volum

e VII

BECKMAN

Separation of DNAby Capillary Electrophoresis

Beckman Instruments, Inc. • 2500 Harbor Boulevard, Box 3100 • Fullerton, California 92634-3100

Sales: 1-800-742-2345 • Service: 1-800-551-1150 • TWX: 910-592-1260 • Telex: 678413 • Fax: 1-800-643-4366

Worldwide Offices: Africa, Middle East, Eastern Europe (Switzerland) (22) 994 07 07. Australia (61) 02 816-5288. Austria (2243) 7292164.

Canada (800) 387-6799. China (861) 5051241-2. France (33) 1 43 01 70 00. Germany (49) 89-38871. Hong Kong (852) 814 7431.

Italy (39) 2-953921. Japan 3-3221-5831. Mexico 525 575 5200, 525 575 3511. Netherlands 02979-85651. Poland 408822, 408833.

Singapore (65) 339 3633. South Africa (27) 11-805-2014/5. Spain (1) 358-0051. Sweden (8) 98-5320. Switzerland (22) 994 07 07.

Taiwan (886) 02 378-3456. U.K. (01494) 441181. U.S.A. 1-800-742-2345.

607397 $14.95 Printed in U.S.A. © 1995 Beckman Instruments, Inc.

BECKMAN

Separation of DNAby Capillary Electrophoresis

Herb Schwartz1

and

Andras Guttman2

1 Palomar Analytical Services, 150 Montalvo Road, Redwood City, CA 94062tel: (415) 365-3711; e-mail: [email protected]

2 Beckman Instruments, Inc., 2500 Harbor Blvd., Fullerton, CA 92634tel: (714) 773-8211; e-mail: [email protected]

iii

Table of Contents

About the Authors .........................................................................................v

Acknowledgments .........................................................................................v

Front Cover ................................................................................................. vi

Acronyms Used .......................................................................................... vii

1 Introduction ............................................................................................1

2 Fundamentals of Capillary Electrophoresis (CE) ..................................42.1 CE Instrumentation.......................................................................42.2 CE Modes .....................................................................................62.3 Theory ...........................................................................................9

2.3.1 Migration Velocity ............................................................92.3.2 Efficiency and Resolution .............................................. 11

3 Separation of DNA by Techniques Other Than CE ........................... 133.1 HPLC ......................................................................................... 133.2 Gel Electrophoresis: Polyacrylamide (PA) and Agarose .......... 15

3.2.1 Instrumentation and Detection ...................................... 153.2.2 Separation Mechanism .................................................. 173.2.3 Polyacrylamide Gel Electrophoresis (PAGE) ............... 193.2.4 Agarose Gel Electrophoresis ......................................... 193.2.5 Pulsed-Field Gel Electrophoresis (PFGE) .................... 19

4 CE Methods: Principles and Strategies .............................................. 214.1 Free Solution Methods .............................................................. 21

4.1.1 CZE with Untreated or Coated Capillaries ................... 214.1.2 CZE with Sample Stacking or ITP Preconcentration ... 234.1.3 Micellar Electrokinetic Capillary Chromatography

(MECC) ......................................................................... 244.2 Capillary Gel Electrophoresis (CGE) Methods ........................ 24

4.2.1 Polyacrylamide .............................................................. 304.2.2 Agarose .......................................................................... 334.2.3 Alkylcellulose and Other Polymers .............................. 344.2.4 Intercalators as Buffer Additives ................................... 374.2.5 Ferguson Plots ............................................................... 394.2.6 Instrument Parameters ................................................... 404.2.7 Sample Injection and Matrix Effects; Quantitation ...... 40

4.2.7.1 Replaceable Gels ............................................. 404.2.7.2 Non-Replaceable Gels .................................... 44

4.2.8 Hybridization; Southern Blotting,Mobility Shift Assays .................................................... 46

iv

4.3 Detection in CE: UV Absorbance vs. Laser-InducedFluorescence (LIF) .................................................................... 494.3.1 DNA Detection with LIF............................................... 52

4.3.1.1 Native and Indirect Fluorescence ................... 524.3.1.2 Intercalators ..................................................... 524.3.1.3 Fluorescent Labeling ....................................... 55

4.4 Fraction Collection: CE as a Micropreparative Tool ................ 574.4.1 Fraction Collection Using Field Programming ............. 57

5 Selected Applications .......................................................................... 595.1 Nucleotides, Nucleosides and Bases with CZE or MECC ....... 59

5.1.1 DNA Adducts; DNA Damage ....................................... 595.1.2 Nucleoside Analog Drugs.............................................. 595.1.3 Nucleotides in Cell Extracts .......................................... 615.1.4 Increasing Detectability: LIF Detection ........................ 63

5.2 Purity Control of Synthetic Oligonucleotides ........................... 635.2.1 Phosphodiester Oligonucleotides .................................. 635.2.2 Antisense DNA.............................................................. 64

5.3 DNA Sequencing ....................................................................... 665.4 dsDNA, PCR Products Analysis (< 2000 bp) ........................... 69

5.4.1 Quantitation of Viral Load in Infectious Diseases ........ 705.4.2 Competitive RNA-PCR by CE-LIF for

Quantitation of Cellular mRNA .................................... 735.4.3 Detection of DNA Polymorphisms and

Mutations in Genetic Diseases ...................................... 755.4.3.1 Point Mutations ............................................... 78

5.4.4 DNA Profiling in Forensic Work .................................. 805.4.5 DNA Profiling of Plants, Bacteria and Fungi ............... 825.4.6 Plasmid Mapping ........................................................... 84

5.5 dsDNA (2 to 20 kbp) by CGE ................................................... 855.5.1 Quantitation of Plasmid Copy Number ......................... 86

5.6 Very Large Chromosomal DNA (> 20 kbp) .............................. 86

6 References ........................................................................................... 87

v

About the Authors

Herbert E. Schwartz is a native of the Netherlands and received a Ph.D. inAnalytical Chemistry from Northeastern University, Boston (under the direc-tion of Prof. B. L. Karger), and a M.S. degree from the Free University, Am-sterdam. He has been working in the analytical instrumentation industry forover 10 years, has authored approximately 50 publications in the field of sepa-ration science, and taught training courses in capillary electrophoresis. Beforestarting his consulting business, Palomar Analytical Services in Redwood City,California (tel: 415-365-3711; e-mail: [email protected]) , Dr. Schwartzmanaged an applications group at Beckman Instruments, Inc., Palo Alto, Cali-fornia, and was involved in the development and marketing of the first fullyautomated, commercial capillary electrophoresis instrument at MicrophoreticSystems. Prior to that, he was employed as a research chemist at AppliedBiosystems and Brownlee Labs. He edited the previous primers (VolumesI–V) on capillary electrophoresis for Beckman.

Andras Guttman is a native of Hungary and a Principal Research Chemistwith Beckman Instruments, Inc., Fullerton, California. His research currentlyfocuses on the development of new chemistry kits for capillary electrophoresis,and size separations of protein and DNA molecules with gel-filled capillarycolumns. He received a Ph.D. in Analytical Biochemistry from the HungarianAcademy of Sciences, a M.S. degree from the University of Chemical Engi-neering at Veszprem, Hungary, and did post-doctoral work in Prof. B. L.Karger’s laboratory at the Barnett Institute, Northeastern University, Boston.Andras has authored 51 publications and holds 4 patents, mainly in the field ofcapillary electrophoresis. He joined the R&D group of the BiotechnologyDevelopment Center of Beckman Instruments, Inc., in 1990. During 1992–1993, he spent half a year in the Analytical Research Laboratory of HafslundNycomed Pharma in Linz, Austria, developing chiral separations by capillaryelectrophoresis.

Acknowledgments

We would like to thank Kathi J. Ulfelder at Beckman Instruments, Inc., Fuller-ton, California, for many suggestions and contributions to this book. The sup-port of Jim McCoy and Dr. Nelson Cooke at Beckman is also highlyappreciated. We thank Judy Strauss, LAHS, Los Altos, California, for review-ing drafts of the manuscript and Gale Leach and Annette Hurst at WordsWorth(Pacifica, California) for the desktop publishing of the manuscript.

vi

All trademarks and registered trademarks are the property of theirrespective owners.

Other Beckman primers (Volumes I, II, III, IV, V, and VI) oncapillary electrophoresis:


I. Introduction to Capillary Electrophoresis 360643

II. Introduction to Capillary Electrophoresisof Proteins and Peptides 266923

III. Micellar Electrokinetic Chromatography 266924

IV. Introduction to the Theory and Applicationsof Chiral Capillary Electrophoresis 726388

V. Separation of Proteins and Peptides byCapillary Electrophoresis: Applicationto Analytical Biotechnology 727484

VI. Introduction to Quantitative Applicationsof Capillary Electrophoresis inPharmaceutical Analysis 538703

Front Cover

The front cover shows EcoRI endonuclease, bound to a single-strand piece ofDNA containing its recognition site sequence. The protein (1rle-pdb) is shownas a blue ribbon highlighting the secondary-structure regions; the space-fillingrepresentation of the DNA backbone is in yellow, and the nucleic bases in red.Those side-chains that interact with the nucleic acid are shown in thick-bondrepresentation, in their element colors, i.e., carbon: green, oxygen: red, andnitrogen: blue. Courtesy of Dr. Don Gregory, Molecular Simulations,Burlington, Massachusetts.

vii

Acronyms Used

The following acronyms are used in this book.

ac alternating currentACE affinity capillary electrophoresisARMS amplification refractory mutation systemASA allele-specific amplificationbp base pairCE capillary electrophoresisCGE capillary gel electrophoresisCMC critical micelle concentrationCTAB cetyltriethylammonium bromideCZE capillary zone electrophoresisdc direct currentDGGE denaturing gradient gel electrophoresisdsDNA double-stranded DNAEB ethidium bromideEDTA ethylenediamine tetraacetic acidEOF electroosmotic flowFITC fluoroscein isothiocyanateGAPDH glyceraldehyde-3-phosphate dehydrogenaseGE gel electrophoresisHPA heteroduplex polymorphism analysisHPCE high-performance capillary electrophoresisHPLC high-performance liquid chromatographyHPMC hydroxypropylmethylcelluloseIEF isoelectric focusingITP isotachophoresiskbp kilobase pairkV kilovoltsLC liquid chromatographyLIF laser-induced fluorescenceMCAD medium-chain coenzyme A dehydrogenaseMECC micellar electrokinetic capillary chromatographyPA polyacrylamidePAGE polyacrylamide gel electrophoresis

viii

PCR polymerase chain reactionPFGE pulsed-field gel electrophoresisPNA peptide nucleic acidRFLP restriction fragment length polymorphismRSD relative standard deviationRT reverse transcription / reverse transcribedSDS sodium dodecyl sulfatessDNA single-stranded DNASSCP single-stranded conformation polymorphismSTR short tandem repeatTTAB tetradecyltrimethylammonium bromideUV ultravioletVNTR variable number tandem repeat

PCR is covered by U.S. patents owned by Hoffmann-La Roche, Inc.

1

1 IntroductionElectrophoresis is the transport process of charged analytes under theinfluence of an electric field. Nucleic acids are generally electrophoresedin neutral or basic buffers—as anions with their negatively chargedphosphate groups (at acidic pH, polynucleotides are insoluble in water).Historically, electrophoresis experiments in glass tubes were reported asearly as the beginning of the 19th century. Before the advent of rod andslab gels, early 20th century electrophoresis was performed in freesolution. In the mid 1930s, the Swedish chemist Arne Tiselius applied afree solution technique—moving boundary electrophoresis—to serumprotein analysis, work for which he would later receive the Nobel Prize(Tiselius, 1937). Studies with “Tiselius”-type apparatus equipped withSchlieren optics to visualize nucleic acid separations continued during the1940s and 1950s (see, for example, Chargaff and Saidel, 1949). Ever sincethe publication of DNA’s double helical structure (Watson and Crick,1953), electrophoresis has been a standard, indispensable analytical tool inmodern biochemistry and molecular biology; electrophoretic proceduresare used in almost every aspect of basic or applied biomedical and clinicalresearch.

Since the 1960s, methods employing supporting media, such as poly-acrylamide or agarose gels, have become the norm for protein and nucleicacid analysis. The gel matrix acts as an anticonvective medium, reducingconvective transport and diffusion so that separated sample componentsremain positioned in sharp zones during the run. In addition, the gel acts asa molecular sieve, separating nucleic acids according to their size. Electro-phoresis, as performed today in the majority of laboratories, is still typi-cally a manual process: the gels are poured, the separation is run, and thebands are visualized by means of a staining/destaining process; a photo-graph of the gel is then kept for records. This makes electrophoretic proce-dures often time consuming and labor intensive. With the exception ofDNA sequencing, electrophoretic procedures typically have not been fullyautomated, unlike chromatographic separation methods such as HPLC.In addition, most electrophoretic techniques are qualitative and accuratequantitation is often problematic.

Capillary electrophoresis (CE) is a relatively new, powerful separationtechnique that is ideally suited for handling small amounts of sample mate-rial. This demand becomes increasingly prominent in bioanalytical re-search, e.g., in biotechnology and in various clinical, diagnostic, genetic,

2

and forensic applications. In its present instrumental configuration, CE ismore than a decade old (Jorgenson and Lukacs, 1981, 1983). Precedingmodern CE, it is interesting to note that, as early as 1953, some experi-menters recognized the importance of miniaturized equipment when deal-ing with certain biological samples. For example, ribonucleotides could beidentified in single cells using “micro-electrophoresis” in fine silk fibers(Edstrom, 1953). Classical electrophoretic techniques were not suitablebecause the sample was too small in size.

CE offers several similarities to high performance liquid chromatogra-phy (HPLC), i.e., ease of use, high resolution, speed, on-line detection, andfull automation capability. CE, having taken essential components fromboth HPLC and electrophoresis, can be viewed as an instrumental approachto electrophoresis. In its relatively short history, CE has found particularapplicability in bioanalysis. Analogous to HPLC, CE is also often referredto as high performance capillary electrophoresis (HPCE). A number ofrecent text books cover the history, theory, instrumentation, and applica-tions of CE (Camillieri, 1993; Guzman, 1993; Landers, 1994). A new jour-nal—J. Cap. Elec.—(Guzman, 1994) has been recently introduced to covernovel applications in the field. Other Beckman primers on CE are listed atthe start of this book.

While the first papers on DNA analysis by CE only appeared in 1988(Kasper et al., 1988; Cohen et al., 1988), their number since then hasgrown exponentially. Partly, this has been fueled by the emergence of newtools in molecular biology, in particular PCR, and the Human GenomeProject, with its spin-offs such as gene hunting technology. It is rapidlybecoming clear that many, if not all, slab gel electrophoresis techniquescan be transferred to a capillary format. Benefits of this transformationwould include fast, high-resolution analyses, full automation and datastorage capability; in addition, only minute amounts of sample are requiredwith CE techniques. With regard to the latter, the recent introduction of ahighly sensitive, laser-induced fluorescence (LIF) detector has opened upnew perspectives for low-level DNA analysis, e.g., detectability into thezeptomole (10-21 mol) range.

This book describes the main principles of the various CE methodsand their applications to nucleic acid analysis, specifically those dealingwith nucleosides, nucleotides, ss oligonucleotides, and dsDNA (PCR)fragments. Whereas the emphasis is on DNA, some RNA separations arealso discussed, as well as applications where RNA is reverse-transcribed

3

into DNA. In addition, a brief overview of other high-resolution, comple-mentary techniques (HPLC, slab gel electrophoresis) will be given. Themain focus is on applications which can be performed with today’s com-mercially available instrumentation (as opposed to those with special,homemade designs). Instrumentation will not be discussed extensively;additional information on CE can be found in the above-mentioned text-books.

4

2 Fundamentals of CapillaryElectrophoresis (CE)

2.1 CE InstrumentationFigure 1 shows a schematic diagram of a CE system such as that used inthe P/ACE™ Capillary Electrophoresis Systems from Beckman. Fusedsilica capillaries, generally 20–100 µm i.d. and 20–100 cm long, are usedas the separation channel. The inner surface of the capillaries can beuntreated or coated, depending on the application, and are encased in atemperature-controlled (± 0.1°C) cartridge. The capillary inlets and outletsare positioned in the sample and/or buffer vials. In P/ACE, sampleintroduction (often called “injection” to maintain the analogy with HPLC)is either by pressure or by electrokinetic means; the samples are loaded inan autosampler tray. In the pressure method, a sample vial is temporarilypressurized to allow the flow of sample into the capillary; theelectrokinetic method utilizes the electric field to draw charged analytesinto the capillary.

Applied Voltage

Sample Flow

Capillary

CapillaryCooling System

(+) (–)

Buffer/Sample Buffer

Detector

Figure 1. Diagram of a CE instrument with normal polarity (anode atinjection side).

5

In contrast to slab gel electrophoresis, CE can take advantage of twotypes of driving forces: 1) the force causing the electrophoretic migration;and 2) the force exerted by the electroosmotic flow (EOF) through thecapillary. The EOF bulk flow results from the charged inner capillary wallduring application of an electric field (Figure 2). With untreated, fused-silica tubing and an aqueous buffer, a negatively charged capillary surfaceis obtained. The magnitude of the EOF is dependent on various experimen-tal factors, most notably the pH of the buffer solution. Capillary coatingscan reverse, reduce, or even totally eliminate the EOF. The latter is thecase in gels (anticonvective medium) when DNA restriction fragments,synthetic oligonucleotides, or DNA sequencing fragments are separated bycapillary gel electrophoresis (CGE).

Electroosmosis

Electrophoresis

Bulk Flow DetectorWindow

Figure 2. Schematic of electrophoresis and electroosmosis in a separationof anionic, neutral, and cationic analytes.

Most modern CE instruments allow voltages of up to 30 kV during theruns. CE voltages are generally much higher than those used in slab gelelectrophoresis. This accounts, in part, for the high resolution and fastanalysis times of CE separations. Joule heat resulting in the capillaries iseffectively dissipated through the capillary wall by the surrounding coolingliquid (as in P/ACE). Detection is achieved by monitoring UV absorbancedirectly on-capillary through a window in the capillary. Other commer-cially available detection options include laser-induced fluorescence (LIF),UV-Vis scanning diode array, and mass spectrometry.

6

The capillary cartridge format used in P/ACE allows for controllingthe temperature (within 0.1°C) of the capillary column, which is very im-portant for obtaining good run-to-run reproducibility (Nelson et al., 1989).Usually, CE separations are performed at 20 to 30°C, i.e., room tempera-ture, although P/ACE permits temperature settings between ≈ 15 to 50°Cfor special applications. Other features of the P/ACE system are the possi-bility of fraction collection and of field programming. While it is primarilyused for analytical separations, CE can also be used for micropreparativeDNA work (see Section 4.4). Similarly to HPLC, on-line system controland computerized data processing are incorporated.

2.2 CE ModesDifferent separation modes, schematically shown in Figure 3, have beendeveloped for CE. For each of these CE modes, Table 1 shows the mainprinciple of operation and the applicability for DNA analysis. Isoelectricfocusing (IEF) will not be discussed as this technique is almost exclusivelyused for peptides and proteins. CZE (capillary zone electrophoresis),MECC (micellar electrokinetic capillary chromatography), and ITP(isotachophoresis) are free-solution techniques. With CZE, only chargedspecies can be separated. MECC is particularly useful for the separation ofrelatively small, neutral molecules such as most pharmaceuticals. Bases,nucleosides, nucleotides, and small oligonucleotides (< 10 bases) also havebeen separated by MECC, as will be discussed in Section 5.1. Thetechnique makes use of surfactants (e.g., sodium dodecyl sulfate [SDS] orcetyltrimethylammonium bromide [CTAB]) which, when added to the runbuffer in sufficiently high enough concentrations, form micelles insolution. During the high-voltage run, the micelles migrate in the oppositedirection to the EOF which drives the separation (Figure 4). Analytespartition into the micelles differently according to their hydrophobicity,thus yielding different elution times. For small DNA fragments, e.g.,nucleosides, nucleotides, and small oligonucleotides, free-solutiontechniques (CZE, MECC) can be applied—generally in conjunction withuncoated capillaries.

7

��

��

��

��

��

��

��

��

��

��

��

BGEt = 0

t = x

t = 0

t = x

t = 0

t = x

BGE

BGE BGE BGEA

A&B

TE LE

TE LEA

A

B

B

B

pH gradient

A&B

A&B

CZEMECC

CGE

CITP

CIEF

Figure 3. Modes of CE. BGE = background electrolyte; LE = leadingelectrolyte; TE = terminating electrolyte; A, B = sample components.t = 0, start of separation (injection); t = x, separation after time x.


= Solute


= Electrophoresis

Figure 4. Schematic of the separations principle of MECC. The detectorwindow is assumed to be positioned near the negative electrode.

8

Table 1. Different CE Separation Modes

Mode Separation ApplicationsMechanism

Free Solution

Capillary Zone Charge-to-Mass Bases, Nucleosides, Nucleo-Electrophoresis (CZE) Ratio tides, Small Oligonucleo-

tides, DNA Damage

Micellar Electrokinetic Charge-to-Mass Same as CZECapillary Chroma- Ratio, Partitioningtography (MECC) into Micelles

Capillary Isotacho- Moving Boundary/ Preconcentration Techniquephoresis (CITP) Displacement for CZE, MECC

Capillary Isoelectric Isoelectric Point ProteinsFocusing (CIEF)

Gel, Polymer Network

Capillary Gel Electro- Molecular Oligonucleotides, Primers,phoresis (CGE) Sieving, Reptation Probes, Antisense DNA,

PCR Products, LargedsDNA, Point Mutations,DNA Sequencing

Isotachophoresis (ITP) or displacement electrophoresis was developedin capillaries in the 1960s and 1970s (Everaerts, 1976). The sample isplaced between carefully selected leading and terminating electrolytes(Figure 3); during migration, the sample concentrations are adapted to thatof the leading zone. ITP can be viewed as a front runner to modern CE—many of its principles apply to modern CE techniques. Except for its utilityas a preconcentration technique with CZE (e.g., in nucleotide separa-tions—see Sections 4.1.2 and 5.1), ITP will not be discussed in this book.For larger DNA (e.g., antisense DNA, sequencing fragments, restrictionfragments), sieving techniques with gels (“polymer networks”) are re-quired. These CGE applications, requiring special pretreated (coated) cap-illaries, are discussed later in Sections 4.2 and 5.

9

2.3 Theory

2.3.1 Migration Velocity

When a uniform electric field (E) is applied to a polyion with a net chargeof Q, the electrical force (Fe) is given by

Fe = QE (1)

In free solution, as well as in viscous media such as a crosslinked gelor a non-crosslinked polymer solution, when the ion is set in motion by theapplied electric field, a frictional force (Ff) acts in the opposite direction:

Ff = f(dx/dt) (2)

where f is the translational friction coefficient and dx and dt are thedistance and time increments, respectively.

Differences in sample properties such as shape, size, or net chargeresult in different electrophoretic mobilities and provide the basis of elec-trophoretic separation. The migration of the charged solute under the elec-tric field is expressed according to Newton’s second law as:

m(d2x/dt2) = EQ - f(dx/dt) (3)

i.e., the product of the mass (m) and acceleration (d2x/dt2) is equal to thedifference of the electrical and frictional forces. When the force from theapplied electric field on the charged solute is counterbalanced by thefrictional force, the solute migrates with a steady state velocity (v):

v = dx/dt = EQ/f (4)

The translational friction coefficient is influenced by the temperature-dependent (T) viscosity of the separation matrix, especially in free solutionand in non-crosslinked polymer networks:

f = C1 e(-1/T) (5)

where C1 is constant for a given shape solute (Guttman et al., 1994). Incapillary gel electrophoresis, the size-dependent retardation of the solute isa primary function of the separation polymer concentration (P%):

v = v0 e(-KRP%) (6)

where v0 is the free solution velocity of the solute and KR is the retardationcoefficient (Chrambach, 1985; Andrews, 1986).

10

As a first approximation, the total electrostatic force (Fe) on the DNAmolecule can be assumed to be constant per unit length, since each residuehas the same net negative charge. Thus, Q is proportional to the number ofbases (n) in the molecule (Cantor et al., 1988) and, therefore, one can con-sider using the number of bases in the DNA molecule (n) instead of the netcharge:

Fe = EQ ≈ Enk (7)

where k, the exponent of n, represents information about the apparentshape of the DNA molecule under the electric field used, i.e., random coil,oriented, stretched, etc. (Lerman and Frish, 1982, Slater and Noolandi,1988, Stellwagen, 1989).

Combining Equations 4–7, and incorporating all the constants togetherin one (Const), the electrophoretic velocity of a migrating DNA moleculecan be described as:

v = Const Enk e(-1/T) e(-KRP%) (8)

where Enk e(-1/T) corresponds to the free solution velocity (v0) at zeropolymer concentration and e(-KRP%) represents the molecular sieving effect.

If there is any special additive in the gel, such as a complexing agent,the solute will have a distribution between the complex and the electrolyte:

v = Const Enk e(-1/T) e(-KRP%) Rp (1+ KL)-1

(9)

where KL is the complex formation constant of the L ion and Rp is themolar ratio of the free solute (Guttman and Cooke, 1991 A). Depending onthe charge and size of the complexing agent, the complex may migratefaster or slower than the free solute. When the complexing agent has acharge opposite to that of the solute, then an uncharged complex may beformed—therefore, the higher the concentration of the complexing agent,the slower will be the migration velocity of the solute.

In capillary gel electrophoresis, molecular sieving can be described bythe Ogston theory (Ogston, 1958): the average pore size of the matrix is inthe same range as that of the hydrodynamic radius of the migrating solute.In this case the logarithm of the velocity of the migrating solute is propor-tional to the size of the solute:

v ≈ E e(-n) (10)

This theory also assumes that the migrating particles behave asunperturbed spherical objects.

11

On the other hand, it is well documented that large biopolymer mol-ecules with flexible chains, e.g., DNA, can still migrate through a polymernetwork which has a pore size that is significantly smaller than the size ofthe solute. This phenomenon can be explained by the reptation model. Thereptation model describes the migration of the polyion, suggesting a “head-first” motion through the pores of the polymer network (Lerman and Frish,1982; DeGennes, 1979; Lumpkin et al., 1985; Viovy and Duke, 1993).According to the reptation model, the velocity of the migrating solute isinversely proportional to the size of the solute, i.e., the chain length:

v ≈ E/n (11)

When higher electric field strengths are used, “biased” reptation oc-curs, in which case the velocity of the solute can be described as:

v ≈ (E/n + bE2) (12)

where b is a function of the mesh size of the polymer network as well asthe charge and segment length (Grossman et al., 1992) of the migratingDNA molecule.

2.3.2 Efficiency and Resolution

According to theory, the major contributor to band broadening, besides theinjection and detection extra-column broadening effects, is the longitudinaldiffusion of the solute in the capillary tube (Jorgenson and Lukacs, 1983;Terabe et al., 1989). The theoretical plate number (N) achieved is charac-teristic of the column efficiency:

N = µe El/2D (13)

where µe is the electrophoretic mobility, D is the diffusion coefficient of thesolute in the separation buffer system, and l is the effective column length.

Resolution (Rs) between two peaks can be calculated from the differ-ences of their electrophoretic mobility (∆µe) (Karger et al., 1989):

Rs = 0.18 ∆µe (El/Dµem)1/2 (14)

where µem is the mean mobility of those species.

Equations 13 and 14 demonstrate that higher applied electric field andlower solute diffusion coefficient would result in higher efficiency (N) andhigher resolution (Rs). The limiting factor in efficiency is mainly the heat

12

generated (Qj, Joule heating) from the power applied (P = V × I) to thenarrow bore capillary tubing (Nelson et al., 1989):

Qj = P/r2IL (15)

where I is the current, L is the total column length and r is the columnradius.

Because of the sensitivity of the electrophoretic mobility to tempera-ture (≈ 2% per °C), as well as the temperature dependence of the complexformation constant, good temperature control is extremely important forachieving good migration reproducibility (Nelson et al., 1989; Karger etal., 1989).

13

3 Separation of DNA by TechniquesOther Than CE

Separation methods of biomolecules can be divided into three major cat-egories: 1) methods which take advantage of size differences of the ana-lytes (e.g., gel filtration, ultrafiltration, centrifugation); 2) methods basedon electrical charge differences (e.g., ion exchange chromatography); and3) methods based on differences in the specific biochemical properties ofanalytes (e.g., hybridization, immunoadsorption techniques).

In electrophoresis, chromatography, and centrifugation the separationmechanism is based on differences in transport velocities of the analytes.Theoretically, therefore, many analogies exist between these separationtechniques. The driving force to achieve the analyte transport is an electricfield in electrophoresis, a hydrodynamic force (flow) in chromatography,and a centrifugal force in centrifugation. Frictional resistance determinesthe extent of retardation during the transport in electrophoresis or centrifu-gation, while adsorption plays a major role in chromatography. Low-reso-lution techniques, e.g., centrifugation, are primarily used for purification ofDNA. Membrane technologies are frequently used to remove unwantedcomponents from the sample (e.g., desalting by ultrafiltration) or to trans-fer DNA bands from gels for further manipulation (e.g., in blotting). Acomprehensive discussion of all separation techniques for DNA is beyondthe scope of this book. In the following sections, the high-resolution sepa-ration techniques (i.e., HPLC and slab gel electrophoresis) are briefly re-viewed, and some comparisons and analogies with CE are made.

3.1 HPLCIn liquid chromatography, the separation mechanism involves partitioningof analytes between a mobile and stationary phase. The selectivity of theseparation is governed by the partitioning of the analytes in the stationaryphase. Diffusion characteristics of the analyte play an important role: gen-erally the higher the molecular mass of the analyte, the lower the separa-tion efficiency (as expressed by the number of theoretical plates). It can betheoretically shown that CE—and electrophoresis in general—should havedistinct advantages over HPLC (Jorgenson and Lukacs, 1983) when largebiomolecules (proteins, DNA, polysaccharides) are separated. This ismainly due to the relatively slow diffusion characteristics of large mol-ecules which work in CE’s favor. The mass transport of large biomolecules

14

in the HPLC mobile phase is slow (relative to the speed of mass transfer),resulting in increased band broadening. In gel electrophoresis, the prob-lems related to mass transfer do not exist. Separation efficiencies inCGE—when expressed as numbers of theoretical plates, N—can be ex-tremely high (e.g., N ≈ 30 million per meter for oligonucleotides, Guttmanet al., 1990). With very large DNA fragments (> 10,000 bp), resolution andefficiency decrease, as predicted (Bae and Soane, 1993).

It is important to note, however, that selectivity often is more impor-tant than “raw” separation power. In electrophoresis, therefore, it is impor-tant to find ways to optimize the difference in electrophoretic mobilities(e.g., with MECC), control the EOF, use certain capillary coatings, etc.Generally, in HPLC, the elution of large molecules can only be achievedby changing the selectivity during the separation process, i.e., by means ofgradient elution. Instrumentation in HPLC is rather complex compared tothe simplicity of a CE system. The separation channel in CE is an open orgel-filled tube; in HPLC, expensive prepacked columns are required whichhave a limited lifetime. Nevertheless, impressive separations of DNA arepossible with HPLC. Size-dependent chromatographic separations ofnucleic acids have recently been reviewed by Kasai (1993). At present, themost popular types of HPLC for DNA separations are anion-exchangechromatography with 2.5 µm, nonporous, polymeric particles (for DNA< 25 kbp), hydrophobic interaction chromatography (for DNA < 3 kbp),and gel permeation chromatography (for DNA < 6 kbp). A recently devel-oped new mode, “slalom chromatography,” appears promising for DNA inthe 5 to 50 kbp range. Other papers by Baba and co-workers (1991),Oefner and Bonn (1994), and Katz (1993) also discuss the state of the artin HPLC of DNA, while comparing HPLC with slab gel electrophoresis orCE.

As mentioned in the Introduction (Section 1), HPLC procedures areoften complementary to those developed by electrophoresis because theseparation mechanisms are different. Both methodologies can be fullyautomated. With the Gold™ Software package from Beckman, HPLC andCE autosamplers can be operated simultaneously from one computer. Ingeneral, HPLC procedures are easier to scale up to (semi)preparative work,although CE also has been used for purifying nanogram to microgramamounts of DNA material (Section 4.4). HPLC of PCR products has beenpresented as being more quantitative than slab gel electrophoresis-densito-metry (Zeillinger et al., 1993). It also has been claimed that HPLC is amore quantitative technique than CE. However, in our opinion, this hasmore to do with CE maturing as an analytical technique: when operational

15

parameters are fully understood and controlled, precision, accuracy, andquantitation are as good as those obtained with HPLC (see also Sections5.4.1 and 5.4.2). Detection sensitivity is dependent on the type of detectorused, as well as the volume amount loaded onto the column: with UV de-tection, minimum detectable concentrations of 300 ng/mL (for a 404 bpDNA fragment, 20 µL sample loop) have been reported, similar to detect-ability feasible with CE-UV detection. HPLC with fluorescence detection,combined with fluorescent labeling, is significantly more sensitive, andcould be applied to low-level DNA detection (Oefner and Bonn, 1994).

3.2 Gel Electrophoresis: Polyacrylamide (PA)and Agarose

3.2.1 Instrumentation and Detection

From all variants of electrophoresis, slab gel techniques have proven oneof the most popular for DNA. In the 1970s, polyacrylamide and agarosebecame the media of choice for separating proteins and nucleic acids. Hori-zontal as well as vertical formats of slab gels are used, as well as cylindri-cal rods for certain applications (Rickwood and Hames, 1983; Andrews,1986). After electrophoresis of the DNA (the process is stopped after themarker reaches the edge of the gel), the gel is typically soaked in a solutionof ethidium bromide (or other fluorescent dye), washed to remove unbounddye, illuminated with UV light, and photographed to reveal the fluores-cence of dye-bound DNA. (Note that in CE the separation and detection ofDNA takes place simultaneously, i.e., in real time. Some slab-gel-basedinstrumentation also allows real-time detection of DNA, e.g., in automatedDNA sequencing).

Another sensitive, albeit time-consuming, detection technique, autora-diography, is often used to identify specific DNA sequences, e.g., those indisease-causing genes or DNA sampled from crime scenes. The procedure,known after its inventor as “Southern blotting” (Southern, 1975), involvesthe use of radiolabeled DNA probes which target complementary DNAstrands present in the original sample. One forensic typing methodologyuses Southern blotting with autoradiography for identification (“DNAfingerprinting”) of materials found at crime scenes. This method, known asrestriction fragment length polymorphism (RFLP) typing, is shown inFigure 5.

16

YieldGel

TestGel

��

��

��

��

��

��

��

��

Whole Bloodor Stains

IsolateNuclei

RecoverDNA

Cut DNA intoFragments

SeparateFragments by Size

Transfer DNAto Membrane

Add LabeledDNA Probe

AnalyzeDNA Profiles

DevelopX-ray Film

Wash Membranes

Figure 5. DNA profiling process using multistrip restriction fragment lengthpolymorphism (RFLP) analysis. Source: Lifecodes Corporation, Stamford, CT.

17

Slab gels have the distinct advantage of running multiple samplessimultaneously. This, of course, increases sample throughput compared toinstrument designs where the sample has to be assayed one at a time, suchas HPLC or CE. However, CE has—because of the higher fields applied—much faster run times than slab gel electrophoresis (≈ 10 to 100 timesfaster). Instrumentation allowing samples to be run simultaneously onmultiple CE columns has also been described recently and applied to PCRfragments and DNA sequencing (see also Section 5.3). Beckman Instru-ments, Inc., has already previewed a CE instrument with seven capillariesdesigned for clinical applications.

3.2.2 Separation Mechanism

While the original purpose of the gel was to provide an anticonvectivemedium for the electrophoresis, the gels resolve DNA fragments (and alsoSDS-complexed proteins) by acting as molecular sieves. The amount ofsieving can be controlled by adjusting the concentration of the gel. Withpolynucleotides, the phosphate group of each nucleotide unit carries astrong negative charge that is much stronger than any of the charges on thebases above pH 7. The mass-to-charge ratio of the polynucleotides is inde-pendent of the base composition and, consequently, nearly the same forclosely related species. For that reason, free-solution techniques (i.e., theelectrophoretic medium contains no gel or polymer network solution) havenot proven successful in the electrophoresis of oligonucleotides or dsDNA.

The actual mechanism involved in DNA separations in electrophoresis hasbeen—and is still—the topic of much discussion (see, for example, Stell-wagen, 1987; Slater and Noolandi, 1989; Smisek and Hoagland, 1990; Barronet al., 1994). Fluorescent video microscopy allows individual DNA moleculesto be monitored as they sieve through a gel under the influence of an electricfield (Smith et al., 1989). However, no single model can fully account for thedependence of DNA mobility on its molecular size and a number of experi-mental parameters (field strength, gel concentration, etc.). Two models havebeen proposed which are shown schematically in Figure 6 (see also Theory,Section 2.3.1, Equation 10). The first, commonly referred to as the “Ogston”mechanism, involves a coil of DNA percolating through a network of polymerfibers. It is thought that the coil moves through the gel as if it were a rigid,spherical particle. Its electrophoretic mobility is proportional to the volumefraction of the pores of a gel that the DNA can enter. Since the average poresize decreases with increasing gel concentration, mobility decreases with in-creasing gel concentration and increasing molecular weight. The Ogstontheory does not account for the fact that a relatively large DNA molecule maystretch or deform so it can squeeze its way through the pores.

18

A B

Figure 6. (A) Schematic diagram of a solute migrating through a polymernetwork by the Ogston mechanism. The DNA percolates through the meshas if it were a rigid particle. (B) Schematic diagram of a solute migratingby the reptation mechanism. In this case, the DNA is forced to squeezethrough the tubes formed by the polymer network. Reproduced with per-mission from Grossman, P. D., and Soane, D. S., Biopolymers 31, 1221(1991).

In the reptation model, the DNA is thought to squeeze through themesh of the gel as if it were a snake going through an obstacle course (seeEquation 11 in Section 2.3.1). The “biased” reptation model (Slater andNoolandi, 1989) can account for the experimentally observed fact that themobility of DNA becomes independent of the molecular weight at highfield strength (Figure 7). At high field strengths, the DNA stretches and thedependence of mobility on molecular size decreases (see Equation 12 inSection 2.3.1). However, the reptation theory cannot explain certain effectsobserved in pulsed-field electrophoresis (Smisek and Hoagland, 1990).

Low Field Moderate Field High Field

Figure 7. Schematic illustration showing the elongation of DNA under theinfluence of the electric field (reptation). Reproduced with permission fromGrossman, P. D., and Soane, D. S., Biopolymers 31, 1221 (1991).

19

Different gel matrices have been used for separating DNA. The twomost common ones are based on either polyacrylamide or agarose and theywill be discussed next. Interestingly, gels are currently not the only me-dium for electrophoresis of DNA: Volkmuth and Austin (1992) recentlydemonstrated—using a microlithography technique—that DNA can elec-trophorese in arrays of micron-sized posts on silicon chips. This “obstaclecourse” is a highly regular and well-characterized medium for DNA study.

3.2.3 Polyacrylamide Gel Electrophoresis (PAGE)

Polyacrylamide gels are most often used for relatively small DNA frag-ments (DNA sequencing gels, restriction fragments), small RNA moleculesand proteins (SDS-PAGE). Polyacrylamide is usually prepared by cross-linking acrylamide with N,N'-methylenebisacrylamide in the cast with theshape of a slab or a rod in which the electrophoresis is to be carried out(“pouring of the gel”). Linear polyacrylamide (i.e., non-crosslinked) hasalso been used with classical slab gel or column techniques (Bode, 1977;Tietz et al., 1986), but has not found widespread utility. However, it hasbeen found to be extremely useful when used in capillaries, i.e., in modernCE, where it is used as a replaceable, relatively non-viscous, polymer net-work in capillaries (see Section 4.2).

3.2.4 Agarose Gel Electrophoresis

Agarose has larger pores than PA and is generally used for the separationof relatively large DNA molecules. It is a natural product (a polysaccha-ride) made from agar-agar which is isolated from algae. Agarose gels tendto have high mechanical strength and biological inertness. Unlike, PA,agarose does not become crosslinked during the gelation process. A 0.8%agarose gel is sufficiently rigid that it can separate DNA fragments withMWs of up to 50 million, while dilute 0.2% agarose gels have been appliedto fragments of up to 150 million MW. Although agarose has been usedwith CE (Bocek and Chrambach, 1991, 1992), it has—in contrast to PA—not yet been widely applied; prepacked capillaries or gel kits with agaroseare not yet commercially available.

3.2.5 Pulsed-Field Gel Electrophoresis (PFGE)

Gel electrophoresis with dc fields can separate DNA fragments up to ≈ 30kb. Larger DNA fragments (up to ≈ 6 Mb in size, chromosomal DNA) canbe resolved by using ac fields, i.e., in pulsed-field gel electrophoresis(PFGE; see Gardiner, 1991, for a review). The principle of PFGE is

20

sketched in Figure 8. It is interesting to mention that the MW range ofelectrophoresis (as extended by PFGE) is much greater than that possiblewith any mode of chromatography; here the liquid flow that drives theseparation causes shearing of the fragile, large DNA. The problems en-countered with “normal” electrophoresis when large DNA is separatedhave been attributed to the stretching and reptation schematically shown inFigure 7. Under these conditions, the molecular sieving mechanism is nolonger operative. However, by pulsing the electric field, the reptation ofDNA is counteracted. At the time of this writing, research in CE withpulsed fields has just begun (Sudor and Novotny, 1994; Kim and Morris,1994). Pulsed-field CE separations with λDNA standards (8.3–48.5 kb)and restriction-digested λDNA (48.5 kb to 1 Mb) were ≈ 10 to 50 timesfaster than typical slab gel separations of this kind.

A – – B

B + + A

Figure 8. Diagram of pulsed-field electrophoresis of DNA in an agarosegel. The dotted lines indicate the sample wells. A and B represent two setsof electrodes. When the A electrodes are on, the DNA is driven downwardand to the right. When the A electrodes are turned off, the B electrodes areactivated, which causes the DNA to move downward and to the left. Repro-duced with permission from Gardiner, K., Anal. Chem. 63, 658 (1991).Copyright: American Chemical Society.

21

4 CE Methods: Principles andStrategies

In this section, a brief overview will be given of the CE systems used in thevarious type of nucleic acid applications. Generally, untreated capillariesare used in conjunction with the free-solution techniques of CZE orMECC. They are applied to the separations of bases, nucleosides, or simpleoligonucleotides. Larger oligonucleotides (> 10 mers), antisense DNA, anddsDNA (restriction fragments, PCR products, etc.) require gel-filled capil-laries for molecular sieving.

4.1 Free Solution Methods

4.1.1 CZE with Untreated or Coated Capillaries

In most applications of CE involving small molecules (also peptides andproteins), untreated, fused-silica capillaries are used. When the capillary isfilled with solution, a negative charge on the wall surface prevails (Figure9A). In P/ACE, the detection window is positioned near the cathode (nega-tive electrode) at ≈ 7 cm from the capillary outlet. This configuration isreferred to as the “normal polarity” mode. The electroosmotic flow (EOF)generated with untreated capillaries is strongly pH dependent, and variesaccording to an S-function (at high pH, EOF is relatively large); the direc-tion of the EOF in an untreated capillary is always toward the cathode(detector side). When a neutral buffer (e.g., phosphate, pH 7) is used fornucleotide separations, the electrophoretic flow forces the negativelycharged nucleotides toward the anode, opposite to the direction of theEOF (the zone of anions in Figure 2). Detection of the nucleotides willonly take place if the existing EOF is larger than the electrophoretic flow(µEOF > µe).

22

A

B

Electroosmotic Flow

Electroosmotic Flow

+-

+-Figure 9. Schematic diagram of the charge-reversal process at the capil-lary wall. (A) No surfactant added. (B) Micellar bilayer formation by hy-drophobic interaction between the nonpolar chains, resulting in a reversalof the electroosmotic flow.

Using a solution of cationic surfactants (e.g., CTAB, DTAB), thesurface of the fused-silica capillary can be modified in situ from a negativeto a positive surface charge (Figure 9B). The capillary wall is said to be“dynamically” modified. When the concentration of the surfactant in thebuffer is below its critical micelle concentration (CMC), CZE conditionsexist. When the CMC is exceeded, a pseudo-stationary, micellar phase iscreated in the capillary and MECC conditions prevail; the charge on thecapillary—and therefore the direction of the EOF—is reversed, as shownin Figure 9B. To drive the EOF again toward the detector, the polarity ofthe power supply must be reversed (cathode at the injection side—“re-versed polarity mode”).

Generally, the strategy with permanently coated capillaries is to mod-erate or entirely eliminate the EOF. The result is often a more efficient andreproducible separation. Polyacrylamide-coated capillaries have been usedfor nucleotide separations in the CZE mode (Huang et al., 1992). TheeCAP™ Neutral capillary (P/N 477441) from Beckman—which also has apolyacrylamide-based, hydrophilic coating—should also prove useful forthese types of separations.

23

4.1.2 CZE with Sample Stacking or ITP Preconcentration

In classical electrophoresis, sample preconcentration (“enrichment”) can beachieved with discontinuous buffer systems (Ornstein, 1964; Andrews,1986). Similar tricks to obtain an increased sample detectability can also beapplied in CE (Chien and Burgi, 1989; Albin et al., 1993). This is espe-cially useful when the analytes’ concentrations are very low (e.g., in bio-logical media) and UV detection is pushed to its limits (note: LIF detectionmay be a good alternative in these cases as will be discussed in Section4.3.1). Dissolving the sample components in a matrix which has a ≈ 10 Xlower ionic strength than the run buffer may result in sharper peaks and 2to 3 X better detectability (Figure 10). Other similar stacking effects can beobtained by manipulating the pH of the sample vs. that of the run buffer(Abersold and Morrison, 1990) or by using electrokinetic injection(Schwartz et al., 1991).

–––

–

––

–

––

––

–

–

Interface

––––

––

– –

– ––

–

–

––––

––

–

(+)

(+)

(–)

(–)

SampleIntroduction

ApplyVoltage

SampleStacking

Sample Plug(low conductivity)

Sample Buffer(high conductivity)

Figure 10. The sample “stacking” mechanism. Sample ions have an en-hanced electrophoretic mobility in a lower conductivity environment (i.e.,elevated local field strength. When a voltage is applied, anions (e.g., poly-nucleotides) in the sample plug instantaneously accelerate toward theadjacent separation buffer zone where, on crossing the boundary, ahigher-conductivity environment (lower field strength) causes a decreasein electrophoretic velocity and “stacking” of the sample ions into a bufferzone smaller than the original sample plug. Adapted with permission fromOda and Landers, Handbook of Capillary Electrophoresis, Landers (Ed.),Boca Raton, CRC Press, 1994.

24

More dramatic increases in sample detectability (≈ 100 times) can beachieved by applying the principles of isotachophoresis (ITP) to CE. ITPcan be used as a preconcentration step prior to separation in the CZE mode.Using homemade systems, Foret and co-workers (1990) have shown theutility of ITP-CE for the separation of nucleotides. ITP-CE can be per-formed in commercial CE instruments such as P/ACE. The leading andterminating electrolytes must be carefully selected to achieve the desiredzone-focusing effect (Schwer and Lottspeich, 1992; Foret et al., 1992,1993).

4.1.3 Micellar Electrokinetic Capillary Chromatography(MECC)

MECC takes advantage of the differential partitioning of analytes into apseudo-stationary phase consisting of micelles. The principle of this tech-nique was discussed earlier (Figure 4). Anionic surfactants such as SDS aretypically used in concentrations of 10 to 200 mM. When cationic surfac-tants such as CTAB are used, the EOF is reversed; therefore, the polarityof the power supply also must be reversed to attain the same flow directionfrom the inlet to the outlet. MECC conditions appear particularly useful inthe analysis of nucleotides in biological media such as cell extracts (seeSection 5.3.1).

4.2 Capillary Gel Electrophoresis (CGE)Methods

A gel is often somewhat vaguely defined: “a form of matter intermediatebetween a solid and a liquid” (Tanaka, 1981). Gels may vary in consis-tency from viscous fluids to fairly rigid solids. In the electrophoresis litera-ture, the nomenclature is also rather ambiguous: terms such as “polymersolutions,” “polymer networks,” “entangled polymer solutions,” “chemicalgels,” “physical gels,” and “liquid gels” all have been used to describe gelmedia. In addition, the term “replaceable” gel recently has been introducedto describe relatively non-viscous gels that can be rinsed in and out of thecapillary. In CE, two types of gel matrices can be distinguished: 1) a rela-tively high-viscosity, crosslinked gel that is chemically anchored to thecapillary wall (“chemical” gel), and 2) a relatively low-viscosity, polymersolution (“physical” gel). With both types of gels, precoated capillaries areused to eliminate the EOF. Table 2 summarizes the main differences be-tween these gels. Figure 11 illustrates the formation of physical and chemi-cal polymer networks.

25

Table 2. Characteristics of Gel Matrices Used in CGE

Chemical Gels• Crosslinked and/or chemically linked to the capillary wall• Well-defined pore structure• Pore size cannot be varied after polymerization• Heat sensitive• Particulates can damage gel matrix• Provide extremely high resolution (oligonucleotides)• Not replaceable; generally high viscosity

Example: eCAP ssDNA 100 for oligonucleotide separations

Physical Gels• Not crosslinked or attached to the capillary wall• Entangled polymer networks of linear or branched hydrophilic

polymers• Dynamic pore structure• Pore size can be varied• Heat insensitive• Particulates can be easily removed• Gel is replaceable when a relatively low-viscosity matrix is used

Example: eCAP dsDNA 1000 for PCR fragment analysis

26

A B

C

Figure 11. Long-chain molecules cause a solution to become viscous (A)because they interfere with one another as the solution flows. As theirconcentration increases, the molecules become entangled, yielding a vis-coelastic behavior that partakes of both solid and liquid traits (B). If theintertwined molecules bond with one another, the result is a crosslinkedgel (C). Adapted with permission from “Intelligent Gels,” Osada, Y., andRoss-Murphy, S. B., Scientific American 268, 82 (1993). Copyright Scien-tific American, Inc. All rights reserved.

27

The chemical gels have a well-defined pore structure, they are rigid,and their pore size can be varied by adjusting the polymerization condi-tions (e.g., by varying the monomer and crosslinker concentration ratio inthe PA gels), as is well known from the practice of classical electrophore-sis. As shown below, with these type of gels in CE, extremely high-resolu-tion separations of oligonucleotides can be achieved (e.g., Guttman et al.,1990; Schomburg, 1993).

The non-crosslinked, replaceable polymer networks have a dynamicpore structure and are more flexible. Polymer networks of variable viscos-ity can be made by carefully selecting the concentration and chain lengthof the linear polymers. With CE, solutions consisting of linear PA, variousalkylcelluloses, low-melting agarose, and other polymers have been em-ployed (see Section 4.2.3 for a listing). Conceivably, many other polymersare also suitable to serve as gel media for CE.

It is interesting to note that very dilute polymer solutions (≈ 0.010 to0.001% hydroxyethylcellulose) are still effective in size-separating DNA(Barron et al., 1994). A fully entangled polymer network does not appearto be a prerequisite for separation. A mechanism other than the Ogston andreptation models (Figure 6) may be operative with DNA electrophoresis inentangled polymer solutions. As envisioned by Soane and co-workers(Barron et al., 1993 and 1994), when DNA migrates through a polymernetwork, entanglement coupling between the DNA and the surroundingpolymer chains occurs, as schematically shown in Figure 12. Ultimately,this coupled entanglement is thought to limit the resolution achievable withrelatively large DNA fragments (Bae and Soane, 1993).

Figure 12. A schematic representation of the entanglement coupling inter-action of DNA with the polymer chains of the sieving matrix. Reproducedwith permission from Barron et al., J. Chromatogr. A 652, 3 (1993).

28

An important advantage of the physical gels is that a low enough vis-cosity can be selected so that the contents of the capillary can be rinsed inand out of the capillary (in P/ACE by 20 psi positive pressure). As men-tioned before, these gels are often referred to as replaceable matrices.When desired, therefore, a fresh gel can be used for every sample injection.In addition, sample introduction is possible by either the pressure or theelectrokinetic mode, in contrast to the chemical gels (and high-viscositypolymer networks) where only the electrokinetic mode is possible (see alsoSection 4.2.7). This advantage of having the pressure injection mode avail-able can be important in work dealing with quantitation. With polymersolutions, CE can be performed at relatively high temperatures (50 to70°C) and field strengths (1000 V/cm) without damaging the gel as wouldbe the case with chemical gels such as crosslinked PA gels.

Beckman offers three, linear-PA-based capillary gel kits; their featuresare described in Table 3. Gel-filled capillaries recently have been reviewedby Baba and Tsuhako (1992), Guttman (1994), Schomburg (1993), andPariat et al. (1993).

29

Table 3. Beckman Capillary Gel Kits

Feature eCAP ssDNA eCAP dsDNA LIFluor™ dsDNA100 Kit 1000 Kit 1000 Kit

Gel Format Fixed, polyacryl- Replaceable Replaceableamide/urea- entangled entangledcontaining gel polymer polymer

DetectionMethod UV or LIF UV UV or LIF

Nucleic Acid RNA and single- Double-stranded Double-strandedType stranded DNA DNA DNA

Size Range Optimal resolution Molecular weight Molecular weightup to 100 bases linearity 100 to linearity 100 to

1000 base pairs 1000 base pairs

Resolution Single base up to 5–15 base pairs 5–15 base pairs100 bases typical for frag- typical for frag-

ments < 400 base ments < 400 basepairs pairs

Injection Electrokinetic Hydrodynamic HydrodynamicTechnique for quantitation for quantitation

without sample without samplepreparation; preparation;Electrokinetic Electrokineticfor maximum for maximumsample loading sample loading

MinimumDNA Con- ≈ 7 ng DNA/mL ≈ 500 ng DNA/mL≈ 1 ng DNA/mLcentration

Detection Not determined ≈ 150 attomoles ≈ 300 zeptomolesLimits per fragment by per fragment by(On Column) UV detection LIF detection

Applications Primer/probe PCR product PCR productpurity and anti- and RFLP and RFLPsense DNA purity analysis analysis

30

4.2.1 Polyacrylamide

The polymerization of crosslinked and high-viscosity (i.e., > 6%T), linearpolyacrylamide (PA) is carried out within the fused-silica capillary tubing.For reasons of stability, the PA gel must be covalently bound to the wall ofthe column, e.g., by means of a bifunctional agent such as (3-methacryl-oxypropyl)-trimethoxysilane. After completion of this pretreatment proce-dure (Guttman, 1994), the polymerization reaction can take place in thecapillary.

Crosslinked PA gels are used mainly for ss oligonucleotide separationsof up to 200 bases, usually under denaturing conditions of 7 to 9 M urea(Section 5.2). The typical gel concentration is 3–6%T with 3–5%C. Theresolving power depends on the length of the capillary (20 to 150 cm).Longer columns give higher resolution at the expense of longer separationtimes. Figure 13 illustrates the ultra-high resolving power feasible withthese types of gel capillaries. The 160-mer in the lower trace of Figure 13has a plate count of 30 million plates per m, while the peak width is only afew seconds! The high resolving power of the crosslinked gels is the rea-son why they often have been used in capillary DNA sequencing (see alsosection 5.3). However, medium-viscosity, linear PA gels (3–6%T) areeasier to work with and can be replaced by pressure rinsing of the CE in-strument.

Denaturing and the non-denaturing systems can be used with PA gelcapillaries. Denaturing PA gel-filled capillary columns are utilized mainlyfor size separation of short ssDNA (up to several hundred bases, e.g., DNAprimers and probes), and in DNA sequencing. The most commonly useddenaturing agent is urea, while formamide is useful in some applications(Ruiz-Martinez et al., 1993). Beckman offers a kit for oligonucleotideanalysis (eCAP ssDNA 100) which contains prepacked, gel-filled capillar-ies, Tris-borate buffers, 7 M urea and standards. Figure 14 illustrates thehigh resolving power of the eCAP ssDNA 100 column using the normalP/ACE configuration (injection at the long end of the capillary). Fast runtimes are possible by using the 7 cm, short end of the capillary with reversesample injection (Figure 15). Non-denaturing PA gels may be useful whensubtle differences based on the shape, size, and charge of the molecules areexploited (Guttman et al., 1992).

31

25Time (min)15

24.8Time (min)24.0

0.001 AU

Figure 13. CGE of polydeoxythymidylic acid, p(dT)20-160. The lower traceshows a blowup of the 24.0 to 24.8 min time interval, with the largest peakshowing an efficiency of 30 million plates per m. Reproduced with permis-sion from Guttman et al, Anal. Chem. 62, 137 (1990). Copyright: Ameri-can Chemical Society.

0.006

0.004

0.002

0.00023 25 27 29 31 33

Time (min)

Abs

orba

nce

254

nm

Figure 14. Single-base separation of p(dA)40-60 achieved using the eCAPssDNA 100 Kit.

32

0.004

0.003

0.002

0.001

0.0003.53.0 4.54.0 5.55.0

Time (min)

Abs

orba

nce

Figure 15. High-speed separation of p(dA)25-30 accomplished using the7-cm portion of the eCAP ssDNA 100 capillary.

The low-viscosity, linear PA polymer networks are not covalentlybound to the capillary wall. In this case, the capillary is pre-coated with asuitable polymer, e.g., with linear PA which forms a monomeric layer onthe inside of the capillary wall surface. With homemade replaceable capil-laries, linear PA is generally used at low concentrations (1.5 to 6%). For thebest polymerization reproducibility, it is recommended to prepare a highconcentration gel (9 to 12%) that can be diluted to the appropriate concen-tration prior to use (Kleemiss et al., 1993; Guttman, 1994). Karger’s groupat Northeastern University recently described the performance characteris-tics of the replaceable columns at various linear PA concentrations (Pariatet al., 1993). With a 6% linear PA capillary, the average peak efficiency wascalculated as 4 million plates per m in the 51 to 267 bp region, making singlebp resolution possible in this range. Comparisons of CGE with agarose slabgel electrophoresis for DNA digest analysis was done by Paulus and Husken(1993). CGE offered better resolution, especially in the < 600 bp range.

Non-denaturing, replaceable gel media are used (1.5 to 6% PA) for theseparation of dsDNA molecules, such as restriction fragments or PCRproducts (see Section 5.4). Figure 16 shows the separation of a pBR322DNA-Msp I digest restriction fragment mixture using a replaceable PA gelmatrix. The addition of a DNA intercalator to the run buffer (ethidiumbromide in Figure 16) improves resolution (see also Section 4.2.4). Run-to-run reproducibility in these type of capillaries is excellent (typically≈ 0.2% RSD for migration times of DNA standards). Once conditioned, thereplaceable capillaries can be used on a daily basis for months, e.g., forPCR product analysis.

33

17

9–11

16

151413

12876

543

2

171615141312

89 101176

543

2

1

1

pBR 322 DNA-Map I digestno EtBr100 V/cm

A

with 1 µg/mL EtBr200 V/cm

B

0.020

0.015

0.010

0.005

0.000

Cha

nnel

A: A

bsor

banc

e

0.010

0.005

0.000

Cha

nnel

B: A

bsor

banc

e

18 20 30 40 50 55Time (min)

Figure 16. Effect of ethidium bromide on the CGE separation of pBR322DNA-Msp I digest restriction fragment mixture. (A) No ethidium bromide;(B) 1 mg/mL ethidium bromide in the gel–buffer system. Peaks (base pairs):1 = 26, 2 = 24, 3 = 67, 4 = 76, 5 = 90, 6 = 110, 7 = 123, 8 = 147,9 = 147, 10 = 160, 11 = 160, 12 = 180, 13 = 190, 14 = 201, 15 = 217,16 = 238, 17 = 242. Conditions: (A) 100 V/cm, (B) 200 V/cm. Reprintedwith permission from Guttman and Cooke, Anal. Chem. 63, 2038 (1991).Copyright: American Chemical Society.

4.2.2 Agarose

Agarose gels are characterized by large pore sizes, high mechanicalstrength, and biological inertness. As agarose is the medium of choice forthe separation of relatively large DNA with slab gels, it seems logical thatagarose would also be tried in capillaries. Compton and Brownlee (1988)showed preliminary results of DNA separations with agarose. Since then,only a few research groups (Bocek and Chrambach, 1991, 1992; Schom-burg, 1993) have reported results with agarose-filled capillaries; they arenot commercially available. Special purified grades of agarose must beused to avoid unwanted EOF in the capillary; in addition, agarose solutionsmust be optically clear for UV detection at 260 nm. Theoretical and practi-

34

cal aspects of DNA sieving in agarose have been published by Stellwagen(1987) and Upcroft and Upcroft (1993), respectively. DNA entanglementas it sieved through low-percent agarose was studied by Smisek andHoagland (1990).

CGE of DNA restriction fragments with agarose concentrations be-tween 0.3 to 2.6% at 40°C were described by Bocek and Chrambach (1991,1992). The advantage of employing liquefied agarose above its “gelling”temperature is that the capillary is replaceable, i.e., it can be easily filled,rinsed, and refilled. Another advantage of this type of agarose is that itsbackground absorbance at 254 to 260 nm is sufficiently low that DNAdetection at the nanogram level is possible. With an agarose sieving ma-trix, the inner surface of the capillary must be coated with a suitable poly-mer, e.g., linear polyacrylamide. Using this technique, the effective sizerange for separation of dsDNA was limited to ≈ 12 kb.

4.2.3 Alkylcellulose and Other Polymers

Apart from polyacrylamide and agarose, size separation of DNA in CE canbe obtained by the use of various other entangled polymer solutions:

• hydroxyethylcellulose (HEC), e.g., Grossman and Soane, 1991;Nathakarnkitkool et al., 1992.

• hydroxymethylcellulose (HMC), e.g., Zhu et al., 1989; MacCrehan etal., 1992; Kim and Morris, 1994.

• hydroxypropylmethylcellulose (HPMC), e.g., Schwartz et al., 1991,1992; Ulfelder et al., 1992; Baba et al., 1993.

• polyacryloylaminoethoxyethanol, e.g., Chiari et al., 1994; Nesi et al.,1994.

• ficoll-400, e.g., Righetti et al., 1991.• polyethyleneglycol, e.g., Zhu et al., 1989; Schwartz et al., 1991.• glucomannan, e.g., Izumi et al., 1993.• polyvinyl alcohol , e.g., Kleemiss et al., 1993.

As suggested by Soane and co-workers (Bae and Soane, 1993; Fig-ure 12), a DNA fragment may be loosely entangled with matrix moleculesas it is pulled through the solution by the electric field. The longer theDNA fragment, the more entanglement points will exist, and the slower thefragment will move through the capillary. Indeed, in CE of DNA, the elec-trophoretic mobility decreases with increasing fragment size (see equationsin Section 2.3.1).

35

As with the PA-based sieving matrices, a precoated capillary is desir-able with other sieving media. Commercially available polysiloxane coatedcapillaries (e.g., DB-17 from J & W Scientific) can be used for this pur-pose, as well as others such as those coated with polyvinyl alcohol(Schomburg, 1993) or polyacrylamide (Strege and Lagu, 1991). It shouldbe noted that the cellulose additives form an additional “dynamic” coatingon the inner surface of the capillary wall—the EOF is significantly de-creased (Schwartz et al., 1991). DNA fragments will migrate in thesecoated capillaries in order of increasing size. (Note: if untreated capillariesare used in conjunction with alkaline buffer conditions, the reverse order ofelution has been observed due to the effect of the EOF—see Grossman andSoane, 1991.)

The sieving of DNA through the medium can be manipulated by vary-ing the chain length and the concentration of the polymer. This is illus-trated in Figure 17 for a number of different polymers by plotting themobility vs. the size (bp) of the DNA fragment on a semi-log scale. S-shaped curves—with a linear middle section—can be observed. The shal-lowness of the slope of the curve is a measure of the sieving power of themedium. The sieving depends on the viscosity of the medium and the poly-mer chain length. For example (panel A), the sieving is better at a higherconcentration of the HPMC polymer (see also Figure 18). Comparison ofpanels A and B reveals that, at the same concentration, the short-chainpolymer (HPMC-100) yields greater mobilities than the long-chain poly-mer (HPMC-4000), in agreement with earlier findings of Bode (1977).Panel C shows a steep curve for another sieving buffer consisting of 5%polyethylene glycol. It appears that this polymer solution would not be aneffective sieving matrix for the DNA separations. Other publications deal-ing with the effect of polymer chain length are from Barron et al. (1993,1994) and Baba et al. (1993).

36

0.7% 0.5% 0.35% 0.1%

HPMC 4000 cps A

10000

1000

100

101.4000 2.4000 3.4000x 10-4

1.0% 0.7% 0.5%

HPMC 100 cps B

10000

1000

100

101.4000 2.4000 3.4000x 10-4

PEG & PAG C

10000

1000

100

101.4000 2.4000 3.4000x 10-4

5% PEG3%T, 0.5%C PAG

BP

BP

BP

Mobility

Figure 17. Effect of different polymer additives on molecular sieving. Theplots (semilogarithmic scale) show the dependence of mobility on the basepair number. DNA fragments from the HaeIII restriction digest of φX wereused as base pair markers. Polymeric additives: (A) HPMC-4000 at 0.1,0.35, 0.5, and 0.7%; (B) HPMC-100 at 0.5, 0.7, and 1.0%; (C) 5% PEGand polyacrylamide (3% T, 0.5% C). Reproduced with permission fromSchwartz et al., J. Chromatogr. 559, 267 (1991).

37

0.001 AU

0.001 AU

0.001 AU

A

B

C134

154 220 298 396506

1018

91628144

71266106

50904072

3054

20361636

101801119812216

517

201 344

10 15 20 min

Figure 18. Effect of the concentration of HPMC-4000 on the CE separa-tion of the 1 kbp DNA ladder. Concentration of HPMC-4000: (A) 0.1%,(B) 0.3%, and (C) 0.7%. Reprinted with permission from Baba et al.,J. Chromatogr. A 653 , 329 (1993).

4.2.4 Intercalators as Buffer Additives

DNA-binding or intercalating dyes have been used for fluorometric DNAassays and in flow cytometry applications. Ethidium bromide (EB) was thefirst of such intercalators to be used for DNA assays (Le Pecq and Paoletti,1967). Since then, a wide variety of even more sensitive dyes have beendeveloped (see, for example, recent work by Glazer and Rye (1992).

Interestingly, the resolution of dsDNA separations in CE can be im-proved by using intercalating dyes (Schwartz et al., 1991 and Guttman andCooke, 1991). This is usually done by adding dye to the running buffer(and/or sample) in concentrations of ≈ 0.5 to 5 µg/mL. The effect of EB onthe separation of DNA restriction fragments was shown earlier in Fig-ure 16. As shown in Figure 19, the dye molecule inserts itself (“interca-lates”) between the base pairs of DNA, changing the molecular persistencelength, conformation, and charge of the DNA. Since EB bears a positive

38

charge (opposite to that of DNA), the EB intercalation increases the migra-tion times of all dsDNA fragments. Because of this complexation (approxi-mately one EB molecule per five bp) and the increasing rigidity of thecomplex, the larger DNA molecules move relatively more slowly. There-fore, the separation time window widens, which increases peak capacity.Intercalators such as EB are, therefore, useful for manipulating migrationtime and separation. For example, it is even possible to achieve baselineseparation of dsDNA species that have an identical chain length but arecomposed of a different sequence (see the two 147 bp fragments and thetwo 160 bp fragments in Figure 16).

N+

C2H5

Br –

NH2H2N

Ethidium bromide

Nucleotide Intercalatedmolecule

Backbone

Figure 19. The intercalation of ethidium bromide into a DNA molecule.Ethidium bromide increases the spacing of successive base pairs, distortsthe regular sugar–phosphate backbone, and decreases the pitch of thehelix. Reproduced with permission from Watson et al., Molecular Biologyof the Gene, Menlo Park, CA, Benjamin/Cummings Publishing Company,1987.

39

It is important to note that the intercalating dyes strongly fluorescewhen excited with an appropriate light source. This opens up the possibil-ity of using (laser-induced) fluorescence detection with CE. As will bediscussed in Section 4.3, extremely sensitive LIF detection methods can beutilized for tracing minute amounts of DNA.

4.2.5 Ferguson Plots

The size selectivity of sieving media (e.g., PA slab gels) is often character-ized by “Ferguson” plots (Andrews, 1986; Stellwagen, 1987). Figure 20,from Heiger et al. (1990), shows the PA gel concentration (expressed as%T) vs. the log mobility for linear PA in a CE system. The larger the DNAfragment, the steeper the slope, in accordance with the sieving theory (seeEquation 6, Section 2.3.1). From the intercept of the plot, the free solutionmobility (i.e., in the absence of a gel matrix) of a DNA fragment can becalculated. From Figure 20 it also appears that the mobility of the DNAfragments in free solution, at 0%T, is virtually independent of the size ofthe DNA fragments examined. Schwartz et al. (1991) showed similarFerguson plots for sieving of DNA fragments in polymer networks ofalkylcellulose.

-3.6

-3.7

-3.8

-3.9

-4

-4.1

-4.2

-4.3

-4.4

-4.5

-4.6

-4.70 2 4 6 108 12 14 16

% T

log

(mob

ility

)

+

+

+

+

+

+

X

X

X

X

X

+

X

X

= 273 base pairs= 310 base pairs= 603 base pairs= 872 base pairs= 1358 base pairs

Figure 20. Ferguson plots for linear polyacrylamide. The lines representthe log mobility of various φX-174 HaeIII fragments as a function ofmonomer composition. Reprinted with permission from Heiger et al.,J. Chromatog. 516, 33 (1990).

40

4.2.6 Instrument Parameters

When gel-filled capillaries are used for DNA separations, the power supplyof the CE system must be in reverse-polarity mode i.e., with the cathode onthe injection side and the anode on the detection side. DNA is monitored inreal time by UV detection at 254 or 260 nm. When fluorescence or LIFdetection is employed, the excitation and emission wavelengths depend onthe fluor label or dye used. The temperature of the gel-filled capillary col-umn is usually maintained at constant temperature during the experiments(± 0.1°C).

Separation of DNA restriction fragments can sometimes be improvedby using different temperatures during the electrophoresis (Guttman andCooke, 1991 B). Migration times decrease with increasing temperature inthe isoelectrostatic (constant voltage) separation mode and maximize in theisorheic (constant current) separation mode. The resolution between theshort DNA fragments (< 300 bp) decreases in the isoelectrostatic separa-tion mode and shows maxima in the isorheic mode at elevated temperature.However, the efficiency in the higher MW range (> 1000 bp) decreases inboth modes with increasing temperature.

Resolution and analysis time can also be optimized by manipulation ofthe electric field. Methods based on pulsed-field electrophoresis (Sudorand Novotny, 1994; Kim and Morris, 1994), analyte velocity modulation(Demana et al., 1991) and field-strength gradient separation techniques(Guttman et al., 1992 B) have been reported. The latter method can di-rectly be performed with P/ACE; the first two require special instrumenta-tion.

4.2.7 Sample Injection and Matrix Effects; Quantitation

4.2.7.1 Replaceable Gels

Quantitation. With replaceable gels (e.g., the eCAP dsDNA 1000 fromBeckman), both pressure (typically: 2 to 20 sec, 0.5 psi) and electrokineticinjection are feasible. In CE, the pressure injection mode is generally rec-ommended for quantitative work: the composition of the sample plug intro-duced into the capillary is exactly that of the sample vial from which theinjection took place. In addition, sample preparation is simplified as nodesalting needs to be performed. Butler et al. (1994) recently reportedprecision results with replaceable gels. Using pressure injection and aninternal standard, peak migration time precision was < 0.1% RSD, whereas

41

the area precision was ≈ 3% RSD (Table 4). It should be noted, however,that electrokinetic injection often yields more efficient peaks than doespressure injection (Schwartz et al., 1991; Butler et al., 1994). Whenelectrokinetically injected from low-ionic-strength sample solutions, DNAfragments are effectively stacked against the relatively viscous, polymernetwork medium. In separations of small molecules, electrokinetic injec-tion may give rise to a sampling “bias” as sample components, because oftheir different mobilities, move into the capillary at different speeds. How-ever, since DNA fragments essentially have the same mass-to-charge ratioin free solution, no such sample bias occurs when these fragments areelectrokinetically injected from an aqueous solution.

Table 4. Peak Area Precision (RSD) with Replaceable GelsSource: Butler et al., J. Chromatogr. B 658, 271 (1994)

Hydrodynamic Electrokinetic

(A) Internal StandardAdjusted Aarea 8.0% 6.0%Height 8.1% 3.0%Migration Time 0.1% 0.1%

(B) No Internal StandardAdjusted Area 8.4% 28%Height 8.5% 23%Migration Time 0.2% 0.3%

(C) Internal StandardAdjusted Area 3.0%Height 6.7%Migration Time 0.07%

(A) 100 bp DNA compared to 200 bp DNA internal standard for 10 runs. Bothinlet and outlet run buffer vials were changed after run 5. (B) Without using inter-nal standard. (C) Same as (A) but outlet buffer vial was changed after every run.

Matrix Effects; Desalting. In the practice of CE, separation performance isoften strongly dependent on the composition of the sample solution. This isparticularly important when samples are electrokinetically injected andvariable amounts of salt are present. Desalting the sample by ultrafiltra-tion—in conjunction with electrokinetic injection—may enhance sampledetectability, as demonstrated by Schwartz et al. (1991), Ulfelder et al.(1992), and Butler et al. (1994). The ultrafiltration procedure removes low-

42

MW sample constituents, resulting in efficient DNA peaks. Figure 21shows the dramatic effect of desalting the sample for two co-amplifiedPCR products from an HIV-1-positive control cell line—no peaks are vis-ible in the trace corresponding to the untreated sample. However, it hasbeen reported that desalting—when used in conjunction with pressureinjection—may also lead to loss of DNA due to adsorption on the filter(Butler et al., 1994). Thus, when possible, pressure injection—withoutdesalting—is preferable.

A

B

4.0 6.0 8.0 10.0Time (min)

HIV-1HLA

Figure 21. Effect of ultrafiltration on the PCR-amplified DNA peaks.(A) Untreated sample (no ultrafiltration), a co-amplified HIV-1, HLA-positive (115 and 242 bp, respectively) control. (B) Desalted sample.Adapted with permission from Schwartz et al., J. Chromatogr. 559, 267(1991).

Recently, van der Schans et al. (1994) studied sample matrix effectsfor analysis of PCR products with replaceable gels and pressure injection.When the sample plug length was increased, decreased efficiency—appar-ent as fronting peaks—was observed. Sharpening of the peaks can be ob-tained by simply injecting a plug of low resistance, 0.1 M Tris-acetate priorto the sample injection (Figure 22). The lower field conditions existing inthe Tris-acetate plug cause electrophoretic stacking of DNA fragments.

Co-injection of PCR products with a standard is a convenient methodof verifying the identity of the sample peaks. It was found that co-injectionof a DNA standard with the PCR sample can lead to sharpening of thesample peaks or the standard peaks, depending on the order in which theplugs were loaded on the capillary. This effect is shown in Figure 23.In the top trace, the 97 bp PCR sample is injected first, followed by theφX-174 HaeIII DNA standard. It can be seen that, while the standard peaks

43

are sharp, the PCR peak is broadened. The opposite is seen in the lowertrace where the injection order was reversed. The broadened peaks are dueto salt migrating from the PCR sample into the plug containing the DNAstandards. During its migration through the capillary, the back end of thestandard zone will migrate at a slower velocity relative to the front as herethe field strength is lower than at the front end of the plug.

201816141210

0.000

0.001

0.002

0.003

0.004

0.005

Time (min)

UV

Abs

orba

nce

(AU

)

2018161412100.000

0.002

0.004

0.006

0.008

Time (min)

UV

Abs

orba

nce

A

B

Figure 22. (A) Electropherogram of a 50 mg/mL φX-174 RF DNA HaeIIIdissolved in 20 mM NaCl. (B) Influence of presample injection of 0.1 MTris-acetate, pH 8.3. Injection procedure: first injection: 10-s pressureinjection of Tris-acetate; second injection: 20-s pressure injection of50 mg/mL φX-174 HaeIII sample in 20 mM NaCl. Reprinted with permis-sion from van der Schans et al., J. Chromatogr. A 680, 511 (1994).

44

60

50

40

30

20

10

012 14 16 18 2010

Time (min)

Flu

ores

cenc

e

30

20

10

012 14 16 18 2010

Time (min)

Flu

ores

cenc

e

A

B

PCR

PCR

Figure 23. (A) Electropherogram of PCR sample and DNA standard. Firstinjection: PCR sample (97 bp); second injection: φX-174 HaeIII 10 mg/mL. (B) Electropherogram of DNA standard and PCR sample. First injec-tion: φX-174 HaeIII 10 mg/mL; second injection: PCR sample (97 bp).Reprinted with permission from van der Schans et al., J. Chromatogr. A680, 511 (1994).

4.2.7.2 Non-Replaceable Gels

When working with capillaries containing high-viscosity gels (e.g., > 6%linear PA), sample introduction can only be performed by electrokineticmeans (typically 0.015 to 0.15 Ws is applied). This is the case for theeCAP ssDNA 100 column from Beckman, as well as with capillaries con-taining highly viscous alkylcellulose or similarly entangled polymer net-works. Pressure injection is limited with highly viscous media; the volumeof the sample injected is inversely proportional to the viscosity of the

45

buffer. Exceedingly long injection times would be required, making pres-sure injection impractical.

When using the eCAP ssDNA 100 capillary and electrokinetic injec-tion mode for quantitative studies, effects due to the presence of salts orother substances in the sample matrix must be carefully considered. Forexample, when electrokinetic injections are made from samples containingdifferent salt concentrations, the amount of analyte introduced into thecapillary will vary. This, in turn, has consequences for the accuracy of adrug assay. As noted by Srivatsa et al. (1994), with many pharmaceuticalsfor intravenous or ophthalmic use, the products are formulated in isotonicsalt solutions. In this case, it is important to use an external reference stan-dard with the same sample matrix in order to accurately assay a drug prod-uct. In CGE with electrokinetic injection, while the peak migration timeprecision generally is excellent (< 0.2% RSD), peak area precision mayexceed tolerable levels. Precision can be greatly improved, however, by theuse of an internal standard (IS), as shown in Table 5 for a series of electro-kinetic injections of an antisense DNA oligonucleotide (ISIS 2922) and itsN-1 deletion sequence on an eCAP ssDNA 100 capillary.

Table 5. Reproducibility of Integrated Peak AreaSource: Srivatsa et al., J. Chromatogr. A 680, 469 (1994)

Observed Peak Area Normalized to IS

ISIS 2922 N-1 ISIS 2922 N-1

0.168487 0.01072 5.61391 0.351840.111345 0.00714 5.47222 1.644520.154746 0.01114 5.83005 0.410760.162939 0.01040 5.73130 0.357860.196860 0.01257 5.71057 0.356750.115623 0.00722 5.70747 0.336180.104393 0.00650 5.74548 0.351770.281972 0.01698 5.71057 0.356750.362852 0.02071 5.70931 0.34879

Mean 0.1844 0.0122 5.6923 0.3572Std. Dev. 0.0863 0.0051 0.0994 0.0212% RSD 46.80 41.73 1.75 5.94

46

4.2.8 Hybridization; Southern Blotting, Mobility ShiftAssays

In Southern blot hybridization, typically a slab gel is run to separate DNAfragments digested from nuclear DNA (Southern, 1975). After blotting thefragments to a membrane, a hybridization reaction is carried out to identifyparticular nucleotide sequences. When used in conjunction with autorad-iography, this can be an extremely sensitive method (for example in foren-sic applications—see Figure 5). It is also a very selective method, as thetarget DNA will only bind with probe DNA that has a complementarysequence. However, the classical method is laborious and time consuming,often taking several days to complete. Brownlee, Sunzeri, and co-workers(1990, 1991) were the first to demonstrate the feasibility of hybridizationwith fluorescently labeled probes for the detection of target DNA (HIV-1and HTLV-1) sequences. Using rhodamine and FITC-labeled probes, thesesequences could be discriminated in a single run by fluorescence diode-array detection at 563 nm and 519 nm, respectively. Also working withHIV-1 genome sequences, Bianchi et al. (1993) analyzed PCR products bypre-capillary hybridization. A 299-nucleotide ssDNA fragment was hybrid-ized with a complementary 28-mer, resulting in mobility shifts in the elec-tropherogram.

Cohen and co-workers have shown that it is possible to transfer South-ern blotting from a slab gel to a CGE format, thereby greatly reducing theanalysis time. In their first paper (Cohen et al., 1991), preliminary resultswere reported with probes labeled with the fluorescent dye, Joe. Analysiswas made by LIF detection, using a 488 nm Ar-ion laser. In a later paper(Vilenchik et al., 1994), antisense DNA (phosphorothioate) was quantifiedas low as 0.1 ng/mL using fluorescein-labeled probes. Although UV detec-tion also was used, LIF detection (488 nm, Ar-ion laser) gives superiorsensitivity. In addition, background DNA in the detector or sample, whichis problematic with UV detection, does not affect LIF detection. The prin-ciple of the CE method used by Cohen and co-workers is illustrated inFigure 24 (UV traces). Sample preparations containing different amountsof target and probe (kept constant here) were injected onto two capillaries:one under non-denaturing CGE conditions (set A), the other under denatur-ing CGE conditions (set B). The electropherograms show three peaks: thatof the target antisense DNA (“GEM”), that of the probe (“COM”), and thatof the hybrid (“duplex”). As the GEM concentration increased, the duplexpeak also increased in size, while the COM peak size decreased. In the setof denaturing electropherograms, the duplex peak—as expected—is not

47

observed. The order of migration of the target and the probe is differentthan that under non-denaturing conditions, most likely due to secondarystructure differences between the two.

6 8 10 12 14 16 18Time (min)

8 10 12 14 16 18 20 22Time (min)

A BCOM

duplex

1

COM

GEM

GEM

duplex

2

COM

duplex

3

COM

1

COM

GEM

GEM

GEM

2

COM

3

Figure 24. The separation of GEM, COM, and duplex by CGE. Electro-pherograms 1, 2, and 3 show different amounts of GEM with constantCOM concentration. (A) Non-denaturing conditions. (B) Denaturing con-ditions. Conditions: (A) 9%T linear polyacrylamide column, effectivelength = 20 cm, applied electric field = 200 V/cm; (B) 13%T linear poly-acrylamide, denaturing conditions, effective length = 15 cm, applied elec-tric field = 400 V/cm. Adapted with permission from Vilenchik et al.,J. Chromatogr. A 663, 105 (1994).

48

LIF detection of the antisense DNA proved to be quantitative andlinear over three orders of magnitude. However, in the CE-LIF electro-pherograms, the fluorescein-labeled probe co-eluted with the duplex (asopposed to the UV traces where resolution is adequate, Figure 24). Usingethidium bromide (0.04 µM) as an intercalating dye, it was possible toincrease the resolution sufficiently to enable quantitative analysis (see alsoSection 4.2.4).

CE can be used as a tool to study biomolecular, non-covalent reac-tions. The term affinity capillary electrophoresis (ACE) has been coined todescribe the CE of receptor-ligand interactions, including those of antigen-antibody (immunoassays). A number of recent papers have demonstratedthe potential of ACE for DNA studies. For example, Heegaard and Robey(1993) used ACE for the study of oligonucleotide-peptide interactions.Using dimeric peptides as probes, the binding was found to depend on boththe size of the oligonucleotide and the specificity of the interaction. Inanother paper, Rose (1993), using CGE (high-viscosity gel), studied thebinding kinetics of a type of antisense DNA (peptide nucleic acids, PNAs)with oligonucleotides. In PNAs, the deoxyribose-phosphate backbone issubstituted for a peptide backbone. The free PNA, oligonucleotide, as wellas the bound heteroduplexes were efficiently separated and quantitated.

Protein-DNA interactions have also been studied with CE. These inter-actions are involved in control of replication, recombination, modification,repair, and transcription processes. Methods for studying DNA-proteininteractions include mobility shift assays, where slab gel electrophoresis isused to detect a change in mobility of DNA when complexed to a protein.CE can be applied to these types of mobility shift assays, as shown byMaschke et al. (1993) for the binding of an endonuclease, EcoRI, witholigonucleotides. A free-solution CE system with LIF detection (Ar-ionlaser, 488 nm) was used. Joe-labeled oligonucleotides with the EcoRI rec-ognition site, GAATTC, interact with the protein; the complex is detectedas a faster-migrating fluorescent peak. Addition of excess unlabeled probedisplaces the labeled probe in the complex, resulting in the disappearanceof the fluorescent signal.

49

4.3 Detection in CE: UV Absorbance vs. Laser-Induced Fluorescence (LIF)

Optical detection techniques for CE recently have been reviewed byPentoney and Sweedler (1994). In the vast majority of DNA as well asother applications of CE, UV-Vis absorbance detection has been used.Practically all commercial CE instruments are equipped with this detector,which is universal (i.e., suitable for many types of analytes) and also hasadequate sensitivity for most applications. However, in bioscience applica-tions, often trace amounts of an analyte need to be determined in the pres-ence of many other sample components, and detection may becomeproblematic. The detection limit with UV detection is—among other fac-tors—related to the small interior diameter of the capillary; for example,a 200 bp DNA fragment typically has a minimal detectable concentrationof ≈ 0.5µg/mL. Optimal stacking and/or ITP preconcentration methods,as well as optimized optics, may improve the detection limits by a factor of2 to 10. Fluorescence detection, however, may yield far lower detectionlimits and has been used for decades with many DNA applications. Fluo-rescence-based assays have the advantage of offering both excellent selec-tivity and very high sensitivity.

Fluorescence or LIF-based detection, in conjunction with fluor label-ing systems, has been developed for many biomedical applications, e.g.,chromosome sorting, DNA sequencing and fingerprinting. With sophisti-cated instrumentation, it has become possible to examine the protein ornucleic acid content of single human cells or even to detect single mol-ecules of stained DNA. Recently, an LIF detector has been developed byBeckman for CE. Figure 25 shows a detail of the optical design used in theP/ACE 5000 Capillary Electrophoresis System.

50

Figure 25. Schematic view of a CE-LIF detector (Beckman Instruments,Inc., Fullerton, CA). Fiber-optic cable transmits laser light from the laserto the detector and illuminates a section of the capillary. Fluorescence iscollected by the ellipsoidal mirror and focused back onto the photomulti-plier tube. To reduce unwanted laser light, a centered hole in the mirrorallows most of the beam to pass. A beam block is used to attenuate scat-tered laser light.

The P/ACE-LIF interface is supplied with a 488 nm Ar-ion laser butcan be connected to various other laser sources. The detector incorporatesan ellipsoidal reflector to maximize the emission light collection effi-ciency. Table 6 shows a list of laser sources and excitation lines whichhave been used in some CE-LIF applications. Thus far, most work in CEhas been performed with the easy-to-use and relatively low-cost Ar-ion,He-Cd, and He-Ne lasers. These laser sources are also well suited for DNAand nucleotide work. The 488 nm emission of the Ar-ion laser matches the490 nm peak of popular fluorescein-based labels (FITC, fluoresceinsuccinimidyl ester). The 325 nm emission of the He-Cd laser matches OPAand dansyl labels. The compact and even less expensive “green or yellow”He-Ne lasers can be used with various rhodamine derivatives and are com-patible with intercalating dyes such as ethidium bromide or propidiumbromide (Kim and Morris, 1994 B; Liu et al., 1995). Of all the lasersources, the mixed Ar-Kr lasers provide the widest wavelength range avail-able (more than 20 emission lines in the 350 to 725 nm region).

51

Table 6. Examples of Laser Light Sources for Detection with CESource: Schwartz et al., J. Cap. Elect. 1, 36 (1994)

Laser Source Wavelength (nm)

Ar-ion (air cooled) .......................................... 457, 472, 476, 488, 496,501, 514

Ar-ion (full frame) .......................................... 275, 300, 305, 333, 351,364, 385, 457, 472, 476,488, 496, 501, 514

Ar-ion (full frame, frequency doubled) .......... 229, 238, 244, 248, 257

Ar-Kr ............................................................... 350-360, 457, 472, 476,488, 496, 501, 514, 521,531, 568, 647, 752

He-Cd .............................................................. 325, 354, 442

He-Ne .............................................................. 543, 594, 604, 612, 633

Excimer

XeCl (pulsed) ............................................. 308

KrF (pulsed) ............................................... 248

Nitrogen (pulsed) ............................................ 337

Nitrogen pumped dye (tunable) ...................... 360-950

Solid state

YAG (frequency doubled) ......................... 532

YAG (frequency quadrupled) .................... 266

Diode lasers

frequency doubled (LiNbO3) ................ 415

frequency doubled (KTP) ...................... 424

frequency tripled (Nd-doped YLiF) ...... 349

Low-cost, semiconductor (“diode”) lasers can also be used with CE.Their wavelengths are in the 635 to 850 nm range; background fluores-cence from biological sample matrices is strongly reduced at these longwavelengths. The analytically relevant blue light (i.e., wavelengths com-patible with the popular fluor labels) can be obtained from these lasers by

52

“frequency doubling” techniques of (infra)red light. In order to extend theapplicability of the red diode lasers to CE, new labeling reagents (based onthiazine-, oxazine-, and cyanine-type compounds) are being developed,because just a few analytes yield native fluorescence in the red or near-IRregion (Jansson et al., 1993). At Beckman Instruments, Inc. (Chen et al.,1993), a cyanine dye (Cy5) has been evaluated for potential use in DNAhybridization and sequencing (see also Section 5.3, Figure 35).

4.3.1 DNA Detection with LIF

4.3.1.1 Native and Indirect Fluorescence

A number of detection schemes have been used for the detection of nucleicacids by fluorescence methods. One approach is based on the native DNAfluorescence in the low-UV region (Milofsky and Yeung, 1993). Nativefluorescence allows analysis of the DNA molecule in its “natural” state,with excimer lasers (e.g., the pulsed KrF, 248 nm laser) as the excitationsource. A later paper (McGregor and Yeung, 1994) describes some im-provement in detectability (lower background signal) with this approachby using a sheath-flow arrangement while separations were performedat pH 2.8.

Another LIF detection scheme for DNA is based on indirect fluores-cence. A fluorogenic CE buffer system such as salicylate can be used inconjunction with laser excitation (e.g., the 325 nm He-Cd laser). Examplesof nucleotides (Kuhr and Yeung, 1988) and dsDNA (Chan et al., 1993)have been demonstrated using this approach. However, the methods basedon native and indirect fluorescence have not been widely used by otherworkers and special, homemade instrumentation is required.

4.3.1.2 Intercalators

The most straightforward and currently popular LIF detection scheme forCE involves the use of fluorescent intercalators (Schwartz and Ulfelder,1992). Beckman has introduced a CE kit for dsDNA analysis which incor-porates a specific intercalator dye, EnhanCE. The nuclear stain is added tothe CE buffer and/or sample and specifically interacts with sample dsDNAor RNA molecules (see also section 4.2.4). The dye inserts itself betweenthe base pairs of DNA, thereby changing its persistence length, conforma-tion, and charge, resulting in a reduction of electrophoretic mobility butproviding enhanced resolution (Figure 26). More importantly, the DNA-dye complex fluoresces strongly when excited by the appropriate laser,

53

whereas the intercalator alone—as well as non-DNA sample components—generally do not. Hence, separation selectivity is often vastly improvedgiving rise to a much “cleaner” electropherogram than possible with UVdetection (see section 5.4.1, Figure 38).

200

150

100

50

00 200 400 600 800 1000

DNA Fragment Size (bp)

Res

olut

ion

(bp)

LIF

UV

Figure 26. A plot of the resolution capability for both the eCAP andLIFluor dsDNA 1000 Kits, expressed as base pairs resolved for a specificDNA fragment size.

The use of fluorescent intercalating dyes leads to 2 to 3 orders of mag-nitude enhanced sensitivity when compared to UV detection (Schwartz andUlfelder, 1992; Ulfelder, 1993; Rossomando et al., 1994; Zhu et al., 1994;Arakawa et al., 1994 B). This is shown in Figure 27 for φX-174 HaeIII RFDNA fragments: the UV trace shows an appreciable noise level whereas, inthe LIF trace, a “clean” baseline is obtained with no visible noise. In addi-tion, the DNA concentration used for the UV detection was 20 times higherthan that of the LIF.

54

0.009

0.007

0.005

0.003

Orange G

1353

1078

872603

281

310271

234194

118720.001

-0.0018 10.5 13 15.5 18

Time (min)

Abs

orba

nce

(254

nm

)

A

64

48

32

16

0

8 10 12 14 16 18 20Time (min)

72118

194234

271281

310

603

872

1078

1353

Flu

ores

cenc

e (e

m 5

30 n

m)

B

Figure 27. Separation of an HaeIII restriction digest of φX 174 RF DNAusing (A) UV and (B) LIF detection. Samples were diluted in water to atotal DNA concentration of 200µg/mL for UV detection; 10µg/mL for LIFdetection. Injection was by pressure to ten seconds. Buffer systems werethe same, except for the addition of EnhanCE for LIF detection. FromUlfelder, K. J., Beckman Application Information Bulletin A-1748 (1993).

55

Monomeric as well as dimeric dyes have been utilized in recent CE-LIF applications (Zhu et al., 1994; Kim and Morris, 1994 B; Figeys et al.,1994). Ethidium bromide, in conjunction with a green He-Ne laser, canalso be used for sensitive LIF detection (Liu et al., 1995). The detection ofPCR products will be discussed later in Section 5.4. The 488 nm Ar-ionlaser incorporated in the P/ACE-LIF instrument is compatible with manyof these DNA and RNA dyes. It has been reported that the dimeric dyes(e.g., ethidium homodimer, TOTO, YOYO) are best used in conjunctionwith monomeric intercalator additives, e.g., 9-aminoacridine (Zhuet al.,1994); otherwise, broad peaks may result from the presence of multipledye-DNA bonding.

4.3.1.3 Fluorescent Labeling

The final LIF detection scheme for DNA involves direct labeling of theanalyte with a suitable fluorophore. Fluorescently labeled probes and prim-ers are used in many molecular biology applications involving hybridiza-tion and PCR (Mansfield and Kronick, 1993). DNA primers and probes areusually synthesized with a fluorescent label attached to the 5' end of themolecule, or with post-synthesis attachment of a dye using commercialDNA labeling kits. Unincorporated dye and/or failure sequences are gener-ally removed by LC methods. For subsequent use in the PCR leading tofluorescent DNA products, conditions can be optimized such that primerpurification is not necessary. Figure 28 shows the UV and LIF traces of afluorescein-labeled primer. While it appears in the UV trace that manyimpurities (failure sequences) are present, the LIF trace reveals that mostof these are not labeled and therefore do not fluoresce. The large (N-1)-meris due to a diasteriomer of the fluorescein phosphoramidite used in thelabeling procedure (Ulfelder, 1994).

56

10 12 14 16 18 20 22 24 26 28 30

Time (min)

A

0.030 AU/FS

B

150 RFU/FS

Figure 28. CE separation of fluorescein-labeled primer, a 26-mer. Detec-tion of UV (A) and LIF (B). From Ulfelder, K. J., Beckman ApplicationInformation Bulletin A-1774 (1994).

In DNA sequencing, fluorescently labeled primers based on fluores-cein, modified fluoresceins, Texas Red, and tetramethylrhodamine areroutinely used (Dovichi, 1994). The fluorescein dyes match the 488 nmline of an Ar-ion laser, whereas the rhodamine dyes are compatible withthe 543.5 nm line of an He-Ne laser or the 514.5 nm line of an Ar-ion la-ser. For the labeling of nucleotides, fluors such as dansyl (Lee et al., 1991)and fluorescein (Li et al., 1993) have been used. Wang and Giese (1993)recently described a phosphate-specific labeling of nucleotides with thefluor BODIPY Fl C3 hydrazide (Molecular Probes, Eugene, OR). Thelabeled nucleotides were subsequently detected by CE-LIF. Adenine-con-taining nucleotides were selectively analyzed by Tseng et al. (1994) usinga chloroacetaldehyde derivatization reaction to convert the analytes tofluorescent products.

57

4.4 Fraction Collection: CE as aMicropreparative Tool

CE is not only an analytical technique but can also be used for micro-preparative purposes. For example, it is possible to collect fractions fromprotein digests and subsequently perform microsequencing to identify thepeptides. Enzyme activity in fractions of a fermentation broth can be deter-mined using CZE in a micropreparative mode (Banke, 1991). In anotherDNA-related application, a small quantity (less than a µg) of a 20-meroligonucleotide primer was collected by using micropreparative CGE; thecollected fraction from a crosslinked gel (8%T, 3.3%C) was subsequentlyused in a dot-blot assay (Cohen et al., 1988). More recently, CGE (replace-able gel, 4% linear PA) was used to collect multiple peaks correspondingto denatured DNA from a mutated, 372 bp PCR product. The collectedfractions were re-amplified by PCR and subsequently analyzed again byCGE. It was found that the different peaks corresponded to different genesequences (Kuypers et al., 1993—see also Section 5.4.2.1).

4.4.1 Fraction Collection Using Field Programming

With automated instruments such as P/ACE, during a micropreparativerun, the outlet of the capillary is switched from the buffer vial to a collec-tion vial which contains a small amount (≈ 2 to 10 µL) of water or dilutebuffer. (Note: relatively large-i.d. capillaries are beneficial in micro-preparative CE as more material can be collected: the loadability of capil-laries is proportional to their cross-sectional area). In the above-mentionedapplications, the electric field was kept constant during the collection ofthe fractions. However, with CGE of DNA, often very narrow peakwidths—a few seconds wide—are obtained. This makes reproducible col-lection of peaks difficult. By programming the electric field during thecollection (“slowing the field down”), the collection process can be simpli-fied (Guttman et al., 1990). This approach of field programming for frac-tion collection is demonstrated in Figure 29.

58

0.005 AU

0.001 AU

A

B

C

19 23

19 23

19 23

0.001 AU

20

20

47

47

Time (min)

Figure 29. (A) Micropreparative CGE separation of a polydeoxyadenylicacid test mixture, p(dA)40–60; (B) analytical run of the isolated p(dA)47spiked with p(dA)20; and (C) analytical run of p(dA)40–60 spiked withp(dA)20. Reprinted with permission from Guttman et al., Anal. Chem. 62,137 (1990). Copyright, American Chemical Society.

The goal of the experiment was to collect the 47-mer from a p(dA)40-60mixture. During the micropreparative run (trace A), the field was maintainedat 300 V/cm until just before the 47-mer reached the end of the capillary.At that point, the field was decreased 10-fold and the fraction was collectedfor 60 s (the calculated peak width under the low-field conditions was 45 s).Trace B shows the reinjected collected fraction: only the 47-mer is visibletogether with the internal standard, a 20-mer. Typically, the micropreparativeruns yield broader peaks than their corresponding analytical runs. The samplesize injected for the micropreparative run resulting in the “overloaded” pro-file (trace A) was 6 times higher than the analytical run (trace C).

59

5 Selected Applications

5.1 Nucleotides, Nucleosides and Bases with CZEor MECC

Nucleotides and their nucleoside and base constituents play an importantrole in many vital biochemical processes. They are the activated precursorsof DNA and RNA. Intracellular nucleoside metabolism is an importanttopic in AIDS research. A number of dideoxynucleoside analog drugs(azT, ddI and ddC) are currently used in the treatment of HIV-1-positiveindividuals or are in human clinical trials (d4T and ddA). The nucleoside“pro-drugs” are converted by intracellular host enzymes (kinases) into theirtriphosphates; the latter inhibit the viral reverse transcriptase enzyme.DNA “damage” at the nucleic acid base level is also actively studied (Ca-det and Weinfeld, 1993). Reliable, high-resolution analytical methods toquantitate nucleotides, nucleosides, and bases in biological samples (oftenat extremely low levels) are highly desirable in these areas. CE has beendemonstrated to be a promising new tool in a number of recent studies. CEmethods may complement existing HPLC methods; however, often supe-rior resolution with shorter analysis time is possible with CE, while a mini-mal amount of sample is required for the process.

5.1.1 DNA Adducts; DNA Damage

DNA damaged by covalent modifications or additions of xenobioticcomponents was studied by Norwood et al. (1993) and Guarnieri et al.(1994). The first group evaluated different CZE and MECC conditions anddemonstrated sample stacking techniques to increase detectability of ben-zo(a)pyrene-DNA adducts. Guarnieri et al. (1994) measured 8-hydroxy-deoxyguanosine by MECC as a marker of DNA oxidation. Single-strandedDNA was incubated in the presence of an oxidizing agent and hydrolyzedby enzymatic digestion. MECC did not, however, permit determination ofextremely low levels of oxidized nucleosides generated by endogenoussources of free radicals (see also Section 5.1.4).

5.1.2 Nucleoside Analog Drugs

Therapeutic drug monitoring of a nucleoside drug was described by Lloydet al. (1991). The antileukemic agent, cytosine-β-D-arabinoside (Ara-C),was determined in human serum. The authors found low-level (i.e., sub-µM)detection of Ara-C problematic. However, by using solid-phase extraction

60

for concentration and sample cleanup, it was possible to determine Ara-C inthe 1 to 10µM range. This procedure removes most of the protein and al-lows doubling of the Ara-C concentration. The assay was validated over aconcentration range of 1 to 10 µM. Response was linear in this range, with acorrelation coefficient of 0.996 for the calibration plot. Compared to HPLC,the proposed assay has a rapid analysis time (no need to run a gradient toremove late eluting compounds) and is free from endogenous substances.

Rogan et al. (1993) showed that CE is well suited to resolve enanti-omers of nucleoside analog drugs. Determining the enantiomeric ratio ofsuch drugs is important to regulatory agencies such as the FDA becauseone enantiomer may exhibit far greater efficacy (or toxicity) than the other.For example, in the manufacturing process of an antiviral drug (Glaxo), theracemic 2'-deoxy-3'-thiacytidine (BCH-189) undergoes an enantiospecificdeamidation to yield the (-) enantiomer. The latter chiral drug is less toxicand therefore preferred for clinical use over the racemic drug. The timecourse of the enzymatic reaction was followed by CE for more than twodays (Figure 30). After 40 hours, less than 1% of the (+) BCH-189 re-mains. Dimethyl-β-cyclodextrin was added to the sodium phosphate, pH2.3, run buffer as the chiral discriminator; baseline resolution was obtainedbetween the (+) and the (-) enantiomers. Compared to HPLC methodswhich typically involve expensive chiral stationary phases, the CE ap-proach is simpler, and reliability and precision reportedly are excellent (seethe Beckman primer on chiral analysis, P/N 726388).

24 hours

45 hours

51 hours

1 hour

0

4 hours

10 20 30Time (min)

Figure 30. Time dependence of percentage of (+) enantiomer remainingduring a biotransformation reaction time course. The two peaks representthe enantiomers of the nucleoside analog drug BCH-189. Electrophero-gram courtesy of Dr. K. Altria, Glaxo Group Research, U.K.

61

5.1.3 Nucleotides in Cell Extracts

Nguyen et al. (1990) used a CZE method for the quantitation of nucleotidedegradation in fish tissues. In most fish, ATP degrades rapidly to IMPwhich, in turn, degrades to inosine and the base hypoxanthine. IMP givesfish a pleasant fresh taste while hypoxanthine accumulation results in an“off” taste. Tissue extracts were assayed by both CZE and an enzymaticmethod. Good correlation between peak area and nucleotide concentrationwas found. The CZE method involved UV detection, untreated, fused-silica capillaries, and a CAPS, pH 11, run buffer generating a high EOF.

Nucleotide profiles in cell extracts were determined by Ng et al.(1992) and Huang et al. (1990) with CZE. Ng et al. (1992) showed nucle-otide profiles in human blood lymphocytes and leukemic cells. Reproduc-ible areas and migration times were obtained using a P/ACE instrument.A simple CZE system using an untreated, fused-silica capillary with aborate, pH 9.4, run buffer was employed. The negatively charged nucle-otides are carried to the detector (cathode) by the EOF. Fourteen of thecommon ribonucleotides were determined in a CZE assay by Huang et al.(1992). In this method, coated (polyacrylamide) capillaries resulted innegligible EOF using the mixed phosphate-Tris, pH 5.3, buffer. Electro-phoretic flow carries the analytes to the detector end (anode, reversed po-larity). Minimum detectable levels of the nucleotides were in the 1 to10 µM range with UV detection at 254 nm. The method was applied to thequantitation of ribonucleotides in HeLa cells. In a later paper (Shao et al.,1994), Ucon-coated capillaries were used for the determination of ribo-nucleotides in lymphoma cells. A similar method, involving (polyacryla-mide) coated capillaries, was described by Takigu and Schneider (1991).The authors discussed validation criteria, i.e., linearity and minimal detect-able concentration (≈ 1 µg/mL per nucleotide without stacking).Beckman’s Neutral coated capillary (P/N 477441) is also suitable for thesetypes of applications.

MECC conditions have also been proposed for the analysis of nucle-otides in cell extracts. It appears that cationic surfactants are more effectivethan their anionic counterparts such as SDS. Perret and Ross (1991) se-lected dodecyltrimethylammonium bromide (DTAB), resulting in a chargereversal on the capillary wall as shown earlier in Figure 9. They also used1 mM EDTA in the run buffer as a metal chelating agent to prevent metal-nucleotide interaction which was thought to result in peak tailing. Fig-ure 31 shows that this method can be applied to acid extracts of cells. Theupper trace shows the separation of a standard mixture of 15 nucleotides

62

with a 50 mM phosphate (pH 7), 100 mM DTAB, 1 mM EDTA run buffer.A neutralized perchloric acid extract of rat tumor is shown in the bottomtrace of Figure 31. The reported method gave a linear response up to200µM; migration time and peak area precision ranged from 2.2 to 5.5%and 3.3 to 6.1%, respectively. Similarly, Ramsey et al. (1994) evaluated anumber of cationic surfactants for the MECC of nucleic acid constituents.Here, optimum resolution was achieved by using tetradecyltrimethyl-ammonium bromide (TTAB) as the micellar reagent. Loregian et al. (1994)compared MECC with HPLC for the quantitation of ribonucleotide triphos-phates in four different cell lines. The MECC method yielded approxi-mately one million theoretical plates with detectability down to 50 fmol.

0.0100

0.0078

0.0056

0.0034

0.0012

-0.00102.00 3.60 5.20 6.80 8.40 10.00

Time (min)

Abs

orba

nce

Abs

orba

nce

0.0010

0.0007

0.0004

0.0001

-0.0002

-0.00051.50 2.80 4.10 5.40 6.70 8.00

Time (min)

A

B

CM

PU

MP

GM

PN

AD

AM

PdA

MP

NA

DP

UD

P

AD

P

UT

P

GT

P

AT

P

cXM

P

GD

P

IMP

NA

DA

MP

AD

P

AT

P

Figure 31. Separation of nucleotides by MECC. (A) Separation of stan-dard mixture of 15 nucleotides at 25 mM each (B) neutralized perchloricextract of rat tumor. Adapted with permission from Perrett, Capillary Elec-trophoresis (Camillieri, Ed.), CRC Press, 1993.

63

5.1.4 Increasing Detectability: LIF Detection

There is increasing interest to measure nucleotides at extremely low levelsin biological matrices or even in single cells. DNA base damage studies(by either HPLC or CE) are often hampered by a lack of sensitivity, espe-cially when only limited sample volume is available. A sensitive assayshould, for example, be capable of detecting at least one DNA base modifi-cation in 104–106 normal bases within a few micrograms of DNA (Cadetand Weinfeld, 1993). CE with LIF detection is, in principle, sensitiveenough to measure these types of modifications. Combined with suitablefluorophore chemistry, LIF detection can, in principle, provide orders ofmagnitude improvement in sensitivity over UV detection. Preliminaryreports (Wang and Giese, 1993; Lee et al., 1991) on nucleotide analysis byCE-LIF support this contention.

5.2 Purity Control of Synthetic OligonucleotidesWhereas the purity requirements may not be that important in applicationswhere the oligonucleotide is used as a hybridization probe, in many otherapplications (including antisense DNA), the purity of the synthetic oligo-nucleotide needs to be ascertained. CGE is ideally suited for this purpose:the technique is fast, reproducible, has high resolving power, and does notinvolve radioactivity or toxic materials.

5.2.1 Phosphodiester Oligonucleotides

For the CE of oligonucleotides with 10 to 150 bases (primers or probes),crosslinked or linear polyacrylamide gels covalently bonded to coated,fused-silica capillaries are used under denaturing conditions (urea,formamide, heat). The prepacked, 100 µm-i.d. capillary available fromBeckman (eCAP ssDNA 100) contains 7 M urea to prevent the formationof secondary structure of oligonucleotides. The capillary is designed foroligonucleotides in the 10 to 150 base range and can be used in the 20 to50°C temperature range. An example illustrating the utility of this columnfor the QC of synthetic oligonucleotides was shown earlier in Figure 28.

The separation of a 119-mer oligonucleotide preparation on the eCAPssDNA 100 column is shown in Figure 32. The upper trace shows thecrude preparation; the lower trace shows the purified oligonucleotide.Separations of this type are difficult to perform with HPLC which gener-ally is limited to an upper range of approximately 70 bases (Warren andVella, 1993). In addition to the main component, the failure sequences arewell resolved by the CGE capillary (N = 565,000 plates per m).

64

46403428221610-0.001

0

0.001

0.002

0.003

Abs

orba

nce

254

nm

Time (min)

Crude

Purified

119-mer

Figure 32. Analysis of a crude and purified synthetic 119-mer oligonucle-otide using the eCAP ssDNA 100 Kit.

5.2.2 Antisense DNA

Antisense therapeutics are synthetic oligonucleotides that have a base se-quence which is complementary to a target sequence on a messenger RNA(mRNA) which encodes for disease-causing proteins or to the double-stranded DNA from which the mRNA was transcribed. The complemen-tary nature of the antisense molecule allows it to hydrogen bond andinactivate the genetic message, inhibiting gene expression. “Normal” oli-gonucleotides with a phosphodiester backbone are very susceptible tocellular nuclease degradation. There is, therefore, much interest in DNAanalogs with phosphorus-modified backbones (e.g., phosphorothioates andmethylphosphonates, Figure 33) which exhibit increased resistance to thesenucleases. Another type of antisense, termed peptide nucleic acid (PNA),in which the deoxyribose-phosphate backbone is substituted for a peptidebackbone composed of (2-aminoethyl)glycine units, shows promise as apotent therapeutic agent and can also be analyzed by CGE (Rose, 1993).Because of their potential use as drugs, stringent purity requirements aretypically required for antisense DNA agents.

65

O

OO

OO

OO

X P

Base

Base

Figure 33. Antisense DNA with phosphorus-modified backbone. X can beO (phosphodiester), S (phosphorothioate), CH3 (methylphosphonate), orOR (alkylphosphotriester). Base = adenine, guanine, cytosine, or thymine.

CGE with the eCAP ssDNA 100 column from Beckman results inexcellent resolution of deletion sequences of a 20-mer phosphorothioateantisense product. Figure 34 shows the electropherogram of a mixture ofthe n-mer and the (n-1)-mer. However, CGE cannot resolve phosphoro-thioates from their corresponding phosphodiesters (anion-exchange chroma-tography is the method of choice here) as the separation mechanism in CGEis based on molecular sieving (size). Using the eCAP ssDNA 100 Kit,Srivatsa, et al. (1994) demonstrated the validity of CGE for routine analysisof drug product formulations. The CGE method was found to be suitable forroutine drug product analysis, as judged by several criteria, i.e., linearity,accuracy, selectivity, precision, and ruggedness—see also Section 4.2.7.2.

66

0.03

0.022

0.014

0.006

-0.0020 8 12 16 20 24 284

Time (min)

Abs

orba

nce

254

nm

Figure 34. Purity analysis of a synthetic phosphorothioate 20-mer, using a47 cm eCAP ssDNA 100 capillary.

5.3 DNA SequencingSeveral research groups are currently exploring the use of CE as an alterna-tive to slab-gel electrophoresis for automated DNA sequence determination(Pentoney et al., 1992; Ruiz-Martinez et al., 1993; Dovichi, 1994). Thelarge surface-area-to-volume ratio of the capillary permits higher electricfields than are used typically with slab gels (due to more efficient heatdissipation), resulting in very rapid and efficient separation of sequencingreaction products. Additionally, the capillary format is readily adaptable toautomated sample loading and on-line data collection. With CE, detectionof separated DNA sequencing fragments is performed by LIF. The sensi-tivity of the LIF detection allows sequencing reactions to be performed onthe same template and reagent scale as that of manual DNA sequencingwith autoradiographic detection. The identity of the terminal base of eachDNA sequencing fragment is encoded in the wavelength and/or the inten-sity of the fluorescent emission.

Sample throughput is a major concern for high-volume sequencingapplications. Automation of sample preparation, sequence reactions (in-cluding electrophoresis), and data interpretation are necessary in order to

67

achieve the ambitious goal of sequencing the entire human genome (ap-proximately 3 billion bp). With CE, the samples are loaded one at a time.Slab gels, on the other hand, can be simultaneously loaded with 24 to 36samples. Instrumentation which would allow the running of several capil-laries in parallel, together with robotics for sample handling, would dra-matically increase the desired sample throughput with CE. Already DNAsequencing has been demonstrated in arrays of multiple (20 to 100) capil-laries (Mathies and Huang, 1992; Ueno and Yeung, 1994; Takahashi et al.,1994). It can be easily envisioned that this type of instrumentation can alsobe incorporated in other applications, e.g., screening for genetic diseases,forensic DNA typing, etc.

A large obstacle to the development of commercial CE-based DNAsequencers has been the stability of gel-filled capillaries. While they canprovide extremely high resolving power, the crosslinked gels typically lastonly a few runs when sequencing reactions are loaded, after which time theentire column must be replaced. Recent developments in CGE columntechnology (in particular, the replaceable gels) should eliminate the time-consuming and laborious procedures of the preparation and alignment ofthe capillaries (Ruiz-Martinez et al., 1993). With the replaceable matrix, itis possible to load a sequencing reaction, rapidly separate the DNA frag-ments at high field strength, and then reload the gel on the capillary priorto the next run. Figure 35 shows the CE separation of a single terminator,Sanger-Coulson reaction using a replaceable linear PA gel. A “red” diodelaser was used for excitation of the fluor- (Cy5) labeled DNA fragments(Chen et al., 1993). The relatively low-viscosity (6%T) gel matrix of thesetypes of capillaries provides reproducible and fast separation of DNA frag-ments with sequence information extending to at least 400 bases. For DNAsequencing applications, typically a CE run buffer containing formamideand/or urea is used. A denaturing buffer of 30% formamide, 3.5 M urea hasa lower viscosity than a 7 M urea buffer and is therefore advantageous touse in a replaceable CGE formulation. In addition, increased decompres-sion of sequences with secondary structures is obtained.

68

9785

6958

363

248

4331

99

Tim

e (m

in)

LIF Signal

Fig

ure

35. E

lect

rop

he

rog

ram

of

fluo

resc

en

tly la

be

led

(C

y5-2

0 p

rim

er)

DN

A f

rag

me

nts

ge

ne

rate

d e

nzy

ma

tica

lly (

M1

3m

p1

8te

mp

late

) u

sin

g d

dG

te

rmin

ato

r. A

La

serM

ax

(Ro

che

ste

r, N

Y)

red

dio

de

lase

r (6

39

nm

exc

itatio

n)

wa

s e

mp

loye

d w

ith a

n in

-h

ou

se-b

uilt

CE

-LIF

inst

rum

en

t. T

he

ca

pill

ary

wa

s fil

led

with

a r

ep

lace

ab

le p

oly

acr

yla

mid

e g

el.

Th

e p

att

ern

is r

eco

gn

iza

ble

be

yon

d 4

00

ba

ses.

Ele

ctro

ph

ero

gra

m c

ou

rte

sy o

f D

r. S

. L

. P

en

ton

ey,

Jr.

, B

eck

ma

n I

nst

rum

en

ts (

19

94

).

69

5.4 dsDNA, PCR Products Analysis (< 2000 bp)With the advent of PCR-related methods, the number of publications in-volving DNA fragment separations by CE is rapidly expanding. Since itsintroduction in 1986, PCR technology is now used in a variety of diversefields, e.g., medicine, biology, forensics, epidemiology, archeology, andnanotechnology, to name a few. Electrophoresis is almost invariably partof this research, i.e., as a tool to visualize DNA fragments as characteristicbanding patterns which could reveal a disease gene in a patient, identify asuspect in a murder case, or establish part of the DNA sequence of a mil-lion-year-old fossil. In the years to come, CE, and especially CE combinedwith LIF, will undoubtedly replace classical electrophoretic techniques inmany PCR-related applications.

The first researchers to investigate the utility of CE for dsDNA analy-sis were Brownlee, and co-workers at Microphoretic Systems (1988) andCohen, Karger, and co-workers at Northeastern University (1988). From1988 to 1991, emphasis was on the development of suitable gels/polymernetworks and capillary coatings, and the fine-tuning of CE conditions tooptimize separation performance (e.g., Zhu et al., 1989; Heiger et al.,1990; Guttman and Cooke, 1991 A; Schwartz et al., 1991). Applications,e.g., the screening of blood for HIV-1, were soon reported (Sunzeri et al.,Brownlee et al., 1990). It also became apparent that detection by UV ab-sorbance was not always sufficiently sensitive, e.g., in the detection of lowcopy DNA for viral screening (Mayer et al. , 1991). Presently, the CEtechnology has advanced in this area and instrumentation (including high-sensitivity LIF detection) is commercially available. In the following sec-tions, some selected applications dealing with the analysis of PCR productsare highlighted. Arbitrarily, we have divided dsDNA analysis into twosections with the bp number smaller or larger than 2000.

70

5.4.1 Quantitation of Viral Load in Infectious Diseases

Quantitation of viral load in patient specimens is important to assess thestage of disease progression or to monitor the effectiveness of drugtherapy. HIV-1 infection is a good case in point. Until recently, there wereno reliable methods available to quantitate HIV-1 in the early-onset phaseof infection, i.e., at low copy numbers. Novel PCR methods (e.g., competi-tive PCR) have recently allowed quantitation of proviral DNA or plasmaviral RNA levels in patients with HIV-1 infection (Piatak et al., 1993)using gel electrophoresis with video scanning of the ethidium bromide-stained DNA. CE is ideally suited to replace slab gel techniques for thesepurposes. Competitive PCR is also adaptable to a CE format, as will bediscussed in the next section.

Brownlee, Sunzeri, Mayer, and co-workers (1990, 1991) developedPCR-based CE methods for the quantitation of multiple retroviral DNAsequences. An instrument was employed which permitted simultaneousUV and fluorescence (non-LIF) diode-array detection (Schwartz et al.,1989). The sensitivity by UV or (non-LIF) fluorescence was not goodenough, however, to allow detection of HIV-1 provirus at very low DNAcopy numbers (e.g., for HIV proviral load in asymptomatic individuals).At that time (1991), LIF detection would have provided the extra sensitiv-ity needed but was not yet commercially available. The large increase insensitivity using an intercalating dye (thiazole orange) and LIF detection(Ar-ion laser) was demonstrated by Schwartz and Ulfelder (1992). SeveraldsDNA fragments (242, 368, and 900 bp) were detected with ≈ 400 Xbetter sensitivity than possible with UV detection.

Recently, a French research team, Lu et al. (1994), used P/ACEwith LIF detection for the quantitative analysis of PCR-amplified HIV-1DNA or cDNA fragments. The LIFluor dsDNA 1000 Kit (containing theEnhanCE intercalating dye) was used in the CE-LIF experiments. Quanti-tation of multi-target PCR fragments was demonstrated. Figure 36 showsthe CE-LIF electropherogram of three HIV-1 sequences (142 bp, 394 bp,and 442 bp from the gag, pol, and gp41 genes, respectively) together witha DNA standard. Figure 37 shows a dilution series of HIV-1 DNA tem-plates ranging from 1 to 25,000 copies subjected to 40 PCR cycles. A lin-ear range of three orders of magnitude was achieved using CE-LIF. Thefigure also shows a dilution series obtained from reverse-transcribed (RT)RNA from HIV-1 virions. Data of virion concentrations in sera of indi-viduals infected with HIV-1 at different stages of infection were presented.The measurements by CE-LIF showed excellent correlation with the dataacquired with the Southern blot hybridization method.

71

20.0

10.0

0.0

12.0 15.0 20.0 24.0Time (min)

RF

U

72 b

p

118

bp

142

bp (

HIV

-1 g

ag)

194

bp

234

bp27

1 bp

281

bp30

1 bp

394

bp (

HIV

-1 p

ol)

442

bp (

HIV

-1 g

p41)

603

bp

872

bp10

78 b

p13

53 b

p

Figure 36. Detection of multi-target PCR-amplified HIV-1 gag, pol, envsequences by CE-LIF using the LIFluor dsDNA 1000 Kit. A φX 174 RFDNA standard was co-injected with the sample. Reproduced with permis-sion from Lu et al., Nature (London) 268, 269 (1994).

103

102

101

100

10-1

1 10 100 1000 10000 100000

Number of HIV DNA copiesNumber of HIV RNA virions

RF

U

Figure 37. Linearity of LIF-CGE analysis of quantitative PCR or RT-PCRproducts. Serial dilutions of HIV-1 DNA templates ranging from 1 to 25,000copies were subjected to a 40-cycle PCR with gag primers SK145/431. RNAextracted from serial dilutions of HIV-1 virions (ranging from 10 to 100,000viral particles) was reverse-transcribed with 20 pmol of 3' primer SK431.Reproduced with permission from Lu et al., Nature (London) 368, 269 (1994).

72

Separations of RT-PCR products from the RNA of polio virus wereshown by Rossomando, White, and Ulfelder (1994). Quantitation wasachieved by comparing the corrected peak area for the RT-PCR product toa standard curve generated from known amounts of template RNA. TheBeckman sieving buffer containing the intercalating dye, EnhanCE, facili-tated excellent separation of 53 bp, 71 bp, 97 bp, and 163 bp DNA frag-ments. The resolution by slab gel electrophoresis, on the other hand, whileadequate for the 163 and 97 bp fragments, was inadequate for the otherDNA fragments. Figure 38 compares UV (260 nm) vs. LIF detection for a53 bp RT-PCR product derived from the Sabin 3 strain of virus from thepolio vaccine. Note that the migration times of the PCR products arelonger for the LIF run as, in this case, the DNA is intercalated with oppo-sitely charged dye. The other important feature to note in Figure 38 isthat—in the LIF trace—interferences are far less prominent and the patternis unambiguous. PCR of target sequences often results in contaminatingby-products which can interfere with UV detection.

Sabin 3

Sabin 3

72 bp

72 bp

UV

LIF

8 9 10 11 12 13 14 15 16 17 18

Time (min)

Figure 38. UV absorbance versus LIF detection of the separation of a53 bp RT-PCR product from the RNA of the Sabin 3 strain of the poliovirus vaccine. AHaeIII-digested φX 174 DNA marker was co-injected withthe PCR product for size determination. The same Sabin 3 concentrationwas used for each analysis. The DNA marker concentration was 200 and10 µg/mL for the UV and LIF analysis, respectively. Reproduced withpermission from Schwartz et al., J. Cap. Elec. 1, 36 (1994). Copyright:ISC Technical Publications, Inc.

73

5.4.2 Competitive RNA-PCR by CE-LIF for Quantitationof Cellular mRNA

Accurate quantitation of PCR products—especially when dealing with lowcopy numbers—is often problematic. A group of methods, termed competi-tive PCR, has been described recently. These methods effectively deal withthe problem of accurate quantitation of PCR products (Gilliland et al.1990; Piatak et al., 1993). In competitive PCR, a known amount of stan-dard template DNA (the “competitor”) competes for the same primers withan unknown amount of target DNA (In the case of RNA-PCR, the DNA isobtained by reverse transcription). The competitor’s sequence is chosensuch that it is largely identical to the target sequence, except for the pres-ence of a mutated restriction site or a small intron. During the amplifica-tion cycles, the target and competitor are exposed to the same PCR-relatedreaction variables; their product ratio should, therefore, remain constant,even after the products have reached a plateau. The amount of target DNA(or RNA) can be obtained through a simple interpolation procedure of anexperimentally generated standard curve. A variation of this method,termed multiplex competitive PCR, involves co-amplification of the cDNAof a “housekeeping” gene (in addition to amplification of the target and itscompetitor) whose RNA does not vary among the different samples to beanalyzed. The expression of the target gene is then calculated in referenceto the housekeeping gene (Apostolakos et al., 1993).

Fasco et al. (1994) recently demonstrated that CE-LIF is an attractivealternative to slab gel techniques. DNA fragments or PCR products, inter-calated with the fluorescent dye YOYO-1 (an oxazole yellow dimer) canbe detected at extremely low levels in real time with high efficiency andprecision. In contrast to the slab-gel-based competitive PCR, the CEmethod is fast and can be fully automated; the computer-generated data arestored on disk. CE-LIF allows accurate and precise quantitation of PCRproducts formed during competitive PCR reactions. With the CE method ofFasco et al., excellent peak efficiency was obtained (approximately 10 bpresolution) and run times were less than 30 minutes. PCR product detect-ability with LIF is adequate for most clinical and diagnostic applications ofcompetitive PCR. The CE-LIF procedure was also applied to multiplexcompetitive PCR. YOYO-1 was used in the CE run buffer for intercalatingthe DNA fragments, as discussed in Section 4.3.1.2. The CE-LIF methodwas applied to reversed-transcribed RNA from glyceraldehyde-3-phos-phate dehydrogenase (GAPDH) and P4501A1 gene sequences.

74

A typical competitive RNA-PCR experiment is illustrated in Fig-ure 39. The formation of product (in pmol per 0.1 mL reaction mixture) isplotted vs. the initial RNA concentration. The initial competitor sequencewas kept constant at 0.1 amol per 0.1 mL reaction mixture. The inset ofFigure 39 shows log-log plot of the ratio of target to competitor product vs.the initial RNA concentration. The intercept at a ratio of 1 corresponds tothe initial competitor concentration, i.e., in this case 0.012µg per 0.1 mLreaction mixture that contains 0.1 amol of target mRNA (assuming 100%reverse transcription efficiency). Representative CE-LIF electrophero-grams for the competitive RNA-PCR experiments involving P4501A1 andGAPDH are shown in Figures 40 A and B, respectively.

0.01 0.10 1.000

2

4

6

8

10

10.00

1.00

0.10

0.01 0.10

12

RNA (µg/0.1 mL RNA-PCR REACTION MIXTURE)

PM

OL

PR

OD

UC

T /

0.1

mL

RN

A-P

CR

RE

AC

TIO

N M

IXT

UR

E

[GA

PD

H] /

[CO

MP

ET

ITO

R G

AP

DH

]

Figure 39. Formation of target (■) and competitor (▲) GAPDH duringcompetitive PCR. Equal concentrations were reverse transcribed, mixed ata 1:1 ratio, and diluted. The competitor concentration was kept constant at0.1 amol/0.1 mL reaction mixture. In the inset plot, the data points ob-tained at 0.25 and 0.5 µg RNA were omitted. Reproduced with permissionfrom Fasco et al., Anal. Biochem. (1995, in press).

75

0

20

40

60

80

100

0

20

40

60

80

100

120CF

CF

IS

IS

3

4

2

1

Rel

ativ

e F

luor

esce

nce

Uni

ts

Time (min)

A

B

5 10 15 20 25 30

Figure 40. (A) CE-LIF electropherogram generated during competitivePCR of P4501A1. 1 = competitor; 2 = target; IS = internal standard;CF = 5-carboxyfluorescein. (B) Competitive PCR of GAPDH. 3 = com-petitor; 4 = target. Reproduced with permission from Fasco et al., Anal.Biochem. (1995, in press).

5.4.3 Detection of DNA Polymorphisms and Mutations inGenetic Diseases

Genetic linkage studies follow the inheritance of a particular trait (pheno-type) in a family over several generations. The goal is to correlate the traitwith the presence of a specific DNA sequence (allele). By finding a markerthat is close (“linked”) to the gene, persons predisposed to certain diseasescan be identified. Often, years of painstaking work are required to identifythe exact location of the gene responsible for the disease on the chromo-some. For example, only recently the genes responsible for an inheritedform of breast cancer were discovered after 20 years of genetic linkage

76

studies (King et al., 1993). Ulfelder et al. (1992) demonstrated the utilityof CGE (with UV detection) for restriction fragment length polymorphism(RFLP) analysis of ERBB2 oncogene. This gene was one of the candidatesfor a breast cancer gene at chromosome 17q and was detected using a PCRmethod by Hall and King (1991). Polymorphic alleles can be identified bythe presence or absence of a specific endonuclease recognition site. RFLPstypically have only two alleles at a given locus. Figure 41 shows the RFLPanalysis of three individuals whose genomic, PCR-amplified DNA wasdigested with MboI. The top two traces represent two homozygous samplescharacterized by the presence of either the 500 bp or the 520 bp fragment.The heterozygous sample shows both fragments present which were sepa-rated in a sieving buffer containing 0.5% HPMC. For comparison, theagarose slab gel of similar samples is also shown in Figure 41.

In a similar study, Del Principe et al. (1993) analyzed PCR-amplifiedproducts of the DXS 164 locus in the dystrophin gene. XmnI digestionyielded a polymorphism generating fragments of 740, 520, and 220 bp.The same sieving buffer as that employed by Ulfelder et al. (1992) wasused. Another Italian research group (Gelfi et al., 1994 A) investigated an8 bp deletion linked to congenital adrenal hyperplasia. CE separations wereperformed with a sieving buffer consisting of 6% linear polyacrylamide.The amplified PCR products were a normal, 135 bp fragment and a dis-ease-linked, 127 bp fragment.

Sensitive LIF detection (P/ACE-LIF combined with a green He-Nelaser) using ethidium bromide (EB) was demonstrated by Liu et al. (1995).The PCR-amplified sequence of the Y-chromosome-specific ZFY gene(307 bp) was detected with much greater sensitivity than possible with EB-stained agarose gel electrophoresis.

DNA microsatellites or short tandem repeats (STRs) are increasinglyused as genetic marker systems in linkage studies. They are characterized bytandemly repeated, short (2 to 10 bases) sequences. A tetranucleotide repeatunit (GATT) linked to cystic fibrosis (CF) was studied by Gelfi et al.(1994 B). The allelic forms, a hexamer (111 bp) and a heptamer (115 bp),were amplified by PCR and separated by polyacrylamide gradient slab gelelectrophoresis and with CGE (6% linear PA). The hexamer was foundlinked to the CF-causing mutation in the gene. A sieving buffer consisting ofpolyacryloylaminoethoxyethanol was used to resolve PCR fragments in the450 to 550 bp range (Nesi et al., 1994). These fragments were derived fromtriplet (CAG) repeats in the androgen receptor gene. An increase in the num-ber of triplet repeats is linked to Kennedy’s disease, a neurological disorder.

77

500

500

520

520

550

550

550

330

330

330

220

220

220Homozygous AI

Homozygous A2

Heterozygous A1/A2

302824201612

Time (min)

-0.001

0.000

0.002

0.004

0.006

0.008

Abs

orba

nce

Figure 41. CE separation of PCR-amplified RFLP samples demonstratingMboI polymorphism. Top, homozygous for allele A1 (520 bp fragment);middle, homozygous for allele A2 (500 bp fragment); bottom, heterozygousfor A1 and A2. Constant fragments of 220, 330, and 550 bp can be seen inall three runs as a result of incomplete PvuII digestion of the 550 bp frag-ment. For comparison, an agarose gel is also shown. Lanes 1, 2, and 8:homozygous (A1). Lanes 3–6: heterozygous (A1 and A2). Lane 10: DNAsize markers. Reproduced with permission from Ulfelder et al., Anal. Bio-chem. 200, 260 (1992).

78

5.4.3.1 Point Mutations

A number of techniques have been described to detect single-point muta-tions in DNA (Perucho, 1994). PCR has a central role in the sample prepa-ration step in most of these techniques. Point mutation studies often requirethat the electrophoretic (or CGE) conditions are chosen such that ds andssDNA can be separated in one run. In denaturing gradient gel electro-phoresis (DGGE), the mobility of a partially melted DNA on the slab gel isreduced compared to an unmelted molecule. A variant of DGGE, termedconstant denaturant gel electrophoresis, was recently adapted in a CE for-mat (Khrapko et al., 1994). CGE was performed in capillaries containing apolymer network (6% linear PA) and a denaturant (3.3. M urea and 20%formamide). In a 10 cm portion of the capillary (the “denaturing zone”),the temperature was elevated; in the rest of the capillary, ambient tempera-ture conditions existed. Detection was by LIF. The critical role of tempera-ture is illustrated in Figure 42. Homo- and heteroduplexes from twofluorescein-labeled DNA fragments (206 bp) originating from human mito-chondria were resolvable by tuning the temperature. The two homodu-plexes differed by a single bp substitution (GC vs. AT). At 31°C, a singlepeak was obtained, indicating that all the duplexes were in the unmeltedform. As the temperature was raised, the duplexes started to separate in theorder of their melting stability. At 40°C, all the species eluted again in onesingle peak as their partially melted duplexes. The sensitivity of the CGEmethod is such that 105 mutant species can be detected among 3 × 108

wild-type sequences.

A similar type of point mutation method, called heteroduplex poly-morphism analysis (HPA), was proposed by Cheng et al. (1994). Duplexesand ssDNA were separated in a polymer network consisting of 0.5%HPMC and 4.8% glycerol. Ethidium bromide (3 µM) was added to in-crease resolution (see Section 4.2.4). The sensitivity of the CE method wasnot as sensitive as the one discussed above, as UV detection instead of LIFwas used. However, the authors contend that their detection system doesnot require the provision of natural or artificial GC-clamp domain. Thelatter are required to provide duplex stability in heteroduplex moleculesunder denaturing conditions.

79

40°C

38°C

36°C

35°C

31°C

GC+AT+GT+AC

GC

GC

AT

AT

GT

GT

GT

AC

AC

AC

GC+AT

GC+AT+GT+AC

0.1

0.0

0.1

0.00.5

0.3

0.1

0.6

0.3

0.01.6

0.8

0.015 20 25 30

Minutes

Vol

tsFigure 42. Constant denaturant CGE separation as a function of columntemperature. The sample, an equimolar mixture of two homoduplexes (GCand AT) and two heteroduplexes (GT and AC), was prepared using fluores-cein-labeled DNA fragments and run at the several temperatures indicated.Reproduced with permission from Khrapko et al., Nucl. Acids Res. 22, 364(1994).

Another technique for the screening of point mutations is single-strandconformation polymorphism analysis (SSCP). This technique, originallydeveloped by Orita et al. (1989), takes advantage of differences in mobili-ties between DNA fragments in non-denaturing gels. Point mutations in theDNA will cause conformational changes resulting in the mobility differ-ences of ssDNA (Guttman et al., 1992 A).

Kuypers et al. (1993) studied the p53 gene located on the short arm ofchromosome 17. CGE was run on control and patient (multiple myeloma)samples, using a 4% linear PA polymer network in the capillary. Denaturedsamples of normal DNA showed two peaks corresponding to the two ssspecies. The control cell line and patient samples revealed more compli-cated patterns of 3 to 5 peaks. Fraction collection by CGE (Section 4.4)was used to confirm the presence of different sequences in these peaks. Arecent paper from the same group (Kuypers et al.,, 1994) dealt with quanti-

80

tation of residual lymphoma cells carrying a translocation between chro-mosomes 14 and 18 in patient blood samples that were amplified by com-petitive PCR.

In the amplification refractory mutation system (ARMS) or allele-specific amplification (ASA), PCR is used to detect point mutations with-out requiring endonuclease digestion or Southern hybridization. The utilityof CE-LIF to detect mutations in the phenylketonuria gene by the ARMSmethod was demonstrated by Arakawa et al. (1994 A).

Another paper by the same group (Arakawa et al., 1994 B) describesthe utility of the CE-LIF method for diagnosis of medium-chain coenzymeA dehydrogenase (MCAD) deficiency, a disorder linked to sudden infantdeath and Rye-like syndromes. In most cases (90% of mutant alleles) theMCAD deficiency is caused by a single, A-to-G nucleotide change at posi-tion 985 in the gene. DNA fragments were amplified by two sets of allele-specific oligonucleotide primers. Mutant alleles yielded a single, 175 bpfragment, normal alleles yielded a 202 bp fragment, whereas heterozygouscarriers produced both fragments. The DNA fragments were well resolvedwithin a 12 minute run time on a capillary filled with low crosslinked PA.The CE-LIF method was linear over 3 orders of magnitude with a detectionlimit of ≈ 10 ng/mL for a 603 bp DNA fragment. Compared to UV detec-tion, LIF was 100 times more sensitive.

5.4.4 DNA Profiling in Forensic Work

CE-LIF has recently been applied in the analysis of genetic markers forhuman identification. Because often extremely low levels of substances areinvestigated, LIF should be the detection method of choice. Therefore,several researchers have studied different polymer matrix-fluorescent dyesystems to optimize separation efficiency and detectability. Variable num-ber tandem repeat (VNTR) analysis of the amplified D1S80 locus (300 to700 bp) with P/ACE-LIF is shown in Figure 43. D1S80 has a repeat unit of16 bp. The EnhanCE dye was used in the polymer-network-containing runbuffer. The figure shows the alleles for homozygous and heterozygousindividuals.

81

200

255

110

65

205 9 13 17 21 25

Time (min)

Flu

ores

cenc

e em

530

nm

FluoresceinT17

T22

T24

Figure 43. VNTR analysis of the amplified D1S80 locus for homozygousand heterozygous individuals using LIFluor dsDNA 1000 Kit and CE-LIFdetection. Courtesy of K. J. Ulfelder, Beckman Instruments, Inc.

Srinivasan et al. (1993) compared two asymmetric cyanine dyes,TOTO-1 and YOYO-1 (Molecular Probes, Eugene, OR), with the 488 nmAr-ion laser. Three genetic marker systems (apolipoprotein B, 700–1000bp with a 14 bp repeat; VNTR locus D1S80, 300 to 700 bp with a 16 bprepeat; and mitochondrial DNA, 130 to 140 bp with a 2 bp repeat), wereinvestigated for forensic applicability by PCR amplification. The PCRproducts were subsequently pre-stained with the fluorescent dye (DNA wasadded to dye at a molar ratio of 5:1 DNA bp to dye and incubated for 20minutes prior to analysis). Capillaries containing easy-to-use, replaceable,polymer network solutions (0.5% methylcellulose) were found superior tocrosslinked PA gels (3%C, 3%T).

McCord et al. (1993) analyzed some other genetic marker systemsof forensic interests, i.e., the human myelin-basic protein gene, the vonWillenbrand Factor gene, and the HUMTH01 gene located on chromo-some 11. PCR-amplified alleles resulting from VNTRs with 4 bp repeatunits in the 100 to 250 bp range were separated with replaceable, polymernetwork solutions. An asymmetrical dye, YO-PRO-1 (Molecular Probes,Eugene, OR) was added to the polymer solution of 1.0% hydroxy-

82

ethylcellulose and to each DNA sample for CE-LIF detection. Later studiesfrom this group (Butler et al., 1994 A and B) focused on quantitative as-pects of the CE-LIF method for forensic DNA typing. Comparisons withother existing methods (slab gel, slot blot, fluorescence spectrophotometry)were also made. With an internal standard, peak migration time precisionwas < 0.1% RSD. Peak area precision—using pressure injection—was≈ 3% RSD (see also Section 4.2.7, Table 4). A single-step voltage gradientallowed shorter run times for the HUMTH01 allelic ladder analysis(< 10 min) while still maintaining ≈ 3 bp resolution in the region of interest(Butler et al., 1994 B).

5.4.5 DNA Profiling of Plants, Bacteria and Fungi

Identification of bacteria in clinical samples using standard culturing tech-niques is both time consuming and cumbersome since the bacteria must begrown in the laboratory and identified on the basis of nutritional develop-ment requirements. In addition, many bacteria are morphologically similarto one another. Avaniss-Aghajani et al. (1994) have developed a methodfor the identification of various bacterial species using PCR amplificationof small subunit ribosomal RNA genes, which vary in sequence amongbacterial species. PCR was accomplished using one set of primers, one ofwhich was 5'-labeled with fluorescein isothiocyanate (FITC). Subsequentfluorescently labeled PCR products were then subjected to digestion withrestriction endonucleases, producing fragments of different length due tovariations in sequence among the bacterial species. When analyzed by CE-LIF (Ar-ion laser at 488 nm), only the DNA fragments containing the ter-minal 5'-FITC label were detected. The length of the labeled restrictionfragment was then used to identify a particular bacterial species. Figure 44shows the CGE analysis of PCR products for four bacterial species(Escherichia coli, Flavobacterium okeanokoites, Klebsiella pneumoneae,and Streptococcus faecalis) after digestion with endonucleases MspI andRsaI. Use of this process will result in clear bacterial identification withmajor time savings.

Size-selective DNA profiling and RFLP analysis of amplified poly-morphic spacers originating from fungus rDNA was performed by Martinet al. (1993) using the CGE conditions of Schwartz et al. (1991). Inter- andintraspecific variation in the size and number of restriction sites of theamplified rDNA spacers from several fungi were examined, allowing thestrains to be genotyped by CGE (UV detection).

83

1.000

0.500

0.000

1.200

0.00 2.00 4.00 6.00 8.00

Time (min)

Cha

nnel

A: F

luor

esce

nce

1

2

3

4

5

6

Figure 44. CE-LIF electropherogram of PCR-amplified SSU rRNA genesof Flavobacterium okeanokoites, Escherichia coli, Streptococcus faecalis,and Klebsiella pneumoneae after digestion with MspI and RsaI. Numberedpeaks correspond to the 5' terminal restriction fragments of the digestedPCR products. Peak 1: MspI digest from F. okeanokoites. Peak 2: RsaIdigest from E. coli. Peak 3: MspI digest from S. faecalis. Peak 4: MspIdigest from K. pneumoneae. Electropherogram courtesy of E. Avaniss-Aghajani, UCLA.

Marino et al. (1994) showed results of soybean genotyping using CGEwith crosslinked gels. Polymorphism was detected in dinucleotide, shorttandem repeat sequences (STRs). Figure 45 shows the allelic STR profilesof two genotypes (Williams and Jackson), together with the F1 progeny.The latter is heterozygous, containing both of the parent alleles. In theelectropherogram, the alleles are located between the molecular massmarkers of 100 bp (29 minutes) and 200 bp (36 minutes).

84

4439342924 49

Williams Soybean

Jackson Soybean

F1 J & W Soybean

Time (min)

UV

Abs

orba

nce

Figure 45. Analysis of the parental genotypes Jackson and Williams andtheir F1 progeny. The molecular mass markers 100 and 200 base pairs areat 29 and 36 minutes, respectively. The SSR-containing fragments for thegenotypes are found between 30 and 33 minutes. F1 generation is heterozy-gous. Reproduced with permission from Marino et al., J. Chromatogr. A676, 185 (1994).

5.4.6 Plasmid Mapping

Restriction enzyme digestion of plasmids (“plasmid mapping”) is oftenused for confirmation of PCR products, in cloning experiments, and inbiotechnology process control, e.g., to monitor genetic stability. Maschkeet al. (1993) employed CGE (UV detection) with replaceable, 6% linearPA gels for the mapping of four closely related plasmids. Various high-resolution plasmid maps were shown, obtained with a number of differentrestriction enzymes. The number of theoretical plates (based on a 242 bpfragment) was calculated as ≈ 3 million per m. In particular, the smallerDNA fragments are better resolved by CGE than with agarose gel electro-phoresis. Compared to the slab gels, CGE for plasmid mapping applica-tions is quantitative, fast (run times < 20 min), and consumes minimalamounts of sample for analysis. Comparisons between CGE (6% linearPA) and agarose slab gels for plasmid restriction digests were also pub-

85

lished by Paulus and Husken (1993). In the analysis of plasmids, the poly-mer network concentration and/or chain length must be optimized to accom-modate the larger sizes of the DNA fragments to be separated (see Section5.5, below).

5.5 dsDNA (2 to 20 kbp) by CGEAs in slab gel electrophoresis, the size range to be separated in CGE can beextended by diluting the polymer concentration of the sieving matrix.Some examples of this were shown earlier for alkylcellulose polymer solu-tions (Figures 17 and 18). With linear-PA-based gels, a 3% polymer net-work was found suitable to separate efficiently a 1 kbp DNA ladder (range:72 to 12216 bp), with some peaks exhibiting in excess of 4 million platesper m (Pariat et al., 1993). Similar capillary separations were obtained byStrege and Lagu (1991) and Baba et al. (1993) on alkylcellulose polymernetworks, by Bocek and Chrambach (1992) on 2% SeaPrep agarose, andby Chiari et al. (1994) on a polyacryloylaminoethoxyethanol replaceablegel. Pariat et al. (1994) showed that low concentrations of linear PA (1.5%)can be used to further extend the size range, e.g., for the separation ofλDNA HindIII restriction fragments. This digest contains fragments rang-ing from 125 bp to 23.1 kbp (Figure 46). Others (Barron et al., 1994; Kimand Morris, 1994 A) have recently demonstrated that ultradilute solutions(≈ 0.01%) of alkylcellulose also can be used for these type of separations.

10 10 10

Time (min)

1

2

3

4

5

6

7

Figure 46. Electropherogram of λHind III DNA using 1.5% linear poly-acrylamide. Peak identification: 1 = 564 bp; 2 = 2.0 kbp; 3 = 2.3 kbp;4 = 4.4 kbp; 5 = 6.6 kbp; 6 = 9.4 kbp; 7 = 23.1 kbp. Reproduced withpermission from Pariat et al., J. Chromatogr. A 652, 57 (1993).

86

5.5.1 Quantitation of Plasmid Copy Number

In recombinant DNA technology, plasmid analysis is often used to controlthe genetic stability during fed-batch culture. Generally, the size range forplasmid analysis is larger than for PCR product analysis, i.e., 2000 to22,000 bp. E. coli is the most frequently used host cell in the production ofrecombinant molecules. However, host cell-plasmid systems have limitedgenetic stability. During cell cultivation, therefore, the plasmid concentra-tion (copy number) may decrease. Hebenbrock et al. (1993) have shownthat CGE is an excellent quantitative tool to monitor the plasmid concen-tration during the cultivation of the E. coli strain containing the plasmid.The plasmid DNA concentration was estimated from the integrated peakareas of an internal standard (4,363 bp) and the plasmid carrying the ge-netic information (a linearized, 13,000 bp restriction fragment). The CGEmethod used 4% linear PA polymer networks for separation of DNA frag-ments ranging from 3,000 to 22,000 bp.

5.6 Very Large Chromosomal DNA (> 20 kbp)Preliminary results of separations of very large DNA fragments wereshown by Chrambach’s group (Guszczynski et al. ,1993). Linear-PA-filledcapillaries (0.1 to 0.9%) were used for three DNA-containing samples inthe size range of 20 to 50 kbp, multiples of 50 kbp, and 3 to 6 Mb, respec-tively. However, as noted by the authors, various problems related to thesampling and the CE instrumentation render their CGE method as yet im-practical for large DNA. Pulsed-field CE, in conjunction with dilute poly-mer solutions (see Section 3.2.5), may, in fact, be a more promisingapproach as recent publications indicate (Sudor and Novotny, 1994; Kimand Morris, 1994).

87

6 ReferencesAbersold, R., Morrison, H. D. J. Chromatogr. 516, 79 (1990)

Albin, M., Grossman, P. D., Moring, S. E. Anal. Chem. 65, 489A (1993)

Andrews, A. T. Electrophoresis: Theory, Techniques and Biochemical andClinical Applications. Oxford, Clarendon Press, 1986.

Apostolakos, M. J., Schuermann, W. H. T., Frampton, M. W., Utell, M. J.,Willey, J. C. Anal. Biochem. 213, 277 (1993)

Arakawa, H., Uetanaka, K., Maeda, M., Tsuji, A. J. Chromatogr. A 664,89 (1994 A)

Arakawa, H., Uetanaka, K., Maeda, M., Tsuji, Matsubara, Y., Narisawa, K.J. Chromatogr. A 680, 517 (1994 B)

Avaniss-Aghajani, E., Jones, K., Chapman, D., Brunk, C. BioTechniques17, 144 (1994)

Baba, Y., Ishimaru, N., Samata, K., Tsuhako, M. J. Chromatogr. A 653,329 (1993)

Baba, Y., Matsuura, T., Wakamoto, K., Tsuhako, M. J. Chromatogr. 558,273 (1991)

Baba, Y., Tsuhako, M. Trends Anal. Chem. 11, 280 (1992)

Bae, Y. C., Soane, D. J. Chromatogr. 652, 17 (1993)

Banke, N., Hansen, K., Diers, I. J. Chromatogr. 559, 325 (1991)

Barron, A. E., Blanch, H. W., Soane, D. S. Electrophoresis 15, 597 (1994)

Barron, A. E., Soane, D. S., Blanch, H. W. J. Chromatogr. 652, 3 (1993)

Bianchi, N., Mischiati, C., Ferriotto, G., Gambari, R. Nucleic Acids Res.21, 3595 (1993)

Bocek, P., Chrambach, A. Electrophoresis 12, 1059 (1991)

Bocek, P., Chrambach, A. Electrophoresis 13, 31 (1992)

Bode, H.-J. Anal. Biochem. 83, 364 (1977)

Brownlee, R. G., Sunzeri, F. J., Busch, M. P. J. Chromatogr. 533, 87(1990)

Butler, J. M., McCord, B. R., Jung, J. M., Allen, R. O. BioTechniques 17,1062 (1994 B)

88

Butler, J. M., McCord, B. R., Jung, J. M., Wilson, M. R., Budowle, B.,Allen, R. O. J. Chromatogr. B 658, 271 (1994 A)

Cadet, J., Weinfeld, M. Anal. Chem. 65, 675A (1993)

Camilleri, P. (Ed.) Capillary Electrophoresis: Theory and Practice. BocaRaton, CRC Press, 1993.

Cantor, R. C., Smith, C. L., Mathew, M. K. Ann. Rev. Biophys. Biophys.Chem. 17, 287 (1988)

Chan, K. C., Whang, C.-W., Yeung, E. S. J. Liq. Chromatogr. 16, 1941(1993)

Chargaff, E., Saidel, H. F. J. Biol. Chem. 177, 417 (1949)

Chen, F.-T. A., Tusak, A., Pentoney, Jr., S., Konrad, K., Lew, C., Koh, E.,Sternberg, J. J. Chromatogr. 652, 355 (1993)

Chen, J. W., Cohen, A. S., Karger, B. L. J. Chromatogr. 559, 295 (1991)

Cheng, J., Kasuga, T., Mitchelson, K. R., Lightly, E. R. T., Watson, N. D.,Martin, W. J., Atkinson, D. J. Chromatogr. A 677, 169 (1994)

Chiari, M., Nesi, M., Righetti, P. G. Electrophoresis 15, 616 (1994)

Chien, R. L., Burgi, D. S. Anal. Chem. 64, 489A (1992)

Chrambach, A. The Practice of Quantitative Gel Electrophoresis.Deerfield Beach, FL, VCH Publishers, 1985.

Cohen, A. S., Najarian, D. R., Paulus, A., Guttman, A., Smith, J. A.,Karger, B. L. Proc. Natl. Acad. Sci. USA 85, 9660 (1988)

Cohen, A. S., Najarian, D., Smith, J. A., Karger, B. L. J. Chromatogr.458, 323 (1988)

Compton, S. W., Brownlee, R. G. BioTechniques 6, 432 (1988)

De Gennes, P. G., in Scaling Concept in Polymer Physics. Ithaca, NY,Cornell University Press, 1979.

Del Principe, D., Iampieri, M. P., Germani, D., Menichelli, A., Novelli, G.,Dallapiccola, B. J. Chromatogr. 638, 277 (1993)

Demana, T., Lanan, M., Morris, M. D. Anal. Chem. 63, 2795 (1991)

Dovichi, N. Handbook of Capillary Electrophoresis, p. 369, Landers, J. P.(Ed.). Boca Raton, CRC Press, 1994.

Edstrom, J. E. Nature 172, 908 (1953)

89

Everaerts, F. M., Beckers, J. L., Verheggen, T. P. E. M. Isotachophoresis:Theory, Instrumentation and Applications. Amsterdam, Elsevier, 1976.

Fasco, M. J., Treanor, C. P., Spivack, S., Figge, H. L., Kamiinsky, L. S.Anal. Biochem., in press (1994).

Figeys, D., Arriaga, E., Renborg, A., Dovichi, N. J. J. Chromatogr. A 669,205 (1994)

Foret, F., Sustacek, V., Bocek, P. J. Microcol. Sep. 2, 299 (1990)

Foret, F., Szoko, E., Karger, B. L. Application Information BulletinA-1740. Beckman Instruments, Inc., Fullerton, CA, (1993)

Foret, F., Szoko, E., Karger, B. L. J. Chromatogr. 608, 3 (1992)

Gardiner, K. Anal. Chem. 63, 658 (1991)

Gelfi, C., Orsi, A., Righetti, P. G., Brancolini, V., Cremonesi, L., Ferrari,M. Electrophoresis 15, 640 (1994 B)

Gelfi, C., Orsi, A., Righetti, P. G., Zanussi, M., Carrera, P., Ferrari, M.J. Chromatogr. B 657, 201 (1994 A)

Gilliland, G., Perrin, S., Blanchard, K., Bunn, H. F. Proc. Natl. Acad. Sci.USA 87, 2725 (1990)

Glazer, A. N., Rye, H.S. Nature 359, 859 (1992)

Grossman, P. D., Menchen, S., Hersey, D. GATA 9, 9-16 (1992)

Grossmann, P. D., Soane, D. S. J.Chromatogr. 559, 257 (1991)

Guarnieri, C., Muscari, C., Stefanelli, C., Giaccari, A., Zini, M., Di Biase,S. J. Chromatogr. B 656, 209 (1994)

Guszczynski, T., Pulyaeva, H., Tietz, D., Garner, M. M., Chrambach, A.Electrophoresis 14, 523 (1993)

Guttman, A. Handbook of Capillary Electrophoresis, p. 129, Landers,J. P. (Ed.). Boca Raton, CRC Press, 1994.

Guttman, A., Cohen, A. S., Heiger, D. N., Karger, B. L. Anal. Chem. 62,137 (1990)

Guttman, A., Shieh, P., Hoang, D., Horvath, J., Cooke, N. Electrophoresis15, 221 (1994)

Guttman, A., Cooke, N. Anal. Chem. 63, 2038 (1991 A)

Guttman, A., Cooke, N. J. Chromatogr. 559, 285 (1991 B)

Guttman, A., Nelson, R. J., Cooke, N. J. Chromatogr. 593, 297 (1992 A)

90

Guttman, A., Wanders, B., Cooke, N. Anal. Chem. 64, 2348 (1992 B)

Guzman, N. A. (Ed.) Capillary Electrophoresis Technology. New York,Marcel Dekker, 1993.

Guzman, N. A. (Ed.) Journal of Capillary Electrophoresis. Shelton, CT,ISC Technical Publications, 1994.

Hall, J. M., King, M.-C. Nucleic Acids Res. 19, 2515 (1991)

Hebenbrock, K., Schugerl, K., Freitag, R. Electrophoresis 14, 753 (1993)

Heegaard, N. H. H., Robey, F. A. J. Liq. Chromatogr. 16, 1923 (1993)

Heiger, D. N., Cohen, A. S., Karger, B. L. J. Chromatogr. 516, 33 (1990)

Huang, M., Liu, S., Murray, B. K.,Lee, M. L. Anal. Biochem. 207, 231(1992)

Huang, X., Shear, J. B., Zare, R. N. Anal. Chem. 62, 2049 (1990)

Izumi, T., Yamaguchi, M., Yoneda, K., Okuyama, T., Shinoda, T.J. Chromatogr. A 652, 41 (1993)

Jansson, M., Roeraarde, J., Laurell, F. Anal. Chem. 65, 2766 (1993)

Jorgenson, J. W., Lukacs, K. D. Anal. Chem. 53, 1298 (1981)

Jorgenson, J. W., Lukacs, K. D. Science 222, 266 (1983).

Karger, B. L., Cohen, A. S., Guttman, A. J.Chromatogr. 492 (1989)

Kasai, K. J. Chromatogr. 618, 203 (1993)

Kasper, T. J., Melera, M., Gozel , P., Brownlee, R. G. J. Chromatogr. 458,303 (1988)

Katz, E. D., in Methods in Molecular Biology, Vol. 15: PCR Protocols,Current Methods and Applications, p. 63, White, B. A. (Ed.). Totowa, NJ,Humana Press, 1993.

Khrapko, K., Hanekamp, J. S., Thilly, W. G., Foret, F., Karger, B. L.Nucleic Acids Res. 22, 364 (1994)

Kim, Y., Morris, M. D. Anal. Chem. 66, 1168 (1994 B)

Kim, Y., Morris, M. D. Anal. Chem. 66, 3081 (1994 A)

King, M.-C., Rowell, S., Love, S. M. JAMA 269, 1975 (1993)

Kleemiss, M. H., Gilges, M., Schomburg, G. Electrophoresis 14, 515(1993)

Kuhr, W. G., Yeung, E. S. Anal. Chem. 60, 2642 (1988)

91

Kuypers, A. W. H. M., Meijerink, J. P. P., Smetsers, T. F. C. M., Linssen,P. C. M., Mensink, E. J. B. M. J. Chromatogr. B 660, 271 (1994)

Kuypers, A. W. H. M., Willems, P. M. W., van der Schans, M. J., Linssen,P. C. M., Wessels, H. M. C., de Bruijn, C. H. M. M., Everaerts, F. M.,Mensink, E. J. B. M. J. Chromatogr. 621, 149 (1993)

Landers, J. P. (Ed.) Handbook of Capillary Electrophoresis. Boca Raton,CRC Press, 1994.

Le Pecq, J.-B., Paoletti, C. Anal. Biochem. 17, 100 (1966)

Lee, T., Yeung, E. S., Sharma, M. J. Chromatogr. 565, 197 (1991)

Lerman, S. L and Frisch, H. L. Biopolymers 21, 995 (1982)

Li, W., Moussa, A., Giese, R. W. J. Chromatogr. 633, 315 (1993)

Liu, M.-S., Zang, J., Evangelista, R. A., Rampal, S. Chen, F.-T. A.BioTechniques (1995, in press)

Lloyd, D. K., Cypess, A. M., Wainer, I. W. J. Chromatogr. 568, 117(1991)

Loregian, A., Scremin, C., Schiavon, M., Marcello, A., Palu, G. Anal.Chem. 66, 2981 (1994)

Lu, W., Han, D.-S., Yuan, J., Andrieu, J.-M. Nature, 368, 269 (1994)

Lumpkin, O. J., Dejardin, P., Zimm, B. H. Biopolymers, 1573 (1985)

MacCrehan, W. A., Rasmussen, H. T., Northrop, D. M. J. Liq.Chrom. 15,1063 (1992)

Mansfield, E. S., Kronick, M. N. BioTechniques 15, 274 (1993)

Marino, M. A., Turni, L. A., Del Rio, S. A., Williams, P. E. J. Chroma-togr. A 676, 185 (1994)

Martin, F., Vairelles, D., Henrion, B. Anal. Biochem. 214, 182 (1993)

Maschke, H. E., Frenz, J., Belenkii, A., Karger, B. L., Hancock, W. Elec-trophoresis 14, 509 (1993)

Maschke, H. E., Frenz, J., Williams, M., Hancock, W. S. Poster T121,Fifth International Symposium on High Performance Capillary Electro-phoresis, Orlando, FL (1993)

Mathies, R. A., Huang, X. C. Nature 359, 167 (1992)

Mayer, A., Sunzeri, F., Lee, T.-H., Busch, M. P. Arch. Pathol. Lab. Med.115, 1228 (1991)

92

McCord, B. R., McClure, D. L., Jung, J. M. J. Chromatogr A 652, 75(1993)

McGregor, D. A., Yeung, E. S. J. Chromatogr. A 680, 491 (1994)

Milofsky, R. E., Yeung, E. S. Anal. Chem. 65, 153 (1993)

Nathakarnkitkool, S., Oefner, P., Bartsch, G., Chin, M. A., Bonn, G. K.Electrophoresis 13, 18 (1992)

Nelson, R. J., Paulus, A., Cohen, A. S. Guttman, A., Karger, B. L.J. Chromatogr., 480, 111 (1989)

Nesi, M., Righetti, P. G., Patrosso, M. C., Ferlini, A., Chiari, M. Electro-phoresis 15, 644 (1994)

Ng, M., Blaschke, T. F., Arias, A. A., Zare, R. N. Anal. Chem. 64, 1682(1992)

Nguyen, A.-L., Luong, J. H. T., Masson, C. Anal. Chem. 62, 2490 (1990)

Norwood, C. B., Jackim, E., Cheer, S. Anal. Biochem. 213, 194 (1993)

Oefner, P. J., Bonn, G. K. Am. Lab. 26, 28C (1994)

Ogston, A. G., Trans. Faraday Soc., 54, 1754 (1958)

Orita, M., Iwahana, H., Kanazawa, H., Hayashi, K., Sekiya, T. Proc. Natl.Acad. Sci. USA 86, 2766 (1989)

Ornstein, L. Ann. N.Y. Acad. Sci. 121, 321 (1964)

Pariat, Y. F., Berka, J., Heiger, D. N., Schmitt, T., Cohen, A. S., Foret, F.,Karger, B. L. J. Chromatogr. 652, 57 (1993)

Paulus, A., Husken, D. Electrophoresis 14, 27 (1993)

Pentoney, S. L., Jr., Konrad, K. D., Kaye, W. Electrophoresis 13, 467(1992)

Pentoney, S. L., Jr., Sweedler, J. V. Handbook of Capillary Electrophore-sis, p. 147, Landers, J. P. (Ed.). Boca Raton, CRC Press, 1994.

Perret, D., Ross, G. R. Human Purine and Pyrimidine Metabolism in Man,Vol. 7, Part B, p. 1, Harkness, R.A. (Ed.). New York, Plenum Press, 1991.

Perucho, M. The Polymerase Chain Reaction, p. 369, Mullis, K. B., Ferre,F., Gibbs, R. A. (Eds.) Boston, Birkhauser, 1994.

Piatak, M., Jr., Luk, K.-C., Williams, B., Lifson, J. D. BioTechniques 14,70 (1993)

Ramsey, R. S., Kerchner, G. A., Cadet, J. HRC 17, 4 (1994)

93

Rickwood, D., Hames, B. D. (Eds.). Gel Electrophoresis of Nucleic Acids.Washingon, D. C., IRL Press, 1983.

Righetti, P. G., Ettori, C., Chiari, M. Electrophoresis 12, 55 (1991)

Rogan, M. M., Drake, D. M., Goodall, D. M., Altria, K. D. Anal. Biochem.208, 343 (1993)

Rose, D. J. Anal. Chem. 65, 3545 (1993)

Rossomando, E. F., White, L., Ulfelder, K. J. J. Chromatogr. B 656, 159(1994)

Ruiz-Martinez, M. C., Berka, J., Belenki, A., Foret, F., Miller, A. W.,Karger, B. L. Anal. Chem. 65, 2851 (1993)

Schomburg, G. Capillary Electrophoresis: Theory and Practice, p. 255,Camillieri, P. (Ed.). Boca Raton, CRC Press, 1993.

Schwartz, H. E., Melera, M., Brownlee, R. G. J. Chromatogr. 480, 129(1989)

Schwartz, H. E., Ulfelder, K. Anal. Chem. 64, 1737 (1992)

Schwartz, H. E., Ulfelder, K. J., Chen, F.-T. A., Pentoney, S. L., Jr.J. Cap. Elec. 1, 36 (1994)

Schwartz, H. E., Ulfelder, K. J., Sunzeri, F. J., Busch, M. P., Brownlee,R. G. J. Chromatogr. 559, 267 (1991)

Schwer, C., Lottspeich, F. J. Chromatogr. 623, 345 (1992)

Shao, X., O’Neill, K., Zhao, Z., Anderson, S., Malik, A., Lee, M.J. Chromatogr. A 680, 463 (1994)

Slater, G. W., Noolandi, J. Am. Chem. Soc. Div. Polym. Chem. 29,416-417 (1988)

Slater, G. W., Noolandi, J. Biopolymers 28, 1781 (1989)

Smisek, D. L., Hoagland, D. A. Science 248, 1221 (1990)

Smith, S. B., Aldridge, P. K., Callis, J. B. Science 243, 203 (1989)

Southern, E. M. J. Mol. Biol. 98, 503 (1975)

Srinivasan, K., Girard, J. E., Williams, P., Roby, R. K., Weedn, V. W.,Morris, S. C., Kline, M. C., Reeder, D. J. J. Chromatogr. A 652, 83 (1993)

Srivatsa, G. S., Batt, M., Schuette, J., Carlson, R. H., Fitchett, J., Lee, C.,Cole, D. L., J. Chromatogr. A 680, 469 (1994)

94

Stellwagen, N. C., in Advances in Electrophoresis, Vol. 1, Chrambach, A.,Dunn, M. J., Radola, B. J. (Eds.). New York, VCH Publishers, 1987.

Stellwagen, N. C. Electrophoresis, 10, 332-344 (1989)

Strege, M., Lagu, A. Anal. Chem. 63, 1233 (1991)

Sudor, J., Novotny, M. V. Anal. Chem. 66, 2446 (1994)

Sunzeri, F. J., Lee, T.-H., Brownlee, R. G., Busch, M. P. Blood 77, 879(1991)

Takahashi, S., Murakami, K., Anazawa, T., Kambara, H. Anal. Chem. 66,1021 (1994)

Takigiku, R., Schneider, R. E. J. Chromatogr. 559, 247 (1991)

Tanaka, T. Sci. Am. 244, 124 (1981)

Terabe, S., Otsuka, K., Ando, T. Anal. Chem., 61, 251 (1989)

Tietz, D., Gottlieb, M. H., Fawcett, J. S., Chrambach, A. Electrophoresis7, 217 (1986)

Tiselius, A. Trans. Faraday Soc. 33, 524 (1937)

Tseng, H. C., Dadoo, R., Zare, R. N. Anal. Biochem. 222, 55 (1994)

Ueno, K., Yeung, E.S. Anal. Chem. 66, 1424 (1994)

Ulfelder, K., Schwartz, H. E., Hall, J. M., Sunzeri, F. J. Anal.Biochem.200, 260 (1992)

Ulfelder, K. J. Application Information Bulletin A-1748A, Beckman In-struments, Inc. (1993)

Ulfelder, K. J. Application Information Bulletin A-1774, Beckman Instru-ments, Inc. (1994)

Upcroft, P., Upcroft, J. A. J. Chromatogr. 618, 79 (1993)

van der Schans, M. J., Allen, J. K., Wanders, B. J., Guttman, A.J. Chromatogr. A 680, 511 (1994)

Vilenchik, M., Belenky, A., Cohen, A. S. J. Chromatogr. A 663, 105(1994)

Viovy, J. L., Duke,T. Electrophoresis 14, 322-329 (1993)

Volkmuth, W. D., Austin, R. H. Nature 358, 600 (1992)

Wang, P., Giese, R. W. Anal. Chem. 65, 3518 (1993)

Warren, W. J., Vella, G. BioTechniques 14, 598 (1993)

95

Watson, J. D., Crick, F. H. C. Nature 171, 738 (1953)

Zeillinger, R., Schneeberger, C., Speiser, P., Kury, F. BioTechniques 15,89 (1993)

Zhu, H., Clark, S. M., Benson, S. C., Rye, H. S., Glazer, A. N., Mathies,R. A. Anal. Chem. 66, 1941 (1994)

Zhu, M. D., Hansen, D. L., Burd, S., Gannon, F. J. Chromatogr. 480, 311(1989)

iii

Table of Contents

About the Author...................................................................... iv

Acknowledgments..................................................................... v

Front Cover................................................................................ v

Acronyms and Symbols Used................................................... vi

I. Introduction..................................................................... 1

II. Electrolyte Systems.......................................................... 4

A. Borate-Based Electrolytes........................................ 4B. Highly Alkaline pH Electrolytes............................... 7C. Carbohydrate-Metal Cation Complexes.................... 8

III. Detection Systems and Precolumn Derivatization........... 10

A. Detection of Underivatized Carbohydrates.............. 10B. Detection of Labeled Carbohydrates....................... 21

IV. Separation Approaches and Selected Applications.......... 33

A. Monosaccharides................................................... 33B. Oligosaccharides.................................................... 45C. Polysaccharides...................................................... 73D. Glycopeptides and Glycoproteins........................... 76E. Glycolipids............................................................ 84

V. Conclusions................................................................... 90

VI. References..................................................................... 92

iv

About the Author

Ziad El Rassi is an Associate Professor at the Department of Chemistry, Okla-homa State University. His current research includes the development ofHPCE and HPLC methods for important biological and environmental speciesand the miniaturization of immobilized enzyme reactors for use in analyticalseparations. He has published over 90 papers and 11 book chapters, and pre-sented several invited lectures at national and international meetings and loca-tions. He has edited

(i) a book entitled Carbohydrate Analysis: High PerformanceLiquid Chromatography and Capillary Electrophoresis forElsevier Science Publishers, Amsterdam, The Netherlands, 1995,

(ii) a symposium paper, 211 pages, for the journal Electrophoresis,1995, on “CE of Amino Acids, Peptides and Proteins” and

(iii) a symposium paper, 162 pages, for the journal Electrophoresis,1996 on “CE of Carbohydrate Species.”

He is currently editing another symposium paper on “CE of Pollutants andToxicants.” He is a member of the scientific committee of the Frederick Con-ference on Capillary Electrophoresis, the International Organizing Committeeof the International Symposium of Asia-Pacific Capillary Electrophoresis, andthe editorial boards of the Journal of Liquid Chromatography, Journal ofChromatography, Journal of Microcolumn Separations and Journal of Capil-lary Electrophoresis. He is the Associate Editor of the international journalElectrophoresis.

v

Acknowledgments

Ziad El Rassi acknowledges previous and current financial support of hisresearch program by the Cooperative State Research Service, U.S. Departmentof Agriculture, under Agreement Nos. 92-34214-7325 and 94-37102-0989.

Front Cover

The illustration on the cover depicts a computer-generated rendition of theN-terminal domain of a “variant surface glycoprotein” from T. brucei. This isa dimer, with a trisaccharide (NAG-NAG-MAN) attached to each monomer.The protein is shown as a ribbon, color coded by structural domain; the twotrisaccharides are shown, one in “space-filling” representation, and the otheras “liquorice bonds.” Courtesy of Don Gregory, Molecular Simulations,San Diego, CA.

Other Beckman primers (Volumes I through VII) on capillary electrophoresis:






Separation of Proteins and Peptides byCapillary Electrophoresis: Application toAnalytical Biotechnology 727484

Introduction to Quantitative Applicationsof Capillary Electrophoresis inPharmaceutical Analysis 538703

Separation of DNA by Capillary Electrophoresis 607397

vi


2-AA 2-aminoacridoneACN acetonitrileADCP amperometric detection at constant potentialAGP α1-acid glycoproteinAHNS 4-amino-5-hydroxynaphthalene-2,7-disulfonic acid3-ANDA 3-aminonaphthalene-2,7-disulfonic acidANDSA 7-aminonaphthalene-1,3-disulfonic acid2,6-ANS 2-anilinonaphthalene-6-sulfonic acid2-ANSA 2-aminonaphthalene-1-sulfonic acid5-ANSA 5-aminonaphthalene-2-sulfonic acidANTS 8-aminonaphthalene-1,3,6-trisulfonic acid2-AP 2-aminopyridineAPTS 9-aminopyrene-1,4,6-trisulfonic acid6-AQ 6-aminoquinolineCBQCA 3-(4-carboxybenzoyl)-2-quinolinecarboxyaldehydeCD cyclodextrinCE capillary electrophoresisCIF capillary isoelectric focusingCZE capillary zone electrophoresisC6MetBr hexamethonium bromideC6MetCl hexamethonium chlorideC10MetBr decamethonium bromideDAB 1,4-diaminobutaneDAP diaminopropaneDM-β-CD 2,6-di-O-methyl-β-CDED electrochemical detectionEDAC 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hy-

drochlorideEOF electroosmotic flow

vii

All trademarks are the property of their respective owners.

GAGs glycosaminoglycansGlcNAc N-acetylglucosaminehCG human chorionic gonadotropinHPAEC-PAD high-performance anion exchange chromatography-

pulsed amperometric detectionHPCE high-performance capillary electrophoresisHPLC high-performance liquid chromatographyIFN-γ interferon-γLIF laser-induced fluorescenceMECC micellar electrokinetic capillary chromatographyMEGA 10 decanoyl-N-methylglucamideMS mass spectrometryNMR nuclear magnetic resonancePAD pulsed amperometric detectionPAGE polyacrylamide slab gel electrophoresisPMP 1-phenyl-3-methyl-5-pyrazolonePMPMP 1-(p-methoxy)phenyl-3-methyl-5-pyrazolonerFVIIa human recombinant factor VIIarhBMP-2 recombinant human bone morphogenetic protein 2rHuEPO recombinant human erythropoietinSA sulfanilic acidTEA triethylamineTRSE 5-carboxytetramethylrhodamine succinimidyl ester

1

I. IntroductionCarbohydrates are polyhydroxylated aldehyde or ketone compounds that makeup most of the organic matter on earth because of their involvement in manylife processes. Carbohydrates have the following important functions:

(i) serve as energy stores, fuels, and metabolic intermediates,(ii) are structural elements in the cell walls of bacteria and plants and in the

exoskeletons of arthropods,(iii) are integral parts of glycoproteins and glycolipids, and(iv) play key roles in cell-cell recognition processes.[1]

Carbohydrates encompass a wide spectrum of compounds, many of whichare isomers or slightly different from each other. While monosaccharides aredivided into several classes (e.g., aldoses, ketoses, alditols, aldonic acids, etc.)and subclasses (e.g., trioses, pentoses, and hexoses), oligo- and polysaccha-rides which are composed of various combinations of monosaccharides areeven more diverse in structures than their monosaccharide constituents, form-ing linear, cyclic, or branched polymeric species. Several oligosaccharidevariations can be formed from a small number of monosaccharide units. Forinstance, while two amino acids can form only two different dipeptides, twomonosaccharides can be joined together in as many as 32 different disaccha-rides since the linkage can (i) occur at any of the four hydroxyl groups permonosaccharide, (ii) exist in either of two anomeric forms, and (iii) involveeither furanose or pyranose rings. Most oligosaccharides occur as side chainsattached to lipids in glycolipids and to polypeptides in glycoproteins and pro-teoglycans. Furthermore, since the carbohydrate chains of glycoproteins areenzymatically generated by processing enzymes that are generally not avail-able in sufficient quantities to yield uniform products, the composition of car-bohydrates at any of the glycosylation sites of a given glycoprotein may varysubstantially, yielding what is known as microheterogeneity.

Because of the multilateral roles of carbohydrates, their analysis has cometo have increasing importance. The complexity of carbohydrate solutes hasmost often engendered the need for an arsenal of analytical techniques and theuse of several chemical and biochemical tools and processes to bring abouttheir separation and structural characterization. The most widely used physico-chemical methods in the analysis of carbohydrates include nuclear magneticresonance (NMR), mass spectrometry (MS), gas-liquid chromatography(GLC), polyacrylamide gel electrophoresis (PAGE), traditional liquid chroma-tography (i.e., low-pressure) and high-performance liquid chromatography(HPLC), and more recently HPCE. NMR and MS are indispensable tools forthe structural elucidation of carbohydrates. With the development of the ion-

2

ization methods, electrospray ionization mass spectrometry (ESIMS) andmatrix-assisted laser desorption ionization time of flight mass spectrometry(MALDI-TOFMS), intact carbohydrates (e.g., oligosaccharides, glycoproteins,glycopeptides, and other glycoconjugates) can be readily measured at thepicomole to the femtomole levels. In addition, ESI is interfaced relativelyeasily with HPLC or CE, thus facilitating the simultaneous separation anddetermination of minute amounts of carbohydrates. However, both NMR andMS as well as CE- and LC-MS involve expensive equipment which generallyrestrict their use to a few specialized laboratories. Furthermore, while NMRstill requires relatively large amounts of samples which most often are difficultto obtain from biological sources, MS usually gives no information aboutlinkage positions and anomeric configurations. Due to these limitations, mostoften the analysis of carbohydrates is accomplished by chromatographic andelectrophoretic methods in conjunction with complimentary biochemical toolssuch as the use of specific lectins or monoclonal antibodies as well as specificexo- and endoglycosidases. While lectins and monoclonal antibodies (forinstance, in the form of lectin chromatography or immuno-chromatography orblotting), serve as biomolecular probes in the tentative structure elucidation ofcarbohydrates, exo- and endoglycosidases provide information about the typeof saccharide liberated and anomericity of the cleaved glycosidic bond as wellas the position to which cleaved sugar residue is linked.

Because of the inherent hydrophilic nature of carbohydrates, aqueous-based separation methods including HPLC, PAGE, and HPCE are very suit-able for their analysis. In this regard, HPCE seems to possess severaladvantages over HPLC and PAGE by (i) offering higher separation efficien-cies, (ii) yielding shorter analysis time, (iii) requiring small sample amounts,and (iv) consuming lower amounts of expensive reagents and solvents.

However, to realize the full benefits of the many sound features of HPCEincluding, among other things, its intrinsically high resolving power in theseparation of complex carbohydrate samples, two major difficulties have to besurmounted. First, with the exceptions of few naturally charged mono- andoligosaccharides, most carbohydrate molecules lack readily ionizable chargedfunctions, a condition that excludes their direct differential migration and, inturn, separation in electrophoresis. Second, most carbohydrate species neitherabsorb nor fluoresce, hindering their sensitive detection by modern analyticalseparation techniques including HPCE.

Various approaches have been introduced to render carbohydrates ame-nable to separation and detection by HPCE. These approaches have exploitedmany of the inherent properties of carbohydrates including (i) the ease with

3

which these molecules can be readily converted in situ to charged species bycomplex formation with other ions such as borate and metal cations which thenensure their differential electromigration and, in turn, separation in an electricfield, and (ii) the reactivity of the reducing end and other functional groups ofthe sugar molecules (e.g., carboxylic acid groups and amino groups) which canbe readily labeled with UV-absorbing or fluorescent tags, thus providing thecenters for sensitive detection. In addition, the electrochemical oxidation ofcarbohydrates at the surface of metallic electrodes provides another means bywhich underivatized carbohydrates can be sensitively detected.

The aims of this primer are to describe the basic aspects of the electrolytesystems used in HPCE of carbohydrates, to discuss the advantages and disad-vantages of the approaches and concepts that are most useful in the separationand detection of carbohydrates by HPCE, and to review important applica-tions.

For recent and detailed reviews of the various aspects of the capillarycolumn technology, the interested reader may consult References 2 and 3.Furthermore, a special issue of the journal Electrophoresis on “Capillary Elec-trophoresis of Carbohydrate Species” will appear at the same time as thisprimer.[4]

4

II. Electrolyte SystemsOnly some saccharides possess charged functional groups in their structureswhich would allow their differential electromigration and eventually separa-tion. These saccharides are the aldonic acids, uronic acids, sialic acids, aminosugars (glucosamine, galactosamine), and compositional sulfated sugars ofchondroitin, dermatan, keratan, and heparin. On the other hand, the separationof “neutral” carbohydrates by HPCE has often required the in situ conversionof these polyhydroxy compounds into charged species via complex formationwith other ions such as borate and metal cations. Furthermore, because of theionization of the hydroxyl groups of the sugars at extremely high pH, highlyalkaline pH electrolyte solutions are also useful for the electrophoresis of car-bohydrates.

Thus far, most HPCE separations of “neutral” carbohydrates have beenachieved by borate complexation and to a lesser extent by ionization at alkalinepH. Only one paper has appeared on the HPCE of sugars as alkaline earthmetal ion complexes.[5]

A. Borate-Based ElectrolytesPolyhydroxy compounds including carbohydrates can reversibly form anioniccomplexes with borate;[6] for recent reviews, see References 2 and 3. In these com-plexation reactions, it is the tetrahydroxyborate ion, B(OH)4

-, rather than boric acidthat undergoes complexation with the polyols. This is because boric acid in aqueousmedia acts as a Lewis acid to form the tetrahedral anion B(OH)4

-, and at alkalinepH, i.e., pH 8-10, where the complexation is most effective, equilibrium (1) is to avery large extent shifted to the right. In equilibria (1), (2), and (3), BL- and BL2

- arethe mono- and diesters, respectively, L is the polyol, and n = 0 or 1. Equilibrium (2)is situated very much to the right whereas Equilibrium (3) is dependent upon theposition of the hydroxyl groups in the polyol.

B(OH) 3 + OH - B(OH)4-

(B) (B-)

(1)

HO CCn

CHO

HOB

HO

O CCn

CO

+ 2H2O

(L) (BL -), Monocomplex

B(OH) 4- + (2)

(B-)

HO CCnCHO

OB

O

O CCn

CO

C

CnC

+ 4H2O

(L)

2

(BL2-), Dicomplex or Spirane

B(OH) 4- +

(B-)

(3)

5

According to the above equilibria, carbohydrates could form cyclic borateesters with either five or six atom rings when n = 0 or 1, respectively (i.e., withhydroxyl groups on adjacent or alternate carbon atoms, respectively).

The complex formation, as described by reactions (2) and (3), is possibleonly if two hydroxyl groups in the polyol molecule are favorably situated. Inthis regard, borate forms more stable complexes with cis- than with trans-oriented pairs of hydroxyl groups. Also, cyclic forms of sugars react lessstrongly with borate ions than do those having open chains (acyclic). In otherwords, alditols form stronger complexes with borate than do their counterpartaldoses under otherwise identical conditions. Increasing the number of hy-droxyl groups increases the strength of borate complexation. This is becausetwo hydroxyl groups will be more and more favorably situated with referenceto one another as the number of adjacent hydroxyl groups increases. The car-bohydrate-borate complex formation is largely influenced by the presence ofsubstituents in the polyol molecule as well as by their charges, locations, andanomeric linkages. Methylated sugars complex less than the parentunsubstituted sugars, and consequently the electrophoretic mobilities of methy-lated sugars in zone electrophoresis are much lower. Also, the complexation isstronger with methyl-β- than methyl-α-D-glucopyranoside and consequentlythe mobility of the α anomer, compared to that of the β anomer, is lower inzone electrophoresis with alkaline borate. This may be due to the fact that inthe α anomer the glycosidic methoxyl group occupies an axial position andwill interact strongly with the axial hydrogen atoms on C3 and C5, thus desta-bilizing the borate complex. This is not the case for the β anomer where thesubstituent occupies an equatorial position and consequently is free fromstrong, non-bonded interactions. Another parameter that must be considered incorroborating mobilities in borate systems is the presence of charged substitu-ents in the polyhydroxy molecule. Generally, a decrease of the stability of aborate ester is observed as a result of Coulombic repulsion between a nega-tively charged substituent (e.g., COO-) and BO4

- moieties.[7]

Usually, mono- and dicomplex (or spirane) borate esters coexist inaqueous solutions, and their molar ratio, among other things, is affected by therelative concentration of borate ions and sugar molecules. Spirane complexespredominate at a high sugar:borate ratio. In CE of carbohydrates, usually0.1-0.2 M borate is added to the running electrolyte and small plugs of10-4-10-5 M sugar samples are introduced into the separation capillary. Underthese conditions, anionic monocomplexes (i.e., BL-) are likely to predominateand thus migrate differentially under the influence of an applied electric field.It should be noted that, whether the injected sugar samples form BL- or BL2

-

or both while migrating in a borate medium, all sugar molecules will beassociated with a negative charge since mono- and dicomplex formations aredynamic. The magnitude of the charge will be influenced by the position of the

6

equilibrium and therefore by the stability of the complex. According toequilibria (1), (2), and (3) at constant sugar concentration, the amount ofcomplex increases with borate concentration according to the law of massaction and also with pH due to a higher concentration of borate ions.

In HPCE, at a given pH, resolution among various sugars increases withincreasing borate concentration. Usually, there is an optimum borate concen-tration at which maximum resolution is obtained for a multicomponent mix-ture[8] (see Figure 1). As seen in Figure 1b, 75 mM borate allowed the fullresolution of the seven sialooligosaccharides derived from gangliosides andderivatized with 7-aminonaphthalene-1,3-disulfonic acid. Also, at constantborate concentration, the complex formation as a function of pH varies amongvarious carbohydrate species with an optimum in the pH range 10-11 for mostmonosaccharides.[9]

Galβ1→3GalNAcβ1→4Galβ1→4Glc3↑2αNeuAc

Sialooligo-GM1 (1)

Galβ1→3GalNAcβ1→4Galβ1→4Glc3↑2αNeuAc8←2αNeuAc

Sialooligo-GD1b (4)

GalNAcβ1→4Galβ1→4Glc3↑2αNeuAc

Sialooligo-GM2 (2)

Galβ1→3GalNAcβ1→4Galβ1→4Glc3↑2αNeuAc8←2αNeuAc

3↑2αNeuAc

Sialooligo-GT1b (5)

Galβ1→3GalNAcβ1→4Galβ1→4Glc3↑2αNeuAc

3↑2αNeuAc

Sialooligo-GD1a (3)

Galβ1→4Glc3↑2αNeuAc

Sialooligo-GM3 (6)

Galβ1→4Glc3↑2αNeuAc8←2αNeuAc

Sialooligo-GD3 (7)

4 8 12 4 8 12 8 12 16

Fluo

resc

ence

a b c

x

3

1

2x

2

1 6

5

7

4

4

56

7

3

x

7

6

54

32

1

Time (min)

Figure 1. Electropherograms of ANDSA derivatives of sialooligosaccharides.Capillary, fused-silica, 50 cm (to detection point), 80 cm (total length) × 50-µm i.d.;running electrolytes, borate buffer of (a) 50, (b) 75, and (c) 125 mM, pH 10.0;voltage, 20 kV. Solutes, X = byproduct, 1 = sialooligo-GM1, 2 = sialooligo-GM2, 3 = sialooligo-GD1a, 4 = sialooligo-GD1b, 5 = sialooligo-GT1b,6 = sialooligo-GM3, 7 = sialooligo-GD3. Reprinted with permission.[8]

7

In summary, under a given set of conditions, various sugars (whethercharged or neutral) would undergo varying degrees of complexation withborate leading to differences in the electrophoretic mobilities of the complexedsolutes and hence separation. Thus, the use of alkaline borate is definitely anelegant approach for the high selectivity separation of saccharides.

B. Highly Alkaline pH ElectrolytesThe differential electromigration of carbohydrates in alkali-metal hydroxidesolutions, such as lithium, sodium, or potassium hydroxide is presumably dueto the ionization of the hydroxyl groups of saccharides at highly alkaline pH,yielding negatively charged species called alcoholates.[10] The ionization con-stants for carbohydrates are in the range of 10-12 to 10-14, i.e., pKa = 12-14.The pKa values of some typical sugars are listed in Table 1. Usually, reducingsugars (e.g., glucose, galactose, mannose, etc.) are the most easily ionizedwhile straight-chain alditols (e.g., glucitol, mannitol) have, on the average,about the same acidity as cyclitols (e.g., inositols) and glycosides (e.g., methyl-glucopyranosides) of similar molecular weight and hydroxyl content. Thehigher acidity of reducing sugars is caused by the higher lability of the hydro-gen atom of the hemiacetal (anomeric) hydroxyl group, a condition that appar-ently stems from an electron-withdrawing polar effect (inductive effect)exerted upon this group by the ring oxygen.

Table 1. Ionization Constants (Hydroxyl Group) of Carbohydratesin Water at 25°C [10]

Compound pKaD-Glucose 12.352-Deoxyglucose 12.52D-Galactose 12.35D-Mannose 12.08D-Arabinose 12.43D-Ribose 12.212-Deoxyribose 12.67D-Lyxose 12.11D-Xylose 12.29Lactose 11.98Maltose 11.94Raffinose 12.74*Sucrose 12.51D-Fructose 12.03D-Glucitol 13.57*D-Mannitol 13.50*Glycerol 14.40

* Measured at 18°C.

8

One important feature of Table 1 is that the greater the number of hy-droxyl groups, the greater the acidity. Glycerol has the same acidity as water;lactose and maltose, which are reducing disaccharides, seem to be somewhatmore acidic than aldopentoses (arabinose, ribose, lyxose and xylose).

Recently, highly alkaline electrolyte solutions such as lithium, potassium,or sodium hydroxide at pH greater than 12 have been shown to be useful in theseparation of underivatized saccharides by CZE.[11] As expected, the resolutionamong the various saccharides increased when moving from pH 12.3 to pH13.0 due to increasing ionization of the separated analytes[11] (see Figure 2).Also, the nature of the alkali-metal influences the resolution of the sugar ana-lytes.[11] It should be noted that separations at extremely high pH can only beperformed on naked fused-silica capillaries since most coated fused- silicacapillaries will undergo hydrolytic degradation under such basic conditions.

C. Carbohydrate-Metal Cation ComplexesIn the complex formation between metal cations and carbohydrates, the hydroxylgroups of carbohydrates are thought to from coordinate bonds with the metalcation. Usually, strong complexing occurs between cations and a contiguous axial(a), equatorial (e), axial sequence of hydroxyl groups in carbohydrates as wasascertained from the electrophoretic movement of compounds containing thissequence, and the immobility of many others lacking such an arrangement.[12]

For acyclic alditols, when three consecutive carbon atoms have the threo-threo configuration, the complex is most favored. An erythro-threo configura-tion is less favored for complex formation, and an erythro-erythro arrangementdoes not give rise to any noticeable complexation with cations. The more threopairs of hydroxyl groups there are in the alditol, the stronger will be its com-plexes.[13] If one of the three hydroxyl groups is replaced by a methoxy group,complexing becomes weaker. If all three hydroxyl groups are methylated,complex formation becomes negligible.[13] This draws similarities to boratecomplex formation.

The complex formation involving cyclic monosaccharides yields triden-tate complexes (see Figure 3), whereby no more than three oxygen atoms cancoordinate to one cation as shown in the following structure, where M+ is themetal cation.

Only in a few cases involving disaccharides, complexation was reportedto occur at more than three oxygen atoms, and tetra- and even pentadentatecomplexation have been described.[13] Further discussion regarding metalcation-carbohydrate complexation can be found in References 2, 3, and 13.

Recently, the use of electrolyte systems containing alkaline-earth metalsfor the separation of neutral carbohydrates by CZE has been reported.[5] Sepa-rations in these media are mainly based on differences in the extent of com-

9

d,e

e

a

bc

d

f

e

abc

df

a,b c

2 nA

Cur

rent

a

b

c

0 10 20 30 40

Time (min)

f

Figure 2. Electropherograms of six saccharides in different concentrations ofNaOH: (a) 20 mM, (b) 50 mM, and (c) 100 mM. The sugars are (a) stachyose,(b) raffinose, (c) sucrose, (d) lactose, (e) galactose, and (f) glucose (concentra-tions between 80 and 150µM). The fused-silica capillary dimensions are50-µm i.d. and 70 cm in length. The separation voltage is 10 kV. Injection is10 s by gravity (10-cm height); the ADCP is performed at 0.6 V (vs. Ag/AgCl).Reprinted with permission.[11]

HO

HOM+

OH

Figure 3. Metal-carbohydrate complex.

plexation of the divalent metals with the carbohydrate solutes and, to a lesserextent, on the bulk and shape of the molecule. Although these systems pro-vided a different selectivity than that achieved with borate buffers, the resolu-tion was in general inferior to that of borate buffers.

10

III. Detection Systems and PrecolumnDerivatization

Most carbohydrates lack chromophores in their structures, a condition thathinders their detection at low concentrations. To circumvent this impediment,several detection strategies have been developed including indirect UV andfluorescence detection, electrochemical detection, dynamic labeling via com-plexation with absorbing or fluorescing ions, and precolumn derivatizationwith a suitable chromophore and/or fluorophore.

A. Detection of Underivatized CarbohydratesA.1. Direct UV DetectionUV detectors, which are the workhorses of HPCE, have been used only occa-sionally in HPCE of underivatized carbohydrates. The major drawback of thisapproach is the limited sensitivity associated with the inherent low molar ab-sorptivities of most carbohydrates.

Recently, Hoffstetter-Kuhn and co-workers[14] performed spectral mea-surements of several simple carbohydrates and found that the addition of bo-rate yielded a two- to fifty-fold increase in the molar absorptivities at 195 nmfor mono- and oligosaccharides as their borate complexes. This increase wasattributed to the fact that borate complexation shifts the equilibrium betweencarbonyl and cyclic sugar forms toward the carbonyl form or to the presenceof additional functional elements such as oxygen bridges between boron andcarbon. However, in that work the sugars were detected from relatively con-centrated solutions, thus rendering the borate complexation not an attractivedetection approach for underivatized carbohydrates.

Only a few carbohydrates exhibit more or less significant absorbance inthe low UV. Oligosaccharides containing N-acetylglucosamine, N-acetyl-galactosamine, and sialic acid residues (i.e., glycans) can be detected at200 nm[15-17,18] or 185 nm.[19] The glycosaminoglycan hyaluronan could bedetected at 200 nm at a rather modest sensitivity.[20] This is facilitated by thepresence of a repeat unit of one glucuronic acid residue and one N-acetyl-glucosamine residue, linked by glycosidic bonds. Also, low-molecular-massheparins and heparins could be detected at moderate sensitivity at 200 nm.[21]

However, such low UV wavelengths impose serious restrictions on the choiceof the composition of the running electrolytes by not allowing the use of manyuseful additives that may absorb extensively in the low UV.

11

Another class of oligosaccharides which can be detected directly in theUV consists of acidic di- and oligosaccharides derived from glycosaminogly-cans (GAGs). These saccharides, which result from the enzymatic depolymer-ization of the large GAGs, bear unsaturated uronic acid residues at thenonreducing end that allow their direct UV detection at 232 nm.[21,22-27]

Synthetic oligosaccharide fragments of heparin including di-, tetra-,penta-, and hexasaccharides which lack the double bond in the uronic acidresidue were detected at 214 nm but the limit of detection was one order ofmagnitude lower than that reached by indirect UV[25] (see below for furtherdiscussion).

A.2. Indirect UV and Fluorescence Detection ofUnderivatized Carbohydrates

Indirect detection schemes are universal and can be used for compounds whichdo not possess the necessary physical properties for direct detection,i.e., chromophores or fluorophores. Indirect detection eliminates the need forpre- or postcolumn derivatization to convert the analyte of interest into a spe-cies that yields an acceptable detector response. In indirect detection, the ana-lyte is thought to displace a component of the running electrolyte which maybe a chromophore or fluorophore. Usually, a co-ion containing detectablefunctions is added to the running electrolyte. Since charge neutrality must bemaintained, an analyte of the same charge as the co-ion will therefore displacethe detectable co-ion. A general scheme illustrating the principles of indirectdetection as was reported by Yeung[28] is shown in Figure 4.

As can be seen in Figure 4A, the detectable co-ion provides a constantbackground signal and, as the analyte zone, which is deficient in the absorbingco-ion, passes through the detector, a decrease in the background signal occurs(see Figure 4B). When the analyte has completely passed through the detectionpoint, the detector response returns to the original baseline (see Figure 4C).Thus, the resulting peak is derived from the detectable background co-ionrather than from the analyte itself. On these bases, almost any detectionscheme in HPCE can be made to function in the indirect mode by altering thecomposition of the running electrolyte and not the actual instrumentation usedfor direct detection.

12

Detector

: Analyte : Detectable Co-ion

Recorder

A

B

C

Figure 4. Illustration of the displacement mechanism for indirect detection.

The mechanism of displacement in indirect detection was described byYeung.[28] The number of electrolyte co-ions displaced (or replaced) by oneanalyte molecule is defined as the transfer ratio (TR). Since a large backgroundsignal is required, the instability of the background signal can have tremendouseffects on the dynamic reserve (DR). DR is defined as the ability to measure asmall change on top of a large background signal. The DR is essentially theratio of the background signal to the background noise. The concentration limitof detection (Clim), expressed in concentration units, is given by[29]

Clim = CM

DR*TR (1)

where CM is the concentration of the detectable co-ion which generates thebackground signal. For a given system, the more stable the background signal(larger DR), the smaller the fractional change one can detect. Likewise, themore efficient the displacement process (large TR), the lower the Clim. Also,the lower the CM is, the greater the fractional change will be. It is desirable thatthe value of TR be close to unity.[29] It has been shown that the best detectionsensitivity is achieved when the analyte ions have an effective mobility closeto that of the detectable co-ion.[30] When optimizing Clim, it must be taken intoaccount that the three parameters are not necessarily independent. For ex-ample, decreasing CM will increase TR, but at the expense of decreasing DR.

13

It should be noted that the TR is not necessarily the same for all analytes. Forin depth discussion of this mode of detection, the interested reader is advised toconsult recent reviews.[28,29,31]

Both indirect photometric and fluorometric detection modes have founduse in CE of carbohydrates. In both types of indirect detection, one of thecritical factors is the selection of the detectable co-ion which should meetseveral criteria including: (i) a high molar absorptivity at the detection wave-length and excitation wavelength used in UV and fluorescence, respectively,(ii) a high quantum efficiency in fluorescence, preferably as close to unity aspossible, (iii) compatibility with the solvent system, i.e., it must be soluble andinert, (iv) non interactive with the capillary wall, and (v) must be charged,preferably with a charge identical to that of the analyte being displaced. Thelast criterion will ensure a value close to unity for the transfer ratio, TR. Torealize the full benefits of indirect LIF detection in terms of limit of detection,the stabilization of the laser power is of primary importance because it greatlyimproves the dynamic reserve, DR, by decreasing fluctuations in the back-ground signal.

The first application of the principle of indirect detection to the area ofHPCE of carbohydrates was demonstrated by Garner and Yeung[32] employ-ing laser-induced fluorescence (LIF) detection. In that work, Coumarin 343was used as the background fluorescing co-ion for indirect LIF. Coumarin 343was selected because of its good solubility, high quantum efficiency, and highmolar absorptivity (ε = 20,000) at 442 nm, which matches the 442 nm line of ahelium-cadmium (He-Cd) laser. Using this indirect LIF detection system,640 femtomoles of three simple sugars could be separated using 1 mM Cou-marin 343, pH 11.5, as the running electrolyte and a capillary of 18µm i.d. Thehigh pH used ensured the partial ionization of the neutral carbohydrates, ren-dering them amenable to electrophoretic separation as well as to indirect detec-tion. In indirect detection, the sensitivity is a function of the fraction of theanalyte that is ionized. This means that, when using indirect detection withsugars, the pH of the running electrolyte must be approaching 12 to have anysubstantial fraction, α, of the sugar solute in the ionized form. When the pH ofthe running electrolyte approaches pH 12, the concentration of hydroxide ionsis no longer negligible relative to the concentration of the detectable co-ion.The effect of the hydroxide ion can be approximated[33] by:

TRtot =α[sugar]

[FL] + [OH-] (2)

where TRtot is the total transfer ratio, α [sugar] is the fraction of the sugarionized, [FL] is the concentration of the detectable fluorophore, and [OH-] isthe hydroxide ion concentration. As can be seen in equation (2), at constantfluorophore (or chromophore) and sugar concentrations, α in the numerator

14

and [OH-] in the denominator are competing functions of pH. The total transferratio goes through a maximum when plotted as a function of pH. This maxi-mum is the most sensitive pH for detection of a given sugar. The optimumdetection pH for simple sugars was found to be 11.65, but 11.5 was used fordetection because the rate of degradation of the fluorophore (i.e., Coumarin343) was decreased without any appreciable decrease of detection efficiency.Using the detection system described above, the absolute limit of detection was2 femtomoles for fructose when using 5 µm i.d. capillaries with exceptionalseparation efficiencies.[33]

Very recently, detection limits in the picogram range are possible forhigh-molecular-weight polysaccharides (e.g., dextran, amylose, amylopectin,etc.) when laser-excited indirect fluorescence detection is employed.[34] Anargon-ion laser source operating at 488 nm was used for excitation. Because ofthe highly alkaline electrolyte (pH 11.5) needed to ionize the polysaccharides,1 mM fluorescein has to be added to the running electrolyte as the fluorophorein order to overcome the competition with the high concentration of hydroxideions and in turn to detect the polysaccharides (see Figure 5). Also, because ofthe weak ionization of polysaccharides at high pH, only 9 out of 11 polysac-charides studied could be detected.[34]

Although the above indirect detection schemes showed exceptional resultsin both sensitivity and efficiency, they involved the use of expensive laserequipment not available in most separation facilities. Indirect UV has beenshown to be feasible for the detection of underivatized carbohydrates. Bonnand co-workers[35,36] have demonstrated the use of 6 mM sorbic acid atpH 12.1 as both the electrolyte and the detectable co-ion in the separation andindirect detection of several simple sugars. The alkaline pH ensured ionizationof the sugars and, hence, their detection by means of charge displacement.Sorbic acid has a high molar absorptivity (ε = 27800 M-1cm-1 at 256 nm), andcarries a single charge, thus ensuring enhanced detectability and a favorabletransfer ratio, respectively. Under these conditions, a detection limit of 2 pico-moles was obtained for glucose.

More recently, Bergholdt et al.[37] employed indirect UV combined withHPCE for the separation and detection of two aldonic acids. Since these ana-lytes are naturally ionized, extreme pH is not required in order to ionize thesugars. The running electrolyte was 6 mM sorbic acid. Optimum resolutionwas obtained at pH 5.0. The detection limit under these conditions was deter-mined to be 18 femtomoles. The lower detection limits, when compared to theprevious example, can be attributed to the lower pH which allows for a larger TR.Moreover, it was shown recently that the use of high pH background electro-lyte resulted in a rapid increase of the low-frequency noise and baseline insta-bility. This was found to be related to the joule heat production and insufficient

15

0 2 4 6 8 10 12 14

Time (min)

Rel

ativ

e Fl

uore

scen

ce

AC

DB

Figure 5. Separation and detection of polysaccharides in CE by indirectfluorescence. Capillary, 80 cm total length, 65 cm to detection window× 26-µm i.d., 140-µm o.d.; injection, electromigration 1 s at 15 kV; run at15 kV with 1 mM fluorescein, pH 11.5. Solutes: A, dextran; B, comb-dextran;C, hydroxyethylamylose; D, amylose. Reprinted with permission.[34]

thermostating of the capillary tubing.[38] Detection could be greatly improvedby using narrow (25 µm i.d.) capillaries and low voltages, and by thermostat-ing the surrounding of the capillary column to allow a uniform heat dissipationalong the capillary.[38] In addition, the noise was influenced by the composi-tion of the background electrolyte. For instance, riboflavin and lithium werefound to be the best chromophore and counterion, respectively, giving the bestperformance in terms of higher signal-to-noise ratio. This was due to the factthat both compounds had the lowest mobility compared to the compoundstested. By optimizing all these factors, the pH range for the separation could beextended to 13 with a limit of detection of 50µM[38] which is one order ofmagnitude lower than previously reported values.[35] Using a riboflavin-NaOHelectrolyte system, the detection limit was improved by approximately 25times at pH 12.3.

Recently, the results of HPCE experiments involving the determination ofcarbohydrates in fruit juices by the HPCE-indirect UV mode of detection werecompared with those obtained by high-performance anion-exchange chroma-tography with pulsed amperometric detection (HPAEC-PAD).[39] In that work,potassium sorbate was chosen as the background electrolyte (pH 12.2-12.3)and chromophore for indirect UV detection at 256 nm. HPAEC-PAD yieldeddetection limits of 2-3 orders of magnitude lower than those with HPCE-indi-rect UV. However, the comparison was in favor of HPCE in terms of mass

16

detection. The absolute amount detectable for HPAEC-PAD was 25-50 pmol,while it was 0.9-1.1 pmol for HPCE-indirect UV. In the application of HPCE-indirect UV to the analysis of sugars in fruit juices, the detection sensitivity isnot an issue since the sugar concentration is relatively high (100 gL-1) and thesample must be diluted 1:50 prior to HPCE-indirect UV analysis.

Other carbohydrates were detected by HPCE-indirect UV including eightheparin disaccharides and some synthetic sulfated disaccharide and oligosac-charide fragments of heparin.[25] HPCE-indirect UV was achieved by usingeither 5 mM 5-sulphosalisylic acid, pH 3.0, or 5 mM 1,2,4-tricarboxylbenzoicacid, pH 3.5, as the running electrolyte and chromophore. In contrast to directUV detection, with indirect UV detection, the signal obtained for various syn-thetic pentasaccharides is nearly independent of their molecular structure andthe sensitivity is at least one order of magnitude higher than that of direct UVdetection. Again, because of the low pH where the transfer ratio is at its opti-mum value, the limit of detection of synthetic pentasaccharide heparin frag-ments was below 5 fmol when performing the detection at 214 nm using 5 mM5-sulphosalicylic acid, pH 2.5.[25] Even for heparin disaccharides possessingunsaturated uronic acid residues at the nonreducing end, HPCE-indirect UVat 214 nm in the presence of 5 mM 1,2,4,-tricarboxybenzoic acid at pH 3.5yielded higher sensitivity than HPCE-direct UV at 230 nm when employing200 mM sodium phosphate, pH 2.5, as the running electrolyte (see Figure 19).

Although the principle of indirect UV (also fluorescence) detection ap-pears relatively simple and is significantly more sensitive than direct lowwavelength UV of underivatized carbohydrates, several drawbacks can bepointed out. First, the instability of the detection system results in drift or dis-turbances of the baseline. Second, an indirect detection system requires work-ing at a low concentration of background electrolyte in order to have efficienttransfer ratios, a condition that results in lower efficiencies at higher sampleconcentrations and the possibility of solute-wall interactions. A third disadvan-tage imposed by indirect detection in HPCE is the limitation in the selection ofthe composition and pH of the background electrolyte. In other words, there isnot much room to manipulate selectivity and optimize separations. In fact,“neutral” carbohydrates are only partially ionized at the optimum pH normallyused in indirect detection, i.e., pH~12, a condition that does not favor theirhigh-resolution separation. Fourth, according to equation (2), quantificationcan become a significant problem with indirect detection methods, especially ifthe analytes possess significantly different electrophoretic mobilities. Finally,another disadvantage is the limited linear dynamic range, typically under twoorders of magnitude. Therefore, this mode of detection should be used only forfairly concentrated samples or whenever analytes are not easily derivatized orcannot be detected otherwise. It should be noted, however, that indirect UVdetection is perhaps the best approach described so far for CE of low-molecu-

17

lar-weight ionic species such as anions, cations, and organic acids[40] wheremore than 30 ions have been separated in less than three minutes. This is be-cause these ionic species are charged in their natural environment and, there-fore, selection of pH and ionic strength of the running electrolyte may not be ascritical for the outcome of the separation as when dealing with weakly ionizedanalytes, e.g., neutral sugars.

A.3. Electrochemical DetectionElectrochemical techniques have proven to be useful methods for the detectionof underivatized carbohydrates. In addition, electrochemical detection (ED) isan ideal method of detection for microcolumn-based separation systems. Thisis because detection is based on an electrochemical reaction at the surface ofthe working electrode so that cell volumes can be made very small with no lossin sensitivity. This is contrary to optical detectors where response is dependenton path length. In particular, amperometric methods are among the most sensi-tive approaches currently available for the detection of underivatized sugars.Amperometric detection is based on the measurement of current resulting fromthe oxidation or reduction of analytes at the surface of an electrode in a flowcell (for detailed discussions, see Reference 41).

Recently, two approaches for ED have been reported as useful methodsfor carbohydrate detection, namely amperometric detection at constant poten-tial (ADCP)[11,42,43] and pulsed amperometric detection (PAD).[44-46] Theprinciples of PAD have been described in detail by Johnson and LaCourse.[47]

The metallic electrodes most widely used to date for the detection of car-bohydrates and related species after HPLC separations have been the platinum(Pt) and gold (Au) electrodes.[47] The success of Pt and Au electrodes is duelargely to the tendency of sugars to adsorb on their surfaces where the sugarsreadily undergo electrochemical reactions at low potentials.[48] Unfortunately,these same adsorption phenomena also constitute one of the major disadvan-tages of Pt and Au in that accumulation of oxidation products generally leadsto electrode poisoning and, unless overcome experimentally, a rapid decreasein analyte response occurs. As a result, anodic detection schemes using Pt andAu electrodes typically include routine desorption and conditioning steps inorder to provide a stable and reproducible response. Most frequently, thesemeasures consist of a continuous pulsing of the surface to extreme positiveand/or negative potentials. This technique is known as pulsed amperometricdetection (PAD).

Recently, O’Shea et al.[44] introduced the PAD concept to the detection ofcarbohydrates after CE separation in an off-column detection format with agold wire microelectrode. The PAD has been successful because the multistepwaveform solves the problem of electrode poisoning typically found with the

18

oxidation of carbohydrates at Au electrodes in the direct amperometric detec-tion mode. This HPCE-PAD allowed a detection limit of 22.5 fmol. However,this detection limit was 1-2 orders of magnitude lower when a 10-µm disk goldelectrode was utilized for the detection of carbohydrates separated in a 10-µmi.d. capillary.[45] Figure 6 shows the separation and detection of eight sugarsusing a gold working electrode. A linear working range was observed from10-6 to 10-4 M for inositol.[45] Recently, Weber et al.[46] reported the applica-bility of PAD to the detection of complex carbohydrates such as glycopeptideswith a detection limit of 2µM (S/N = 3). Although this approach has solvedthe problem of electrode poisoning usually encountered with the oxidation ofcarbohydrates on Pt or Au electrodes in the direct amperometric detectionmode, the PAD detection system requires specialized pulse sequences, thusentailing expensive instrumentation. Also, other major drawbacks of PADinclude problems involving charging currents and surface changes associatedwith potential pulsing which do not allow the ultimate detectability to beachieved.

Cur

rent

(nA

)

Time (sec)400 450

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

1.6

3

46

5

78

9

12

500 550 600 650 700 750

Figure 6. Electropherogram of carbohydrates with PAD. Experimental condi-tions: separation voltage, 30 kV over 10 µm i.d. × 60 cm capillary; electrode,10µm (in diameter) Au disk; electrode potential, 300 mV (vs. SCE) for 165 ms(sampling at 111-165 ms), 1200 mV for 55 ms, and -1000 mV for 165 ms;electrolyte, 0.1 M NaOH; electromigration injection, 30 kV for 3 s; sampleconcentrations, 1 × 10-4 M for inositol and 2 × 10-4 M for others.Peaks: 1 = inositol, 2 = sorbitol, 3 = unknown, 4 = maltose, 5 = glucose,6 = rhamnose, 7 = arabinose, 8 = fructose, and 9 = xylose. Reprinted withpermission.[45]

19

An alternative to PAD is using amperometric detection at a constant po-tential (ADCP). ADCP has long been proven a useful approach for the detec-tion of electroactive species at the trace level using carbon electrodes. Theprincipal difficulty encountered in this approach is that carbohydrates exhibit alarge overpotential for oxidation at the carbon electrodes used in conventionalliquid chromatography with electrochemical detection. This phenomenondrastically increases the potential required for the oxidation and thereby com-promises both the selectivity and sensitivity of the detection. To overcome thisproblem, attention has been focused on the development of new electrodematerials[49] that permit the oxidation of carbohydrates at high pH and at rela-tively low potentials to provide optimum detector performance. In general,the Cu electrode was found to provide superior detection capabilities in termsof its range of response, detection limits, and especially stability, even in the0.10 M NaOH at which the studies were performed. The Cu electrode had adetection limit of 3 × 10-8 M for glucose at +0.58 V (vs. Ag/AgCl) with alinear response range over four decades.[49]

Zare and co-workers[11] employed ADCP with a Cu microelectrode forthe detection of carbohydrates after separation by HPCE. The separation ofsugars was performed in strongly alkaline solutions (i.e., pH 13) without priorderivatization or complexation with borate ions. The Cu microelectrode at+0.6 V (vs. Ag/AgCl) could be employed for hundreds of runs without deterio-ration. Because the pKs of most sugars are in the vicinity of 12-13, they areionized at high pH and separated by HPCE under such conditions. Figure 7illustrates the separation and detection of 15 different carbohydrates.

The limits of detection were calculated to be below 50 femtomoles for the15 sugars studied with a linear dynamic range that extended over 3 orders ofmagnitude (e.g., mM-mM). However, the reproducibility of this system is verylow due to the difficulty associated with the electrode/capillary alignmentduring an electrophoresis run and from run-to-run. Recently, Ye andBaldwin[42] reported the design and characterization of a simple wall-jet elec-trochemical detector which allows the use of normal size working electrodes,thus increasing the reproducibility of amperometric detectors without introduc-ing significant peak broadening. In this approach, a disk-shaped electrodeconsisting of metal wire with only its tip cross section exposed was positionedimmediately in front of the capillary outlet. Detection was performed on thesolution exiting the capillary and flowing radially across the face of the copperelectrode. This design was shown to exhibit 50 times improvement in detectionlimit (ca. 1 fmol) over the conventional ED, and 5-6 times reproducibilityimprovement.[42]

20

Time (min)

Cur

rent

0 15 30 45

0.25 nA

a

b

cd

e

h

i

f

g

k

j

l

mn

o

Figure 7. CZE/ADCP electropherogram of a mixture containing 15 differentcarbohydrates (80-150 µM). Electrolyte, 100 mM NaOH; capillary,fused-silica, 73 cm (total length) × 50-µm i.d.; 10 s hydrodynamic injection,10-cm height; voltage, 11 kV; analytes, a = trehalose, b = stachyose,c = raffinose, d = sucrose, e = lactose, f = lactulose, g = cellobiose,h = galactose, i = glucose, j = rhamnose, k = mannose, l = fructose,m = xylose, n = talose, o = ribose. Reprinted with permission.[11]

Regardless of whether the PAD or ADCP approach is used for the am-perometric detection of carbohydrates, both approaches suffer from (i) thelimitations imposed by the alkaline conditions needed for sensitive detectionand differential electromigration which restrict the useful pH to a very narrowrange, i.e., pH >12, and (ii) the non-discriminative nature of amperometricdetection which is known to yield a response not only for carbohydrates but

21

also for other analytes including amino acids, peptides, organic acids, simplealcohols, and aliphatic amines,[47] a condition that may render peak assign-ment difficult and may lead to less accurate quantitative measurements.Nevertheless, HPCE-ED permitted the rapid determination of sugars in bever-ages,[11,43] the detection of glycopeptides,[46] the determination of glucose inhuman blood,[44] and the monitoring of the enzymatic oxidation of glucosewith time.[43] Unfortunately, at the present time this type of detector is notavailable from a commercial source. This may become a reality as soon asfurther improvements are realized in the design of electrochemical detectors.

B. Detection of Labeled CarbohydratesB.1. Dynamically Labeled CarbohydratesRecently, mixtures of α-, β- and γ-cyclodextrins (CDs) were separated anddetected by HPCE-LIF using 2-anilinonaphthalene-6-sulfonic acid (2,6-ANS)as the background electrolyte.[50] This scheme exploits the ability of cyclo-dextrins to form inclusion complexes with 2,6-ANS, thus allowing their simul-taneous differential electromigration and LIF detection. As the fluorophorecomplexed with the hydrophobic cavity of CDs, its fluorescence was en-hanced, and consequently the CD-2,6-ANS adducts could be detected by thefluorescence increase as positive peaks. The LIF detection was performed withan argon ion laser operating at 363.8 nm for excitation and the emission wascollected at 424 nm. Under these conditions, the detection limits were deter-mined to be 62, 2.4 and 24µM for α-, β- and γ-cyclodextrins, respectively.This may reflect that β-CD forms the strongest complex with 2,6-ANS. In fact,they eluted and separated in the order of increasing strength of complexation(see Figure 8).

Very recently, large polysaccharides such as amylopectin and amylosewere shown to complex with iodine which provides both the charge needed fordifferential electromigration and the chromophore necessary for direct visibledetection at 560 nm.[51] The primary basis for this process is iodine bindingaffinity to carbohydrates which can be manipulated through control of tem-perature and iodine concentration (see section IV.C.2 for more details).

B.2. Precolumn DerivatizationCarbohydrates are generally tagged with a suitable chromophore or fluoro-phore to allow their detection at low levels. As the tagging process bringsabout dramatic changes in the structure of the carbohydrate analytes, it is wiseto carefully select the tag not only to allow the sensitive detection of the de-rivatized carbohydrates but also to produce the changes needed for the subse-quent separation step. In other words, it is preferred that the tag also supplies

22

7.87.67.47.27.06.86.66.4Time (min)

Neu

tral

s

Fluo

resc

ence

(ar

bitr

ary

units

)

α

γ

β

Figure 8. CE separation of α-, β- and γ-CD. Analysis buffer: 40 mM phos-phate, pH 11.76, 1 mM 2,6-ANS. The analysis was carried out in a capillary ofdimensions 50-µm i.d., 360-µm o.d. and 1 m in length in a field of 300 Vcm-1.The sample was introduced into the capillary by electrokinetic injection; 5 kVfor 2 s from a sample containing 1.44 mg/mL α-CD, 0.017 mg/mL β-CD, and0.24 mg/mL γ-CD. Detection was by fluorescence excited at 363 nm and moni-tored at 424 nm. Reprinted with permission.[50]

the charge necessary for electrophoresis over a relatively wide range of pHs orthat the tag imparts a hydrophobic character to the derivatives so that the prin-ciples of MECC can be applied to the separation of derivatized carbohydrates.Other essential criteria for a successful precolumn derivatization include (i)high yields, (ii) the formation of a single product for each species, (iii) nodetectable side products, (iv) minimum sample workout and cleanup, and (v)no cleavage of an essential sugar residue, e.g., sialic acid residue. In precolumnderivatization reactions, it is generally preferred that the tagging occurs at onlyone reactive functional group of the analyte and should be complete so that asingle derivative is obtained in high yields. The polyhydroxy nature of sugarsis attractive as far as the attachment of a tag to the molecule is concerned. Thisroute to derivatization has been used extensively in gas-liquid chromatographyin order to increase the volatility of carbohydrates and consequently facilitatetheir separation. However, derivatizing the hydroxyl groups would lead tomultiple tagging of an analyte and, because hydroxyl groups vary in theirrelative reactivities, a distribution of derivatives rather than a single productwould be obtained. In order to prevent multiple derivatization, other functionalgroups on the sugar molecule must be considered. The most popular sites for

23

tagging include (i) the carbonyl group in reducing sugars, (ii) amino group inamino sugars, and (iii) the carboxylic moiety in acidic sugars. To produce asingle product in a given precolumn derivatization, the tag must possess onlyone reactive site for attachment to the analyte. For UV detection, the taggingagent should exhibit a high molar absorptivity at a given wavelength withminimal interferences from the running electrolyte to ensure highly sensitivedetection of the derivatized carbohydrates. Likewise, in fluorescence a taggingagent should exhibit high quantum efficiencies at a given excitation wave-length. These requirements become crucial when dealing with extremely smallsample volumes encountered in nano-scale separation techniques such asHPCE.

Thus far, five different precolumn derivatization schemes have been intro-duced for the tagging of carbohydrates: (1) reductive amination (the mostwidely used)[52,53] (see Scheme I); (2) condensation of carboxylated carbohy-drates with aminated tags in the presence of carbodiimide[8,54-56] (see SchemeII); (3) base-catalyzed condensation between the carbonyl group of reducingcarbohydrates and the active hydrogens of 1-phenyl-3-methyl-5-pyrazolone(PMP) or 1-(p-methoxy)phenyl-3-methyl-5-pyrazolone (PMPMP), formingbis-PMP and bis-PMPMP derivatives, respectively[5,9,57,58] (see Scheme II);(4) reductive amination of reducing carbohydrates with amines to yield1-amino-1-deoxyalditols followed by reaction with 3-(4-carboxybenzoyl)-2-quinolinecarboxyaldehyde (CBQCA) in the presence of potassium cyanide[59]

(see Scheme IV); (5) reductive amination of reducing carbohydrates withamines to yield 1-amino-1-deoxyalditols followed by reaction with 5-carboxy-tetramethylrhodamine succinimidyl ester (TRSE)[60] (see Scheme V).

With the exception of precolumn derivatization in Scheme III which yields UVabsorbing derivatives, all other precolumn derivatization schemes can produceboth UV-absorbing and fluorescing derivatives depending on the spectralproperties of the tag used. It should be noted that the precolumn derivatizationaccording to Scheme III will yield fluorescing derivatives if a fluorescent tagsimilar to PMP in terms of chemical reactivity becomes available. Whileprecolumn derivatization according to Schemes IV and V necessitates two ormore distinct workout steps, the remaining are much simpler, requiring onlyone derivatization step.

Charts 1 and 2 list the most important tags for the derivatization of carbo-hydrates. Those which yield neutral or ionizable sugar derivatives are listed inChart 1 while those yielding permanently charged sugar derivatives are listedin Chart 2.

24

O

OH

NH-C-CH3

OH

CH2OH

O

OH

N

H2N

OH

CH2-NHOH

NH-C-CH3

CH2OH

ON

+NaBH3CN

H+

6-AQGlcNAc 6-AQ-GlcNAc

HO

Scheme I. Illustration of reductive amination where the reducing sugar isN-acetylglucosamine (GlcNAc) and the tag is 6-aminoquinoline (6-AQ).[61]

CRN

C NR'

CR1O

RHN

C NHR"

O

+ R1 C NHR'RHNR"NH 2 +

NR'

Carbodiimide

+ R1COOH

IsoureaCarbohydrate Derivative

CarboxylatedCarbohydrate

Intermediate

Derivatizing Agent

C NR'

O

CR1O

RHN O

O

Scheme II. Illustration of the selective precolumn derivatization of carboxy-lated carbohydrates via a condensation reaction between the carboxylic groupof the saccharide and the amino group of the derivatizing agent in the pres-ence of carbodiimide.[54]

OHO

HOOH

OH

OHHO

HOOH

OH NN

NN

H3C

OH

CH3

HO

NN

CH3

O+

H,OH

OH-CH

Reducing sugar PMP

Bis-PMP Derivative

Scheme III. Illustration of condensation reaction with PMP.[58]

25

OCH2OH

OH

OHOH

OHCH2OH

OH

OHOH

NH2

N

N

CN

COOH

OHCH2OH

OH

OHOHCBQCA/CN-

NH4+/NaBH3CN

H,OH

Reducing Sugar 1-Amino-1-deoxyalditol CBQCA-Sugar Derivative

Scheme IV. Illustration of precolumn derivatization with CBQCA.[59]

O N+(CH3)2(H3C)2N

COO-

C

O

O N

O

O

OHOH

HOHO

HO

NH2

O N+(CH3)2(H3C)2N

COO-

C

O

OHOH

HOHO

HON

HTRSE-Aminoglucitol

Derivative TRSEHydroxysuccinimide

HO N

O

O

1-Amino-1-deoxy-D-glucitol

+ +

Scheme V. Illustration of the derivatization of aminated sugar with TRSE.Since it was not provided in the pertinent reference, i.e., 60, this scheme wasinspired from Reference 62 which describes the linking of peptides to activatedmethoxy polyethylene glycol succinimidyl succinate.

26

N

NH2

NH2

COOH

NH2

CN

N CHO

C

O

COOH

NN

CH3

ON

ONH2

H

3-(-4-Carboxybenzoyl)-2-quinolinecarboxaldehyde (CBQCA)

λem = 552 nm Argon ion laser 457 nm or 488 nm

He-Cd laser 442 nm

p-Aminobenzoic acidλmax = 285 nm

p- Aminobenzonitrileλmax = 285 nm

(I)

2-Aminopyridine (2-AP)λmax = 240 nm λem = 375 nm

He-Cd laser 325 nm

6-Aminoquinoline (6-AQ)λmax = 270 nm λem > 495 nm

He-Cd laser 325 nm

(II)

CH NH2

CH3

(III)

(IV)

1-Phenyl-3-methyl-5-pyrazolone (PMP)

λmax = 245 nm

2-aminoacridone (2-AA)λexc = 425 nm, λem = 520 nm

Argon ion laser 488 nm

S-(-)-1-Phenethylamineλ = 200 nm

(V) (VI)

(VII) (VIII)

N

H2N

Chart 1. Structures, names and abbreviations of tags yielding neutral andionizable sugar derivatives. The spectral data were taken from Ref. 58 fortag I; from Refs. 138 and 110 for tag II concerning the fluorescence propertiesand the use of LIF detection, respectively; from Ref. 68 for tag III; from Refs. 9and 65 for tag IV regarding UV detection and LIF detection, respectively;from Refs. 61 and 65 for tag V concerning UV detection and LIF detection,respectively; from Ref. 69 for tag VI; from Refs. 77, 79, and 82 for tag VII;from Ref. 67 for tag VIII. The spectral values listed in the chart correspond tothe sugar derivatives of the indicated tags. As can be seen in this chart, most ofthe derivatives were excited at a wavelength dictated by the line of the lasersource available, where in most cases the laser line yields only a fraction ofthe maximum absorption of the derivatives. The interested reader is advised toconsult the listed references to find more details concerning UV and LIF detec-tion of the various derivatives.

27

NH2

SO3-

H2NSO3

-

SO3-

SO3-

SO3--O3S

NH2

SO3-

NH2

NH2

SO3-

SO3-

NH2

-O3S

OH NH2

-O3S SO3-

Sulfanilic acid (SA)λmax = 247 nm

7-Aminonaphthalene-1,3-disulfonicacid (ANDSA), λmax = 247 nmλexc = 315 nm λem = 420 nm

Xenon-Mercury Lamp

-O3S SO3-

NH2-O3S

3-Aminonaphthalene-2,7-disulfonic acid (3-ANDA)

λ = 235 nm

8-Aminonaphtha-lene-1,3,6-trisulfonic acid (ANTS)

λexc = 370 nm, λem = 520 He-Cd laser 325 nm

2-Aminonaphthalene-1-sulfonic acid (2-ANSA)

λ = 235 nm

5-Aminonaphthalene-2-sulfonic acid (5-ANSA)

λ = 235 nm λem = 475 nm

He-Cd laser 325 nm

4-Amino-5-hydroxynaphthalene-2,7-disulfonic acid (AHNS)

λem = 475 nm He-Cd laser 325 nm

(XI)

(XVII)

(X)

(XII)

O

(XIII)

N+(CH3)2

(XIV)

(H3C)2N

(XV)

COO-

(XVI)

C

9-Aminopyrene-1,4,6-trisulfonic acid (APTS)

λexc = 455 nm, λem = 512 nmArgon ion laser 488 nm

O

O N

O

O

5-Carboxytetramethylrhodaminesuccinimidyl ester (TRSE)

λem = 580 nm He-Ne laser 543 nm

(IX)

Chart 2. Structures, names and abbreviations of tags yielding permanentlycharged sugar derivatives. Spectral data were taken from Refs. 60 and 84 fortag IX; from Ref. 54 for tag X; from Ref. 98 for tag XI; from Refs. 98 and 65for tag XII concerning UV and LIF detection, respectively; from Ref. 54 for tagXIII; from Ref. 65 for tag XIV; from Ref. 98 for tag XV; from Refs. 74 and 97for tag XVI; from Refs. 70 and 71 for tag XVII. As in Chart 1, the spectralvalues listed in the chart correspond to the sugar derivatives of the indicatedtags. For more details, see the listed references. Here also most of the deriva-tives were excited at a wavelength dictated by the line of the laser sourceavailable, where in most cases the laser line yields only a fraction of the maxi-mum absorption of the derivatives.

28

Besides their different spectral characteristics in terms of molar absorptivi-ties and quantum efficiencies, and whether they are UV absorbing and/or fluo-rescing tags, the different tags listed in Charts 1 and 2 will yield sugarderivatives of varying electrophoretic mobility, separation efficiency, andselectivity. Limiting our discussion to labeling only “neutral” carbohydrateswith the various tags, and distinguishing between the various schemes ofprecolumn derivatization listed above, we can state the following guidelines.PMP (tag I), which can only label carbohydrates according to Scheme III,usually yields neutral sugar derivatives which become negatively charged inaqueous basic solutions due to the partial dissociation of the enolic hydroxylgroup of the PMP tag. Usually, the weakly ionized derivatives do not separatewell. Thus, the sugar derivatives of tag I will require borate-based electrolytesand/or micellar phases such as SDS to bring about differential electromigrationand, in turn, separation. Tags II and III will lead to neutral sugar derivatives viaScheme I, thus requiring borate-based electrolytes and/or micellar phases forseparation. Using Scheme I, tags IV, V, and VI will lead to sugar derivativesthat can acquire a positive charge at acidic pH. In addition, tag VI is an enan-tiomeric reagent that allows the separation of sugar enantiomers. Sugar deriva-tives of tags IV, V, and VI obtained via Scheme I will electrophorese in boratebuffers at alkaline pH. Since the sugar derivatives of tags IV, V, and VI areneutral at basic pH, one can envision that they can be electrophoresed byMECC at basic pH. CBQCA (tag VII), which is exclusively used in precolumnderivatization according to Scheme IV, will yield derivatives that can becomenegatively charged via the ionization of the carboxylic group. TRSE (tag IX)used for labeling carbohydrates according to Scheme V will produce sugarderivatives that can acquire a net positive charge at acidic pH and become azwitterion at neutral and basic pH. Tag VIII is an amphoteric solute and, there-fore, when used to tag carbohydrates via Scheme I, will give derivatives whichwill charge positively at pH < 3.0 and negatively at ≥ 3.8. Using Scheme I, allthe other tags (X to XVII) will produce sugar derivatives that are negativelycharged over a wide range of pH due to their strong sulfonic acid groups andthe very weak ability of their amino groups to become protonated (pKa values≤ 3). Thus, under a given set of conditions and for a given set of saccharides,different tagging agents will lead to sugar derivatives with different electro-phoretic behavior and selectivity.

It should be emphasized that the sugar derivatives of all the tags will, ofcourse, electrophorese at high pH in the presence of borate buffers. This willbecome clear as this discussion progresses and more details are provided inPart IV. Due to the presence of permanently ionized strong sulfonic acidgroups in the tags listed in Chart 2, their sugar derivatives will electrophoreseover a wide range of pH. Although they do not require borate complexation toundergo differential electromigration, the sugar derivatives of the sulfonic

29

acid-based tags will certainly exhibit different electrophoretic behavior asborate complexes and, in turn, different selectivity may be obtained.

Due to the inherent high sensitivity of fluorescence, it is not surprising tosee that laser-induced fluorescence (LIF) is the current trend for high-sensitiv-ity detection of labeled carbohydrates after HPCE separation. In addition,fluorescence is characterized by a good specificity and a large linear dynamicrange. The specificity of fluorescence is the result of two main factors. First,only a small percentage of compounds fluoresce because not all compoundsthat absorb radiation are capable of emitting radiation upon returning to theground state. In fact, fewer than 10% of all absorbing compounds will emitradiation via any luminescence scheme. Second, two wavelengths are used influorometry, but only one in spectrophotometry. Two compounds that absorbradiation at the same wavelength are not likely to emit radiation at the samewavelength unless they are structurally very similar. Likewise, two compoundsthat emit radiation at the same wavelength will probably absorb radiation attwo different wavelengths. Another important advantage is the extended lineardynamic range. It is not unusual to encounter six to seven orders of magnitudelinearity with fluorescence spectrometry, but only two to three orders of mag-nitude with spectrophotometric procedures.

One major drawback of fluorescence detection is that most analytes do notpossess satisfactory fluorophores and must be chemically tagged in order toachieve high-sensitivity detection. This is especially true when working withcarbohydrates. Also, another principal disadvantage of fluorescence as ananalytical tool lies in its serious dependence on environmental factors such astemperature, pH, ionic strength, and the presence of dissolved oxygen. Photo-chemical decomposition is rarely a problem when using fluorescence as amode of detection because the analyte is only briefly exposed to the intenseexcitation radiation. Quenching, which is the reduction of fluorescence by acompeting deactivating process resulting from the specific interaction betweena fluorophore and another substance present in the system, can introduce sig-nificant errors when using fluorescence as an analytical technique. The mostnoticeable forms of quenching in fluorescence detection are temperature, oxy-gen, and analyte concentration. As the temperature is increased, the fluores-cence decreases. This is due to the resulting increase in molecular motion andcollisions which rob a molecule of energy through collisional deactivation. Thechange in fluorescence is typically 1% per 1°C. However, for some com-pounds, such as tryptophan or rhodamine B, it can be as high as 5% per 1°C. InHPCE, changes in the buffer composition, pH, ionic strength, and the operat-ing voltage can all lead to large changes in temperature within the capillary.For quantitative analysis, using fluorescence as the mode of detection, caremust be taken such that the temperature at which the separation is carried out is

30

similar to that at which calibration was performed. Oxygen is another source ofquenching. Oxygen present in solutions at a concentration of 10-3 M normallyreduces fluorescence by 20%. In HPCE, oxygen quenching can be reduced oreliminated by degassing solvents and sample solutions before use. Concentra-tion quenching causes many problems during fluorometric determinations. Inorder for fluorescence to be observed, absorption must occur first. When theabsorption is too high, light can not pass through the entire flow cell to causeexcitation. In order for a linear relationship to be observed between fluores-cence and concentration, the absorbance must be kept below 0.05, otherwise adecrease in fluorescence can be observed. For a detailed description of theinstrumentation used in fluorescence detection with HPCE, see recent reviewsby Amankwa et al.[63] and Li.[64]

So far, precolumn derivatization by reductive amination (Scheme I) hasbeen the method of choice for the labeling of reducing saccharides.[65-71] Thishas been facilitated by the availability of a large number of chromophores withamino functions. Typically, 2-aminopyridine (2-AP, tag IV) derivatives ofsugars formed via reductive amination[52] are usually detected in the UV at240 nm at the 10 pmol level. They also can be detected by fluorescence. Also,6-aminoquinoline (6-AQ, tag V)[61] is an efficient tag for linear and branchedoligosaccharides and the subsequent separation of the 6-AQ derivatives inHPCE. The 6-AQ derivatives showed a maximum absorbance at 270 nm, andthe signal obtained was 8 times higher compared with 2-AP derivatives underotherwise identical conditions.[61] The removal of excess 2-AP and 6-AQ isnot required in HPCE since the derivatized carbohydrates elute after the excesstag or, in the case of acidic sugars,[72] the derivatives and the derivatizing agentmove in opposite direction. Linhardt and co-workers[73] have described thederivatization via reductive amination of N-acetylchitooligosaccharides with anegatively charged tag, 7-aminonaphthalene-1,3-disulfonic acid (ANDSA, tagXIII). The ANDSA-derivatives were excited at 250 nm using an arc-lamp andthe fluorescence was collected at 420 nm with detection limits in the femto-mole range. UV detection was also described for sugar derivatives ofANDSA.[73] ANDSA has a fairly strong absorbance at 247 nm (ε = 3100 M-1

cm-1), but more importantly it supplies the charge necessary for rapid electro-phoretic analysis of sugars in HPCE. Chiesa and Horváth[74] utilized a similartag, 8-aminonaphthalene-1,3,6-trisulfonic acid (ANTS, Tag XVI), for thederivatization of malto-oligosaccharides separated in HPCE which was firstdemonstrated in the derivatization of sugars separated in polyacrylamide slabgel electrophoresis.[75] ANTS not only supplies a strong chromophore but alsoprovides multiple negative charges even at low pH, which allow rapid electro-phoretic analysis. Using ANTS-derivatized glucose, Chiesa and Horváth[74]

were able to detect as little as 15 fmol at 214 nm. Using HPCE-LIF with a

31

He-Cd laser at 325 nm, the limits of detection lie in the low attomole range,three orders of magnitude lower than that of the UV detection. Other aminon-aphathalene sulfonic acid-based tags were introduced for the derivatization ofcarbohydrates by reductive amination.[65] They were the fluorescent tags5-aminonaphthalene-2-sulfonic acid (5-ANSA, tag XII) and 4-amino-5-hy-droxynapthalene-2,7-disulfonic acid (AHNS, tag XIV). A very recent develop-ment in fluorescent labeling of carbohydrates has been the introduction ofAPTS (tag XVII)[70,71,76] for the HPCE-LIF of mono- and oligosaccharides(see Part IV for more discussions).

The selective precolumn derivatization according to Scheme II, whichwas introduced recently by El Rassi’s research group for the derivatization ofcarboxylated carbohydrates,[8,54,56] offers several important features including(i) the formation of a stable amide bond between the amino group of the de-rivatizing agent and the carboxyl group of the carbohydrate molecule by acidcatalyzed removal of water in the presence of 1-ethyl-3-(3-dimethylamino-propyl) carbodiimide hydrochloride (EDAC), (ii) the selective precolumnderivatization of sialylated saccharides at room temperature, thus avoiding thecleavage of the sialic acid residue from the carbohydrate molecule being de-rivatized, (iii) quantitative yield as deduced from mass spectrometry data,[54]

and (iv) the replacement of the weak carboxylic acid group of the saccharideby one or more strong sulfonic acid groups when tagging with sulfanilic acid(SA, tag X) or 7-aminonaphthalene-1,3-disulfonic acid (ANDSA), a conditionthat allows the electrophoresis over a wider range of pH. Furthermore, usingUV absorbance at 247 nm, the detection limit of acidic monosaccharides la-beled with ANDSA or SA was 15 fmol or 30 fmol, respectively. For ANDSAderivatives, as low as 0.6 fmol could be detected with on-column fluorescencedetection using a detector operated with a 200 W xenon-mercury lamp.[54]

The labeling procedure with CBQCA (Scheme IV, tag VII) which wasfirst introduced by Novotny and co-workers[77-80] allows the detection ofsubmicromolar concentration of CBQCA-sugar derivatives by LIF. CBQCA-sugar derivatives have an excitation maximum at 456 nm, which convenientlymatches the 442 nm line of a He-Cd laser, and an emission maximum near552 nm. Using CBQCA, attomole levels of amino sugars have been analyzedby LIF-HPCE.[81] Very recently, Zhang et al.[82] utilized CBQCA for thelabeling of aminated monosaccharides. The CBQCA-sugar derivatives weredetected by LIF at the nanomolar concentration level, three orders of magni-tude lower than the limit of detection reported by Novotny and co-workers.The nanomolar limit of detection was achieved by utilizing a detection schemethat was based on a low scattering sheath flow cuvette as a postcolumn detec-tor and two photomultiplier tubes that mutually exclude wavelength ranges toprevent the water Raman band from contributing to the background system.

32

Such an arrangement had a limit of detection of 75 zeptomoles of fluorescentlylabeled 1-glucosamine.[83] One of the virtue of CBQCA derivatization is thatthe excess derivatizing agent does not have to be removed since the fluoro-genic CBQCA does not interfere with the analysis.

The derivatization according to Scheme V was used to label the six mostabundant hexoses (e.g., glucose, galactose, mannose, fucose, glucosamine,galactosamine) found in mammalian carbohydrates with TRSE, tag IX.[60] Thedetection limit of these derivatives was 100 molecules utilizing a postcolumnlaser-induced fluorescence detection in a sheath flow cuvette.[60] AlthoughTRSE is a useful fluorescent label for aminated sugar monomers, the labelingreaction requires high sugar concentration to overcome the competition be-tween the labeling reaction and the hydrolysis of the dye, thus rendering theanalysis of some real sugar samples not feasible. In fact, the dye yielded twohydrolysis products as was reported by the authors.[60] In another report, sixstructurally similar oligosaccharides were labeled according to the aforemen-tioned procedure and were baseline resolved by CE. Laser-induced fluores-cence detection of these derivatized oligosaccharides allowed a detection limitof 50 molecules.[84]

33

IV. Separation Approaches and SelectedApplications

A. Monosaccharides

A.1. Underivatized MonosaccharidesIn general, intact monosaccharides are separated by CZE using either alkalineborate buffers or high pH electrolyte systems. The borate electrolyte systemshave been shown to be compatible with direct UV detection at 195 nm[14] andRI detection.[85] The high pH electrolyte systems allowed indirect detection ofmonosaccharides by fluorescence[32] or UV[35,36,38,39] and also direct detec-tion by amperometry.[11,42-45]

As discussed above (see part II), ED of monosaccharides requires highlyalkaline pH electrolyte systems ( e.g., alkali-metal hydroxide solutions, pH 12-13)for achieving high sensitivity detection.[11,42-45] At this high pH, the monosac-charides become negatively ionized due to their weakly acidic properties(pKa of most sugars is ca. 12-13). The concentration and nature of the alkali-metal hydroxide largely affected the analysis time and the resolution ofunderivatized saccharides.[11] This separation approach was successfully usedin the determination of sugars in two common beverages.[11] When usingelectrochemical detection for the sensing of underivatized saccharides, theaddition of borate to bring about a better separation of closely related sugarswas reported to decrease the anodic response which might be due to the re-duced availability of oxidation sites present on carbohydrates.[11]

Again, and for achieving high sensitivity detection of underivatizedmonosaccharides by ED, electrolyte solutions at highly alkaline pH were usedeven for the HPCE of charged monosaccharides including glucosaminic acid,glucosamine-6-sulfate, and glucosamine-6-phosphate.[44] Separation and de-tection were achieved by using a 10 mM sodium hydroxide solution containing8 mM sodium carbonate, pH 12. This electrolyte system was shown useful inthe determination of glucose in biological samples as complex as human bloodafter 1:50 dilution (85 µM); the only sample pretreatments were centrifugationand filtration.[44] Ye and Baldwin[43] reported the separation of mixtures con-taining glucose and galactose as well as their respective alditol and aldonic,uronic, and aldaric acid derivatives using 50 mM sodium hydroxide and ED ata copper disk electrode. A higher sodium hydroxide concentration (250 mM),however, was used in order to achieve a baseline resolution of a set of eightalditols whose pKa values are in the 13-14 range. These separation and detec-tion approaches have been applied to the analysis of the sugar contents of

34

commercial apple juice as well as for monitoring the activity of glucose oxi-dase enzyme.[43]

Although they adversely affect the sensitivity of indirect UV or fluores-cence detection, highly alkaline pH electrolytes ought to be used in the HPCE-indirect photometric or fluorometric detection of underivatized saccharides toensure optimum differential electromigration. The relatively low detectionsensitivity of HPCE-indirect UV detection at highly alkaline pH is not an issuewhen the concentration of the saccharides to be determined is relatively high,such as the determination of the sugar contents of fruit juices.[39]

While with borate buffers the separation of underivatized sugars is basedon differences in the extent of complexation among the various solutes, theseparation in the presence of high-pH electrolyte systems can be explained bythe lability of the proton of the monosaccharides and by the charge-to-massratio. Using aqueous sodium hydroxide solution, pH 11.5, as the running elec-trolyte,[32] sucrose (pKa = 12.51) was detected first followed by glucose(pKa = 12.35) and then fructose (pKa = 12.03). This is the expected elutionorder when using positive polarity (anode to cathode). The stronger acid, fruc-tose, is moving at a higher velocity upstream against the EOF, thus eluting last.Glucose, a weaker acid, eluted before fructose. Sucrose, a disaccharide, movesupstream against the flow at a lower velocity than glucose due to its highermolecular weight, thus eluting first. The influences of acidity and size of themolecule were also observed with a mixture of raffinose, deoxyribose, galac-tose, glucose, and mannose.[35] They eluted in the order of increasing acidity:raffinose (a trisaccharide, pKa =12.74) < deoxyribose (pKa = 12.52) < galac-tose (pKa = 12.35) < glucose (pKa = 12.35) < mannose (pKa =12.08).

Using the proper pH for ionized acidic monosaccharides, even two chiral4-epimers, D-galactonic acid and D-gluconic acid, in an underivatized formcould be separated with a resolution of 1.2 at pH values of 4.1 to 5.0. Theaddition of a chiral selector (β-cyclodextrin) did not improve the separation ofthe two aldonic acids.

A. 2. Derivatized Monosaccharides

The separation of closely related derivatized monosaccharides by HPCE mayrequire the use of borate buffers even when the tag contains ionizable functionalgroups. For instance, twelve monosaccharides derivatized with 2-aminopyridine(2-AP) were separated as anionic borate complexes at pH 10.5 in ca. 25 min[52]

(see Figure 9). Obviously, the migration velocity of each derivative was prima-rily affected by the extent of complexation with borate. As typical examples,arabinose and ribose with cis-oriented hydroxyl groups at C3/C4 positions weremore retarded than lyxose and xylose with trans-oriented hydroxyl groups.

35

The same behavior was also observed with aldohexoses, e.g., galactose (cis) andglucose (trans), and hexuronic acids, e.g.,galacturonic acid (cis) and glucuronicacid (trans). However, N-acetylhexosamines, e.g., N-acetylgalactosamine (cis)and N-acetylglucosamine (trans), showed the reverse effect concerning the3,4-orientation of hydroxyl groups. This was attributed to the contribution of theN-acetyl substituent at the C2 position.[52] The borate buffer system was usefulin the quantitative determinations of the monosaccharide composition of variousdi- and oligosaccharides such as lactose, melibiose, rutin, digitonin, and arabicgum, which were in good agreement with the theoretical values.[52]

10 200

Time (min)

Reag

1

23

4567

8

9 10I.S.

11 12

Figure 9. Separation of N-2-pyridylglycamines derived from various reducingmonosaccharides. Electrolyte, 200 mM borate buffer, pH 10.5; applied voltage15 kV; detection UV at 240 nm. Peak assignment of parent saccharides:Reag, reagent (2-AP), 1 = N-acetylgalactosamine; 2 = lyxose; 3 = rhamnose;4 = xylose; 5 = ribose; 6 = N-acetylglucosamine; 7 = glucose; 8 = arabinose;9 = fucose; 10 = galactose; I.S. (internal standard), cinnamic acid; 11 = glu-curonic acid; 12 = galacturonic acid. Reprinted with permission.[52]

Monosaccharides derivatized with 2-AP can be migrated in electrophore-sis at low pH in the presence of noncomplexing buffer (i.e., direct CZE) sincethe 2-AP tag imparts the analytes with a positive charge arising from the proto-nation of the amino group of the 2-AP-sugar derivative. However, due to equalcharge-to-mass ratios, the 2-AP tagging of closely related monosaccharideisomers would only bring about group separation of the derivatives by directCZE, i.e., in the absence of complex formation with a suitable complexingion.[86] The positively charged 2-AP derivatives of galactose (Gal), mannose(Man), N-acetylglucosamine (GlcNAc), and N-acetylgalactosamine (GalNAc)

36

were separated into groups using phosphate buffer containing small amountsof tetrabutylammonium bromide, pH 5.0. In other words, 2-AP-Gal and 2-AP-Man, differing only in the position of hydroxyl groups, emerged together asone peak separated from the single peak of 2-AP-GalNAc and 2-AP-GlcNAc,which have in their structures an additional acetyl group. The poor resolutionof 2-AP-sugar derivatives at pH 5.0 may be due to the partial ionization ofthese rather weak bases. The 2-AP-sugar mixture was resolved when a200 mM borate buffer, pH 10.5, was used as the running electrolyte. AtpH 10.5, the 2-AP sugar derivatives become neutral, thus requiring a relativelyhigh borate concentration in order to sufficiently complex them with borateand, in turn, allow their differential migration under the influence of an appliedelectric field.

Although monosaccharides derivatized with PMP may be charged nega-tively in aqueous basic solutions due to the dissociation of the enolic hydroxylgroup, the PMP derivatives of isomeric aldopentoses or aldohexoses could notbe separated because they have the same charge-to-mass ratio.[58] With a run-ning electrolyte containing alkaline borate, the PMP derivatives of aldohexosesor aldopentoses were readily converted to anionic borate complexes and sepa-rated on the basis of the extent of complexation with the borate. The separationefficiency was quite high with a plate height of 4.1 µm for PMP-xylose. Themechanism of CZE separation of the PMP-derivatives was the same as that ofthe 2-AP derivatives, although the optimum pH for borate complexation wasshifted to 9.5 as opposed to 10.5 in the case of the 2-AP-derivatives, presum-ably due to the participation of the PMP substituent group in the complexation,perhaps through its hydroxyl groups. An electropherogram depicting a typicalseparation of a mixture of six PMP-aldohexoses of the D-series is illustrated inFigure 10. When run alone, the aldopentoses were also well separated. How-ever, the peaks of a few species of the pentose and hexose derivatives over-lapped when a mixture of pentoses and hexoses was analyzed by CZE.

Thirteen monosaccharides derivatized with CBQCA were separated byHPCE-LIF in less than 22 minutes with separation efficiencies that rangedfrom 100,000 to 400,000 per meter (see Figure 11).[77] Besides tagging thesugar molecule with a fluorogenic group, the derivatization of monosaccha-rides with CBQCA provides each sugar derivative with an ionizable weakcarboxylic acid group, thus allowing their electromigration. Therefore, theinclusion of only 10 mM borate in the running electrolyte (pH 9.4) was enoughto magnify small steric differences between closely related isomers and bringabout their separation by CZE. Also, despite the fact that the introduction of anegatively charged group would weaken the complex formation by virtue ofCoulombic repulsion, the extent of borate complexation with the hydroxylgroups of CBQCA sugar derivatives was still enough to produce differentialmigration among the various derivatized monosaccharides.

37

100 20

Time (min)

0.01

AU

FS0

met

hano

l

AB

Rea

gG

lc All

Alt

Man

Ido

Gul

Tal

Gal

Figure 10. Separation of PMP-aldohexoses by CZE. Capillary, fused silicatube, 63 cm (to the detection point), 78 cm total length × 50-µm i.d.; electro-lyte, 0.2 M borate solution, pH 9.5; voltage, 15 kV; detection, UV at 245 nm.AB, amobarbital (internal standard); Reag, excess reagent (PMP); Glc, glu-cose; All, allose; Alt, altrose; Man, mannose; Ido, idose; Gul, gulose; Tal,talose; Gal, galactose. Reprinted with permission.[58]

The concentration of borate in the running electrolyte required to bringabout the separation of a given set of monosaccharide derivatives varies withthe nature of the sugar derivatives being separated. For instance, the separationof six ANDSA-monosaccharides was best achieved in the presence of a0.10 M borate buffer, pH 10, see Figure 12.[55] The presence of two strongsulfonic acid groups in each derivative would weaken borate complexation to amuch larger extent than in the preceding case (i.e., CBQCA-sugar derivatives)due to a higher Coulombic repulsion. Under this condition, higher borate con-centration is needed to overcome Coulombic repulsive forces and bring aboutsufficient complexation with borate. The ANDSA-sugar shown in Figure 12was obtained by the new and specific precolumn derivatization reaction foracidic monosaccharides which was recently introduced by El Rassi’s labora-tory, Scheme II.[8,54,56] In addition to the improved detection sensitivity, thederivatization offered the advantage of replacing the weak carboxylic acidgroup of the sugar by the stronger sulfonic acid group of the tag, which is fully

38

ionized at all pHs. This allowed the electrophoresis of the sugar derivativesover a wide pH range and permitted the determination of acidic carbohydratesat the low femtomole levels by UV and fluorescence detection.[54]

Time (min)10 200

13 6 9

8

5

11

74

2

10

13

12

Figure 11. Electrophoretic separation of a derivatized monosaccharide mix-ture. The derivatizing agent was CBQCA. Sample concentrations were 6.2 µMfor glucosamine and galactosamine, 5.5 µM for galacturonic acid, and 4.4 µMfor other sugars. Peaks: 1 = D(+)-glucosamine; 2 = D(+)-galactosamine;3 = D-erythrose; 4 = D-ribose; 5 = D-talose; 6 = D-mannose; 7 = D-glucose;8 = D-galactose; 9 = impurity; 10 = D-galacturonic acid, 11 = D-glucuronicacid, 12 = D-glucosaminic acid; 13 = D-glucose-6-phosphate. Electrophoreticconditions were: buffer, 10 mM Na2HPO4/10 mM Na2B4O7.10H2O, pH 9.40;capillary 50-µm i.d. × 88 cm (58 cm effective length; applied voltage, 20 kV.Reprinted with permission.[77]

39

5.00 10.000.00

Time (min)

Abs

orba

nce

0.0000

0.0500

0.1000

ANDSA

1 2

3

4

56

Figure 12. Electropherogram of ANDSA derivatives of acidic monosaccha-rides obtained on a dextran 150 kDa-coated capillary. Capillary, 47 cm totallength (40 cm effective length) × 50-µm i.d.; electrolyte: 0.10 M borate,pH 10.0; pressure injection, 1 s; applied voltage, -15 kV; detection UV at250 nm. Samples: 1 = D-glucuronic acid, 2 = D-glyceric acid, 3 = D-galactonicacid, 4 = D-galacturonic acid, 5 = D-gluconic acid, 6 = N-acetylneuraminicacid. Reprinted with permission.[55]

40

A different situation was encountered in terms of the separation require-ments when the derivatized monosaccharides had a sizable triply chargedpolyaromatic tag. This was the case of ten monosaccharides includingN-acetylgalactosamine, N-acetylglucosamine, rhamnose, mannose, glucose,fructose, xylose, fucose, arabinose, and galactose which were derivatized withAPTS by the standard reductive amination procedure (Scheme I) and subse-quently separated by HPCE-LIF.[70,71] The various APTS-monosaccharidederivatives were more or less fully separated by four different buffer systemsbut, of course, with different migration patterns and selectivities. The fourbuffer systems include: 135 mM borate, pH 10.2; 100 mM acetate buffer, pH 5.0;120 mM Mops buffer, pH 7.0; and 50 mM sodium phosphate, pH 7.4.[70]

Using 135 mM borate, pH 10.2, did not resolve APTS-arabinose and APTS-fucose. These two APTS derivatives were separated at much higher borateconcentration but at the expense of much longer separation time. Again, theformation of anionic borate complexes with the APTS sugar derivatives maybe inhibited by the three negatively charged sulfonic acid groups of the APTStag. With the exception of the borate buffer where separation among the differ-ent APTS-monosaccharides is due to difference in borate complexation, theseparation in the other buffers are based on the difference in the relative stere-ochemistry of the hydroxyl groups in aldoses which ultimately determines thehydrodynamic radius of the derivatized species which is inversely proportionalto the electrophoretic mobility. Increasing the Mops buffer concentration wasshown to improve the resolution substantially; this might be due to the de-crease in the EOF caused by increasing buffer concentration. As shown inFigure 13, the order of migration of the APTS derivatives with the Mopsbuffer (positive or normal polarity) was opposite to that observed with theacetate buffer (negative or reverse polarity) and the migration of the analytestoward the detector (at the cathode end) was due to the high cathodal EOFwhich was strong enough to overcome the electrophoretic mobility of theanalytes in the opposite direction. Phosphate buffer provided a good resolutionand the migration order paralleled that obtained with the Mops buffer. In allcases, the analysis time in these buffer systems was shorter than that of theborate buffer system.[71] The tagging of sugars with APTS (tag XVII) is aninteresting development in the area of fluorescent labeling of carbohydrates forLIF detection.[70,71] Due to the higher aromaticity and number of fused ringsin APTS, APTS-sugar derivatives have substantially higher molar absorptivityand quantum efficiency than most of the commonly used fluorophore sugarderivatives. Also, the presence of three negatively charged groups seems toensure the separation of closely related structures such as those in Figure 13.

In general, and as discussed above, borate buffers find wide applicabilityin the separation of derivatized monosaccharides. An important developmentin the area of HPCE of monosaccharides has been the separation of enantio-

41

5

6

4

3

RFU

2

1

00 5 10 15

Time (min)

B10

97

64 3

21

58

A10

97

64

32

1

ba

5 8

Figure 13. (A) Electropherograms of 10 APTS-derivatized monosaccharides.Conditions: fused-silica capillary, 20-µm i.d. × 27 cm; laser source, 488 nmargon-ion, 2.5 mW; emission filter, 520 ± 20 nm and a notch filter at 488 nm;buffer, 100 mM sodium acetate, pH 5.0; outlet, anode; applied potential,25 kV/14 µA; peak identification, 1 = xylose, 2 = arabinose, 3 = ribose,4 = fucose, 5 = rhamnose, 6 = glucose, 7 = galactose, 8 = mannose,9 = N-acetylglucosamine, 10 = N-acetylglucosamine. (B) Electropherogramof 10 APTS-derivatized monosaccharides. Conditions; same as in (A). Buffer,120 Mops, pH 7.0; outlet, cathode; applied potential 25 kV/19 µA; peak identi-fication, same as in (A). Reprinted with permission.[70]

meric sugars using the principle of borate complexation in the presence ofchiral selectors. D- and L-monosaccharides derivatized according to Scheme Iwith different fluorophores, namely 2-aminopyridine (2-AP, tag IV), 5-amino-naphthalene-2-sulfonic acid (5-ANSA, tag VII) and 4-amino-5-hydroxynaph-thalene-2,3-disulfonic acid (AHNS, tag XIV), were enantiomerically separatedby capillary electrophoresis as borate complexes in the presence of linear orcyclic dextrins.[65] 5-ANSA was shown to be the most suitable tag for theenantiomeric separation of the of D- and L- forms of the carbohydrate deriva-tives when β-CD was used as the chiral selector.[65] Systematic studies on theeffects of different CDs and modified β-CDs revealed the importance of thehydroxyl groups of the CD for chiral recognition since enantioselectivity wasonly observed with the underivatized CDs and hydroxypropyl-β-CD.

42

In another report, monosaccharides were also enantiomerically separatedin a borate buffer after their derivatization with S-(-)-1-phenylethylamine[69]

(tag VI). The derivatization followed the general reductive amination proce-dure (Scheme I) and the derivatives were detected in the UV at 200 nm. Theenantiomeric separation was based on the formation of diastereoisomers by thederivatization. Optimum enantiomeric separation of 16 aldohexoses wasachieved in 50 mM borate, pH 10.3, 23% acetonitrile and 30 kV separationvoltage (see Figure 14).[69]

Other avenues have been explored in the HPCE separation of derivatizedmonosaccharides in order to provide an alternative to borate complexation. Forinstance, five PMP-monosaccharide derivatives, namely arabinose, ribose,galactose, glucose, and mannose were separated by CZE as complexes withdivalent metal ions.[5] The five PMP-monosaccharides, initially coelutingusing 100 mM sodium acetate as the background electrolyte, were fully re-solved in an electrolyte system containing 20 mM calcium acetate. The separa-tion was presumably due to the relative ease of complexation of thesederivatives with the metal ion. In fact, the PMP-pentoses eluted in the order ofincreasing complexing ability with the metal ion which depends on the orienta-tion of their hydroxyl groups. Ribose-PMP with erythro-erythro-orientedhydroxyl groups eluted first, followed by lyxose-PMP (erythro-threo) andarabinose-PMP (threo-erythro) and then xylose-PMP (threo-threo). This inter-action gives rise to a positive charge around the metal nucleus of the sugar-metal complexes and, consequently, causes a relative reduction in the totalnegativity of the sugar derivatives, a condition that favors their migration to-ward the anode at a much slower rate than the unreacted reagent PMP. How-ever, divalent metal ions have the tendency to adsorb electrostatically on thenegatively charged fused-silica wall, leading to a gradual inversion in the di-rection of the electroosmotic flow, from cathodal to anodal passing by zeroEOF, as the capillary surface charge changes from negative to positive passingby neutral with time.[5] With anodal flow (cathode to anode), the electro-phoretic migration of the derivatives is in the same direction as the EOF, acondition that leads to rapid separation. The inversion of EOF is rather slowand can take up to six hours to occur. To speed up the inversion and to ensurereproducible separations, the capillary column should be preconditioned beforeuse with a more concentrated metal ion solution than the one used as the sepa-ration electrolyte.[5] Other alkaline earth metal salts including barium, stron-tium and magnesium acetate were also investigated.[5] While the elution orderof selected PMP-aldopentoses stayed the same upon varying the nature of themetal ion in the electrolyte, as expected, the electrophoretic mobility of thesugar-metal complexes was slightly higher with Ba2+ than with Sr2+ and Ca2+,which is consistent with the fact that Ba2+ has a slightly larger ionic radius.

43

0.0000

2.0000

Abs

orba

nce

(x 1

0-3)

D-e

ry;

L-e

ry[a

]

reag

ent

D-t

hr+

L-t

hr[a

]

L-f

ul[a

]; D

-gul

D-t

alL

-tal

[a] L-g

lc[a

]; D

-glc

D-m

an;

L-m

an[a

]D

-ido L

-ido

[a];

L-a

ll[a]

D-a

ll

D-a

ltL

-alt[

a]L

-gal

[a]

D-g

al20.00 40.00 60.00 80.00

Figure 14. Separation of a mixture of eight D-aldohexoses and twoaldotetroses with rac-1-phenylethylamine. Capillary: fused-silica, 107 cm(total length), 100 cm (to detector) × 50-µm i.d.; electrolyte, 50 mM borate,pH 10.3, and 23% acetonitrile; voltage 30 kV; temperature, 25°C; UV detec-tion, 200 nm; injection, 3.0 s by pressure (3.45 kPa). In the electropherogramthe signals of derivatives of D-sugars with R-(+)-1-phenylethylamine are as-signed as L-sugars marked with [a]. They are enantiomers of the derivativesof L-sugars and S-(-)-1-phenylethylamine and have the same electrophoreticmobility in the separation system. Reprinted with permission.[69]

The separation efficiencies decreased in the following order: Ba+2 > Ca+2 >Sr+2 >> Mg+2.[5] Also, it seems that the binding of Ba2+ ions to the capillarysurface is slightly stronger than that of Ca2+, while the binding of Sr2+ is theweakest. This was manifested by a higher anodal flow in the presence ofbarium acetate electrolyte and, consequently, the separation was faster.

A given precolumn derivatization of saccharides is usually performed toallow the high-sensitivity detection of the analytes. Often, precolumn labelingmay also impart the saccharide with a charge and, in addition, improve its

44

compatibility with certain electrolyte systems. For instance, monosaccharidesare generally too hydrophilic to be solubilized in ionic surfactant-based micel-lar systems; however, their labeling with hydrophobic tags allows their separa-tion by MECC. In fact, eight different 2-aminoacridone- (2-AA, tag II) deriva-tized sugars (Scheme I), namely N-acetylglucosamine, galactose, mannose,fucose, glucose, N-acetylgalactosamine, ribose, and lyxose were separatedby MECC in the presence of an electrolyte consisting of sodium taurodeoxy-cholate and sodium borate.[66] The mechanism of migration of the 2-AA la-beled monosaccharides in the MECC system was described as complex andmay involve partition equilibria of the neutral species into the micellar surfaceand electrophoretic movement of these molecules complexed with borate.This same system was also employed in the enantiomeric separations of2-AA-galactose, -fucose and -ribose enantiomers by the addition of β-CD tothe sodium taurodeoxycholate and sodium borate buffer system. However,baseline resolution was only achieved for fucose enantiomers and the analysistime was long (ca. 43 min). The replacement of taurodeoxycholate by a non-chiral surfactant such as sodium dodecyl sulfate (SDS) did not cause the enan-tiomeric separation.[66]

Another precolumn derivatization, which seems to impart hydrophilicsaccharides with sufficient hydrophobicity to allow their separation by MECC,is PMP labeling (Scheme III, tag I).[87] This precolumn labeling, which wasfirst introduced by Honda et al.,[9] yields bis-PMP adducts that absorb stronglyin the UV. Various PMP derivatives of monosaccharides were shown to sepa-rate quite nicely in MECC in the presence of SDS at pH 7.5 using Tris-phos-phate buffer. The applicability of the SDS micellar system was extended to theidentification and quantitation of monosaccharides obtained from carbohydratehydrolyzates from glycoproteins.[87]

Fourteen different saccharides including monosaccharides were deriva-tized by reductive amination (Scheme I) with 4-aminobenzonitrile (a hydro-phobic derivatizing agent, tag III) and subsequently separated in about 5 minby MECC using an SDS micellar phase[68] (see Figure 15). The separation isbased on the differential distribution of the neutral derivatives between theaqueous mobile phase and the micellar phase. While hydrophilic carbohydratederivatives partitioned slightly inside the micelle and migrated first, hydropho-bic derivatives such as deoxyaldohexoses partitioned strongly in the micelleand migrated very slowly toward the detector. Derivatives with intermediatehydrophobicity migrated within this migration time window. The finalmigration order exhibited the strength of partition in the micellar phase and itincreased in the order ketohexoses < aldohexoses < aldopentoses < deoxyaldo-hexoses (see Figure 15).[68]

45

10 3 32

Time (min)

M

1a

2a

2b

1b

3

45

6

7

8

9

10

11

1213

14

R

Figure 15. High-speed analysis of 4-aminobenzonitrile derivatives by meansof MECC. Capillary, fused silica, 35 cm to detection, 55 cm total length× 50-µm i.d.; electrolyte, 215 mM Tris-phosphate, pH 7.5, 100 mM SDS; volt-age, 30 kV; current, 55 µA; detection. UV at 285 nm; temperature, 30°C;injection, vacuum, 1 s. Peaks: M = methanol; 1a/b = fructose, 2 a/b = sorbose;3 = lactose; 4 = melibiose; 5 = cellobiose; 6 = maltotriose; 7 = maltose;8 = mannose; 9 = glucose; 10 = galactose; 11 = ribose; 12 = lyxose;13 = arabinose; 14 = xylose; R = reagent. Reprinted with permission.[68]

B. Oligosaccharides

B.1. Underivatized Oligosaccharides

B.1.1. Simple Oligosaccharides

As their underivatized monosaccharide counterparts, simple and short un-derivatized oligosaccharides have been analyzed in CE with highly alkaline pHelectrolytes.[11,38,42] Typical examples include the analysis of simple disaccha-rides (e.g., trehalose, sucrose, lactose, lactulose, cellobiose), trisaccharides(e.g., raffinose) and tetrasaccharides (e.g., stachyose) by HPCE-ED using highpH electrolyte systems[11,42] (see Figure 7). In another approach, the disaccha-rides sucrose and maltose were analyzed at high pH by HPCE-indirect UVdetection.[38]

46

Arentoft et al.[88] demonstrated the separation of underivatized oligosac-charides of the raffinose family using borate complexes. These oligosaccha-rides are α-(1→6)-galactosides linked to C-6 of the glucose moiety of sucrose.Raffinose is the template of this homologous series with only one galactosideunit attached. By successive adding of one, two, and three additional α-galac-toside units to C-6 of the terminating galactose of raffinose, the compoundsstachyose, verbascose, and ajugose are formed. These sugars were separatedand detected as borate complexes at 195 nm. The influence of various separa-tion conditions including voltage, pH, temperature, and buffer composition onresolution, separation efficiency, migration time, and quantitative aspects wereexamined. These oligosaccharides are synthesized in various plants and accu-mulate in appreciable amounts in legume seeds.

B.1.2. Glycosaminoglycan-Derived Oligosaccharides

Glycosaminoglycans (GAGs), also called mucopolysaccharides, are un-branched polysaccharides of alternating uronic acid and hexosamine residues.They occur naturally in cartilage and other connective tissues, which are col-lectively called the ground substance. GAGs exhibit a variety of biologicalfunctions and can be altered in disease states. In GAG, while the amino groupof the hexosamine residue is either N-acetylated or N-sulfated, the uronic acidmay be either D-glucuronic acid or L-iduronic acid. Moreover, the repeatingdisaccharide units (i.e., uronic acid-glucosamine disaccharide) are O-sulfatedto varying degrees at C6 and/or C4 of the various glucosamine residues and atC2 of the uronic acid residues. Hyaluronic acid, chondroitin sulfates (chon-droitin-4-sulfate and chondroitin-6-sulfate), dermatan sulfate, keratan sulfate,heparin, and heparan sulfate are the most common GAGs.

One approach for determining structural differences among various gly-cosaminoglycans (GAGs) is to analyze their disaccharide constituents. Thisinvolves first the enzymatic degradation of GAGs with polysaccharide lyases,e.g., heparinases, chondroitinases, etc., thus yielding disaccharides bearingunsaturated uronic acids at C4-C5 which allow their direct UV detection at232 nm. Due to their ionic nature, the GAG-derived disaccharides are thenanalyzed readily by HPCE.

HPCE proved useful in the separation and quantitative determination ofthe disaccharides derived from chondroitin sulfate, dermatan sulfate, and hy-aluronic acid.[89] Exhaustive treatment of these GAGs with polysaccharidelyases released nine different disaccharides bearing unsaturated uronic acids(Figure 16) that can be detected by UV absorbance at 232 nm without priorderivatization (molar absorptivity, ε = 5000-6000 M-1 cm-1).

47

OH

OH

CH2OH

OH

OH OH

NHCOCH3

OOCOO-

OH

OX2

OX6

X4O

OH OH

NHCOCH3

OOCOO-

1 [∆di-HA]

2 [∆di-0S] where X2, X4, X6 = H3 [∆di-6S] where X2, X4 = H, X6 = SO3

-

4 [∆di-4S] where X2, X6 = H, X4 = SO3-

5 [∆di-UA2S] where X4, X6 = H, X2 = SO3-

6 [∆di-SE] where X4, X6 = SO3-, X2 = H

7 [∆di-SD] where X2, X6 = SO3-, X4 = H

8 [∆di-SB] where X2, X4 = SO3-, X6 = H

9 [∆di-triS] where X2, X4, X6 = SO3-

Figure 16. Disaccharides derived from chondroitin sulfates, dermatan sulfate,and hyaluronic acid by enzymatic depolymerization.

These disaccharides, having a net charge from -1 to -4, were well resolved byCZE using a borate buffer, pH 8.8, primarily on the basis of their net chargeand to a lesser extent on the basis of charge distribution. Despite the fact thatthe two nonsulfated disaccharides, structures 1 and 2, differed only by thechirality at the C4 in the hexosamine residue, these disaccharides were wellresolved by CZE. In another study,[90] the electrophoretic behavior of chon-droitin disaccharides (for structures, see Figure 16) was examined under vari-ous conditions including different pH, borate concentration, buffer ionicstrength, and the inclusion of SDS micelles in the running electrolytes. Al-though the disaccharides are highly charged and too polar to partition in theSDS micelles, the presence of SDS in the electrolyte improved the resolutionof the electrophoretic system. Also, the SDS-based electrolyte proved useful inthe rapid separation (ca. 15 min) of six oligosaccharides derived fromhyaluronan by digestion with testicular hyaluronidase (see Figure 17).

48

50 10 15

Time (min)

0.0015 AU

1

3

45

6

7

Figure 17. Electropherogram of oligosaccharides derived from hyaluronan bydigestion with testicular hyaluronidase. Applied voltage, 15 kV; temp., 40°C;electrolyte, 40 mM phosphate containing 40 mM SDS and 10 mM borate,pH 9.0; capillary, 72 cm × 50-µm i.d.; the column was monitored at 200 nm.Peak 1 is the unsaturated disaccharide of hyaluronan (∆di-HA), peaks 3, 4, 5,and 6 are the saturated hexa-, octa-, deca-, dodeca- and tetradecasaccharideof hyaluronan, respectively. Reprinted with permission.[90]

Also, the electrophoretic behavior of eight commercial disaccharide stan-dards derived from heparin, heparan sulfate, and derivatized heparins of thestructure ∆UA2X(1→4)-D-GlcNY6X (where ∆UA is 4-deoxy-α-L-threo-hex-4-enopyransyluronic acid, GlcN is 2-deoxy-2-aminoglucopyranose, S is sul-fate, Ac is acetate, X may be S, and Y is S or Ac) were investigated in HPCEunder various operating conditions.[91] Heparin and heparan sulfate are struc-turally similar GAGs differing primarily in their relative content of N-acetyl-glucosamine, O-sulfation, and glucuronic acid. Using heparinases I, II, and IIIas the degrading enzymes, heparin and heparan sulfate can be depolymerizedthrough an eliminative mechanism to yield 8 different disaccharides shown inFigure 18.[91] Using a borate buffer pH 8.8, two of the standard heparin/heparan sulfate disaccharides, having an identical charge of -2, ∆UA2S(1→4)-D-GlcNAc (structure 3) and ∆UA(1→4)-D-GlcNS (structure 4), were not fullyresolved. The resolution of these two saccharides could be improved by pre-paring borate buffer in deuterated water or eliminating boric acid. Surprisingly,baseline resolution was achieved in a micellar solution of sodiumdodecylsulfate (SDS) in the absence of buffer. Since the two saccharides(structures 3 and 4, see Figure 18) are charged and polar, it is unlikely that theseparation of these solutes was caused by differential partitioning in the SDSmicelles. These electrophoretic systems were then applied to the determination

49

OH

OX2

OX6

OH OH

NHY

O

O

OCOO-

1. X2 = X6 = H, Y = Ac2. X2 = H, X6 = SO3

-, Y = Ac3. X2 = SO3

-, X6 = H, Y = Ac4. X2 = X6 = H, Y = SO3

-

5. X2 = X6 = SO3-, Y = Ac

6. X2 = H, X6 = Y = SO3-

7. X4 = H, X6 = Y = SO3-

8. X2 = SO3-, X6 = H, Y = SO3

-

9. X2 = X6 = Y = SO3-

Figure 18. Disaccharide fragments obtained by enzymatic depolymerization ofheparin and heparan sulfate using heparinases I, II, and III.

of disaccharide composition of porcine mucosal heparin and that of bovinekidney heparan sulfate. Both GAGs were found to have an equimolar contentof disaccharide ∆UA2S(1→4)-D-GlcNAc (structure 3) and ∆UA(1→4)-D-GlcNS (structure 4).

An important application of HPCE in the area of GAGs has been in thequality control of natural and synthetic heparin fragments.[92] Because of theanti-blood-clotting activity of heparin, the production of natural and syntheticheparin fragments for pharmaceutical use relies on the availability of analyticalprocedures for the efficient characterization of intermediates and final prod-ucts. Using a low-pH electrolyte system, namely 0.2 M phosphate, pH 4.0, andcontrolling the capillary column temperature at 40°C allowed the separation ofthe nine most common heparin disaccharides (for structures, see Figure 18),the mapping of the oligosaccharides derived from heparin after heparinasetreatment, and the assessment of the quality of synthetic heparin pentasaccha-ride preparations.[92]

Recently, HPCE has been shown useful in determining structural differ-ences between various GAGs. For instance, heparin and heparan sulfate aretwo GAGs which are structurally related species, yet heparan sulfate has a farmore variable composition with a fewer N- and O-sulfate groups and moreN-acetyl groups. The structural differences between heparin and heparan sul-

50

fate have been determined by HPCE analysis of the disaccharides resultingfrom the action of heparinases I, II, and III on the polysaccharides.[22] Usingsodium phosphate buffer, pH 2.5, as the running electrolyte and a negative-polarity applied electric field, the resulting electropherograms of the digestedsamples were interpreted with the aid of a reference mixture of disaccharidesstandards derived from heparin.[22]

Another recent application of HPCE in GAG analysis has been the eluci-dation of the structural differences between different low-molecular-weight(LMW) heparins ( e.g., Fraxiparine, Fluxum, Fragmin, Sandoparin, andEnoxaparin) using oligosaccharide compositional analysis.[23] This was ac-complished after complete depolymerization of heparin and LMW heparinswith a mixture of heparin lyase I, II, and III followed by CE analysis using10 mM sodium borate buffer, pH 8.81, containing 50 mM SDS. According tothe authors,[23] the major mode of separation for such a system is zone electro-phoresis, while MECC mode, resulting from the presence of SDS, is a minorcontributor. As determined by CE, the oligosaccharide composition for thedifferent LMW heparins varied, suggesting that LMW heparins have a signifi-cantly different proportion of antithrombin III binding sequence, which mayexplain their different biological activity.

As discussed above, the addition of SDS, an anionic surfactant, to therunning electrolyte at a concentration above the CMC produced little change inthe migration time of disaccharides from GAG and slightly improved the over-all resolution. When SDS is replaced by cetyltrimethylammonium bromide,MECC seems to be a good choice for the separation of anionic GAG disaccha-rides [93]. Resolution improved with increasing CTAB concentration. TheGAG disaccharides eluted in the order of increasing number of charged groupsof the disaccharides. The method proved useful in the determination ofsamples of chondroitin sulfates and mink skin.

Very recently, a comparative study on compositional analysis of two setsof eight unsaturated disaccharide standards derived from heparin/heparansulfate (see Figure 18) and chondroitin/dermatan sulfate (see Figure 16) wascarried out[24] using both normal- and reverse-polarity capillary electrophore-sis. Reverse-polarity CE completely resolved disaccharide mixtures into allcomponents using a single buffer system composed of sodium phosphate,pH 3.48. At this pH, the EOF is negligible and the solutes migrate by their ownelectrophoretic mobilities toward the grounded anode. In the same report, theseparation of 13 heparin-derived oligosaccharides of sizes ranging from di- totetrasaccharides using both normal and reverse polarities was reported. Mix-tures containing oligosaccharides primarily differing in size (i.e., number ofsaccharide units) were better resolved by normal polarity[24] using 10 mMsodium borate buffer, pH 8.8, containing 50 mM SDS. This may be due to

51

some partitioning into the SDS micelles and also to the fact that the EOF is inthe opposite direction to the electrophoretic mobility of the analytes, thus slow-ing their apparent migration and, in turn, allowing a better resolution.

More recently, variously sulfated disaccharides derived from hyaluronanand chondroitin/dermatan sulfates were analyzed by CE using phosphatebuffer at low pH and negative polarity applied electric field.[27] Under theseconditions, baseline separation of the nine different sulfated disaccharides wasobtained while the two nonsulfated disaccharides exhibited peak splitting dueto the anomeric forms of the hexosamines present in the reducing terminal ofthe nonsulfated disaccharides.

Other applications of HPCE in the area of GAGs include:

(i) the assay of sulfoesterase activity on sulfated disaccharides derived fromchondroitin sulfate, dermatan sulfate, and heparin;[94] the high resolutionof capillary electrophoresis allowed the use of the assay on impure en-zyme preparations containing high protein concentrations;

(ii) The use of CE as an analytical tool for monitoring chemical reactions oftrisulfated disaccharides;[26] the reactions monitored were the acylation,pivaloylation, and benzylation of hydroxyl groups on heparin-derivedtrisulfated disaccharides; the progress of these reactions was monitoredusing the borate/SDS buffer system, pH 8.8,[23,24] and UV detection at232 nm.

While the above oligosaccharides could be readily detected at 232 nm viathe unsaturated bond in the uronic acid residues, the HPCE analysis of sulfatedsynthetic low-molecular-weight heparin fragments necessitated the use ofindirect UV.[25] This is due to the fact that the sulfated synthetic oligosaccha-rides exhibit low molar absorptivities as a result of the absence of the doublebond in their structures. The indirect UV detection involved the use of5-sulfosalisylic acid or 1,2,4-tricarboxybenzoic acid as the running electrolyteand background chromophore. Sulfated disaccharides with unsaturated uronicacid residues derived form heparin were also analyzed by CE at low pH withthe indirect-UV mode of detection. The inherent charge possessed by mostGAG disaccharides allowed CE-indirect UV detection to be conducted usingbuffers at low pH, thus eliminating the negative effect of the hydroxideions.[25] The sensitivity of indirect UV detection was reported to be at least oneorder of magnitude higher than that of direct UV detection.[25] Due to the factthat the buffer systems used for CE-direct detection (phosphate buffer, pH 2.5)and CE-indirect UV detection (i.e., 5 mM 1,2,4-tricarboxybenzoic acid,pH 3.4) were different, the resolving power of the two systems was not thesame.[22] Three of the disulfated disaccharides were not totally resolved in theindirect UV system, whereas they were resolved in the direct UV system.

52

However, two monosulfated disaccharides coeluted in the direct UV system,while they were baseline separated in the indirect UV system (see Figure 19).

1

0

2

3

4

5

6

776

5

4

3

2

1

1 2

3 45 6

78

8

10 20 30 40 50 60Migration Time (min)

-8

-9

-7

-6

-5

-4

-3

10 20 30 40 50 60Migration Time (min)

B

A

Figure 19. Capillary electrophoresis of eight heparin disaccharides usingdirect (A) or indirect (B) UV detection. Electrolytes: in (A), 200 mM sodiumphosphate, pH 2.5; in (B), 5 mM 1,2,4-tricarboxybenzoic acid, pH 3.5. Injec-tions: in (A) 9 nL from a solution containing 0.16 mg/mL of each saccharide;in (B) 1.8 nL from the same solution as in (A), except that the concentrationwas 0.1 mg/mL for each saccharide. Applied voltages: 131.5 Vcm-1 in (A)and 87.7 Vcm-1 in (B). Temperature, 25°C. Solutes: 1 = δUA2S→GlcNS6S,2 = δUA2S→GlcNS, 3 = δUA1→GlcNS6S, 4 = δUA2S→GlcNAc6S,5 = δUA2S→GlcNCOEt6S, 6 = δUA2S→GlcNAc, 7 = δUA→GlcNS,8 = δUA→GlcNAc6S. Reprinted with permission.[25]

53

B.1.3. Glycoprotein-Derived OligosaccharidesHermentin et al.[95] analyzed the reducing oligosaccharides released fromα1-acid glycoprotein (AGP) using both high-pH anion exchange chromatogra-phy with pulsed amperometric detection (HPAEC-PAD) and CZE with UVdetection at 190 nm. According to the authors, the CZE analysis proved to be4000 times more sensitive than HPAEC-PAD. In fact, the carbonyl function ofthe N-acetyl and carboxyl groups present in the molecules enabled their directUV detection at concentrations in the femtomole region. This approach has theadvantage of avoiding derivatization and sample clean up processes. In thatstudy, the authors also compared the mapping profiles of AGP glycans re-leased by conventional hydrazinolysis or by digestion with peptide-N-glycosi-dase F (PNGase F). Hydrozinolysis proved best with practically no loss ofN-acetylneuraminic acid while the PNGase F digestion resulted in partialdesialylation of the liberated N-glycans in the presence of SDS.[95]

N-linked oligosaccharides released from recombinant tissue plasminogenactivator (rt-PA) after N-glycanase digestion were separated by MECC usingSDS surfactant and direct UV detection at 200 nm.[15] The oligosaccharidesconsisted of neutral (high mannose) and mono- to tetraantennary negativelycharged oligosaccharides. As one can expect, the neutral oligosaccharidesseparated on the basis of differential partitioning into the micelles, whereas theseparation of the sialylated oligosaccharides was due mainly to differences inelectrophoretic mobilities among the negatively charged glycans. The additionof a divalent ion (Mg2+) to the SDS electrolyte system provided an effectivemeans of enhancing the selectivity of separation through both an increase ofthe migration time window of the micellar systems and the differential com-plexation of carbohydrates with the divalent metal ion.[15] This electrolytesystem was further utilized in the N-glycosylation mapping of rt-PA[16] todetermine the difference between the oligosaccharides distribution of the twort-PA variants which differ by the presence (type I) or the absence (type II) ofoligosaccharides at the Asn-184 site.

N-Oligosaccharides from fetuin, tissue plasminogen activator (t-PA) andα1-acid glycoprotein (AGP) were separated by CZE on the basis of their sialicacid content and their structures.[96] The monosialylated fraction obtained formion-exchange chromatography eluted first, followed by di-, tri- and tetrasialy-lated glycans. As the number of sialic acid residues increased in the oligosac-charides, the UV absorbance at 200 nm of the underivatized analytes wasgreatly enhanced. Within each group of sialylated N-glycans a significantseparation was still attainable, indicating that the separation relies not only on acharge difference but also on a structural difference between sugar chainsbearing the same number of sialic acid residues. Variations in the type of link-age (α-2,4 or α-2,6) between the sialic acids and the galactose, in the oligosac-

54

charide size, or in the peripheral fucose residue may have facilitated the sepa-ration within each class.

Approximately eighty underivatized sialooligosaccharides derived fromglycoproteins were analyzed by HPCE at 194 nm,[17] and a carbohydrate-mapping database was established which would enable a carbohydrate struc-tural analysis by simple comparison of migration times. Reproduciblemigration times could be achieved (RSD < 0.20%) when mesityl oxide andsialic acid were included as two internal standards for the correction of migra-tion time using a triple-correction method. The suitability and reliability of thedatabase for the structural determination of sialylated N-glycans by compari-son of corrected migration time was established by analyzing N-glycan poolsof various glycoproteins such as recombinant human urinary erythropoietin(baby hamster kidney), bovine serum fetuin, and α1-acid glycoprotein.[17]

Since there is no enzymatic cleaving process available for the cleavage ofO-linked oligosaccharides, the chemical process used for their cleavage(i.e., treatment with alkali in the presence of borohydride) results in reducingthem to alditols. Thus, the released O-linked oligosaccharides lack a site forfluorescent labeling by reductive amination and their detection is only possibleby measuring UV absorbance at 185 nm.[19] Using this detection approach,several O-glycosidically linked monosialooligosaccharides were analyzed astheir alditols by HPCE.[19] Alkaline borate buffer yielded a migration profile forthe oligosaccharides that was basically similar to that obtained in alkaline phos-phate buffer, indicating no significant contribution of borate complex formation.However, neither electrolyte provided enough resolving power to separateN-acetyl and N-glycolylneuraminic acid containing oligosaccharide pairs. Theywere only resolved after the addition of 100 mM SDS to the borate buffer (seeFigure 20). The separation mechanism is based on changing the conformation ofthese oligosaccharides, thus resulting in variation of the molecular size.[19] Thisbuffer system was utilized in a microscale analysis of sialooligosaccharides inbovine submaxillary mucin and swallow nest material.[19]

B.2. Derivatized Oligosaccharides

B.2.1. Simple Oligosaccharides

Several disaccharides including gentibiose, maltose, lactose, cellobiose andmelibiose were labeled with APTS by reductive amination and subsequentlyseparated by CE using Mops and borate buffers.[70] In Mops buffer, thegentibiose disaccharides migrated first followed by maltose, lactose, cello-biose, and melibiose. This migration order might be governed by the differ-ences in the hydrodynamic volume arising from varying degrees of hydrationdue to varying positions of hydroxyl groups in the nonreducing end pyra-nose.[70] In borate buffer, the migration order was gentibiose, maltose,

55

20 4 6 8 10 12

0.00

1 A

U

6 53 4

12

NeuAcα26GalNAc-OH

NeuAcα2

GlcNAcβ1

Galβ(1→4)GlcNAc(β1→6)

NeuAcα(2→3)Gal(β1→3)

GalNAc-OH

GalNAc-OH

Galβ(1→4)Galβ(1→4)GlcNAc(β1→6)

NeuAcα(2→3)Gal(β1→3)

GalNAc(α1→3)GalNAc-OH

63

1

3

NeuGcα26GalNAc-OH

NeuGcα2

GlcNAcβ1

GalNAc-OH63

2

4

5

6

Figure 20. Separation of an equimolar mixture of sialooligosaccharide alditolstandards. Peak numbers corresponds to the structures included with the fig-ure. Analytical conditions: capillary, fused silica (50 cm × 50-µm i.d.); electro-lyte, 200 mM borate buffer, pH 9.6, containing 0.1 M SDS; applied voltage,17 kV; detection UV absorbance at 185 nm. Reprinted with permission.[19]

melibiose, cellobiose, and lactose. This migration order is dictated by the mag-nitude of the stability constant of the disaccharide-borate complexation. In thesame report, two APTS-derivatized glucose tetrasaccharide isomers differingonly in one linkage at the nonreducing end were well resolved in Mops bufferbut not in borate buffer. This means that borate forms a weak complex withboth isomers and thus has no significant effect on the relative electrophoreticmobility of each species. These isomers were maltotetraose-APTS [glc-α-(1-4)glc-α-(1-4)glc-α-(1-4)glc-APTS] and glc-α-(1-6)glc-α-(1-4)glc-α-(1-4)glc-APTS.

56

Recently, HPCE-LIF was applied to monitor enzyme products formedduring the incubation of yeast cells with the trisaccharide α-D-Glc(1→2)α-D-Glc(1→3)α-D-Glc-O(CH2)8CONHCH2CH2NHCO-tetramethylrhodamine(-TMR). TMR is the fluorescent arm attached to the trisaccharide throughreductive amination of the analyte.[84] After 5 hr of incubating the yeast cellswith the trisaccharide, the lysed yeast spheroplasts were injected and the com-ponents were separated and detected by HPCE-LIF. Most of this trisaccharidewas converted to linker arm, intermediate disaccharide and monosaccharide(see Figure 21). This resulted from the sequential activity of α-glucosidase Iand II inside the yeast cell which act specifically on α-D-Glc(1→2) and α-D-Glc(1→3) linkages, respectively.[84]

B.2.2. Homologous Oligosaccharides

The separation of derivatized homologous, ionic oligosaccharides, or homolo-gous oligosaccharides labeled with an ionic tag by HPCE is most often accom-plished in the presence of regular noncomplexing electrolytes, e.g., phosphate,MES, Tris, etc. This is because the members of both types of derivatized ho-mologous oligosaccharides will exhibit significant differences in the charge-to-mass ratios among each other, thus ensuring sufficient differential migrationand, in turn, separation. This is only true up to a certain degree of polymeriza-tion.

The high resolving power of CZE in the separation of oligosaccharideswas first demonstrated by Nashabeh and El Rassi[86] who reported the separa-tion of the pyridylamino derivatives of maltooligosaccharides having a degreeof polymerization (d.p.) from 4 to 7 using untreated fused-silica capillaries.The positively charged sugar derivatives migrated ahead of the EOF markerand were separated according to their size in the pH range 3.0-4.5 using0.1 M phosphate solutions as the running electrolytes. The inclusion of 50 mMtetrabutylammonium bromide in the electrolyte solution decreased slightly theEOF, and consequently allowed the separation of the maltooligosaccharides atpH 5.0. However, as the pH approached the pKa value of the derivatives(pKa= 6.71), the homologues practically coeluted and moved virtually to-gether with the EOF.

To examine the effect of the nature of the derivatizing agent on the spac-ing pattern between the migrating zones of homologues, a series ofN-acetylchitooligosaccharides derivatized with either 2-AP or 6-AQ wereseparated by CZE[61] using the buffer system established for the maltooligo-saccharides[86] as the running electrolyte, and a capillary having polyetherinterlocked coating (see Figure 22). As can be seen in Figure 22, since 2-Apand 6-AQ have similar characteristic charges, the spacing between two neigh-boring homologues is virtually independent of the tagging agent.

57

0.5

1.0

1.5

2.0

2.5

13.0 13.0 13.0 13.0 13.0

DM

T D

ML

T

L

Migration Time (min)

Fluo

resc

ence

Sig

nal (

volt)

Figure 21. Electropherograms obtained from the analysis of lysed yeastspheroplasts (top) and a standard solution containing 10-9 M of each compo-nent (bottom). Electrolyte, 10 mM each of phosphate, borate, phenylboronicacid and SDS, pH 9.3; capillary, 42 cm × 10-µm i.d.; voltage, 400 V/cm.Peaks:T = α-D-Glc(1→2)α-D-Glc(1→3)α-D-Glc-O(CH2)8CONHCH2CH2NHCO-TMR;D= α-D-Glc(1→3)α-D-Glc-O(CH2)8CONHCH2CH2NHCO-TMR;M = α-D-Glc-O(CH2)8CONHCH2CH2NHCO-TMR;L = H-O(CH2)8CONHCH2CH2NHCO-TMR. Reprinted with permission.[84]

58

10 20 3000

0.001

Time (min)

10 20 3000

0.001

Abs

orba

nce

at 2

40 n

m

1

2 34 5 6

n=

n=

2-AP

A6-AQ

12

3

4

56

B

Figure 22. Electropherograms of pyridylamino (A) and quinolylamino(B) derivatives of N-acetylchitooligosaccharides. Capillary, fused silica tubewith polyether interlocked coating on the inner walls, 50 cm (to the detectionpoint), 80 cm total length × 50 µm i.d.; electrolyte, 0.1 M phosphate solutioncontaining 50 mM tetrabutylammonium bromide, pH 5.0; voltage, 18 kV.2-Ap = 2-aminopyridine; 6-AQ = 6-aminoquinoline. Reprinted with permis-sion.[61]

On the contrary, when various tagging agents with different characteristiccharges were used, the spacing between the oligosaccharides was largelyinfluenced by the nature of the derivatizing agent.[97,98,99] In one study, threederivatizing agents were used in the tagging of a sample of dextran oligomersincluding 2-AP, 5-ANSA, and ANTS.[97] At the pH of the experiment,i.e., pH 8.65, the 2-AP, 5-ANSA, and ANTS derivatives possess the negativecharges of zero, one, and three sulfonic acid groups, respectively. The averagemigration times for the individual ANTS oligomers was roughly one-third ofthose observed with 2-AP derivatives, with 5-ANSA derivatives beingintermediate. In addition, the ANTS derivatives exhibited narrower peaks,greater resolution, shorter analysis time, and higher peak detection.

59

Another study examined the effects of the structure and charge of naphtha-lene sulfonic acid-based derivatizing agents, such as ANTS, ANDSA, 3-amino-naphthalene-2,7-disulfonic acid (3-ANDA, tag XV), 2-aminonaphthalene-1-sulfonic acid (2-ANSA, tag XI), and 5-ANSA, on the CE analysis of theirmaltooligosaccharides derivatives using a running electrolyte consisting ofsodium phosphate, pH 2.5, in the presence of TEA.[98] As can be seen inFigure 23a, ANTS-derivatized maltooligosaccharides were separated for up toa d.p. of more than 30 glucose units in less than 30 min. 3-ANDA-derivatizedmaltooligosaccharides showed the same resolution, yet the separation was onlyachievable up to d.p. 30 (Figure 23b). On the other hand, for 2-ANSA and5-ANSA-derivatized maltooligosaccharides (Figure 23c and 23d, respectively),only 20 components were resolved and loss of efficiency was observed, whichmight be attributed to the longer analysis time (ca. 40 min). These findingsshow the importance of having permanent multiple charges on the tag such asANTS and 3-ANDA. Although 2-ANSA and 5-ANSA tags are structural iso-mers, the migration time of 2-ANSA is almost four times lower than that of5-ANSA (see Figures 23c and d). The pKa value for the primary amino groupof 5-ANSA is higher than that of 2-ANSA due to the fact that the sulfonic acidand amino groups are further apart. This high pKa value decreases the net nega-tive charge on the molecule and, in turn, its mobility. Despite its smaller size,5-ANSA tagging agent migrated slower than the 5-ANSA oligosaccharidederivatives. This might be due to the presence of a less basic secondary aminogroup on the 5-ANSA oligosaccharide derivatives, thus possessing a higher netnegative charge and mobility than the tagging agent.

Also, the effect of the electrical charge of the fluorescent tags on the sepa-ration of negatively charged oligosaccharides, derived from partially hydro-lyzed k-carrageenan, was studied using capillaries coated with a layer of linearpolyacrylamide.[99] This was accomplished by comparing the separation ofANTS- and 6-AQ-derivatized k-carrageenan oligosaccharides. When thecharge-to-friction ratio of oligosaccharides is increased by the end-label(i.e., ANTS), the migration order is from smaller to larger oligomers, and theseparation of larger oligosaccharides is improved by using a sieving medium.The separation is based on the differences in the charge-to-mass ratio of thedifferent oligomers. The derivatizing agent, ANTS, has three negative chargesand, as a result, it will alter the charge-to-mass ratio of the analytes; however;this alteration is more significant in the case of small oligomers and diminishesas the degree of polymerization of the oligomers increases. The migrationorder can be entirely reversed when the charge-to-friction ratio is decreased bythe end-label. This is the case of oligosaccharides tagged with 6-aminoquino-line, where the migration order starts with large oligomers and proceeds tosmaller ones. This is due to the fact that the tag decreases the charge-to-massratio of the small oligomers more significantly than the larger ones.[99]

60

100 20 30 min

30

20

10

1

R

a

100 20 30 min

20

10

1

R

b

90 18 27 36 min

30

20

10

1

R

c

100 20 30 40 50 min

20

10

1

R

d

Figure 23. Electropherograms of maltooligosaccharides labeled with(a) ANTS, (b) 3-ANDSA, (c) 2-ANSA, (d) 5-ANSA. Capillary, fused -silica,50 cm (to detection point), 72 cm (total length) × 50-µm i.d.; running electro-lyte, 50 mM phosphate, 36 mM TEA, pH 2.5; voltage, -20 kV; detection, UV,235 nm; temperature, 25°C. Solutes, (R) reagent, in (c) the presence of a con-taminant in the reagent generates a second peak, also designated as R. Num-bers indicated reflect the number of glucose residues in the linearmaltooligosaccharides. Reprinted with permission.[98]

An important operating parameter that largely affects the resolution ofhomologous oligomers is the magnitude of the EOF. In fact, the use of coatedcapillaries having very low or virtually no electroosmotic flow improved theresolution of homologous oligosaccharides with a higher degree of polymer-ization. For instance, the separation of pyridylamino derivatives of oligogalac-turonide homologous series with d.p. in the range 1 to 18 was best achieved ona coated capillary having a switchable (anodal/cathodal) EOF using 0.1 Mphosphate solution, pH 6.5, as the running electrolyte since, at this pH, the

61

EOF is very low.[100] Similarly, pyridylamino derivatives of isomaltooligosac-charides were completely separated from each other, at least up to a d.p. of 20,using fused-silica capillaries in which the EOF was suppressed by chemicallycoating the capillary inner wall with linear polyacrylamide.[101]

Furthermore, in a recent study, Chiesa and Horváth[74] arrived at the con-clusion that the use of polyacrylamide gel-filled capillaries does not contributeto enhancing the separation of derivatized homologous oligomers which con-tradicts what was first shown by Novotny and coworkers.[78,102] In theirwork,[74] ANTS-derivatized maltooligosaccharides were separated at pH 2.5by open-tubular CZE and the results were compared to those reported in theliterature involving ANTS-derivatized maltooligosaccharides separated bypolyacrylamide slab gel[75] and CBQCA-derivatized maltooligosaccharidesseparated in capillaries filled with highly concentrated polyacrylamide gelaccording to the procedure described by Liu et al.[78,102] The authors con-cluded that the presence of the gel shows no enhancement in the resolution ofoligosaccharide derivatives containing at least up to 20 glucose units.[74] Onthe contrary, the mobility of the ANTS derivatives appears to be lower by afactor of 22 in the cross-linked gel than in free solution, with no change inresolution. The separation of ANTS-maltooligosaccharides in open-tubularCZE was achieved with a background electrolyte of 50 mM sodium/triethyl-ammonium phosphate buffer, pH 2.50, containing 10.8 mM triethylamine(TEA).[74] Under these conditions, nearly 30 homologues were well resolvedin less than 5 min (see Figure 24). This separation electrolyte system was usedelsewhere in the separation of ANTS-derivatized dextran and galacturonic acidladders by CE.[103] The effect of the cationic additive (TEA) was mainly at-tributed to its interaction with the capillary wall (electrostatic binding). In fact,at sufficiently high concentrations of TEA (e.g., 50 mM), an inversion in thedirection of EOF from cathodal to anodal was observed.[74] Since the electro-phoretic migration of the triply negatively charged ANTS-derivatives takesplace in the same direction, the EOF has the beneficial effect of increasing thespeed of separation. However, the resolution of the system decreased for d.p.greater than 12. This system allowed the ultrafast separation (in less than 10 s)of three short-chain maltooligosaccharides derivatized with ANTS.[74]

62

2 4 60

3020

10

ANTS10

5

0

Abs

orba

nce

at 2

14 n

m x

103

Time (min)

Figure 24. Electropherogram of ANTS-derivatized maltooligosaccharidesobtained with 50 mM sodium/triethylammonium phosphate buffer, pH 2.5,containing 10.8 mM TEA. Capillary, 270 mm × 50-µm i.d.; temperature, 25°C;voltage, 22 kV; 80 ng of sample. Reprinted with permission.[74]

Very recently, Stefansson and Novotny[97] demonstrated the separationof large oligosaccharides of dextrans (although these oligosaccharides arebranched, their behavior is first described here simply for the completeness ofthe discussion) in open-tubular capillaries coated with linear polyacrylamide(i.e., zero-EOF capillaries) using running electrolytes based on Tris-borate,pH 8.65. This work has clearly demonstrated the importance of zero-flowcapillaries in achieving high resolution and high separation efficiencies (excessof 1 million theoretical plates/m) for large oligosaccharides. Also in this work,open-tubular CZE with zero-flow capillaries was shown to be useful in sepa-rating the various oligomers of corn amylose as well as in preliminary ex-amples of applications to monitoring the action of hydrolytic and synthesizingenzymes (for further discussion, see next section).

ANTS-derivatized dextran oligosaccharides were separated by CE usingpolymer networks under various operating conditions. As a model system, ANTS-labeled wheat starch digest was analyzed by CE using 25 mM sodium acetatebuffer, pH 4.75, containing 0.5% polyethylene oxide.[104] Although the polymerconcentration used was above the entanglement threshold value, separation was notbased on a sieving mechanism. This was evident from Ferguson plots where the

63

logarithm of the electrophoretic mobility of the analytes shows no dependence onthe polymer concentration in the range 0.2-1.0%. Therefore, the separation attainedwas caused by the charge-to-mass ratio differences of the oligosaccharides and, aswas found by Chiesa and Horváth,[74] the effect of the polymer network was only toslow down the velocity of the derivatives.

As in the case of monosaccharides, oligosaccharides derivatized withneutral or weakly ionizable tags necessitate the use of high pH electrolytesolutions or borate buffers to be electrophoresed. This is the case of PMP-derivatives of homologous oligoglucans,[58] such as α-(1→3)-linked(laminara-) oligoglucans, α-(1→6)-(isomalto-) oligoglucans and β-(1→4)-(cello-) oligoglucans, which were separated as borate complexes. All thehomooligoglucans eluted in the order of decreasing size and, because of theunfavorable mass-to-charge ratio at high degrees of polymerization, the resolu-tion between the homologues decreased as the number of recurring units in-creased. As expected, the rate of migration varied among series since theextent of their complexation with borate is largely influenced by the orientationof hydroxyl groups, i.e., by the type of interglycosidic linkage of the variousoligosaccharides. In another report from the same laboratory,[5] the PMP de-rivatives of a series of isomaltooligosaccharides were separated by CZE usingan aqueous barium salt solution as the running electrolyte. The PMPisomaltooligosaccharides were separated from each other up to d.p. of 9 asopposed to d.p. of 13 in the presence of borate.[58]

B.2.3. Branched Oligosaccharides Derived from Plants

The high resolution separation that HPCE provides for homologous linearoligosaccharides was also exploited in the separation of branched heterooligo-saccharides derived from large xyloglucan polysaccharides (XGs) by enzy-matic digestion.[61] Figure 25 illustrates the CZE separation of pyridylaminoderivatives of xyloglucan oligosaccharides (2-AP-XG) obtained from cottoncell walls by cellulase digestion.[105] XGs possess a basic backbone identicalto that of cellulose, a (1→4)-β linked D-glucan. Variations in XGs are causedby the differences in the nature and distribution of xylose, galactosyl-xylose,fucosyl-galactosyl-xylose and, in some cases, arabinosyl-xylose side chains onthe glucan backbone. Cellulase, a complex of enzymes, is able to digest thebackbone of XGs after any glucosyl residue which does not subtend a sidechain, thus liberating fragments of the polymer that reflect its branching pat-terns. The peak numbering on the electropherogram (see Figure 25) reflects theelution order obtained in reversed-phase chromatography (RPC).[105] In CZE,the elution order was mainly governed by the number of sugar residues and thedegree of branching, whereas in RPC the elution order was mainly influencedby the size of the oligosaccharide and the hydrophobic character of the sugarresidues.

64

100 10 10

Time (min)

A24

0 nm

0

0.0025

2PA1

2PA2

2PA3

2PA5

2PA6

2PA8

2PA7

2PA42-AP

1

4

2 37

6

5 8

Figure 25. Capillary zone electrophoresis mapping of pyridylamino deriva-tives of xyloglucan oligosaccharides from cotton cell walls. Capillary, fused-silica tube with polyether interlocked coating on the inner walls, 50 cm (to thedetection point), 80 cm total length × 50-µm i.d.; electrolyte, 0.1 M sodiumphosphate solution containing 50 mM tetrabutylammonium bromide, pH 4.75;running voltage, 20 kV. Symbols: 2-AP, 2-aminopyridine; ●, glucose; , xy-lose; ■ , galactose; ◊, fucose. Reprinted with permission.[61]

In order to interpret the electrophoretic behavior of the various 2-AP-XGand to quantitatively describe the effects of the various sugar residues on theirelectrophoretic mobility, Nashabeh and El Rassi[61] have introduced a mobilityindexing system for the branched xyloglucan oligosaccharides with respect tothe linear pyridylamino derivatives of N-acetylchitooligosaccharides (2-PA-GlcNAcn) homologous series, shown in Figure 22. The indexing system re-vealed that the addition of a glucosyl residue to the linear core chain of theoligosaccharide showed a similar change in the mobility index decrement asthe addition of a xylosyl residue at the glucose loci and behaved as one half ofa GlcNAc residue in terms of its contribution to the electrophoretic mobility of

65

the 2-AP-XG (see structures in Figure 25). However, the addition of a galacto-syl residue to an already branched xylosyl residue exhibited less retardationthan the addition of a glucosyl or xylosyl unit to the backbone of the xylo-glucan oligosaccharide. The same observation was made about adding afucosyl residue to a branched galactosyl residue. Thus, as the molecule be-comes more branched, the addition of a sugar residue does impart a slightlysmaller decrease in its mobility. This approach may prove valuable in correlat-ing and predicting the effects of several parameters such as the nature, posi-tion, and number of sugar residues on the mobilities of complex carbohydrates.

Very recently, the ability of HPCE to distinguish complex oligosaccha-rides of very similar structures has been further exploited in the oligosaccha-ride mapping of laminarin[97] (a branched polysaccharide) after enzymaticcleavage with laminarinase, cellulase, and endoglucanases. The variousANTS-oligosaccharide maps obtained with the three different β-1,3-glucose-hydrolases show that cellulase and laminarinase seem to cause more completehydrolyses than the endoglucanases (EG I and EG II). Also, EG I and EG IIappeared to differ somewhat in their structural preferences. This shows againthe importance of the high sensitivity and high resolution of HPCE in the char-acterization of structural preferences for different enzymes in the hydrolysis ofa given polysaccharide.

Finally, oligosaccharides of α-D-glucans (amylose, amylopectin andpullulan) and β-D-glucans (exemplified by lichenan) were also derivatized withANTS and analyzed by CE to evaluate their complexity.[106] The separationcapillary consisted of a fused-silica capillary coated with a linear layer of poly-acrylamide. The oligosaccharide maps were obtained after selective debranch-ing using isoamylase, laminarinase, and cellulase enzymes using variousborate-based electrolyte systems. According to the authors, “complex glucanchains with numerous residual branches can potentially be assessed using thisoligosaccharide mapping procedure.” Since the capillary column was of thetype with reduced EOF, a baseline separation of an intact amylose sample withd.p. close to 70 could be achieved.

B.2.4. Glycosaminoglycan-Derived Oligosaccharides

Although underivatized GAG-derived di- and oligosaccharides can be readilyelectrophoresed by HPCE and detected in the UV at 232 nm, precolumn de-rivatization of these species with a suitable chromophore or fluorophore willcertainly improve their detectability. In addition, the tagging may provideadditional properties to the GAG sugars which may enhance the separationpotential.

66

Thus far, only two research papers have appeared on the separation ofderivatized GAG-derived oligosaccharides. Unsaturated disaccharides derivedfrom GAGs by digestion with chondroitinase AC and ABC were labeled withPMP and subsequently separated by CZE.[57] The conversion of these saccha-rides to their PMP derivatives improved the system sensitivity (see Figure 26).The electrophoretic system also proved suitable for the quantitative estimationof human urinary chondroitin sulfates (see Figure 26b).

0.00

50

AU

Time (min)

0 1

a0.

005

0A

Ub

1

2

3

4

56

7

12

3

5

6

7

8

9

2 3 4 5 6 7

Time (min)

0 105

Figure 26. (a) Electropherogram of chondroitin ABC-digested mixture ofchondroitin sulfates A-E, chondroitin, and hyaluronic acid by CZE after de-rivatization with PMP. Capillary, fused silica (51 cm × 50-µm i.d.); electrolyte,100 mM borate buffer, pH 9.0; applied voltage 25 kV. Peaks: 1 came from thebuffer for enzymatic digestion; 2, PMP (excess reagent); 3, PMP derivative of∆di-0S; 4, PMP derivative of ∆di-HA; 5, sodium benzoate (internal standard);6, PMP derivative of ∆di-4S; 7, PMP derivative of ∆di-6S; 8, PMP derivativeof ∆di-SD; 9, PMP derivative of ∆di-SE. (b) Analysis of the PMP derivatives ofunsaturated disaccharides derived from the GAG fraction of a urine sampledigestion with chondroitinase ABC by CZE. Conditions and peak assignmentare as in (a). Reprinted with permission.[57]

Liu et al.[102] demonstrated the advantages of using gel-filled capillarieswith high gel concentration in the separation of enzymatically degraded hyalu-ronic acid from human umbilical cords. CZE with gel-filled capillaries af-forded the high resolution separation of hyaluronic acid-derived oligosac-charides tagged with CBQCA.

67

B.2.5. Glycoprotein-Derived Oligosaccharides

HPCE is increasingly used in the separation and mapping of the oligosaccha-ride fragments of glycoproteins, i.e., glycans. Using ovalbumin as a modelprotein, Honda et al. [101] demonstrated the separation of glycans by CZE withon-column fluorometric detection. The oligosaccharides of ovalbumin (hybridand high mannose type) were released with anhydrous hydrazine and taggedwith 2-AP after re-N-acetylation. As shown in Figure 27, the oligosaccharideswere electrophoresed using two different electrolytes, an acidic phosphatebuffer whereby the derivatized glycans are positively charged due to the proto-nation of the amino group of the tag (i.e., direct CZE) and an alkaline boratebuffer which allows the in situ conversion of the derivatives to anionic boratecomplexes (i.e., indirect CZE). Because of differences in their separationmechanisms, direct and indirect CZE yielded different selectivities. As can beseen in Figure 27a, direct CZE gave good separation among the oligosaccha-ride derivatives that are different in their molecular size (i.e., oligosaccharideshaving different number of monosaccharide units), but could not resolve sol-utes having the same degree of polymerization. On the other hand, CZE asborate complexes separated the oligosaccharide derivatives based on structuraldifferences of the outer monosaccharide residues (Figure 27b). The greater thenumber of unsubstituted mannose units, the more retarded are the deriva-tives.[101] However, the borate system failed to resolve high mannose typeoligosaccharides having the same number of outer mannose residues, such asstructures h1 and h2 (see Figure 27). Similarly, hybrid-type oligosaccharideshaving the same number of peripheral mannose or galactose residues, butdiffering in the total number of monosaccharide units, were not resolved (seestructures g1 and g2 in Figure 27). In both cases, however, the use of directCZE with phosphate buffers gave satisfactory separations.

Also, capillary electrophoresis proved to be useful in elucidating the dif-ferences in glycan structures of the same glycoprotein but from differentsources.[107] The CZE mapping of the pyridylamino derivatives of the oli-gosaccharides derived from human and bovine α1-acidglycoprotein (AGP)yielded two different electropherograms each containing well-defined peaksand a few minor peaks, eluting after the excess 2-AP (see Figure 28). Bothhuman and bovine AGP have been found to have the same sialic acid, galac-tose, and mannose content. The major differences are such that 50% of thesialic acid in bovine AGP are N-glycolylneuraminic acid and the fucose con-tent is very low. These differences were unveiled by CZE mapping of bothtypes of glycans. Based on these results, CZE will play an important role inthe field of glycan separation and characterization.

68

Manα1–6

Manβ1–4GLcNAcβ1–4GlcNAcManα1–3

Manα1

Manα16

3i

Manα1Manα1–3

GLcNAcβ1–4Manβ14GlcNAcβ1–4GlcNAc

Manα1GLcNAcβ1–2

6

3

a

Manα1

Manα1GLcNAcβ14Manβ14GlcNAcβ1–4GlcNAc

Manα1–3

GLcNAcβ1–4

GLcNAcβ1–2 d

6

3

Manα1–6

Manα1–3Manα1

GLcNAcβ1–4

GLcNAcβ1–2Manα1

Manβ1–4GlcNAcβ1–4GlcNAc6

3

f

Manα1–3Manα1

GLcNAcβ1–4Manβ1–4GlcNAcβ1–4GlcNAc

GLcNAcβ1–2

Galβ1–4GlcNAcβ14Manα1

63

e

Manα1–6

Manα1–3Manα1

GLcNAcβ1–4Manβ1–4GlcNAcβ1–4GlcNAc6

3Manα1

GLcNAcβ1–2 g1

Manα1–6

Manα1–3

Manα1–2Manα1

Manα16

3Manβ1–4GLcNAcβ1–4GlcNAc

h1

Manα1–2Manα1–6

Manα1–3

Manα1–2Manα1

Manα16

3Manβ1–4GLcNAcβ1–4GlcNAc

h2

Manα1–6Manα1

Manα1–3 6GLcNAcβ1–4Manβ1–4GlcNAcβ1–4GlcNAc

3Galβ1–4GlcNAcβ1–4

G;cMAcβ1–2Manα1

g2

a b

AP2

3

1

4

5

0 10 20

Time (min)

0 10 20

Time (min)

5 15 25

AP G-AP

abc

de

f

g

h

i

Figure 27. Analysis of the reductively pyridylaminated oligosaccharides de-rived from ovalbumin (a) by direct CZE or (b) CZE as borate complexes.(a) Capillary, 20 cm × 25 µm i.d.; electrolyte, 0.1 M phosphate, pH 2.5; run-ning voltage, 8 kV; detection, 240 nm. (b) Capillary, 95 cm × 50-µm i.d.; elec-trolyte, 200 mM borate buffer, pH 10.5; running voltage, 20 kV; fluorescencedetection with excitation at 316 nm and emission at 395 nm. Peak assignmentsin (a): 1, heptasaccharide; 2, octasaccharides; 3, nonasaccharide; 4, deca-saccharide; 5, undecasaccharide. Peak assignments in (b) refer to the illustra-tion provided above the electropherograms; peaks b and c were not assigned.Reprinted with permission.[101]

69

0.004

0.008

Human

0 20 40

Time (min)

0.004

0.008

0 20 40

Time (min)

Abs

orba

nce

at 2

40 n

ma Bovine

2-AP

2-AP

b

Figure 28. Capillary zone electrophoresis mapping of pyridylamino deriva-tives of human (a) and bovine (b) AGP oligosaccharides. Capillary, fused-silica tube with hydrophilic coating on the inner walls, 45 cm (to the detectionpoint), 80 cm total length × 50-µm i.d.; electrolyte, 0.1 M phosphate solution,pH 5.0, containing 50 mM tetrabutylammonium bromide; running voltage,18 kV; current. 80 µA; injection by electromigration for 2 s at 18 kV. Reprintedwith permission.[107]

High mannose oligosaccharides released from RNase B by digesting theprotein with PNGase F were also analyzed by CZE[61] as their 2-AP deriva-tives. One major peak on the CZE map was identified as (GlcNAc)2-Man5using an oligosaccharide standard. The polypeptide chain of RNase B isknown to have only one glycosylation site which can accommodate five differ-ent high mannose glycans. This may explain the presence of several peaks inthe CZE map besides that of (GlcNAc)2-Man5 oligosaccharide which is themost predominant carbohydrate moiety of bovine RNase B.

Very recently, an electrophoretic system based on capillary gel electro-phoresis was described for the profiling of oligosaccharides enzymaticallyderived from ribonuclease B and labeled with APTS. The capillary gelelectrophoresis system involving the use of an entangled polymer networkexhibited a higher resolving power than open-tubular CZE.[76] Optimumseparation was attained using 25 mM acetate buffer, pH 4.75, containing0.4% polyethylene oxide polymer and a neutrally coated capillary. This systemyielded a high resolution of all the major components of the ribonuclease

70

N-glycan pool, and a baseline separation of the three positional isomers ofmannose-7 and mannose-8 oligosaccharides (see Figure 29). According to theauthors, the separation is not related to sieving effect but rather to the changein the hydrodynamic volumes of the labeled glycans and viscosity of theseparation medium.[76]

To compare capillary to planar electrophoresis, various carbohydratesreleased from several glycoproteins including bovine fetuin, human α1-acidglycoprotein, HIV envelope, and bovine ribonuclease B were labeled withANTS and analyzed by HPCE-LIF and high-concentration polyacrylamideslab gel electrophoresis (PAGE). HPCE in the open-tubular format (i.e., in theabsence of an entangled polymer network) using acetate buffer, pH 4.75,yielded comparable results to those obtained with PAGE in terms of number ofmigrating bands. In both cases, high resolution separations of the released andlabeled carbohydrates were achieved.[108]

Moreover, ANTS-derivatized complex oligosaccharides, both neutral andsialylated, were separated by HPCE-LIF.[109] The separation of the derivativeswith good resolution was achieved in less than 8 min using phosphate buffer atpH 2.5. The linear relationship between the electrophoretic mobility andcharge-to-mass ratios of the ANTS-derivatized oligosaccharides was used forpeak assignment.

The effect of the structure and charge of the derivatizing agent on the CEanalysis of branched oligosaccharides was also investigated.[98] High-man-nose-type oligosaccharides from bovine pancreatic ribonuclease B were la-beled with ANTS, ANDSA, and 2-ANSA, and separated by CZE using phos-phate buffer, pH 2.5, containing TEA. A baseline resolution of the maincomponents was obtained in all cases. The resolution of two of the three struc-tural isomers of Man7 indicates that the mechanism of separation is not strictlybased on differences in charge-to-mass ratio but also on the three-dimensionalstructure which also affects electrophoretic mobility.[98]

The nature of the derivatizing agent influences the choice of the HPCEmode to be used in the subsequent separation step. In fact, MECC proveduseful in the analysis of complex oligosaccharides derivatized with a hydro-phobic tag, 2-aminoacridone. The derivatized oligosaccharides derived fromribonuclease B, hen egg albumin, and fetuin were separated[110] using a boratebuffer containing taurodeoxycholate surfactant that has been shown useful forthe analysis of 2-aminoacridone-derivatized monosaccharides.[66] The separa-tion mechanism of such a system is based more on the borate complex forma-tion, as determined by the absence of resolution when bicarbonate buffer wasused, than on the partitioning of the derivatives in the micellar phase. How-ever, the addition of the surfactant was shown to improve the separation effi-ciency. The pattern of separation of the major components of the 2-AA

71

derivatized oligosaccharides derived from ovalbumin is similar to that obtainedby gel permeation.[110]

11.0

6.0

1.0

10.00 12.00

Time (min)

14.00

Man-9Man-8Man-7Man-6Man-5Rib-B

Rel

. Uni

t (R

U, x

1E-0

1)

Man-5

Man-6

Man-8

Man-9

Man-7

6 6 43 3

6 6 43 3

2

6 6 43 3

22

6 6 43 3

222

6 6 43 3

22

22

Figure 29. CGE separation of the APTS-labeled high-mannose type oligosac-charides released from bovine ribonuclease B (upper trace) and the individualstandard structures (lower traces). Inset: structural representation of thehigh-mannose type N-linked oligosaccharides: squares = GlcNAcβ1→4;circles #4 = Manβ1→4; circles #6 = Manα1→6; circles #3 = Manα1→3;circles #2 = Manα1→2. Conditions: 57 cm neutrally coated column (50 cmto detection point) × 50-µm i.d.; LIF detection: argon ion laser, excitation488 nm, emission: 520 nm; separation buffer: 25 mM acetate buffer, pH 4.75,0.4% polyethylene oxide, applied electric field = 500 V/cm.; i = 19 µA; 20˚C.Reprinted with permission.[76]

Additional CZE analysis of glycans was reported by Liu et al.[77] N-linkedoligosaccharides from bovine fetuin were released through hydrazinolysis andthen derivatized with CBQCA. With on-column LIF detection, this tag permit-ted the CZE of subpicogram amounts in a phosphate-borate buffer, pH 9.5.Four major peaks as well as few minor peaks were well resolved.

Finally, to provide an HPCE methodology that makes it possible to drawinferences about structural characteristics of complex glycans and, in turn,expedite subsequent structural analyses, a novel method for identifying andquantifying 2-AP derivatives of desialylated N-glycosidically linked oligosac-charides in glycoproteins was introduced. It is based on two-dimensional map-

72

ping of different oligosaccharides by HPCE[111,112] using two different elec-trolyte systems so that the individual modes of separation are as different fromeach other as possible.

B.2.6. Glycolipid-Derived Oligosaccharides

Very recently, Mechref et al.[56] reported the most suitable conditions for theselective precolumn derivatization of sialooligosaccharides, derived fromgangliosides, and the subsequent separation of the derivatives by HPCE.Seven sialooligosaccharides, whose structures and abbreviations are illustratedin Figure 1, were cleaved from gangliosides by ceramideglycanase and deriva-tized with ANDSA according to Scheme II.[8,54] This precolumn derivatiza-tion, which involves the formation of a stable amide bond between the aminogroup of the ANDSA tag and the carboxylic acid group of the analyte, is veryattractive for the labeling of sialooligosaccharides since it is readily achieved inan aqueous medium and at room temperature. Sialooligosaccharides are proneto desialylation at high temperature and in very acidic media. The ANDSA-sialooligosaccharide derivatives, which fluoresce at 420 nm when excited at315 nm, were readily detected in HPCE at the low femtomole levels using anon-column lamp-operated fluorescence detector. The various ANDSA-sialooligosaccharide derivatives are charged at all pH due to the fact that theprecolumn derivatization with ANDSA replaces each weak carboxylic acidgroup of the parent sugar by two strong sulfonic acid groups. Using 100 mMsodium phosphate, pH 6.0, as the running electrolyte and an untreated fused-silica capillary allowed the resolution of six of the sialooligosaccharides inves-tigated. The two structural isomers sialooligo-GD1a and sialooligo-GD1b (seeFigure 1) were not resolved, suggesting that their hydrodynamic volumes arenot significantly different. The separation of the seven ANDSA-sialooligosac-charides with an uncoated fused-silica capillary was best achieved when75 mM borate, pH 10.0, was used as the running electrolyte (see Figure 1).[56]

In another report by Mechref and El Rassi,[55] the seven ANDSA-sialooligosaccharides were perfectly separated in a dextran-coated capillaryusing 100 mM sodium phosphate buffer, pH 6.0, and a negative polarity (seeFigure 30). The dextran-coated capillary used in that study exhibited a reducedEOF with respect to an untreated fused silica capillary and the reduced EOFwas in the opposite direction to the intrinsic electrophoretic mobility of theanalytes. This condition could have favored a better differential migration ofthe two structural isomers.

73

0.0100

0.00501

2

3

4

5

6

7

Abs

orba

nce

0.0000

7.50 10.00

Time (min)

15.00

Figure 30. Electropherogram of ANDSA derivatives of sialooligosaccharidesderived from gangliosides obtained on a dextran 150 kDa-coated capillary.Capillary, 47 cm total length (40 cm effective length) × 50-µm i.d.; runningelectrolyte: 0.10 M phosphate, pH 7.0; pressure injection, 1 s; applied voltage,-15 kV; detection, UV at 250 nm. Sample: 1 = Sialooligo-GD3; 2 = Sialooligo-GM3; 3 = Sialooligo-GT1b; 4 = Sialooligo-GD1b; 5 = Sialooligo-GD1a;6 = Sialooligo-GM2; 7 = Sialooligo-GM1. Reprinted with permission.[55]

C. PolysaccharidesC.1. Underivatized PolysaccharidesThus far, only a few attempts have been made for the application of HPCE topolysaccharides. Recently, Richmond and Yeung[34] reported the HPCE sepa-ration and detection of high-molecular-weight native polysaccharides (seeFigure 5). To partially ionize the various analytes and in turn achieve differen-tial electromigration, an electrolyte of pH 11.5 was used. Detection was madepossible by laser-excited indirect fluorescence detection. In general, migrationtimes were reproducible to 0.05 min for consecutive injections.[34]

74

HPCE was utilized for the quantitative analysis of the glycosaminoglycanhyaluronan in human and bovine vitreous with UV detection at 200 nm.[20] Arunning electrolyte consisting of 50 mM disodium hydrogen phosphate,10 mM sodium tetraborate, and 40 mM SDS, pH 9.0, was found to be opti-mum for assaying hyaluronan and its oligomers. This alkaline electrolyte en-sures that hyaluronan migrates as a polyanion and minimizes the possibility ofwall adsorption of both protein and polysaccharide components of the vitreoushumor, while SDS binds to the proteins which then migrate away fromhyaluronan. The signal corresponding to hyaluronan was confirmed by depoly-merization of the native mucopolysaccharide by hyaluronidase. This resultedin the loss of the hyaluronan peak and the appearance of several new peakscorresponding to the oligomeric fragments which had shorter migration times(see Figure 31).

C.2. Derivatized PolysaccharidesSudor and Novotny[80] reported the separation of neutral polysaccharides(e.g., chitosan, dextran, various water-soluble cellulose derivatives, etc.) labeledwith CBQCA using capillaries coated with polyacrylamide and filled with anappropriate polymer solution such as linear polyacrylamide. In the presence ofborate, the derivatized polysaccharides migrated readily in open-tubular CZEbut showed little tendency to separate. Due to their large molecular size, poly-saccharides were not amenable to sieving using gel capillary electrophoresis asthey did not penetrate the gel network. This hindrance was solved by causingthe polysaccharides to migrate through solutions of entangled polymers. Butthe extent of separation in such sieving media was complicated by reptationeffect. This was overcome by using pulsed-field conditions where a potentialgradient along the separation capillary was periodically inverted at a 180°angle which brought about shape transitions and, in turn, favored the separa-tion of polysaccharides according to molecular size. This approach resembleswhat is known as pulsed-field electrophoresis or field-inversion gel electro-phoresis which was originally introduced for the separation of large DNAfragments in traditional gel slab electrophoresis.

The electrophoretic migration of neutral and highly charged polysaccha-rides, such as chemically modified celluloses and heparins labeled withCBQCA, was regulated by secondary thermodynamic equilibria during capil-lary electrophoresis by using suitable buffer additives.[113] Electrophoreticmigration of uncharged chemically modified cellulose was induced by theadsorption of hydrophobic and charged detergents onto the analyte compo-nents, and the process could be described by Langmuir adsorption isotherms.Different polymers were found to contain different adsorption sites, includingmultilayer formation. On the other hand, reduction of a high electrophoreticmobility of highly charged heparin was attained by the addition of ion-pairing

75

reagents to the running electrolyte. Under these conditions, migration velocityand selectivity were influenced by the concentration and number of charges ofthe additive.[113]

0.02

0.01

0.0010 12 14

Time (min)

Abs

orba

nce

b

a

Figure 31. Electropherogram of (a) bovine vitreous, compared with (b) hyalu-ronidase digest of Hyaluronan. Peaks labeled according to the number ofdisaccharide units in the oligomer. Separation buffer: 50 mM disodium hydro-gen phosphate, 40 mM SDS, 10 mM sodium tetraborate; capillary, 50 cm ×75-µm i.d.; voltage, 15 kV; detection at 200 nm. Reprinted with permission.[20]

Another approach for the electrophoresis and sensitive detection ofpolysaccharides has been the dynamic or in situ labeling of the analytes. Infact, underivatized starch components were separated and detected by CE astheir iodine complexes.[51] Iodine complexation with carbohydrates impartscharge and optical detection sensitivity at 560 nm. The starch-iodine complexconsists of a helix of sugar residues surrounding a linear I-5 core.[114,115] Un-like borate complexation, iodine binding is a cooperative interaction whichexhibits strong chain-length dependence in both complexation and opticalproperties. The iodine binding constant increases nearly exponentially withglucan chain length, reaching a plateau at approximately 125 residues.[116]

Moreover, the wavelength of maximum absorbance exhibits a red shift withincreasing chain length.[117] These facts could be utilized to reveal informationon the size and structure of the analytes independent of their electrophoreticbehavior. Amylopectin and amylose were well resolved from each other in lessthan 10 min using iodine-containing buffers in unmodified capillaries. Amy-lopectin electrophoretic mobility is dependent on the iodine concentration as

76

well as on the separation temperature, while that of amylose is not. This re-flects the fact that the long chains of amylose are essentially saturated withiodine even at low iodine concentration. This system was also shown to beuseful in the analysis of potato starch, amizo V, corn starch, and malto-dextrin.[51]

D. Glycopeptides and GlycoproteinsD.1. GlycopeptidesThe high resolving power and unique selectivity of CZE were also exploited inthe separation and characterization of peptide and glycopeptide fragments ofglycoproteins. Figure 32 illustrates the CZE mapping of the tryptic peptidefragments of human α1-acid glycoprotein (AGP) as well as the submapping ofits glycosylated and nonglycosylated fragments.[107] Prior to CZE runs, thewhole digest was first fractionated into peptide and glycopeptide fragments byhigh-performance affinity chromatography on a silica-bound concanavalin A(Con A) column. Three pooled fractions were obtained, the first two beingCon A non-reactive and Con A slightly reactive eluted with 20 mM phosphatepH 6.5 containing 0.1 M NaCl, while the third fraction interacted strongly withthe Con A column and eluted with the heptanic sugar, i.e., methyl-α-D-manno-pyranoside. The three fractions were then analyzed by CZE using a fused-silica capillary with fuzzy polyether coating and 0.1 M phosphate, pH 5.0, asthe running electrolyte. As seen in Figure 32, the CZE mapping of the wholedigest reveals the microheterogeneity of the glycoprotein as manifested by theexcessive number of peaks.

One of the unique aspects of the primary structure of the polypeptidechain of pooled human AGP is its peculiar structural polymorphism. Substitu-tions were found at 21 of the 181 amino acids in the single polypeptide chain,which is responsible in part for multiple peptide and glycopeptide fragments inthe tryptic digest. Another source of multiple fragments in the tryptic digest isthe microheterogeneities of the oligosaccharide chains attached, which can benoticed by inspecting Figure 33. In fact, the variation in the terminal sialic acidcauses charge heterogeneity in the glycopeptide fragments cleaved at the samelocation by trypsin, the differences in the extent of glycosylation among apopulation of the protein molecules lead to fragments having the same peptidebackbone but with or without carbohydrate chains, and the variation in thenature of the oligosaccharide chains at each glycosylation site yields severalglycopeptides that have the same peptide backbone but are different in theiroligosaccharide structures.[107]

77

0 20 40

Time (min)

60

Abs

orba

nce

at 2

00 n

m

Wholedigest

0

0'

1

2

A BB2 C2C1 C3B1

C

Figure 32. Capillary zone electrophoresis tryptic mapping and submapping ofhuman AGP. Capillary, fused-silica tube with hydrophilic coating on the innerwalls, 45 cm (to the detection point), 80 cm total length × 50-µm i.d.; electrolyte,0.1 M phosphate solution, pH 5.0; running voltage, 22.5 kV; injection by electromi-gration for 4 s at 22.5 kV. Symbols: fraction 0, con A non-reactive (excluded fromthe column); fraction 0', Con A non-reactive (unretained by the column); fraction 1,Con A slightly reactive (eluted with buffer); fraction 2, Con A strongly reactive(eluted with the haptenic sugar). Reprinted with permission.[107]

As can be seen in Figure 32, the CZE submapping of Con A reactivepeptides produced peaks that are missing from the submaps of all other col-lected fractions, i.e., 0, 0' and 1, but whose components are found in the wholemap (see area C1, Figure 32). This approach allows the monitoring of a groupof peptides as well as the assessment of glycosylated fragments in the wholemap. This methodology is expected to work also with other glycoproteins, andthe CZE submapping of all the glycosylated tryptic fragments with differenttypes of glycans may require the use of more than one lectin column in theprefractionation step.[107]

78

Also, HPCE has been shown to be useful in evaluating the glycopeptidemicroheterogeneity of recombinant human erythropoietin (rHuEPO) expressedfrom Chinese hamster ovary (CHO) cells. This was achieved by using100 mM heptanesulfonic acid (ion-pairing agent) in 40 mM sodium phosphatebuffer, pH 2.5.[118] The negatively charged heptanesulfonic acid forms ionpairs with basic amino acid residues, thus reducing analyte-wall interaction aswell as altering analyte electrophoretic mobility. This led to improved peptideresolution, which allowed the evaluation of the heterogeneity of glycopeptidesderived from rHuEPO. The total tryptic map exhibited two regions: nongly-cosylated and glycosylated peptides. Since rHuEPO glycoprotein has threeglycosylation sites, three glycopeptides are expected to result from trypticdigestion of this glycoprotein. However, the aforementioned electrophoreticsystem revealed the microheterogeneity of these glycopeptides by yielding atleast 12 glycopeptide peaks in the peptide map.[118]

CFC

BFB

A

Fucα1

Galβ1-4GlcNAcβ1

Galβ1-4GlcNAcβ1-2Manα1

3

4

∗

∗Galβ1-4GlcNAcβ1-2Manα1

Ga;β1-4G;cMAcα1

Manβ1–4GlcNAcβ1–4GlcNAcβ1-Asn36

6

Figure 33. Primary structure of the carbohydrate classes of AGP. There are fivecarbohydrate classes attached to AGP having different degrees of branching andsialylation. Classes A, B and C are the bi-, tri- and tetraantennary complex N-linkedglycans, respectively, whereas BF and CF are the fucosylated B and C structures.Two additional glycans exist: one has two additional fucose linked to the GlcNAcresidues marked with asterisks, and one has an outer chain prolonged by Galβ1-4GlcNAc at either of the Gal residues marked with an arrow.

Very recently, Weber et al.[46] reported the characterization of glycopep-tides from recombinant coagulation factor VIIa by HPLC and HPCE usingUV and pulsed electrochemical detection (PED). The combination of the moretraditional methods of HPLC-UV and HPAEC-PED with capillary electro-phoresis methods based on CZE-UV and CZE-PED allowed a better character-ization of the glycopeptides’ heterogeneity. In addition, this report demon-strated the potential of CZE-PED in the analysis of PNGase F-treated glyco-peptides where, in contrast to UV detection, both the peptides and the releasedcarbohydrates can be detected simultaneously.[46]

79

D.2. Glycoprotein GlycoformsProtein glycosylation can occur at two or more positions in the amino acid se-quence, and the glycans at even a single position may be heterogeneous or maybe missing from some molecules. This leads to populations of glycosylated vari-ants of a single protein, usually referred to as glycoforms, whose relative propor-tions are found to be reproducible and not random. However, the glycoformsmay be affected by several factors including the environment in which the pro-tein is glycosylated, the manufacturing process, and the isolation procedures.This would affect the function of a glycoprotein, thus engendering the need forhigh-resolution separation methods to allow the monitoring of glycoform popula-tions especially for genetically engineered glycoprotein pharmaceuticals.

Several HPCE approaches have been described for profiling glycoproteinglycoforms. Transferrin glycoforms were separated by CZE and CIF.[119] Inboth modes of HPCE, at least five components corresponding to the di-, tri-,tetra-, penta-, and hexasialo-transferrins differing from each other by one nega-tive charge were resolved. The capillary columns used in this study werecoated with a layer of linear polyacrylamide on the inner wall to suppress EOFand consequently provide better resolution and sharper focusing of the closelyrelated glycoforms by CZE and CIF, respectively. To assess the presence ofvarying degree of sialylation among the various glycoforms, the action patternof neuraminidase on the electrophoretic behavior of the various isoforms wasmonitored by HPCE. Neuraminidase is an exoglycosidase that liberates spe-cifically the negatively charged sialic acids from the terminal non-reducingpositions in glycans. As shown in Figure 34, the electrophoretic analyses ofsamples taken from the enzymatic digestion at various time intervals demon-strated the gradual removal of sialic acid as manifested by the changes in therelative proportions of the different isoforms with time. The electrophoreticpattern of the final product was completely different from the starting materialand showed one main component, the asialo-transferrin. Thus, in the case oftransferrin, the major source of microheterogeneity seems to be the variation inthe terminal sialic acid of the glycans.

The microheterogeneity of glycoproteins results from sugar residues otherthan the sialic acid residues. This is the case of ribonuclease B (RNase B)whose glycans portions are of the high-mannose type. The separation of the fivedifferent glycoforms of RNase B has been achieved through the formation ofanionic borate complexes with the hydroxyl groups of the glycan moiety (Fig-ure 35).[120] The relative proportions of the various glycoforms correlated withthe relative proportions of the high mannose glycan populations, i.e., Man9-Man5, determined by other more established analytical methods, e.g., massspectrometry, high performance anion exchange chromatography (HPAEC)and size exclusion chromatography on Bio-Gel P4. To further substantiate thepresence of the various glycoforms of RNase B, the time course for the diges-

80

tion of the protein with A. saitoi α(1-2) mannosidase was monitored by CZE.Mannosidase is an exoglycosidase that specifically cleaves mannose from thenon-reducing end of glycans. Figure 35b shows that after 25 hrs the glycoformpopulations carrying Man9-Man6 structures were all reduced to a single popula-tion carrying Man5. Thus, HPCE offers a direct method for analyzingglycoforms at the protein level with high resolution and precision.

0 2 4 6 0 2 4 6

66 2

3

5

0' 1'

0.01

AU

a b

0.01

AU

0 2 4 6 0

00

2 4 6

2

2

2

3

3

3

1

14

4

45

5

5

10' 15'

0.01

AU

c d

0.01

AU

0 2 4 6 0

0

2

30

2 4 6

2

1

12

3

25' 45'

0.01

AU

e f

0.01

AU

0 2 4 6 0

2 2

11

2 4 6 0 2 4 6

0 0

1

0

2

200' 500' 1200'

0.01

AU

g h i

0.01

AU

0.01

AU

Time (min)Figure 34. CZE of iron-free transferrin following incubation with neuramini-dase. Capillary, 18.5 cm × 50-µm i.d.; electrolyte, 18 mM Tris-18 mM boricacid-0.3 mM EDTA, pH 8.4; running voltage, 8 kV. The samples for electro-phoresis were taken after various incubation times: (a) 0; (b) 1; (c) 10; (d) 15;(e) 25; (f) 45; (g) 200; (h) 500; (i) 1200 min. The proportions of the transferrinisoforms (asialo, mono-, di-, trisialo, etc., marked 0, 1, 2, 3, etc.) changed withtime. The sample taken after 20 hr still contained transferrin molecules havingone and two sialic acids. The small peak (labeled with a star) appeared after50-80 min, but did not increase in size on prolonged incubation time (g-i).Reprinted with permission.[119]

81

12 13 14 15Time (min)

Abs

orba

nce

Man

9M

an8

Man

7M

an6

Man

5

Ribonuclease A

a b

12 131110Time (min)

Abs

orba

nce

units

(λ=

200)

0 hr

1 hr

24 hr

48 hr

2 hr

6 hr

72 hr

96 hr

12 131110

Figure 35. (a) CZE profile of ribonuclease showing the nonglycosylated formof the protein, ribonuclease A, and the glycoforms of the same protein, collec-tively known as ribonuclease B. RNase A is a contaminant of RNase B as sup-plied by Sigma. RNase B remained unaffected during the digestion of theoligosaccharide component of RNase B with A. saitoi α(1-2) mannosidase.(b) CZE profile of RNase B showing the time course for the digestion of theglycoprotein with the exoglycosidase, A. saitoi α(1-2) mannosidase. Capillary,fused silica 72 cm × 75-µm i.d.; applied voltage 1 kV for 1 min and 20 kV for19 min, temp., 30°C; detection, UV at 200 nm; injection 1.5 s; electrolyte,20 mM phosphate containing 50 mM sodium dodecylsulfate, 5 mM borate,pH 7.2. Reprinted with permission.[120]

Another example of high-mannose related microheterogeneity has beenthe various glycoforms of recombinant human bone morphogenetic protein 2(rhBMP-2).[121] The separation of rhBMP-2 glycoforms by CE necessitatedthe use of simple phosphate buffer containing no additives. Under this condi-tion, the rhBMP-2 sample yielded nine peaks which have been identified to beglycoforms of rhBMP-2. The difference between any adjacent peaks is onlyone mannose residue (Mr = 162). The nine peaks obtained with intactrhBMP-2 reduced into one major peak when the endo-H digested rhBMP-2was analyzed by CE. Endo-H is an endoglycosidase specific for the cleavageof high-mannose glycans from glycoproteins. This confirms that the microhet-erogeneity of the glycoprotein is due to the high mannose-type carbohydrates.The migration order of the glycoforms was found to follow the increasingnumber of mannose residues in the analyte molecules. Mannose residue canaffect the separation by providing higher friction coefficient rather than by acharge shielding effect.[121]

82

Phosphorylation of glycans seems to be another source of glycoproteinmicroheterogeneity. In fact, CE analysis of proteinase A glycoforms, bothnative and underglycosylated, revealed charge heterogeneities attributedto differences in the phosphorylation level of the carbohydrate moiety atAsn-68.[122] The CE analysis was performed using an electrolyte consistingof 100 mM acetate/phosphate buffer, pH 3.2. Both forms were separated intothree distinct peaks that probably correspond to charge heterogeneities dueto differences in carbohydrate phosphorylation.

The microheterogeneity of a glycoprotein is mostly the result of glycosy-lation and is largely unaffected by the presence of other functionalities in theprotein. This is the case of ovalbumin, a phosphorylated glycoprotein. Thisprotein has one asparagine residue that can accommodate at least nine differentcarbohydrate structures of the high mannose and hybrid type N-glycans. Thereare also two potential phosphorylation sites at two serine residues, one at posi-tion 68 and the other at position 344. The various glycoforms were separatedvia borate complex formation with the hydroxyl groups of the carbohydratemoieties of the protein using an untreated fused silica capillary.[123] This gly-coprotein is a strongly acidic species and therefore would not undergo adsorp-tion onto the naked capillary surface when using alkaline borate. To improvethe resolution of the ovalbumin glycoforms, putrescine (i.e., 1,4-butanediamine),a doubly charged cationic species, was added in small amounts (1 mM) to theborate buffer. Using these conditions, five major protein peaks were separated,indicating the presence of protein glycoforms. Upon dephosphorylation of theglycoprotein with calf intestinal alkaline phosphatase or potato acid phos-phatase, the five peaks were still resolved but shifted in the position to a morerapid migration time, a behavior consistent with a loss of negative charge.Based on this observation, it was suggested that all ovalbumin glycoforms arephosphorylated to the same degree, and heterogeneity among ovalbuminisoforms resides solely in the carbohydrate structures. Also, the same electro-phoretic system was shown to permit the separation of pepsin glycoforms.

Realizing the benefit of including an amine additive (e.g., 1,4-butane-diamine, DAB) into the running electrolyte in achieving the separation ofovalbumin glycoforms prompted the investigation of other amine additivessuch as α,ω-bis-quaternary ammonium alkanes.[124] Three of the α,ω-bis-quaternary ammonium alkane additives, namely hexamethonium bromide(C6MetBr), hexamethonium chloride (C6MetCl), and decamethonium bromide(C10MetBr), were examined and the results were compared to those obtainedusing DAB as a buffer additive.[124] The alkyl chain length and the cationgroup of α,ω-bis-quaternary ammonium alkanes strongly influence the analy-sis time and resolution.[124] Under identical separation conditions, C10MetBrwas shown to yield a better resolution and shorter analysis time than C6MetBr.Originally, with DAB, it was thought that such an additive exerts its effect by

83

mainly altering the EOF.[123] The fact that the additives repressed EOF simi-larly but the quaternary ammonium alkane additives allowed the resolution ofovalbumin glycoforms in half the time required with DAB suggests that themechanism of action of these additives is not solely related to EOF repression.Other mechanisms may be involved including protein-additive interactions,protein-wall interactions, additive-wall coating interactions, or any combina-tion of these. In the same report, seven of the eight glycoforms of human chori-onic gonadotropin (hCG) were resolved by CE using 1 mM C6MetBr and25 mM borate pH 8.4.[124] However, the eight glycoforms of the hCG werenear-baseline resolved using 25 mM borate and 5 mM diaminopropane andseparated in less than 50 min.[125]

Along the above separation strategies which include (i) the employmentof either borate complexation or amine additives and (ii) the use of specific en-zymes directed either toward the glycan moieties or other protein functional-ities, HPCE was applied successfully to the analysis of recombinant glyco-protein glycoforms. In fact, the fractionation of the human recombinant tissueplasminogen activator (rt-PA) by CZE[126] was recently reported. The CZEanalysis of two main glycosylation variants (type I and II) of the same glyco-protein showed different electrophoretic migration patterns. The study furtherelucidated the microheterogeneity of the glycoprotein as was manifested by thepartial resolution of almost 15 glycoforms in a protein that has only four pos-sible N-glycosylation sites. This report compared the CZE profile of an rt-PAsample to that of a desialylated rt-PA obtained through neuraminidase treat-ment, an approach similar to that introduced by Kilàr and Hjertén for humantransferrin.[119] The desialylated rt-PA exhibited a much simpler CZE profileindicating that the glycoprotein microheterogeneity is mostly the result of dif-ferent levels of sialylation. Along the same lines, Watson and Yao[127] ex-tended the use of CZE to the separation of glycoforms of recombinant humangranulocyte-colony-stimulating factor (rhGCSF) produced in Chinese hamsterovary cells. This glycoprotein contains only two O-linked carbohydrate moi-eties that differ only in having one or two sialic acid residues. Due to its rela-tive simplicity compared to other more complex glycoproteins, the rhGCSFyielded two wellresolved and equally sized peaks using phosphate-boratebuffer, pH 8.0, and an untreated fused-silica capillary. Under these conditions,the acidic protein was repelled from the negatively charged capillary wall andno apparent solute adsorption was observed. When the glycoforms were incu-bated with neuraminidase, a single peak was obtained eluting earlier than ei-ther of the original two sialylated glycoforms.

The high selectivity of CZE was also demonstrated in the separation of theglycoforms of recombinant human erythropoietin (rHuEPO),[128,129] a glyco-protein hormone produced in the kidney of adult mammals that acts on bone

84

marrow erythroid progenitor cells to promote their development into matureblood cells.

James and co-worker[130] demonstrated the usefulness of MECC in re-solving recombinant interferon-γ (IFN-γ) glycoforms produced by Chinesehamster ovary cells. Separations were performed in uncoated fused-silicacapillaries at alkaline pH in the presence of SDS micelles. Optimal separationwas obtained with 400 mM borate buffer containing 100 mM SDS, pH 8.5. Itwas noted that optimum separation of IFN-γ glycoforms occurred at a pH closeto the pI of the protein (8.5-9.0) since it is most susceptible to hydrophobicinteraction at this pH. However, this electrolyte system did not resolve bovineserum fetuin nor α1-acid glycoprotein glycoforms and showed partial resolu-tion for ribonuclease B and horseradish peroxidase glycoforms.[130]

Very recently, a combination of HPLC and HPCE was found useful in theanalysis of the glycoforms of human recombinant factor VIIa (rFVIIa).[131]

Again the use of DAB was found to be essential for the separation of the vari-ous glycoforms. The separation is reported to be based primarily upon thedifferent content of N-acetylneuraminic acid of the oligosaccharide structuresof rFVIIa.

Another application of HPCE in the area of glycoforms has been the iden-tification and determination of the isoforms of monoclonal F(ab')2 fragmentobtained after pepsin proteolysis of IgG.[132] The presence of these isoforms isusually attributed to post-translational modification of the IgG molecule. It wasfound that the variation in the pH of the background electrolyte can be effec-tively employed in modulating selectivity of isoform separations with opti-mum resolution observed at alkaline pH (9.50).

E. GlycolipidsRecently, the potentials of HPCE have also been demonstrated in the separa-tion of gangliosides, the sialic acid-containing glycosphingolipids.[56,133,134]

As shown in Figure 36, a ganglioside molecule has a hydrophilic sialooligo-saccharide chain and a hydrophobic moiety, i.e. ceramide, that consists of asphingosine and fatty acid. Due to this inherent structural feature, the ganglio-sides are amphiphilic solutes forming stable micelles in aqueous solutions withvery low critical micellar concentration values (10-8-10-10 M).[135] Further-more, gangliosides most often exist at low concentrations and their structureslack strong chromophores. Thus, two major obstacles must be overcome whendeveloping an HPCE method for the analysis of gangliosides, namely to beable to (i) separate them as monomers and (ii) detect them at low levels. Thefirst obstacle has been addressed recently by Yoo et al.[134] and further elabo-rated by Mechref et al.[56] while the second obstacle was overcome byMechref et al.[56]

85

GM1

GM2

GM3

Galactose N-Acetylgalactosamine

GalactoseAcNH

CH2OH

CH2OH

CH2OH

CH2OHGlucose

HOOH

OHOHHOH2C HO

AcNH N-Acetylneuraminic acid

OR2

O

-OOC

OO

O

O

OOH

O

O

R1O

OH

O

OH

O

H

C C C HH2 NH CH

CH

CH3 CH3

3)16

O C

(CH (CH2)12

Stearic Acid SphingosineGM1, R1 = R2=HGM2, R2 = HGM3, R2 = R2=HGD1a, R1 = N-Acetylneuraminic acid (NeuNac) , R2 = HGD1b, R1 = H, R2 = NeuNAcGD3, = GM3, R2 = NeuNAcGT1b, R1= R2 = NeuNAc

Figure 36. Structures of the gangliosides. Reprinted with permission.[56]

On the other hand, Liu and Chan[133] applied HPCE to the study of the behav-ior of gangliosides in aqueous solutions. These researchers[133] demonstratedthat CZE can separate some ganglioside micelles and consequently permittedstudies of the micellar properties of these amphiphilic species using untreatedfused-silica capillaries and on-column direct UV detection at 195 nm. Theganglioside micelles were successfully analyzed within 10 min with masssensitivity in the order of 10-11 mol. Baseline resolution of a mixture of threeganglioside micelles, namely, GM1, GD1b, and GT1b, was achieved using2.5 mM potassium phosphate, pH 7.40, as the running electrolyte (see Fig-ure 37). The separation was mainly facilitated by the varying content of thesialic acid residues in the ganglioside micelles, which imparted them withdifferent electrophoretic mobilities.

86

0 2 4 6 8 10 0 2 4 6 8 10

a

b

GT1b

GD1b

GM1

GT1b

GD1b

GM1

150 min

90 min

60 min

30 min

0 min

Time (min)

Abs

orba

nce

(195

nm

)

Figure 37. (a) Capillary electrophoresis of a mixture of GM1, GD1b, and GT1b.Individual GM1, GD1b, and GT1b micelles and a mixture of these three ganglio-sides shortly after mixing were analyzed by CE. The buffer was 2.5 mM potas-sium phosphate, pH 7.40. Detection was by UV at 195 nm. (b) Time course ofmixed micelle formation between GD1b and GT1b. Equimolar concentrations ofpolysialogangliosides GD1b and GT1b (165 µM) in 2.5 mM potassium phos-phate, pH 7.40, were mixed by vortexing and incubated in a water bath at37°C. At time intervals, the electrophoretic patterns of the ganglioside mix-tures were analyzed. Electrophoretic conditions are as in (a). Reprinted withpermission.[133]

87

However, upon incubation at 37°C, complete fusion between both micellarpeaks could be observed in less than 2.5 hrs (see Figure 37b). The fusion pro-cess was accelerated by raising the incubation temperature. Also, it was foundthat polysialogangliosides (e.g., GD1a and GT1b) may have higher propensitiesfor micellar fusion than monosialogangliosides (e.g., GM1 and GM2). Thus thehigh resolution, high speed, and quantitative aspects of CZE were clearly dem-onstrated in monitoring processes that may have important implications in thedistribution and function of gangliosides in biological membranes.

One of the many elegant features of HPCE is the ease with which theelectrophoretic behavior of the analytes can be altered through the simpleaddition of specific reagents to the running electrolyte. Among the many bufferadditives described so far,[136] two kinds of additives are suitable for HPCE ofgangliosides, namely cyclodextrins (CDs) and acetonitrile (ACN). While CDscan alter the electrophoretic behavior of a wide variety of compounds viainclusion complexes, the extent of which is determined by the solute hydro-phobicity and size, ACN improves analyte solubility as well as selectivity andcontrols electroosmotic flow.[136] On this basis, Yoo et al.[134] demonstratedthe utility of CDs in the separation of native gangliosides at 185 nm. Amongthe various CDs, α-CD was the best buffer additive as far as the separation isconcerned. This may be due to the size of the cavity of α-CD that best fit thelipid moiety of the gangliosides.

Although the gangliosides could be detected at 185 nm,[134] the sensitivityis rather low to allow their detection in biological matrices where they are nor-mally found in minute amounts. To overcome this difficulty, Mechref et al.[56]

have expanded the utility of the selective precolumn derivatization procedurewhich has been developed originally for the tagging of carboxylated monosac-charides[54] (see Scheme II) to include the derivatization of gangliosides in orderto improve their detectability. Moreover, novel electrolyte systems were also intro-duced for the analysis of derivatized gangliosides by HPCE. The derivatizationinvolved the tagging of gangliosides with either SA (a UV-absorbing tag) orANDSA (a UV absorbing and also fluorescent tag). The derivatization wasshown to occur uniformly on the sialic acid residues by monitoring the enzy-matic digestion of SA-GT1b with neura-minidase. The SA-GT1b yielded threepeaks upon digestion, which corresponded to be SA-GM1, SA-GD1b and SA-GT1b, thus indicating that all three sialic acid residues of GT1b were labeled.[56]

To separate the derivatized gangliosides in their monomeric forms, ACNand CD were added to the running electrolyte to break up ganglioside micelles.As can be seen in Figure 38, HPLC-grade acetonitrile at 50% (v/v) broughtabout the separation of three ANDSA-derivatized gangliosides in their mono-meric forms while, in the absence of ACN, the three ganglioside derivatives

88

(i.e., ANDSA-GM1, ANDSA-GD1a, and ANDSA-GT1b) migrated as a singlebroad peak.[56]

The advantage of using ACN, whose UV cutoff is at 185 nm, was alsoillustrated in the ability to separate and detect underivatized gangliosides at195 nm.[56] The use of ACN allowed the separation of the differently sialy-lated gangliosides shown in Figure 38; however, it did not provide enoughselectivity to cause the separation of structural isomers ANDSA-GD1a andANDSA-GD1b. Partial separation was attained using α-CD and 100 mM so-dium borate, pH 10.0.[56] Other additives such as polyvinyl alcohol and hy-droxypropyl cellulose brought about a better resolution of the two isomers, i.e.,ANDSA-GD1a and ANDSA-GD1b. Complete baseline separation of the twodisialylated ganglioside isomers was attained by utilizing an in situ chargedmicellar system composed of decanoyl-N-methylglucamide (MEGA 10)/borate in the presence of α-CD[56] (see Figure 39).

A24

7

0.00

16 A

.U.

A24

7

0.00

32 A

.U.

1,2,31,2,3

1

X X

X

X

a

Time (min) Time (min)

b c d

2

3

1

2 3

20 20 10 20

Figure 38. Electropherograms of standard gangliosides labeled with ANDSAat neutral (a and b) and high (c and d) pH in the presence (b and d) and ab-sence (a and c) of acetonitrile in the running electrolyte. In (a) and (b): run-ning electrolyte, 25 mM sodium phosphate, pH 7.0, at 0% (a) and 50% v/v(b) acetonitrile; voltage, 25.0 kV. In (c) and (d): running electrolyte, 10 mMsodium phosphate, pH 10.0, at 0% (c) and 50% v/v (d) acetonitrile; voltage,20 kV; capillary, fused-silica, 50 cm (to detection point), 80 cm (total length)× 50-µm i.d. Solutes, (1) GM1, (2) GDa, (3) GT1b. For structures refer to Fig-ure 36. Reprinted with permission.[56]

89

A24

7

0.00

32 A

.U.

X

Time (min)

12 3

10

Figure 39. Electropherogram of standard ANDSA-1. Running electrolyte,50 mM borate, pH 6.0, containing 5.0 mM MEGA surfactant and 15.0 mMα-CD; running voltage, 18.0 kV; capillary, fused-silica, 50 cm (to detectionpoint), 80 cm (total length) × 50-µm i.d. Peaks: 1 = GD1a; 2 = GD1b; 3 = GD3.Reprinted with permission.[56]

90

V. ConclusionsAs this review reveals, HPCE in its various modes of separation and detectionis suitable for the analysis of a wide variety of carbohydrate species includingmono-, oligo-, and polysaccharides, glycopeptides, glycoproteins, and glycolipids.This was facilitated in part by the progress made in the capillary column tech-nology and by the introduction of novel electrolyte systems. However, themajor factors behind the advances made in HPCE of carbohydrates have beenthe development of various detection systems and approaches which include

(i) indirect UV and LIF detection,(ii) electrochemical detection, and(iii) precolumn labeling with suitable chromophores and fluorophores for the

sensitive UV and fluorescence detection, respectively.

In surveying the literature, precolumn derivatization seems to be the mostelegant approach for the separation and detection of carbohydrates. In fact, fivedifferent reaction schemes have been introduced for the labeling of carbohy-drates with various kinds of tags. All of these tagging processes have yieldedthe sensitivity required for the analysis of carbohydrates at moderate and lowlevels. Furthermore, the tagging of carbohydrates imparted the derivatizedcarbohydrates with charges and/or hydrophobic functional groups that facili-tated their efficient separation by various separation principles, thus leading tovarying degree of selectivity. It should be noted that multiply charged tagssuch as ANDSA, ANTS and APTS are excellent labels for the HPCE of sugarderivatives not only because of their high detection sensitivity by either UV orLIF but also because they yield derivatives that are readily separated by HPCE.Although major progress has been made in the area of precolumn labeling ofcarbohydrates, there is still room for improvements regarding the introductionof other tagging agents and optimizing the existing reaction schemes.

Although noncomplexing electrolyte systems have found some use in theelectrophoresis of a wide range of carbohydrate species, the bulk of HPCEseparation is still accomplished primarily by borate complexation regardlesswhether the carbohydrates are derivatized or underivatized. Borate complex-ation magnifies small structural differences among closely related carbohy-drates, thus leading to a better resolution for multicomponent mixtures.

The ease with which the electrolyte systems can be modified and tailoredto fit a given separation problem is another important feature of HPCE. In fact,glycolipids such as gangliosides which are not compatible with purely aqueouselectrolyte solutions were readily separated in their monomeric forms in hydro-organic electrolyte systems.

91

Furthermore, the unique selectivity and high separation efficiencies ofHPCE have proved extremely useful in the separation of all kinds of carbohy-drate species. The technique provides an unsurpassed resolving power forprofiling and mapping closely related oligosaccharides cleaved from glycopro-teins, glycolipids, and glycosaminoglycans. This high resolving power has alsoallowed the efficient fingerprinting of complex glycoprotein glycoformswhich, in other separation techniques such as ion-exchange HPLC and tradi-tional gel electrophoresis, would yield smeared, unresolved bands.

The advantages of HPCE over other separation techniques such as HPLCand traditional polyacrylamide gel electrophoresis reside in its higher separa-tion efficiencies, shorter analysis time, small sample requirements and, moreimportantly, lower consumption of expensive reagents and solvents. With theintroduction of precolumn labeling with suitable fluorescent tags for LIF detec-tion, HPCE can reach nanomolar detection limits, thus making the techniqueextremely suitable for the analysis of minute amounts of carbohydrates. How-ever, the major drawback of the technique is its limited preparative capability.

92

VI. References1. Stryer, L. Biochemistry. 4th ed., New York, W. H. Freeman and Co.,

1995.2. El Rassi, Z. Capillary Electrophoresis of Carbohydrates. Adv.

Chromatogr. 34, 177-250 (1994)3. El Rassi, Z., Nashabeh, W. High Performance Capillary Electrophoresis

of Carbohydrates and Glycoconjugates, in Carbohydrate Analysis: HighPerformance Liquid Chromatography and Capillary Electrophoresis,267-360. Z. El Rassi, Ed., Amsterdam, Elsevier, 1995.

4. El Rassi, Z. Ed. Capillary Electrophoresis of Carbohydrate Species. Elec-trophoresis, Vol. 17, pp. 275-437 (1996)

5. Honda, S., Yamamoto, K., Suzuki, Kakehi, U. M., Kakehi, K. High-Performance Capillary Zone Electrophoresis of Carbohydrates in thePresence of Alkaline Earth Metal Ions. J. Chromatogr. 588, 327-333(1991)

6. Foster, A. B. Zone Electrophoresis of Carbohydrates. Adv. Carbohydr.Chem. 12, 81-116 (1957)

7. van Duin, M., Peters, J. A., Kieboom, A. P. G., van Bekkum, H. Studieson Borate Esters II. Tetrahedron 41, 3411-3421 (1985)

8. Mechref, Y., Ostrander, G. K., El Rassi, Z. Capillary Electrophoresis ofCarboxylated Carbohydrates. Part 2. Selective Precolumn Derivatizationof Sialooligosaccharides Derived from Gangliosides with 7-Aminonaph-thalene-1,3-disulfonic Acid Fluorescing Tag. Electrophoresis 16,1499-1504 (1995)

9. Honda, S., Akao, E., Suzuki, S., Okuda, M., Kakehi, K., Nakamura, J.High Performance Liquid Chromatography of Reducing Carbohydrates asStrongly Ultraviolet-Absorbing and Electrochemically Sensitive 1-Phe-nyl-3-methylpyrazolone Derivatives. Anal. Biochem. 180, 351-357(1989)

10. Rendleman, J. A. Ionization of Carbohydrates in the Presence of MetalHydroxides and Oxides. Adv. Chem. Ser. 117, 51-69 (1971)

11. Colón, L. A., Dadoo, R., Zare, R. N. Determination of Carbohydrates byCapillary Zone Electrophoresis with Amperometric Detection at a CopperMicroelectrode. Anal. Chem. 65, 476-481 (1993)

12. Angyal, S. J., Davies, K. P. Complexing of Sugars with Metal Ions.Chem. Commun. 10, 500-501 (1971)

13. Angyal, S. J. Complexes of Metal Cations with Carbohydrates in Solu-tion. Adv. Carbohydr. Chem. Biochem. 47, 1-43 (1989)

93

14. Hoffstetter-Kuhn, S., Paulus, A., Gassmann, E., Widmer, H. M. Influenceof Borate Complexation on the Electrophoretic Behavior of Carbohy-drates in Capillary Electrophoresis. Anal. Chem. 63, 1541-1547 (1991)

15. Taverna, M., Baillet, A., Baylocq-Ferrier, D. Analysis of Neutral andSialylated N-Linked Oligosaccharides by Micellar Electrokinetic Capil-lary Chromatography with Addition of a Divalent Cation.Chromatographia 37, 415-422 (1993)

16. Taverna, M., Baillet, A., Schlüter M., Baylocq-Ferrier, D. N-glycosyla-tion Site Mapping of Recombinant Tissue Plasminogen Activator byMicellar Electrokinetic Capillary Chromatography. Biomed. Chromatogr. 9,59-67 (1995)

17. Hermentin, P., Doenges, R., Witzel, R., Hokke, C. H., Vliegenthart, J. F. G.,Kamerling, J. P., Conradt, H. S., Nimtz, M., Brazel, D. A Strategy for theMapping of N-Glycans by High-Performance Capillary Electrophoresis.Anal. Biochem. 221, 29-41 (1994)

18. Hughes, D. E. Capillary Electrophoretic Examination of UnderivatizedO-Linked and N-Linked Oligosaccharide Mixtures and ImmunoglobulinG Antibody-Released Oligosaccharide Libraries. J. Chromatogr. B 657,315-326 (1994)

19. Kakehi, K., Susami, A., Taga, A., Suzuki, S., Honda, S. High-Perfor-mance Capillary Electrophoresis of O-Glycosidically Linked Sialic Acid-Containing Oligosaccharides in Glycoproteins as Their AlditolDerivatives with Low-Wavelength UV Monitoring. J. Chromatogr. A680, 209-215 (1994)

20. Grimshaw, J., Kane, À., Trocha-Grimshaw, J., Douglas, A., Chakravarthy,U., Archer, D. Quantitative Analysis of Hyaluronan in Vitreous HumorUsing Capillary Electrophoresis. Electrophoresis 15, 936-940 (1994)

21. Malsch, R., Harenberg, J., Heene, D. L. High-Resolution Capillary Elec-trophoresis and Polyacrylamide Gel Electrophoresis of Heparins.J. Chromatogr. A716, 259-268 (1995)

22. Damm, J. B. L., Overklift, G. T., van Dedem, G. W. K. Determination ofStructural Differences in the Glycosaminoglycan Chains of Heparin andHeparan Sulfate by Analysis of the Constituting Disaccharides with Capil-lary Electrophoresis. Pharm. Pharmacol. Lett. 3, 156-160 (1993)

23. Desai, U. R., Wang, H., Ampofo, S. A., Linhardt, R. J. OligosaccharideComposition of Heparin and Low-Molecular-Weight Heparins by Capil-lary Electrophoresis. Anal. Biochem. 213, 120-127 (1993)

24. Pervin, A., Al-Hakim, A., Linhardt, R. J. Separation of Glycosaminogly-can-Derived Oligosaccharides by Capillary Electrophoresis Using Re-verse Polarity. Anal. Biochem. 221, 182-188 (1994)

94

25. Damm, J. B. L., Overklift, G. T. Indirect UV Detection as a Non-Selec-tive Detection Method in the Qualitative and Quantitative Analysis ofHeparin Fragments by High-Performance Capillary Electrophoresis.J. Chromatogr. A 678, 151-165 (1994)

26. Kerns, R. J., Vlahov, I. R., Linhardt, R. J. Capillary Electrophoresis forMonitoring Chemical Reactions: Sulfation and Synthetic Manipulation ofSulfated Carbohydrates. Carbohydr. Res. 267, 143-152 (1995)

27. Karamanos, N. K., Axelsson, S., Vanky, P. Tzanakakis, G. N., Hjerpe, A.Determination of Hyaluronan and Galactosaminoglycan Disaccharides byHigh-Performance Capillary Electrophoresis at the Attomole Level. Ap-plications to Analyses of Tissue and Cell Culture Proteoglycans.J. Chromatogr. A 696, 295-305 (1995)

28. Yeung, E. S. Indirect Detection Methods: Looking for What Is NotThere. Acc. Chem. Res. 22, 125-130 (1989)

29. Yeung, E. S., Kuhr, W. G. Indirect Detection Methods for CapillarySeparations. Anal. Chem. 63, 275A-282A (1991)

30. Foret, F., Fanali, S., Boceck, O. L., Bocek, P. Indirect Photometric Detec-tion in Capillary Zone Electrophoresis. J. Chromatogr. 470, 299-308(1989)

31. Yeung, E. S. Optical Detectors for Capillary Electrophoresis. Adv.Chromatogr. 35, 1-51 (1995)

32. Garner, T. W., Yeung, E. S. Indirect Fluorescence Detection of SugarsSeparated by Capillary Zone Electrophoresis with Visible Laser Excita-tion. J. Chromatogr 515, 639-644 (1990)

33. Kuhr, W. G., Yeung, E. S. Optimization of Sensitivity and Separation inCapillary Zone Electrophoresis with Indirect Fluorescence Detection.Anal. Chem. 60, 2642-2646 (1988)

34. Richmond, M. D., Yeung, E. S. Development of Laser-Excited IndirectFluorescence Detection For High-Molecular-Weight Polysaccharides inCapillary Electrophoresis. Anal. Biochem. 210, 245-248 (1993)

35. Vorndran, A. E., Oefner, P. J., Scherz, H., Bonn, G. K. Indirect UV De-tection of Carbohydrates in Capillary Zone Electrophoresis.Chromatographia 33, 163-168 (1992)

36. Oefner, P. J., Vorndran, A. E., Grill, E., Huber, C., Bonn, G. K. CapillaryZone Electrophoretic Analysis of Carbohydrates by Direct and IndirectUV Detection. Chromatographia 34, 308-316 (1992)

37. Bergholdt, A., Overgaard, J., Colding, A., Frederiksen, R. B. Separationof D-Galactonic and D-Gluconic Acids by Capillary Electrophoresis.J. Chromatogr. 644, 412-415 (1993)

95

38. Xu, X., Kok, W. T., Poppe, H. Sensitive Determination of Sugars byCapillary Electrophoresis with Indirect UV Detection under Highly Alka-line Conditions. J. Chromatogr. A 716, 231-240 (1995)

39. Klockow, A., Paulus, A., Figueiredo, V., Amado, R., Widmer, H. M.Determination of Carbohydrates in Fruit Juices by Capillary Electrophore-sis and High-Performance Liquid Chromatography. J. Chromatogr. A680, 187-200 (1994)

40. Weston, A., Brown, P. R., Jandik, P., Heckenberg, A. L., Jones, W. R.Optimization of Detection Sensitivity in the Analysis of InorganicCations by Capillary Ion Electrophoresis using Photometric Detection.J. Chromatogr. 608, 395-402 (1992)

41. Rocklin, R., in A Practical Guide to HPLC Detection, 145. D. Parriott,Ed. New York, Academic Press, 1993.

42. Ye, J., Baldwin, R. P. Amperometric Detection in Capillary Electrophore-sis with Normal Size Electrodes. Anal. Chem. 65, 3525-3527 (1993)

43. Ye, J., Baldwin, R. P. Determination of Carbohydrates, Sugar Acids,Alditols by Capillary Electrophoresis and Electrochemical Detection at aCopper Electrode. J. Chromatogr. A 687, 141-148 (1994)

44. O’Shea, T. J., Lunte, S. M., LaCourse, W. R. Detection of Carbohydratesby Capillary Electrophoresis with Pulsed Amperometric Detection. Anal.Chem. 65, 948-951 (1993)

45. Lu, W., Cassidy, R. M. Pulsed Amperometric Detection of CarbohydratesSeparated by Capillary Electrophoresis. Anal. Chem. 65, 2878-2881(1993)

46. Weber, P., Kornfelt, T., Klausen, N. K., Lunte, S. M. Characterization ofGlycopeptides from Recombinant Coagulation Factor VIIa by High-Performance Liquid Chromatography and Capillary Zone ElectrophoresisUsing Ultraviolet and Pulsed Electrochemical Detection. Anal. Biochem.225, 135-142 (1995)

47. Johnson, D. C., LaCourse, W. R. Pulsed Electrochemical Detection ofCarbohydrates at Gold Electrodes Following Liquid ChromatographySeparation, in Carbohydrate Analysis: High Performance Liquid Chro-matography and Capillary Electrophoresis, 391-429. Z. El Rassi, Ed.Amsterdam, Elsevier, 1995.

48. Johnson, D., LaCourse, W. Liquid Chromatography with Pulsed Electro-chemical Detection at Gold and Platinum Electrodes. Anal. Chem. 62,589A-597A (1990)

49. Luo, P., Zhang, F., Baldwin, R. Comparison of Metallic Electrodes forConstant-Potential Amperometric Detection of carbohydrates, AminoAcids and Related Compounds in Flow Systems. Ana. Chim. Acta 244,169-178 (1991)

96

50. Penn, S. G., Chiu, R. W., Monnig, C. A. Separation and Analysis ofCyclodextrins by Capillary Electrophoresis with Dynamic FluorescenceLabeling and Detection. J. Chromatogr. A 680, 233-241 (1994)

51. Brewster, J. D., Fishman, M. L. Capillary Electrophoresis of PlantStarches as the Iodine Complexes. J. Chromatogr. A 693, 382-387 (1995)

52. Honda, S., Iwase, S., Makino, A., Fujiwara, S. Simultaneous Determina-tion of Reducing Monosaccharides by Capillary Electrophoresis as theBorate Complexes of N-2-Pyridylglycamines. Anal. Biochem. 176, 72-77(1989)

53. Hase, S. Pre- and Post-Column Detection-Oriented Derivatization Tech-niques in HPLC of Carbohydrates, in Carbohydrate Analysis: High Per-formance Liquid Chromatography and Capillary Electrophoresis,555-575. Z. El Rassi, Ed. Amsterdam, Elsevier, 1995.

54. Mechref, Y., El Rassi, Z. Capillary Electrophoresis of Derivatized AcidicMonosaccharides. Electrophoresis 15, 627-634 (1994)

55. Mechref, Y., El Rassi, Z. Fused-Silica Capillaries with Surface-BoundDextran Layer Crosslinked with Diepoxypolyethylene Glycol for Capil-lary Electrophoresis of Biological Substances at Reduced ElectroosmoticFlow. Electrophoresis 16, 617-624 (1995)

56. Mechref, Y., Ostrander, G. K., El Rassi, Z. Capillary Electrophoresis ofCarboxylated Carbohydrates I. Selective Precolumn Derivatization ofGangliosides with UV Absorbing and Fluorescent Tags. J. Chromatogr.A 695, 83-95 (1995)

57. Honda, S., Ueno, T., Kakehi, K. High-Performance Capillary Electro-phoresis of Unsaturated Oligosaccharides Derived from Glycosaminogly-cans by Digestion with Chondroitinase ABC as 1-Phenyl-3-methyl-5-pyrazolone Derivatives. J. Chromatogr. 608, 289-295 (1992)

58. Honda, S. Suzuki, S. Nose, A., Yamamoto, K. Kakehi, K. Capillary ZoneElectrophoresis of Reducing Mono- and Oligosaccharides as the BorateComplexes of Their 3-Methyl-1-phenyl-2-pyrazolin-5-one Derivatives.Carbohydr. Res. 215, 193-198 (1991)

59. Novotny, M. V., Sudor, J. High-Performance Capillary Electrophoresisof Glycoconjugates. Electrophoresis 14, 373-389 (1993)

60. Zhao, J. Y., Diedrich, P., Zhang, Y., Hindsgaul, O., Dovichi, N. J. Sepa-ration of Aminated Monosaccharides by Capillary Electrophoresis withLaser-Induced Fluorescence Detection. J. Chromatogr. B 657, 307-313(1994)

61. Nashabeh, W., El Rassi, Z. Capillary Zone Electrophoresis of Linear andBranched Oligosaccharides. J. Chromatogr. 600, 279-287 (1992)

97

62. Lee, C. K., Sandler, S. I. Vancomycin Partitioning in Aqueous Two-Phase Systems: Effects of pH, Salts, and an Affinity Ligand. Biotechnol.Bioeng. 35, 408-416 (1990)

63. Amankwa, L., Albin, M., Kuhr, W. Fluorescence Detection in CapillaryElectrophoresis. Trends Anal. Chem. 11, 114-120 (1992)

64. Li, S. F. Y. Capillary Electrophoresis. Principles, Practice and Applica-tions, Ch. 3. Amsterdam, Elsevier, 1992.

65. Stefansson, M., Novotny, M. Electrophoretic Resolution of Monosaccha-rides Enantiomers in Borate-Oligosaccharide Complexation Media.J. Am. Chem. Soc. 115, 11573-11580 (1993)

66. Greenaway, M., Okafo, G. N., Camilleri, P., Dhanak, D. A Sensitive andSelective Method for the Analysis of Complex Mixtures of Sugars andLinear Oligosaccharides. J. Chem. Soc. Chem. Commun., 1691-1692(1994)

67. Huber, C., Grill, E., Oefner, P., Bobleter, O. Capillary ElectrophoreticDetermination of the Component Monosaccharides in Hemicelluloses.Fresenius J. Anal. Chem. 348, 825-831 (1994)

68. Schwaiger, H., Oefner, P., Huber, C., Grill, E., Bonn, G. K. CapillaryElectrophoresis and Micellar Electrokinetic Chromatography of4-Aminobenzonitrile Carbohydrate Derivatives. Electrophoresis 15,941-952 (1994)

69. Noe, C. R., Freissmuth, J. Capillary Zone Electrophoresis of AldoseEnantiomers: Separation after Derivatization with S-(-)-1-Phenylethyl-amine. J. Chromatogr. A 704, 503-512. (1995)

70. Chen F., Evangelista, R. A. Analysis of Mono- and OligosaccharideIsomers Derivatized with 9-Aminopyrene-1,4,6-trisulfonate by CapillaryElectrophoresis with Laser-Induced Fluorescence. Anal. Biochem. 230,273-280 (1995)

71. Evangelista, R. A., Liu, M., Chen, F. Characterization of 9-Aminopyrene-1,4,6-trisulfonated-Derivatized Sugars by Capillary Electrophoresis withLaser-Induced Fluorescence Detection. Anal. Chem. 67, 2239-2245(1995)

72. Smith J. T., El Rassi, Z. Capillary Zone Electrophoresis of BiologicalSubstance with Fused Silica Capillaries Having Zero or Constant Elec-troosmotic Flow. Electrophoresis 14, 396-406 (1993)

73. Lee, K., Kim, Y., Linhardt, R. Capillary Electrophoresis for the Quantita-tion of Oligosaccharides Formed Through the Action of Chitinase. Elec-trophoresis 12, 636-640 (1991)

74. Chiesa, C., Horvath, C. Capillary Zone Electrophoresis of Malto-Oli-gosaccharides Derivatized with 8-Aminonaphthalene-1,3,6-trisulfonicacid. J. Chromatogr. 645, 337-352 (1993)

98

75. Jackson, P. Polyacrylamide Gel Electrophoresis of Reducing SaccharidesLabeled with the Fluorophore 2-Aminoacridone: Subpicomolar DetectionUsing an Imaging System Based on a Cooled Charge-Coupled Device.Anal. Biochem. 196, 238-244 (1991)

76. Guttman, A., Pritchett, T. Capillary Gel Electrophoresis of High-Man-nose Type Oligosaccharides Derivatized by 1-Aminopyrene-3,6,8-trisulfonic Acid. Electrophoresis 16, 1906-1911 (1995)

77. Liu, J., Shirota, O., Wiesler, D., Novotny, M. Ultrasensitive FluorescenceDetection of Carbohydrates as Derivatives in Mixtures Separated by Cap-illary Electrophoresis. Proc. Natl. Acad. Sci. USA 88, 2302-2306 (1991)

78. Liu, J., Shirota, O., Novotny, M. Separation of Fluorescent Oligosaccha-rides Derivatives by Microcolumn Techniques Based on Electrophoresisand Liquid Chromatography. J. Chromatogr. 559, 223-235 (1991)

79. Liu, J., Shirota, O., Novotny, M. Sensitive, Laser-Assisted Determinationof Complex Oligosaccharides Mixtures Separated by Capillary Gel Elec-trophoresis. Anal. Chem. 64, 973-975 (1992)

80. Sudor, J., Novotny, M. Electromigration Behavior of Polysaccharides inCapillary Electrophoresis Under Pulsed-Field Conditions. Proc. Natl.Acad. Sci. USA 90, 9451-9455 (1993)

81. Liu, J., Shirota, O., Novotny, M. Capillary Electrophoresis of AminoSugars with Laser-Induced Fluorescence Detection. Anal. Chem. 63,413-417 (1991)

82. Zhang, Y., Arriaga, E., Diedrich, P., Hindsgaul, O., Dovichi, N. J. Nano-molar Determination of Aminated Sugars by Capillary Electrophoresis.J. Chromatogr. A 1716, 221-229 (1995)

83. Zhang, Y., Le, X., Dovichi, N. J., Compston, C. A., Palcic, M. M.,Diedrich, P., Hindsgaul, O. Monitoring Biosynthetic Transformations ofN-Acetyllactosamine Using Fluorescently Labeled Oligosaccharides andCapillary Electrophoretic Separation. Anal. Biochem. 227, 368-376(1995)

84. Le, X., Scaman, C., Zhang, Y., Zhang, J., Dovichi, N. J., Hindsgaul, O.,Palcic, M. M. Analysis of Capillary Electrophoresis-Laser-Induced Fluo-rescence Detection of Oligosaccharides Produced from Enzyme Reaction.J. Chromatogr. A 716, 215-220 (1995)

85. Bruno, A. E., Krattiger, B. On-Column Refractive Index Detection ofCarbohydrates Separated by HPLC and CE, in Carbohydrate Analysis:High Performance Liquid Chromatography and Capillary Electrophore-sis, 431-446. Z. El Rassi, Ed. Amsterdam, Elsevier, 1995.

86. Nashabeh, W., El Rassi, Z. Capillary Zone Electrophoresis ofPyridylamino Derivatives of Maltooligosaccharides. J. Chromatogr. 514,57-64 (1990)

99

87. Chiesa, C., Oefner, P. J., Zieske, L. R., O’Neill, R. A. Micellar Electroki-netic Chromatography of Monosaccharides Derivatized with 1-Phenyl-3-methyl-2-pyrazolin-5-one. J. Cap. Elec. 2, 175-183 (1995)

88. Arentoft, A. M., Michaelsen, S., Sørensen, H. Determination of Oligosac-charides by Capillary Zone Electrophoresis. J. Chromatogr. A 652,517-524 (1993)

89. Al-Hakim, A., Linhardt, R. J. Capillary Electrophoresis for the Analysisof Chondroitin Sulfate- and Dermatan Sulfate-Derived Disaccharides.Anal. Biochem. 195, 68-73 (1991)

90. Carney, S. L., Osborne, D. J. The Separation of Chondroitin Sulfate Dis-accharides and Hyaluronan Oligosaccharides by Capillary Zone Electro-phoresis. Anal. Biochem. 195, 132-140 (1991)

91. Ampofo, S. A., Wang, H. M., Linhardt, R. J. Disaccharide CompositionalAnalysis of Heparin and Heparan Sulfate Using Capillary Zone Electro-phoresis. Anal. Biochem. 199, 249-255 (1991)

92. Damm, J. B. L., Overklift, G. T., Vermeulen, B. W. M., Fluitsma, C. F,.van Dedem, G. W. K. Separation of Natural and Synthetic Heparin Frag-ments by High Performance Capillary Electrophoresis. J. Chromatogr.608, 297-309 (1992)

93. Michaelsen, S., Schroder, M.-B., Sorensen, H. Separation and Determina-tion of Glycosaminoglycan Disaccharides by Micellar ElectrokineticCapillary Chromatography for Studies of Pelt Glycosaminoglycans.J. Chromatogr. A 652, 503-515 (1993)

94. Pervin, A., Kenan, G., Linhardt, R.J. Capillary Electrophoresis to Mea-sure Sulfoesterase Activity on Chondroitin Sulfate and Heparin DerivedDisaccharides. Appl. Theor. Electrophoresis 3, 297-303 (1993)

95. Hermentin, P., Witzel, R., Doenges, R., Bauer, R., Haupt, H., Patel, T.,Parekh, R. B., Brazel, D. The Mapping by High-pH Anion-ExchangeChromatography with Pulsed Amperometric Detection and CapillaryElectrophoresis of the Carbohydrate Moieties of Human Plasma α1-AcidGlycoprotein. Anal. Biochem. 206, 419-429 (1992)

96. Taverna, M., Baillet, A., Biou, D., Schlüter, M., Werner, R., Ferrier, D.Analysis of Carbohydrate-Mediated Heterogeneity and Characterizationof N-Linked Oligosaccharides of Glycoproteins by High PerformanceCapillary Electrophoresis. Electrophoresis 13, 359-366 (1992)

97. Stefansson, M., Novotny, M. Separation of Complex OligosaccharideMixtures by Capillary Electrophoresis in the Open-Tubular Format. Anal.Chem. 66, 1134-1140 (1994)

98. Chiesa, C., O’Neill, R. A. Capillary Zone Electrophoresis of Oligosac-charides Derivatized with Various Aminonaphthalene Sulfonic AcidsElectrophoresis 15, 1132-1140 (1994)

100

99. Sudor, J., Novotny, M. V. End-Label, Free-Solution Capillary Electro-phoresis of Highly Charged Oligosaccharides. Anal. Chem. 67, 4205-4209 (1995)

100. Smith, J. T., El Rassi, Z. Capillary Zone Electrophoresis of BiologicalSubstances with Surface-Modified Fused Silica Capillaries with Switch-able Electroosmotic Flow. J. High Resolut. Chromatogr. 15, 573-578(1992)

101. Honda, S., Makino, A., Suzuki, S., Kaheki, K. Analysis of the Oligosac-charides in Ovalbumin by High-Performance Capillary Electrophoresis.Anal. Biochem. 191, 228-234 (990)

102. Liu, J., Dolnik, V., Hsieh, Y.-Z., Novotny, M. Experimental Evaluationof the Separation Efficiency in Capillary Electrophoresis Using OpenTubular and Gel-Filled Columns. Anal. Chem. 64, 1328-1336 (1992)

103. Klockow, A., Widmer, H. M., Amadò, R., Paulus, A. Capillary Electro-phoresis of ANTS Labeled Oligosaccharides Ladders and ComplexCarbohydrates. Fresenius J. Anal. Chem. 350, 415-425 (1994)

104. Guttman, A., Cooke, N., Starr, C. M. Capillary Electrophoresis Separa-tion of Oligosaccharides: I. Effect of Operational Variables. Electro-phoresis 115, 1518-1522 (1994)

105. El Rassi, Z., Tedford, D., An, J., Mort, A. High-Performance Reversed-Phase Chromatographic Mapping of 2-Pyridylamino Derivatives ofXyloglucan Oligosaccharides. Carbohydr. Res. 215, 25-38 (1991)

106. Stefansson, M., Novotny, M. Resolution of the Branched Forms ofOligosaccharides by High-Performance Capillary Electrophoresis.Carbohydr. Res. 258, 1-9 (1994)

107. Nashabeh, W., El Rassi, Z. Capillary Zone Electrophoresis of α1-AcidGlycoprotein Fragments from Trypsin and Endoglycosidase Digestions.J. Chromatogr. 536, 31-42 (1991)

108. Guttman, A., Starr, C. Capillary and Slab Gel Electrophoresis Profilingof Oligosaccharides. Electrophoresis 16, 993-997 (1995)

109. Klockow, A., Amado, R., Widmer, H. M., Paulus, A. Separation of8-Aminonaphthalene-1,3,6-trisulfonic Acid-Labeled Neutral andSialylated N-Linked Complex Oligosaccharides by Capillary Electro-phoresis. J. Chromatogr. A 716, 241-257 (1995)

110. Camilleri, P., Harland, G. B., Okafo, G. High Resolution and RapidAnalysis of Branched Oligosaccharides by Capillary Electrophoresis.Anal. Biochem. 230, 115-122 (1995)

111. Suzuki, S., Honda, S. Two-Dimensional Mapping of N- GlycosidicallyLinked Oligosaccharides in Glycoproteins by High-Performance Capil-lary Electrophoresis. Trends Anal. Chem. 14, 279-288 (1995)

101

112. Suzuki, S., Kakehi, K., Honda, S. Two-Dimensional Mapping ofN- Glycosidically Linked Asialo-Oligosaccharides from Glycoproteinsas Reductively Pyridylaminated Derivatives using Dual SeparationModes of High-Performance Capillary Electrophoresis. Anal. Biochem.205, 227-236 (1992)

113. Stefansson, M., Novotny, M. Modification of the Electrophoretic Mo-bility of Neutral and Charged Polysaccharides. Anal. Chem. 66, 3466-3471 (1994 )

114. Teitelbaum, R. C., Ruby, S. L., Marks, T. J. J. Am. Chem. Soc. 100,3215-3217 (1961)

115. Rundle, R. E., French, D. The Complexation of Starch and Starch-Iodine Complex. II. Optical Properties of Crystalline Starch Fractions.J. Am. Chem. Soc. 65, 558-561 (1943)

116. Thoma, A., French, D. Starch-Iodine-Iodine Interaction. I. Spectropho-tometric Investigations. J. Am. Chem. Soc. 65, 1825-1828 (1961)

117. Banks, W., Greenwood, C. T., Khan, K. M. Interaction of Linear Amy-lose Oligomers with Iodine. Carbohydr. Res. 17, 25-33 (1971)

118. Rush, R. S., Derby, P. L., Strickland, T. W., Rohde, M. F. Peptide Map-ping and Evaluation of Glycopeptide Microheterogeneity Derived fromEndoproteinase Digestion of Erythropoietin by Affinity High-Perfor-mance Capillary Electrophoresis. Anal. Chem. 65, 1834-1842 (1993)

119. Kilàr, F., Hjerten, S. Separation of the Human Transferrin Isoforms byCarrier Free High-Performance Zone Electrophoresis and IsoelectricFocusing. J. Chromatogr. 480, 351-357 (1989,)

120. Rudd, P. M, Scragg, I. G., Coghil, E., Dwek, R. A. Separation andAnalysis of the Glycoform Populations of Ribonuclease B Using Capil-lary Electrophoresis. Glycoconjugate J. 9, 86-91 (1992)

121. Yim, K., Abrams, J., Hsu, A. Capillary Zone Electrophoretic Resolutionof Recombinant Human Bone Morphogenetic Protein 2 Glycoforms. AnInvestigation into the Separation Mechanisms for an Exquisite Separa-tion. J. Chromatogr. A 716, 401-412 (1995)

122. Pedersen, J., Biedermann, K. Characterization of Proteinase AGlycoforms from Recombinant Saccharomyces cerevisiae. Biotechnol.Appl. Biochem. 18, 377-388 (1993)

123. Landers, J. P., Oda, R. P., Madden, B. J., Spelsberg, T. C. High Perfor-mance Capillary Electrophoresis of Glycoproteins: The Use of Modifiersof Electroosmotic Flow for Analysis of Microheterogeneity. Anal.Biochem 205, 115-124 (1992)

124. Oda, R. P., Madden, B. J., Spelsberg, T. C., Landers, J. P. α,ω-Bis-Quaternary Ammonium Alkanes as Effective Buffer Additives for En-hanced Capillary Electrophoretic Separation of Glycoprotein. J.Chromatogr. A 680, 85-92 (1994)

102

125. Morbeck, D. E., Madden, B. J., McCormick, D. J. Analysis of the Mi-croheterogeneity of the Glycoprotein Chorionic Gonadotropin withHigh-Performance Capillary Electrophoresis. J. Chromatogr. A, 680,217-224 (1994)

126. Yim, K. W. Fractionation of the Human Recombinant Tissue Plasmino-gen Activator (rtPA) Glycoforms by High-Performance Capillary ZoneElectrophoresis and Capillary Isoelectric Focusing. J. Chromatogr. 559,401-410 (1991)

127. Watson, E., Yao, F. Capillary Electrophoretic Separation of Recombi-nant Granulocyte-Colony-Stimulating Factor Glycoforms. J.Chromatogr. 630, 442-446 (1993)

128. Tran, A. D., Park, S., Lisi, P. J., Huynh, O. T., Ryall, R. R., Lane, P. A.Separation of Carbohydrate-Mediated Microheterogeneity of Recombi-nant Human Erythropoietin by Free Solution Capillary Electrophoresis.J. Chromatogr. 542, 459-471 (1991)

129. Watson, E. ,Yao, F. Capillary Electrophoretic Separation of HumanRecombinant Erythropoietin (r-HuEPO) Glycoforms. Anal. Biochem.210, 389-393 (1993)

130. James, D. C., Freedman, R. B., Hoare, M., Jenkins, N. High-ResolutionSeparation of Recombinant Human Interferon-γ Glycoforms by MicellarElectrokinetic Capillary Chromatography. Anal. Biochem. 222, 315-322(1994)

131. Klausen, N. K., Kornfelt, T. Analysis of the Glycoforms of HumanRecombinant Factor VIIa by Capillary Electrophoresis and High-Perfor-mance Liquid Chromatography. J. Chromatogr. A 718, 195-202 (1995)

132. Vincentelli, R., Bihoreau, N. Characterization of Each Isoform ofF(ab')2 by Capillary Electrophoresis. J. Chromatogr. 641, 383-390(1993)

133. Liu, Y., Chan, J. K.-F. High-Performance Capillary Electrophoresis ofGangliosides. Electrophoresis 12, 402-408 (1991)

134. Yoo, Y. S., Kim, Y. S., Jhon, G.-J., Park, J. Separation of GangliosidesUsing Cyclodextrin in Capillary Zone Electrophoresis. J. Chromatogr.A 652, 431-439 (1993)

135. Ulrich-Bott, B., Wiegandt, H. Micellar Properties of Glycophingolipidsin Aqueous Media. J. Lip. Res. 25, 1233-1245 (1984)

136. Issaq, J. I., Janini, G. M., Chan, K. C., El Rassi, Z. Approaches for theOptimization of Experimental Parameters in Capillary Zone Electro-phoresis. Adv. Chromatogr. 35, 101-169 (1995)

An Introduction to Chiral Analysis by

Capillary Electrophoresis

U.S. (800) 4BIORAD California (510) 741-1000 New York (516) 756-2575 Australia 02-805-5000 Austria (1) 877 89 01 Belgium 09-385 55 11 Canada (905) 712-2771China (01) 2046622 Denmark 39 17 9947 Finland 90 804 2200 France (1) 49 60 68 34 Germany 089 31884-0 India 91-11-461-0103 Italy 02-21609 1 Japan 03-5811-6270 Hong Kong 7893300 The Netherlands 0318-540666 New Zealand 09-443 3099 Singapore (65) 4432529 Spain (91) 661 70 85 Sweden 46 (0) 735 83 00Switzerland 01-809 55 55 United Kingdom 0800 181134

Life Science Group

SIG 051995 Printed in USA Bulletin 1973 US/EG REVA 95-0284 0695


95-0284 Chiral Guide Cover 3/11/99 9:20 AM Page CVR1

Salvatore Fanali

Istituto di Cromatografia del Consiglio Nazionaledelle Ricerche

Area della Ricerche di Roma P.O. Box 10 00016 Monterotondo Scalo (Roma) Italy



95-0284 Chiral Guide Text 3/11/99 9:37 AM Page a

Table of Contents

About the Author ..........................................................ii

Section 1 Introduction....................................................................11.1 Abbreviations Used in the Text................................................1

1.2 Introduction to Chiral Analysis ................................................2

Section 2 Chirality and Optical Isomers Separation.................3

Section 3 Separation of Diastereoisomers by Indirect Methods......................................................4

Section 4 Direct Methods for Chiral Separation in Capillary Electrophoresis ........................................8

4.1 Inclusion Complexation .........................................................12

Cyclodextrins and their Derivatives in Capillary Zone Electrophoresis .........................................12

Cyclodextrins and their Derivatives in Capillary Gel Electrophoresis, Isotachophoresis, and Electrochromatography...................................................25

Chiral Crown-Ether................................................................26

4.2 Ligand Exchange ...................................................................27

4.3 Chiral Micelles........................................................................28

4.4 Affinity Interactions................................................................30

Proteins as Chiral Selectors ...................................................30

Interactions with Saccharides (Linear Polymers)and Antibiotics ........................................................................33

4.5 Combination of Chiral/Chiral or Chiral/Achiral Selectors ...33

Section 5 Optimization of the Method.......................................35

Section 6 Quantitation .................................................................39

Section 7 Applications..................................................................40

Section 8 Conclusions ..................................................................42

Section 9 References.....................................................................43

Acknowledgments .......................................................50

i

95-0284 Chiral Guide Text 3/11/99 9:37 AM Page i

About the Author

Dr. Salvatore Fanali is a researcher at the Istituto di Cromatografia delConsiglio Nazionale delle Ricerche in Montelibretti (Rome) Italy. After experience in high performance paper electrophoresis and paper chromatog-raphy, his research concerned the study of capillary isotachophoresis in thepharmaceutical field. Since 1988 his research has focused on electromigrationmethods, mainly applied to the separation of enantiomers. He is author or co-author of seventy publications, two booklets, and chapters in two books.He is a member of the advisory editorial boards of Bollettino ChimicoFarmaceutico, Chromatographia, Electrophoresis, Journal of CapillaryElectrophoresis, Journal of Chromatography, and Journal of High ResolutionChromatography.

ii

95-0284 Chiral Guide Text 3/11/99 9:37 AM Page ii

Section 1Introduction

1.1 Abbreviations Used in the TextAPA aryl propionic acidBGE background electrolyteCD cyclodextrinCGE capillary gel electrophoresisCM-β-CD carboxymethylated-β-cyclodextrinCM-β-CD pol carboxymethylated-β-cyclodextrin polymerCMC critical micellar concentrationCrown-ether 18-crown-6-ether tetra-carboxylic acidCSP chiral stationary phaseCZE capillary zone electrophoresisDAT (+)-O,O’-diacetyl-L-tartaric anhydrideDBT (+)-O,O’-dibenzoyl-L-tartaric anhydridedi-OMe-β-CD 2,6-di-O-methyl-β-cyclodextrinEHC electro-chromatographyeof electro-osmotic flowGC gas chromatographyHEC hydroxyethylcelluloseHP-β-CD hydroxypropyl-β-cyclodextrinHPLC high performance liquid chromatographyHPMC hydroxypropylmethylcelluloseITP isotachophoresisLE leading electrolyteLEC ligand exchange chromatographyMEKC micellar electrokinetic chromatographyNDA naphthalene-2,3-dicarboxyaldehydeNSAIDs non-steroidal anti-inflammatory drugsPhL phenyl lactic acidPVA polyvinyl alcoholPVP polyvinyl pyrrolidoneSBE-β-CD sulfobutyl-ether(IV)-β-cyclodextrinSDAla sodium N-dodecanoyl-L-alaninateSDS sodium dodecyl sulfateSDVal sodium N-dodecanoyl-L-valinateSTC sodium taurocholateSTDC sodium taurodeoxycholateTLC thin layer chromatographytri-OMe-β-CD 2,3,6-tri-O-methyl-β-cyclodextrin

1

95-0284 Chiral Guide Text 3/11/99 9:37 AM Page 1

1.2 Introduction to Chiral AnalysisThe separation of enantiomers is widely studied in analytical

chemistry, especially in the pharmaceutical and biological fields,since chiral drugs are administered either as enantiomers or as racemicmixtures. Very often two enantiomers of the same racemic drug possess different pharmacological effects. For example, S(-)-propra-nolol is considerably more active than its enantiomers. The anestheticketamine is administered as racemate, and the S(+)-ketamine form ismore potent than the R(-) form. In addition, the R(-) form may causepost-operative effects. Because of the side effects that could be causedby the presence of an undesirable component in a racemic drug, thecurrent tendency in the pharmaceutical industry is to prepare drugs withone enantiomer only. However, the production of such drugs throughstereoselective reaction or preparative enantiomeric separation canprovide impure material. Thus rapid, sensitive analytical methods ofhigh resolving power are required to control the synthetic chiral process of the pharmaceutical.

Analytical methods which have been used for the separation of chiral compounds include high performance liquid chromatography(HPLC),1–3 gas chromatography (GC),4,5 thin layer chromatography(TLC),6 and recently, capillary electrophoresis (CE).7–15 CE is especially useful for the analysis of different classes of compounds,including organic and inorganic ions, peptides, proteins, saccharides,drugs, optical isomers, and others. In CE the separation will takeplace if the analytes, under the influence of an applied electric field,move toward the detector with different velocities (different effectivemobilities). Thin capillaries, with internal diameters of 10–100 µm,allow application of relatively high electric fields (200–1,000 V/cm)permitting analytes to be separated in short times with high resolutionand high efficiency, because the Joule heat, the main cause of zone dis-persion, is easily controlled.

The different separation modes available in CE, namely zoneelectrophoresis (CZE), micellar electrokinetic chromatography(MEKC), gel electrophoresis (CGE), and isotachophoresis (ITP),combined with the high resolving power and high efficiency, make theCE technique competitive with others.16 CE offers additional advantages over other techniques such as HPLC, including

• A relatively small volume of sample and buffer (nl and µl, respectively) is required

• Expensive chiral columns can be avoided because the chiral selector can be easily added to the BGE

2


• The separation is highly reproducible because the buffer with thechiral selector is replenished after each run

The use of CE for analysis of chiral compounds is well documented and, considering the enormous increase in publications,it can be expected that in the near future chiral analysis by CE will bewidely applied to pharmaceutical, clinical, and environmental sam-ples.

Published reviews,17–24 and the special issue edited by us25 dealingwith the separation of optical isomers by CE, are available for furtherreading. The article by Fanali et al. provides a list of enantiomericcompounds which have been separated using capillary electrophoresis.24

Section 2Chirality and Optical Isomers Separation

When four different ligands are bound to a tetravalent carbon,an asymmetric molecule is generated in which the carbon is the asymmetric center. As shown in Figure 1, two optical isomers canbe generated due to the different spatial orientations of the ligandsaround the chiral center.

Fig. 1. Mirror image of two optical isomers.

Enantiomers are two stereoisomers which exhibit non-superim-posable mirror images. Diastereomers generally possess at least twoasymmetric centers (one of them has the same configuration), andare not mirror images. The most common chiral center is representedby tetrahedral carbon, although other atoms, like nitrogen, sulfur, andphosphorous, can be found in stereoisomers. Compounds possessingat least two enantiomers are chiral compounds.

BX

Y

A C * B

X

Y

A

C *

3


The main property of stereoisomers is represented by the rotationof polarized light in different directions, counter-clockwise (levo-rotatory) and clockwise (dextro-rotatory). They are l(-)- andd(+)-isomers, respectively. The optical rotation of two enantiomers (ina racemic mixture) is of the same magnitude, while in the case ofdiastereomers the two parameters may not be the same due to different physico-chemical properties. Unlike enantiomers, diastereomers may have different melting points, boiling points, andsolubility.

Although the d, or (+), and l, or (-), symbols show a very important physical property of the molecule, they unfortunately do notgive any information concerning the spatial arrangement of the chiral center. The Fisher convention, widely employed for sugars andamino acids, makes use of D and L symbols and is based on the comparison of the substituents of the chiral center of the compoundunder investigation with that of (+)-glyceraldehyde. However, the D and L convention can be confused with d and l terminology, and thusthe currently recognized convention is the Cahn-Ingold-Prelog.26

Here the priority of the ligand to the asymmetric center based on theatomic number is controlled, and the group with lower priority ispositioned far away from the observation point. Then the priority ofthe other substituents is examined. If it decreases clockwise, the R configuration is assigned, otherwise the S configuration is given.

Section 3Separation of Diastereoisomers byIndirect Methods

Because two diastereoisomers of the same compound have different physico-chemical properties, they can, in principle, be separated from each other by CE using a non-chiral electrophoreticsystem. As an example, consider the chemical structure of nor-ephedrine and ψ-nor-ephedrine (Figure 2). The two compoundspossess the same molecular weight and the same substituent groupsat the two asymmetric carbons, and the configuration of only one ofthe asymmetric centers is the same; thus they are diastereoisomers.Each diastereoisomer exists as a pair of enantiomers (d- and l-norephedrine and d- and l-ψ-norephedrine, respectively).

4


Fig. 2. Chemical structure of nor-ephedrine and ψ-nor-ephedrine.

When a mixture of the two compounds is injected for an electrophoretic run using a non-chiral background electrolyte (phosphate buffer at pH 2.5, see Figure 3), two peaks, corresponding tothe two diastereoisomers, are obtained; each peak represents a coupleof enantiomers, but the electrophoretic system is unable to separatethem.

The above principle can be advantageously used for the resolutionof a pair of enantiomers. Thus it is necessary to use a chiral environ-ment that will modify the physico-chemical properties of the twoanalytes, transforming enantiomers into diastereoisomers. By adjustingthe selectivity of the separation, the resolution can be improved.

NHHO

HH

CH

H

2

3NH

HO

HH

CH

H

2

3

NH

HO

H

H

CH

H

2

3

NHHO

H

H

CH

H

2

3

1R,2S-(-) -norephedrine 1S,2R-(+)-norephedrine

1R,2R-(-)-ψ-norephedrine 1S,2S-(+)-ψ-norephedrine

5


Fig. 3. Electropherogram of the separation of 1) ψ-norephedrine and 2)norephedrine. Apparatus, BioFocus¤ 3000 system; capillary, (coated) 35 cm x 0.05 mm ID; background electrolyte, 50 mM phosphate buffer, pH 2.5;applied voltage 15 kV; injection, electrokinetic 7 kV, 7 s of 2 x 10-5 M of eachcompound; carousel and capillary temperature 25 ¡C; detection at 206 nm.

The indirect separation method is based on the reaction, before theanalysis, of a racemic mixture with a chiral reagent (R or S), producing a mixture of two diastereoisomers that can be resolvedusing a non-chiral electrophoretic system. The product of the reactionis a mixture of two stable compounds where relatively strong bonds(covalent) are involved in the process.

The derivatization of α-amino- or α-hydroxyacids with (+)-O,O’-diacetyl-L-tartaric anhydride (DAT) or (+)-O,O’-dibenzoyl-L-tartaricanhydride (DBT) can serve as an example. Figure 4 shows the chemical structure of the two diastereoisomers formed between DBTand L- and D-amino acids. The configuration of the two carbons ofthe tartrate is the same, and that of the amino acid chain is different.

1

2

6 8min

6


Fig. 4. Chemical structure of the two diastereoisomers formed in thereaction between (+)-O,O’-dibenzoyl-L-tartaric anhydride (DBT) and L-,D-amino acid.

The separation of such derivatives was obtained by CE using aBGE at pH 5.8 containing polyvinyl pyrrolidone (PVP) as a physicalnetwork.27 The effect of the concentration of PVP added to the BGEfor the separation of mandelic acid and tryptophan after derivatizationwith DBT28 is shown in Figure 5.

Fig. 5. Effect of the concentration of polyvinylpyrrolidone (PVP) on theseparation of diastereoisomers of DPT-D,L-mandelic acid and DBT-D,L-tryptophan. Capillary, (coated) 56 cm x 0.1 mm ID; background electrolyte,25 mM phosphate buffer, pH 5.8; applied voltage 12 kV; injection, hydro-dynamic (5 s at a height distance of 10 cm) of 10-4 M of sample; detectionat 233 nm (modified from reference 28).

L D

2%

0%

2%

L D

0%

8 9 11 12 13 t14 15 167 8 t

DBT-trp DBT-mandelic acid

O

O

O

O

O

OO

O

OC

CC

C

C

CH

CH

CH

R

NH*

*

* _

_

DBT amino acid derivative(+)-O,O′ -dibenzoyl-L-tartaric anhydride

7


Although the indirect separation method using the appropriatechiral reagent (R or S) can be used advantageously when an inversionof migration order is required, the method presents several draw-backs. It is time consuming, it requires the use of very pure chiralreagent, the enantiomers must contain reacting groups (hydroxyl,amino), and the kinetics of the reaction can lead to different peakareas for the two diastereomers formed. Probably due to these and otherdrawbacks of the indirect separation method, the direct separationmethod is becoming very popular for enantiomeric separation by CE.

Section 4Direct Method for Chiral Separation inCapillary Electrophoresis

In the direct separation method the chiral selector can be addedto the BGE, bound to the capillary wall, or included in a gel matrix.It interacts with the two enantiomers during the electrophoretic process, forming labile diastereoisomeric complexes. Relatively weakbonds are involved, e.g., hydrogen, π-π, or hydrophobic.

The separation of two enantiomers can take place only if the twodiastereoisomers formed possess different stability constants, causingthe two analytes to move with different velocities. The effective mobility of the most complexed enantiomer is lower than that of itsisomer (this is not true if the chiral selector is negatively charged). Forexample, Figure 6 shows the separation of the four optical isomers ana-lyzed in Figure 3, using the same background electrolyte supplementedwith a chiral selector (2,6-di-O-methyl-β-CD) that allows the separationnot only of the diastereoisomers but also of the enantiomers presentin the mixture.

8


Fig. 6. Electropherogram of the separation of the four optical isomersof norephedrine and ψ-norephedrine. Apparatus, BioFocus 3000 system;background electrolyte, 50 mM phosphate buffer, pH 2.5, and 20 mM of 2,6-di-O-methyl-β-cyclodextrin. For other experimental conditions, see Figure 3.

The direct separation of enantiomers by CE is easier to performthan the indirect separation. This is demonstrated by the increasingnumber of publications reported in the literature (see references 24and 29). The direct method is less time consuming since derivatizationand purification are not required, a wide number of chiral selectorsare commercially available, and small amounts of chiral selectors canbe used. Furthermore the chiral purity of the system is not of paramountimportance. In fact, the presence of impure chiral selectors will onlycause a reduction of resolution of the two enantiomers in the directmethod. In the indirect method, the presence of impure chiral selectorswill produce four diastereoisomers (more difficult to separate). This isdemonstrated in Figure 7, where a mixture of (+)-Co(en)3

3+ and (-)-Co(en)3

3+ is separated using sodium L(+)-tartrate; when the BGEcontained the same chiral selector with 10% D(-)-tartrate, the resolutionof the two enantiomers was reduced, while use of 100% D(-)-tartrateallowed baseline resolution with inversion of migration order.

10 12 14 min1R

,2S

-(-)

-nor

-Eph 1R

,2R

-(-)

-ψ-n

or-E

ph

1S,2

R-(

+)-

nor-

Eph

1S,2

S-(

+)-

ψ-n

or-E

ph

9


Fig. 7. Electropherograms of the separation of (+/-)-Co(en3)3+. Apparatus,

BioFocus 3000 system; capillary, (coated) 24 cm x 0.05 mm ID; backgroundelectrolyte, 50 mM sodium tartrate, pH 5.1 [A= 100% L(+), B= 90% L(+)and 10% D(-), C= 100% D(-)]; applied voltage 5 kV, 9.4 A; injection, 5 psi*s 35 mg racemic Co-complex + 9 mg (+)-antipode in 100 ml; detectionat 230 nm.

Table 1 summarizes the main chiral selectors used in CE for thedirect separation of enantiomers, as well as the different CE modes usedand the main mechanisms involved in the separation process.

0 3 6

min

C(+)

(-)

B (+)

(+)(-)

(-)

A

A23

0

10


Table 1. Chiral Inclusion-Complexing Agents Used InCapillary Electrophoresis.Compound CE Type Mechanism References

Uncharged Compounds

α-cyclodextrin (α-CD) CZE, ITP, inclusion 14, 30–32CGE, MEKC

β-cyclodextrin (β-CD) CZE, ITP, inclusion 15, 33–35CGE, MEKC

γ-cyclodextrin (γ-CD) CZE, ITP inclusion 14, 36–38

glycosylated-α-cyclodextrin CZE inclusion 39(G-α-CD)

heptakis-2,6-di-O-methyl-β- CZE, ITP inclusion 13, 15, 30, 40cyclodextrin (di-OMe-β-CD)

heptakis-2,3,6-tri-O-methyl-β- CZE, ECH inclusion 40–43cyclodextrin (tri-OMe-β-CD)

hydroxyethyl-β-cyclodextrin CZE inclusion 41(HE-β-CD)

hydroxypropyl-β-cyclodextrin CZE inclusion 33, 39, 41, 44(HP-β-CD)

allyl derivatized-β-CD (Al-β-CD) CGE inclusion 45

Uncharged-β-cyclodextrin polymer CZE inclusion 46

Chargeable Compounds

6A-methylamino-β-cyclodextrin CZE inclusion 47(6A-NH-β-CD)

6A,6D-dimethylamino-β- CZE inclusion 47cyclodextrin (6A,D-di-NH-β-CD)

4-sulfobutyl-ether-β-cyclodextrin CZE inclusion 33, 48–52(SBE-β-CD)

mono-(6-β-aminoethylamino- CZE inclusion 536-deoxy)-β-cyclodextrin (β-CDen)

carboxymethyl-β-cyclodextrin CZE inclusion 9, 54(CM-β-CD)

carboxymethyl-β-cyclodextrin CZE inclusion 55polymer

18-crown-6-ether tetracarboxylic CZE inclusion 32, 56acid (18-crown ether)

copper(II)/L-proline or L-hydro- CZE, MEKC ligand exchange 57–59xyproline or aspartame

maltodextins CZE affinity 60

bile salts MEKC affinity-MEKC 7, 61, 62

proteins CZE affinity 63–65

antibiotics CZE inclusion 66, 67

L-tartrate CZE outer-sphere 68

11


4.1 Inclusion ComplexationThe resolution of enantiomers has been obtained in CE through

inclusion-complexation mechanisms using either cyclodextrins orcrown-ether derivatives. Here the two analytes (guest compounds)fit the cavity of the chiral selector (host compound), forming host-guestor inclusion-complexes. The separation of the two enantiomers cantake place only if the two diastereomeric complexes, formed duringthe electrophoretic process, possess different stability constants. The chiral separation is obtained due to the formation of secondary bondsbetween the substituent groups on the chiral center of the analytesand those of the chiral selectors positioned outside the cavity (hydroxyl or modified hydroxyl and carboxylic groups for cyclodex-trins and crown-ethers, respectively). To clarify the recognitionmechanism of chiral resolution using inclusion-complexation, it isnecessary to describe the structure and the main properties of the chiral selectors involved in the electrophoretic process.

Cyclodextrins and their Derivatives in Capillary ZoneElectrophoresis

Cyclodextrins (CD) are neutral and natural cyclic oligosaccharidescomposed of several glucopyranose units. They are chiral due to thepresence of asymmetric carbons on the glucose units. In spite of thefact that cyclodextrins containing 6–12 D(+)-glucopyranose units are reported in the literature, only 6, 7, and 8 unit-formed molecules,named with the Greek alphabet letters α, β, and γ, respectively, are infrequent use in analytical chemistry.69,70

The shape of cyclodextrins is similar to that of a truncated conewith a cavity of different dimensions depending on the CD type (thenumber of the glucose units and the substituent groups). Their cavityis relatively hydrophobic and able to accept guest compounds of different types, particularly those with non-polar groups. The out-side is relatively hydrophilic due to the presence of hydroxyl groups(primary and secondary).

Figure 8 shows the chemical structure of α-cyclodextrin, whileTable 2 gives the main properties of the most commonly used CDs.

12


Fig. 8. Chemical structure of α-cyclodextrin.

Table 2. The Main Properties of Native Cyclodextrins 69,71

Cyclodextrin type α β ψ

Number of glucopyranose 6 7 8

Molecular weight 973 1,135 1,297

Internal diameter (nm) 0.47–0.52 0.60–0.64 0.75–0.83

Depth (nm) 0.79–0.80 0.79–0.80 0.79–0.80

Specific rotation [ α ]25 D 150.5 162.5 177.4

Melting point (K) 551 572 540

Solubility in water g/100 ml 25 °C 14.50 1.85 23.20

The solubility of β-CD is relatively low when compared to thatof α and γ; this could be a problem when using the chiral selector asan additive to the background electrolyte. When higher concentrationsof β-CD must be used, water-methanol, -ethanol, or -urea mixtures canbe successfully used. In fact the solubility of β-CD in aqueous solutions of 4 and 8 M urea increases up to 0.089 and 0.226 M, respectively.72

OO

O

O

O

OO

O

OO

O

OOH

OH

OH

OHOH OH

OH

HO

HO

HO

HO

HO

23

6

HOH C2

HOH C2

HOH C2

CH OH2

CH OH2

CH OH2

13


The hydroxyl groups present on the two entrances of CD at position 2, 3, and 6 of each glucopyranose can be modified, by chemical reactions, to produce CD derivatives with different propertiesthan those of the parent ones. By manipulating the reaction type, theoperational reaction conditions, and the ratio of the reagents, themodification can be obtained in only one or more than one hydroxylgroup. Reaction with hydroxyl groups alters the properties of the CD,allowing

• improved solubility of the CD

• formation of different bonds with analytes that can improve theinclusion-complexation

• analysis of uncharged optical isomers

For example, the solubility of 2,6-di-O-methyl-β-CD and of thenegatively charged β-cyclodextrin polymer in water were found to be57 g/100 ml73 and >40 g/100 ml, respectively, relatively high in comparison to that of β-CD (1.85 g/100 ml). Furthermore, the cavity of di-OMe-β-CD is deeper than its parent due to the presenceof methoxy groups at the entrances, and thus is more hydrophobic.Considering other derivatives like carboxymethyl, methyl amino,phosphate, and sulfate, it is easy to recognize the advantages in analytical chemistry, especially in CE. In fact, such types of CDs canbe charged to allow the optimization of the separation, e.g., migrationin the opposite direction of the analyte, or introduction of charge toneutral compounds through complexation (for a comprehensive reviewof CD derivatives, see reference 74).

Inclusion complexation and stereoselectivity are influenced by sev-eral experimental parameters, such as CD type and concentration,applied voltage, capillary temperature and length, ionic strength,organic solvent, electro-osmotic flow, and polymeric additives. Someeffects of these parameters are discussed in this section, with the aimof describing the appropriate experimental conditions when enan-tiomers must be separated using cyclodextrins.

CD Type and Concentration

The first requirement for inclusion complexation is fitting theanalyte into the CD cavity, and thus CD selection must be made considering the shape of the analyte. As previously discussed, theinternal diameters of native CDs increase by increasing the numberof glucose units γ>β>α. Thus, selecting the appropriate CD is related to the shape and dimensions of the analyte. As an example,

14


consider the chemical structure of terbutaline and nicergoline (see Figure 9a and 9b). Terbutaline is currently used in the treatmentof asthma diseases, and nicergoline is used as a vasodilator in acutemyocardial infarction with diastolic hypertension. Terbutaline containsonly one aromatic ring and one asymmetric carbon at the α position,while nicergoline is formed by more than one aromatic ring and possesses several chiral centers.

Fig. 9. Chemical structure of (A) nicergoline and (B) terbutaline.

Figures 10a, 10b, and 10c show the electropherograms of theseparation of terbutaline and nicergoline into their enantiomers bycapillary zone electrophoresis when α, β, and γ-CD were separatelyadded to a BGE at pH 2.5.

OH

OHHO

_ _ _*CH CH NH CH(CH )2 23

B

CH2

CH3

CH3

H

H

CH3O

N

N

810

5

OCON

BrA

15


Fig. 10. Electropherograms of the enantiomeric separation of racemicterbutaline and nicergoline using different types of cyclodextrins aschiral selectors by CE. Apparatus, BioFocus 3000 system; capillary, (coated)35 cm x 0.05 mm ID; background electrolyte, 0.1 M phosphate buffer +a) 30 mM α-CD, b) 20 mM β-CD and c) 30 mM γ-CD; applied voltage 15 kV;injection, electrokinetic 7 kV, 7 s of 5 x 10-5 M of racemic terbutaline or nicergoline.

12 13 14 15 16

D,L-nicergoline(+.-)-terbutaline

α-cyclodextrin

6 7 8 9 10

time (min) time (min)

LD

nicergolineterbutaline

(-)(+)

9 10 11 12 13 10 11 12 13 14

time (min)time (min)

β-cyclodextrin

(+,-)-terbutaline nicergoline

L

D

γ-cyclodextrin

12 13 14 15 16

time (min)time (min)

6 7 8 9 10

B

C

A

16


Clearly α-CD was not able to separate both racemic mixturesinto their enantiomers because its cavity is too small to guest the twocompounds. β-CD allowed the resolution of terbutaline, while nicer-goline was partly resolved, and γ-CD was not able to guest terbutaline(the molecule is too small) but was very effective for nicergoline.

In a previous study it has been shown that for the enantiomeric separation of terbutaline, either β-CD and dimethylated-β-CD canbe successfully used in CE but the latter forms the strongest inclusioncomplexes and is more stereoselective than the former. The complexation of terbutaline increased with increasing concentrationof both β-CD and its derivative, while the resolution showed a maxi-mum at 15 mM and 5 mM for β-CD and dimethylated-β-CD,respectively.15 Figure 11 shows the effect of the concentration of β-CDand di-O-methyl-β-CD on the resolution of (+) and (-)-terbutaline.

Fig. 11. Effect of the concentration of β-cyclodextrin and di-O-methyl-β-cyclodextrin on the resolution of racemic terbutaline into theirenantiomers.

It is important to consider the degree of substitution (DS) of thecyclodextrin used. The effect of DS was demonstrated for the chiralseparation of some hydroxyacids using different β-CDs modifiedwith methyl amino47 or hydroxypropyl groups75 where the presenceof more than one substituent can strongly affect the chiral recognition.The CD type can also influence the migration order of the two enan-tiomers; this effect was observed for the separation of dansyl-norvalineenantiomers using trimethylated-β-CD and β-CD.76 Other authorshave discussed the influence of the concentration of CD in CZE, indicating that the concentration of the CD can play a very importantrole for chiral resolution by CE, and thus this parameter should be

β-CD

0 5 10 15 20

1

2

3

0

cyclodextrin (mM)

reso

luti

on

(R

)

1

2

3

0

2,6-di-OMe-β-CD

17


carefully controlled in order to find the optimal experimental conditions.8,34,38 Wren and Rowe discussed the optimum concentrationof CD for enantiomeric separation and found that it is dependentupon the formation constants of the diastereoisomers formed duringthe electrophoresis (equation 1).77

where CCD is the optimum concentration of the CD used, and K isthe formation constants of the two enantiomers (+) and (-). The mathematical model was verified for the enantiomeric separation ofpropranolol using β-CD or its dimethylated derivative.

Analyte shape

The shape of the analytes is another important parameter to be considered for inclusion-complexation with cyclodextrin. This effectwas studied for the resolution of several tryptophan derivatives (methyl,hydroxyl) where the position of the substituent group on the indole ringstrongly influenced the stability of the complex formed, and thus thechiral recognition.30 The shape of the enantiomeric pairs can be modified, for example, by derivatizing with hydrophobic achiralcompounds that can be included into the CD cavity. Amino acids asdansyl-, naphthalene-2,3-dicarboxaldhyde- (NDA), and naphthy-lamide derivatives, and monosaccharides derivatized as naphtalenesulfonate, were resolved into their enantiomers using cyclodextrins.78–81

Charge of CD

The charge of the cyclodextrin can play an important role in theresolution mechanism when the electrophoretic separation of enan-tiomers has to be carried out. In this case the electrostatic interactionwith the analytes, the movement of the chiral selector in the oppositedirection of the two enantiomers, and the possibility for separatinguncharged compounds (inducing the charge) are the main advantagesfor using such chiral selectors.

The use of charged cyclodextrins was first demonstrated byTerabe,53 who separated several racemic dansyl-amino acids using apositively charged β-cyclodextrin. In our laboratory we used a methylamino-β-cyclodextrin for the enantiomeric resolution of several α-hydroxyacids47 and a chargeable β-cyclodextrin polymer(carboxymethylated-β-CD polymer, CM-β-CD pol)55 for the separa-tion of racemic basic compounds into their enantiomers.

[ C CD ] = _______1[ K (+) K (-) ] 1/2

(1)

18


The inversion of migration order of studied enantiomers wasdemonstrated55,82 using a charged chiral selector in the presence ofelectro-osmotic flow; this approach can be advantageously employedwhen the enantiomeric impurity migrates behind the main isomer. Inthis case, it was necessary to increase the injected amount of the sample in order to detect the impurity (< 1%), and a loss of resolutionoccurred.

Electro-osmotic flow, concentration of the chargeable cyclodex-trin, and the pH of the BGE have to be controlled to optimize the separation method.

When carboxymethylated-β-CD polymer is used in CE, the weaklyacidic carboxylic groups can be charged or uncharged, dependingupon the pH of the BGE. At low pH, 2–3, the carboxylic groups ofthe CD are protonated, the chiral selector behaves as a quasi stationary phase, and the groups form hydrogen bonds with analytes.At pH >3.5, the CD is charged due to the dissociation of the car-boxylic substituents and migrates with its own mobility, while forminginclusion complexes and allowing ion-pair interactions with analytes.

Recently a new modified β-cyclodextrin (sulfobutyl-ether(IV)-β-cyclodextrin, SBE-β-CD) has been synthesized and characterized byCE with indirect UV detection.83 This modified β-CD has been usedin CE for the enantiomeric separation of several classes of compounds (positively charged, negatively charged, or uncharged).As shown in Figure 12, the SBE-β-CD contains four modified primaryhydroxyl groups (position 6 of four glucose units) with a butyl chainand sulfonic groups. Due to its chemical properties, the modified CDis negatively charged at any commonly used pH in CE. These featuresallow its use in a charged mode over a wide pH (2–11) range.

Fig. 12. Chemical structure of sulfobutyl-ether(IV)-β-cyclodextrin (SBE-β-CD).

O

OO

OSO3

SO3

SO3

SO3

-

-

-

-

HOHO

32

6

19


The chiral selector was used at pH 10, close to the pK of severalephedrines, to obtain good enantiomeric resolution.48 Operation atpH 2.5 allowed the enantiomeric separation of several basic com-pounds of pharmaceutical interest, including thalidomide,dimethindene, and mefloquine.49 Recently we showed the usefulnessof this new chiral selector for the enantiomeric separation of severalpositively and negatively charged compounds of pharmaceuticalinterest as well as uncharged phenyl-alcohols and dansyl-amino acids.52

In the case of acidic compounds, ion-pair interactions are not involvedin the chiral resolution mechanism, but electrostatic repulsion can beresponsible for the different stability constants of the two diastereoiso-mers formed during the electrophoretic process. Figure 13 shows theelectropherograms of the enantiomeric separation of warfarin, bupivacaine, 2-phenyl-2-butanol and dansyl-phenylalanine.

Fig. 13. Electropherograms of the enantiomeric separation of warfarin,bupivacaine, 2-dansyl-phenylalanine, and phenyl-2-butanol. Experimentalconditions: apparatus, BioFocus 3000 system; capillary, 50 cm x 0.05 mmID; applied voltage, 15 kV. Background electrolyte: A and B, 50 mM phos-phate buffer (pH 6) and SBE-β-CD 6 and 10 mg/ml, respectively; C, 50 mMTris/HCl, pH 8, with 3 mg/ml of SBE-β-CD; D, 50 mM borate buffer, pH 9, and20 mg/ml SBE-β-CD; injection, 5 psi*s 10-4 M of each racemic compound.

9 10 11 12 13 14 4 5 6

bupivacainewarfarin BA

time (min) time (min)

2-phenyl-2-butanolDNS-phenylalanine

DL

C D

eof

6 7 8 9 5.8 8.8 11.8time (min)time (min)

20


Composition of the Background Electrolyte

Chiral resolution can be strongly influenced by the compositionof the BGE, and thus the selection of the appropriate buffer system,its concentration, its ionic strength, its pH, the presence of polymericadditives, and the content of the organic additive should be careful-ly considered. Decreasing the ionic strength of the BGE generally causesa reduction of migration time and resolution; peak tailing has alsobeen observed.34 The negative effect related to the low ionic strengthof the buffer is due to electromigration dispersion that can be avoid-ed using a BGE at a concentration 102 times higher than that of the sample. The peak shape can be controlled by selecting a co-ion withsimilar electrophoretic mobility to that of the analyte.

The addition of an organic solvent to the BGE can produce a neg-ative effect on the binding constant of the inclusion-complex withcyclodextrins due to the competition between the organic additive andthe analyte. The affinity of the analytes for the organic additive mustbe considered. However, organic solvents can improve the selectivi-ty of the enantiomeric separation by differential influence on thebinding constants of the two enantiomers. Propranolol enantiomerswere not separated even using 40 mM of β-CD (urea was added to theBGE in order to increase the solubility of the CD), while separation wasobtained when 30% (v/v) methanol was employed as a component ofthe buffer.15 Recently we demonstrated that the use of methanol cangreatly improve the stereoselectivity of the separation of racemic non-steroidal anti-inflammatory drugs (NSAIDs). Racemic flurbiprofenwas baseline resolved into its enantiomers at pH 5 using 30 mM of2,3,6-tri-O-methyl-β-cyclodextrin, but the drug was poorly resolvedwhen the CD concentration was lowered to 5 mM. The addition ofmethanol to the BGE containing 5 mM of chiral selector caused anincrease in migration time as well as an increase in resolution, allowingbaseline enantiomeric separation when the buffer contained 20%organic additive. The effect of methanol on the enantiomeric separationof propranolol and flurbiprofen is shown in Figure 14.

A theoretical model of the effect of organic solvent in the BGEwhen cyclodextrins are used as chiral selector in CE has been discussed by Wren.84 When the CD concentration was at or belowthe optimum value (maximum resolution), the addition of organicsolvent (methanol or acetonitrile) caused a reduction of resolution. Thiseffect was due to the change in the formation constants of the inclusion-complexes, modifying the optimum CD concentration.

21


Fig. 14. Effect of methanol added to the BGE on the resolution of R,S-propranolol and R,S-flurbiprofen. A, (without methanol) and B, (with 30%methanol); apparatus HPE 100 system, capillary, 20 cm x 0.025 mm ID;background electrolyte, 0.1 M phosphate pH 2.5, 4 M urea and 40 mM β-CD.C, (without methanol) and D, (with 20 % methanol); apparatus BioFocus3000 system; capillary, (coated) 36 cm x 0.05 mm ID; 0.1 M MES, pH 5,and 5 mM tri-OMe-β-CD; applied voltage 20 kV.

The influence of the pH of BGE was discussed by several authorsand a theoretical model has also been described.85

12.7 14.9 15.5 20.7

DS

RC

R, S - flurbiprofen

4 8 4 8 10

R

S

R S

BA

R,S - propranolol

A20

6

time (min)

22


Strong electro-osmotic flow is not recommended for good enantiomeric separation by CE, because the time spent by the analytefor inclusion-complexation equilibrium is reduced. A coated capillaryof short length can be used to improve the resolution or to perform theanalysis in a reasonable time.

For many applications in the separation of enantiomers of basiccompounds, acidic pH values have been used. At low pH the electro-osmotic flow is reduced, increasing the resolution.29 Similar effects canbe obtained by adding polymers such as hydroxyethylcellulose (HEC),hydroxypropylmethylcellulose (HPMC), or polyvinylalcohol (PVA)to the BGE, or by using capillaries coated internally, for examplewith polyacrylamide. The addition of polymers will reduce the eof andincrease the viscosity. In all cases, the adsorption of analytes is alsoavoided or reduced.

The pH of the BGE not only influences the electro-osmotic flowbut can also affect the charge of the analytes, and thus their effectivemobility. When the analysis of weakly acidic compounds must becarried out in the presence of electro-osmotic flow and a native CDis used, both CD and analytes (uncharged and charged, respectively)are transported to the detector by the electro-osmotic flow. The migra-tion time of the analyte in the presence of the CD is shorter than in itsabsence due to inclusion-complexation. Increasing the pH will causean increase in migration time due to increased analyte charge, andthis phenomenon can help the resolution. Of course, for the optimumexperimental conditions, the magnitude of the electro-osmotic flowmust be considered.

Capillary Temperature

Efficient temperature control is recommended to prevent loss inresolution when cyclodextrins are used as chiral selectors for enantiomeric separations. This can be easily obtained using a BioFocuselectrophoresis system, which provides a system for circulating liquidcoolant into the cartridge containing the capillary.

Temperature increases cause a decrease in buffer viscosity, and thusa decrease in migration time. Furthermore, a change in the capillarytemperature can strongly influence the stability of the inclusion complex formed between analyte and cyclodextrin (generally anincrease in temperature causes a decrease in the binding constant).86

This effect is shown in Figure 15, where the separation of nore-pinephrine, epinephrine, and isoproterenol enantiomers was obtainedusing a dimethylated-β-CD.87

23


Fig. 15. Influence of the capillary temperature on the resolution ofracemic norepinephrine (n-EPI), epinephrine (EPI), and isoproterenol (i-PRO) into their enantiomers. Apparatus, BioFocus 3000 system; capillary,(coated) 30 cm x 0.05 mm ID; background electrolyte, 50 mM phosphatebuffer, pH 2.5, and 20 mM of di-OMe-β-cyclodextrin; applied voltage 10kV; column temperature, A) 20 ¡C, B) 40 ¡C. C) resolution (R) vs. column temperature (modified from reference 87).

▲

▲▲

▲

i-PROEPIn-EPI

15 20 30 40 Co

R

3

2

1

0

A

A20

6

n-E

PI

EP

I

i-PR

O

0 3 6 9 12 min

A20

6

B

n-E

PI

EP

I

i-PR

O

0 3 6 9 min

C

24


Cyclodextrins or their Derivatives in Capillary GelElectrophoresis, Isotachophoresis, andElectrochromatography

Cyclodextrins or their derivatives have also been used as chiralselectors for enantiomeric separations with other capillary electrophoretic techniques, although there are fewer publications onthis than on CZE and micellar electrokinetic chromatography. Theuse of capillary gel electrophoresis (CGE) for the separation of enantiomers was first described by Guttman et al.8 A capillary, filledwith crosslinked polyacrylamide gel and also containing unbound α- or β- or γ-cyclodextrin, was used for the enantiomeric separationof several dansyl-amino acids. The electro-osmotic flow was suppressed by coating the capillary wall, and the separations wereperformed at pH 8.3. The CD type and its concentration, as well as thecolumn temperature, influenced the migration time and the resolutionof the amino acid derivatives studied. The addition of 10% methanolimproved the separation of enantiomers. After this pioneering work,gel filled capillaries were not employed for chiral separations in CEuntil 1992, when Cruzado45 used a modified β-CD (allyl carbamoylated-β-CD) copolymerized with acrylamide for the enan-tiomeric separation of several dansyl-amino acids and homatropine.The gel-modified cyclodextrin presented several drawbacks, includ-ing short lifetime and poor reproducibility.

The feasibility of enantiomeric separation using a chiral stationary phase (CSP) in CE has been shown by Mayer and Schurig.43

The CSP was bound to the capillary wall after static thermal coatingwith Chirasil-Dex containing monokis-6-O-octamethylene-per-methyl-β or γ-cyclodextrin. The coating stability was tested bygas chromatography and by CE.

The separation of several racemic mixtures into their enantiomerswas obtained with a borate/phosphate buffer at pH 7 in presence ofelectro-osmotic flow. Recently, the enantiomeric separation of bothneutral and charged analytes, e.g., benzoin, hexobarbital, and sever-al amino acid derivatives, has been achieved using a β-CD bonded chiral stationary phase (CSP with 5 µm particle diameter employed inHPLC).88 The type of BGE influenced the direction of the electro-osmotic flow. The results obtained suggested perspectives on the useof a wide number of chiral phases.

Enantiomeric separations have also been performed by capillaryisotachophoresis (ITP) employing cyclodextrins or their derivativesas chiral additives to the leading electrolyte (LE). The first paper bySmolkova-Keulemansova’s group89 dealt with the separation of some

25


pseudo ephedrine alkaloid enantiomers using β-CD or di-OMe-β-CD as a chiral additive to the LE. In this study, the CD type andconcentration, and the pH of the LE, influenced the efficiency of theseparation. After this pioneering study the same group showed theenantiomeric resolution of some phenothiazines with β- or γ-CD anddiscussed the importance of the counter-ion type on the ITP separation.90

Ketotifen and its intermediate enantiomers91 have also been separatedby this group using ITP. Furthermore the possibility of using a coupled column system, in which the two capillaries contained twodifferent cyclodextrins, for a two-step separation has been demon-strated. The authors called the system two-dimensional separation.Unfortunately, this technique did not become popular, even though itoffered several advantages over CZE, including a concentrating effect,analysis of compounds in traces, and the possibility of working withlarger volumes of sample. Considering the advantages, we are convinced that in the near future ITP will be widely used, at least incombination with CZE, in order to improve the sensitivity of the separation method.

Chiral Crown-Ether

Another class of compounds used for enantiomeric resolution bythe inclusion-complexation mechanism is represented by crown-ethers. These macromolecules, discovered in 1967 by Pederson,93 areable to form inclusion complexes with several inorganic and organicions in which the guest compound fits the cavity, forming weak bonds(ion-dipole) with the etheroatoms (O, S) of the crown. Here the inclusion-complexation is completely different than that of CDs; thehydrophilic part of the analyte is included, while with CDs it is thehydrophobic part. Although the inclusion of the amino group is necessary for enantiomeric separation, other interactions are requiredin order to perform chiral recognition. Figure 16 shows the chemicalstructure of the chiral 18-crown-6-ether tetra carboxylic acid used incapillary zone electrophoresis.

Fig. 16. Chemical structure of 18-Crown-6-ether tetracarboxylic acid.

O

O

O

O

O

O

HOOC

HOOCCOOH

COOH

26


Two different chiral resolution mechanisms have been proposedby Kuhn94 in which the four carboxylic groups of the crown wereinvolved in the chiral recognition process, either forming electro-static interactions or causing a steric barrier effect with the includedmolecule.

The chemical structure of the analyte plays a very important rolefor chiral recognition with the modified crown-ether; in fact the presence of a primary amino group is fundamental for inclusion-complexation. For the barrier mechanism, nonpolar substituents on theasymmetric center have a strong influence. For example, naph-thylethylamine was resolved, while phenylethylamine was not at all.Enantiomers of aliphatic amino acids with branched chains were better separated than those with linear chains (valine > norvaline).Finally the distance of the amino group from the chiral center has tobe considered. Better resolution is obtained with short distances.

Other parameters which must be controlled for the optimizationof the chiral separation using chiral crown ether by CZE include thepH of the BGE, the organic additive, and the buffer composition (K+, NH4

+, Na+ ions should be avoided because they can competewith analytes in the inclusion process while Tris seems to be the mosteffective cation).94 Applications using 18-crown-6-ether tetra carboxylic acid as chiral selector in capillary zone electrophoresiscover a wide range of amino derivatives, including racemic aminoacids,32 amines and peptides,56 and amino alcohols and sympath-omimetic drugs.95

4.2 Ligand ExchangeThe ligand-exchange mechanism, first introduced by Davankov

in chromatography,96 was successfully applied in CE for the enan-tiomeric separation of dansyl amino acids57,58 and α-hydroxy acids.59

The separation is based on the interaction between a metal complexwith a chiral ligand, added to the BGE, and the two enantiomers to beresolved. L- or D-proline, L- or D-hydroxyproline and aspartame arethe ligands which have been employed in CE.

Figure 17 shows the separation of D and L-phenyllactic (3-PhL)acid when Cu(II)-L-hydroxyproline complexes were used as thechiral selector. The two analytes migrate as anions, toward the anodein the coated capillary, while the positively charged metal chelatemigrates in the opposite direction. The 3-PhL forms a ternary complexwith Cu(II), causing a reduction of the effective mobility of both enan-tiomers; the two complexes possess different stability constants and thusare separated by the end of the electrophoretic run. The optimization

27


of the separation method requires the control of several parameters,namely the metal and chiral ligand type, their ratio and concentration,the pH, the ionic strength, the organic additive and the column temperature.

Fig. 17. Electropherograms of the separation of D- and L-phenyllacticacid using a ligand-exchange mechanism by CE. Apparatus, BioFocus3000 system; capillary, (coated) 35 cm x 0.05 mm ID; background electrolyte,0.02 M phosphate buffer, pH 4.4, and copper(II) acetate (2 mM)/ L-hydrox-yproline (4 mM); applied voltage 15 kV, 15 A; injection 10 psi*s of 10-4 M ofracemic 3-phenyllactic acid; carousel and capillary temperature, 25 ¡C.

One drawback that can be expected when ligand exchange isused in CE is the UV absorption of the metal complex, which reducesthe sensitivity of the method. This drawback can be negated by usingdifferent approaches, e.g., by filling only part of the capillary or byusing a different detector such as laser induced fluorescence57 orconductivity.

4.3 Chiral MicellesThe use of micelles in capillary electrophoresis was first introduced

by Terabe et al. for the separation of neutral compounds. They calledthis technique micellar electrokinetic chromatography (MEKC).97

Surfactants such as sodium dodecyl sulfate (SDS) were added to theaqueous solution in order to improve the selectivity of the separationof neutral analytes. With neutral-alkaline solutions, the electro-osmoticflow is relatively high and the bulk solution (analytes and micelles)moves toward the cathode. The migration of the negatively-chargedmicelles is retarded relative to neutral compounds. The analyte interacts with the two phases (aqueous and micellar) and thus its

0.0045

0.0035

0.0025

0.0015

0.0005

0

-0.00052 4 6 8

time (min)

abso

rban

ce (

206

nm

) DL

28


migration is also retarded. The extent of the retardation is a functionof the percentage of analyte distributed into the micellar phase.

Chiral recognition can be achieved in MEKC by including achiral selector in the micellar system, either bound to the micellesor added to the electrolyte.

Sodium N-dodecanoyl-L-valinate (SDVal) and sodium N-dode-canoyl-L-alaninate (SDAla) were applied as chiral micelle selectorsfor the separation of several amino acid derivatives, namely 3,5-dini-trobenzoyl, 4-nitrobenzoyl and benzoyl-O-isopropyl esters.98 SDValwas also used for the enantiomeric separation of phenylthiohydantoinamino acids by Terabe’s group. Addition of methanol to the SDValmicellar solution as an organic modifier improved resolution, andthe addition of urea significantly improved peak shapes.99, 100

Other very interesting chiral micelles are represented by bilesalts, introduced for the first time in MEKC by Terabe et al.7 These surfactants, which possess a steroid nucleus and a side chain carryinga carboxylic group that can be conjugated with taurine, have bothhydrophobic and hydrophilic properties. When taurine is conjugated,due to the presence of sulfonate groups, the chiral selector is chargedeven at low pH. Figure 18 shows the chemical structure of the mainbile acid salts used in CE.

Fig. 18. Chemical structure of bile acid salts.

OHH

OH

COR2

1R

Bile acidsAbbreviation R R1 2(sodium salts)

Cholate SC OH ONa

Taurocholate STC OH NHC H SO Na2 4 3

Deoxycholate SDC H ONa

NHC H SO Na2 4 3HTaurodeoxycholate STDC

29


Sodium taurocholate (STC) and taurodeoxycholate have beenused successfully as chiral selectors for the separation of several dansyl amino acids into their enantiomers.7 The reduction of the electro-osmotic flow by decreasing the pH to 3 produced good separations not obtainable under other experimental conditions. Thesame class of chiral surfactants proved to be useful chiral selectors forthe separation of several racemic drugs into their enantiomers, e.g., diltiazem and trimetoquinol,61,101 fonoldopam, 4-hydrox-ymephenytoin,102 laudanosine,101 tetrahydropapaveroline,103 naphtoland binaphthyl derivatives.62 Digitonin,100 glycirrhizic acid, and β-escin104 are natural surfactants which have also been used for chiral separations at a relatively low pH, and also combined withSDS in mixed micelle systems.

A very promising approach in MEKC is the use of chiral micellepolymers. Poly(sodium N-undecylenyl-L-valinate) has been shown tobe a good chiral selector for the separation of 1-1’-bi-2-naphtol andlaudanosine into their enantiomers. The advantages of such polymersover the monomers are their enhanced stability and rigidity.Furthermore, since the polymicelles have no CMC, they can be usedat any concentration. Resolution was influenced by the polymer concentration and the pH. At low pH the charged polymerized micelleshave a compact conformation, while at higher pH they lose it.105

The addition of cyclodextrins to the micellar phase (CD-modifiedMEKC) enables the enantiomeric separation of uncharged compoundswhich could also be analyzed with charged cyclodextrins or withelectrochromatography.

4.4 Affinity Interactions

Proteins as Chiral Selectors

Proteins are natural biopolymers with helical conformation ableto interact selectively with a wide number of compounds of smallsize, such as pharmaceuticals. They have been used successfullyas chiral selectors in capillary electrophoresis. The enantiomeric separation is based on stereoselective interactions between the protein and the two analytes, with the formation of two labilediastereoisomeric complexes during the electrophoretic run. Tanakaet al.64 discussed a theoretical model for enantiomeric separationwith proteins where the electrophoretic mobility of the free analyte(µs) and of the protein (µ p), and the concentration of the protein (P),are correlated by the following equation:

30


where µapp, µeo, and K are the apparent mobilities of the analyte, electro-osmotic flow, and binding constant, respectively.

Calculating ∆µapp= µapp1- µapp2, the following equation providesthe optimum concentration of protein for the best chiral resolution (see equation 3)

The theoretical model is similar to that discussed by Wren, wherean uncharged chiral selector was used.77 The Tanaka model was verified for the enantiomeric separation of several acidic compoundsusing avidin as a chiral selector. Avidin, a basic protein with a pI of10.0–10.5, is positively charged at acidic/neutral pH and migrates ina direction opposite that of the analytes.

When proteins are used as chiral selectors in CE, several draw-backs have to be expected. The protein may adsorb on the capillarywall, the sensitivity in the low UV region can be reduced due to strongabsorbtion of the chiral selector, and the protein may not be stable underthe operating conditions. Coated capillaries were used in order toreduce the adsorption of the protein to the wall and to reduce theelectroosmotic flow.

Figure 19 shows schematically the separation principle of affinityEKC for analysis of an acidic compound with avidin at acidic pH.

A-A-

A-

A-

A-

HA

++

++

++

+++ +

++ +

+ –

[P] op = (K1 K2)-1/2(3)

µapp = µeo + ______1 + K[P]

µs + µpK[P] (2)

31


Fig. 19. Electrophoretic separation scheme for anions using avidin asa chiral selector (with permission from reference 48).

The separated enantiomeric compounds were vanilmandelicacid, warfarin, ibuprofen, ketoprofen, flurbiprofen, and the twodiastereoisomers of folinic acid. Several parameters were investigated,e.g., the concentration of the protein, the pH of the BGE, the content ofthe organic solvent and the capillary temperature. As shown in Figure20, increasing the concentration of avidin from 0 to 25 µM gave bet-ter resolution of the enantiomeric pair of vanilmandelic acid at pH4.

A0.002

0.001

0.000

-0.001

Abs

orba

nce

8.0 9.0 10.0 11.0 12.0 13.0Time (min)

B0.002

0.001

0.000

-0.001

Abs

orba

nce

8.0 9.0 10.0 11.0 12.0 13.0Time (min)

C0.001

0.000

-0.001

Abs

orba

nce

8.0 9.0 10.0 11.0 12.0 13.0Time (min)

32


Fig. 20. Effect of the concentration of avidin on the enantiomeric separation of racemic vanilmandelic acid. Apparatus, BioFocus 3000system; capillary, (coated) 36 cm x 0.05 mm ID; background electrolyte,0.05 M phosphate buffer, pH 4, and avidin A) 0 M; B) 10 M and C) 25 M;applied voltage -12 kV; detection at 240 nm; injection, 1 psi*2s (sample 0.4 mg/ml five fold dilution) (with permission from reference 48).

It is worthwhile to mention here an interesting approach used byValtcheva et al., employing the enzyme cellulase for the enantiomericseparation of several pharmaceutical compounds such as propranolol,pindolol, metoprolol, and labetalol.10 The chiral selector was negativelycharged at the operating pH (pH=5.1), and the analytes migrated inthe direction opposite to the enzyme. To optimize the method, thedetector window was chiral selector-free, improving the detectionlimit and allowing use of a relatively high concentration of buffer(0.4 M) supplemented with 2-propanol.

Interactions with Saccharides (Linear Polymers) andAntibiotics

Linear polymers of α-(1–4)-linked D-glucose, termed maltodextrins, have proven to be good chiral selectors in CZE forthe enantiomeric separation of non-steroidal anti-inflammatory drugs 2-APA NSAIDs (flurbiprofen, ibuprofen, carprofen, suprofen, andindoprofen) as well as coumarinic anticoagulant drugs.60,106

An anionic biopolymer, heparin, was studied for the enantiomericseparation of several drugs including antimalarials and antihistamines.Heparin is a di-, tetra-, or hexasaccharide composed of uronic acid andglucosamine with a helical conformation in aqueous solutions and amolecular weight of 10–30,000. The enantiomeric separations wereobtained at pH 4 and 5 and the chiral recognition was influenced bythe size of the analyte and by the electrostatic interactions.107

Another interesting class of compounds that can be used in CE aschiral selectors for enantiomeric separations is antibiotics. Vancomycin,a glycopeptide, was effective for the separation of enantiomeric dipep-tides D-Ala-D-Ala/L-Ala-L-Ala and their acetyl derivatives108 andfor the separation of racemic derivatized tri- and tetrapeptides;10 in bothstudies the binding constants have been calculated. A recent paperdemonstrated the applicability of the macrocyclic antibiotic rifamycinb for the separation of chiral compounds by affinity electrophoresis.110 Indirect UV detection was used because the antibiotic had strong absorption at low wavelengths.

33


4.5 Combination of Chiral/Chiral or Chiral/Achiral Selectors

Improvement in the selectivity of an enantiomeric separation canbe obtained in CE by employing mixed selector systems. This approachconsists of the use of a mixture of either chiral/achiral or chiral/chiralcompounds. In a wide number of chiral separations, the combinationof cyclodextrins with micelles was effective for the resolution of several uncharged compounds. Figure 21 shows schematically theelectrophoretic separation when CD-MEKC is employed.

Fig. 21. Scheme of the electrophoretic separation when CD-modifiedMEKC is used.

= cyclodextrin = SDS E = analyte = micelles

The system is composed of an ionic micelle (usually SDS) and anaqueous phase (buffer + cyclodextrin). The analytes move with the eof,and are distributed between the micellar and the aqueous phase. Thestereoselectivity is due to the interactions with CDs. The higher theinteraction of the analyte with the chiral selector, the lower the migration time.

The selectivity can be modified by controlling several parameters,including surfactant type and concentration, CD type and concentration,organic additive, buffer type and concentration, and pH.

CD-MEKC allows the separation of enantiomers of amino acids,as dansyl-78 or naphthalene-2,3-dicarboxaldehyde (NDA) deriva-tives,111 using β- and/or γ-CD, as well as compounds of pharmaceuticalinterest, such as mephenytoin and its hydroxy derivative in urine.112

Figure 22 shows the electropherogram of an extracted undeglu-coronidated urine sample from an extensive metabolyzer(phenotyping). The presence of S-4-hydroxymephenytoin and R-mephenytoin was confirmed by spiking the sample mixture with theracemic

0

0

0--

----

O

OO0

-

0

O--

0

0

0

0

0

00

0

O

O

O

O

-

--

-

-- -

-

--

----

O

O

O0

-

Eeof

E

EE

+E

E

34


compounds (Figure 22 A and B) and comparing the UV spectra of thepure enantiomers with the separated peaks (Figure 22 D and E).

Fig. 22. Electrophoretic separation and characterization of S-4-hydrox-ymephenytoin (S-4-OH-MEPH) and R-mephenytoin (R-MEPH).Undeglucoronidated urine sample from an extensive metabolizer. Singleand multiwavelength detection (A and C, respectively); in B the sample from A was spiked with the racemic compounds, while in D and E the spectra ofthe two detected compounds in the sample are compared with those of thepure enantiomers. (with permission from reference 112).

Another example of combining two chiral selectors, namely β-CDand L-tartrate, has been shown by us for the separation of severalstereoisomers of cobalt(III) complexes with ethylenediamine andamino acids like Gly, Pro, and Phe. The L(+)-tartrate alone did notallow the resolution of the enantiomers of Co[(en)2D-Phe]2+ andCo[(en)2L-Phe]2+ and the addition of β-CD provided satisfactoryresults. In this separation both inclusion- and outer-sphere complex-ation are involved in the separation mechanism.68

Section 5Optimization of the Method

When a racemic mixture has to be separated into its enantiomers,it is necessary to consider several parameters in order to select theoptimum electrophoretic experimental conditions. First the chemicalstructure has to be inspected to determine the following information:

S-4-OH-MEPH

R-MEPH

Wavelength (nm)200 240 280

E

D

Nor

mal

ized

Spe

ctra

24.016.08.0

Time (min)

-0.001

0.0154-OH-MEPH

-0.001

Abs

orba

nce

0.015

B

AS-4-OH-MEPH

S

R

R

S

R-MEPH

MEPH

Uric acid

Uric acid

192

192

Abs

orba

nce

0.010

-0.0018.0 16.0 24.0

Time (min)

C S-4-OH-MEPH

R-MEPH

192

297

Wav

elen

gth

(nm

)

35


36


When basic compounds must be analyzed (this case represents awide number of pharmaceutical compounds), an acidic pH of the BGEis selected and the separation can be performed either in a coated oruncoated capillary. The recommended buffers are phosphate or citrateat pH 2.5 and 3.5, respectively (50–100 mM). In some cases a pH oneunit lower than the pKa of the enantiomers allows the migration of analytes as cations, and a pH higher than 2.5–3.5 could be considered.

The use of a pH >4–5 can be helpful because it enables the enan-tiomeric separation and allows sufficient electro-osmotic flow for theanalysis of achiral impurities in the sample. Moreover the electro-osmotic flow should be controlled because it reduces the migration timeand thus minimizes the time spent by the analytes in contact with thechiral selector.

Increasing the BGE concentration generally improves the resolution of enantiomers, but increases the migration times due to ion-pairing effects and reduction of electro-osmotic flow. Furthermore,peak-tailing and electromigration dispersion can be avoided. An upperlimit for buffer concentration is dictated by the increase in currentand resulting Joule heating which reduces the efficiency of the sep-aration. Use of a zwitterionic buffer can be considered as a goodsolution to this problem.

When the enantiomeric separation of negatively charged compounds must be performed, the general rules for basic compoundsmust be considered. Selecting a pH for the BGE in the range 4.5–8 willcharge (negatively) the two enantiomers and produce a sufficientelectro-osmotic flow for movement of the analytes toward the cathode. The use of a coated capillary permits operation in a reversedpolarity mode, and inversion of migration can be obtained. This

Basic

ANALYTE

Acidic Uncharged

is

Possessaromaticrings

* no* one* several

αβ

Positionof thechiral centre

37


approach can be advantageously used when a minor component in amixture of enantiomers has to be quantified.

The separation of uncharged enantiomers does not represent aproblem in CE, because different electrophoretic mechanisms andchiral selector types are available. With uncharged analytes, a relativelystrong electro-osmotic flow is necessary to perform the analysis withina reasonable time, and a charged chiral selector should be employed.Carboxymethylated or ethylated-β-cyclodextrin, sulfobutyl-ether-β-cyclodextrin, or a charged β-cyclodextrin polymer could be effectivefor chiral recognition. By inducing a negative charge, a selectiveincrease in the migration times allows separation of the enantiomers.

Another approach for the enantiomeric separation of unchargedcompounds is the use of surfactants with chiral selectors, e.g., SDS-cyclodextrins. The partition of the analytes into the micellar phasecauses a selective migration of neutral compounds, while the chiralselector allows chiral resolution. A buffer at a pH in the range 6–9 containing 25–100 mM of SDS supplemented with variable amountsof cyclodextrins is recommended. The effective mobility should be controlled by changing the composition of the buffer, to select theappropriate experimental conditions. The electro-osmotic flow shouldbe controlled, and in some cases the addition of organic solvents canhelp the separation of enantiomers. If the uncharged compounds cannot be resolved into their enantiomers, different chiral selectorsshould be employed, e.g., bile salts.

The capillary temperature should also be controlled when per-forming enantiomeric separations because, in general, an increase intemperature will result in a reduction of resolution. This parameterinfluences both the viscosity of the BGE (reduction of migrationtime) and mass transfer kinetics of analytes and chiral selectors. Thuspreliminary studies should be carried out at 20–25 °C, and then alower temperature should be used to optimize the chiral separation.

The capillary length can also influence the resolution of two enan-tiomers, and thus its selection should be done considering parameterssuch as applied voltage and capillary type. The use of longer capillariescan increase the resolution as well as the analysis time. When coatedcapillaries are used and the electro-osmotic flow is reduced or eliminated, good enantiomeric separations can be obtained even withcapillaries of 20 cm length. For coated capillaries, lengths in the rangeof 20–35 cm are recommended, while with uncoated capillaries 50–70 cm lengths should be tried.

38


Selection of the Chiral Selector

The most simple chiral selectors that can be used for enantiomericseparation by CE are represented by native cyclodextrins becausethey are commercially available, not expensive, and several applica-tions have been published.

If the analyzed compound does not contain aromatic rings, α-cyclodextrin or its derivatives can be selected. The same choicecould be made if the analyte contains only one aromatic ring, unless substituents in ortho or meta position are present. Such substituentswill cause steric hindrance effects, responsible for poor inclusioncomplexation.

When the compounds possess in their structure two aromaticrings (e.g. condensed), β-CD has to be selected, while for more thantwo aromatic groups γ-CD could be the appropriate chiral selector.

If successful enantiomeric separation cannot be achieved usingnative cyclodextrins, the wide number of modified ones (e.g., dimethy-lated, trimethylated, hydroxypropylated) should be considered. Chargedcyclodextrins may also be investigated, selecting the appropriate electrophoretic mechanism and experimental parameters.

As a general suggestion, run the compound initially in the absenceof CDs and calculate the effective mobility. Then perform the runwith the same BGE, containing a relatively low amount of CD(2.5–5 mM) in order to verify a decrease in effective mobility. Ifcomplexation is observed but insufficient resolution is obtained,increase the concentration of the chiral selector until satisfactoryseparation is achieved. In the case of β-CD, concentration screeningis limited by its solubility (< 20 mM); solubility can be increased byaddition of urea (4–8 M) or methanol or ethanol (<30% v/v) to theBGE. Alternatively, modified cyclodextrins can be selected to improvethe solubility or to introduce different secondary stereoselective bonds.

When compounds containing primary amines in their chemicalstructures, (e.g., amino acids, peptides, drugs) have to be separated intotheir enantiomers, 18-crown-6-ether tetracarboxylic acid can be usedadvantageously. In this case attention should be paid to the selectionof the BGE (pH and cation type, the latter should not compete withanalytes in the inclusion-complex mechanism).

Bile salts with planar structures, namely sodium cholate, tauro-cholate, deoxycholate, and taurodeoxycholate, are recommended forchiral compounds with rigid skeletons. Optimization of the enan-tiomeric separation can be performed by modifying those parameters

39


involved in micellar systems, e.g., buffer type and concentration, concentration of the bile salt, pH, and organic solvent.

For the enantiomeric separation of amino acids or hydroxy carboxylic acids, cyclodextrins are the simplest chiral selector to beused; however a ligand exchange electrophoretic system can also beconsidered. Copper(II)-amino acid or aspartame complexes could beselected, but unfortunately the detection limit is reduced due to theabsorption of the chiral BGE.

Section 6Quantitation

Quantitative analysis of chiral compounds is a very interesting andimportant topic that, although it has been shown to be possible byCE, unfortunately up to now is not yet well established. However,considering the good results obtained by several authors, the appli-cability of CE will be extended.

When a racemic mixture of a certain compound is analyzed by CEusing an appropriate chiral selector, it should be separated into twopeaks whose area ratio should be 1.00. The electrophoretic analysisof a racemic mixture of epinephrine14 showed two peaks with two different areas (49% of (-)-epinephrine and 51% of its antipode). Inorder to explain the observed differences, we have to consider that during the electrophoretic process the formation of two diastereoiso-mers takes place, and due to the different association constants, adifferent UV absorbance co-efficient could be expected.14 Furthermore,the analyte residence time in the detection path is different, the second enantiomer spending a longer time than the first one becauseit possesses a lower velocity.113 In order to solve this problem, it hasbeen suggested that the measured areas be normalized to migrationtime,14,113 and/or that an internal standard (I.S.) be used.14

It has been shown that CE can be successfully used for quantita-tion of chiral impurities of drugs with good precision and accuracy.Quantitative analysis of trimetoquinol enantiomers by MEKC hasbeen carried out with STDC as a chiral selector. The calibration graphwas linear for concentrations of the R-antipode in the range 2.5–15%and less than 1% of this compound could be detected.101

Peterson and Trowbridge performed quantitative determination of(-)-epinephrine/(+)-epinephrine with and without an internal standard(pseudopehedrine). The correlation coefficient and the RSD werefound to be 0.9998 and 1.3%, respectively using the I.S., while

40


without I.S. they found 0.9992 and 3.3%, respectively. The authors analyzed a pharmaceutical formulation with different ages (5, 10,and 29 months) and found the presence of (+)-epinephrine in therange 1.2–2.3% (this antipode should not be present at all).

The analysis of the antiviral compound 2’-deoxy-3’-thiacytidine(BCH189), a nucleoside analog containing 0.3% of the undesired (-)-antipode, was done using a BGE at pH 2.5 and 50 mM of di-OMe-β-CD with a detection limit < 0.1%.114

A capillary electrophoretic method has been validated for thedetermination of the chiral purity of fluparoxan. Detector linearitywas quite good in the studied range 1.5–125% of a target concentration(1.25 mg/ml) and the limit of quantitation and detection were foundto be 1% and 0.3%, respectively. The concentration of β-CD usedfor the method was relatively high (150 mM) and thus urea and isopropanol were added to the BGE.115

Section 7Applications

Chiral capillary electrophoresis has been applied, until now, mainlyfor enantiomeric separation of compounds of pharmaceutical interest.Resolution mechanisms, binding constants, and parameters affectingchiral analysis have been widely discussed. Several applications deal-ing with biological fluids are also available, as well as those dealingwith quantitative analysis. The former topic is only at the beginning,and we are convinced that in the near future will be extensively devel-oped.

For simplicity and to avoid repetition of some information givenpreviously, Table 2 presents selected applications for enantiomericseparation of compounds of pharmaceutical interest.

41


Table 2. Selected Applications on the EnantiomericSeparation of DrugsCompound Chiral selector CE type References

adrenaline di-OMe-β-CD CZE 14, 116

atenolol di-OMe-β-CD CZE 41, 42tri-OMe-β-CD

basic drugs macrocyclic CZE 67(terbutaline, isoproterenol, antibioticmetaproterenol, synephrine, (rifamycin B)metanephrine, salbutamol, epinephrine, norphenylephrine, ephedrine, ψ-ephedrine, octopamine, )

basic drugs negative β-CD CZE 55(nor-phenylephrine, ketamine, polymernor-ephedrine, ephedrine, epinephrine, terbutaline, propranolol)

basic drugs (S)-N- MEKC 117(atenolol, bupivacaine, dodecanoylvalineephedrine, homatropine, (S)-N-dodecoxy-ketamine, metoprolol, carbonylvalineN-methylpseudoephedrine, norephedrine, octopamine, pindolol, terbutaline)

bupivacaine di-OMe-β-CD CZE/MEKC 40

carvediol di-OMe-β-CD CZE 40

chloramphenicol di-OMe-β-CD CZE 38

chlorpheniramine β-CD CZE or MEKC 118, 119

cicletanine γ-CD MEKC 120

clenbuterol β-CD CZE 34

clenbuterol SBE-β-CD CZE 49

cyclophosphamide glycoprotein affinity 121

diltiazem STDC MEKC 103

dimethindene SBE-β-CD CZE 49

ephedrine di-OMe-β-CD CZE 13

epinephrine di-OMe-β-CD CZE 13, 116

ergot alkaloids γ-CD CZE 36(lisuride, meluol, nicergoline,terguride)

herbicides α-, β-, γ-, di-OMe- MEKC 122(phenoxy acids) β-CD

hexobarbital Chirasil-β-CD EHC 123

mefloquine SBE-β-CD CZE 49

mephenytoin SDS-β-CD MEKC 102, 112hydroxy metabolite β-CD-STDC

methotrexate vancomycin CZE 66

mianserine SBE-β-CD CZE 49

42


(continued on the next page)

Table 2. (continued)Compound Chiral selector CE type References

naproxen HP-β-CD CZE 124

NSAIDs (APA) vancomycin CZE 66(flurbiprofen, indoprofen, carprofen, ketoprofen, suprofen, fenoprofen, ibuprofen, naproxen)

NSAIDs (APA) tri-OMe-β-CD CZE 125(flurbiprofen, ketoprofen, fenoprofen, ibuprofen, suprofen)

NSAIDs (APA) Chirasil-β-CD EHC 123(ibuprofen, cicloprofen, flurbiprofen, carprofen)

NSAIDs (APA) avidin affinity 64(warfarin, ibuprofen, ketoprofen, flurbiprofen)

octopamine di-OMe-β-CD CZE 87

quinagolide crown-ether CZE 94

racemethorphan, racemorphan SDS-β-CD MEKC 126

salbutamol di-OMe-β-CD CZE 127(and its chiral and achiral impurities)

thalidomide SBE-β-CD CZE 33, 49

trimetoquinol dextran sulfate affinity MEKC 128and its isomers

trimetoquinol STDC MEKC 103

warfarin BSA affinity 63

warfarin SD-Val MEKC 129

warfarin SBE-β-CD CZE 52

warfarin di-OMe-β-CD CZE 130

Section 8Conclusions

The enantiomeric separation of different classes of compounds canbe obtained easily and rapidly by CE techniques, with good repro-ducibility and at a relatively low cost. Direct and indirect separationmethods can be successfully applied, but the former seems to be themost popular due to its simplicity.

Cyclodextrins have become the most popular chiral selectors inCE, since they are stable and quite soluble in the buffers commonly

43


used in CE, and because the faster equilibration of the CD-solutecomplex provides high efficiency and good peak symmetry. Thehydroxyl groups of the CD (primary and secondary) can be easilymodified, using chemical reactions, to obtain CDs with different prop-erties, thus extending the application range by allowing differentelectrophoretic separation mechanisms. However, several other chiralselectors are available, e.g., chiral crown ethers for compounds containing amino groups, proteins, chiral surfactants, saccharides, andrecently, antibiotics. It has been shown that the enantiomeric resolu-tion is strongly influenced by several parameters, including chiralselector type and concentration, composition of the BGE (ionic strength,ion type and concentration, pH, organic solvent), polymeric additivesin the BGE, applied voltage, and capillary temperature.

The advantages of CE over other separation methods, e.g. HPLC,include higher efficiency, shorter analysis time, and lower costs (onlya few µL of chiral buffer are required). Poor detection limits and limited preparative capability are the main disadvantages not yetresolved; in fact the loading of relatively high amounts of sample isnot possible in CE. Some solutions to this problem have been shown,e.g., the combination of ITP and CE.

Section 9References

1. Debowski, J., Sybilska, D. and Jurczak, J., Resolution of some chiral mandelicacid derivatives into enantiomers by reversed-phase high performance liquidchromatography via α- and β-cyclodextrin inclusion complexes, J. Chromatogr.,282, 83 (1983).

2. Ward, T. J. and Armstrong, D. W., Improved cyclodextrin chiral phases: acomparison and review, J. Liq. Chromatogr., 9, 407 (1986).

3. Blaschke, G., Chromatographic resolution of chiral drugs on polyamides andcellulose triacetate, J. Liq. Chromatogr., 9, 341 (1986).

4. Singh, N. H., F. N. Pasutto, F. N., Coutts, R. T. and Jamali, F., Gas chro-matographic separatin of optically active anti-inflammatory 2-arylpropionicacids using (+) or (-)-amphetamine as derivatizing reagent, J. Chromatogr., 378,125 (1986).

5. Kobor, F. and Schomburg, G., 6-tert-butyldimethylsilyl-2,3-dimethyl-alpha-cyclodextrin, beta-cyclodextrin, and gamma-cyclodextrin, dissolved inpolysiloxanes, as chiral selectors for gas chromatography – influence of selectorconcentration and polysiloxane matrix polarity on enantioselectivity, HRC-J.High. Res. Chromatogr., 16, 693 (1993).

6. Armstrong, D. W., Faulkner, Jr. and Han, S. M., Use of hydroxypropyl andhydroxyethyl-derivatized β-cyclodextrins for the thin-layer chromatograph-ic separation of enantiomers and diastereomers, J. Chromatogr., 452, 323(1988).

7. Terabe, S., Shibata, H. and Miyashita, Y., Chiral separation by electrokineticchromatography with bile salt micelles, J. Chromatogr., 480, 403 (1989).

44


8. Guttman, A., Paulus, A., Cohen, A. S., Grinberg, N. and Karger, B. L., Useof complexing agents for selective separation in high performance capillaryelectrophoresis, chiral resolution via cyclodextrins incorporated within poly-acrylamide gel column, J. Chromatogr., 448, 41 (1988).

9. Schmitt, T. and H. Engelhardt, H., Charged and uncharged cyclodextrins as chi-ral selectors in capillary electrophoresis, Chromatographia, 37, 475 (1993).

10. Valtcheva, L., Mohammed, J., Pettersson, G. and Hjerten, S., Chiral separa-tion of b-blockers by high performance capillary electrophoresis based onnon-immobilized cellulase as enantioselective protein, J. Chromatogr., 638,263 (1993).

11. Khun, R., Erni, F., Bereuter, T. and Hausler, J., Chiral recognition and enan-tiomeric resolution based on host-guest complexation with crown-ethers incapillary zone electrophoresis, Anal. Chem., 64, 2815 (1992).

12. Belder, D. and Schomburg, G., Chiral separations of basic and acidic com-pounds in modified capillaries using cyclodextrin-modified capillary zoneelectrophoresis, J. Chromatogr. A., 666, 351 (1994).

13. Fanali, S., Separation of optical isomers by capillary zone electrophoresisbased on host-guest complexation, J. Chromatogr., 474, 441 (1989).

14. Fanali, S., and Bocek, P., Enantiomeric resolution by using capillary zoneelectrophoresis: resolution of racemic tryptophan and determination of theenantiomer composition of commercial pharmaceutical epinephrine,Electrophoresis, 11, 757 (1990).

15. Fanali, S., Use of cyclodextrins in capillary zone electrophoresis. Resolutionof terbutaline and propranolol enantiomers, J. Chromatogr., 545, 437 (1991).

16. Foret, F., Krivankova, L. and Bocek, P., Capillary Zone Electrophoresis, VCHVerlagsgesellschaft mbH, Weinheim-New York-Basel-Cambridge-Tokyo,1993.

17. Snopek, J., Jelinek, I. and Smolkova-Keulemansova, E., Chiral separation byanalytical electromigration methods, J. Chromatogr., 609, 1 (1992).

18. Snopek, J. and Smolkova-Keulemansova, E., in D. Duchene (Editor), Newtrends in cyclodextrins and derivatives, Edition de Santé, Paris, p. 483, 1991.

19. Fanali, S., in N. A. Guzman (Editor), Capillary electrophoresis technology,Marcel Dekker, Inc, New York-Basel-Hong Kong, p. 731, 1993.

20. Ward, T. J., Chiral media for capillary electrophoresis, Anal. Chem., 66, A632(1994).

21. Novotny, M., Soini, H. and Stefansson, M., Chiral separation through capillaryelectromigration methods, Anal. Chem., 66, 646A (1994).

22. Rogan, M. M., Altria, K. D. and Goodall, D. M., Enantioselective separa-tions using capillary electrophoresis, Chirality, 6, 25 (1994).

23. Terabe, S., Otsuka, K. and Nishi, H., Separation of enantiomers by capillaryelectrophoretic technique, J. Chromatogr., 666, 295 (1994).

24. Fanali, S., Cristalli, M., Vespalec, R. and Bocek, P., in A. Chrambach, M. J.Dunn and B. J. Radola (Editors), Advances in electrophoresis, VCHVerlagsgesellschaft mbH, Weinheim-New York-Basel-Cambridge-Tokyo, p.3, 1994.

25. Fanali, S., Chiral separations by capillary electrophoresis, Electrophoresis,15, 753 (1994).

26. Cahn R. S., Ingold, C. K. and Prelog, V., Specification of molecular chirality,Angew. Chem. Int. Ed. Engl., 5, 385 (1966).

45


27. Schutzner, W., Fanali, S., Rizzi, A. and Kenndler, E., Separation of diastere-omeric derivatives of enantiomers by capillary zone electrophoresis with apolymer network: use of polyvinylpyrrolidone as buffer additive, J. Chromatogr.,639, 375 (1993).

28. Schutzner, W., Caponecchi, G., Fanali, S., Rizzi, A. and Kenndler, E., Improvedseparation of diastereomeric derivatives of enantiomers by a physical networkof linear polyvinylpyrrolidone applied as pseudophase in capillary zone electrophoresis, Electrophoresis, 15, 769 (1994).

29. Terabe, S., Otsuka, K. and Nishi, H., Separation of enantiomers by capillaryelectrophoretic techniques, J. Chromatogr. A., 666, 295 (1994).

30. Nardi, A., Ossicini, L. and Fanali, S., Use of cyclodextrins in capillary zoneelectrophoresis for the separation of optical isomers. Resolution of racemic tryp-tophan derivatives, Chirality, 4, 56 (1992).

31. Tanaka, M., Asano, S., Yoshinago, M., Kawaguchi, Y., Tetsumi, T. and Shono,T., Separation of racemates by capillary zone electrophoresis based on complexation with cyclodextrins, Fresemius J. Anal. Chem., 339, 63 (1991).

32. Kuhn, R., Stoecklin, F. and Erni, F., Chiral separations by host-guest com-plexation with cyclodextrin and crown-ether in capillary zone electrophoresis,Chromatographia, 33, 32 (1992).

33. Aumatell, A., Wells, R. J. and Wong, D. K. Y., Enantiomeric differentiationof a wide range of pharmacologically active substances by capillary elec-trophoresis using modified cyclodextrins, J. Chromatogr. A, 686, 293 (1994).

34. Altria, K. D., Goodall, D. M. and Rogan, M. M., Chiral separation of β-aminoalcohols by capillary electrophoresis using cyclodextrins as buffer additives.I. Effect of varying parameters, Chromatographia, 34, 19 (1992).

35. Nielen, M. W. F., Chiral separation of basic drugs using cyclodextrin-modified capillary zone electrophoresis, Anal. Chem., 65, 885 (1993).

36. Fanali, S., Flieger, M., Steinerova, N. and Nardi, A., Use of cyclodextrins forthe enantioselective separation of ergot alkaloids by capillary zone electrophoresis, Electrophoresis, 13, 39 (1992).

37. Belder, D. and Schomburg, G., Modification of silica surfaces for CZE byadsorption of nonionic hydrophilic polymers or use of radial electric fields,J. High Resol. Chromatogr., 15, 686 (1992).

38. Snopek, J., Soini, H , Novotny, M., Smolkova-Keulemansova, E. and Jelinek,I., Selected applications of cyclodextrin selectors in capillary electrophoresis,J. Chromatogr., 559, 215 (1991).

39. Sepaniak, M. J., Cole, R. D. , and Clark, B. K., Use of native and chemicallymodified cyclodextrin for the capillary electrophoretic separation of enantiomers, J. Liq. Chromatogr., 15, 1023 (1992).

40. Soini, H., Riekkola, M. L. and Novotny, M. V., Chiral separation of basicdrugs and quantitation of bupivacaine enantiomers in serum by capillary elec-trophoresis with modified cyclodextrin buffers, J. Chromatogr., 608, 265(1992).

41. Peterson, T. E., Separation of drug stereisomers by capillary electrophoresiswith cyclodextrins, J. Chromatogr., 630, 353 (1993).

42. Wren, S. A. C. and Rowe, Theoretical aspects of chiral separation in capillaryelectrophoresis. III. Application to β-blockers, J. Chromatogr., 635, 113 (1993).

43. Mayer, S. and Schurig, V., Enantiomer separation by electrochromatographyon capillaries coated with chirasil-dex, J. High Resol. Chromatogr., 15, 129(1992).

46


44. Penn, S. G., Goodall, D. M. and Loran, J. S., Differential binding of tiocona-zole enantiomers to hydroxypropyl-beta-cyclodextrin studied by capillaryelectrophoresis, J. Chromatogr., 636, 149 (1993).

45. Cruzado, I. D. and Vigh, G., Chiral separations by capillary electrophoresisusing cyclodextrin-containing gels, J. Chromatogr., 608, 421 (1992).

46. Ingelse, B. A., Everaerts, F. M., Desiderio, C. and Fanali, S., A study on theenantiomeric separation by capillary electrophoresis using a soluble neutral β-cyclodextrin polymer, J. Chromatogr. A, (1994).

47. Nardi, A., Eliseev, A., Bocek, P. and Fanali, S., Use of charged and neutralcyclodextrins in capillary zone electrophoresis: enantiomeric resolution ofsome 2-hydroxy acids, J. Chromatogr., 638, 247 (1993).

48. Dette, C., Ebel and, S. and Terabe, S., Neutral and anionic cyclodextrins in capillary electrophoresis: enantiomeric separation of ephedrine and relatedcompounds, Electrophoresis, 15, 799 (1994).

49. Chankvetadze, B., Endresz, G. and Blaschke, G., About some aspects of theuse of charged cyclodextrins for capillary electrophoresis enantioseparation,Electrophoresis, 15, 804 (1994).

50. Tait, R. J., Thompson, D. O., Stella, V. J. and Stobaugh, J. F., Sulfobutyl etherb-cyclodextrin as a chiral discriminator for use with capillary electrophoresis,Anal. Chem., 66, 4013 (1994).

51. Lurie, I. S., Klein, R. F. X., Dalcason, T. A., LeBelle, M. J., Brenneisen, R. and.Weinberger, R. E., Chiral resolution of cationic drugs of forensic interest bycapillary electrophoresis with mixtures of neutral and anionic cyclodextrins,Anal. Chem., 66, 4019 (1994).

52. Desiderio, C. and Fanali, S., Use of negatively charged sulfobutyl-ether-β-cyclodextrin for enantiomeric separation by capillary electrophoresis, J. Chromatogr. A, (submitted) (1995).

53. Terabe, S., Electrokinetic chromatography: an interface between electrophoresisand chromatography, Trends Anal. Chem., 8, 129 (1989).

54. Schmitt, T. and Engelhardt, H., Derivatized cyclodextrins for the separationof enantiomers in capillary electrophoresis, J. High Resol. Chromatogr., 16,525 (1993).

55. Aturki, Z. and Fanali, S., Use of beta-cyclodextrin polymer as a chiral selector in capillary electrophoresis, J. Chromatogr. A., 680, 137 (1994).

56. Kuhn, R., Erni, F., Bereuter, T. and Hausler, J., Chiral recognition and enantiomeric resolution based on host-guest complexation with crown ethersin capillary zone electrophoresis, Anal. Chem., 64, 2815 (1992).

57. Gozel, P., Gassmann, E., Michelsen, H. and Zare, R. N., Electrokinetic reso-lution of amino acid enantiomers with copper(II)- aspartame support electrolyte,Anal. Chem., 59, 44 (1987).

58. Gassmann, E., Kuo, J. E. and Zare, R. N., Electrokinetic separation of chiralcompounds, Science, 230, 813 (1985).

59. Desiderio, C., Aturki, Z. and Fanali, S., Separation of α-hydroxyacid enan-tiomers by high performance capillary electrophoresis using copper(II)-L-aminoacid and copper(II)-aspartame complexes as chiral selectors in the backgroundelectrolyte, Electrophoresis, 15, 864 (1994).

60. D’Hulst, A. and Verbeke, N., Chiral separations by capillary electrophoresiswith oligosaccharides, J. Chromatogr., 608, 275 (1992).

61. Nishi, H., Fukuyama, T., Matsuo, M. and Terabe, S., Chiral separation ofoptical isomeric drugs using micellar electrokinetic chromatography and bilesalts, J. Microcol. Sep., 1, 234 (1989).

47


62. Cole, R. O., Sepaniak, M. J. and Hinze, W. L., Optimization of binaphthyl enan-tiomer separations by capillary zone electrophoresis using mobile phasescontaining bile salts and organic solvents, J. High Resol. Chromatogr., 13,579 (1990).

63. Busch, S., Kraak, J.C. and Poppe, H., Chiral separations by complexationwith proteins in capillary zone electrophoresis, J. Chromatogr., 635, 119(1993).

64. Tanaka, Y., Matsubara, N. and Terabe, S., Separation of enantiomers by affini-ty electrokinetic chromatography using avidin, Electrophoresis, 15, 848 (1994).

65. Yang, J. and Hage, D.S., Chiral separations in capillary electrophoresis usinghuman serum albumin as a buffer additive, Anal. Chem., 66, 2719 (1994).

66. Armstrong, D.W., Rundlett, K. L. and Chen, J. R., Evaluation of the macro-cyclic antibiotic vancomycin as a chiral selector for capillary electrophoresis,Chirality, 6, 496 (1994).

67. Armstrong, D.W., Tang, Y. B., Chen, S. S., Zhou, Y. W., Bagwill, C. and . Chen,J. R., Macrocyclic antibiotics as a new class of chiral selectors for liquid chromatography, Anal. Chem., 66, 1473 (1994).

68. Fanali, S., Ossicini, L., Foret, F. and Bocek, P., Resolution of optical isomersby capillary zone electrophoresis. Study of enantiomeric and diastereoiso-meric Co(III)-complexes with ethylenediamine and amino acid ligands, J. Microcol. Sep., 1, 190 (1989).

69. Szejtli, J., Cyclodextrins and their inclusion complexes, Akademiai Kiado,Budapest, 1982.

70. Ward, T. J. and Armstrong, D. W., in M. Zief and L. J. Crane (Editors),Chromatographic chiral separation, Marcell Dekker, New York, p. 131, 1988.

71. Snopek, J., Jelinek, I. and. Smolkova-Keulemansova, E., Micellar, Inclusionand metal-complex enantioselective pseudophases in high performance electromigration methods, J. Chromatogr., 452, 571 (1988).

72. Pharr, D. Y., Fu, Z. F., Smith, T. K. and Hinze, W. L., Solubilization ofcyclodextrins for analytical applications, Anal. Chem., 61, 275 (1989).

73. Czugler, M., Eckle, E. and Stezowski, J., Crystal and molecular structure ofa 2,6-tetradeca-O-methyl-b-cyclodextrin-adamantanol 1:1 inclusion complex,J. Chem. Soc. Chem. Comm., 1291 (1981).

74. Croft, A. P. and Bartsch, R. A., Synthesis of chemically modified cyclodex-trins, Tetrahedron, 39, 1417 (1983).

75. Valko, I. E., Billiet, H. A. H., Frank, J. and Luyben, K. C. A. M., Effect of thedegree of substitution of (2-hydroxy)propyl-beta-cyclodextrin on the enan-tioseparation of organic acids by capillary electrophoresis, J. Chromatogr.A., 678, 139 (1994).

76. Yoshinaga, M., Asano, S., Tanaka, M. and Shono, T., Separation of racematesby capillary zone electrophoresis based on complexation with cyclodextrin-derivatives, Analytical Science, 7, 257 (1991).

77. Wren, S. A. C. and Rowe, R. C., Theoretical aspects of chiral separation in cap-illary electrophoresis. I.Initial evaluation of a model, J. Chromatogr., 603,235 (1992).

78. Terabe, S., Miyashita, Y., Ishihama, Y. and Shibata, O., Cyclodextrin-modifiedmicellar electrokinetic chromatography – separation of hydrophobic and enantiomeric compounds, J. Chromatogr., 636, 47 (1993).

79. Ueda, T., Mitchell, R., Kitamura, F., Metcalf, T., Kuwana, T. and Nakamoto,A., Separation of naphthalene-2,3-dicarboxaldehyde-labeled amino acids byhigh-performance capillary electrophoresis with laser-induced fluorescencedetection, J. Chromatogr., 593, 265 (1992).

48


80. Yamashoji, Y., Ariga, T., Asano, S. and Tanaka, M., Chiral Recognition andEnantiomeric Separation of Alanine b-naphthylamide by Cyclodextrins, Anal. Chim. Acta, 268, 39 (1992).

81. Stefansson, M. and Novotny, M., Electrophoretic resolution of monosac-charide e-Enantiomers in borate-oligosaccharide complexation media, J. Am. Chem. Soc., 115, 11573 (1993).

82. Schmitt, T. and Engelhardt, H., Charged and uncharged cyclodextrins as chiralselectors in capillary electrophoresis, Chromatographia, 37, 475 (1993).

83. Tait, R. J., Skanchy, D. J., Thompson, D. P., Chetwyn, N. C., Dunshee, D.A., Rajewsky, R. A., Stella, V. J. and Stobaugh, J. F., Characterization ofsulphoalkyl ether derivatives of b-cyclodextrin by capillary electrophoresis withindirect UV detection, J. Pharm. Biomed. Anal., 10, 615 (1992).

84. Wren, S. A. C., Theory of chiral separation in capillary electrophoresis, J. Chromatogr., 636, 57 (1993).

85. Rawjee, Y. Y., Williams, R. L. and Vigh, G., Capillary electrophoretic chiralseparations using cyclodextrin additives. III. Peak resolution surfaces foribuprofen and homatropine as a function of the pH and the concentration ofbeta-cyclodextrin, J. Chromatogr. A., 680, 599 (1994).

86. Hinze, W. L., Applications of cyclodextrins in chromatographic separationsand purification methods, Sep. Purif. Meth., 10, 159 (1981).

87. Schutzner, W. and Fanali, S., Enantiomers resolution in capillary zone elec-trophoresis by using cyclodextrins, Electrophoresis, 13, 687 (1992).

88. Li, S. and Lloyd, D. K., Packed-capillary electrochromatographic separationof the enantiomers of neutral and anionic compounds using beta-cyclodextrinas a chiral selector – effect of operating parameters and comparison with free-solution capillary electrophoresis, J. Chromatogr. A., 666, 321 (1994).

89. Snopek, J., Jelinek, I. and Smolkova-Keulemansova, E., VIII: Use of cyclodex-trins in isotachophoresis. IV. The influence of cyclodextrins on the chiralresolution of ephedrine alkaloid enantiomers, J. Chromatogr., 438, 211 (1988).

90. Jelinek, I., Dohnal, J., Snopek, J. and Smolkova-Keulemansova, E., Use ofcyclodextrins in isotachophoresis. VII. Resolution of structurally related andchiral phenothiazins, J. Chromatogr., 464, 139 (1989).

91. Jelinek, I., Snopek, J. and Smolkova-Keulemansova, E., Use of cyclodextrins inisotachophoresis. VIII. The separation of ketotifen and its polar intermedi-ate enantiomers, J. Chromatogr., 439, 386 (1988).

92. Snopek, J., Jelinek, I. and Smolkova-Keulemansova, E., Use of cyclodextrinin isotachophoresis. VIII. Two dimensional chiral separation in isota-chophoresis, J. Chromatogr., 472, 308 (1989).

93. Pedersen, C., Cyclic polyethers and their complexes with metal salts, J. Am. Chem. Soc., 89, 2495 (1967).

94. Kuhn, R., Wagner, J., Walbroehl, Y. and Bereuter, T., Potential and limita-tions of an optically active crown ether for chiral separation in capillary zoneelectrophoresis, Electrophoresis, 15, 828 (1994).

95. Hohne, E., Kraus, G. J. and Gubitz, G, Capillary zone electrophoresis of theenantiomers of aminoalcohols based on host-guest complexation with a chiral crown ether, J. High Resol. Chromatogr., 15, 698 (1992).

96. Davankov, V. A. and Rogozhin, S. V., Ligand chromatography as a novelmethod for the investigation of mixed complexes: stereoselective effects in α-amino acid copper(II) complexes, J. Chromatogr., 60, 280 (1971).

49


97. Terabe, S., Otsuka, K., Ichikawa, K., Tsuchiya, A. and Ando, T., Electrokineticseparations with micellar solution and open-tubular capillaries, Anal. Chem.,56, 111 (1984).

98. Dobashi, A., Ono, T., Hara, S. and Yamaguchi, J., Enantioselective hydrophobicentanglement of enantiomeric solutes with chiral functionalized micelles byelectrokinetic chromatography, J. Chromatogr., 480, 413 (1989).

99. Otsuka, K. and Terabe, S., Effects of methanol and urea on optical resolutionof phenylthiohydantoin DL-amino acids by micellar electrokinetic chromatography with sodium-N-dodecanoyl-L-valinate, Electrophoresis, 11,982 (1990).

100. Otsuka, K. and Terabe, S., Enantiomeric resolution by micellar electrokineticchromatography with chiral surfactants, J. Chromatogr., 515, 221 (1990).

101. Nishi, H., Fukuyama, T., Matsuo, M. and Terabe, S., Chiral separation oftrimetoquinol hydrochloride and related compounds by micellar electroki-netic chromatography using sodium taurodeoxycholate solutions and applicationto optical isomers, Anal. Chim. Acta, 236, 281 (1990).

102. Okafo, G. N., Bintz, C., Clarke, S. E. and Camilleri, P., Micellar electrokineticcapillary chromatography in a mixture of taurodeoxycholic acid andβ-cyclodextrin, J. Chem. Soc. Chem. Commun., 1189 (1992).

103. Nishi, H., Fukuyama, T., Matsuo, M. and Terabe, S., Chiral separation of diltiazem, trimetoquinol and related compounds by micellar electrokineticchromatography with bile salts, J. Chromatogr., 515, 233 (1990).

104. Ishihama, Y. and Terabe, S., Enantiomeric separation by micellar electrokineticchromatography using saponins, J. Liq. Chromatogr., 16, 933 (1993).

105. Wang, J. A. and Warner, I. M., Chiral separations using micellar electrokineticcapillary chromatography and a polymerized chiral micelle, Anal. Chem., 66,3773 (1994).

106. D’Hulst, A. and Verbeke, N., Quantitation in chiral capillary electrophoresis:theoretical and practical considerations, Electrophoresis, 15, 854 (1994).

107. Stalcup, A. M. and Agyei, N. M., Heparin: A chiral mobile-phase additive forcapillary zone electrophoresis, Anal. Chem., 66, 3054 (1994).

108. Carpenter, J. L., Camilleri, P., Dhanak, D. and Goodall, D. M., A study ofthe binding of vancomycin to dipeptides using capillary electrophoresis, J. Chem. Soc. Chem. Commun., 804 (1992).

109. Chu, Y. H. and Whitesides, G. M., Affinity capillary electrophoresis cansimultaneously measure binding constants of multiple peptides to vancomycin,J. Org. Chem., 57, 3524 (1992).

110. Armstrong, D. W., Rundlett, K. and Reid, G. L., Use of a macrocyclic antibi-otic, rifamycin b, and indirect detection for the resolution of racemic amino alcohols by CE, Anal. Chem., 66, 1690 (1994).

111. Ueda, T., Kitamura, F., Mitchell, R., Metcalf, T., Kuwana, T. and Nakamoto,A., Chiral separation of naphtalene-2,3-dicarboxaldehyde labeled amino acidenantiomers by cyclodextrins modified micellar electrokinetic chromatographywith laser induced fluorescence, Anal. Chem., 63, 2979 (1991).

112. Desiderio, C., Fanali, S., Kupfer, A. and Thormann, W., Analysis of mepheny-toin, 4-hydroxymephenytoin and 4-hydroxyphenytoin enantiomers in humanurine by cyclodextrin micellar electrokinetic capillary chromatography: Simpledetermination of a hydroxylation polymorphism in man, Electrophoresis, 15,87 (1994).

113. Altria, K. D., Essential peak area normalisation for quantitative impurity con-tent determination by capillary electrophoresis, Chromatographia, 35, 177(1993).

50


114. Rogan, M. M., Drake, C. D., Goodall, D. M. and Altria, K. D., Enantioselectiveenzymatic biotransformation of 2’-deoxy-3’-thiacytidine (BCH 189) monitoredby capillary electrophoresis, Anal. Biochem., 208, 343 (1993).

115. Altria, K. D., Walsh, A. R. and Smith, N. W., Validation of a capillary electrophoresis method for the enantiomeric purity testing of fluparoxan, J. Chromatogr., 645, 193 (1993).

116. Peterson, T. E. and Trowbridge, D., Quantitation of l-epinephrine and deter-mination of the d/l- epinephrine enantiomers in a pharmaceutical formulationby capillary electrophoresis, J. Chromatogr., 603, 298 (1992).

117. Mazzeo, J. R., Grover, E. R., Swartz, M. E. and Petersen, J. S., Novel chiralsurfactant for the separation of enantiomers by micellar electrokinetic capillary chromatography, J. Chromatogr. A., 680, 125 (1994).

118. Ong, C. P., Ng, C. L., Lee, H. K. and Li, S. F. Y., Determination of antihis-tamines in pharmaceutical by capillary electrophoresis, J. Chromatogr., 588,335 (1991).

119. Otsuka, K. and Terabe, T., Optical resolution of chlorpheniramine by cyclodex-trin added capillary zone electrophoresis and cyclodextrin modified micellarelectrokinetic chromatography, J. Liq. Chromatogr., 16, 945 (1993).

120. Prunonosa, J., Obach, R., Diez-Coscon, A. and Gouesclou, L., Determinationof cicletaline enantiomers in plasma by high performance capillary electrophoresis, J. Chromatogr., 574, 127 (1992).

121. Li, S. and Lloyd, D. K., Direct chiral separations by capillary electrophoresisusing capillaries packed with an alpha(1)-acid glycoprotein chiral stationaryphase, Anal. Chem., 65, 3684 (1993).

122. Nielen, M. W. F., (Enantio-)separation of phenoxy acid herbicides using capillary zone electrophoresis, J. Chromatogr., 637, 81 (1993) .

123. Mayer, S. and Schurig, V., Enantiomer separation using mobile and immobilecyclodextrin derivatives with electromigration, Electrophoresis, 15, 835(1994).

124. Guttman, A. and Cooke, N., Practical aspects in chiral separation of phar-maceuticals by capillary electrophoresis. II. Quantitative separation of naproxenenantiomers, J. Chromatogr. A, 685, 155 (1994).

125. Fanali, S. and Aturki, Z., Use of cyclodextrins in capillary electrophoresisfor the chiral resolution of some 2-arylpropionic acid non-steroidal anti-inflammatory drugs, J. Chromatogr. A, (in press) (1995)

126. Aumatell, A. and Wells, R. J., Chiral differentiation of the optical isomers ofracemethorphan and racemorphan in urine by capillary zone electrophoresis,J. Chromatogr. Sci., 31, 502 (1993).

127. Rogan, M. M., Altria, K. D. and Goodall, D. M., Enantiomeric separation ofsalbutamol and related impurities using capillary electrophoresis,Electrophoresis, 15, 808 (1994).

128. Nishi, H., Nakamura, K., Nakai, H., Sato T. and Terabe, S., Enantiomericseparation of drugs by affinity electrokinetic chromatography using dextransulphate, Electrophoresis, 15, 1335 (1994).

129. Otsuka, K., Kawahara, J., Tatekawa, K. and Terabe,S., Chiral separations bymicellar electrokinetic chromatography with sodium N-dodecanoyl-L-valinate, J. Chromatogr., 559, 209 (1991).

130. Gareil, P., Gramond, J. P. and Guyon, F., Separation and determination ofwarfarin enantiomers in human plasma samples by capillary zone electrophoresis using methylated b-cyclodextrin-containing electrolyte, J. Chromatogr., 615, 317 (1993).

51


AcknowledgmentsWe thank Drs. Zeineb Aturki, Emanuela Camera, Claudia

Desiderio, Wolfgang Schutzner, Mr. Gabreile Caponecchi, and Michele Cristalli for their cooperation in the research work.

52




U.S. (800) 4BIORAD California (510) 741-1000 New York (516) 756-2575 Australia 02-805-5000 Austria (1) 877 89 01 Belgium 09-385 55 11 Canada (905) 712-2771China (01) 2046622 Denmark 39 17 9947 Finland 90 804 2200 France (1) 49 60 68 34 Germany 089 31884-0 India 91-11-461-0103 Italy 02-21609 1 Japan 03-5811-6270 Hong Kong 7893300 The Netherlands 0318-540666 New Zealand 09-443 3099 Singapore (65) 4432529 Spain (91) 661 70 85 Sweden 46 (0) 735 83 00Switzerland 01-809 55 55 United Kingdom 0800 181134

Life Science Group

SIG 051995 Printed in USA Bulletin 1973 US/EG REVA 95-0284 0695


95-0284 Chiral Guide Cover 3/11/99 9:20 AM Page CVR1

iii

Table of Contents

About the Authors...................................................................................vAcknowledgment....................................................................................vFront Cover.............................................................................................vAcronyms Used.................................................................................... vii

1 Introduction......................................................................................1

2 Determination of Drug-Related Impurities.......................................22.1 Overview................................................................................22.2 Benefits of CE Methods for Related Impurity

Determinations.......................................................................32.3 Applications.........................................................................10

3 Main Component Assay.................................................................153.1 Overview..............................................................................153.2 Benefits of CE Methods for Drug Assay...............................163.3 Applications.........................................................................19

4 Chiral Separations..........................................................................244.1 Overview..............................................................................244.2 Benefits of CE Methods for Chiral Analysis.........................244.3 Applications.........................................................................29

5 Stoichiometric Determinations.......................................................335.1 Overview..............................................................................335.2 Benefits of CE Methods for Stoichiometric Determinations ..345.3 Applications.........................................................................34

6 Quantitative Procedures.................................................................396.1 Overview..............................................................................396.2 Impurity Determinations.......................................................406.3 Main Component Assay........................................................406.4 Chiral Analysis.....................................................................406.5 Stoichiometric Analysis........................................................41

iv

7 Method Validation Considerations..................................................417.1 Overview..............................................................................417.2 Accuracy..............................................................................417.3 Precision...............................................................................427.4 Reproducibility.....................................................................427.5 Linearity...............................................................................427.6 Robustness...........................................................................43

8 Future Developments.....................................................................458.1 Overview..............................................................................458.2 Detection..............................................................................478.3 Electrochromatography.........................................................48

9 Conclusions....................................................................................49

Practical Hints.......................................................................................50Electrolyte/Sample.........................................................................50Sensitivity......................................................................................50Quantitation...................................................................................51Capillary Care................................................................................51

References.............................................................................................52

v

About the Authors

Kevin Altria commenced his Ph.D. studies at the University of London in 1986studying various aspects of capillary electrophoresis. Highlights of his researchinclude fundamental studies concerning electroosmotic flow, the first report ofpharmaceutical analysis by CE, and the development of a novel radioactivitydetector. Kevin has since joined Glaxo Research and Development and iscurrently concentrating on developing applications of CE for pharmaceuticalanalysis. He has authored or co-authored more than 40 CE-related publicationsand has edited a book on CE methodology and applications which is currentlyin press with Humana Press.

Manus Rogan graduated from Dublin City University with a B.Sc. inAnalytical Science in 1989. Since then, he has been working as an AnalyticalChemist at Glaxo Research and Development. Manus is near completion ofpart-time Ph.D. studies concerning use of CE for pharmaceutical analysis atthe University of York. He has been a co-author on over 10 papers and chap-ters on CE and has a particular interest in the theoretical aspects of chiral CE.

The authors have previously written a primer on the theory and applica-tions of chiral capillary electrophoresis.

Acknowledgment

Thanks are extended to Mrs. Larraine Horwood of Glaxo Research and Devel-opment for careful typing and continued patience.

Front Cover

Separation of 11 basic drugs by CE using a low-pH electrolyte with UV detec-tion at 200 nm (Altria, 1993f).

vi

Other Beckman primers (Volumes I, II, III, IV, and V) oncapillary electrophoresis:






Separation of Proteins and Peptides byCallipary Electrophoresis: Applicationto Analytical Biothechnology 727484

All trademarks and registered trademarks are the property of their respectiveowners.

vii

Acronyms Used

The following acronyms are used in this book.

CE capillary electrophoresisCIA capillary ion analysisDAD diode array detectorEDTA ethylene diamine tetra acetic acidEOF electroosmotic flowFSCE free-solution capillary electrophoresisGC gas chromatographyHPLC high-performance liquid chromatographyICP inductively coupled plasmaIEC ion-exchange chromatographyLIF laser-induced fluorescenceLOD limit of detectionLOQ limit of quantitationMECC micellar electrokinetic capillary chromatographyMS mass spectroscopyppb parts per billionppm parts per million (mg/L)RSD relative standard deviationr2 correlation coefficientSDS sodium dodecyl sulphateTLC thin-layer chromatography

1

1 IntroductionCapillary electrophoresis was popularized in the early 1980’s by Jorgensonand co-workers (Jorgenson and Lukacs, 1983) who demonstrated that excep-tional efficiencies could be obtained by performing electrophoresis in capillar-ies. Their early work on the electrophoretic separation of biomolecules spurreddevelopment of commercial instrumentation and the investigation of applica-tion areas where use of conventional electrophoresis was infrequent. One sucharea was that of pharmaceutical analysis where few applications of electro-phoresis were in routine use due to the cumbersome and semi-quantitativenature of this technique. However, the introduction of the capillary formatenables full quantitative and automated analysis to be conducted.

To ensure the safety and efficacy of the final marketed product, it is im-portant to characterize drug substance material and formulations. Full charac-terization involves the assessment of the drug material by both physical andchemical methods. The measurement of chemical properties such as purity,assay, chiral purity, inorganic ion content, and identity confirmation is rou-tinely performed by HPLC and other chromatographic techniques. Recently,capillary electrophoresis (CE) has been developed to perform these tests andshown to be a complementary and attractive alternative to the more establishedmethods.

The majority of drugs are either acidic and/or basic water-soluble com-pounds. The basis for separation in free solution CE (FSCE) relies upon anexploitation of differences between the analytes’ electrophoretic mobilities,which are related to the solutes’ charge and size. Consequently, the separationof many drugs is possible by FSCE. For example, an acidic drug may be ana-lyzed in its anionic form at high pH and basic drugs may be tested at low pH intheir cationic form. The front cover depicts the efficient separation of 11 basicdrugs using a low-pH electrolyte. Zwitterionic drugs (those containing bothacidic and basic groups) may be analyzed at either end of the pH range. Amixture of neutral drugs would be unresolved by FSCE. However, ionic,charged micelles can be incorporated into the electrolyte solution to add apartitioning element to the separation. This is the basic idea behind MECC(Terabe et al., 1984).

An important aspect of drug analysis involves the determination of drug-related impurities. This is generally performed by HPLC which has an estab-lished methodology and highly automated instrumentation available. It isimportant to ensure that a thorough and accurate impurity profile is generated.

2

This can be supplemented by cross-correlation of results obtained with thosefrom TLC or an alternative HPLC method. In all cases, the selectivity reliesupon a chromatographic interaction where co-elution or irreversible adsorptionmay occur. CE offers a completely different selectivity process and is, there-fore, truly a complementary and orthogonal technique to HPLC.

Given that an analytical method may be developed for long-term use andthus may be applied to several thousand samples, speed, simplicity, and auto-mation are key requirements. Both CE and HPLC can offer these facilities formany applications. Since the first application of CE to drug analysis in 1987(Altria and Simpson, 1987; Fujiwara and Honda, 1987), there have been over150 papers on this subject in the literature. This primer will cover aspects ofthe validation, application, and performance of CE methods in the analysis ofpharmaceuticals as reported in the literature.

A basic introduction to the theory and various CE separation modes of CEis covered in “Introduction to Capillary Electrophoresis.” Further primers givedetailed discussion on MECC and chiral separations. Review articles of phar-maceutical analysis (Altria 1993a; Rabel and Stobaugh, 1993) and chiral sepa-rations (Kuhn and Hoffstetter-Kuhn, 1992; Rogan et al., 1994) have recentlybeen published.

2 Determination of Drug-RelatedImpurities

2.1 OverviewThe determination of drug-related impurities is currently the principal role ofCE within pharmaceutical analysis and presents a challenge to both selectivityand sensitivity. The main component and structurally related impurities havesimilar chemical properties and thus make resolution difficult. However, anadvantage of CE over its chromatographic counterparts is that high separationefficiencies are achievable. The resulting peak sharpness often translates asmall degree of selectivity to acceptable resolution. A detection limit of 0.1%area/area is widely accepted as a minimum requirement for a related impuritiesdetermination method. This 0.1% level is possible by CE. For example, Figure1 (Swartz, 1991) shows determination of salicylamide-related impurities at0.1% area/area and below. HPLC can routinely decrease this level by up to anorder of magnitude (≈ 0.01%). Owing to the possibility that a formulation or

3

drug substance may contain a great number of low-level impurities, a methodcapable of resolving the required peaks within a defined analysis time isneeded. The high separation efficiencies offered by CE means that this is nowa real possibility.

5.85.65.45.25.04.8

Time (min)

4.64.44.24.04.4

4.8

5.2

5.4

mV

5.8

Figure 1. Salicylamide CE impurity profile. Electrolyte: 0.02 M sodium phos-phate, pH 11.0, 0.075 M SDS; capillary: 60 × 50 µm; voltage 20 kV; sampleconcentration: 0.1 mg/mL in water; injection: 10 s. Reproduced with permis-sion from Swartz, 1991.

When calculating impurity levels, it is necessary to divide the observedarea of each peak by its migration time (Altria, 1993b). This normalization isnecessary since faster-migrating peaks move through the detector at a greaterspeed than their slower counterparts. Therefore, faster-moving peaks havesmaller peak widths and correspondingly smaller peak areas. The sum of these“normalized peak areas” is used to calculate impurities as % area/area. Whenimpurity determinations are to be expressed as % w/w through the use of exter-nal standards, this normalization process is not required, providing the preci-sion of migration time is acceptable.

2.2 Benefits of CE Methods for Related ImpurityDeterminations

When adopting a CE method for determination of drug related impurities,possible features include complementary data to HPLC, low wavelength detec-

4

tion, speed, and simplicity. Other facilities include the ability to carry out on-line spiking experiments and the use of UV diode array detectors.

2.2.1 Complementary Data to HPLC

The data generated on impurity profiling is often compared to that obtained bychromatographic methods (typically HPLC). The principles of separation inCE are entirely different to HPLC and, therefore, a good agreement betweenthe two techniques strongly supports the integrity of the data. This techniquecombination is now established in many laboratories and has become a suitablereplacement for the conventional use of TLC and HPLC in combination. Apartfrom routine investigations, this combined use is of particular importanceduring method validation.

The differences in selectivity between CE and HPLC can result in discrep-ancies in results with one technique showing an underestimation in impuritylevels. This occurrence signifies that further method optimization is required.Critical events in the development of a formulation such as synthetic route orprocess changes or at key drug stability timepoints are times when this mayoccur. Literature examples include additional tetracycline (Zhang et al., 1992)and domperidone impurities (Pluym et al., 1992) resolved by CE. Figure 2shows an impurity profile for domperidone in which peaks 1 and 2 co-elutedusing both HPLC and TLC.

5

O O

N-CH2-CH2-CH2-NH-N N N-H

Cl

domperidone R 33812

b

c

a

4 53

21

0 17Time (min)

Figure 2. CE separation of Domperidone (R33812) and major known impuri-ties. (a) Batch G1A041; (b) 0.1% reference mixture; (c) 1% reference mixture.Peak numbers: 1 = R29676, 2 = R45771, 3 = domperidone R33812,4 = R48557, 5 = R52211. Separation conditions: citrate-phosphate, pH 4,+25kV, 30°C, 250 nm, 50 µm × 72 cm. Reproduced with permission fromPluym et al., 1992.

2.2.2 Low Wavelength Detection

Aqueous-based electrolytes (which have low UV absorbance coefficients) areoften employed in CE, allowing detection wavelengths such as 200 nm to beroutinely employed. Many impurities or small intermediates have poor chro-mophores making their quantitation at traditional HPLC wavelengths difficultor impossible. This may be of particular importance for degradative processeswhere reactions may lead to the loss of the functionality providing the chro-mophore.

Alternatively, the use of low UV wavelengths may compensate for theinherent poor sensitivity in CE. For example (Altria, 1993c), when operating at

6

200 nm there is a ten-fold increase in signal for salbutamol and its impuritiescompared with 276 nm which is the HPLC wavelength.

2.2.3 Speed

Use of short capillaries coupled with high voltages can allow extremely shortanalysis times to be attained. For example, Figure 3a (Altria, 1993d) shows theseparation of fluparoxan impurities within two minutes using a 27-cm capil-lary. Figure 3b shows the same separation using a 57-cm capillary. Althoughsome degree of resolution is sacrificed, a good indication of the purity of thetest substance is rapidly obtained.

12.00

11.00

10.00

9.00

8.00

7.00

6.00

5.00 6.00 7.00 8.00 9.00 10.00

18.00

16.00

14.00

12.00

10.00

8.00

1.40 1.60 1.80 2.00

a


UV

Abs

orba

nce

UV

Abs

orba

nce


b

(V)

(V)

Figure 3. High-speed separation of fluparoxan and related impurities.Conditions: 50 mM borax, pH 2.2, with conc H3PO4, sample concentration0.5 mg/mL in water, 214 nm, 10 s pressure injection, 10 kV.Peak V = fluparoxan. Figure 3a: 27-cm capillary. Figure 3b: 57-cm capillary.Reproduced with permission from Altria, 1993d.

7

2.2.4 Simplicity

The majority of methods employed involve use of aqueous-based electrolytesand uncoated fused-silica capillaries. This ensures that the methods are simpleto operate and relatively easy to transfer between laboratories.

2.2.5 On-Line Spiking

For a complex impurity separation, relative migration times are not alwayssufficient for making a positive identification of a peak. Confirmation of peakidentity can be achieved (Altria and Luscombe, 1993) by on-line spiking witha solution of the impurity. This can be programmed into the separationmethod. Figure 4a shows the separation of a salbutamol sample solution whichalso contains two dimeric impurities. Figure 4b shows the separation obtainedfrom a 5-second injection of the salbutamol sample solution followed immedi-ately by a 5-second injection of a solution of the impurity of interest. Follow-ing both injections, the voltage was applied and the separation given in Figure4b was produced. The impurity identity is clearly confirmed and no loss inresolution is observed from this dual-injection procedure.

8

10.00

9.00

8.00

7.00

6.00

5.00

10.00

9.00

8.00

7.00

6.00

5.00

7.00 8.00 9.00

7.00 8.00 9.00

Retention Time (min)

I

II

III

I

II

III

a

b

Figure 4. Determination of salbutamol impurities. (a) CE separation of a1 mg/mL salbutamol solution; (b) separation of a 1 mg/mL salbutamol solutionspiked with 0.5% w/w/ bis-ether by co-injection. Electrolyte: 20 mM Na cit-rate, pH 2.5; voltage: 30 kV; detector: 200 nm; capillary: 75 cm × 57 µm.Reproduced with permission from Altria and Luscombe, 1993.

9

2.2.6 UV Diode Array Detection

The recent advent of UV diode array detector (DAD) technology greatly as-sists in method development and peak assignment confirmation. Figure 5shows the use of this detector to measure the spectrum of a 0.1% impurity. Inconjunction with the on-line spiking procedure discussed previously, this givesadded confidence in peak assignments.

Figure 5. Spectra of an 0.1% area/area impurity. Reproduced with permissionfrom Altria, 1994 (unpublished results).

The principal disadvantages of the use of CE for determining relatedimpurities are the possible requirements for higher sample concentrations.When operating HPLC and CE at the same wavelength, it may be necessary touse two to five times more concentrated samples for CE to obtain an equiva-lent limit of detection. This may represent problems for poorly soluble drugs.

10

Due to the small sample volumes employed in CE, the possibility of con-ducting micropreparative scale separations are very limited. However a num-ber of reports have been published in the literature (Camilleri et al., 1991;Altria and Dave, 1993).

2.3 ApplicationsA number of applications have been reported, many at the 0.1% area/areadetection level. Figure 6 shows the impurity profile of ranitidine, using a two-second and ten-second injection. In this example (Altria, 1993e), a good corre-lation between CE and HPLC was obtained for the total number and level ofimpurities. Smaller impurities were quantified at the 0.1% level. Use of ahigher sample loading can extend this detection limit even lower. If a shortinjection time (i.e., one second) can be employed to produce a separation withthe main peak on-scale, a longer injection time (i.e., 10 seconds) will produce aseparation with the main peak off-scale but with considerably enhanced sensi-tivity for the minor components. The peak area of the off-scale peak is thencalculated by multiplying the peak area of the on-scale injection by the ratio ofthe injection times (i.e., 10:1). Impurities are quantified as % area/area of thecalculated area for the off-scale peak. This is demonstrated (Altria 1993e) inpractice for fluparoxan impurities where the LOD for the off-scale separationis 0.01%.

11

a

9.00

8.50

8.00

7.50

7.00

5.00 6.00 7.00 8.00 9.00 10.00

Migration time (min)

UV

Abs

orba

nce

(arb

. uni

ts)

9.00

8.50

8.00

7.50

7.00UV

Abs

orba

nce

(arb

. uni

ts)

5.00 6.00 7.00 8.00 9.00 10.00


b

I

I

Figure 6. (a) 2-s loading of degraded ranitidine solution; (b) 10-s loading.I = ranitidine. Electrolyte: 50 mM borax, pH adjusted to 2.5 with H3PO4 and 2mM hydroxypropyl-β-CD. Reproduced with permission from Altria, 1993e.

The separation of a basic drug and its impurities at low pH can beachieved by exploiting fundamental differences in the charge and size of thecomponents. If this selectivity is insufficient, incorporation of an additive such

12

as cyclodextrin (Ng et al., 1992; Altria, 1993d) or an ion-pair reagent (Nishiet al., 1990a) into the electrolyte may achieve the desired effect. Componentswill selectively interact with these additives, resulting in overall changes inseparation selectivity. Figure 7 shows the separation of ranitidine and relatedimpurities employing a low-pH electrolyte containing cyclodextrin. A 27-cmcapillary was utilized and, therefore, a rapid analysis was achieved.

30.00

25.00

20.00

15.00

10.00

0.60 0.70 0.80 0.90 1.00 1.10 1.20

UV

Abs

orba

nce


(IV)

Figure 7. HSCE separation of degraded ranitidine syrup sample. Separationconditions: 25 mM borax with 2 mM dimethyl-β-cyclodextrin, pH 2.4, withconc. H3PO4; sample concentration, 15 mg/mL; 230 nm; 2-s pressure injec-tion; 15 kV. Peak IV = ranitidine. Reproduced with permission from Altria,1993d.

The complementary nature of HPLC and CE was highlighted in a study(Altria, 1993c) of the determination of dimeric salbutamol impurities. Theserelatively large impurities were strongly adsorbed onto the HPLC column andtherefore had lengthy retention times. However, in CE the dimeric impuritieshad a charge Z = +2 and therefore migrated before the salbutamol (Z = +1).Drug substance batches were tested by CE and HPLC using external standardsof the impurities for quantitation. Table 1 shows the good correlation betweenthe two techniques.

13

Table 1. Levels of Salbutamol Impurities inDrug Substance Determined by CE and HPLC

Good cross-correlation was obtained between the two techniques.

Bis Ether Dimer Batch (%w/w) (%w/w)

CE HPLC CE HPLC

1 0.14 0.16 0.08 0.080.14 0.16 0.08 0.08

2 0.10 0.11 0.06 0.070.10 0.11 0.07 0.06

3 0.20 0.19 0.13 0.110.20 0.19 0.14 0.10

4 0.12 0.13 0.07 0.060.15 0.14 0.08 0.05

5 0.13 0.14 0.08 0.060.12 0.13 0.07 0.05

6 0.31 0.28 0.18 0.170.31 0.26 0.19 0.15

7 0.07 0.09 0.05 0.040.08 0.10 0.06 0.03

8 0.38 0.38 0.20 0.180.44 0.38 0.22 0.19

9 0.37 0.38 0.19 0.180.37 0.35 0.19 0.19

10 0.75 0.66 0.39 0.330.77 0.67 0.40 0.35

(Altria, 1993c)

A worker from the FDA demonstrated (Flurer and Wolnik, 1994) thathigher levels of gentamycin impurities were detected using CE compared withthose achieved by the USP-registered HPLC method. Gentamycin has a poorchromophore and, therefore, needs to be derivatized prior to HPLC analysis.However, not all impurities appear to be derivatized to an equal extent. CEallowed the direct analysis of gentamycin and impurities with low UV wave-length detection.

14

The time-dependent degradation of an impurity of ranitidine was moni-tored by CE (Altria and Connolly 1993). Sample solution was reinjected sev-eral times over the course of 9.25 hours. After this storage, less than 2% of theoriginal substance remained. The initial sample solution was analyzed by bothCE and HPLC (8.2 and 8.7% area/area impurity content, respectively).

A stability-indicating method for enalapril has been reported (Qin et al.,1992) which gave detection limits of 0.2% for the monitored impurities.

Impurities have been determined in a water-insoluble quinolone antibiotic(Altria and Chanter, 1993). The compound was only soluble at pH extremes ofless than 2 and greater than 10. The sample was dissolved in NaOH solutionand analyzed with a pH-1.5 electrolyte. A detection limit of 0.1% was demon-strated during validation. Linearity was measured in two exercises (1 to 150%and 20 to 150% of target concentration); correlation coefficients of 0.9990 and0.9997 were obtained, respectively. A single sample was injected 10 times andprecision values of 0.4 and 0.6% RSD were obtained for migration time andpeak area, respectively.

Resolution of acidic or neutral drugs and impurities requires operation athigh pH and possibly the addition of a surfactant (SDS). Salicylamide impuri-ties at levels below 0.1% area/area have been shown (Swartz, 1991) usingSDS-based separation. Repeated analysis produced an RSD of 9.3% of thearea of a peak relating to an impurity spiked at the 0.1% level.

Nishi and Terabe (1990) showed use of MECC to determine impuritylevels in dilitazem drug substance. They reported detection limits of 0.1%.

Separation and quantitation of tetracycline impurities were achieved(Zhang et al., 1992) by on-capillary derivatization with EDTA. The anioniccomplexes were resolved and directly quantified. No performance data wasshown.

Levels of domperidone impurities were quantified by TLC, HPLC, andCE in drug substances (Pluym et al., 1992). Table 2 shows the results fromHPLC and CE for selected impurities determined in three drug substancebatches. Good agreement between the three techniques was obtained for totalimpurity levels. However, CE resolved two additional components which co-eluted in HPLC and TLC determinations.

15

Table 2. Levels of Domperidone Impurities in Drug Substance Determinedby CE and HPLC

CE resolved an additional unknown impurity.

Batch R45571 R48557 Unknown No. Content Content Impurity

CE HPLC CE HPLC CE HPLC

1 0.24 0.26 0.15 0.35 0.17 -2 0.22 0.23 0.15 0.34 0.24 -3 0.26 0.27 0.15 0.30 0.18 -

(Pluym et al., 1992)

3 Main Component Assay

3.1 OverviewThis is an important application area in drug analysis. Several reports haveappeared concerning the successful correlation of CE and HPLC assay resultsindicating that CE produces comparable results. The major drawback to CEbeing more widely adopted in this area is the perceived lack of precision sug-gested in early papers (5 to 10% RSD). However, it should be recognized thatthis data was typically generated on homemade apparatus or using instrumen-tation which has subsequently been developed considerably further. ExistingHPLC instrumentation can routinely achieve <1% RSD. The newly introducedcommercial CE equipment can approach this level of precision (Watzig andDette, 1993).

The major source of imprecision remaining when using commercial in-strumentation is injection volume variability. Reproducible nanoliter samplevolumes presents a formidable engineering challenge. Use of an internal stan-dard will reduce this effect (Dose and Guiochon, 1991).

Injection precision is also related (Ryder, 1992) to sample concentration,being worse at low sample concentrations. This is due to increased variancecontributions from factors such as integration errors and solute adsorption onto

16

the capillary surface. It is suggested, therefore, that a high sample concentra-tion and injection volume should be employed. This has the effect of reducingseparation efficiencies but improving precision. This use of high concentra-tions also allows the simultaneous determination of related impurities.

The volume of sample solution injected onto the capillary is related to thesample solution viscosity (Watzig and Dette, 1993). This does not represent aproblem when analyzing drug substances since equivalent sample weights aredissolved in an identical matrix. Tablet excipients such as cellulose, starch, orcyclodextrins can considerably increase solution viscosity and, therefore, lowerinjection volumes and potency for tablet assay solutions. This would be high-lighted by recovery experiments from tablet excipient mixtures. Employmentof an internal standard will alleviate this problem.

3.2 Benefits of CE for Drug AssayThe features of adopting a CE method for main peak assay include generationof data complementary to HPLC, possible reductions in sample preparation,reduced operating costs, and simplicity.

3.2.1 Complementary Data to HPLC

As with related impurities determinations, CE can be used to confirm assaydata generated by HPLC. Assay of identical samples by both CE and HPLCcan play an important part of method validation or result confirmation foreither technique. Several reports (Ackermans et al., 1992a; Altria and Filbey,1993; Pluym et al., 1992; Tsai et al., 1992) have shown equivalence betweenassay results obtained by HPLC and CE for a range of formulations. Theseresults are discussed later in section 3.3, “Applications.”

3.2.2 Sample Pretreatment Reductions

Many formulations contain components which are strongly retained and mayunduly affect the chromatographic performance of HPLC columns. Thereforeit is often necessary to pretreat sample solutions prior to HPLC analysis. Typi-cal procedures include solid phase extraction, filtration, and centrifugation.Use of guard columns can reduce this necessity.

However, in the CE analysis of a formulation containing a basic drug atlow pH, the majority of excipients, being neutral, will not migrate. These ex-cipients will remain at the injection end of the capillary and will be removed

17

during a rinse step. For example, direct injection of diluted syrup samples waspossible, minimizing sample pretreatment (Altria and Rogan, 1990).

Components with poor chromophores are generally not suitable for directanalysis by HPLC without sample derivatization. However, the use of low UVwavelengths or indirect detection in CE can overcome this difficulty. Seepage 13 for an example involving the antibiotic gentamycin. Severalaminoglycoside antibiotics including neomycin, streptomycin, and sisomycinwere separated (Ackermans et al., 1992b) by CE (Figure 8).

0.001

-0.000

-0.001

-0.002

-0.003

-0.0040 5 10 15

EOF

12

34

5

6

Time (min)

Abs

orba

nce

(AU

)

Figure 8. Electropherogram for the separation of (1) dihydrostreptomycin, (2)lividomycin, (3) amigacin, (4) kanamycin, (5) tobramycin, and (6) sisomycin(all 1.0 mg/mL) in the anionic mode with the reversed EOF applying a back-ground electrolyte of 0.1 M imidazole acetate at pH 5.0 containing the additiveFC 135 (50 µl/mL). Capillary length, 67 cm; applied voltage, 12.5 kV; pres-sure injection time, 2 s; UV detection wavelength, 214 nm .Reproduced withpermission from Ackermans et al., 1992b.

3.2.3 Operating Costs

In CE, consumable expense is relatively low. Typical methods employ aque-ous electrolyte solutions with 10 to 20 mL being a common daily requirement.Costs are minimal when compared to HPLC solvent purchase and disposal.Uncoated capillaries are generally employed which are a fraction of the cost ofa HPLC column.

18

3.2.4 Simplicity

Often a single set of operating parameters can be applied to a wide variety ofdrugs. Figure 9 shows the separation (Chee and Wan, 1993) of 17 basic drugsat a low pH. An additional 11 basic drugs are resolved (Altria, 1993f) usingsimilar conditions (see front cover for separation).

4 5 6 7 8 9 10 11

4

6

8

10

12

14

Time (min)

1

2

3

4

5

6

78

9

10

11

12

13

14

15

i

16

17

Figure 9. Electropherograms of a mixture of 17 basic drugs.(1) methapyrilene, (2) brompheniramine, (3) amphetamine, (4) methamphet-amine, (5) procaine, (6) tetrahydrozoline, (7) phenmetrazine, (8) butacaine, (9)medazepam, (10) lidocaine, (11) codeine, (12) acepromazine, (13) meclizine,(14) diazepam, (15) doxapram, (16) benzocaine, (17) methaqualone. Separa-tion conditions: 0.05 M sodium dihydrogenphosphate-phosphoric acid, pH2.35; capillary: 60 cm × 75 µm; injection: 10 s; hydrostatic loading; 22 kV;detection: 214 nm. Peak i is an artifact of benzocaine (peak 16). Reproducedwith permission from Chee and Wan, 1993.

Use of an MECC method to quantify seven ingredients of theophyllinetablets (and two possible internal standards) has been reported (Dang et al.,1993). Agreement between label claim and MECC results was demonstratedfor several components simultaneously.

19

A pH-7.0 electrolyte was shown to resolve 16 common sulphonamides(Ackermans et al., 1992c). Figure 10 shows the separation with an analysistime of 32 minutes. The method was applied to the testing of sulphonamidelevels in pork meat.

Abs

orba

nce

(arb

. uni

ts)

Time (min)

0 8 16 24 32

14

14

5 6

13

7 10

12 1116

2

15

Figure 10. Electropherogram of the separation of 16 sulphonamides(0.1 mg/mL) in the Beckman standard capillary and detection 50 cm, pressureinjection time 2 s (39 nL), applied voltage 10 kV of standard mixture electro-lyte 0.02 M imidazole acetate at pH 7. Reproduced with permission from Ack-ermans et al., 1992c.

3.3 ApplicationsMany reports have appeared and have been reviewed (Rabel and Stobaugh,1993; Altria, 1993g) on the use of CE for quantitative analysis of drug formu-lations. CE has been applied to the testing of a wide variety of formulationsincluding tablets, solutions for injection, infusion solutions, syrups, eardrops,creams, and rotacaps. The majority of reports have demonstrated equivalencebetween CE, HPLC, and/or label claim for the formulation.

Some noteworthy examples include the determination of threebronchodilators (fenoterol, salbutamol and terbutaline) in six different dosageforms (Ackermans, et al., 1992a). Comparisons were made between the CEdata, HPLC, isotachophoresis (an alternative electrokinetic separation tech-

20

nique), and label claim and all results obtained were deemed to be satisfactory.Correlation coefficients of better than 0.999 were reported between HPLC andCE results. Acceptable performance of the CE methods was also reported forlinearity (R = < 0.999) and precision for migration times and peak areas rangedbetween 0.3 to 2.2% RSD.

MECC was used (Nishi and Terabe, 1990) to test paracetamol, caffeineand ethenzamide content in tablets from two manufacturers. The results ob-tained were between 99 to 102% of the label content with precision of 1.0 to2.3% RSD for peak area and 0.5 to 0.9% RSD for migration time. An internalstandard was employed to achieve this precision.

In general, CE is not the first technique of choice when faced with theneed to separate non-water-soluble compounds due to problems with on-capil-lary precipitation. In some cases, MECC conditions can be appropriate. Forexample, the separation of benzothiazepines and corticosteroids has been re-ported (Nishi et al., 1990b). Dilitazem content in tablets by MECC was foundto be within 99 to 102% of label claim with area precision of 2.2% RSD. Inthis work, methanolic solutions of the samples (e.g., as a cream formulation)and standards were employed and injected directly on capillary.

Ackermans et al. (1992b) quantified various aminogylcoside antibioticsusing indirect UV detection. They reported correlation coefficients of < 0.9995in the range 0.1 to 1.0 mg/mL.

Alendronate levels in formulations were determined by CE. Separationand detection was achieved by virtue of on-capillary complexation withCa2+ ions in the running buffer (Tsai et al., 1992). HPLC, involving a chemicalderivatization, was also employed. An average recovery of 100.7% was ob-tained by CE in the range of 80 to 120% of target concentration. Equivalentassay results were obtained by both techniques. Ten tablets were analyzed byboth HPLC and CE to produce average assay results of 2.47 and 2.45 mg/tablet, respectively, with precision data of 0.8 and 0.7% RSD, respectively.However, the sample analysis took 8 minutes using CE, whereas HPLC re-quired 4 hours.

The content uniformity of enalapril-containing tablets has been assessedby a stability-indicating CE method (Qin et al., 1992). Detection limits of< 0.2% were reported with good precision (0.6% RSD). Detector linearity overthe range 50 to 150% of target concentration was shown (R2 = 0.999).

21

Ackermans et al. (1991a) determined levels of water-insoluble dapsonein tablets using methanol as the sample-dissolving solvent and fenbendazole asan internal standard. Average assays of 105.2 mg/tablet were obtained for a100 mg/tablet (no indication was available of possible tablet content overagevalues). Correlation coefficients of 0.999 were obtained for detector linearityin the range 10 to 100% of target concentration.

Various analgesics have been determined by MECC (Fujiwara andHonda, 1987). Recoveries of 98.7 to 101.0% were reported from excipients.Precision data for peak area ratios was shown to be 0.8 to 1.8% RSD. Assay oftablets gave precision data for drug content ranging from 0.9 to 1.4% RSD.The results obtained were in good agreement (99.3 to 100.6%) with the labelclaim for tablet content.

Levels of the anti-migraine agent sumatriptan in injection solutions weredetermined by CE and HPLC (Altria and Filbey, 1993). An internal standardwas employed to provide improved precision. Results generated by the twotechniques were in good agreement (Table 3). Detector linearity in the range of5 to 150% of target concentration was 0.9993. Peak area ratio data was 0.1 to0.8% and 0.5 to 0.7% RSD for sample and standard solutions, respectively.Inter-day repeatability gave assay data repeatability within 1%. A range ofsynthetic and degradative impurities was simultaneously separated and detec-tion limits of <0.1% were possible.

Table 3. Sumatriptan Contents in Injection Solutions asDetermined by CE and HPLC

Good agreement was obtained between the two techniques.

SampleSumatriptan content

(mg/mL)

CE HPLC

Batch 2 Sample 1 11.5, 11.6 11.6, 11.6Batch 2 Sample 2 11.6, 11.6 11.7, 11.7

Batch 3 Sample 1 11.7, 11.8 11.8, 11.8

Batch 3 Sample 2 11.6, 11.6 11.7, 11.7Batch 4 Sample 1 11.7, 11.8 11.8, 11.8

Batch 4 Sample 2 11.7, 11.6 11.7, 11.7

(Altria and Filbey 1993)

22

An inter-company collaboration program has been established to investi-gate the transfer and application of CE methods. Seven independent pharma-ceutical companies are involved in this study. One exercise involved thequantitative analysis of the paracetamol content in a commercial capsule for-mulation containing both paracetamol and caffeine. Using an internal standardand an SDS-based MECC method, all seven participating companies were able(Altria, 1994a) to repeat the separation. Figure 11 gives three representativeelectropherograms showing the repeatability of the separation. All companiesreported CE assay values equivalent to the HPLC data and the label claim(298.3, 302.5, and 300 mg/capsule, respectively). Relative migration precisiondata ranged between 0.3 to 0.8% RSD. The average precision data for thecalibration response factor was 1.5% RSD, while the assay results gave anRSD of 1.0% for 56 individual determinations.

23

0.020

0.010

0.000

-0.0050.00 4.00 8.00 12.00 16.00 20.00

Cha

nnel

A: A

bsor

banc

e

Time

Abs

orba

nce

(214

nm

)

0.030

0.020

0.010

0.000

0.00 5.00 10.00 15.00 20.00

11.6

4

13.9

3

17.6

3

5.0 10.0 15.0 20.0Time (min)

-1.00

100.00

UV

Abs

orba

nce

(arb

itrar

y sc

ale)

1

2

3

Figure 11. Separation of (1) paracetamol, (2) caffeine, and (3) 4-hydroxy-acetophenone (internal standard) by MECC. Instrument-specific settings:(a) ABI HT, (b) Beckman P/ACE, (c) Spectra Physics. Conditions: electrolyte,40 mM disodium tetraborate, 125 mM sodium lauryl sulphate; detection,210 or 214 nm; capillary, fused silica, 72 cm × 50 µm (a, c) 57 × 50 µm (b);temperature, 40°C. Reproduced with permission from Altria et al., 1994b.

b

a

c

24

A fully validated quality control method for hydrochlorothiazide andchlorothiazide drug substances has been reported (Thomas et al., 1994). TheMECC method also allowed the quantitation of a selection of synthetic impuri-ties at the 0.1% level. Injection precision of < 1% RSD was obtained withoutan internal standard. Careful control of rinse cycles and buffer replenishmentwere necessary to achieve this performance. The method was shown to con-form to USP validation guidelines. Validation included measurements of preci-sion, repeatability between analysts, capillaries and instruments. Othermeasurements included robustness testing and linearity of detector response(R2 = 0.995 to 0.999). Repeated analysis of various drug substance batches bydifferent analysts showed a little variability between assay values.

4 Chiral Separations

4.1 OverviewThis is an important application of CE in drug analysis as CE can offer distinctadvantages over alternative techniques. Several approaches to achieving chiralseparations by CE have been reviewed (Kuhn and Hoffstetter-Kuhn, 1992;Rogan et al., 1994). These include the use of chirally selective cyclodextrins,micelles, proteins, and crown ethers. The separation principles and theory aredescribed in depth in a previous primer.

4.2 Benefits of CE Methods for Chiral AnalysisChiral pharmaceuticals are generally synthesized as racemates or single enanti-omers. Analysis of a single enantiomeric form requires determination of levelsof the undesired enantiomer at less than 1% or even 0.1% in certain circum-stances. Quantitation at these levels presents a considerable challenge to cur-rent techniques since both appropriate sensitivity and selectivity are required.In fact, due to the poor peak efficiency and subsequent quantitative difficultiesin this type of HPLC, chiral analysis by CE is one area where CE typicallymatches, if not outperforms, HPLC.

When compared to HPLC, CE has several attractive features for chiralanalysis. These include the potential for speed of method development andanalysis, robustness, simplicity, and cost. An example of the type of perfor-mance achievable by CE is shown (Altria et al., 1994c) in Figure 12.

25

The undesired enantiomer of the BCH 189 racemate is detected at 0.3% by CEand the signal-to-noise ratio indicates a possible LOD below 0.1%.

4.00 9.00 14.00 19.00 24.00

13.6

71

19.4

29

17.8

04

1 2

UV

Abs

orba

nce

Figure 12. Analysis of a BCH 189 drug substance batch containing 0.3% ofthe (-) enantiomer (2). Separation conditions: electrolyte, 50 mM dimethyl-β-cyclodextrin in 50 mM borax, adjusted to pH 2.5 with conc. H3PO4; detection,214 nm; voltage, 13 kV; capillary, 47 cm × 50 µm. Reproduced with permis-sion from Altria et al., 1994b.

4.2.1 Speed of Method Development

Chiral HPLC method development can be a time-consuming process and caninvolve laborious investigations of numerous stationary phase and mobilephase combinations. Extensive column equilibration may also be required.When operating in normal phase conditions, sample pretreatment may berequired in HPLC to convert ionic drugs back to their base forms for reasons ofsolubility.

A particular attraction of the use of CE for chiral analysis is the simplicityand speed at which separation conditions may be assessed. With a knowledgeof the aqueous solubility and ionizable groups of the compound, initial separa-tion conditions can be selected (the previous primer on chiral separations con-tained sections on method development and optimization). Several potential

26

electrolyte compositions can be assessed in an overnight sequence and themost appropriate optimized further.

4.2.2 Speed of Analysis

Short analysis times are possible when operating with short capillaries andcertain electrolyte compositions. Figure 13 (Sepaniak et al., 1992) shows theresolution of phenylalanine enantiomers within 90 seconds. The enantiomersof the bronchodilator picumeterol have been resolved (Altria, 1993f) in 2.5minutes using a 27-cm capillary.

Retention Time (sec) Retention Time (sec)

30 120 30 60 90

Figure 13. (a) Separation of DNS phenylalanine enantiomers from the compo-nents in a commercial formulation (35 kV applied). (b) Rapid separation ofDNS phenylalanine (25 kV). Conditions: electrolyte, 0.01 m disodium phos-phate, 0.006 M disodium borate (pH 9); detection, laser-induced fluorescence;capillary, 50 cm × 25 µm. Reproduced with permission from Sepaniak et al.,1992.

a b

27

4.2.3 Robustness

Chiral HPLC is flexible in terms of the number of stationary phases and mo-bile phase additives that can be employed (Allenmark, 1988). However, theruggedness of these methods can be questionable, especially when sensitiveprotein phases are employed.

In most reported circumstances, chiral separations in CE have beenachieved using untreated fused-silica capillaries which are simple and repro-ducible. The robustness of chiral CE methods was highlighted (Altria et al.,1993a) in an inter-company cross-validation of a method for the enantiomericresolution of clenbuterol. The separation was achieved using a low-pH electro-lyte containing a derivatized cyclodextrin. All seven independent pharmaceuti-cal companies achieved baseline resolution, or better, for the clenbuterolenantiomers. Figure 14 shows three specimen separations. Acceptable data forlinearity, migration time, and peak area precision was found in all cases. Apeak area ratio precision of < 1% RSD and a 50:50 enantiomeric ratio werealso shown.

28

9.98

10.2

9

En 1 En 2



0.0000

0.0100

0.0150 11.4

511

.90

En 1 En 2

Abs

orba

nce

(arb

itrar

y un

its)

Abs

orba

nce

(arb

itrar

y un

its)

0.0000

50.0000

100.0000

Cur

rent

0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.00.022

0.027

0.031

0.035

0.040

0.044

0.048

0.052

0.057

0.061

En 1 En 2

Abs

orba

nce

(arb

itrar

y un

its)


Figure 14. Specimen chiral separations of clenbuterol achieved at three com-panies. Electrolyte: 30 mM hydroxylpropyl-β-cyclodextrin (typically 0.83 g per20 mL) in 50 mM disodium tetraborate, pH adjusted to 2.2 with concentratedorthophosphoric acid. Reproduced with permission from Altria et al., 1993a.

29

4.2.4 Cost

Chiral HPLC columns generally employ specialized stationary phases and aretherefore expensive (≈ $600). On the other hand, uncoated fused-silica capillar-ies are relatively inexpensive. Daily electrolyte requirements may be as little as20 mL of electrolyte containing a chiral additive at millimolar concentrations.This compares favorably to liters of an organic solvent-based mobile phase.

The principal disadvantage when using CE for chiral analysis is the sepa-ration of water-insoluble compounds. Separations have been attempted withchirally selective bile salt micelles (Nishi et al., 1990a) or with electrolytecontaining both SDS and cyclodextrins (Terabe et al., 1993).

4.3 ApplicationsThe principal application area is the enantiomeric purity testing of drug sub-stances. For example CE is used for the optical purity testing of Sandoz drugEN792 with an LOD of 0.2% for the undesired enantiomer (Kuhn et al.,1992a). Similar separation conditions were employed (Nielen, 1993) for chiralseparation of norephedrine and ephedrine and allowed the enantiomeric puritytesting of both compounds at the 1% level with acceptable precision (1.2 to2.5% RSD for areas and 0.2 to 1.5% RSD for optical purity results) and linear-ity better than 0.9997.

A chirally selective MECC method was developed for the optical puritytesting of a trimequinol drug substance with an LOD of 1% (Nishi et al.,1990a).

A CE method was used to monitor an enantioselective enzymaticbiotransformation reaction (Rogan et al., 1993). The reaction produced therequired (+) enantiomer by selective deamination of the undesired (-) formusing an enzyme. The reaction of the (-) was monitored over a 51-hour periodand the final product found to contain less than 0.5% of the (-) enantiomer.Reaction rate and half-life information was generated.

A chiral CE method was shown (Soini et al., 1992a) to be capable ofdetermining 0.1% of the S-bupivacaine in the presence of R-bupivacaine.

CE and HPLC have been employed (Altria et al., 1994c) to test the opticalpurity of picumeterol drug substance batches. Detector linearity of 0.9993 wasobtained for peak area ratio data plotted against prepared mixtures of picu-meterol enantiomers. Good agreement between the results was obtained con-

30

firming the accuracy of both techniques (less than 0.4% difference betweentechniques).

Enantiomeric ratios can alter following storage. For example L-epi-nephrine can undergo enantiomeric inversion to its D-isomer during storage.CE has been employed (Peterson and Trowbridge, 1992) to test enantiomericpurity of stored formulations containing L-epinephrine and acceptable preci-sion (1.4 to 1.8% RSD for peak area ratios), linearity (0.9988 to 0.9998), andrecoveries (99 to 101%) were obtained.

A validated method for the chiral analysis of fluparoxan drug substanceshas been reported (Altria et al., 1993b). A typical separation of a test mixtureof a fluparoxan enantiomer spiked with 1% of its stereoisomer is given inFigure 15. The test mixture was a synthetic mixture of 60:40% w/w of the (+)enantiomer: (-) enantiomer) and the peak area ratio obtained accurately con-firmed the spiking level. Repeated analysis of mixtures containing enantiomersspiked at the 1% level confirmed this spiking (0.8 to 1.1% peak area). Thevalidation of this method included precision peak area (1.6 to 2.0% RSD),linearity (0.994), and limits of detection (0.3%) and quantitation (1.0%). Themethod was capable of determining 0.3% of either enantiomer in the presenceof the other.

31

10.00 12.00 14.00

14.60

14.40

14.20

14.00

13.80

16.00 18.00 20.00 22.00 24.00

(+)

(-)


UV

Abs

orba

nce

(214

nm

)

Figure 15. Electropherogram of 1% (-) enantiomer in presence of (+) enanti-omer of fluparoxan. Separation conditions: 57 cm × 50 µm; electrolyte,10 mM borax, 10 mM tris, 150 mM β-cyclodextrin, 6 M urea-isopropanol(80:20) v/v; adjusted to pH 2.5 with H3PO4; temperature, 25°C; detection,15 kV. Reproduced with permission from Altria, 1993b.

Hohne et al. (1992) chirally resolved several aminoalcohols using electro-lyte-containing crown ethers. They reported a detection limit of 0.5% for theinactive enantiomer of methoxamine.

An inter-company cross-validation (Altria et al., 1993a) of a method forthe enantiomeric resolution of clenbuterol between seven independent pharma-ceutical companies showed good resolution (Figure 14). Linearity data fordetector response was assessed over the range 10 to 150% of target concentra-tion with correlation coefficients of 0.990 to 0.999. Precision for migrationtime and relative migration time was 0.2 to 1.3 and < 0.1% RSD respectively.Peak area and peak area ratio precision was found to be 0.8 to 2.5 and 0.2 to0.9% RSD. An average 50:50 enantiomeric ratio was obtained by the sevencompanies with an RSD of 0.6%.

32

Chiral separations of drugs in biofluids have also been reported (Gareilet al., 1993; Heuermann and Blaschke, 1993; Shibukawa et al., 1993). Thesimplicity and robustness of CE make it particularly attractive in this area.Examples include the quantitation of dimethindene enantiomers in urine, fol-lowing a single 4-mg dose (Heuermann and Blaschke, 1993), and the separa-tions of leucovorin and its major metabolite at therapeutic levels in plasma(Shibukawa et al., 1993).

In the body, enantiomers are often preferentially metabolized and a suit-able method is required to monitor these reactions. A chiral CE method wasused (Gareil et al., 1993) to confirm that the (-) enantiomer of warfarin is pref-erentially metabolized in patients undergoing warfarin therapy. Figure 16shows a typical separation with a LOD of 0.2 mg/liter for each enantiomer. Awarfarin homologue was used as an internal standard and an RSD of 2.1% wasobtained for peak area ratios for the standards.

8 10 12

0.001 AUWf

CIWf

min

Figure 16. Electropherogram of a plasma sample of a patient under warfarintherapy. Conditions: electrolyte, 100 mM sodium phosphate buffer (pH 8.35),8 mM Me-β-Cd-methanol (98:2, v/v); capillary, 72 cm × 50 µm ID; voltage,20 kV (I=70 µA); temperature, 25°C; UV detection, 310 nm; hydrodynamicinjection time, 1 s. The additional unlabeled peaks were not identified. Thetotal warfarin concentration determined was 2.0 mg/mL with an S/R enantio-meric ratio of 34:66. Wf = warfarin; ClWf = 5-chlorowarfarin. Reproducedwith permission from Gareil et al., 1993.

A validated method for the determination of cicletanine in both plasmaand urine has been reported (Soini et al., 1992b).

33

5 Stoichiometric Determinations

5.1 OverviewTo modify the physiochemical properties such as solubility, bioavailabilty, andbiocompatability, many drugs are prepared as salts. Typically, acidic drugs areprepared as their sodium or potassium salts. The range of salt forms availablefor basic drugs is more diverse, including both inorganic counter-ions such aschloride or sulphate and organic counter-ions such as succinate or maleate.

Many reports of the determination of inorganic ions or small organic acidsby CE have appeared (Jackson and Haddad, 1993a) in what is commonlyknown (Jandik et al., 1991) as capillary ion analysis (CIA). Generally, themethods employ indirect UV detection and have short (< 5-minute) analysistimes. Figure 17 shows (Altria et al., 1994b) the separation of a range of inor-ganic anions present at low ppm levels. Chromate is added to the electrolyte toprovide the necessary background UV response and the detector polarity isreversed.

2.50 3.00 3.50 4.00

6.30

6.10

5.90

5.70

Bro

mid

eC

hlor

ide

Sul

phat

eN

itrite

Nitr

ate

Flu

orid

e

Pho

spha

te

Figure 17. Separation of an anionic test mixture. Separation conditions: elec-trolyte, 5 mM chromate with 0.5 mM tetradecyl trimethyl ammonium bromide;capillary, 75 µm × 57 cm; voltage, -15 kV; detection, indirect at 254 nm. Re-produced with permission from Altria et al., 1994b.

34

5.2 Features of CE Methods for StoichiometricDeterminations

Drug:counter-ion stoichiometry is traditionally determined by ion-exchangechromatography (IEC) or titrimetry. CE represents an alternative techniqueand can offer advantages in terms of simplicity and cost.

5.2.1 Simplicity

In general uncoated fused-silica capillaries are employed in CIA which, asdiscussed in previous sections, are inexpensive. Aqueous-based electrolytes areused which can be prepared and stored for a long period of time prior to reuse.This compares to the preparation of IEC mobile phases and the use of IECcolumns which often require regeneration procedures. This testing is per-formed on standard CE equipment. Automated testing and data handling ispossible with both IEC and CIA, and both are therefore more desirable thantitrimetry.

5.2.2 Cost

Both sample solutions and electrolyte are aqueous and volume requirementsare minimal. The cost of a capillary, compared to a column, represents a fur-ther cost saving. A further advantage that may be overlooked is that the testingcan be performed on standard unmodified CE instrumentation.

The principal disadvantage of CE compared to IEC may be limits of de-tection. Low ppm levels of both cations and anions are possible by CIA whenemploying standard injection techniques. Single figure ppb levels are possible(Jackson and Haddad, 1993b) when electrokinetic injection is used. This sensi-tivity issue is irrelevant in stoichiometric determinations as high sample con-centrations (i.e., 100 to 200 ppm) are employed.

5.3 ApplicationsThis is a new CE application area and to-date few reports of the CIA analysisof pharmaceuticals have been reported. Undoubtedly more will appear as thetechnology matures. To illustrate possibilities, references to other applicationareas and, where possible, pharmaceuticals are given.

35

5.2.1 Anions

The two pharmaceutical application reports involve the determination of drugstoichiometry (Altria et al., 1994b) and monitoring of inorganic contaminantsin drug substances (Nair and Izzo, 1993). Other examples include the quantita-tive determination of inorganic anions in tap water (Motomizu, 1992) andbread (Ackermans, 1992c).

Stoichiometric testing of several drug substances has been conducted(Altria et al., 1994b). Figure 18 shows separations of a batch of a drug pre-pared as a chloride salt. Clearly this batch largely contains chloride but alsocontains low levels of other inorganic and organic anions as contaminants.

3.00 3.50 4.00 4.50

6.44

6.53

6.63

6.73

Chl

orid

e

Time (min)

Indi

rect

UV

at 2

54 n

m

Figure 18. Electropherogram of chloride assay of GGR1 drug substance.Conditions as per Figure 17. Reproduced with permission from Altria et al.,1994b.

Table 4 shows results for chloride and sulphate content in three drug sub-stances (denoted GRD1–3). Good agreement between CE, theoretical anioncontent, and microanalysis results are obtained (Altria et al., 1994b). Themethod performance in terms of linearity (R2 > 0.999) and RSD values of 1 to2% for peak area and migration time was also acceptable.

36

Table 4. Levels of Anionic Counter-Ions inDrug Substances Determined by CE

Good agreement was obtained between CE, microanalysis, and theoreticalcontent.

Sample Theoretical Microanalysis CE ResultsContent

Chloride (%w/w)

GRD 1 batch A 8.0 - 8.0 , 7.9

GRD 2 batch A 9.6 9.5 9.3 , 9.3

GRD 2 batch B 9.6 9.6 9.4 , 9.4

GRD 2 batch C 9.6 9.5 9.2 , 9.7

GRD 2 batch D 9.6 9.4 9.9 , 9.6

GRD 2 batch E 9.6 9.5 9.3 , 9.5

Sulphate (% w/w)

GRD 3 batch A 16.6 - 16.7

GRD 3 batch A 16.6 - 16.8

(Altria et al., 1994b)

When fully characterizing a drug substance, it is necessary to determinethe inorganic ion content which originates from synthetic reagents. If present,inorganic ions will contribute to the mass of the substance and will need to beassayed. CE has been employed (Nair and Izzo, 1993) to determine a range ofinorganic and organic ions present in drug substances. Limits of detection of< 0.1% w/w were reported. The method allowed the determination to be con-ducted on both water-soluble and insoluble drug substance materials. Insolubledrugs were dissolved in acetonitrile:water mixtures. Figure 19 shows the sepa-ration of a test mixture of inorganic ions dissolved in acetonitrile:water andanalysis of a 1-mg/mL solution of deuterated pravachol. Various validationcriteria were applied to the method including linearity (typically 0.999 in therange of 1 to 100 ppm), precision (1.3 to 6% RSD for anions spiked at the limitof quantitation, 0.1% w/w), and sensitivity (LOD of 0.05% w/w impurity inthe drug substance).

37

2.0 2.5 3.0 3.5 4.0 4.5 6.56.05.55.0

103.0

102.0

101.0

100.0

99.0

98.0

Inte

nsity

(m

V)

Time (min)

1mg/mL Deuterated Prava

1 µg/mL Anion Screen Standard

a

b

Figure 19. Electropherogram of (a) 1 µg/mL anion screen standard. Themigration order of the anions is bromide, chloride, sulphate, nitrite, citrate,fluoride, phosphate, carbonate and acetate. (b) 1 mg/mL deuterated pravachol.Separation conditions: electrolyte, chromate, dilute sulphuric acid and Wa-ters’ Anion-BT OFM; voltage, 20 kV; injection, 20 s, hydrostatic; detection,indirect UV at 254 nm. Reproduced with permission from Nair and Izzo, 1993.

5.3.2 Cations

Typical separation conditions employed (Beck and Engelhardt, 1992) involveuse of a low-pH electrolyte containing imidazole (or similar) as the UV back-ground electrolyte.

Figure 20 shows a typical separation of a standard metal ion test mixused in the determination of metal ions in vitamin tablets (Swartz, 1993).The method allowed low-ppm LODs to be obtained. Aspects of validationsuccessfully evaluated include linearity (better than 0.998), precision (< 1% formigration time and < 2% RSD for peak area), and sensitivity (1 ppm). Resultsfrom CE compared well with those obtained by ICP.

38

2.69

2.68

2.67

2.66

2.00 2.50 3.00 3.50

X 1

0 v

olts

-1

Time (min)

K

Na

Ba

Ca

MgLi

Figure 20. Separation of alkali metal and alkaline earth metal ions. Separa-tion conditions: 5 mM imidazole, pH 4.5; voltage, 25 kV; sample concentra-tion, 1 ppm (Swartz, 1993).

Levels of potassium and sodium content have been determined in variousacidic drug substances (Altria et al., 1994b). Table 5 shows that the resultsobtained by CE for three drug substance materials (denoted GRD4–6) agreewell with the theoretical and IEC results. Good method performance in termsof linearity (0.9999 over the range 10 to 500 mg/L), precision (0.3 to 1.5%RSD), and sensitivity (LOD of 1 ppm) was also obtained.

39

Table 5. Levels of Cationic Drug Counter-Ionsin Drug Substances Determined by CE

Good agreement was obtained between CE, HPLC (IEC) and theoretical con-tent.

Sample Theoretical HPLC CE ResultsContent

Sodium

GRD 4 batch A 5.2 - 5.2, 5.1

GRD 4 batch B 5.2 - 5.0, 4.8

GRD 4 batch C 5.2 - 4.9, 4.7

GRD 5 batch A 3.6 - 3.3, 3.4

Potassium

GRD 6 batch A 6.0 6.1 6.1

GRD 6 batch B 6.0 - 5.9

GRD 6 batch C 6.0 - 6.0

(Altria et al., 1994b)

Given the simplicity and robustness of the CE methods used for the deter-mination of drug stoichiometry, it is expected that there will be a significantincrease in application and perhaps even replacement of existing methods oftesting.

6 Quantitative Procedures

6.1 Overview

The options for conducting quantitative analysis are similar to those adopted inHPLC and have been reviewed (Altria, 1993g). The output format is similar toan HPLC chromatogram (i.e., a plot of UV absorbance versus time). There-fore, HPLC data handling and peak integration packages are generally appli-cable to CE. As discussed in the Introduction, it is important (Altria, 1993b) tonormalize peak areas in appropriate circumstances. Some important consider-ations and reminders essential for achieving good quantitative data are detailedbelow.

40

6.2 Impurity DeterminationsImpurity levels may be calculated and reported (Swartz, 1991) as a % of thetotal peak area of the electropherogram. This assumes an equal response factorfor all known and unknown impurities. Alternatively (Altria, 1993c), if iso-lated standards of impurities are available, then external standards can be pre-pared and response factors obtained. These response factors are then used tocalculate impurity levels as % w/w.

6.3 Main Component AssayAs with HPLC, sample solutions are analyzed against standards of knownconcentration. If an internal standard is employed as a means of increasingprecision (Dose and Guiochon, 1991), then ratios of the drug and internalstandard peak areas are used in the calculations.

6.4 Chiral AnalysisEnantiomeric content can be reported in several ways. The most commonexpresses the peak area of the undesired enantiomer as a percentage of totalpeak area. Alternative approaches determine the enantiomeric excess or useexternal standards of either enantiomer. Irrespective of which method is used,it is essential to use normalized peak areas (Altria, 1993b). This is exemplifiedby the data obtained for the repeated analysis of a racemic drug (Table 6). Apeak-area ratio of 1.00 should be obtained for a racemic compound. The un-normalized areas indicate a ratio of 1.04, which suggests that more of the sec-ond detected enantiomer is present. However, the expected ratio of 1.00 isobtained (Altria, 1993b) when calculating using normalized areas. If low levelsof undesired enantiomer are to be quantified, it may be necessary to employsufficient sample concentration such that the main peak is off-scale. In thiscase, an external standard of either enantiomer at an appropriate concentrationwould be employed to generate a response factor.

41

Table 6. Effect of Peak Area Normalizationon Peak Area Data for an Enantiomeric Separation

Normalization of the data confirms the correct area ratio for a racemic com-pound.

Enantiomer 1 Enantiomer 2 Peak area ratio

Peak area (observed) 1199781 1246293 1.04

Peak area (normalized) 103827 103841 1.00

Migration time (min.) 11.55 12.00

(Altria, 1993b)

6.5 Stoichiometric AnalysisAnalar-grade inorganic materials such as NaCl are employed (Altria et al.,1994b) as standards to generate a response factor for the ion concerned.Sample weights are adjusted to give an equivalent concentration of the ionbeing analyzed.

7 Method Validation Considerations

7.1 OverviewThe approach to the validation of a CE method is similar to that employed forHPLC methods (Altria and Chanter, 1993). Aspects such as accuracy, preci-sion, reproducibility, linearity, robustness, and method transfer are evaluatedduring validation. This section covers each of these criteria with illustratedexamples and references.

7.2 AccuracyGood agreement between CE results and the label claim or an alternative testmethod is often used to demonstrate accuracy. It is also recommended thatrecovery-type experiments are also performed, thus establishing the effect onquantitation of the matrix, if any.

42

7.3 PrecisionPrecision in CE, as with HPLC, is dependent on the type of method and itsdegree of optimization. Poor peak shape and low loadings will obviously resultin unacceptable performance. Precision in CE is typically of the order 0.5 to2% RSD for main peak assay, with or without internal standard (Altria andFilbey, 1993, Ackermans et al., 1992a). For trace impurities, the precisionwould be expected to be < 10% RSD (Swartz, 1991; Altria and Chanter,1993). Migration time and relative migration times should be about 1% RSDand < 1% RSD respectively. For example, in the transfer of a method for deter-mining the enantiomeric ratio of clenbuterol (Altria et al., 1993a), peak areaprecision was 1 to 2% RSD, peak area ratio precision was < 1%, and migrationtime precision was ≈ 0.1% RSD across seven companies.

7.4 ReproducibilityIt is necessary to demonstrate that the method can be reproduced in a varietyof situations. As with HPLC, reproducibility between capillaries, instruments,analysts, days, and laboratories is required. As the method may be employedon CE instruments from various manufacturers, it is important to demonstratethat acceptable performance can be attained using different sample introduc-tion modes and capillary dimensions.

Examples of reproducibility studies include assessments of CE methodsfor the determination of sumatriptan (Altria and Filbey, 1993), and hydrochlo-rothiazide (Thomas et al., 1994).

7.5 LinearityIt is important to show that a linear relationship between detector response andsample concentration exists within the working range required, in both directand indirect modes. Typically, a range of 50 to 150% for main peak assay andof 0.1 to 10% for related impurities should be assessed. The literature containsnumerous reports of the good linearity that may be expected for direct andindirect measurements. For example, correlation coefficients of better than0.999 have been reported for sumatriptan (Altria and Filbey, 1993), severalbronchodilators (Ackermans et al., 1992a), and alendronate (Tsai et al., 1992).

When determining impurity levels or trace enantiomer content, it is alsonecessary to demonstrate detector rectilinearity. This involves maintaining themain component at a constant concentration and varying the content of theminor component. Acceptable rectilinearities have been reported for impurities

43

of salbutamol (Altria, 1993c) and a quinolone antibiotic (Altria and Chanter,1993). Correlation coefficients of > 0.99 have been reported (Altria et al.,1993b) for fluparoxan enantiomer levels of 1 to 10% of the desired enantiomercontent.

7.6 RobustnessMany definitions exist for this term but, in the context of this primer, it relatesto the sensitivity of the method to small, deliberate deviations from themethod. For instance, if the method states an operating temperature of 30°C,will acceptable performances be maintained at either 25°C or 35°C? Twoapproaches to method robustness are possible: uni-variate or multi-variate.The uni-variate approach involves systematically varying each parametersequentially. This “one-by-one” approach has been performed (Thomas andGhodbane, 1993) for the determination of enalapril content in tablets by CE.The multi-variate assessment involves simultaneous evaluation of severalparameters using a predefined matrix. Typical experimental designs that maybe employed include Placket-Burman (Vindevogel and Sandra, 1991) andCentral Composites.

CE has been employed (Altria and Filbey, 1994) to determine levels of4-guanidino-Neu5Ac2en and related impurities. Figure 21 shows separation ofthe main component from six impurities (denoted I1 to I6). Both fractionalfactorial and central composite designs were employed in the robustness test-ing of this method. Figure 21 also shows (Altria and Filbey, 1994) a Paretoplot of the resolution of 4-guanidino-Neu5Ac2en form I2 as a function ofelectrolyte pH and voltage.

44

16

15

14

11

12 13

main peak

12.00

0.7631

0.5593

0.3555

0.1517

12.0011.20

10.409.60

8.808.00

RES2 (1)

B: VOLTAGE 1.9001.980

2.0602.140

2.2202.300

12.50

13.00

13.50

14.00

14.50

0.00 10.00 20.00 30.00 40.00

A: pH

Figure 21. Application of experimental design in robustness testing of a re-lated impurity determination method. (a) Separation of 4-guanidino-Neu5Ac2en and related impurities. (b) Response surface showing thecombined effect of pH and voltage on the resolution between the main peakand impurity I2. Separation conditions: electrolyte, 50 mM NaH2PO4, pH 2.1;injection, 10 s; detection, 230 nm; voltage, 10 kV; temperature, ambient; cap-illary, 50 µm × 47cm; concentration, 1 mg/mL. Reproduced with permissionfrom Altria and Filbey, 1994.

The need to demonstrate repeatability on instruments from at least twoinstrument manufacturers is stressed. Given the infancy of CE, the need forcomplete validation of methods is essential to generate further confidence andincreased momentum.

a

b

45

8 Future Developments

8.1 OverviewThere has been a rapid development of CE since the launch of commercialinstrumentation. Improvements in detector sensitivity and the introduction ofdiode array technology are assisting continued developments. This sectioncontemplates potential development areas that will influence the use of CE forpharmaceutical analysis. It is anticipated that the principal areas are new sepa-ration options, detection possibilities, capillary technology, andelectrochromatography.

Currently, the separation options available in CE are not as extensive asHPLC. The introduction of new electrolyte additives and surface-modifiedcapillaries are expected to have a major part to play. The flexibility and lowcost involved in CE separations will facilitate these developments.

Some examples include reports on the use of ionizable cyclodextrins(Nardi et al., 1993) and crown ethers (Kuhn et al., 1992b). Ion-pair reagentsand non-aqueous solvent systems will also extend the possibilities to water-insoluble drugs. The authors also recognize that the area of MECC is as yetrelatively untapped. The variety of off-the-shelf surfactants and possible per-mutations are astronomical. Novel chiral selectors such as proteins (Li andLloyd, 1993) will receive increased attention and will expand the area of chiralanalysis further (Figure 22). These investigations will give the analytical chem-ist of the future many options for solving a separation problem.

46

0 3 6 9 12Time (min)

a

b

c

d

e

Figure 22. Electrochromatograms showing the enantiomeric separations ofthe following: (a) disopyramide (15% 2-propranol/4 mM phosphate, pH 6.8;potential, 12 kV; current, 2 µA); (b) pentobarbital (2% 2-propranol/2 mMphosphate, pH 5.5; potential, 20 kV; current, 2 µA); (c) hexobarbital(2% 2-propranol/2 mM phosphate, pH 5.5; potential, 18 kV; current, 2 µA);(d) cyclophosphamide (3% 2-propranol/2 mM phosphate, pH 6.5; potential,25 kV; current, 2 µA); (e) benzoin (5% 1-propranol/5 mM phosphate, pH 6.5;potential, 15 kV; current, 3 µA). Reproduced with permission from Li andDavid, 1993.

Surface-modified capillaries are also expected to become important. Ini-tially, capillaries were coated to eliminate electroendosmotic flow, but increas-ingly specific coatings are being developed for selectivity purposes. Forexample, a CE column coated with immobilized cyclodextrin (Mayer andShurig, 1993) has been used for chiral separations of various non-steroidalanti-inflammatories.

Beckman has developed an amine-coated capillary which offers interest-ing selectivity possibilities. The coating results in the capillary surface having a

47

positive charge, whereas an uncoated capillary is normally negatively chargeddue to dissociation of surface silanols. Application of a voltage across theamine coated capillary generates a strong EOF in the opposite direction to thatobtained with an uncoated capillary and thus offers the potential for a reversalof migration order to that normally obtained. Figure 23 shows a test mixture offour basic drugs separated at pH 4.6 using an amine-coated capillary and anegative voltage. This separation could not be achieved on an uncoated capil-lary at the same pH. Undoubtedly the development and application of coatedcolumns will expand the possibilities in CE method development.

33.00

32.00

31.00

30.00

29.00

28.00

27.00

5.00 6.00 7.00 8.00 9.00

3

2

1

4


Figure 23. Separation of 4 basic drugs on amine-coated capillary. 25 mMNaH2PO4 (pH unadjusted), 200 nm, 2-second injection, 50 cm × 37 cm amine-coated capillary, -10 kV. Altria, 1994, unpublished results.

8.2 DetectionCurrently the vast majority of CE applications in pharmaceutical analysis haveinvolved the use of UV absorbance detection. Separation of pharmaceuticalshaving poor chromophores such as aminogyclosides (Figure 8) have involvedthe use of indirect UV detection and low UV wavelength detection.

48

New detection options on commercial instrumentation include diode array(DAD) and laser-induced fluorescence. DAD technology offers a number ofbenefits in areas such as peak identity confirmation and peak homogeneitytesting.

Currently the use of laser-induced fluorescence (LIF) detection is some-what limited in pharmaceutical analysis as few drugs have the required naturalfluorescence. The additional steps of preseparative sample derivatization isunattractive from a routine method perspective. However, appropriate pro-gramming of the separation method (Houben et al., 1993) can allow unat-tended on-column sample derivatization. The sample and the requiredderivatization reagents are injected into the capillary, allowed to react, and thederivatized sample components are then separated.

It is expected that more detection options will become commerciallyavailable to supplement those currently available. Most HPLC detection modeshave been reported for CE and include electrochemical (Yik and Li, 1992) andradioactivity detectors (Altria et al., 1990). The potential sensitivity and selec-tivity these offer will further widen the scope for application. The hyphenationof CE to MS is well established and a number of reports of CE-MS to druganalysis (Johansson et al., 1991) have appeared. Again, significant progress inthis area is expected.

8.3 ElectrochromatographyElectrochromatography is a technique in which a CE capillary packed withHPLC stationary phase material is employed for separations. The capillarycolumn is filled with electrolyte and a voltage applied. Application of thevoltage causes electroendosmotic flow to occur which effectively pumps elec-trolyte through the column. Solutes separate by virtue of their chromatographicinteractions with the packing and also by differences in electrophoretic mobili-ties. The technique is therefore essentially a combination of CE and HPLCand, as such, offers yet another dimension to method development options.

Electrochromatography can be operated on standard CE instruments andhas many of the features of CE in terms of reduced reagent purchase and sol-vent disposal costs but with the added advantage of good peak efficiencies.Little attention has been paid to this area as yet and the reports are limited(Yamamoto et al., 1992). Figure 24 shows the separation of israpidin whichillustrates the high separation efficiencies that are possible usingelectrochromatography using 3 mm ODS particles.

49

Abs

orba

nce

(220

nm

)

0.5 1.0 1.5 2.0

0.002 AU


isra

dipi

din

thio

urea

Figure 24. Separation of isradipidin by electrochromatography. Reproducedwith permission from Yamamoto et al., 1992.

9 ConclusionsCapillary electrophoresis is now firmly established as a viable option for theanalysis of pharmaceuticals. Specific application areas include the determina-tion of drug-related impurities, drug potency, chiral analysis, and determina-tion of drug counter-ion content. CE is often viewed as an alternative orcomplementary technique to HPLC.

Validated CE methods are in routine use in many industrial pharmaceuti-cal analysis laboratories. Validation criteria for CE methods are similar tothose employed in evaluation of HPLC methods.

Use of CE for specific analysis such as chiral analysis can have benefits interms of method robustness and ruggedness, cost, and time. Current disadvan-tages are largely poorer sensitivity when directly compared to HPLC and thelimited preparative options. Undoubtedly, technological developments andadvances in methodology will strengthen and endorse the position of CEwithin pharmaceutical analysis. Of considerable note are the possibilities thatthe further development of new detector options, coated capillaries, andelectrochromatography may bring.

Without doubt, CE has been recognized as a valid means of testing phar-maceuticals. To date, over 150 references have appeared on the subject. Thisnumber will no doubt double or triple in the next two to three years as thenumber of application areas expands.

50

Practical HintsThe following section is intended to highlight a number of practical aspectswhich should be considered during the development and subsequent use ofCE-based methods in pharmaceutical analysis. While these are by no meansexhaustive, they serve to alert the user to potential problem areas, their possibleremedies, and relevant literature references containing further details.

Electrolyte/SampleAnalyte peak shape can be dramatically improved by:

• matching the mobility of the analyte and the buffer co-ion (Mikkers et al.,1979);

• having the sample a factor of 102 lower than the buffer concentration(Mikkers et al., 1979);

• dissolving the sample in a low concentration buffer (water if possible) andanalyzing with a high-concentration electrolyte (Chien and Burghi 1992);

• to avoid band broadening due to Joule heating effects, buffers with lowspecific conductance should be used (generally operation above ≈ 100 µAcan lead to problems of internal heating in the capillary);

• buffers and samples should be filtered prior to use to remove particulates.

SensitivityThe sensitivity of a method using standard CE equipment (i.e., UV detector)can improved by:

• employing a “high-low” injection technique (see Section 2.3, Altria,1993e);

• using low UV wavelengths such as 200 nm where many components havesignificantly enhanced UV activity (Altria, 1993d);

• employing large sample volumes (i.e., long injection times) under stack-ing conditions (Chien and Burghi, 1992);

• using larger-bore capillaries than the standard 50 or 75µm. Bores as highas 180µm have been employed (Altria, 1993f) to dramatically increasesensitivity. However, deleterious heating effects may occur and the bufferconcentration, voltage, and capillary length should be adjusted to give anacceptable level of current.

51

QuantitationThe following measures may improve peak area precision:

• methods should be run at constant temperature as the volume injected isproportional to temperature (Rose and Jorgenson, 1988).

• peak area normalization (Altria, 1993b) should always be conducted. Theexception to this is where migration time precision is very good and exter-nal standardization is employed.

• internal standards can be used to increase precision (Dose and Guiochon,1991).

• the volume injected onto the capillary is proportional (Rose andJorgenson, 1988) to the viscosity of the sample and calibration solutions.Viscosity differences can be introduced when components such as cellu-lose are in the sample which may lead to apparent imprecision. Use of aninternal standard eliminates this difficulty.

• high sample concentrations should be used to reduce the effect of integra-tion errors.

Capillary CareIt is recommended that:

• the capillary is rinsed for 30 minutes with 0.1 M NaOH on its initial use;• the capillary is flushed with 0.1 M NaOH (or acid) and equilibrated with

the running buffer prior to each injection;• the capillary is rinsed with water and air blown through it at the end of

each day. This prevents buffer precipitation and subsequent capillaryblockage;

• individual capillaries be dedicated to specific methods to avoid “memoryeffects” which may lead to non-reproducible separations;

• an initial injection of a blank, or sample solution, be performed prior tocommencing an analytical sequence to allow the system to equilibrate.

52

ReferencesAckermans, M. T., Everaerts, F. M., Beckers, J. L. J. Chromatogr. 585,123-131 (1991a)

Ackermans, M. T., Everaets, F. M., Beckers, J. L. J. Chromatogr. 549,345-355 (1991b)

Ackermans, M. T., Beckers, J. L., Everaerts, F. M., Seelen, I. G. J. A. J.Chromatogr. 590, 341-353 (1993a)

Ackermans, M. T., Everaerts, F. M., Beckers, J. L. J. Chromatogr. 606, 229-235 (1992b)

Ackermans, M. T., Beckers, J. L., Everaerts, F. M., Hoogland, H., Tomassen,M. J. H. J. Chromatogr. 596, 101-109 (1992c)

Ackermans, M. T., Ackermans-Loonen, J. C. J. M., Beckers, J. L.J. Chromatogr. 627, 273-279 (1992d)

Allenmark, S. G. in Chromatographic Enantioseparation Methods and Appli-cations. New York: John Wiley and Sons, 1988.

Altria, K. D., Simpson, C. F. Anal. Proc. 23, 453 (1986)

Altria, K. D., Simpson, C. F. 1st Int. Symp. Pharm. Biomed. Analysis,Barcelona (1987) J. Biomed. Pharm. Anal. 6, 801-807 (1988)

Altria, K. D., Rogan, M. M. J. Pharm. Biomed. Anal. 8, 1005-1008 (1990)

Altria, K. D., Simpson, C. F., Bharij, A. K., Theobald, A. E. Electrophoresis11, 732-734 (1990)

Altria, K. D., Harden, R. C., Hart, M., Hevizi, J., Hailey, P. A., Makwana, J.V., Portsmouth, M. J. J. Chromatogr. 641, 147-153 (1993a)

Altria, K. D. J. Chromatogr.652, 459-463 (1993a)

Altria, K. D. Chromatographia 35, 177-182 (1993b)

Altria, K. D., Walsh, A. R., Smith, N. W. J. Chromatogr. 645, 193-196(1993b)

Altria, K. D. J. Chromatogr. 634, 323-328 (1993c)

Altria, K. D. J. Chromatogr. 636, 125-132 (1993d)

Altria, K. D. Chromatographia 35, 493-496 (1993e)

Altria, K. D. LC-GC Int. 6, 616-620 (1993f)

53

Altria, K. D. J. Chromatogr. 646, 245-257 (1993g)

Altria, K. D., Chanter, Y. L. J. Chromatogr. 652, 459-463 (1993)

Altria, K. D., Clayton, N. G., Harden, R. C., Hart, M., Hevizi, J., Makwana,J. V., Portsmouth, M. J. Chromatographia 39, 180-184 (1994a)

Altria, K. D., Rogan, M. M., Goodall, D. M. Chromatographia 38, 637-642(1994b)

Altria, K. D., Goodall, D. M., Rogan, M. M. Electrophoresis 15, 824-827(1994c)

Altria, K. D., Connolly, P. Chromatographia 37, 176-178 (1993)

Altria, K. D., Dave, K. J. Chromatogr. 633, 221-225 (1993)

Altria, K. D., Filbey, S. D. J. Liq. Chromatography 16, 2281-2292 (1993)

Altria, K. D., Luscombe, D. C. M. J. Pharm. Biomed. Analysis 11, 415-420(1993)

Beck, W., Engelhardt, H. Chromatographia 33, 313-316 (1992)

Buckberger, W., Haddad, P. R. J. Chromatogr. 608, 59 (1992)

Camilleri, P., Okafo, G. N., Southan, C., Brown, R. Analytical Biochemistry198, 36-42 (1991)

Chee, G. I., Wan, T. S. M. J. Chromatogr. 612, 172-177 (1993)

Chien, R. L., Burghi, D. S. Anal. Chem. 64, 489a (1992)

Dang, Q., Yan, L., Sun, Z., Ling, D. J. Chromatogr. 630, 363-369 (1993)

Dose, E. V., Guiochon, G. A. Anal. Chem. 63, 1154-1158 (1991)

Flurer, C. L., Wolnik, K. A. J. Chromatogr. 663, 259-263, (1993)

Fujiwara, S., Honda, S. Anal. Chem. 59, 2773-2776, (1987)

Gareil, P., Gramond, J. P., Guyon, F. J. Chromatogr. 615, 317-325 (1993)

Heuermann, M., Blaschke, G. J. Chromatogr. 648, 267-274 (1993)

Hohne, E., Krauss, G.-J., Gubitz, G., JHRCC 15, 698-700 (1992)

Honda, S., Iwase, S., Fujiwara, S. J. Chromatog. 404, 313-320 (1987)

Houben, R. J. H., Gielen, H., Van der Wal, S. J. Chromatogr. 634, 317-322(1993)

Jackson, P. E., Haddad, P. R. TRAC 12, 231-238 (1993a)

54

Jackson, P. E., Haddad, P. R. J. Chromatogr. 640, 481-487 (1993b)

Jandik, P., Jones, W. R., Weston, A., Brown, P. R. LC-GC 9, 634 (1991)

Johnasson, M., Pavelka, R., Henion, J. J. Chromatogr. 559, 515-528 (1991)

Jorgenson, J. W., Lukacs, K. D. Science 222, 266 (1983)

Kuhn, R., Hoffstetter-Kuhn, S. Chromatographia 34, 501-512 (1992)

Kuhn, R., Stoecklin, F., Erni, F. Chromatographia 33, 32-36 (1992a)

Kuhn, R., Erni, F., Bereuter, T., Hausler, J. Anal. Chem. 64, 2815-2820(1992b)

Li, S., Lloyd, D. Anal. Chem. 65, 3684-3690 (1993)

Mayer, S., Schurig, V. J. Liq. Chromatogr. 16, 913-931 (1993)

Mikkers, F. E. P., Everaerts, F. M., Verheggen, T. P. E. M. J. Chromatogr.169, 1-10 (1979)

Motomizu, S., Oshima, M., Matsuda, S., Obata, Y., Tanaka, H. AnalyticalSciences 8, 619 (1992)

Nair, J. B., Izzo, C. G. J. Chromatogr. 640, 445 (1993)

Ng, Cl., Lee, H. K., Li, S. F. Y. J. Chromatogr. 598, 133-138 (1992)

Nielen, M. W. F. Anal. Chem. 65, 885-893 (1993)

Nishi, H., Fukuyama, T., Matsuo, M. J. Chromatogr. 515, 245-255 (1990)

Nishi, H., Fukuyama, T., Matsuo, M., Terabe, S. J. Chromatogr. 498, 313-323(1990)

Nishi, H., Fukuywara, T.,Matsuo, M., Terabe, S. J. Chromatogr. 515, 233-243(1990a)

Nishi, H., Fukuyama, T., Matsuo, M., Terabe, S. J. Chromatogr. 513, 279-295(1990b)

Nishi, H., Terabe, S. Electrophoresis 11, 691 (1990)

Nishi, H., Tsumagari, N., Kakimoto, T., Terabe, S. J. Chromatogr. 477,259-270 (1989)

Nishi, H., Tsumagari, N., Terabe, S. Anal. Chem. 61, 2434-2439 (1989a)

Peterson,T. E., Trowbridge, D. J. Chromatogr. 603, 298-301 (1992)

Pluym, A., Van Ael, W., De Smet, M. TRAC 11, 27 (1992)

55

Qin, X. Z., Ip, D. P., Tsai, E. W. J. Chromatogr. 626, 251-258 (1992)

Rabel, S. R., Stobaugh, J. F. Pharm. Res. 10, 171-186 (1993)

Rogan, M. M., Altria, K. D. Presentation at 2nd Int. Symp. Chiral Discr.Rome: May, 1991.

Rogan, M. M., Altria, K. D., Goodall, D. M. Chirality 6, 25-40 (1994)

Rogan, M. M., Drake, C., Goodall, D. M., Altria, K. D. Anal. Biochem. 208,343-347 (1993)

Rose, D. J., Jorgenson, J. W. Anal. Chem. 60, 642 (1988)

Ryder, S. J. Chromatogr. 605, 143 (1992)

Sepaniak, M. J., Cole, R. O., Clark, B. K. J. Liq.Chromatogr. 15, 1023-1040(1992)

Shibukawa, A., Lloyd, D. K., Wainer, I. W. Chromatographia 35, 419-429(1993)

Soini, H., Snopek, J., Novotony, M. V. Beckman Application Data Sheet DS-836 (1992a)

Soini, H., Riekkola, M. L., Novotony, M. V. J. Chromatogr. 608, 265 (1992b)

Swartz, M. E. J. Liq. Chromatogr. 14, 923-938 (1991)

Terabe, S., Otsuka, K., Ichikawa, K., Tsuchiya, A., Ando, T. Anal. Chem. 56,111-113 (1984)

Terabe, S., Miyashita, Y., Ishihama, Y., Shibata, O. J. Chromatogr. 636, 47-55 (1993)

Thomas, B. R., Ghodbane, S. J. Liq. Chromatogr. 16, 1983 (1993)

Thomas, B. R., Fang, X. G., Chen, X., Tyrell, R. J., Ghodbane, S.J. Chromatogr. 657, 383-394 (1994)

Tsai, E. W., Singh, M. M., Lu, H. H., Ip, D. P., Brooks, M. A. J. Chromatogr.626, 245-250 (1992)

Vindevogel, J., Sandra, P. Anal. Chem. 63, 1530 (1991)

Watzig, H., Dette, C. J. Chromatogr. 636, 31-38 (1993)

Yamamoto, H., Baumann, J., Erni, F. J. Chromatogr. 593, 313-319 (1992)

Yik, Y. F., Li, S. F. Y. TRAC 11, 325-332 (1992)

Zhang, C. X., Sun, Z. P., Ling, D. K., Zhang, Y. J. J. Chromatogr. 627, 281-286 (1992)

C a p i l l a r y E l e c t r o p h o r e s i s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

T e c h n i c a l ␣ I n f o r m a t i o n

T-1823A

BECKMAN

Richard H. Palmieri, Ph.D.Beckman Instruments, Inc.

Comparison of Air and Liquid Cooling in Capillary Electrophoresis

SummaryData presented in this article demonstrate the advan-tages of a liquid-cooled capillary electrophoresissystem when compared with air-cooled instruments.Liquid cooling will accommodate a wide variety ofcommon experimental conditions (moderate to highionic strengths; low electroosmotic flow) without theneed to compromise on experimental design.

I. Introduction

Current Status

Although capillary electrophoresis (CE) has gainedacceptance in the competitive separations market,the reproducibility of the technique has at timesbeen questioned.(1) Compared to HPLC (CVs <1%),CE can have values which are greater than 5%. Thelarger-than-expected variances have been attributedto inadequate system equilibration,(2) variable elec-troosmotic flow (EOF),(3) and/or inefficient coolingsystems.(4) In most cases, the equilibration and EOFissues have been addressed and remedied. This bul-letin will address the relationship between coolingmethods and system performance.

In the following sections, we will define an im-portant experimental criterion to evaluate coolingeffectiveness, review the principles of heat genera-tion and transfer, analyze the relationship betweensystem parameters and heat generation, and examinethe effect of temperature on system performance.In the process we will demonstrate that:1) Experimental compensations for heating effects

can be limiting;2) One method of cooling is not just as effective as

another;3) Air-cooled systems tend to run hotter than the

corresponding liquid cooled systems; and4) Temperature fluctuations encountered in air-

cooled systems can compromise system valida-tion.

II. Criteria for Cooling Effectiveness

A. General Features

A viable cooling system must be able to accommo-date typical capillary electrophoresis applicationswithout compromising experimental conditions.The requirements of the sample, not the hardware,

2

should define the experimental parameters. To ac-commodate the needs of the pharmaceutical and bio-tech companies, the cooling systems must bevalidatable, i.e., ensure reliability and reproducibleperformance.

B. Evaluation Conditions

As we examine the effects of different cooling sys-tems in the following sections, it will be importantto evaluate the conditions under which the compari-sons were made (buffer pH, ionic strength, EOF,and temperature). Along the way, we will point outthe limitations of the low ionic strength “DesignerBuffers” systems and include examples from highionic strength buffers (>100 mM) since they are in-tegral to a number of CE techniques. Such condi-tions are required to prevent analyte adsorption, tofacilitate sample stacking, to control analyte/analyteinteractions, to optimize the MEKC method, or toincrease buffering capacity of the system.

III. Fundamental PrinciplesA. Joule Heating and System Parameters

Before considering the effects of different coolingmodes, it is appropriate to examine the theoreticalbasis for heat generation and cooling during capil-lary electrophoresis.

When current passes through a solution, electri-cal energy is partially converted into Joule heating(Figure 1).(5) Depending on the application/electro-phoresis conditions, temperature changes may rangefrom a few degrees above the surroundings to boil-ing/outgassing.(6) Unless this heat is removed (Fig-ure 1), it can have an impact upon the electrophore-sis results primarily though changes in the migrationtime (viscosity, net charge, EOF) and changes in theresolution/profile (diffusion coefficient and stabil-ity). In some cases, quantitation is affected fromlosses due to thermal expansion.(7)

It is important to keep in mind that changes inJoule heat may also occur randomly. These factorscan be induced by changes in the buffer (ion de-pletion, ionization) and/or the EOF (low buffercapacity, analyte adsorption, hydrostatic flow). Itis imperative to control Joule heating since thisparameter is directly linked to analyte mobility,resolution, and stability, as well as systemreproducibility.

32

1

Voltage

Ionic Strength

µapp = µs + µe

Heat

Current

(4) Net chargechanged

Mobility

EOF Cooling

Viscosity

(2) Chargesuppression

EOF

pH

4

(1&2)

Temperature

(1) Adsorption/chargemodification

(3) Hydration/sizealtered

Figure 1. Relationship between temperature andsystem parameters in capillary electrophoresis.

During electrophoresis, thermal gradients de-velop radially and transversely (the long capillaryaxis).(4,8) This occurs because heat is not removedrapidly or uniformly enough from the capillary viaconduction, convection, radiation, or displacement(EOF). These processes are not uniformly distrib-uted even in the best of systems. In addition, theheat transfer efficiency differs throughout the sys-tem. A diagram of this process is shown inFigure 2A. Since the temperature is highest at thecenter of the capillary and only somewhat lower atthe capillary wall, most of the temperature drop oc-curs between the outside wall and the cooling me-dium.

In the body of the capillary, heat is transferredby conduction across the capillary wall, and by con-vection with respect to the coolant which may be airor liquid (see Figure 2B).(10) The buffer vial func-tions as a component of the cooling system, i.e., itacts as a heat sink for the respective ends of the cap-illary and serves as a source of cooled buffer for the“pumping” action of EO flow.

B. Experimental Compensation

A common empirical approach has been to compen-sate for heating effects by modification of the ex-perimental parameters.(9) “First Principles” havebeen used to justify modification of one or more ofthe experimental conditions, i.e., the buffer ionicstrength, the capillary i.d., or the run voltage, in or-der to reduce the effects of Joule heating. Whilethese changes result in a reduction in heat, they donot necessarily produce the most efficient or desiredresults and can represent an unsatisfactory compro-mise (see Table 1).

3

A

B

parabolic

∆T

i.d.o.d.

tube bore

capillary walls

logarithmic

surroundingair

PeltierControl

Coolant Bath

Capillary

Mandrel Cartridge

Slit

Vials

• Capillary/cartridge design allows effective heatdissipation and temperature control

• Peltier controlled, recirculating fluid range: 15–50°C

Figure 2. (A) Example of the thermal gradientsencountered during electrophoresis (from Knox, J.,Chromatographia 26, 329-337 (1988)(B) Cross-section of a liquid-cooled capillaryelectrophoresis system (Burolla et al., Am Biotech-nol. Lab. 7 (1989)).

IV. Cooling Effectiveness and SystemParametersA. Ohm’s Law Relationship

Ohm’s Law can be used to evaluate the relative ef-fectiveness of different cooling methods.(11)

The voltage (V) is linearly related to the current (I)as long as the resistance remains constant (V = IR).If the temperature changes as a result of the elec-trophoretic process, corresponding changes will oc-cur in the resistance and the current will deviatefrom linearity. This condition typically prevails asa consequence of excess Joule heating.

It is possible to use this relationship to define alimiting value for the field strength (E, volts/cm) fora given capillary, buffer, temperature, and coolingcombination. A plot of the potential (kV) vs. current(µamps) will be linear as long as the temperature iscontrolled. As the voltage increases, excess heat willnot be completely removed; the internal temperaturewill rise, viscosity will decrease, and the current willincrease. A comparison of both air and liquid cool-ing is shown in Figure 3.(12)

It is obvious that liquid cooling allows signifi-cantly higher field strength (E, volts/cm) to be em-ployed which means faster run times and better effi-ciency. In addition, the higher allowable “runningcurrents” (linear portion) observed with the forced-air system correspond to higher internal tempera-tures for a chosen value of E. This difference whichis subject to experimental variation can be expectedto affect the reproducibility and reliability of thesystem (see Figure 1 for the interrelationship of thevarious experimental parameters).

250

200

150

100

50

00 5 10 15 20 25 30

Voltage (kV)

Cur

rent

(µA

)

Liquid

No CoolingForced Air

Figure 3. Ohm’s Law Plot for a 75 µm i.d. by 57 cmfused-silica capillary. The buffer was 50 mM sodiumphosphate, pH 4.3; the temperature was set at 25°C.(12)

Table 1. Typical Adverse System EffectsCaused by Varying Experimental Parameters

to Reduce Joule Heating

Options Effect

1. Lower voltage • Longer run times

• Increased band broadening

• Decreased throughput

2. Smaller i.d. • Decreased sensitivity

capillary • Manipulation problems

• Reduced fraction collectioncapability

3. Decreased • Faster EOFs

ionic strength • Lower sensitivity

• Increased adsorption

• Decreased resolution

• Longer regeneration times

While all three options result in lower overallcurrent, choices #1 and #3 can have a direct impacton the separation, while the second option adds fur-ther restrictions to a technique that is alreadysample-limited for preparative studies. Obviously,a more efficient cooling system is the solution ofchoice over experimental modifications.

4

While changes in current are indicative of tem-perature changes for this set of moderate buffers,i.e., medium ionic strengths and low to mediumEOFs, other more taxing conditions are also em-ployed which cause even larger current changes.What then is the relationship between the currentchange and the internal temperature?

B. Run Buffers and Capillary Temperature

In this section, we will examine the impact of vari-ous buffers on the internal temperature of the capil-lary vs. a fixed external set temperature for an air-cooled system.

The Joule heat generated by a given buffer canbe evaluated in system terms, i.e., the difference intemperature between the center of the capillary andthe capillary wall or between the center of the capil-lary and the surroundings.(13) Somewhat unexpect-edly, the system can “efficiently” transfer heat fromthe central lumen to the outside capillary wall undera variety of conditions (Figure 4A and Equation 1).At most, the maximum temperature changes repre-sents only a few degrees.

By contrast, the difference in temperature be-tween the central lumen and the surrounding cool-ing “fluid” is rather large (Figure 4B and Equations2 and 3). Using typical buffer conditions, the inter-nal temperatures ranged from 5°C to greater than40°C above the set temperature when a fieldstrength of 300 V/cm was used! For 100 mM so-dium phosphate, pH 7.0, these changes were solarge that one or more of the experimental condi-tions must be changed before continuing (Section III B).

For “typical” buffer conditions (curves 2 and 3,Figure 4), temperature changes in the 5 to 10°Crange might be expected when operating at 300 V/cm.How significant are these changes and what impactdo they have on system performance, i.e., on migra-tion time or on mobility? These questions will beaddressed for both types of cooling systems in thenext section.

V. Cooling Effectiveness and SystemPerformanceCritical Factors

The effect(s) of temperature on the system param-eters (migration time, EOF, and analyte mobility)will be considered under two different sets of condi-tions, i.e., moderate pH and low-moderate ionicstrength or under low pH and moderate-high ionicstrength. In the latter case, the influence of EOF has

40

30

20

10

00 100 200 300 400

B 1

2

3

0.6

0.4

0.2

0

A

1

2

3

T0-

T1(

C)

T0-

Ta(

C)

E (V/cm)

Figure 4. Actual vs. set temperature in an air-cooledsystem. Calculated curves showing the change intemperature that occurs (A) between the centralcapillary lumen and the capillary wall (T0-T1) and(B) between the center of the capillary and thecoolant (T0-Ta) as a function of electrical fieldstrength for three different buffers. The curves weregenerated using fundamental equations andexperiment results from the following buffers:(1) 100 mM sodium phosphate, pH 7.0; (2) 50 mMsodium citrate, pH 2.5; (3) 20 mM CHAPS, pH 11.0Paul Grossman, in Capillary Electrophoresis, p. 14.P. D. Grossman and J. C. Colburn, eds., AcademicPress, San Diego, 1992.

been removed so that analyte mobility must be usedto evaluate the relative effectiveness of the coolingsystems. It will be readily apparent that buffer EOFalong with ionic strength must be taken into consid-eration when drawing any conclusions from thiscomparison.

1. Temperature and Migration Time

From a theoretical consideration, electrophoretic-mobility changes exponentially with the temperature(Equation 5).(14) However, over a narrow range, themobility increases approximately 1-2% per degreeKelvin. The experimental results in Figure 5 illus-trate this relationship. As the temperature increases,there is an approximately linear decrease in the mi-gration time. It is apparent that the relatively smallchanges in temperature (± 5°C) cannot be neglectedsince they correspond to significant changes in mi-gration time.

5

into the system from the buffer vial. Localized“hot spots” do not develop since these sections arecontinuously moving. Thus, higher fields can be em-ployed without experiencing failure. However, whenEOF is eliminated, and more robust buffers are em-ployed, greater system differences are apparent.

3. Effects of Temperature on Analyte Mobility inthe Absence of EOF

To compare the cooling efficiencies of different sys-tems, low EOF and moderate to high ionic strengthsconditions (>100 mM) should also be employedsince they challenge the system. These conditionsare representative of a number of key CE applica-tions (see Section II. B). To monitor the differences,analyte rather than EOF mobilities will be used inthese studies since the EOF is low.

Figure 5. The effect of temperature on theelectrophoretic behavior of peptides. Separationof a standard mix of peptides [50 µg/mL each, in5 mM phosphate buffer, pH 2.5; 5 sec. pressureinjection at0.5 psi]. Separation conditions: capillary,75 µm × 50 cm (effective length) fused silica; buffer,100 mM sodium tetraborate, pH 8.3; voltage,350 V/cm; detection, 214 nm. Adapted from Rush,R. S., Cohen, A. S., and Karger, B. L., Anal. Chem. 63,1346 (1991).

2. Temperature and EOF

Comparisons of cooling systems are often madewith moderate EOFs (pH 6-7) and moderate-lowionic strengths (< 50 mM) [see Figure 6, and refer-ence 15]. Under these conditions, there seems to beno particular advantage to either cooling systemsince they behave similarly. However, the actual dif-ferences have been minimized due to EOF cooling.With this process, cool buffer is continually pumped

15

10

5

µeo. 1

04[c

m2 /

Vs]

0 100 200 300 400 500

E [V.cm-1]

cb

a

Figure 6. A plot of electroosmotic mobility vs. fieldstrength for a capillary electrophoresis system witheither (A) no cooling, (B) air-cooling, or with(C) liquid cooling. Experimental conditions: fusedsilica capillary, 75 µm i.d. by 50 × 57 cm; injection,1 sec. using 0.5 psi; buffer, 50 mM sodium phos-phate, pH 7.0; temperature, 30 C; detection, 200 nm.Benzyl alcohol was used as the EOF marker. Takenfrom Kuhn, R., in Capillary Electrophoresis:Principles and Practice, p. 48, R. Kuhn and S.Hoffstetter-Kuhn, eds. New York, Springer-Verlag,1993.

6

Figure 7. Effect of field strength on mobility andpeak width using either (A) a liquid or (B) an air-cooled system. For these analyses, the buffer was150 mM Phosphate, pH = 2.8, detection was at214 nm. Run conditions (V/cm) for panel A, top tobottom, were 175, 263, 350 and 439; conditions forpanel B, top to bottom, were 153, 255, 357 and 459.The capillary (liquid-cooled, Beckman) was 75 µmi.d. fused-silica, with an effective length of 50 cm× 57 cm while the air cooled system was 75µm i.d.which was 40 cm × 49 cm in length. In each case,aliquots of a peptide mixture (50 µg/mL each ofangiotensin I and II, bradykinin, and neurotensin in0.1% TFA) were injected for 5 sec. at 0.5 psi. Forboth instruments, the cooling systems were thermo-stated at 25°C while the buffer vials were kept atambient temperature. Buffers were changed after threeruns in order to avoid “Buffer Depletion effects.(3)

A

B

100 200 300 400 500

Field Strength (Volts/cm)

Mob

ility

(cm

2 /vo

lt se

c X

10-

4 )

1

2

3

4

5Effect of Field Strength on Mobility

Liquid

Air

1 2 3 4 5 6 7 8 9 10 11

Power (Watts/Meter)

2

3

4

5Effect of Power on Mobility

Air

Liquid

Mob

ility

(cm

2 /vo

lt se

c X

10-

4 )

Figure 8. (A) Comparison of analyte mobility atdifferent field strengths using either a liquid- or anair-cooled system. Conditions are as in Figure 7.In each case, the apparent mobility was calculatedfor peak 3 and plotted against the field strength.(B) Comparison of analyte mobility at differentpower settings using either an air- or liquid-cooledsystem. The conditions for this experiment are thesame as Figure 7. This method of plotting the dataallows us to include the current along with thevoltage in our consideration.

A

B

All analyses were conducted with a four-com-ponent “standard” peptide mix which included an-giotensin I and II, bradykinin, and neurotensin.The buffer was chosen for its low EOF and moder-ate-high ionic strength (150 mM sodium phosphate,pH 2.8) (Figure 7A & B). The successive electro-pherograms were obtained by increasing the runvoltage by 5 kV between runs. Only a portion of theanalysis could be completed at approximately400 V/cm (Figure 7B) since the air-cooled systemcrashed after a few minutes.

While both cooling systems demonstrated fieldstrength dependence, the liquid-cooled unit clearlyoutperformed the air-cooled system (Figure 8A).

7

In each case, the apparent mobilities were plottedagainst the field strength. The mobilities from theair-cooled system were approximately 20% and30% higher at 300 V/cm and 400 V/cm respec-tively! More significantly, the initial portion of thecurrent traces tended to be irregular throughout theexperiments with the air-cooled system. This condi-tion finally led this system to fail during the analysisat 400 V/cm. By contrast, no current irregularitieswere observed with the liquid-cooled unit. This unitcontinued to operate until the limiting current of250µamps was reached.

It is informative to consider the effect of power(watts/meter) on cooling system performance. Thisis shown in Figure 8B where the mobility is plottedagainst power. As expected, both curves show a lin-ear correlation but the slopes are significantly differ-ent. This difference is reflected in higher internaltemperatures for the air-cooled system, resulting inshorter migration times, higher mobilities, broaderpeaks, and poorer resolution for the same analytewith the same buffer using the same conditions.

Since internal temperature is a system param-eter, validation becomes increasingly difficult to as-sign since the actual temperature is unknown andsubject to experimental fluctuations from changes incapillary diameter, EOF, ionic strength, and hydro-static head pressure.

VI. ConclusionsFrom a comparison of the air- and liquid-cooled sys-tems, it is possible to show that:• Heat is more efficiently transferred to a liquid

than to air no matter how fast the air is moving;• Higher internal temperatures result in shorter mi-

gration times, broader peaks, and variable results;• Using so-called “Designer Buffers” for compara-

tive studies can be misleading since the EOFcooling tends to minimize actual system differ-ences;

• Comparison studies should be performed with abroad range of buffers that include low EOF andmoderate to high ionic strength buffers which canchallenge the cooling system;

• Liquid-cooled systems possess the distinct advan-tage in cooling efficiency since they accommo-date a broader range of CE application require-ments.

EquationsEquation 1. Temperature differences can be ex-pressed in terms of watts/volume generated, capil-lary radii (inside & outside), and the thermal con-ductivity of the buffer. The temperature differencewithin the buffer is of concern since its potentialparabolic profile would tend to promote decreasedefficiency. This relationship is shown as:

T0-T1 = PmR12/4kb

where

Pm = Watts/capillary volume

R1 = Capillary radius (inside)

T0 = Temperature at capillary center

T1 = Temperature at capillary wall (inside)

kb = Thermal conductivity of buffer

Equation 2. The temperature change between thecenter of the capillary (T0) and the surroundingcoolant (Ta) can be approximated from the follow-ing relationship:

T0-Ta = PmR12/R3h

where

Ta = Temperature of surroundings

R3 = The outside radius

h = The heat transfer coefficient

T0 = Temperature at capillary center

It can be difficult to determine the ∆T with thisequation since the value of “h” is a function of theparticular cooling mode, configuration of the capil-lary, and the net EO flow. It is possible, however, toestimate ∆T from another relationship (Equation 3)and then derive a value for “h” which should be in-strument/capillary specific.

Equation 3. The internal temperature can be calcu-lated from a measurement of the electroosmotic mo-bility at two different voltages according to theequation:

where T2 is the internal temperature, µeo1 and µeo2

are the EO mobilities at voltages 1 and 2, andA = Ea/R ≈ 1820K and B = Ea/RT1 [B = 6.11 whenT1 is measured at room temperature, or 298K] areconstants which were derived from the temperaturedependence of the mobility.(14) The first voltagemust be chosen well below the Ohm’s Law break sothat the internal and set temperature are identical.

T2 = A

ln(µeo1) - ln(µeo2) + B

Note: This expression should also beused for estimating the internal temperatureparameter for various methods (MEKC, in-clusion CE, etc.) and for system validation. Itis not reasonable to rely on the external set-ting of the temperature.

Equation 4. Mobility is related to tempera-ture through the viscosity (µ=Q/6πηr) whileresolution is directly proportional to the ap-plied voltage (V) and inversely proportionalto the diffusion coefficient (D) according tothe following equation:

Rs = C1 [V/D (mav + meo)]1/2

Rs = C2 [Vh/T (mav + meo)]1/2

where “C” is a constant, and µav, µeo are theaverage mobility and EO mobility respec-tively. Temperature is incorporated into thisexpression from Fick’s law, i.e., D = kT/6πηr(k = Boltzman constant, T = absolute tem-perature, η = viscosity, and r = hydrody-namic radius).

Equation 5. The influence of temperature onmobility is attributed to changes in the viscos-ity. The dependence of viscosity on tempera-ture is expressed by the relationship:

η = CeEa/RT

where C is a constant, Ea is the activation en-ergy for viscous flow, R is the gas constant,and T is the temperature in Kelvin.(15) Sincethe mobility is inversely related to the viscos-ity, it is apparent that it will increase expo-nentially with increases in temperature.

References1. Eby, M. The Reality of Capillary Electro-

phoresis. Biotechnology 7, 903 (1989)2. Landers, J. P., Oda, R. P., Madden, B. J.,

Sismelich, T. P., Spelsberg, T. C. Repro-ducibility of Sample Separation UsingLiquid or Forced-Air Convection Thermo-stated High Performance CapillaryElectrophoresis. J. High Res. Chro-matogr. 15, 517 (1992)

3. Strege, M. A., Lagu, A. L. Studies ofMigration Time Reproducibility of CEProtein Separations. J. Liq. Chromatogr.16, 51 (1993)

4. Grushka, E., McCormick, R. M.,Kirkland, J. J. Effect of TemperatureGradients on the Efficiency of CapillaryZone Electrophoresis Separations. Anal.Chem. 61, 241 (1989)

5. Rush, R. S., Cohen, A. S., Karger, B. L.Influence of Column Temperature on theElectrophoretic Behavior of Myoglobinand aplha-Lactalbumin in High-Perfor-mance Capillary Electrophoresis. Anal.Chem. 63, 1346 (1991)

6. Burgi, D. S., Salomon, K., Chien, R. L.,Methods for Calculating the InternalTemperature of Capillary Columns Dur-ing Capillary Electrophoresis. J. Liq.Chromotogr. 14, 847-867 (1991)

7. Kuhn, R., in Capillary Electrophoresis:Principles and Practice, p. 48. R. Kuhnand S. Hoffstetter-Kuhn, eds. New York,Springer-Verlag, 1993.

8. Knox, J. Thermal Effects and BandSpreading in Capillary Electro-Separa-tion. Chromatographia 26, 329-337(1988)

9. McCormick, R. M. Capillary Zone Elec-trophoretic Separation of Peptides andProteins Using Low pH Buffers in Modi-fied Silica Capillaries. Anal. Chem. 60,2322 (1988)

10. Burolla, V. P., Pentoney, S. L., Jr., Zare,R. High Performance Capillary Electro-phoresis. Am. Biotechnol. Lab. 7, 10(1989)

11. Nelson, J. R., Burgi, Dean, S. Tempera-ture Control in Capillary Electrophoresis,in Handbook of Capillary Electrophore-sis, Chapter 21, J. Landers, ed. Ann Ar-bor, CRC Press, 1994.

8

9

12. Allen, J. Unpublished results, BeckmanInstruments, 1993.

13. Grossman, P. D. Factors Affecting thePerformance of Capillary ElectrophoresisSeparations: Joule Heating, Electroosmo-sis, and Zone Dispersion, in CapillaryElectrophoresis, Chapter 1. P. D.Grossman and J. C. Colburn, eds. SanDiego, Academic Press, 1992.

14. Burgi, D. S., Salomon, K., Chien, R. L.Methods for Calculating the InternalTemperature of Capillary Columns Dur-ing Capillary Electrophoresis. J. Liq.Chromotogr. 14, 847-867 (1991)

15. Kuhn, R. and Hoffstetter-Kuhn, S. Fac-tors Influencing Performance, in Capil-lary Electrophoresis: Principles andPractice, Chapter 3. R. Kuhn and S.Hoffstetter-Kuhn, eds. New York,Springer-Verlag, 1993.


BECKMAN Beckman Instruments, Inc. • 2500 Harbor Boulevard, Box 3100 • Fullerton, California 92634-3100

Sales: 1-800-742-2345 • Service: 1-800-551-1150 • Internet: http://www.beckman.com • Telex: 678413 • Fax: 1-800-643-4366

Worldwide Offices: Africa, Middle East, Eastern Europe (Switzerland) (41) 22 994 07 07. Australia (61) 2 844 6000. Austria (43) 222 7292164.Canada (800) 387-6799. China (861) 5051241-2. France (33) 1 43 01 70 00. Germany (49) 89-358700. Hong Kong, PRC (852) 814 7431.Italy (39) 2-953921. Japan 3-5352-2825. Mexico (525) 559-1635. Netherlands 0297-230630. Poland (48) 22 408822, 408833.Singapore(65) 339 3633. South Africa (27) 11-805-2014/5. Spain (1) 358-0051. Sweden (8) 445 11 20. Switzerland (41) 22 994 07 07.Taiwan (886) 02 378-3456. U.K. (01494) 441181. U.S.A. 1-800-742-2345.

BR95-2890-CB-10 © 1996 Beckman Instruments, Inc. Printed in U.S.A. on recycled paper.

iii

Table of Contents

About the Author ................................................................... iv

Acronyms and Symbols Used................................................. v

I. Introduction ............................................................................. 1

II. Separation Principle/Fundamentals ........................................ 4

Capacity Factor ................................................................ 5

Effective Mobility ........................................................... 7

Operating Conditions....................................................... 8

Composition of the Micellar Solution ........................... 10

Factors Affecting Reproducibility ................................. 11

Marker of the Electroosmotic Flow and the Micelle .... 12

Resolution ...................................................................... 13

III. Selectivity Manipulation ....................................................... 20

IV. Enantiomeric Separations ..................................................... 38

V. Strategy for Optimizing Resolution ..................................... 43

iv

About the Author

Shigeru Terabe is a professor of analyti-cal chemistry at the Himeji Institute ofTechnology in Kamigori, Hyogo (Ja-pan). His research interests include thedevelopment of high-resolution separa-tion methods such as capillary electro-phoresis, electrokinetic chromatography,and open-tubular liquid chromatography.His scientific publications span 65 re-search papers, 19 review papers, and 2books. Dr. Terabe is well known for hispioneering work in the field of micellarelectrokinetic chromatography. He is amember of the editorial advisory boardsof the Journal of Microcolumn Separa-tions, Analytical Chemistry, the Journalof Chromatography, Chromatographia,and the Journal of Biochemical and

Biophysical Methods.

Acknowledgements

We would like to thank Jeff Allen, Ron Biehler, Phyllis Browning, AndyJacobs, and Kathi Ulfelder at Beckman Instruments for reviewing drafts ofthe manuscript and for their many suggestions. We are also grateful to DonGregory at Polygen/Molecular Simulations, Sunnyvale, CA, for thecomputer drawing of the SDS micelle (front cover) and to Gale Leach atWordsWorth, Pacifica, CA, for the desktop publishing.

Herb Schwartz (technical editor)Palomar Analytical Services

Redwood City, CA

v


The following acronyms and symbols are used in this book.

α separation factorCD cyclodextrinCDEKC cyclodextrin electrokinetic chromatographyCDMEKC cyclodextrin-modified micellar electrokinetic chromatogra-phyCE capillary electrophoresisCMC critical micelle concentrationCsf surfactant concentrationCTAB cetyltrimethylammonium bromideCZE capillary zone electrophoresisDTAB dodecyltrimethylammonium bromideE electric field strengthEKC electrokinetic chromatographyEOF electroosmotic flowHPLC high performance liquid chromatographyIXEKC ion exchange electrokinetic chromatographyK distribution coefficientKp Kraft pointk’ capacity factorLMT N-lauroyl-N-methyltaurateMECC micellar electrokinetic capillary chromatographyMEEKC microemulsion EKCMEKC micellar electrokinetic chromatographyµeo electroosmotic mobilityµep electrophoretic mobilityµep(mc) electrophoretic mobility of the micelleµep* effective electrophoretic mobilityN number of theoretical platesn aggregation numbernaq number of moles of the analyte incorporated into the aqueous

phase

vi

v

nmc number of moles of the analyte incorporated into the micelleRPLC reversed phase liquid chromatographySDS sodium dodecyl sulfateSDVal sodium N-dodecanoyl-L-valinateSTS sodium tetradecyl sulfatet0 migration time of the bulk solutiontmc migration time of the micelletR migration time of the analyteUV ultravioletVaq volume of the nonmicellar phaseVmc volume of the micellar phasev migration velocity

partial specific volume of the micelle

1

I: Introduction

Electrokinetic chromatography (EKC) is a family of electrophoresis tech-niques named after electrokinetic phenomena, which include electroosmo-sis, electrophoresis, and chromatography. Micellar electrokineticchromatography (MEKC) is a mode of EKC in which surfactants (mi-celles) are added to the buffer solution. Surfactants are molecules whichexhibit both hydrophobic and hydrophilic character. They have polar“head” groups that can be cationic, anionic, neutral, or zwitterionic andthey have nonpolar, hydrocarbon tails. The formation of micelles or“micellization” is a direct consequence of the “hydrophobic effect.” Thesurfactant molecules can self-aggregate if the surfactant concentrationexceeds a certain critical micelle concentration (CMC). The hydrocarbontails will then be oriented toward the center of the aggregated molecules,whereas the polar head groups point outward. Micellar solutions may solu-bilize hydrophobic compounds which otherwise would be insoluble inwater. The front cover picture shows an aggregated SDS molecule. In thecenter of the aggregate, p-fluorotoluene is situated depicting the partition-ing of a neutral, hydrophobic solute into the micelle. Every surfactant has acharacteristic CMC and aggregation number, i.e., the number of surfactantmolecules making up a micelle (typically in the range of 50-100). (See alsoTable 1 and the discussion on page 10). The size of the micelles is in therange of 3 to 6 nm in diameter; therefore, micellar solutions exhibit proper-ties of homogeneous solutions. Micellar solutions have been employed in avariety of separation and spectroscopic techniques. In 1980, Armstrongand Henry pioneered the use of micellar solutions as mobile phases forreversed-phased liquid chromatography (RPLC).

In the literature, MEKC is also often referred to as MECC (micellarelectrokinetic capillary chromatography) since the separations are mostoften performed in a capillary tube. Other modes of EKC are cyclodextrinEKC (CDEKC), ion-exchange EKC (IXEKC), and microemulsion EKC(MEEKC). Cyclodextrin derivatives, polymer ions, and microemulsionsare used in CDEKC, IXEKC, and MEEKC, respectively, instead of themicelles used in MEKC. The references listed on page 3 provide furtherdetail on the differences between the various kinds of EKC techniques. Inthe following chapters, relevant references are listed in reverse chronologi-cal order after each chapter. All EKC techniques are based on the same

2

separation principle: the differential partitioning of an analyte between atwo-phase system (i.e., a mobile/aqueous phase and a stationary phase).

The same instrument that is used for capillary zone electrophoresis(CZE) is also used for MEKC. Both MEKC and CZE are modes of capil-lary electrophoresis (CE), as are capillary gel electrophoresis, capillaryisoelectric focusing, and capillary isotachophoresis (for an introduction toCE, see the Beckman Primer Introduction to Capillary Electrophoresis,part number 360643). MEKC is different in that it uses an ionic micellarsolution instead of the simple buffer salt solution used in CZE. The micel-lar solution generally has a higher conductivity and hence causes a highercurrent than the simple buffer does in CZE. MEKC can separate both ionicand neutral substances while CZE typically separates only ionic sub-stances. Thus MEKC has a great advantage over CZE for the separation ofmixtures containing both ionic and neutral compounds. However, inMEKC the size of the sample molecules is limited to molecular weights ofless than 5000, whereas CZE has virtually no limitation in molecular size.The separation principle of MEKC is based on the differential partition ofthe solute between the micelle and water; CZE is based on the differentialelectrophoretic mobility.

3


Janini, G. M., Isaaq, H. J. Micellar electrokinetic capillary chromatogra-phy: basic considerations and current trends. J. Liq. Chromatogr. 15, 927-960 (1992)

Kuhr, W. G., Monnig, C. A. Capillary Electrophoresis. Anal. Chem. 64,389R-407R (1992)

Watarai, H. Microemulsion capillary electrophoresis. Chem. Lett., 391-394(1991)

Nishi, H., Terabe, S. Application of micellar electrokinetic chromatogra-phy to pharmaceutical analysis. Electrophoresis 11, 691-701 (1990)

Terabe, S., Isemura, T. Ion-exchange electrokinetic chromatography withpolymer ions for the separation of isomeric ions having identical electro-phoretic mobilities. Anal. Chem. 62, 650-652 (1990)

Terabe, S. Electrokinetic chromatography: an interface between electro-phoresis and chromatography. Trends Anal. Chem. 8, 129-134 (1989)

Khaledi, M. G. Micellar reversed phase liquid chromatography.Biochromatography 3, 20-35 (1988)

Burton, D. E., Sepaniak, M. J. Analysis of B6 vitamers by micellar electro-kinetic capillary chromatography with laser-excited fluorescence detection.J. Chromatogr. Sci. 24, 347-351 (1986)

Terabe, S., Ozaki, H., Otsuka, K., Ando, T. Electrokinetic chromatographywith 2-O-carboxymethyl-β-cyclodextrin as a moving "stationary" phase.J. Chromatogr. 332, 211-217 (1985)

Terabe, S., Otsuka, K., Ando, T. Electrokinetic chromatography with mi-cellar solution and open-tubular capillary. Anal. Chem. 57, 834-841 (1985)

Terabe, S., Otsuka, K., Ichikawa, K., Tsuchiya, A., Ando, T. Electrokineticseparations with micellar solution and open-tubular capillaries. Anal Chem.56, 111-113 (1984)

Armstrong, D. W., Henry, S. J. Use of an aqueous micellar mobile phasefor separation of phenols and polynuclear aromatic hydrocarbons viaHPLC. J. Liq. Chromatogr. 3, 657-662 (1980)

4

II: Separation Principle/Fundamentals

Figure 1 shows a schematic representation of the separation principle ofMEKC. When an anionic surfactant such as sodium dodecyl sulfate (SDS)is employed, the micelle migrates toward the positive electrode by electro-phoresis. The electroosmotic flow transports the bulk solution toward thenegative electrode due to the negative charge on the surface of fused silica.The electroosmotic flow (EOF) is usually stronger than the electrophoreticmigration of the micelle under neutral or alkaline conditions and, therefore,the anionic micelle also travels toward the negative electrode at a retardedvelocity.


= Solute


= Electrophoresis

Figure 1. Schematic of the separation principle of MEKC. The detectorwindow is assumed to be positioned near the negative electrode.

When a neutral analyte is injected into the micellar solution, a fractionof it is incorporated into the micelle and it migrates at the velocity of themicelle. The remaining fraction of the analyte remains free from the mi-celle and migrates at the electroosmotic velocity. The migration velocity ofthe analyte thus depends on the distribution coefficient between the micel-lar and the non-micellar (aqueous) phase. The greater the percentage ofanalyte that is distributed into the micelle, the slower it migrates. Theanalyte must migrate at a velocity between the electroosmotic velocity andthe velocity of the micelle (see Figure 2A), provided the analyte is electri-cally neutral. In other words, the migration time of the analyte, tR, is lim-ited between the migration time of the bulk solution, t0, and that of the

5

micelle, tmc (see Figure 2B). This is often referred to in the literature as themigration time window in MEKC.

Water Solute Micelle

Micelle Solute Water

inj. column det.

Time0 t0 tR tmc

(A)

(B)

Figure 2. Schematic of the zone separation in MEKC (A) and chromato-gram (B). Reproduced with permission from Terabe, et al., Anal. Chem.57, 834 (1985).

Capacity Factor

We can define the capacity factor, k', similarly to that of chromatographyas

k' = nmc

naq(1)

where nmc and naq are the amount of the analyte incorporated into the mi-celle and that in the aqueous phase, respectively. We can obtain the rela-tionship between the capacity factor and the migration times as

tR = 1 + k'1+ (t0 / tmc )k'

t0 (2)

The migration time of the analyte is equal to t0 when k' = 0, or whenthe analyte does not interact with the micelle at all; the migration timebecomes tmc when k' is infinity or the analyte is totally incorporated intothe micelle. Thus, the migration time window is limited between t0 andtmc.

6

When t0 is infinity (electroosmosis is completely suppressed), equa-tion (2) becomes

tR = (1 + 1 / k' )tmc (3)

In this case, the bulk solution remains stationary in the capillary andthe micelle migrates only by electrophoresis. If we define the capacityfactor as the reciprocal of equation (1), equation (3) becomes identical withthe relationship between tR, t0, and k' in conventional chromatography.

Figure 3 shows a typical example of MEKC separation. Eight electri-cally neutral compounds were successfully resolved in 17 min. The capac-ity factor scale is inserted in the figure to indicate the relationship betweenthe migration time and the capacity factor. The capacity factor of infinitymeans that analyte has the same migration time as the micelle. Theoreticalplate numbers calculated from the peak widths range from 200,000 to250,000 which is typical for MEKC separations.

0.004 AU

0 1 2 6 10 20 50∞

0 5 10 15Time (min)

Capacity Factor

1

2

3

4

5

6

7

8

Figure 3. Micellar electrokinetic chromatogram of a test mixture:1 = methanol; 2 = resorcinol; 3 = phenol; 4 =p-nitroaniline;5 = nitrobenzene; 6 = toluene; 7 = 2-naphthol; 8 = Sudan III. Conditions:capillary, 50µm i.d. × 65 cm (effective length 50 cm); run buffer, 30 mMSDS in 50 mM phosphate/100 mM borate (pH 7.0); applied voltage, 15 kV;current, 33µA; detection, UV absorbance at 210 nm; temperature, 35°C.Reproduced with permission from Terabe, Trends Anal. Chem. 8, 129 (1989).

7

Effective Mobility

The capacity factor is a fundamental term in chromatography and equation(2) is derived from a chromatographic perspective. We can derive a similarequation for electrophoretic processes. In CZE, the migration velocity ofthe analyte, vs, is expressed as

vs = µeo + µep (s)[ ]E (4)

where µeo and µep(s) are the electroosmotic mobility and electrophoreticmobility of the analyte, respectively, and E is the electrical field strength.We can apply this equation to MEKC by defining the effective electro-phoretic mobility, µep*(s), for the neutral analyte as

µep * (s)k'

1 + k'µep (mc) (5)

where µep(mc) is the electrophoretic mobility of the micelle and k'/(1+k') isthe fraction of the analyte incorporated into the micelle, as shown by thefollowing equation:

k'1 + k'

= nmc

naq + nmc(6)

Thus, the velocity of the neutral analyte in MEKC is given as

vs = µeo + µep * (s)[ ]E (7)

The effective mobility tells us that even a neutral analyte has an appar-ent mobility. We can use the effective mobility in a similar way as theelectrophoretic mobility is used in CZE to derive a resolution equationwhich can assist in optimizing separations (see also the discussion regard-ing equation 8). The effective mobility can simply be calculated accordingto equation 7, provided veo (= µeoE) or µeo is known. However, in MEKCthe calculation of the capacity factor requires both t0 andtmc. This is theadvantage of using the effective mobility as compared to using the capacityfactor. It should be emphasized that the capacity factor provides quantita-tive information about the distribution equilibrium, whereas the effectivemobility only gives qualitative information about it. In other words, thecapacity factor cannot be obtained according to equation 5 unless µep(mc)is known.

8

Operating Conditions

Capillary: 25–75 µm i.d. × 20–75 cm lengthRun buffer: A solution of an ionic micelle in a buffer solution.

The surfactant concentration must be higher thanits critical micelle concentration (CMC).

Applied voltage: 10 to 25 kVCurrent: Below 75 µA, preferably below 50 µA

Since the micellar solution has a relatively high conductivity, a capil-lary with a small diameter is favored to prevent excessive Joule heating.The length of the capillary usually is not very important, but a longer onecan accommodate a larger amount of sample solution at the expense oftime. When using relatively large i.d. capillaries for increased sample ca-pacity, efficient capillary cooling (such as the liquid cooling system used inthe P/ACE™) becomes important.

The micellar solution is prepared by dissolving a surfactant into abuffer solution at a concentration higher than its CMC (Table 1). Thebuffer solution is required to keep the pH constant. Concentrations of 30 to100 mM are typically employed for both the surfactant and buffer compo-nents. The separation solution must be filtered through a membrane filterto remove particulates. A disposable, cartridge-type membrane filter can beused with a syringe because the amount of the solution necessary for anMEKC run is usually less than 10 mL. Degassing of the micellar solutionis troublesome because of bubbling and is typically unnecessary. When acationic surfactant (e.g., CTAB) is employed at a high enough concentra-tion, the direction of the electroosmotic flow is reversed. In this case, thepolarity of the power supply must be reversed (see page 16).

The applied voltage must be kept to a level that is not too high in orderto avoid excessive current. It is also desirable to control the capillary tem-perature because the migration time in MEKC is even more sensitive totemperature than in CZE (S. Terabe, J. Microcol. Sep., (1992) in prepara-tion).

9

Table 1. Critical Micelle Concentration, Aggregation Number (n),and Kraft point ( Kp) of Selected Ionic Surfactants

Surfactant CMCa/10-3 M n Kp

Sodium dodecyl sulfate (SDS) 8.1 62 16Sodium tetradecyl sulfate (STS) 2.1 (50°C) 138b 32Sodium decanesulfonate 40 40 —Sodium dodecanesulfate 7.2 54 37.5Sodium N-lauroyl-N-methyltaurate (LMT) 8.7 — <0Sodium polyoxyethylene dodecyl ether sulfate 2.8 66 —Sodium N-dodecanoyl-L-valinate (SDVal) 5.7 (40°C) — —Sodium cholate 13–15 2–4 —Sodium deoxycholate 4–6 4–10 —Sodium taurocholate 10–15 5 —Sodium taurodeoxycholate 2–6 — —Potassium perfluoroheptanoate 28 — 25.6Dodecyltrimethylammonium chloride (DTAC) 16 (30°C) — —Dodecyltrimethylammonium bromide (DTAB) 15 56 —Tetradecyltrimethylammonium bromide (TTAB) 3.5 75 —Cetyltrimethylammonium bromide (CTAB) 0.92 61 —

a 25°Cb In 0.10 M NaCl

10

Composition of the Micellar Solution

Ionic surfactants are essential for MEKC. Numerous ionic surfactants arecommercially available. The surfactants suitable for MEKC should meetthe following criteria:

1. The surfactants must have enough solubility in the buffersolution to form micelles.

2. The micellar solution must be homogeneous and UVtransparent.

3. The micellar solution must have a low viscosity.

Table 1 lists the CMC, aggregation number, and Kraft point of someselected ionic surfactants available for MEKC. The Kraft point is the tem-perature above which the solubility of the surfactant increases steeply dueto the formation of micelles. In order to obtain a micellar solution, theconcentration of the surfactant must be higher than its CMC. The surfac-tant has enough solubility to form micelles only at temperatures above theKraft point asmentioned above. The counter ion of the ionic surfactant does affect theKraft point. For example, the Kraft point of sodium dodecyl sulfate (SDS)is 16°C but potassium dodecyl sulfate has a Kraft point of approximately35°C. Therefore, if SDS is dissolved in a buffer containing potassium ions,the solubility of SDS will be less than its CMC at ambient temperaturebecause of the exchange reaction of the counter ions. The actual CMC inthe buffer solution is usually lower than the value listed in Table 1, as thesevalues were obtained with pure water as solvent.

High concentrations of surfactant (>200 mM) result in relatively highviscosities (and high currents) and should therefore be avoided in MEKC.It is recommended that the concentration of the buffer salt should not belower than 10 mM. A higher concentration of buffer relative to that ofsurfactant is preferred to keep the pH constant during the run. It should beremembered that, in electrophoresis, electrolysis occurs at the electrodes,resulting in a reduction of the pH of the anodic solution. This solution mayenter the capillary during the run when an anionic micelle (e.g., SDS) isemployed. The cathodic solution becomes alkaline during electrophoresisand may enter the capillary when a cationic surfactant (e.g., CTAB) isemployed. In this case the EOF is reversed.

11

Factors Affecting Reproducibility

In our experience, MEKC usually yields better reproducibility in migrationtimes and peak areas than does CZE. Occasionally, however, the reproduc-ibility is poor, probably due to a contaminated capillary surface. In suchcases, cleaning of the capillary is necessary. It is recommended to use astrong cleaning solution, e.g., a product used for cleaning lab glassware.Organic solvents such as methanol, acetone, acetonitrile, or tetrahydrofu-ran are also often effective.

In a clinical application relevant to therapeutic drug monitoring,Nakagawa et al. (1989) found that cleaning of the capillary between runswith 0.1 M NaOH was essential to obtain good performance and reproduc-ibility. In this study, an antibiotic, cefpiramide, was determined in plasmaby means of a direct injection method in an SDS-containing buffer. That is,no sample pretreatment such as deproteination or extraction was necessary.Without SDS (i.e., in the CZE mode), plasma protein peaks interfered withthe peaks of interest (see Figure 4A). With SDS, the migration times of theprotein peaks had shifted relative to cefpiramide and the internal standard(antipyrine), thus enabling a quantitative plasma assay (see Figure 4B).The plasma proteins solubilized by the SDS have a strong negative chargeresulting in a slow net migration whereas the negatively chargedcefpiramide and antipyrine will not interact with the SDS. Hence, the mi-gration times of the drugs are not affected by the presence of SDS in thebuffer as indeed is observed from Figure 4. Recoveries were found to benear 100%, indicating that the method could be used to determine the totalconcentration of the drug in plasma (i.e., the SDS enabled rapid release ofprotein-bound drug). Furthermore, the assay is fast (≈ 6 min run time) as itis not necessary to wait for the elution of the plasma proteins from thecapillary. A rinse with 0.1 M NaOH can be started as soon as the antibioticpeak has been detected. As pointed out above, rinsing of the capillary alsoaids in generating reproducible surface conditions, which benefits preci-sion and accuracy. A similar direct sample injection assay for the penicillinantibiotic aspoxicillin in human plasma was developed by Nishi et al.(1990). Both assays showed good linearity and covered the plasma levelstypically encountered in clinical therapy.

12

1 2

0 4 8 0 4 8 12

A

Time (min)

B1 2

Figure 4. (A) Separation of cefpiramide (peak 2) and antipyrine (peak 1)in human plasma with a 50 mM phosphate, pH 8.0 run buffer; (B) with therun buffer of A and 10 mM SDS. Detection wavelength, 280 nm. Repro-duced with permission from Nakagawa et al., Chem. Pharm. Bull. 36, 1622(1988).

Marker of the Electroosmotic Flow and the Micelle

In order to calculate the capacity factor according to equation (2), it isnecessary to know the migration time of the bulk solution, t0, the migrationtime of the micelle, tmc, as well as the migration time of the analyte, tR.Since the whole capillary is filled with micellar solution, the markers of thebulk solution and the micelle are required to obtain t0 and tmc. Strictlyspeaking, no ideal marker is available for MEKC.

The marker for the bulk solution must be electrically neutral as well astotally excluded from the micelle. Mesityl oxide, often used in CZE tomeasure t0, is not an appropriate choice in MEKC, because it is partiallyincorporated into the micelle. Methanol often serves to measure t0, becauseits distribution coefficient is almost negligible. Furthermore, it can be de-tected by UV absorption due to a change in refractive index as the metha-nol peak passes through the detection zone.

The marker for the micelle must be totally incorporated into the mi-celle. Sudan III or IV are often used to measuretmc. Both solutes are notsoluble in water and can be dissolved in methanol or in the micellar solu-tion. However, because of the poor solubility in water, it is not alwayspossible to observe the peaks of Sudan III or IV in the electropherograms.As an alternative, compounds that are insoluble in water but soluble in themicellar solution can be employed to measuretmc. Timepidium bromide orquinine hydrochloride are good markers for anionic SDS micellar systems.

13

Resolution

The resolution equation for MEKC is given as

Rs = N

4α − 1

α

k2'

1 + k2'

1 − t0 / tmc

1+ t0 / tmc( )k1'

(8)

where N is the theoretical plate number, α the separation factor equal tok2'/k1', and k1' and k2' are capacity factors of analytes 1 and 2, respectively.The equation predicts the effect of N, α, k', and t0/tmc on resolution. Eacheffect is briefly discussed below.

Plate NumberResolution increases in proportion to the square root of the plate number.The higher the applied voltage, the higher the plate number, unless condi-tions are such that the applied voltage generates too much Joule heating.Average plate numbers for most analytes are usually in the range of100,000 to 200,000. If the plate number is considerably lower, analytes arelikely to be adsorbed on the capillary wall. In such cases, experimentalconditions must be optimized to produce more efficient separations. Clean-ing of the capillary is a possible procedure, as is changing the pH of the runbuffer. Hydrophobic analytes, or those having longer migration times,typically yield high theoretical plate numbers because the micelle has asmaller diffusion coefficient. The plate number does not depend signifi-cantly on the capillary length. With short capillaries, however, the amountof sample volume injected must be minimized to avoid zone broadening.

Separation FactorThe separation factor, α, is the most important and most effective term tomaximize resolution. The separation factor reveals the relative differenceof the distribution coefficient between the two analytes and can be manipu-lated by chemical means. Since the distribution coefficient is a characteris-tic of a given separation system consisting of a micellar and an aqueousphase, we can manipulate the separation factor by changing either the typeof micelle or by modifying the aqueous phase. Various factors affecting theselectivity are discussed later. Generally in MEKC it is not very difficult toseparate a pair of analytes with α =1.02.

14

Capacity FactorIt can be calculated that the optimum value of the capacity factor is equalto (tmc/t0)1/2. Under conditions of pH above 6, the optimum k' value isclose to 2 for most long alkyl chain surfactants. Under most conditions, thecapacity factors must be adjusted to be between 0.5 and 10. A large capac-ity factor means that the major fraction of the analyte is incorporated intothe micelle. It is necessary for the analyte to be distributed evenly betweenthe micellar and the aqueous phase, i.e., the analyte must not spend most ofits time in one phase.

The capacity factor is related to the distribution coefficient, K, by

k' = KVmc

Vaq(9)

where Vmc/Vaq is the phase ratio and Vmc and Vaq are volumes of the mi-celle and the remaining aqueous phase. The capacity factor is approxi-mately related to the surfactant concentration, Csf, by

k' = Kv (Csf - CMC) (10)

wherev is the partial specific volume of the micelle. Equation 10 indicatesthat the capacity factor increases linearly with an increase of the surfactantconcentration. It should be noted that the capacity factor is not linearlycorrelated to the migration time (see axes in Figure 3). Equation 10 sug-gests that we can easily vary the capacity factor by adjusting the surfactantconcentration, provided the CMC is known. As mentioned above, the sur-factant concentration should preferably be in the range of 20 to 200 mM toavoid excess currents.

Electroosmotic VelocityThe effect of the electroosmotic flow velocity on resolution can be dis-cussed in terms of the migration time ratio, t0/tmc, which can be expressedas

t0 / tmc = [1 + µep (mc) / µeo]E (11)

where E is the electrical field strength. The mobilities µeo and µep(mc)usually have different signs and the ratio µep(mc)/µeo is smaller than zeroand larger than minus one. Therefore, t0/tmc is less than one. The t0/tmc isalso directly related to the width of the migration time window. Thesmaller the value of t0/tmc, the wider the migration time window, hence the

15

higher resolution. A longer run time is required, however. The value of themigration time ratio t0/tmc is in the range of 0.2 to 0.3 for most ionic mi-celles under the conditions of pH above 6.

In order to reduce the value of t0/tmc, it is necessary to reduce µeo rela-tive to µep(mc) because, in practice, increasing µep(mc) is rarely possible.The addition of an organic solvent, e.g., methanol or 2-propanol (<20%) isa possible way to reduce the electroosmotic flow velocity. It is also pos-sible to reduce veo by changing the pH of the buffer to acidic conditions.Additives such as methylcellulose derivatives or ethylene glycol are oftenused in CE to increase the viscosity of the run buffer. Although an increasein viscosity of the solution reduces veo, it also reduces vep(mc) and, there-fore, it does not help to increase resolution as predicted from equation (8).

Use of Coated Capillaries to Control the EOFCoated capillaries can be used in CZE and MEKC to reduce the zeta poten-tial at the capillary wall and, consequently, the electroosmotic flow. Cur-rently, coated capillaries for CE are commercially available from severalmanufacturers (e.g., J & W Scientific, Supelco, Scientific Glass Engineer-ing, Chrompack). In MEKC, untreated fused silica capillaries have mainlybeen used. An example of the utility of coated capillaries is presented inFigure 5. A mixture of nucleobases was separated under identical condi-tions on a nonpolar, polymethylsiloxane (OV-1) capillary (panel A) and ona polar, polyethylene glycol (CW20M) coated capillary (panel C). Forcomparison, panel B ofFigure 5 shows the separation on an untreated fused silica capillary. It canbe seen that with the polymethylsiloxane capillary actually an increase inthe electroosmotic flow is obtained with a corresponding decrease in reso-lution. Presumably this is because of an increase in negative charge densityat the capillary wall as the micellar reagent, SDS, binds to the nonpolarpolysiloxane network. This is not the case with a relatively polar coating.The polyethylene glycol capillary shows a increase in migration times(decrease in the electroosmotic flow) and superior resolution. For example,the resolution for the cytidine/guanosine pair is 3.9 with thepolymethylsiloxane capillary, 6.9 with the untreated capillary, and 9.7 withthe polyethylene glycol capillary. Another way to manipulate the electroos-motic flow is to use cationic surfactants as will be discussed in the nextsection.

16

A

1

23

45

B2

4

5

3

C

0

1

2

3

4

5

5 10 15

Figure 5. MEKC separations of nucleobases using capillaries coated withdifferent polymers: 1 = water; 2 = uridine; 3 = cytidine; 4 = guanosine;5 = adenosine. (A) polymethylsiloxane OV-1 coating; (B) uncoated;(C) polyethylene glycol CW20M coating. Run buffer, 0.02 M phosphate,0.05 M SDS. Reproduced with permission from Lux et al., J. High Resolut.Chromatogr. HRC 13, 145 (1990).

EOF ReversalCationic surfactants such as cetyl-, dodecyl-, and hexadecyltrimethylam-monium salts can be used in MEKC to reverse the charge on the capillarywall. These surfactants are absorbed on the capillary wall surface by amechanism involving electrostatic attraction between the positivelycharged ammonium moieties and the negatively charged Si-O-groups. The

17

non-polar chains (C10, C14, C16, etc.) create a hydrophobic layer and, at ahigh enough surfactant concentration, the negative surface charge will becompletely neutralized. At even higher surfactant concentrations (e.g., 0.35mM for CTAB), a bilayer is formed through hydrophobic interaction be-tween the nonpolar chains. Schematically, this situation is depicted inFigure 6. In this case, the cationic head groups are facing the buffer solu-tion and the charge at the capillary wall is reversed from negative to posi-tive. Consequently, under the influence of an electric field, a reversal of theEOF takes place.

+

++ +

+

+

+

+++

+

+

+

+

+

+-

--

-

-

+

(–)

(+)

Vep1

Vep2

EOF

+ + + + + + + + +

++++++++ +

Figure 6. Separation mechanism with EOF reversal. A bilayer is formed atthe capillary wall. The EOF is directed toward the anode. Proteins with anet positive charge will electrophoretically migrate (Vep1) toward the cath-ode. Net migration is toward the anode if EOF > Vep1. Negatively chargedproteins migrate toward the anode by both EOF and electrophoretic flow(Vep2).Positively charged micelles are formed and may interact with proteins.Reproduced with permission from Emmer et al., J. High Resolut.Chromatogr. HRC 14, 738 (1991).

Figure 7 shows examples of the charge reversal mode of MEKC. Opiatesand ephedrine are separated with high efficiency and relatively short analy-sis times, made possible by operating at high pH conditions. It should benoted that with charge reversal MEKC, the polarity of the power supplymust be reversed since the bulk flow is now towards the positive electrode,i.e., the bulk flow must always go in the direction of the detector. Whereasthe majority of applications of charge reversal MEKC have involved smallmolecules (e.g., ions, pharmaceuticals, and drugs), applications of biologi-cally important molecules such as peptides and proteins have also beendemonstrated (Liu et al., 1990; Emmer et al., 1991). For example, in the

18

separation of basic proteins, Emmer et al. (1991) suggested a fluorocarbonsurfactant, FC-134 (commercially available from 3M Company, St. Paul,MI). This surfactant has a non-sticking, Teflon*-like chain, thus minimiz-ing interactions with many proteins. In addition, proteins at a pH belowtheir pI will be repelled from the capillary surface. As shown in Figure 6,the electrophoretic migration of a positively and a negatively charged pro-tein is in opposite direction but in both cases, the net migration is in theanodic direction (assuming that the EOF is greater than electrophoreticmigration). Interactions between the micelles and the protein may enhancethe above described repellant effect.

O

HO

HOcodeine

H

NCH3

O

H3CO

HOcodeine

H

NCH3

O

O

HO

pholcodine

NCH3

H3CO

dextromethorphan

NCH3

N

O

0.150

-0.0203.0 10.0

Time (min)

Abs

orba

nce

(230

nm

)

2

1

3

4

5

Figure 7. Separation of ephedrine and opiates with EOF reversal.1 = ephedrine; 2 = pholcodine; 3 = codeine; 4 = morphine;5 = dextromethorphan. Run buffer, 50 mM CTAB, 50 mM triethylamine,pH 12.0 in 20% acetonitrile. Capillary, 75 µm i.d. × 57 cm (50 cm to de-tector). Reproduced from Kerr and Jung, Beckman Application Note DS-783, 1990.

* A registered trademark of E. I. du Pont de Nemours & Company.

19


Emmer, A., Jansson, M., Roeraarde, J. Improved CZE separations of basicproteins using a fluorosurfactant buffer additive. J. Chromatogr. 547, 544-550 (1991)

Gowsi, K., Foley, J. P., Gale, R. J. Micellar electrokinetic capillary chro-matography theory based on electrokinetic parameters: optimization forthree modes of operation. Anal. Chem. 62, 2714-2721 (1990)

Liu, J., Cobb, K. A., Novotny, M. V. Capillary electrophoretic separationsof peptides using micelle-forming compounds and cyclodextrins as addi-tives. J. Chromatogr. 519, 189-197 (1990)

Lux, J. A., Yin, H., Schomburg, G. Influence of polymer coating of capil-lary surfaces on migration behavior in micellar electrokinetic capillarychromatography. J. High Resolut. Chromatogr. 13, 145-147 (1990)

Nishi, H., Fukuyama, T., Matsuo, M. Separation and determination ofaspoxicillin in human plasma by micellar electrokinetic chromatographywith direct sample injection. J. Chromatogr. 515, 245-255 (1990)

Nakagawa, T., Oda, Y., Shibukawa, A., Fukuda, H., Tanaka, H. Electroki-netic chromatography for drug analysis. Separation and determination ofcefpiramide in human plasma. Chem. Pharm. Bull. 37, 707-711 (1989)

Otsuka, K., Terabe, S. Effects of pH on electrokinetic velocities in micellarelectrokinetic chromatography. J. Microcol. Sep. 1, 150-154 (1989)

Terabe, S., Otsuka, K., Ando, T. Band broadening in electrokinetic chro-matography with micellar solutions and open-tubular capillaries. Anal.Chem. 61, 251-260 (1989)

Hinze, W. L. Organized surfactant assemblies in separation science, in"Ordered media in chemical separations," Hinze, W. L., Armstrong, D. W.,eds. ACS Symposium Series 342, American Chemical Society, WashingtonD.C., pp. 2-82 (1987)

Terabe, S., Utsumi, H., Otsuka, K., Ando, T., Inomata, T., Kuze, S.,Hanaoka, Y. Factors controlling electroosmotic flow in open-tubular capil-laries in electrokinetic chromatography. HRC CC, J. High Resolut.Chromatogr. Chromatogr. Commun. 9, 666-670 (1986)

Armstrong, D. W. Micelles in separations: a practical and theoretical re-view. Separation and Purification Methods 14, 213-304 (1985)

20

III: Selectivity Manipulation

From the viewpoint of selectivity manipulation, the micellar phase inMEKC corresponds to the stationary phase in reversed-phase liquid chro-matography (RPLC) as does the surrounding aqueous phase to the mobilephase in RPLC. There are five factors we can control to manipulate selec-tivity:

1. Temperature2. Choice of surfactant3. Modification of the micelle4. Choice of the aqueous phase5. Modification of the aqueous phase

Each variable is discussed below.

1. TemperatureThe distribution coefficient is significantly dependent on temperature. Anincrease in temperature causes a reduction in migration time because of thedecrease in the distribution coefficient as well as the viscosity. Since the de-pendence of the distribution coefficient on temperature is different amonganalytes, temperature will also affect selectivity. The temperature effect onselectivity is not dramatic but temperature does, however, affect the migra-tion time. For reproducible results, it is essential to keep the run tempera-ture constant. It should be noted that even if the temperature of the capillaryis carefully controlled, a high current will cause a substantial rise in thetemperature of the solution inside the capillary. This is because heat dissi-pation under these conditions is not complete and instantaneous. Run condi-tions which result in high currents should therefore be avoided. A CE sys-tem which has efficient capillary cooling (such as that used in P/ACE) ismandatory under high current conditions.

2. Choice of SurfactantA surfactant molecule has a hydrophobic part and hydrophilic part andboth groups affect the selectivity in MEKC. Since most analytes interactwith the micelles at their surfaces, the hydrophilic group, or ionic group, isgenerally more important in terms of selectivity. That is, SDS and sodiumtetradecyl sulfate (STS) show very similar selectivity, but SDS and sodiumN-lauroyl-N-methyltaurate (LMT) yield considerably different selectivity,provided the analytes are polar (see Figure 8 for example). In practice, it iseasy to change the micellar solution. When the resolution is not adequate

21

with a particular surfactant solution, it is recommended to try another sur-factant solution.

101211

1

2 6

3

(A)

54

78

9

0 5 10 15 20 25 30Time (min)

(B)

10 11

12+

94

785

2 31

6

0 5 10 15 20 25

Time (min)

Figure 8. Comparison of selectivity between surfactants with differentpolar groups: 1 = caffeine; 2 = acetaminophen; 3 = sulpyrin;4 = trimetoquinol; 5 = guaifenesin; 6 = naproxen; 7 = ethenzamide;8 = phenacetin; 9 = isopropylantipyrine; 10 = noscapine;11 = chlorpheniramine; 12 = tipepidine. Conditions: Run buffer, (A) 0.1 MLMT in 20 mM phosphate-borate (pH 9.0); (B) 0.1 M SDS in the samebuffer as in A; applied voltage, 20 kV. Reproduced with permission fromNishi et al., J. Pharm. Sci. 79, 519 (1990).

Bile salts such as sodium cholate, sodium deoxycholate, and theirtaurine conjugates are natural surfactants. Bile salts form helical micellesand yield selectivities significantly different from the long alkyl chainsurfactants. In particular, bile salts are useful for the separation of very

22

hydrophobic compounds. In Figure 9, for example, corticosteroids werenot successfully separated with SDS but they were completely resolvedwith sodium cholate as the micelle.

1050 15 20

Time (min)

b

a

c

d

ef

g

h

Figure 9. Separation of eight corticosteroids: a = hydrocortisone;b = triamcinolone; c = betamethazone; d = hydrocortisone acetate;e = dexamethasone acetate; f = fluocinolone; g = fluocinolone acetonide;h = fluocinonide. Conditions: Run buffer, 100 mM sodium cholate in20 mM phosphate-borate (pH 9.0). Reproduced with permission from Nishiet al., J. Chromatogr. 513, 279 (1990).

Although MEKC is used mainly for separating neutral compounds, itsometimes also allows an improved separation of ionic analytes. MEKCmay work for the separation of ionic analytes when they are not success-fully separated by CZE. Since the micelles used in MEKC are charged onthe surface, an analyte with the opposite charge of the micelle will stronglyinteract with the micelle through electrostatic forces and an analyte withthe same charge as the micelle will interact weakly, due to the electrostaticrepulsion. In case of ionic analytes, hydrophobicity and charge affect thedistribution coefficient. Therefore, the use of a cationic surfactant willresult in an entirely different selectivity than when an anionic surfactant isused for the separation of ionic analytes. This is illustrated in Figure 10 forthe separation of PTH amino acids.

Some classes of surfactants possess very specific selectivity. For ex-ample, sodium N-dodecanoyl-L-valinate (SDVal) and many bile salts en-able enantiomeric separations (see Section IV). Surfactants withperfluorinated alkyl chains may exhibit enhanced selectivity towards flu-

23

orinated compounds. It is expected that in the near future, surfactants spe-cifically designed for the separation of certain groups of analytes will becommercially available.

0.002 AU

0 10 20

(A)

T

S

N

G

Q

AE

D

Y

P

δ

T

H

I

V

M

n-V

L

W

n-L

F

K

R

T

(B)

0 10 20

Time (min)

S

H

GD

N•Q

R

A P

E

Mn-V

n-L

LV

Y

I

WF

K

δ

T

0.001 AU 0.002 AU

Figure 10. Comparison of selectivity between anionic and cationic surfac-tants for the separation of 22 phenylthiohydantoin (PTH)-amino acids byMEKC: The peaks are labeled with one-letter abbreviations for aminoacid. Conditions: (A) Run buffer, 50 mM SDS in 50 mM phosphate-borate(pH 7.0);(B) 50 mM DTAB in 100 mM Tris-HCI buffer (pH 7.0); applied voltage,(A) 10 kV, (B) 15 kV; detection, UV absorbance at 260 nm; temperature,35°C. Reproduced with permission from Otsuka et al., J. Chromatogr. 332,219 (1985).

24

3. Modification of the MicelleThe micellar phase can be modified by adding a second surfactant to forma mixed micelle or by selecting a different counter ion. Since a mixed mi-celle of an ionic and a nonionic surfactant has a lower surface charge and alarger size, its electrophoretic mobility will be lower than a single ionicmicelle. The addition of a nonionic surfactant to an ionic micellar solutioncauses a narrower migration time window. In Figure 11, the separation ofseveral pharmaceuticals is shown. As shown in panel B, the addition of thenonionic surfactant Tween 60 to the buffer resulted in several peak rever-sals and a shorter analysis time. Typically, the selectivity is dramaticallyaffected by adding a nonionic surfactant because the polar group of non-ionic and ionic surfactants are very different. The effect of the counter ionis generally less important, unless the counter ion is replaced by an organicion.

10 20 30 10 20 30

Time (min) Time (min)

(A) (B)

7

8

9

10

11

12

8

7

911

10

12

Figure 11. Effect of nonionic surfactant addition: 7 = acetaminophen;8 = caffeine; 9 = guaifenesin; 10 = ethenzamide;11 = isopropylantipyrine; 12 = trimetoquinol. Conditions: (A) Run buffer,100 mM SDS in 50 mM phosphate/100 mM borate (pH 7.0); capillary,75 µm i.d. × 57 cm (effective length 50 cm); applied voltage, 18 kV; detec-tion, UV absorbance at 214 nm. (B) as in (A) but with 30 mM Tween 60added to the run buffer. Unpublished data from Ishihama and Terabe(1991).

25

A relevant application of the mixed micellar phase is the separation ofenantiomers with a nonionic chiral surfactant. Several nonionic chiral sur-factants are effective for the separation of enantiomers. In order to givethese surfactants electrophoretic mobilities, an ionic surfactant such asSDS can be employed to form the mixed micelle. For example, digitonin issuccessfully used with SDS for the enantiomeric separation of PTH aminoacids (Figure 12). The method involving the mixed micelle can also beuseful when surfactants are used which become nonionic under acidicconditions.

20 30 40 90

Time (min)

1

2

3

4

5

6

Figure 12. Enantiomeric separation of six PTH-DL-amino acids byMEKC: 1 = Trp; 2 = Nle; 3 = Nva; 4 = Val; 5 = Aba; 6 = Ala. Condi-tions: Run buffer, 25 mM digitonin and 50 mM SDS in 50 mM phosphatebuffer (pH 3.0); capillary, 50µm i.d. × 63 cm (effective length 59 cm);applied voltage, 20 kV; detection, absorbance at 260 nm. Reproduced withpermission from Otsuka and Terabe, J. Chromatogr. 515, 221 (1990).

26

4. Choice of the Aqueous PhaseThe constituents of the buffer solution have very little effect upon selectiv-ity. Organic buffers usually have a relatively low conductivity and, there-fore, are recommended if they are stable and UV transparent. Care shouldbe taken not to replace the counter ion of the ionic surfactant with thebuffer ion.

The pH of the buffer is a critical parameter for the separation of ioniz-able analytes. In separations with closely spaced peaks, it is often essentialto find the optimum pH. This is demonstrated in Figure 13 for the separa-tion of chlorinated phenols by plotting the apparent capacity factor vs. thebuffer pH. Here, the apparent capacity factor is calculated according theequation (2), regardless of whether the samples are ionized or not. It shouldbe noted that changing the buffer pH (especially in the acidic region)causes a notable change in the electroosmotic velocity.

k app

'

1

5

10

50

100

6 7 8 9pH

1

2

20

7

14

17

Figure 13. Dependence of apparent capacity factors of chlorinatedphenols on pH: 1 = phenol; 2 = 2-chlorophenol; 7 = 2,5-dichlorophenol;14 = 2,4,5-trichlorophenol; 17 = 2,3,4,5-tetrachlorophenol; 20 = pen-tachlorophenol. Conditions: Run buffer, 100 mM SDS in 50 mM phos-phate-borate; capillary, 50µm i.d. × 65 cm (effective length, 50 cm);applied voltage, 15 kV; detection, UV absorbance at 220 nm. Reproducedwith permission from Otsuka et al., J. Chromatogr. 348, 39 (1985).

27

5. Modification of the Aqueous PhaseThe use of additives to modify the aqueous phase is very effective in ma-nipulating selectivity. Modification of the mobile phase by additives iswell documented in HPLC and many of the HPLC additives are also appli-cable to MEKC. However, we have to consider the difference between themicellar phase in MEKC and the stationary phase in HPLC to take advan-tage of such additives. Five classes of additives are applicable in MEKC:

1. Cyclodextrins (CDs)2. Ion-pair reagents3. Urea4. Organic solvents5. Metal salts

Cyclodextrins

CDs are oligosaccharides with truncated cylindrical molecular shapes.Their outside surfaces are hydrophilic, while their cavities are hydropho-bic. Traditionally, they have been used to improve the therapeutic useful-ness, in animals and humans, of drugs and biologicals that are relativelyinsoluble in water. CDs tend to include compounds which fit their cavitiesby hydrophobic interaction. Three CDs, α-, β-, and γ-CD, are widely usedand some of their characteristics are listed in Table 2. The size of the cav-ity differs significantly among the α, β and γ CDs. Only β-CD has a rela-tively low solubility in water. Many CD derivatives (e.g., O-methylatedβ-CD derivatives) have been developed for increased solubility in water aswell as to modify the cavity shape. CDs can be obtained from a variety ofcommercial sources. For example, in the U.S., Sigma (St. Louis, MO) andPharmatec (Alachua, FL) offer a wide choice in chemically modified CDs.

Table 2. Characteristics of Cyclodextrins

Cyclodextrin α-CD β-CD γ-CD

Number of glucose units 6 7 8Molecular weight 972.9 1135.0 1297.2Internal diameter of the cavity/nm 0.47–0.52 0.62–0.64 0.75–0.83Outside diameter/nm 1.46 1.54 1.75Height of the cavity/nm 0.79–0.80 0.79–0.80 0.79–0.80Solubility in water at 25°C 14.50 1.85 23.2

28

In MEKC, CDs are electrically neutral and have no electrophoreticmobility. They are assumed not to be incorporated into the micelle, be-cause of the hydrophilic nature of the outside surface of the molecules. Asurfactant molecule may, however, be included into the CD cavity. Theseparation principle of CD-modified MEKC (CDMEKC), in which CD isadded to the micellar solution, is shown schematically in Figure 14.

Anionic Surfactant

Analyte

Cyclodextrin

Electroosmotic Flow

ElectrophoreticMigration

Figure 14. Schematic of the separation principle of CDMEKC. The detec-tor window is assumed to be positioned near the negative electrode..

The analyte molecule included by CD migrates at the same velocity asthe electroosmotic flow because, electrophoretically, CD behaves as thebulk aqueous phase. Therefore, the addition of CD reduces the apparentdistribution coefficient and enables the separation of highly hydrophobicanalytes, which otherwise would be almost totally incorporated into themicelle in the absence of CD. The higher the concentration of CD, thesmaller the capacity factor. In CDMEKC, therefore, the capacity factor ischanged by varying the concentrations of both the surfactant and CD. Anexample of this approach for hydrophobic compounds is shown in Fig-ure 15. All of the isomers of the trichloro- biphenyls migrated with thesame velocity as that of the micelle in MEKC. However CDMEKC al-lowed the baseline separation of the eleven hydrophobic isomers.

29

0 10 20 30

Time (min)

BIPH

Figure 15. Separation of eleven trichlorobiphenol isomers by CDMEKC.BIPH = biphenyl. Run buffer, 60 mM γ-CD, 100 mM SDS, and 2 M urea in100 mM borate-50 mM phosphate (pH 8.0). Reproduced with permissionfrom Terabe et al., J. Chromatogr. 516, 23 (1990).

It should be mentioned that CDMEKC is a different technique fromCDEKC. In CDMEKC a neutral CD is added to the micellar solution,while in CDEKC an ionic CD derivative without the micelle is employed.CDMEKC is also very effective for enantiomeric separations because ofthe chirality of CD itself. This will be discussed later in Section IV onEnantiomeric Separations.

The solubility of β-CD is relative low. Addition of a high concentra-tion of urea such as 2 M increases the solubility. It is our experience that,for many compounds, γ-CD is a more effective additive than β-CD inCDMEKC. This is even the case for the compounds successfully separatedin HPLC with the β-CD-bonded phase. Conceivably, co-inclusion of thesurfactant molecule into the cavity of γ-CD would result in less availablespace for the analyte molecule. The use of a different CD typically causeschanges in selectivity.

Ion-Pair Reagents

There is an essential difference in the separation of ionic compounds be-tween MEKC and RPLC, although both have similar characteristics in theseparation of nonionic compounds. The difference is due to the electriccharge on the micelle. When a tetraalkylammonium salt is added to theSDS solution, migration times of anionic analytes increase with an in-

30

crease in the concentration of the ammonium salt, because the ammoniumion interacts with the anionic analyte to form the paired ion. Hence, theelectrostatic repulsion between the anionic SDS micelle and the anionicanalyte is reduced. In contrast, migration times of cationic analytes de-crease, because the ammonium ion competes with the cationic analyte inpairing to the anionic micelle. The effect of the ion-pair reagent on selec-tivity depends significantly on the structure of the reagent, e.g., the lengthof the alkyl chain. Figure 16 demonstrates the effect of the addition oftetramethylammonium ion to the SDS micellar solution on the selectivityfor cephalosporin antibiotics. Also shown in Figure 16 is the CZE separa-tion (i.e., with no SDS). In this case, poor selectivity is obtained. The bestseparation is obtained when the tetramethylammonium salt is added to thebuffer (panel C). It should be mentioned that all cephalosporins arebaseline resolved when 100 mM SDS instead of 50 mM is used.

1

24

6

9

0 5 10

3+5

7+8

Time (min)

(A)

2

3+4

7+8

6

1

5

50 10

Time (min)

(B)

31

9

8

76

45

321

1550 10

Time (min)

(C)

Figure 16. Separation of cephalosporin antibiotics by CZE (A), MEKCwith SDS (B), and MEKC with SDS and tetramethylammonium salt (C):1 = C-TA; 2 = ceftazidime; 3 = cefotaxime; 4 = cefmenoxime;5 = cefoperazone; 6 = cefpiramide; 7 = cefpimizole; 8 = cefminox;9 = ceftriaxone. (A) Run buffer, 20 mM phosphate-borate (pH 9.0); (B)with 50 mM SDS added to run buffer A; (C) with 40 mM tetramethylammo-nium bromide added to run buffer B. Capillary, 50µm i.d. × 65 cm (effec-tive length 50 cm); applied voltage 20 kV; detection, UV absorbance at210 nm. Reproduced with permission from Nishi et al., Anal. Chem. 61,2434 (1989).

32

Urea

A high concentration of urea is known to increase the solubility of hydro-phobic compounds in water. Urea also breaks down hydrogen-bond forma-tion in the aqueous phase. In MEKC, the addition of a high concentrationof urea to the SDS solution improves the separation of highly hydrophobiccompounds. In Figure 17, the effect of urea is shown for the separation oflipophilic corticosteroids. Urea slightly reduces the electroosmotic velocityand considerably reduces the migration velocity of the micelle, resulting ina reduced capacity factor. Although the addition of urea causes a slightincrease in viscosity, the decrease of vmc is more significant than would beexpected from the viscosity change alone. Selectivity is not remarkablyaltered by the addition of urea, but minor changes are noticeable, espe-cially for the separation of closely related compounds.

(A)

(B)

0 5 10 15 20 25 30

Time (min)

1

23 4

5

6

7 8

Figure 17. The effect of urea addition to the SDS solution: 1 = hydrocorti-sone; 2 = hydrocortisone acetate; 3 = betamethasone; 4 = cortisone ac-etate; 5 = triamcinolone acetonide; 6 = fluocinolone; 7 = dexamethasoneacetate; 8 = fluocinonide. (A) without urea; (B) with the addition of 6 Murea to the run buffer. 50 mM SDS in 20 mM phosphate-borate (pH 9.0);capillary, 50µm i.d.× 65 cm (effective length 50 cm); applied voltage,20 kV; detection, UV absorbance at 210 nm. Reproduced with permissionfrom Terabe et al., J. Chromatogr. 545, 359 (1991).

33

Organic Solvents

Organic solvents miscible with water are widely used as mobile phasemodifiers in HPLC. The main objective of using organic solvents in RPLCis to adjust the capacity factor close to the optimum value. The result is areasonable analysis time, even for very hydrophobic compounds. We mayexpect a similar effect with the use of organic solvents in MEKC. How-ever, it should be noted that a high concentration of the organic solventmay break down the micellar structure. Generally, concentrations of up to20% organic solvent can be used without difficulty in MEKC. The use oforganic solvents contributes to the improvement of resolution or the alter-ation of selectivity. In general, the addition of methanol, 2-propanol oracetonitrile reduces the electroosmotic velocity and, hence, expands themigration time window. This is illustrated in Figure 18 for the separationof a number of aromatic sulfides with the addition of methanol to the runbuffer.

Metal Salts

It has been reported that the addition of certain metal salts to the SDS mi-cellar solution improves selectivity. In particular, the MEKC separation ofoligonucleotides is enhanced by the addition of magnesium, zinc, or copper(II) ion. Metals ions are electrostatically attracted to the surface of themicelle where they can be selectively complexed with analytes. (Careshould be taken not to cause precipitation of metal salts with SDS.) Theanalyte/metal ion/micelle complexation results in a widened migrationtime window and enhanced selectivity. An example of this approach isshown in Figure 19. 3 mM Zn (II) was added to the buffer to significantlyenhance the resolution of oligonucleotides. Relatively low concentrationsof metal salt (0.3–3.0 mM) give a good compromise between zone broad-ening and resolution.

34

0

5 16

2

8

4

3,7

9

10

1112

10 20 30

Time (min)

0.004 AU

(A)

(B)

10 30 50

Time (min)

1,56

2

8

4

73

11

109

0.003 AU

12

0

Figure 18. Effect of methanol addition to the SDS solution: 1 = benzylmethyl sulfide; 2 = benzyl ethyl sulfide; 3 = benzyl propyl sulfide; 4 = ben-zyl isopropyl sulfide; 5 = methyl phenyl sulfide; 6 = ethyl phenyl sulfide;7 = phenyl propyl sulfide; 8 = isopropyl phenyl sulfide; 9 = butyl phenylsulfide; 10 = isobutyl phenyl sulfide; 11 =s-butyl phenyl sulfide;12 = Sudan III. (A) without methanol; (B) with 20% methanol added. Runbuffer, 20 mM SDS in 50 mM phosphate-borate (pH 7.0); capillary, 50µmi.d. × 90 cm (effective length 75 cm); applied voltage 21 kV; detection, UVabsorbance at 210 nm; temperature, 35°C. Reproduced with permissionfrom Otsuka et al., Nippon Kagaku Kaishi, no. 7, 950 (1986).

35

10 20 30 400

Figure 19. Separation of a mixture of 18 oligonucleotides, each with 8bases. Run buffer, 7 M urea, 20 mM Tris, 5 mM sodium phosphate, 50 mMSDS, 3 mM Zn (II). Reproduced with permission from Cohen et al., Anal.Chem. 59, 1021 (1987).


Ackermans, M. T., Everaerts, F. M., Beckers, J. L. Determination of somedrugs by micellar electrokinetic capillary chromatography: the pseudo-effective mobility as parameter for screening. J. Chromatogr. 585, 123-131(1991)

Amankwa, L. N., Kuhr, W. G. Indirect fluorescence detection in micellarelectrokinetic chromatography. Anal. Chem. 63, 1733-1737 (1991)

36

Cole, R. O., Sepaniak, M. J., Hinze, W. L., Gorse, J., Oldiges, K. Bile saltsurfactants in micellar electrokinetic capillary chromatography: applicationto hydrophobic molecule separations. J. Chromatogr. 577, 113-123 (1991)

Lecoq, A.-F., Leuratti, C., Marafante, E., Di Biase, S. Analysis of nucleicacid derivatives by micellar electrokinetic capillary chromatography.J. High Resolut. Chromatogr. 14, 667-671 (1991)

Thormann, W., Meier,P., Marcolli, C., Binder, F. Analysis of barbituratesin human serum and urine by high performance capillary electrophoresis-micellar electrokinetic capillary chromatography with on-column multi-wavelength detection. J. Chromatogr. 545, 445-460 (1991)

Terabe, S., Ishihama, Y., Nishi, H., Fukuyama, T., Otsuka, K. Effect ofurea addition in micellar electrokinetic chromatography. J. Chromatogr.545, 359-368 (1991)

Weinberger, R., Lurie, I. S. Micellar electrokinetic capillary chromatogra-phy of illicit drug substances. Anal. Chem. 63, 823-827 (1991)

Nishi, H., Fukuyama, T., Matsuo, M., Terabe, S. Effect of surfactant struc-ture on the separation of cold medicine ingredients by micellar electroki-netic chromatography. J. Pharm. Sci. 79, 519-523 (1990)

Nishi, H., Fukuyama, T., Matsuo, M., Terabe, S. Separation and determina-tion of lipophilic corticosteroids and benzothiazepin analogues by micellarelectrokinetic chromatography using bile salts. J. Chromatogr. 513, 279-295 (1990)

Rasmussen, H. T., Goebel, L. K., McNair, H. M. Micellar electrokineticchromatography employing sodium alkyl sulfates and Brij 35. J. Chroma-togr. 517, 549-555 (1990)

Terabe, S., Miyashita, Y., Shibata, O., Barnhart, E. R., Alexander, L. R.,Patterson, D. J., Karger, B. L., Hosoya, K., Tanaka, N. Separation ofhighly hydrophobic compounds by cyclodextrin- modified micellar electro-kinetic chromatography. J. Chromatogr. 516, 23-31 (1990)

Nishi, H., Tsumagari, N., Kakimoto, T., Terabe, S. Separation of β-lactamantibiotics by micellar electrokinetic chromatography. J. Chromatogr. 477,259-270 (1989)

Nishi, H., Tsumagari, N., Kakimoto, T., Terabe, S. Separation of water-soluble vitamins by micellar electrokinetic chromatography. J. Chroma-togr. 465, 331-343 (1989)

Nishi, H., Tsumagari, N., Terabe, S. Effect of tetraalkylammonium salts onmicellar electrokinetic chromatography. Anal. Chem. 61, 2434-2439 (1989)

37

Balchunas, A. T., Sepaniak, M. J. Gradient elution for micellar electroki-netic capillary chromatography. Anal. Chem. 60, 617-621 (1988)

Fujiwara, S., Iwase, S., Honda, S. Analysis of water-soluble vitaminsby micellar electrokinetic capillary chromatography. J. Chromatogr. 447,133-140 (1988)

Gorse, J., Balchunas, A. T., Swaile, D. F., Sepaniak, M. J. Effects of or-ganic mobile phase modifiers in micellar electrokinetic capillary chroma-tography. HRC CC, J. High Resolut. Chromatog. Chromatogr. Commun.11, 554-559 (1988)

Snopek, J., Jelinek, I., Smolkova-Keulmansova, E. Micellar, inclusion andmetal-complex enantioselective pseudophases in high-performanceelectromigration methods. J. Chromatogr 452, 572-590 (1988)

Balchunas, A. T., Sepaniak, M. J. Extension of elution range in micellarelectrokinetic capillary chromatography. Anal. Chem. 59, 1466-1470(1987)

Cohen, A. S., Terabe, S., Smith, J. A., Karger, B. L. High-performancecapillary electrophoresis separation of bases, nucleosides, and oligonucle-otides: retention manipulation via micellar solution and metal additives.Anal. Chem. 59, 1021-1027 (1987)

Szejtli, J., Zsadon, B., Cserhati, T. Cyclodextrin use in separations, in"Ordered Media in Chemical Separations," Hinze, W. L., Armstrong,D. W., eds. ACS Symposium Series 342, American Chemical Society,Washington, D.C., pp. 200-217 (1987)

Otsuka, K., Terabe, S., Ando, T. Separation of aromatic sulfides by elec-trokinetic chromatography with micellar solution. Nippon Kagaku Kaishi,No. 7, 950-955 (1986)

Otsuka, K., Terabe, S., Ando, T. Electrokinetic chromatography with mi-cellar solutions: separation of phenylthiohydantoin-amino acids.J. Chromatogr. 332, 219-226 (1985)

Otsuka, K., Terabe, S., Ando, T. Electrokinetic chromatography with mi-cellar solutions: retention behavior and separation of chlorinated phenols.J. Chromatogr. 348, 39-47 (1985)

38

IV: Enantiomeric Separations

In the pharmaceutical industry, the determination of the optical purity orseparation and determination of enantiomers is becoming increasinglyimportant. The non-active enantiomer in a drug formulation may be con-sidered as an impurity. High resolution separation methods are required toachieve chiral separations. The approaches used in CE, discussed next, arerelatively simple compared to HPLC methods in which expensive, chiralstationary phases are frequently used. In CE only minute amounts of chiralselectors are required to determine enantiomeric purity.

Two approaches can be used to perform enantiomeric separations inMEKC:

1. Use of chiral surfactants2. Use of chiral additives

1. Chiral SurfactantsBile salts are widely available commercially and have shown to be usefulchiral surfactants. Sodium cholate (see the separation in Figure 9) or so-dium deoxycholate can be used under neutral or alkaline conditions toionize the carboxyl group of the surfactant. Taurine conjugates of bile saltscan also be used in acidic conditions because taurine has a sulfonic acidgroup. Amino acid-derived surfactants (e.g., sodium N-dodecanoyl-L-valinate, SDVal) are another group of chiral surfactants that are commer-cially available. They also must be used under neutral or alkalineconditions. In order to use these surfactants under acidic conditions, SDScan be added to form mixed micelles with appreciable electrophoreticmobilities. The addition of a small amount of methanol and/or a relativelyhigh concentration of urea often improve resolution while sharpening peakprofiles. An example of the enantiomeric separation of PTH amino acidswith SDVal is shown in Figure 20. Alternatively, digitonin can be used toform mixed micelles for these types of applications (see Figure 12).

39

0 20 40

5

6

4

3

0

1

2

Figure 20. Enantiomeric separation of six PTH-DL-amino acids by MEKCwith SDVal: 0 = acetonitrile; 1 = Ser; 2 = Aba; 3 = Nva; 4 = Val;5 = Trp; 6 = Nle. Conditions: Run buffer, 50 mM SDVal-50 mM SDS/0.5 M urea in 50 mM borate buffer (pH 9.0) containing 10% methanol;detection, UV absorbance at 260 nm. Reproduced with permission fromOtsuka et al., J. Chromatogr. 559, 209 (1991).

40

2. Chiral AdditivesThe second, more popular, method of enantiomeric separation by MEKC isto add CD to the micellar solution (see also discussion on page 27,CDMEKC). The SDS micelle may be conveniently used for this approach.Various CDs or CD derivatives may be tried. The concentrations of SDSand CD should be optimized to yield optimal capacity factors (see ChapterV). An example of this approach is shown in Figure 21. Cicletanine is amember of a new class of antihypertensive drugs, the fusopyridines. Chiralselectivity was obtained by using a buffer consisting of 100 mM borate,50 mM SDS, pH 8.6 and 50 mM β-CD. The addition of methanol and/orurea often enhances solubility and improves resolution. When γ-CD isemployed, a second chiral component, such as d-camphor-10-sulfonate orl-menthoxyacetic acid, may enhance resolution. This is illustrated in Fig-ure 22 for the separation of several barbiturates.

Abs

orba

nce

(214

nm

)

0.040

0.030

0.020

0.010

0.000

-0.005

14.0 16.0 18.0 20.0

Time (min)

Figure 21. Enantiomeric separation of R(-) and S(+) cicletanine. Runbuffer, 100 mM sodium borate, 50 mM SDS, pH 8.5. Reproduced fromPruñonosa et al., Beckman Application Note DS-798, 1990.

41

1

2

5

6

34

0 5 10 15 20 25

Time (min)

Figure 22. Enantiomeric separation of several barbiturates by CDMEKCwith d-camphor-10-sulfonate: 1 = thiopental (sodium salt); 2 = pentobar-bital (calcium salt); 3 = 2,2,2-trifluoro-1-(9-anthryl)ethanol); 4 = 2,2'-dihydroxy-1,1'-dinaphthyl; 5 = phenobarbital; 6 = barbital (sodium salt).Run buffer, 50 mM SDS, 30 mM γ-CD, 20 mM sodium d-camphor-10-sulfonate in 20 mM phosphate-borate (pH 9.0); capillary, 50µmi.d. × 65 cm (effective length 50 cm); applied voltage, 20 kV; detection, UVabsorbance at 220 nm. Reproduced with permission from Nishi et al.,J. Chromatogr. 553, 503 (1991).

Since the separation factor, α, of enantiomeric pairs is generally small,the separation conditions must be carefully optimized to maximize resolu-tion. This can be achieved, for example, by using a longer capillary, ahigher voltage, adjustment of the capacity factors, expansion of the migra-tion time window, etc. Suppression of electroosmosis often enhances reso-lution (see page 14, equation 11), although analysis times may becomelong. The use of acidic buffer conditions or the addition of an organicsolvent is useful in reducing the electroosmotic flow.

42


Nishi, H., Fukuyama, T., Terabe, S. Chiral separation by cyclodextrin-modified micellar electrokinetic chromatography. J. Chromatogr. 553,503-516 (1991)

Otsuka, K., Kawahara, J., Takekawa, K., Terabe, S. Chiral separation bymicellar electrokinetic chromatography with sodium N-dodecanoyl-L-valinate. J. Chromatogr. 559, 209-214 (1991)

Otsuka, K., Terabe, S. Enantiomeric resolution by micellar electrokineticchromatography with chiral surfactants. Chromatogr. 515, 221-226 (1990)

Dobashi, A., Ono, T., Hara, S., Yamaguchi, J. Optical resolution of enanti-omers with chiral mixed micelles by electrokinetic chromatography. Anal.Chem. 61, 1984-1986 (1989)

Nishi, H., Fukuyama, T., Matsuo, M., Terabe, S. Chiral separation of opti-cal isomeric drugs using micellar electrokinetic chromatography.J. Microcol. Sep. 1, 234-241 (1989)

Terabe, S., Shibata, M., Miyashita, Y. Chiral separation by electrokineticchromatography with bile salt micelles. J. Chromatogr. 480, 403-411(1989)

43

V: Strategy for Optimizing Resolution

After discussion of the separation principles, selectivity effects, and vari-ous applications, this final section of the primer provides a method devel-opment guide for MEKC. In many cases, the novice practitioner of MEKCwill be surprised how easy it is to get a good separation on the first try.This is often the case with other modes of CE as well. Indeed, many sepa-ration scientists feel that, in general, it is easier and quicker to get satisfac-tory results in CE than in HPLC. Many separation problems can be solvedwith a “standard” MEKC run buffer and operating conditions. These stan-dard conditions are listed in Table 3. When an electropherogram obtainedwith the standard operating conditions is not satisfactory, other optionsmust be pursued. These are discussed below and are summarized in theflow chart shown in Figure 23.

Table 3. Suggested Standard Operating Conditions

Running solution: 50 mM SDS in 50 mM borate buffer*(pH 8.5 – 9.0)

Capillary: 50 – 75 µm i.d. × 20 – 50 cm(from the injection end to the detector)

Applied voltage: 10 – 20 kV (keep current below 100 µA)Temperature: 25°C or ambientSample solvent: water or methanolSample concentration: 0.1 – 1 mg/mLInjection end: the positive or anodic endInjection volume: below 2 nL (or less than 1 mm from the end of

the capillary)Detection: 200 – 210 nm (depends on the sample)

* The buffer solution in which to dissolve SDS may be a 50 mM phosphate(pH 7.0) or 20 mM Good’s buffer (pH 7.0), if neutral conditions are de-sired.

Optimize the Capacity FactorThe optimum value of the capacity factor, k', is around 2 under the aboveconditions. Useful and practical capacity factors are in the range of 0.5 to10. When the migration time of the micelle, tmc, is difficult to measure,assume tmc is roughly equal to four times t0.

44

No

Yes

No

Yes

Run with standardMEKC conditions(refer to Table 3)

Does the separationrequire furtheroptimization?

Endor optimize

migration times

Anionicanalyte?

Increase[SDS]

Is the separationsuccessful?

End

No

Yes

Cationicanalyte?

Use an ion-pairreagent

(an ammoniumsalt) or acationic

surfactant

Use bile saltsinstead of SDS,

add CD intoan SDS solution

(CDMEKC),or add an

organic solventor urea to anSDS solution

Yes

No

Yes

No

No

Yes

Optimize k'to ≈2

and/or usea longercapillary

Yes

Rs < 0.5?

EstimateRs

Use a differentsurfactant

(including amixed micelle)

k' >10?k' < 0.5?

Calculatek'

No

Figure 23. Optimization scheme for MEKC separations.

45

If k' is less than 0.5, the concentration of SDS may be increased up to200 mM, (see equation 10). When k' is still too small even when the SDSconcentration is relatively high, the surfactant should be altered. In such acase, the analytes are negatively charged, and pH manipulation should beconsidered as will be described later.

When k' is too large (i.e., k'>10), many options are available. For ca-pacity factors less than 10, a reduction of the SDS concentration should beeffective. However, the SDS concentration must be kept higher than10 mM, as low SDS concentrations limit the sample loading capacity.When k'>10 or when the migration times are close to tmc, other parametersshould be changed.

Change the SurfactantEven when the capacity factors are in the optimum range, separations maynot always be successful. One of the easiest ways to improve resolution isto change the surfactant. At present, it is hard to predict which surfactantwill be suitable for the separation of particular analytes. In general, thepolar group of the surfactant molecule affects selectivity more than thenonpolar group. Therefore we recommend using surfactants with differentpolar groups to change selectivity. When the capacity factors are too large,the use of a bile salt may give successful results. Cationic surfactants showsignificantly different selectivities from anionic surfactants. The substitu-tion of a cationic surfactant for an ionic surfactant should dramaticallychange selectivity for ionic analytes. The addition of a nonionic surfactantto the ionic micellar solution also affects selectivity but, as we have seen inFigure 12, this approach results in an increase of the capacity factors and anarrower migration time window.

Optimize the pH of the BufferAs in CZE, the pH of the micellar solution is an important parameter inoptimizing the separation of ionizable analytes. The pH does not signifi-cantly affect the selectivity of neutral analytes. In general, when an anionicsurfactant is employed, the increase in pH will decrease the capacity fac-tors of acidic compounds. Acids ionize at a high pH and have their ownelectrophoretic mobilities which are usually less than the mobility of theSDS micelle. That is, the acids will migrate at velocities that are slowerthan the electroosmotic flow at a high pH, although their interaction withthe SDS micelles will be weak. The simplest way to find the optimum pHis to change the pH by one or two pH units and to run the separation in acertain pH range.

46

Modify the Aqueous PhaseCyclodextrin is one of the most useful additives, not only for the separationof very hydrophobic or enantiomeric analytes but also for changing theselectivity. Among the various CDs, γ-CD in general will be most effectivefor most analytes in MEKC. The other CDs also change selectivity andshould be used in case γ-CD is not very effective. The concentration of CDaffects resolution and should be optimized. In CDMEKC, the CD concen-tration is dependent on the k' (in the absence of CD): for the separation ofwater-soluble compounds or enantiomers, 10 mM of CD is generally suffi-cient if the k's are in the proper range without CD. For the separation ofhighly hydrophobic compounds (k' usually >10), CD concentrations of>40 mM will be necessary.

Methanol or acetonitrile are other useful additives for changing selec-tivity as well as expanding the migration time window. It is recommendedthat the concentration of the organic solvent remains below 20% in order toavoid breaking up the micellar structure. Other additives (ion-pair reagents,urea, metal salts) were discussed earlier and may also be useful in certainapplications.


Khaledi, M. G., Smith, S. C., Strasters, J. K. Micellar electrokinetic capil-lary chromatography of acidic solutes: migration behavior and optimiza-tion strategies. Anal. Chem. 63, 1820-1830 (1991)

Strasters, J. K., Khaledi, M. G. Migration behavior of cationic solutesin micellar electrokinetic capillary chromatography. Anal. Chem. 63,2503-2508 (1991)

Vindevogel, J., Sandra, P. Resolution optimization in micellar electroki-netic chromatography: use of Plackett-Burman statistical design for theanalysis of testosterone esters. Anal. Chem. 63, 1530-1536 (1991)

Foley, J. P. Optimization of micellar electrokinetic chromatography.Anal Chem. 62, 1302-1308 (1990)

A p p l i c a t i o n ␣ I n f o r m a t i o n

P r o t e i n. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A-1777A

BECKMAN

Characterization of Proteases by Capillary Electrophoresis withLaser-Induced Fluorescence DetectionFu-Tai A. ChenBeckman Instruments Inc., Fullerton, CA

IntroductionProteases are enzymes which catalyze the hydrolysisof peptide bonds. Proteases are often present incomplex biological mixtures such as fermentationbroths, serum, or urine. In biotechnology, proteasecontamination may present a serious problem whendesirable proteins are purified. Traditionally, analy-sis of protease activity involves chromogenic orfluorogenic substrates to assay enzyme activity.(1)

Most protease substrates are peptides linked toortho- or para-nitroanilide through amide bonds.Protease digestion of the substrate results in the for-mation of ortho- or para-nitroaniline which absorbslight significantly more strongly than does the sub-strate. By designing a nitroanilide peptide with aspecific amino acid sequence, a protease may beprobed. However, at very low protein concentrations(i.e., 10-12 M), the enzymatic activity is difficult, ifnot impossible, to determine with traditional spec-trophotometric methods. This is because the conver-sion of substrate to product is insignificant com-pared to the existing background of the substrate.

In this Application Bulletin (and in a recent pa-per(2)), we propose to use a fluorescently labeledpeptide as a stable substrate for the analysis of low-

level proteases. The enzyme hydrolysis is monitoredby P/ACE capillary electrophoresis with laser-in-duced fluorescence detection (CE-LIF). Thefluorophore used to label the substrate is Cy3, a cya-nine-type dye which has an absorption maximumclose to the helium-neon laser (543 nm) or a solid-state, frequency-doubled diode laser (532 nm). Thelaser light sources can be easily connected to theP/ACE system via standard fiber optic couplers.

Protease assays using CE with UV absorbancedetection have been described.(3,4) However, LIFdetection is orders of magnitude more sensitive thanUV, and should be, therefore, well suited for low-level analyte detection.

Experimental

Materials

Cy3, a carboxyl activated cyanine dye, was pur-chased from Biological Detection System (Pitts-burgh, PA). All buffer components, angiotensin I,angiotensin II, and trypsin were products of SigmaChemical Co. (St. Louis, MO). Proteinase K, car-boxypeptidase P and Y were purchased fromBoeringer Mannheim (Indianapolis, IN).

2

CE-LIFA P/ACE 2100 instrument equipped with laser-in-duced fluorescence (LIF) detection (Beckman In-struments, Inc., Fullerton, CA), was used for the CEruns. Post-run data analysis was performed withGold® (version 7.0) software. Capillary columns,27 cm length (20 cm to detector window) × 20 µmi.d. (Polymicro Technologies, Phoenix, AZ) wereassembled in a P/ACE LIF cartridge. This cartridgecontains an ellipsoidal mirror to collect fluores-cence. A 15-mW, frequency-doubled diode laseremmiting at 532 nm was kindly provided by ScottMiller of Amoco Laser (Naperville, IL). A laserheadcoupler to a standard SMA-906 fiber connectorto the P/ACE system with LIF detector was a prod-uct of OZ optics (Ontario, Canada). The fluores-cence signal was collected through a narrow-bandfilter of 590 nm ± 9 nm (Oriel, Stratford, CT) whilethe laser beam was rejected by a notch filter at532 nm (Applied Physics, Torrance, CA). A 5-mW,“green” helium-neon laser (543 nm) was purchasedfrom Particle Measurement Systems (Boulder, CO)and a notch filter at 543.5 nm was purchased fromBarr Associates (Westford, MA).

Synthesis of Cy3-AngiotensinsCy3, an activated carboxyl cyanine dye, wascoupled to the N-terminal of the peptides. Angio-tensin I was dissolved in 50 mM phosphate buffer(pH 7.5) at a final concentration of 1.0 mg/mL.50 µL of the angiotensin I solution (equivalent to46 nmol) was added to a vial of the activated Cy3(80 nmol) at room temperature for 30 minutes. Theresulting mixture was chromatographed on a C-18reversed-phase column (4.1 mm × 25 cm, BeckmanODS Spherogel). Peaks containing peptide andCy3 were collected from the liquid chromatograph(System Gold). Purity was assessed by LC, CE,and UV/Visible spectrophotometry. Similarly,Cy3-angiotensin II and Cy3-aspartic acid were syn-thesized by the above procedures.

Protease DigestionCy3-angiotensin I (10-8 M, in 90 µL of 0.05 M Tris-HCl buffer, pH 8.0) was mixed with 10 µL protein-ase K (0.1 U/mL) at room temperature. The reactionmixture was immediately monitored by CE-LIF.Samples were introduced by pressure injection for20 s. Electrophoresis was performed in a 200 mMborate buffer, pH 10.2, with a field strength of

740 V/cm (20 kV/30 µA). The capillary was main-tained at ambient temperature (23°C) in the P/ACEcartridge. Between runs, the capillary was sequen-tially rinsed under high pressure with 1.0 N sodiumhydroxide and water (12 s each), followed by recon-ditioning with borate buffer for 60 s.

Trypsin and carboxypeptidase catalyzed hy-drolysis of Cy3-angiotensin I and II were performedsimilarly.

Results

Digestion of Cy3-Angiotensin I and II withTrypsin and Carboxypeptidase PThe purity of the labeled substates, Cy3-angiotensinI (asp-arg-val-tyr-ile-his-pro-phe-his-leu) andCy3-angiotensin II (asp-arg-val-tyr-ile-his-pro-phe),was first confirmed by CE. As can be seen inFigure 1A, the migration times of Cy3-angiotensin I(peak 1) and Cy3-angiotensin II (peak 3) are 2.1 and2.25 minutes, respectively. Cy3 diacid (migrationtime 4.2 min) is present in each sample as a refer-ence. Addition of trypsin to the above mixture re-sults in the formation of a single Cy3-peptide, asshown in Figure 1B. The specificity of trypsin di-gestion suggests that this peptide is Cy3-asp-arg(peak 9). Addition of carboxypeptidase P to thefinal tryptic digest mixture led to the formation ofCy3-asp (peak 10, Figure 1C).

00

0.05

0.1

0.15

0.2

0.25

0.3

1 2 3Time (min)

4 5 6 7

Rel

ativ

e F

luor

esce

nce

Inte

nsity

A

B

C

1

3

9

R

R

R

10

Figure 1. A: CE-LIF electropherogram ofCy3-angiotensin I and II, 10 nM each; Cy3 diacid,1.0 nM. Laser light source: frequency-doubledsemiconductor laser, emitting at 532 nm (AmocoLaser). B: Tryptic digest of Cy3-angiotensin I and II,10 nM each; Cy3 diacid, 1.0 nM.C: Carboxypeptidase digest of the reactionmixture in Figure 1B. Peak i.d.: see Table 1;run voltage: 20 kV.

3

Digestion of Cy3-Angiotensin I with Proteinase KProteinase K is a powerful protease capable of de-grading native proteins, even at high pH or in thepresence of SDS or urea. The enzyme acts as anendo- as well as an exopeptidase. The time profileof enzymatic digestion of the substrate (Cy3-angio-tensin I) by proteinase K can be monitored in Fig-ure 2. Immediately after the addition of proteinase K(approximately 0.1 min), about 30% Cy3-angio-tensin I is hydrolyzed to Cy3-angiotensin II (datanot shown). Eight minutes after the addition of pro-teinase K, nearly all Cy3-angiotensin I is convertedto Cy3-angiotensin II (peak 3, migration time2.25 min) along with the formation of a minor prod-uct, Cy3-asp-arg-val-tyr, appearing at 2.8 minutes(peak 7). Further digestion leads to more Cy3-asp-arg-val-tyr formation (peak 7) and its degradationproduct, Cy3-asp-arg-val (peak 8). A new2 species,Cy3-asp-arg (peak 9) begins to show up, as seen inFigure 2C. (Note: the structural assignments weremade based on the electrophoretic mobilities of thespecies). Further digestion (124 min, Figure 2D) re-

sults in the disappearance of peak 7 while peak 9continues to grow at the expense of peak 8. At310 min (Figure 2E), peak 9 is the only one remain-ing. The structure representing peak 9 is identical tothe product of the tryptic digestion of Cy3-angio-tensin I and Cy3-angiotensin II, shown in Figure 1B.Addition of the enzyme carboxypeptidase P to thedigest of E yields Cy3-asp (Figure 2F) identical tothe product (peak 10) obtained in Figure 1C.

Digestion of Cy3-Angiotensin I withCarboxypeptidase YCarboxypeptidase Y is a serine protease whichsucessively releases all amino acids from the C-ter-minal of peptides or proteins. One minute after theaddition of the carboxypeptidase Y to a 10 nM solu-tion of Cy3-angiotensin I, the formation of four ma-jor species can be observed, i.e., Cy3-asp-arg-val-tyr-ile-his-pro-phe (peak 3, Cy3-angiotensin II),Cy3-asp-arg-val-tyr-ile-his (peak 5), Cy3-asp-arg-val-tyr (peak 7), and Cy3-asp-arg-val (peak 8).

0

0.5

0.6

0.7

0.8

0.9

0

0.1

0.2

0.3

0.4

1.0

1 2 3Time (min)

4 5 6 7

Rel

ativ

e F

luor

esce

nce

Inte

nsity

A

B

C

1

3

R

R71

R

87

9

D

E

F

8

9

9

R

R

R

10

Figure 2. A: CE-LIF electropherogram ofCy3-angiotensin I, 20 nM, and Cy3 diacid, 1.0 nM.LIF source: same as that in Figure 1.B: Proteinase K digest of Cy3-angiotensin I at 8 min.C: Digest at 102 min. D: Digest at 124 min.E: Digest at 310 min. F: Carboxypeptidase Pdigest of the reaction mixture of Figure 2E.Peak i.d.: see Table 1; run voltage: 20 kV.

0.5

0.6

0.7

0.8

0.9

1.0

1

Rel

ativ

e F

luor

esce

nce

Inte

nsity

A

B

C

1

3

R

R4

5

6 82

R

8

7

7

0

0.1

0.2

0.3

0.4

1 2 3Time (min)

4 5 6

D

E

F

8

7

8

9

R

R

R

Figure 3. A: CE-LIF electropherogram ofCy3-angiotensin I, 20 nM, and Cy3 diacid, 1.0 nM.Laser light source: A “green” helium-neon laser,emitting at 543 nm (Particle Measurement Systems).Peak i.d.: see Table 1; run voltage: 20 kV.B: Carboxypeptidase Y digest of Cy3-angiotensin Iat 1.0 min. C: Digest at 20 min. D: Digest at 60 min.E: Digest at 250 min. F: Addition of proteinase K tothe reaction mixture of E.

XXXX-XXXX-XX-X © 1994 Beckman Instruments, Inc. Printed in U.S.A. on recycled paper.

Worldwide Offices: Africa, Middle East, Eastern Europe (Switzerland) (22) 994 07 07. Australia (61) 02 816-5288. Austria (2243) 85656-0.Canada (800) 387-6799. China (861) 5051241-2. France (33) 1 43 01 70 00. Germany (49) 89-38871. Hong Kong (852) 814 7431.Italy (39) 2-953921. Japan 3-3221-5831. Mexico 525 575 5200, 525 575 3511. Netherlands 02979-85651. Poland 408822, 408833.Singapore (65) 339 3633. South Africa (27) 11-805-2014/5. Spain (1) 358-0051. Sweden (8) 98-5320. Switzerland (22) 994 07 07.Taiwan (886) 02 378-3456. U.K. (0494) 441181. U.S.A. 1-800-742-2345.

BECKMAN

Beckman Instruments, Inc. • 2500 Harbor Boulevard, Box 3100 • Fullerton, California 92634-3100Sales: 1-800-742-2345 • Service: 1-800-551-1150 • TWX: 910-592-1260 • Telex: 678413 • Fax: 1-800-643-4366

Minor species are Cy3-asp-arg-val-tyr-ile-his-pro-phe-his (peak 2), Cy3-asp-arg-val-tyr-ile (peak 4),and Cy3-asp-arg-val-tyr-ile (peak 6), shown in Fig-ure 3B. Each of the species is the result of sequen-tial hydrolysis from the C-terminal end of the pep-tide. Continued digestion after 20 min leads to theaccumulation of two major species, peaks 7 and 8(Figure 3C). Further digestion yields peak 8 at theexpense of 7 (Figure 3D). Exhausitive digestion re-sults in the formation of peak 8, Cy3-asp-arg-val, asthe end product (Figure 3E). Addition of proteinaseK to the Cy3-asp-arg-val (peak 8) leads to the for-mation of Cy3-asp-arg, Figure 3F, identical to theproduct of trypsin-catalyzed digestion of Cy3-an-giotensin I and II shown in Figure 2B.

ConclusionProtease-catalyzed hydrolysis of a substrate can bestudied by CE. Labeling the substrate with the cya-nine dye Cy3 permits highly sensitive P/ACE-LIFdetection in conjunction with economical laser lightsources. The CE technique provides high-resolutionseparation (and potential quantitation) of substrateand digestion products. Furthermore, the dynamicrange of LIF detection is relatively large, typically5–6 orders of magnitude. This should allow, for ex-ample, routine detection of 10-9 M product in thepresence of 10-6 M substrate. The present methodcan be readily adapted by employing well-definedpeptide sequences to probe the specificity of pro-teases in real samples (e.g., fermentation broths, se-rum). Kinetics of relative hydrolytic reactivity can

also be studied with a similar approach. The conceptshould also be applicable to monitoring the activityof other types of enzymes such as synthetases.

Table 1. Peak Identification of Figures 1–3.

1 Cy3-asp-arg-val-tyr-ile-his-pro-phe-his-leu(Cy3-angiotensin I)

2 Cy3-asp-arg-val-tyr-ile-his-pro-phe-his3 Cy3-asp-arg-val-tyr-ile-his-pro-phe

(Cy3-an-giotensin II)4 Cy3-asp-arg-val-tyr-ile-his-pro5 Cy3-asp-arg-val-tyr-ile-his6 Cy3-asp-arg-val-tyr-ile7 Cy3-asp-arg-val-tyr8 Cy3-asp-arg-val9 Cy3-asp-arg

10 Cy3-aspR Cy3-diacid

References:1. Orlowski, M., Wilk, S. Peptides, Structure and

Biological Function, Gross, E. and Meinhofer, J.(eds.), Rockford, IL: Pierce Chemical Co., p. 925,1980.

2. Chen, F.-T. A. Submitted for publication toAnal. Biochem.

3. Mulholland, F., Movahedi, S., Hague, G. R.,Kasumi, T. J. Chromatogr. 636, 63 (1993)

4. Vinther, A., Adelhorst, K., Kirk, O. Electro-phoresis 14, 486 (1993)

T · H · E

B E C K M A N

S C I E N C E

S U P P O R T

S O L U T I O N S


N u c l e i c A c i d s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A-1788A

Michael J. Fasco,1 Chris P. Treanor,1 S. Spivack,2 Helen L. Figge,2 and Laurence S. Kaminsky1

1Wadsworth Center, New York State Department of Health, Albany, NY2Albany Medical College, Albany, NY

IntroductionPCR-based methods are extremely useful for quali-tative detection of RNA or DNA in a wide variety ofspecimens. However, accurate quantitation—espe-cially when dealing with low copy numbers—is of-ten problematic. This is the case, for example, inclinical PCR* applications involving the determina-tion of low-level RNA viral load in plasma.(1) Agroup of methods, termed competitive PCR, havebeen described recently which effectively deal withthe problem of accurate quantitation of PCR prod-ucts. In competitive PCR,(1-4) a known amount ofstandard template DNA (the “competitor”) competesfor the same amplification with an unknown amountof target DNA (in the case of RNA-PCR, the DNAis obtained by reverse transcription). The competi-tor’s sequence is chosen such that it is largely iden-tical to the target sequence, except for the presenceof a mutated restriction site or a small addition/dele-tion sequence which allows separation of the twoproducts. During the amplification cycles, the targetand competitor are exposed to the same PCR-relatedreaction variables; their product ratio should, there-fore, remain constant, even after the products havereached a plateau. The amount of target DNA (or

RNA) can be obtained from the competitor concen-tration when the ratio of the simultaneous amplifica-tion products is 1. A variation of this method,termed multiplex competitive PCR, involves co-am-plification of the cDNA of a “housekeeping” gene(in addition to amplification of the target and itscompetitor) whose RNA does not vary among thedifferent samples to be analyzed. The expression ofthe target gene is then calculated in reference to thehousekeeping gene.(3)

Slab gel electrophoresis in conjunction with au-toradiography and scanning densitometry is typi-cally used to separate and detect the DNA frag-ments. However, most of these methods are laborintensive, time consuming, and not readily auto-mated. Furthermore, they lack the precision and ac-curacy of high performance separation techniquessuch as HPLC or capillary electrophoresis (CE). Re-cently, capillary electrophoresis with laser-inducedfluorescence (CE-LIF) detection has emerged as aattractive alternative to slab gel techniques: DNAfragments or PCR products, intercalated with fluo-rescent dyes, can be detected at extremely low levelsin real time with high efficiency and precision.(5,6)

BECKMAN

Competitive RNA-PCR by Capillary Electrophoresisand␣ Laser-Induced Fluorescence (LIF) Detection forQuantitation␣ of␣ Cellular mRNA

2

The object of this Application Bulletin and aforthcoming publication(4) is to explore the utility ofCE-LIF for quantitative competitive RNA-PCR ap-plications. The dimeric fluorescent dye YOYO-1was used in the CE run buffer for intercalating theDNA fragments, thus allowing LIF detection withhigh resolution and detectability down to 0.5 fmoleper mL of PCR product. The CE-LIF method wasapplied to reversed transcribed RNA from glyceral-dehyde-3-phosphate dehydrogenase (GAPDH) andcytochrome P450-1A1 gene sequences.

Experimental

CE-LIF

A P/ACE™ 2200 equipped with a 488-nm, Ar ionLIF detector (Beckman Instruments, Inc., Fullerton,CA) was used for all CE runs. The instrument wasused in the reversed polarity mode. Coated capillar-ies (eCAP™ dsDNA 1000), 47 cm (40 cm to detec-tor) × 100 µm i.d. (Beckman), were maintained at25°C. The sieving buffer consisted of TBE (89 mMTris, 89 mM boric acid, 2 mM EDTA, pH 8.5),0.5% HPMC, and 0.67 nM YOYO-1. A linear poly-acrylamide-based sieving buffer (eCAP dsDNA1000 gel buffer, Beckman) was found equally suit-able. Prior to a run, the capillary was rinsed (re-versed, high pressure) with sieving buffer, followedby a 12-s, high-pressure forward rinse of 5 µMYOYO-1 in TBE. Next, the sample was introducedby low pressure for 10–18 s. During the CE runs,the field strength was 200 V/cm.

Material and MethodsHydroxypropyl methylcellulose (HPMC, cat. no.H-7509) and 5-carboxyfluorescein were from SigmaChemical Co. (St. Louis, MO). YOYO-1 (oxazoleyellow dimer) was from Molecular Probes (Eugene,OR). φX 174 RF DNA Hae III digest was fromBeckman Instruments.

Reverse transcription and amplification reac-tions were performed without oil in a Perkin Elmer9600 thermal cycler (PE Applied Biosystems Divi-sion, Foster City, CA). Primers for P450-1A1 andGAPDH were designed from their cDNA with theaid of a computer program.(4) Competitors weremade from the cDNA of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD)-treated HepG2 cells. Details onRNA isolation, cell cultures, RNA-PCR conditions,primer and competitor sequences, and human actininternal standard can be found in reference 4.

Concentrations of internal standard solutionswere determined by mixing aliquots with knownquantities of φX 174 RF DNA Hae III and subse-quently comparing the CE peak areas. Concentra-tions of GAPDH and P450-1A1 competitor solu-tions were determined by comparing their peak ar-eas with calibrated actin standards.

Results and DiscussionPreliminary experiments were designed to optimizethe CE conditions for consistent run-to-run repro-ducibility and quantitation. The CE sieving buffer isa polymer network solution of 0.5% HPMC, similarto the one described by Schwartz and Ulfelder(5) oran eCAP dsDNA 1000 gel buffer (Beckman Instru-ments). Both polymer network solutions yield goodresults. The DNA intercalator we selected,YOYO-1, is a dimer of oxazole orange and belongsto the group of asymmetric cyanine dyes. Thesedyes have a very high binding affinity to DNA, en-hanced fluorescence upon binding, and minimalbackground fluorescence. However, in previousstudies, Srinivasan et al.(6) found it necessary topremix the DNA sample with YOYO-1 (in a molarratio of five bp to one dye ) prior to separation byCE. Hence, this method requires prior knowledge ofthe DNA sample concentration and is, therefore, notsuitable for quantitative PCR methods involving alarge number of samples of unknown concentration.We have found that the premixing step may simplybe eliminated by incorporating an extra step duringthe automated CE run sequence: by loading a plugof buffer—without the sieving matrix but with arelatively high concentration of YOYO-1 (5 µM)—onto the capillary prior to the sample introduction,the sample DNA is effectively mixed with the dyewithin the narrow capillary. The resulting electro-pherograms of DNA standards and PCR productsyield sharp peaks, comparable to those obtained bypremixing YOYO-1 with DNA samples. Figure 1shows the electropherogram of a φX 174 RF DNAHae III digest. The sample was introduced by a 10 spressure injection. In our experience, with the ma-jority of competitive RNA-PCR applications, pres-sure loading is the preferred sample introductionmethod. It provides adequate sensitivity, and ismore reproducible than the electrokinetic methodwhere the sample needs to be desalted prior to load-ing.(5,6) In addition, sample preparation is simpli-fied.

3

17.00

0.0000

20.0000

40.0000

30.0020.00 40.00 50.00

Rel

ativ

e F

luor

esce

nce

Uni

ts

72 b

p 118

bp

194

bp

234

bp

271

bp28

1 bp

307

bp

603

bp

872

bp10

78 b

p13

53 b

pTime (min)

Figure 1. Separation of φX 174 RF DNA Hae IIIdigest (10 ng/µL ) by CE-LIF. Sample introductionwas by pressure. A 6-s pre-injection of a 5 µMYOYO-1 plug was made, followed by a 6-s sampleintroduction. eCAP dsDNA 1000 buffer contained66 nM YOYO-1.

Using the DNA standard of Figure 1, linear re-lationships were obtained by plotting the number ofbp vs. their peak areas. Plots of the concentration ofthe φX 174 RF DNA standard (ng/mL ) vs. peakarea were also linear. However, some day-to-dayvariability was observed with these plots, indicatingthe need for an internal standard to compensate forvariations in YOYO-1 intercalation and/or fluores-cence signals. We selected a 328 bp human actinDNA fragment which was intermediate in size be-tween the target and competitor DNA fragments.A non-DNA internal standard (5-carboxyfluo-rescein) may be incorporated in the samples tocheck the sample loading efficiency.

Cytochrome P450-1A1 and GADPH were se-lected for model studies to test the CE-based com-petitive RNA-PCR system for the capability to de-tect different levels of mRNA. Cytochrome P450-1A1 is an important enzyme that metabolizes a widevariety of xenobiotics. It is virtually undetectable innormal tissue, but it is highly inducible by poly-nuclear aromatic hydrocarbons and chlorinated aro-matics such as TCDD. In contrast, GAPDH mRNAis not affected by inducing agents. By mixing equalconcentrations, but different proportions, of RNAfrom untreated and TCDD-treated human HepG2cells, a concentration gradient of P450-1A1 mRNAis created in the presence of a constant amount ofGAPDH mRNA.

A basic requirement of competitive PCR is thatthe target and competitor are amplified in proportionto their concentration during the PCR reaction. Fig-ure 2 shows a plot of PCR product concentration vs.cycle number for P450-1A1 target and competitor.Indeed, the two plots are very similar, indicatingthat similar amplification efficiencies are obtained.

4

3

2

1

0

2220 24 26 28 30 32

Cycle Number

pmol

1A

1 P

rodu

ct/0

.1 m

L R

eact

ion

Mix

ture

Figure 2. Amplification efficiencies of P450-1A1 (■)and its competitor (●) as a function of PCR cyclenumber. The competitor concentration was0.03 amol/0.1 mL PCR reaction mixture. The RNAfrom TCDD-treated HepG2 cells was 0.05 µg/0.1 mLreaction mixture. Reprinted with permission fromFasco, et al. (reference 4).

A typical competitive RNA-PCR experiment isillustrated in Figure 3. The formation of product(in pmol per 0.1 mL reaction mixture) is plotted vs.the initial RNA concentration. The initial competitorconcentration was kept constant at 0.1 amol per0.1 mL reaction mixture. The inset of Figure 3shows a log-log plot of the ratio of target to com-petitor product vs. the initial RNA concentration.

4

0.01 0.10 1.000

2

4

6

8

10

10.00

1.00

0.10

0.01 0.10

12

RNA (µg/0.1 mL RNA-PCR REACTION MIXTURE)

PM

OL

PR

OD

UC

T /

0.1

mL

RN

A-P

CR

RE

AC

TIO

N M

IXT

UR

E

[GA

PD

H] /

[CO

MP

ET

ITO

R G

AP

DH

]

Figure 3. Formation of target (■) and competitor(▲) GAPDH during competitive PCR. Equalconcentrations were reverse transcribed, mixed ata 1:1 ratio, and diluted. The competitor con-centration was kept constant at 0.1 amol/0.1 mLreaction mixture. In the inset plot, the data pointsobtained at 0.25 and 0.5 µg RNA were omitted.Reprinted with permission from Fasco, et al.(reference 4).

The x-intercept at a ratio of 1 corresponds tothe initial RNA concentration, i.e., in this case,0.012 µg per 0.1 mL reaction mixture (assuming100% reverse transcription efficiency) that contains0.1 amol of target mRNA. This value can also beobtained from the cross-over point of the graphs inFigure 3, but the inset plot allows a more accuratedetermination of the initial competitor RNA becausethe data points are obtained by linear regression.Representative CE-LIF electropherograms for thecompetitive RNA PCR experiments involvingP450-1A1 and GAPDH are shown in Figures 4Aand B, respectively.

0

20

40

60

80

100

0

20

40

60

80

100

120CF

CF

IS

IS

3

4

2

1

Rel

ativ

e F

luor

esce

nce

Uni

ts


A

B

5 10 15 20 25 30

Figure 4. A: CE-LIF electropherogram generatedduring competitive PCR of P450-1A1.1=competitor; 2=target; IS=internal standard;CF=5-carboxyfluorescein. B: Competitive PCR ofGAPDH. 3=competitor; 4=target. Reprinted withpermission from Fasco, et al. (reference 4).

Results from the experiments to detect changesin cytochrome P450-1A1 mRNA concentration areshown in Table 1. The measured values of P450-1A1 mRNA in the untreated/TCDD-induced RNAmixtures were in excellent agreement with the pre-dicted values. Two-fold differences in concentra-tions between different P450-1A1 mRNA solutionscould be detected by the method used in this study.Moreover, in untreated cell RNA, as little as 0.002amol per µg of P450-1A1 mRNA could be detected,demonstrating the high sensitivity of the CE-LIFmethod.

Table 1: Competitive PCR of Mixtures of RNAObtained from Untreated and TCDD-Treated Cells

RNA Mixture P450-1A1Untreated/TCDD-Treated Cells amol µg RNA

0/100 1.2050/50 0.6790/10 0.1599/1 0.02

100/0 0.002

5

ConclusionCompetitive PCR is the preferred method for abso-lute quantitation of specific DNA or RNA se-quences. In contrast to the slab gel-based methodsused for the separation and quantitation of PCRproducts, the CE method is fast and can be fully au-tomated; the computer-generated data are stored ondisk. CE-LIF allows accurate and precise quantita-tion of PCR products formed during competitivePCR reactions. With the described CE method,sharp peaks were reproducibly obtained (≈ 10 bpresolution) and run times were less than 30 min.Product detectability with LIF should be more thanadequate for most clinical and diagnostic applica-tions of competitive PCR. The CE-LIF procedurecan also be applied to multiplex competitive PCR.(4)

References1. Piatak, M., Jr., Luk, K.-C., Williams, B., Lifson,

J. D. BioTechniques 14, 70 (1993)2. Gilliland, G., Perrin, S., Blanchard, K., Bunn,

H. F. Proc. Natl. Acad. Sci. USA 87, 2725 (1990)3. Apostolakos, M. J., Schuermann, W. H. T.,

Frampton, M. W., Utell, M. J., Willey, J. C.Anal. Biochem. 213, 277 (1993)

4. Fasco, M. J., Treanor, C. P., Spivack, S., Figge,H. L., Kaminsky, L. S. Anal. Biochem. 224,No. 1 (1995, in press)

5. Schwartz, H. E., Ulfelder, K. J. Anal. Chem. 64,1737 (1992)

6. Srinivasan, K., Girard, J. E., Williams, P., Roby,R. K., Weedn, V. W., Morris, S. C., Kline, M. C.,Reeder, D. J. J. Chromatogr. 652, 83 (1993)

* PCR is covered by U.S. patents owned byHoffman-La Roche Inc.

XXXX-XXXX-XX-X © 1994 Beckman Instruments, Inc. Printed in U.S.A. on recycled paper.

Worldwide Offices: Africa, Middle East, Eastern Europe (Switzerland) (22) 994 07 07. Australia (61) 02 816-5288. Austria (2243) 85656-0.Canada (800) 387-6799. China (861) 5051241-2. France (33) 1 43 01 70 00. Germany (49) 89-38871. Hong Kong (852) 814 7431.Italy (39) 2-953921. Japan 3-3221-5831. Mexico 525 575 5200, 525 575 3511. Netherlands 02979-85651. Poland 408822, 408833.Singapore (65) 339 3633. South Africa (27) 11-805-2014/5. Spain (1) 358-0051. Sweden (8) 98-5320. Switzerland (22) 994 07 07.Taiwan (886) 02 378-3456. U.K. (0494) 441181. U.S.A. 1-800-742-2345.


BECKMAN


N u c l e i c A c i d s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A-1774A

BECKMAN

Quantitative Capillary Electrophoretic Analysis of PCR ProductsUsing Laser-Induced Fluorescence DetectionKathi J. UlfelderBeckman Instruments, Inc.

IntroductionWhile excelling at amplification of genomic mate-rial, the polymerase chain reaction (PCR) suffersfrom difficulties in quantitation of the reaction prod-ucts. One reason for this is the method currentlyused for separation and detection of PCR products:slab gel electrophoresis in conjunction with autorad-iography. In addition, the method cannot be easilyautomated. Recently, capillary electrophoresis (CE)has been used successfully for the separation ofDNA restriction fragments and PCR products withhigh reproducibility and efficiency.(1-3) With the in-troduction of laser-induced fluorescence (LIF), thesensitivity of CE has increased a thousand-fold,(4,5)

which is required if one wishes to compete with auto-radiography for low-level detection. Given these ad-vances in CE methodology, CE was used as themethod for separation and quantitation of PCR prod-ucts.

In previous studies,(6,7) RNA has been reversetranscribed and PCR amplified, with the resultingDNA products analyzed by CE-LIF. As a first steptowards quantitation of the PCR products in un-known samples, it was possible to generate a stan-dard curve demonstrating a linear relationship be-tween amount of template and corrected peak area.

AbstractA more sensitive approach to compete with slab gelelectrophoresis and autoradiography for low-leveldetection and quantitation of PCR*-amplified DNAfragments is demonstrated using capillary electro-phoresis (CE) with laser-induced fluorescence (LIF)detection. In the first system, addition of a fluoro-genic intercalator to the buffer was used to identifyand quantify the PCR products. In a modification ofthis technique, fluorescein-labeled primers wereused to create fluorescently-labeled PCR products.The DNA was analyzed by CE using coated capil-laries and a replaceable polyacrylamide buffer. CEdetection was on-line LIF using the 488 nm line ofan argon-ion laser as the excitation source. Quantita-tion of PCR product was possible through the gen-eration of a standard curve showing linearity be-tween the amount of DNA template present prior toamplification and the peak area of resulting PCRproduct. The enhanced resolution and sensitivity ob-served with CE, together with the ease of quantita-tion, make it a powerful alternative to slab gels forthe separation and quantitation of PCR products.

2

Identification of the products based on size was pos-sible through the use of co-injected DNA size stan-dards.

In the present study, two methods of CE fluo-rescence detection and quantitation are compared.Bacteriophage lambda DNA was used in a PCR,producing a 500 base-pair (bp) fragment. The reac-tion product was generated using one primer of theset fluor-labeled, thus producing a fluorescent PCRproduct. The DNA was analyzed by CE using LIFand compared to UV detection. In addition, unla-beled PCR product was also analyzed by CE-LIF,using a fluorescent intercalator in the buffer sys-tem.(5) A curve demonstrating linearity betweenamount of template DNA and corrected peak areacould be generated for both LIF methods.

Materials and Methods

TemplateWhole bacteriophage lambda DNA was obtainedfrom the GeneAmp PCR Reagent Kit (Perkin-Elmer, Norwalk, CT) at a concentration of 1µg/mL.Used as the template, this DNA was titrated from1000 pg to 0 pg.

PrimersBoth primers were synthesized on an Oligo 1000DNA Synthesizer from Beckman at the 30 nmolscale of synthesis. The FAST Synthesis method withfinal detritylation and the UltraFAST Cleavage andDeprotection chemistries were used. The primerswere dried in a SpeedVac (Savant, New York, NY),resuspended in deionized, sterile water, and usedwithout further purification. Final concentration inthe PCR was 0.5 µM (50 pmol per 100 µL reactionvolume).

One primer (#1) was labeled at the 5' end withfluorescein. This primer was prepared in a singlestep by using fluorescein-labeled CED-phosphor-amidite (Glen Research, Sterling, VA) in the lastcoupling step.

Primer #1:5' - FL - GAT GAG TTC GTG TCC GTA CAA CTG GG - 3'

Primer #2:5' - GGT TAT CGA AAT CAG CCA CAG CGC CC - 3'

PCR amplificationThe PCR amplification reaction was performed us-ing 0.2 mL microcentrifuge tubes in a 50 µL reac-tion mixture with no oil overlay on a PE-9600thermocycler (Perkin-Elmer). Taq polymerase and

dNTPs were obtained from Perkin-Elmer and usedat the concentrations recommended in the GeneAmpPCR Reagent Kit.

The buffer was modified as follows:Buffer(Final Composition): 50 mM Tris-HCl, pH 9.0

20 mM (NH4)2SO4

2 mM MgCl20.004% Tween-20

The temperature profile was also modified:Initial Denaturation: 94°C for 2 minutesCycling: 94°C for 45 seconds

50°C for 30 seconds72°C for 1 minute(25 cycles)

Final Extension: 72°C for 5 minutes

Capillary ElectrophoresisAll CE analyses were conducted on either aP/ACE™ 2200 or a P/ACE 5010 from Beckman.Separations were carried out in the reversed polaritymode (anode at detector side). UV detection was at260 nm; LIF detection used an argon-ion lasersource (Beckman) with excitation at 488 nm andemission at 520 nm (for fluorescein) or 530 nm (forEnhanCE intercalator). Data was collected and ana-lyzed using Gold software, version 8.1, from Beck-man with the Molecular Weight option.

Analysis of primers was accomplished using theeCAP ssDNA 100 Kit: a linear polyacrylamide-filledcapillary (100 µm i.d., 37 cm), in a Tris-borate, 7 Murea buffer. Run temperature was set at 30oC. Oligo-nucleotides were injected electrokinetically for 3–5 sat 7 kV following sample resuspension. Separationswere carried out at 300 V/cm.

Analysis of PCR products was conducted usingthe eCAP dsDNA 1000 Kit: a coated capillary(100µm i.d., 47 cm), filled with Tris-borate-EDTAcontaining replaceable linear polyacrylamide. To vi-sualize unlabeled PCR products, the fluorescent in-tercalator EnhanCE was added to the buffer systemat 0.4 µg/mL. Run temperature was set at 20°C.PCR products were injected hydrodynamically for10 s at 0.5 psi, directly following amplification andwithout further sample preparation. This results in a7.8 nL injection of sample into the capillary. Insome analyses, the PCR sample and DNA sizemarkers—200 and 1000 bp fragments combined,1 µg/mL each in water (GenSura, Del Mar, CA)—were injected sequentially and allowed to co-mi-grate in the capillary. Separations were carried outat 200 V/cm.

3

Results and DiscussionFor most applications, synthetic oligonucleotides donot require purification. However, since quantitationof labeled PCR products will depend greatly on thedegree of primer labeling, purity assessment ofprimer #1 was performed. CE was carried out underidentical conditions using either UV or LIF detec-tion. Figure 1 shows a comparison of the two detec-tion methods. At first inspection by UV (Figure 1A),there appear to be many failure sequences, espe-cially one resembling a large N-1-mer. By LIF, itappears that most of the failure sequences do notfluoresce and therefore are not labeled (Figure 1B);however, there is still present the large N-1-mer. Infact, this fragment is not a failure sequence butrather the N-mer labeled with a diastereomer of thefluorescein phosphoramidite. A positional isomer offluorescein in the phosphoramidite may also accountfor additional fluorescent peaks.

In Figure 2, a 500 bp product was generatedwith 500 pg lambda template DNA and using thefluorescein labeled primer, but was detected by UVfor comparison to LIF methods. The electrophero-gram showed many peaks, including dNTPs, prim-ers, and primer–dimer.

In an electropherogram, the area under the peakfor a particular DNA fragment can be correlated tothe quantity of that fragment. Correction of peakarea becomes critical when one realizes that thepeak area of a DNA fragment is related to its resi-dence time in the detector. Slower migrating (large)fragments will remain in the detector window longerthan a faster migrating (small) fragment, generatinga larger area that is not representative of the truequantity of DNA passing through the detector. TheGold algorithm that corrects for area compensatesfor this anomaly by normalizing analyte velocity.This corrects for differences in fragment mobility.

12 14 16 18 20 22 24 26 28

A

0.030 AU/FS

B

150 AU/FS

Figure 1. CE separation of labeled primer.

10 12 14 16 18 20 22 24

00E+00

.00E-03

.00E-03

.00E-03

.00E-03dNTPs

primers andprimer-dimer 500 bp

Figure 2. CE-UV separation of the PCR product generated from 500 pg of template lambda DNA.

4

The corrected peak area for a DNA fragmentcan be compared to a standard concentration curveto determine absolute amount. Alternatively, a ratioof the unknown’s peak area relative to that of an in-ternal standard can be determined (e.g., for competi-tive PCR). To develop this standard curve for quan-titation of product in a PCR reaction with an un-known amount of template, a serial dilution oflambda DNA template was performed. In this reac-tion, only unlabeled primers were used, with LIF de-tection through the use of a fluorogenic intercalator.The replaceable buffer system has been formulatedto contain a “dye,” EnhanCE, which specifically in-teracts with DNA (and RNA). This molecule acts asa mono-intercalator, inserting itself between everytwo base pairs of DNA. Intercalation changes themolecular length, conformation, and charge on theDNA molecules, resulting in a change in electro-phoretic behavior. Moreover, this DNA-dye com-plex will fluoresce at 530 nm when excited by the488 nm line of an argon-ion laser, whereas the dyealone (as well as non-DNA sample components)will not. Figure 3 shows CE-LIF analysis of thePCR products using decreasing amounts of tem-plate. The limit of detection was determined to be10 pg template or ≈190,000 copies. When the resultsobtained were compared to UV detection, LIF withintercalator produced a >100-fold increase in sensi-tivity. In addition, the PCR mixture contained com-ponents such as dNTPs and Taq polymerase, whichare detected by UV methods and can complicatepattern interpretation. These peaks are not detected

by LIF with intercalator since only the complex offluorogenic intercalator + oligonucleotide will fluo-resce.

An estimation of relative molecular size is pos-sible by comparing an unknown sample’s mobilityto a standard size curve (log of base pair number vs.migration time) generated from the mobility of frag-ments from a marker diluted in water. However,since DNA fragments from restriction digests andthe PCR typically are contained in a high salt ma-trix, their mobility will vary depending on samplesalt concentration. Therefore, as with slab gel analy-sis, size determination compared to a standard inwater is not accurate. To correct this, a double injec-tion technique is employed, whereby an unknownand standard are injected sequentially and allowedto co-migrate in the capillary. Gold software cor-rects for any migration time variations from run-to-run by use of reference peaks. Corrected migrationtimes for the co-injected marker fragments generatethe molecular size curve, which is then used to de-termine the relative size of the unknown DNA frag-ment. Using the co-injected 200 and 1000 bp markerfragments as reference, the lambda PCR productwas determined to be 496 bp. Alternatively, an in-ternal standard may be included in the sample (suchas PCR co-amplification of a second target se-quence, or addition of a known quantity of DNA toeach sample).

Peak area for the 500 bp lambda product wascorrected for transit time through the detector byGold software and plotted as a function of the

7

6

5

4

3

2

1

0

0 12 14 16 18 20 22

200 bp500 bp 1000 bp

1000 pg

100 pg

50 pg

10 pg

Figure 3. CE-LIF separation of unlabeled lambda PCR product using EnhanCE intercalator.

5

amount of DNA used in the PCR. A linear relation-ship is observed from 10 to 1000 pg DNA template.PCR with >1000 pg template was not studied (Fig-ure 4).

In Figure 5, the same labeled PCR products asin Figure 2 were analyzed by CE-LIF, in this casedetecting only the fluor from the PCR product. Withincreasing amounts of DNA template, peak heightand area of the product also increased up to a point,whereupon the “plateau effect”—the leveling off of

the rate of PCR amplification—occurs.(8) The limitof detection was determined to be 1 pg templateor ≈19,000 copies. Note the increase in primer andprimer–dimer peaks as template availabilitydecreases. Unfortunately, PCR product size couldnot be determined without the use of similarlylabeled DNA size standards.

Peak area for the labeled 500 bp lambda prod-uct was again corrected for transit time through thedetector and plotted as a function of the amount of

101100

101

102

103

104

102 103

Amount Template (pg)

Cor

rect

ed P

eak

Are

a

R2=0.993

Figure 4. Quantitation of unlabeled lambda PCR product with EnhanCE intercalator. Plot of corrected peak areavs. DNA template amount.

12 17 22

-0.1

1.9

3.9 primer andprimer-dimer

500 bp

1000 pg

100 pg

10 pg

0 pg

Figure 5. CE-LIF separation of labeled lambda PCR product.

6

DNA used in the PCR (Figure 6). A linear relation-ship is observed up to 250 pg DNA template. Inclu-sion of data points with DNA template >250 pgdemonstrates PCR plateauing. When compared toCE-LIF with the use of intercalators in the buffer,CE-LIF of fluor-labeled PCR products shows>10-fold lower detection and quantitation limits. Al-though it seems unusual that a single fluor moleculewould display greater signal that a molecule interca-lated with many fluors, there are three possible ex-planations for this better sensitivity.

First, fluorescence intensity is directly propor-tional to the product of a fluor’s molar extinctioncoefficient (ε) for absorbance and its quantum yield(Φ) for fluorescence (see reference 9 for a review offluorescence). Under the conditions used for PCRproduct separation and detection, both ε and Φ aregreater for a single fluorescein molecule than for theEnhanCE + DNA complex.(10)

Second, although EnhanCE intercalator has ahigh quantum yield when complexed to DNA, it stillexhibits some low background fluorescence. At highDNA concentrations (>1 µg/mL), this backgroundappears negligible; however, at the DNA levels inthe present investigation (producing a signal <1relative fluorescent unit full scale), backgroundfluorescence becomes significant for low-levelquantitation. While it is possible to reduce the dyeconcentration in the buffer in order to decreasebackground, this will affect the signal linearity forthe highest DNA concentrations (>100 µg/mL) dueto insufficient amount of intercalator.

Third, the presence of excess dye in the buffermay cause self-quenching—the reduction in fluores-cence intensity caused by interactions between indi-vidual fluorophores. Again, although signal reduc-tion is insignificant at high DNA concentrations, itcannot be neglected for the DNA levels in thisstudy. Removal of the dye from the buffer andprestaining of samples with a higher affinity interca-lator such as TOTO-1 or YOYO-1(10) requires thatthe sample concentration be known a priori sincethe correct DNA bp:dye ratio is critical for signallinearity. For samples of unknown DNA concentra-tion, this is not a viable option. Because of this re-duction in signal-to-noise ratio, sensitivity may becompromised for low-level DNA, especially whencompared to single fluor molecule detection withoutthe use of intercalators.

SummaryUsing this separation buffer system, together withCE-LIF, no sample preparation is required for CE-LIF analysis. Marker standards may be used in mi-gration time correction for confirmation of peakidentity and size determination. CE-LIF of PCRproducts with the use of intercalators in the buffershows >100 times lower detection and quantitationlimits than CE with UV detection and no added dye.Furthermore, CE-LIF of fluor-labeled PCR productsgives an additional ten-fold greater sensitivity. Thismay be due to background fluorescence or quench-ing of the dye in the buffer. However, linearity ofthe standard curve for both CE-LIF detection methodssuggests its feasibility for automated PCR quantitation.

10-1 100100

101

102

101 102 103

Amount Template (pg)

Cor

rect

ed P

eak

Are

a

R2=0.997

Figure 6. Quantitation of labeled lambda PCR product. Plot of corrected peak area vs. DNA template amount.

7

AcknowledgmentI wish to thank my colleague, Christian Oste, for hisassistance in primer synthesis, PCR optimization,and for many helpful discussions.

References1. Schwartz, H. E., Ulfelder, K. J. Analysis of

DNA restriction fragments and polymerasechain reaction products towards detection of theAIDS (HIV-1) virus in blood. J. Chromatogr.559, 267-283 (1991)

2. Ulfelder, K. J., Schwartz, H. E., Hall, J. M.,Sunzeri, F. J. Restriction fragment length poly-morphism (RFLP) analysis of ERBB2 oncogeneby capillary electrophoresis. Anal. Biochem.200, 260-267 (1992)

3. Landers, J. P., Oda, R. P., Spelsberg, T. C.,Nolan, J. A., Ulfelder, K. J. Capillary electro-phoresis: a powerful micro-analytical techniquefor biologically-active molecules.BioTechniques 14, 98-111 (1993)

4. Schwartz, H. E., Ulfelder, K. J. Capillary elec-trophoresis with laser-induced fluorescence de-tection of PCR fragments using thiazole orange.Anal. Chem. 64, 1737-1740 (1992)

5. Ulfelder, K. J. Capillary electrophoresis ofdsDNA fragments with UV and laser-inducedfluorescence detection. Beckman ApplicationInformation Bulletin A-1748, Beckman Instru-ments, Inc., Fullerton, CA (1993)

6. Rossomando, E. F., White, L., Ulfelder, K. J.Capillary electrophoresis: separation and quan-titation of RT-PCR products from polio virus.J. Chromatogr. B 656, 159-168 (1994)

7. Lu, W., Han, D.-S.,Yuan, J., Andrieu, J.-M.Multi-target PCR analysis by capillary electro-phoresis and laser-induced fluorescence.Nature 368, 269-271 (1994)

8. Siebert, P. D. Quantitative RT-PCR: Methodsand applications Book 3. Clonetech Laborato-ries, Inc., 1993.

9. Guilbault, G. G. General aspects of lumines-cence spectroscopy, in Practical Fluorescence,pp. 1-40. Edited by G. G. Guilbault. NewYork: Marcel Dekker, Inc., 1990.

10. Haugland, R. P. Nucleic acid stains, in Molecu-lar Probes: Handbook of fluorescent probesand research chemicals, pp. 221-228. Edited byK. D. Larison. Eugene, OR: Molecular Probes,Inc., 1992.

*PCR is covered by U.S. patents owned by Hoffmann-La Roche Inc.


Worldwide Offices: Africa, Middle East, Eastern Europe (Switzerland) (22) 994 07 07. Australia (61) 02 816-5288. Austria (2243) 7292164.Canada (800) 387-6799. China (861) 5051241-2. France (33) 1 43 01 70 00. Germany (49) 89-38871. Hong Kong (852) 814 7431.Italy (39) 2-953921. Japan 3-3221-5831. Mexico 525 575 5200, 525 575 3511. Netherlands 02979-85651. Poland 408822, 408833.Singapore (65) 339 3633. South Africa (27) 11-805-2014/5. Spain (1) 358-0051. Sweden (8) 98-5320. Switzerland (22) 994 07 07.Taiwan (886) 02 378-3456. U.K. (01494) 441181. U.S.A. 1-800-742-2345.


BR94-8551-CB-10 © 1994 Beckman Instruments, Inc. Printed in U.S.A. on recycled paper.

BECKMAN


D N A. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A-1748

BECKMAN

Capillary Electrophoresis of dsDNA Fragmentswith UV␣ and␣ Laser-Induced Fluorescence DetectionKathi J. UlfelderBeckman Instruments, Inc.Fullerton, CA

Introduction

Capillary Electrophoresis (CE) can be used as ahigh resolution separation method for the analysisof dsDNA and PCR products.(1-6) With CE, poly-mer buffer solutions, known as “entangled poly-mer networks,”(4) simulate the pore structure of agel and serve as a sieving medium for the DNAfragments. A coated capillary is used in conjunc-tion with the sieving buffer to control electroos-motic flow and enhance separation efficiency.With the P/ACE™ 2100 Series Capillary Electro-phoresis System from Beckman, on-line detectionis feasible with either UV or Laser-Induced Fluo-rescence (LIF). In the latter mode, fluorescent, in-tercalating dyes are added to the run buffer en-abling high sensitivity analysis.(3) In this Applica-tion Information Bulletin, several examples ofdsDNA separations are shown using two recentlyintroduced DNA capillary gel kits available fromBeckman. With these kits, CE of PCR fragmentscan be performed without desalting the samplesusing the pressure injection mode of the P/ACEinstrument.

Experimental Conditions

CE Instrument: P/ACE System 2100 with Gold™

(Version 7.12) SoftwarePolarity: Reversed (anode at detector end)Sample Injection: See text and figure captionsApplied Voltage: 7.4 kVCapillary: Coated, 37 cm (30 cm to detector)

× 100µm (supplied in the eCAP™ dsDNA1000 Kit, P/N 477410) or LIFluor dsDNA1000 Kit (P/N 477407)

Gel Buffer: As supplied in eCAP dsDNA 1000 Kit;in case the LIF detection, buffer including theEnhance fluorescent dye was used

Temperature: 20°CUV Detection: 254 nmLIF Detection: P/ACE LIF System Ar ion 488 nm

laser, 530 nm emission filter

Results and Discussion

Figure 1 shows the enhancement in detectabilitywhen LIF detection is used instead of UV. TheHae III restriction fragments of φX-174 RF DNAare a good reference standard for PCR fragmentscovering the range of 72-1353 bp. Note that inFigure 1A (UV detection) the concentration of the

2

0.009

0.007

0.005

0.003

Orange G

1353

1078

872603

281

310271

234194

118720.001

-0.0018 10.5 13 15.5 18

Time (min)

Abs

orba

nce

(254

nm

)

64

48

32

16

0

8 10 12 14 16 18 20Time (min)

72118

194234

271281

310

603

872

1078

1353

Flu

ores

cenc

e (e

m 5

30 n

m)

Figure 1. Separation of a Hae III restriction digest of φX 174 RF DNA using (A) UV and (B) LIF detection.Samples were diluted in water to a total DNA concentration of 200 µg/mL for UV detection; 10 µg/mL for LIFdetection. Injection was by pressure for ten seconds. Buffer systems were the same, except for the addition ofEnhance for LIF detection.

A B

sample was 20× higher than that in Figure 1B (LIFdetection). Depending on the conditions, LIF pro-vides 2-3 orders of magnitude higher sensitivitythan UV detection. The small shift towards longermigration times in Figure 1B is due to the bindingof the intercalating fluorescent dye to the DNAfragments. It can be observed that in both cases(i.e., with and without the dye) the 271 bp and 281bp fragments are baseline resolved, exemplifyingthe high resolution feasible with CE.

The separation shown in Figure 2 covers awider bp range and demonstrates the kit’s abilityto also partially resolve DNA fragments greaterthan 1000 bp. The sample was composed of a 1-kbDNA ladder (concatemer of a 1018-bp fragmentwith sticky ends of four bases) with fragments of1018, 2036, 3054, 4072 bp, etc. Also present is a1636-bp fragment and up to ten other fragments ofless than 1000 bp generated from pBR322. Notethat, in contrast to typical agarose slab gel electro-phoresis patterns, the 506 bp and 517 bp frag-ments are baseline resolved with CE.

Figure 3 shows the analysis of a sample con-taining a 97-bp dsDNA fragment amplified byPCR. In the case of the UV detection (Figure 3A),a pronounced group of peaks can be seen in the 8-10 minute elution range. This is due to the pres-ence of PCR reaction products in the sample, i.e.,primers, primer-dimers and deoxyribonucleotidetriphosphates (dNTPs). Figure 4 shows the elec-tropherogram of each of the four dNTPs. The peakheight ratio of the 97-bp peak vs. the dNTP groupis very different when LIF detection is utilized(Figure 3B). The fluorescent dye appears to inter-act preferentially with dsDNA making the signalof the 97-bp PCR fragment relatively larger. As inthe case ofFigure 1, a much larger signal-to-noise ratio canbe observed in the LIF trace compared to the UV.It should also be noted that the PCR sample wasdirectly run by CE (using the pressure injectionmode) without any sample preparation (e.g., de-salting by ultrafiltration). Electrokinetic injectionmay be useful for desalted samples or when theDNA concentration is very low.(5)

3

75 134154

201220

298 344396

506517

1636

10182036

3054

4072

5090

0.0017

0.0013

0.0009

0.0005

0.000110.8 13.3 15.8 18.3

Time (min)

Abs

orba

nce

(254

nm

)

Figure 2. Separation of a 1-kb DNA ladder under UV detection conditions. Sample was at 50 µg/mL total DNAconcentration in water and was injected electrokinetically for ten seconds at 40 V/cm.

0.0015

0.0009

0.0003

-0.00037 9.5 12

Time (min)

14.5

97 bp product

Abs

orba

nce

(254

nm

)

8

6

4

2

0

6 8 10 12 14

Time (min)

97 bp product

Flu

ores

cenc

e (e

m 5

30 n

m)

Figure 3. PCR product analysis using (A) UV and (B) LIF detection. Sample was amplified for 30 cycles (targetsequence 97 bp) and pressure-injected for ten seconds with no further pretreatment. Buffer systems were the sameexcept for the addition of Enhance for LIF detection.

A B

without further sample manipulation. Using theEnhance fluorescent dye, LIF detection of DNAallows 2-3 orders of magnitude improvement indetectability compared to UV detection.

Conclusion

A new “polymer network” gel kit is applicable todsDNA fragments. Pressure injection as well aselectrokinetic injection are feasible. Pressure in-jection allows direct analysis of PCR products

0.0035

0.0025

0.0015

0.0005

-0.00055 7.5 10 12.5 15

C

T

A

G

250 µM each component

Time (min)

Abs

orba

nce

(254

nm

)

Figure 4. Separation of dNTPs (dATP, dGTP, dCTP, dTTP) found in PCR reaction mix. Equal volumes of eachcomponent were combined and diluted to 250 µM with water. Injection was by pressure for ten seconds. UV buffersystem conditions were used.

Worldwide Offices: Africa, Middle East, Eastern Europe (Switzerland) (22) 994 07 07. Australia (61) 02 816-5288. Austria(2243) 85656-0. Canada (905) 387-6799. China (861) 5051241-2. France (33) 1 43 01 70 00. Germany (49) 89-38871. HongKong (852) 814 7431. Italy (39) 2-953921. Japan 3-3221-5831. Mexico (52)5 264 0667. Netherlands 02979-85651. Poland408822, 408833. Singapore (65) 339 3633. South Africa (27) 11-805-2014/5. Spain (1) 358-0051. Sweden (8) 98-5320.Switzerland (22) 994 07 07. Taiwan (886) 02 378-3456. U.K. (0494) 441181. U.S.A. 1-800-742-2345.

BM93-3080-CB-10 © 1993 Beckman Instruments, Inc. Printed in U.S.A. on recycled paper.

Beckman Instruments, Inc. • Bioanalytical Systems Group • 2500 Harbor Boulevard, Box 3100 • Fullerton, California 92634-3100

Sales: 1-800-742-2345 • Service: 1-800-551-1150 • TWX: 910-592-1260 • Telex: 678413 • Fax: 1-800-643-4366

BECKMAN

References1. Heiger, D. N., Cohen, A. S., Karger, B. L.

J. Chromatogr. 516, 33 (1990)

2. Guttman, A., Cohen, A. S., Heiger, D. N.,Karger, B. L. Anal. Chem. 62, 137 (1990)

3. Schwartz, H. E., Ulfelder, K. J. Anal. Chem. 64,1737 (1992)

4. Grossman, P. D., Soane, D. S. J. Chromatogr.559, 257 (1991)

5. Schwartz, H. E., Ulfelder, K., Sunzeri, F. J.,Busch, M. P., Brownlee, R. G. J. Chromatogr.559, 267 (1991)

6. Ulfelder, K. J., Schwartz, H. E., Hall, J. M.,Sunzeri, F. J. Anal. Biochem. 200, 260 (1992)

P r o t e i n s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

BECKMAN

IntroductionThe ability to monitor intermolecular disulfidebond (dimer) formation in biological systems is ofparamount importance for two reasons: 1) in manypeptides and proteins, disulfide bonds are neces-sary for biological activity; 2) where disulfidebond formation is undesirable, gentle oxidation ofa peptide can be performed without generation ofthe dimer which would form under stronger condi-tions. In both cases, the ability to rapidly measuredisulfide bond formation is required. Typically,colorimetric assays are used to determine theseprocesses. However, while giving qualitative “yes/no” answers, they are limited with respect toquantitative information and the type of bond for-mation (i.e., homo- or hetero-dimer). Capillaryelectrophoresis (CE) is a relatively new analyticaltechnique capable of resolving subtle differencesbetween proteins and peptide conformations(1,2).In this Application Information Bulletin, the util-ity of CE to monitor peptide dimer formation isdemonstrated.

Monitoring Disulfide Formationwith P/ACE Capillary Electrophoresis

Robert P. Oda, Jane A. Liebenow, T. C. Spelsberg, and James P. LandersMayo Clinic, Rochester, MN 55905

Experimental ConditionsCE instrument: P/ACE™ System 2050 with Gold™

(version 7.11) software

Polarity: Normal (cathode at detector end)

Capillary: Uncoated, 57 cm (50 cm to detec-tor) x 50 µm i.d.

Temperature: 28°C

Run buffer: 20 mM sodium citrate or 50 mMsodium phosphate, pH 2.5

Applied voltage: 25 kV

Injection: pressure, 3 s

CE run method 3 column volumes rinse withsequence: run buffer, pressure injection, sepa-

ration, 5 column volumes rinsewith 0.1 M NaOH, 5 column vol-umes rinse with run buffer

Detection: UV, 200 nm

Ntc primary CFLGIPFAEPPVGSRRFMPPEPstructure: KRPWSGVL

Ctc primarystructure: TFQTNPDGTIQFRC

Samplepreparation: See reference 3 for details


A-1757

2

Results and DiscussionTwo synthetic peptides were used to evaluate theutility of CE for the separation of peptide monomersfrom their disulfide-linked dimers. One peptide (mo-lecular mass 3369), termed N-terminal cysteine orNtc peptide, consisted of a 30-mer which containedresidues 32-60 of mouse mRNA acetylcholineesterase plus an N-terminal cysteine. The other pep-tide (molecular mass 1629) was called Ctc peptide,and consisted of a 14-mer with 13 residues from aproprietary protein and one C-terminal cysteine resi-due. The disulfide-linked homo-dimers of these pep-tides were generated under controlled air oxidizingconditions(3) and subsequently purified. It was foundthat the purified homo-dimers electrophoresed with

faster migration times than their correspondingmonomers. Mixtures of monomers and dimers werenext subjected to strong oxidizing (with H2O2) andreducing conditions (with dithiotreitol—DTT). Theresults of the CE analysis of Ntc peptide are shownin Figure 1. Reducing a 2:1 mixture of dimer/mono-mer with 1 mM DTT leaves only 4% of the Ntcdimer peak (Figure 1, lower left panel). Likewise,oxidizing a 1:2 dimer/monomer mixture with 0.015%H2O2 reduces the monomer peak by 90% (Figure 1,lower right panel). The conversion between the Ntcmonomeric and dimeric forms is rapid and virtuallycomplete within 30 min. In the experiment of Figure1, CE analysis was commenced 2-3 min after addingthe reagent to the dimer/monomer mixture.

Monomer Dimer

No Treatment

+ DTT + H2O2

Dimer Monomer

1 5 10 151 5 10 15

Time (minutes)

Abs

orba

nce

(200

nm

)

Figure 1. Oxidization and reduction of an Ntc monomer/dimer mixture. A 1:2 mixture of monomer/dimer is reducedin the presence of 1 mM DTT (left panels) while a 2:1 mixture monomer/dimer is oxidized in the presence of 0.015%H2O2 (right panels). Separation was carried out in 20 mM citrate buffer, pH 2.5. Bar represents 0.002 AU. Arrowindicates the sulfonic acid derivative.

3

The time-course of peptide dimerization moni-tored by CE is shown in Figures 2 and 3. The dimer-ization process is temperature-dependent.

Figure 2 shows that approximately 32% of theNtc monomer (peak at 8.2 min) was converted to thedimer (peak at 7.2 min) after 2 h of incubation at27oC (Panel C). Approximately the same conversionrate was observed after 24 h incubation at 4oC, as

shown in panel B. The purified Ntc dimer was sub-jected to the same oxidizing conditions as the mono-mer; however, in this case no decrease in peak heightis shown (see right hand panels of Figure 2), sup-porting the argument that the loss in peak height iscaused by the dimerization and not to a degradationprocess.

A D

B E

C F

Monomer

2 hours, 27°C

24 hours, 4°C

0 hours

Dimer

1 5 10 15 1 5 10 15

Time (min)

Abs

orba

nce

(200

nm

)

Figure 2. Oxidization of Ntc peptide. Purified Ntc monomer and dimer were dissolved in water (2 mg/mL), analyzedimmediately (0 h) and after incubation at 27°C for 2 h and 4°C for 24 h. Bar represents 0.005 AU.

4

Figure 3 shows the time-course of CE analysiswith Ctc peptide. The dimerization process of Ctcpeptide is markedly slower than that observed withNtc peptide. After 8 h under mild oxidizing condi-tions, only negligible conversion to the dimer (indi-cated with an arrow in Figure 3) was observed. Aswas the case with the Ntc peptide, the Ctc dimer hasa faster electrophoretic mobility than its monomer.

The two forms are baseline resolved by CE. To in-duce substantial oxidation, 0.015% H2O2 was added.It can be seen that another 8 h of incubation caused a59% conversion of the monomer to the dimer. It ap-pears that the rate of dimerization is influenced bythe peptide structure (size, amino acid compositionand sequence) and/or the location of the cysteineresidues (N- vs. C-terminal).

1 5 10 15


Time at 27°C

immediate

3 h 28 min

8 h 02 min

immediate

2 h 29 min

8 h 07 min

H2O2

Abso

rban

ce (2

00 n

m)

Figure 3. CE time-course analysis of the Ctc dimerization process. Ctc peptide (1 mg/mL in water) was incubated at27°C and analyzed at 0 min (immediately), 3 h 28 min, and 8 h 2 min. Hydrogen peroxide was added and analysiscarried out at 0 min (immediately), 2 h 29 min, and 8 h 7 min. Separation was carried out in 20 mM citrate buffer,pH 2.5. Bar represents 0.005 AU.

5

Evidence for hetero-dimer formation is pre-sented in Figure 4. A mixture of the purified Ntc andCtc monomers was incubated at 27°C and analyzed at0 min (panel A), 2 h 46 min (panel B), 5 h 31 min(panel C) and 11 h (panel D). Substantial Ntc:Ntchomo-dimer formation can be observed in panel B,

whereas Ctc:Ctc dimerization is relatively slow. Thepeak at 8.6 min, appearing between the Ntc monomerand dimer, is the Ntc:Ctc hetero-dimer. Only after 11h (panel D) has a definable amount of the Ctc homo-dimer been formed. This result would be expected inview of the slow dimerization rate of the Ctc peptide.

Time (minutes)

Abs

orba

nce

(200

nm

)

1 5 10 15 1 5 10 15

A Ntc

Ctc

C

DB

Ntc:Ctc

Ctc:Ctc

Ctc:CtcCtc:Ctc

Ntc:Ntc

Figure 4. Co-oxidization of the Ctc and Ntc peptides. A mixture of the purified Ntc and Ctc monomers wasincubated at 27°C and analyzed at 0 min (Panel A), 2 h 46 min (Panel B), 5 h 31 min (Panel C) and 11 h 1 min(Panel D). Separation was carried out in 20 mM citrate buffer, pH 2.5. Bar represents 0.005 AU.


Beckman Instruments, Inc. • Bioanalytical Systems Group • 2500 Harbor Boulevard, Box 3100 • Fullerton, California 92634-3100Sales: 1-800-742-2345 • Service: 1-800-551-1150 • TWX: 910-592-1260 • Telex: 678413 • Fax: 1-800-643-4366

Worldwide Offices: Africa, Middle East, Eastern Europe (Switzerland) (22) 994 07 07. Australia (61) 02 816-5288. Austria(2243) 85656-0. Canada (800) 387-6799. China (861) 5051241-2. France (33) 1 43 01 70 00. Germany (49) 89-38871. HongKong (852) 814 7431. Italy (39) 2-953921. Japan 3-3221-5831. Mexico 525 575 5200, 525 575 3511. Netherlands 02979-85651.Poland 408822, 408833. Singapore (65) 339 3633. South Africa (27) 11-805-2014/5. Spain (1) 358-0051. Sweden (8) 98-5320.Switzerland (22) 994 07 07. Taiwan (886) 02 378-3456. U.K. (0494) 441181. U.S.A. 1-800-742-2345.

BECKMAN

ConclusionCE appears to be an excellent tool to monitor thedimerization process of proteins and peptides. Asthis study with two model peptides has demon-strated, CE can potentially resolve the monomericand dimeric forms of a species, as well as dis-criminate between homo- and hetero-dimers. WithCE, time-course analysis of the oxidation of apeptide can be easily automated yielding quantita-tive information on dimerization kinetics.

References1. Schwartz, H. E., Palmieri, R. H., Brown, R.

Separation of Proteins and Peptides by Capil-lary Electrophoresis. Capillary Electrophoresis:Theory and Practice, pp. 201-253. Edited byP. Camillieri. CRC Press, Boca Raton, 1993.

2. Palmieri, R. H., Nolan, J. Protein CapillaryElectrophoresis: Theoretical and ExperimentalConsiderations for Method Development. Cap-illary Electrophoresis: A Practical Approach,pp. 313-362. Edited by J. Landers. CRC Press,Boca Raton, 1993.

3. Landers, J. P., Oda, R. P., Liebenow, J. A.,Spelsberg, T. C. J. Chromatogr. (in press)

APPLICATION INFORMATION

Biotechnology products, especially glycoproteinpharmaceuticals, provide many challenges to boththe analytical and process chemist. Most of thesechallenges are due to the complexity and heterogene-ity of proteins in general. The manufacture, release,and characterization of a recombinant biopharmaceu-tical product require many analytical tools.

Flatbed isoelectric focusing (IEF) has been usedin biotechnology to monitor the consistency, stabilityand semi-quantitative levels of charged isoforms ofprotein products. Even though this technology islabor intensive, time consuming and semi-quantita-tive, it has become one of the most routine analyticalmethods in protein analysis.1-5

Capillary isoelectric focusing (cIEF) extends theutility of IEF into an automated and quantitative for-mat and clearly has an important role in analyticalbiotechnology. In our laboratory we have used cIEFin combination with cation ion-exchange chromatog-raphy and mass spectrometry not only to assesscharge isoform heterogeneity but also to determinethe identity and the quantity of isoforms.

Single Amino Acid Variants

Isoforms generated from the differential processingof carboxy (C)-terminal lysine (Lys) and arginine(Arg) residues have been reported in proteins isolat-ed from mammalian cell culture.6 In this paper wedemonstrate the use of cIEF to separate and identifysingle amino variants of the recombinant humanmonoclonal antibody Ab1.

We also briefly examine the impact of differentproduction methods on these variants by comparingAb1 produced both by the fermentation of Chinesehamster ovary cells (CHO) and by transgenic goats (G).

In addition to separating the isoforms by isoelec-tric point, we have employed several enzymatictreatments to assist us in determining the identity ofthe isoforms. By using enzymes which cleave at spe-cific sites we can generate evidence to help identifythese isoforms as C-terminal Lys variants, differingfrom 0 to 2 Lys residues. In this experiment carboxy-peptidase C (CPB) was used to remove the C-termi-nal Lys, while papain was used to digest the anti-body into Faband Fc fragments.

ExperimentalApparatus and Chemicals

All separations were performed on the P/ACE™

MDQ System (Beckman Coulter, Inc., Fullerton,CA). The eCAP™ neutral capillary (Beckman CoulterPart No. 477441), with dimensions of 50 µm (i.d.) x31 cm (20 cm to detector), is used to eliminate elec-troosmotic flow. NaH2PO4, NaCl, hydroxypropyl-methycellulose (HPMC) and tetramethylethylene-diamine (TEMED) are from Sigma Chemical Co. (St.Louis, MO). All enzymes are from BoehringerMannheim (Indianapolis, IN). CPB enzymatic reac-tion: 0.1 mg/mL protein was mixed with 0.05 mgcarboxypeptidase B and incubated at 37ºC for twohours. Anolyte, catholyte and mobilizer are fromBioRad Laboratories (Hercules, CA). The Pharmalyte8-10.5 is from Pharmacia Biotech (Piscataway, NJ),and Bio-Lyte 3-10 is from BioRad Laboratories.

C a p i l l a r y E l e c t r o p h o r e s i s

A-1861A

USING CIEF TO CHARACTERIZE RECOMBINANT HUMANMONOCLONAL ANTIBODIES

L. C. Santosa,1 I. S. Krull,1 and K. L. Grant21 Department of Chemistry, Northeastern University, Boston, MA2 Process Development, BASF Bioresearch Corp., Worcester, MA

2

All recombinant human IgGs (CHO Ab1 and GAb1—BASF Bioresearch Corp., Worcester, MA)were diluted with HPLC-grade water to a concen-tration of 0.25 mg/mL for the cIEF analysis.

cIEF Conditions

cIEF was performed with “normal” polarity (cathodenearest detector). Detection was with UV at 280 nm.The anolyte, catholyte, and mobilizer were 20 mMphosphoric acid, 40 mM sodium hydroxide and azwitterionic solution, respectively. The ampholytesolution was a mixture of diluted Pharmalyte pH 8-10,Bio-Lyte 7-9 and Bio-Lyte pH 3-10 containing 0.4%TEMED (v/v) and 0.2% HPMC (w/v). The dilutionfactor was 20. The mixing ratio (v/v) was 8:1:1.

The pI standard markers (pI 10.1, 8.4 and 7.9)were diluted 1:20 with HPLC-grade water. Samplesfor cIEF consisted of 70 µL of 0.25 mg/ml mAbwith 100 µL ampholyte solution plus 4 µL of eachpI standard marker, as above.

cIEF capillary preconditioning consisted ofrinsing first with HPLC-grade water for 2 minutesat 20 psi, followed by a rinse with 20 mM H3PO4for 2 minutes at 20 psi, and then another water rinsefor 2 minutes. Sample was introduced by filling theentire capillary using positive pressure at 20 psi.The focusing electric field was 580 V/cm for 8 min-utes, followed by a mobilization field of 645 V/cmfor 28 minutes or until all of the proteins of interestand standard markers migrated past the detector.

Results and Discussion

The BASF human recombinant mAb (Ab1) hasC-terminal Lys isoforms, 0-Lys, 1-Lys and 2 Lys.This antibody was first run on conventional gel IEFand three bands were observed (data not shown). TheC-terminal 0-Lys isoform had the lowest pI, while theC-terminal 2-Lys isoform had the highest pI value.

The cIEF electropherogram of the CHO Ab1 hasthree peaks, due to Ab1 (Figure 1). The peak migrat-ing at approximately 21 minutes represents the 2-Lysvariant with a C-terminal lysine on both heavychains. The second isoform with only 1 C-terminalLys migrated a little later while the major componenthaving no C-terminal Lys residues migrated third.

The “standard” markers act as internal calibra-tors to allow us to automatically calculate theisolectric points of our isoforms. We calculatedthe pI for 0-C-terminal Lys to be pH 8.85, 1-C-ter-minal Lys to be pH 8.98 and the 2-C-terminal Lysto be pH 9.11. These isoelectric points are all inagreement with their expected values: the 2-Lys

is the most basic with the greatest number of Lysresidues, while the others have lower pI values,being less basic. The precision of this determinationis very good, with same-day reproducibility of theanalytes’ isoelectric point being typically 0.1%R.S.D, and day to day and lot to lot reproducibilityless than 1% R.S.D (Tables 1, 2 and 3).

cIEF is highly quantitative, unlike gel elec-trophoresis—which is only semi-quantitative.Quantitative analysis of the three isoforms basedon peak area percent were as follows: 71.9% 0-Lys,21% contained 1-Lys, while the 2-Lys isoformaccounted for 7.1%. This data agreed very wellwith a cation exchange chromatography assay thatwe had developed in parallel. (Figure 2).

Compared to CHO Ab1 isoforms, there aremore transgenic goat Ab1 (G Ab1) isoforms thanCHO Ab1 isoforms (Figure 3). The most significantdifference is the increased acidic peaks, which webelieve is due to sialic residues on the G Ab1.

Figure 1. cIEF analysis of CHO Ab1 for the deter-mination of isoelectric points (pI). The pIs are auto-matically calculated and annotated on the peak.

3

Enzymatic DigestionPapain is a proteolytic enzyme known to split anti-body molecules into two separate and identical Fabfragments, each with one antigen-binding site andone Fc fragment. The fragments can be purified orseparated by Protein A. If the isoforms are due toC-terminal Lys variation, then they should be pre-sent on the Fc portion of the antibody, rather thanthe Fabfragments. Our results demonstrate that Fabregion shows one peak and the Fc region elutes asthree peaks which are clearly generated from the Fcregion of Ab1 (Figure 4).

CPB is known to cleave Lys and Arg residuesfrom the C-termini of proteins. Therefore, if the iso-forms are indeed single C-terminal Lys variants, CPBtreatment should collapse the three variants into onethat co-migrates with the 0-Lys variant. Indeed, thethree protein peaks collapse into the most acidic peak,which co-migrates with the 0-Lys variant (Figure 5).

Figure 2. A plot of the percentages of antibody C-terminal lysine isoforms quantitated by both CIEX-HPLC and cIEF assays (n=8, y-error bar is SD).

Figure 3: Comparison of the cIEF-charged isoformdistribution in CHO Ab1 and G Ab1.

Figure 4: cIEF analysis of CHO Ab1 after digestionwith the proteolytic enzyme Papain. The isoforms ofAb1 appear on the Fc region, not on the Fab region.

Figure 5: Comparison by cIEF analysis of CHO Ab1before and after digestion with Carboxypep-tidase B(CPB). CPB removes the C-terminal Lysine residues,collapsing the three Ab1 peaks into one major peak.

4

Table 2: Summary of the overall reproducibility of cIEF analysis and pIdeterminations for the CHO derived antibody Ab1 from different days.

These three samples were all taken from the same lot #802.

Table 3: Summary of the overall reproducibility of cIEF analysis and pIdeterminations for the CHO derived antibody Ab1 from different batches

analyzed on different days.Ten separate batches are indicated (ID), allanalyzed on different days, with average values for all lots reported.

Table 1. Summary of the overall reproducibility of cIEF analysis and pIdeterminations for the CHO derived antibody Ab1 from different injectionson the same day.Also indicated are the average pI values of each peak in

the sample and standard deviation (SD) and percent relative standard devi-ation (%R.SD).These three samples were all taken from the same lot #604.

5

ConclusionsWe have been able to develop a very repro-ducible cIEF method for determining the pI val-ues and quantifying the levels of charged iso-forms present in any typical recombinant mono-clonal antibody. In combination with proteinsequencing, CIEX and LC/MS have been ableto confirm that the charged isoforms seen hereare indeed due to C-terminal Lys variants.Capillary electrophoresis (CE) is clearly a valu-able tool for the automation of isoelectric focus-ing, providing much better reproducibility andquantitation than traditional gel methodology.We plan to extend the CE studies further byexamining the microheterogeneity of antibodyglycosylation, using APTs derivatization of theglycans and CE-LIF detection.

References1. Gordon, A. H. Electrophoresis of Proteins

in Polyacrylamide and Starch Gels.Amsterdam: North-Holland/AmericanElsevier, 1969; Righetti, P. G. IsoelectricFocusing: Theory, Methodology, andApplications, in Laboratory Techniques inBiochemistry and Molecular Biology,T. S.Work and R. H. Burdon, General Editors.Amsterdam: Elsevier Biomedical, 1983.

2. Fundamentals of Protein Biotechnology,edited by S. Stein. New York: MarcelDekker, 1990.

3. Bhown, A. S. Protein/Peptide SequenceAnalysis: Current Methodologies.BocaRaton, FL: CRC Press, Inc., 1988.

4. Bollag, D. M., Rozycki, M. D., Edelstein,S. J. Protein Methods, Second Edition.New York: Wiley-Liss, 1996.

5. Protein Purification Protocols, edited by S.Doonan, Totowa, NJ: Humana Press, 1996.Methods in Molecular Biology,Volume 59.

6. Harris, R. J. Processing of C-terminal lysineand arginine residues of proteins isolatedfrom mammalian cell culture. J.Chromatogr. A, 705,129 (1995)


Developing innovative solutions in genetic analysis, drug discovery, and instrument systems.

Beckman Coulter, Inc.• 4300 N.Harbor Boulevard,Box 3100 • Fullerton,California 92834-3100Sales:1-800-742-2345 • Service:1-800-551-1150 • Telex:678413 • Fax:1-800-643-4366 • www.beckmancoulter.com

Worldwide Bioresearch Division Offices: Australia (61) 2 9844-6000 Canada (905) 819-1234 China (86) 10 6515 6028 Eastern Europe, Middle East, Africa (41) 22 994 07 07 France 01 49 90 90 00 Germany (89) 35870-0 Hong Kong (852) 2814 7431 / 2814 0481 Italy 02-953921 Japan 03-5404-8359 Mexico 525-559-16-35 Netherlands 0297-230630 Singapore (65) 339 3633 South Africa (27) 11-805-2014/5 Spain 91 728 7900 Sweden 08-98 53 20 Switzerland 0800 850 810 Taiwan (886) 2 2378 3456 U.K. 01494 441181 U.S.A. 1-800-742-2345.

B99-3222-CB-5 © 1999 Beckman Coulter, Inc. Printed in U.S.A.on recycled paper.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A-1744

BECKMAN

Guy G. Yowell, Steven D. Fazio, Richard V. VivilecchiaSandoz Pharmaceuticals CorporationE. Hanover, NJ 07936

IntroductionRecent advances in capillary isoelectric focusing(CIEF) offer the potential of replacing conventionalslab gel isoelectric focusing for the analysis of pro-teins. CIEF depends on the formation of a stable pHgradient within the capillary. The need to control orreduce the electroosmotic flow (EOF) during the fo-cusing process has led to the development of capil-lary wall modifications (1-4). The applications ofCIEF to recombinant tissue plasminogen activatorglycoforms (5) and other proteins (6, 7) have beendescribed. However, these CIEF methods consistedof a two-step process: the first was focusing the pro-teins in the capillary; the second required a changeof either the cathode or anode buffers with a salt so-lution to mobilize the proteins past the detectionwindow. Improvements and optimization of the CIEFseparation parameters have included the use oftetramethylethylenediamine (TEMED) to adjust thepH range as described by Yao-Jun and Bishop (8),the use of alternate mobilization agents, and the ad-dition of non-ionic surfactants to minimize proteinprecipitation as described by Zhu et al. (9). Theseimprovements still required a salt mobilization of thefocused bands for detection. This mobilization stepand a lack of stable capillary wall chemistries havemade CIEF unattractive for routine testing.

The Analysis of a Recombinant Granulocyte Macrophage ColonyStimulating Factor (GM-CSF) by Capillary Isoelectric Focusing

Recently, Mazzeo and Krull (10, 11) described aone-step CIEF method in which the EOF was used tomobilize the focused protein zones. Improved sepa-rations were later demonstrated with a commerciallyavailable coated capillary (12). This method elimi-nated the need for salt mobilization. These modifica-tions have advanced the CIEF technique from a basicinvestigational tool to a readily automated method tostudy proteins in the pharmaceutical industry. Thisprocedure has been shown to be successful for theanalysis of a recombinant granulocyte macrophagecolony stimulating factor (GM-CSF) to quantitate im-purities and to study deamidation.

ExperimentalThe octyl-bonded, CElect1 H150 capillaries werepurchased from Supelco, Inc. (Bellefonte, PA). ThePharmalyte2 3-10 ampholyte mixture was purchasedfrom Pharmacia LKB Biotechnology (Piscatawary,NJ). The hydroxypropylmethylcellulose (HPMC),4000 CPs, was purchased from Sigma Chemicals(St. Louis, MO). The TEMED was purchased from

1 CElect is a registered trademark of Supelco, Inc.2 Pharmalyte is a registered trademark of Pharmacia/LKBBiotechnology.

A p p l i c a t i o n␣ ␣ ␣ ␣ ␣ I n f o r m a t i o n ␣

P r o t e i n

2

Bio-Rad Laboratories (Hercules, CA). The capillaryelectrophoresis experiments described were carriedout on a Beckman P/ACE™ 2050 system (Fullerton,CA).

The capillary isoelectric focusing was accom-plished using a method described by Mazzeo andKrull (12). For the CIEF analysis, the Beckmancartridge was fitted with the coated capillary(Supelco) with a total length of 47 cm (40 cm tothe detector) and an internal diameter of 50 µm.The volume of the capillary was 923 nanolitersand the mass of GM-CSF loaded was 23 nano-grams. The capillary was rinsed for 1 hour with0.5% (w/v) HPMC followed by water for 5 min-utes. The GM-CSF lyophilizate was reconstitutedwith 1.0 mL of water. It was then diluted to 50µg/mL with water and again diluted 1:1 with a 2×concentrate of the CIEF-ampholyte mixture whichconsisted of 970 µL of deionized water, 800µL of1% HPMC, 200 µL of Pharmalyte 3/10, and 30 µLof TEMED. The final running concentration ofeach of the components of the CIEF-ampholytemixture was 0.2% HPMC, 2% Pharmalyte 3-10,and 0.75% TEMED. The capillary was rinsed for 4minutes with 10 mM phosphoric acid and then theentire capillary was filled with the sample prepa-ration for 2 minutes using the high-pressure rinsecapability of the P/ACE instrument. The runningfield strength was 300 V/cm at 14.1 kV and thepolarity was reversed (positive at detector end).Detection was carried out at 280 nm; the capillarycartridge was maintained at a temperature of23°C. The catholyte, 20 mM NaOH, was placed atthe inlet and the anolyte, 10 mM phosphoric acid,was placed at the outlet. The total analysis timewas 14 minutes.

Results and DiscussionPrinciple of the Method

Initial work by Mazzeo and Krull (10, 11) withuntreated fused-silica capillaries demonstrated thata one-step isoelectric focusing procedure was pos-sible. Addition of HPMC to the sample matrix re-sulted in a reduction of electroosmotic flow whichallowed the focused protein bands to migrate pastthe detector window. Both the focusing and mobi-lization steps occur simultaneously, but the focus-ing step occurs more rapidly than mobilization.

Mazzeo and Krull (12) improved this methodby employing reverse polarity (thus switching theanodic and cathodic reservoirs), using an octyl-bonded capillary (which further reduced EOF),

and adjusting the amount of TEMED. TEMEDacts as a blocking agent, allowing the proteins tofocus past the detection window. Using this con-figuration, the TEMED blocked the 40 cm regionfrom the inlet to just past the detector window.The proteins focused in the 20-cm region past thedetector window and the EOF moved in the re-verse direction towards the inlet. The proteins thendrifted past the detector window, resulting inanalysis times of typically less than 10 minutes.

The improved CIEF configuration describedabove was used in the present bulletin to quanti-tate the GM-CSF dosage forms as shown in theelctropherogram in Figure 1. Adjusting theTEMED concentration allowed over 40 cm (fromthe inlet to just past the detector) of a 47-cm capil-lary to be blocked. The pH gradient from 3 to 10was located in the 7 cm region from the detectorto the outlet. The proteins focused in the 7-cm sec-tion of the capillary from the detector window tothe outlet. The reduced electroosmotic flow in theoctyl-bonded capillary allowed the focused pro-teins, GM-CSF and human serum albumin (HSA),to drift back towards the inlet reservoir past thedetector window. The shortest capillary length ob-tainable from the detector to the outlet is a func-tion of the manufacturer’s instrument design. Itappears that a short capillary section from the de-tector to the outlet (i.e., 7 cm in P/ACE versus20 cm used in the reference) is advantageous sincethis results in a shorter analysis time for a givenelectric field strength. In addition, the proteins areexposed less to the capillary wall during the mobi-lization process, resulting in a minimal protein-wall interaction. Since the pH gradient is forcedinto a smaller section of the capillary (i.e., 7 cmversus 20 cm), the protein focuses in a narrowerband, which results in lower detection limits (≈ 2µg/mL for the selected protein standard).

Linearity

The linear range for GM-CSF was observed from10 µg/mL to 60 µg/mL using a 47 cm capillary.GM-CSF concentrations above 60 µg/mL showedanomalous peaks which may represent precipita-tion (9) of the GM-CSF in the capillary. Concen-trations above 250µg/mL showed no peaks at alland again may be due to protein precipitation. Therange of protein concentrations which may beused in CIEF vary from protein to protein andmust, therefore, be optimized. Detection can beachieved in CIEF at 280 nm because of the highconcentration of the focused bands (lower wave-

3

0.025

0.020

0.010

0.000

-0.0050.50 4.00 8.00 12.00 16.00

0.023

0.020

0.010

0.000

-0.003

GCSF (Neupogen)

GM-CSF

HSA

Time (min)

Cha

nnel

A (

Abs

orba

nce)

Cha

nnel

B (

Abs

orba

nce)

Figure 1. A CIEF electropherogram of GM-CSF andG-CSF (Neupogen) showing different pI values as amethod for product identification

GM-CSF. Assignment of a pI value to a protein orestimation of pIs for degradation products can beachieved much more rapidly using CIEF than con-ventional slab gel isoelectric focusing as previ-ously mentioned (10). A reference standard curveis shown in Figure 2 and the corresponding elec-tropherogram in Figure 3.

Reproducibility

The reproducibility of the standard migrationtimes was less than 0.5% relative standard devia-tion for 8 injections, and the resolution was be-tween 0.07 and 0.18 pI units (Table 1).

10

9

8

7

6

5

46 8 10 12 14 16 18

Isoe

lect

ric P

oint

(pI

)

Migration Time (minutes)

Chymotrypsinogen, pI 9.0

Myoglobin, pI 7.2

Myoglobin, pI 6.8

Carbonic Anyhydrase II, pI 5.9

Lactoglobulin A, pI 5.1

Carbonic Anhydrase I, pI 6.6

Figure 2. Standard protein calibration curveshowing migration time versus pI

400

350

300

250

200

150

100

504 6 8 10 12 14 16 18 20

Time (min)

mV

α-C

hym

otry

psin

igen

(pI

9.0

)

Myo

glob

in (

pI 7

.2)

Myo

glob

in (

pI 6

.8)

Car

boni

c A

nhyd

rase

I (p

I 6.6

)

Car

boni

c A

nhyd

rase

II (

pI 5

.9)

β-La

ctog

lobu

lin A

(pI

5.1

)

Figure 3. A CIEF electropherogram of a proteinstandard mixture

Table 1. Resolution as a Function of pH*

Proteins between ResolutionpI Ranges (in pI Units)9.3 - 7.2 0.187.2 - 6.8 0.106.8 - 6.6 0.076.6 - 5.9 0.075.9 - 5.1 0.08

* Standard protein mixture calculated assuming baselineseparation of proteins.

3 Neupogen is a registered trademark of Amgen.

lengths are not recommended due to of the absor-bance of the ampholytes). When GM-CSF at 50µg/mL is injected into a 47 cm capillary, the fo-cused band has a concentration of 11.4 mg/mL.Above these concentrations, protein precipitationoften occurs due to the fact that the proteins are attheir pI, the point of least solubility. The CIEFmethod concentrates the original sample of 25 µg/mL to 5.7 mg/mL, which is 228 times the originalsample concentration.

The amount of TEMED used to extend thepH range as described by Yao-Jun and Bishop (8)must also be considered, as well as the range ofampholytes and the concentration of HPMC in thedevelopment of the method. Migration time vari-ability from capillary to capillary was observedand is caused by differences in EOF. This mayrepresent different degrees of capillary wall modi-fication achieved by the manufacturer. Improve-ments in coating procedures and/or the use of dif-ferent mobile phase additives may reduce thisvariability. Also, the purity of ampholytes influ-enced the resolution and resulted in the appear-ance of additional peaks.

Applicability

Although CIEF is being used to check GM-CSFconcentration of dosage forms, it also offers theopportunity to distinguish recombinant proteins ofdifferent pI values. Quick identifications can beaccomplished as shown in Figure 1 which distin-guishes recombinant granulocyte colony stimulat-ing factor (G-CSF), trade name Neupogen,3 from

References

1. Hjerten, S., Liao, J. L. Protides Biol. Fluids 34,

727-730 (1986)

2. Hjerten, S., Zhu, M. D. J. Chromatogr. 346,265-270 (1985)

3. Hjerten, S., Elenbring, K., Kilar, F., Liao, J.L., Chen, A. J. C., Siebert, C. J., Zhu, M. D.J. Chromatogr 403, 47-61 (1987)

4. Kilar, F., Hjerten, S. J. Chromatogr. 480,351-357 (1989)

5. Yim, K. W. J. Chromatogr. 559, 401-410(1991)

6. Wehr, T., Zhu, M., Rodriguez, R., Burke, D.,Duncan, K. Am. Biotechnol. Lab, 8, 22-29(1990)

7. Zhu, M., Hansen, D. L., Burd, S., Ganknon, F.J. Chromatogr. 430, 311-319 (1989)

8. Yao-Jin, G., Bishop, R. J. Chromatogr. 234,459-462 (1982)

9. Zhu, M., Rodriguez, R., Wehr, T.J. Chromatogr. 559, 479-488 (1991)

10. Mazzeo, J. R., Krull, I. S. Anal. Chem. 63,2852-2857 (1991)

11. Mazzeo, J. R., Krull, I. S. J. Chromatogr.606, 291 (1992)

12. Mazzeo, J. R., Krull, I. S. Methods 4,205-212 (1992)


Beckman Instruments, Inc. • Bioanalytical Systems Group • 2500 Harbor Boulevard, Box 3100 • Fullerton, California 92634-3100Sales: 1-800-742-2345 • Service: 1-800-551-1150 • TWX: 910-592-1260 • Telex: 678413 • Fax: 1-800-643-4366

Worldwide Sales Offices: Africa, Middle East (Switzerland, Nyon) (22) 994 07 07. Australia , Gladesville (61) 02 816-5288. Austria , Vienna (2243) 85656-0. Canada, Mississauga (800) 387-6799. France, Gagny (33) 1 43 01 70 00. Germany,Munich (49) 89-38871. Hong Kong, Aberdeen (852) 814 7431. Italy , Milan (39) 2-953921. Japan, Tokyo 3-3221-5831.Mexico, Mexico City (52)5 264 0667. Netherlands, Mijdrecht 02979-85651. Poland, Warszawa 408822, 408833.Singapore (65) 339 3633. South Africa, Johannesburg (27) 11-805-2014/5.Spain, Madrid (1) 358-0051. Sweden, Bromma (8) 98-5320. Switzerland, Nyon(22) 994 07 07. Taiwan, Taipei (886) 02 378-8000. U.K., High Wycombe (0494)441181. U.S.A. 1-800-742-2345.

BECKMAN

Problems

In our experience, the CIEF methods have specificproblems that are protein dependent. For example,it was observed that, after analyzing GM-CSF,other protein standards did not separate in thesame capillary as they had previously: the migra-tion times were longer and the efficiency was sub-stantially less. GM-CSF may be adsorbing to thecapillary wall, thereby changing the wall charac-teristics. Designating individual capillaries for aparticular method or protein may prevent thisproblem.

ConclusionCIEF has worked well in our laboratory for allproteins tested thus far, yielding high efficiencyand resolution comparable to slab gel isoelectricfocusing. Compared to slab gel isoelectric focus-ing, CIEF offers reduced analysis times, directtransfer of data to a computer database, and is 10times more sensitive, on a concentration basis,than Coomassie-stained gels. The one-step mobili-zation method utilizing moderate EOF is simple,readily automated, and reproducible.


P r o t e i n. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A-1771

BECKMAN

Hemoglobin Analysis by Capillary Isoelectric Focusing (CIEF)J. M. HempeDept. of Pediatrics, LSU School of Medicineand Children’s Hospital, New Orleans

IntroductionHemoglobins (Hb) are polypeptide tetramers consist-ing of two pairs of unlike globin chains (e.g., α, β, γ,δ), each of which is covalently linked to four differ-ent heme groups (ferroprotoporphyrin IX). Over 95%of normal adult human hemoglobin is Hb A (α2β2),with Hb A2 (α2δ2) and Hb F (α2γ2) present as minorconstituents. Many congenital and acquired hemato-logic disorders exist. Congenital defects in hemoglo-bin synthesis include both abnormal production ofnormal globin chains (e.g., α and β thalassemia syn-dromes), and the production of abnormal or variantglobin chains (e.g., Hb S and more than 600 otherstructurally different human hemoglobins). Accurateidentification and quantitation of normal and abnor-mal hemoglobin variants is thus clinically importantfor the diagnosis of hemoglo-binopathies and man-agement of many diseases. For example, determina-tion of Hb S level is important for diagnosis of sicklecell disease and as a follow-up to transfusiontherapy. Quantitation of normal hemoglobins like HbA2 and Hb F can provide important diagnostic infor-mation since the levels of these minor constituentsare frequently altered in patients with thalassemiasyndromes. In addition to congenital hemoglobino-pathies, post-translational hemoglobin modifications

are also clinically important, including Hb A1c, aglycosylated form of Hb A that increases in diabetes,and is widely used to monitor long-term glycemiccontrol in diabetic patients.

To meet these varied diagnostic needs, mostclinical laboratories employ a battery of different he-moglobin assays. This includes electrophoresis atboth alkaline and acid pH for common normal andabnormal Hb variants, ion exchange or affinity chro-matography for Hb A2 and Hb A1c, alkali denatur-ation for low levels of Hb F, and/or isoelectric focus-ing (IEF) for other abnormal variants. The advent ofcomputerized HPLC operation and data collectionhas increased Hb analysis by ion-exchange HPLCwhich is useful for a wider variety of Hb than mostother methods. Although the remarkable resolvingpower of IEF has long been recognized, the labor-in-tensive and time-consuming operation of classicalgel IEF has precluded its routine use in most labora-tories. Capillary electrophoresis (CE) is a rapidlyevolving and versatile analytical electrophoresis withwidespread application for the analysis of many dif-ferent types of analytes by a variety of separationmodes, including capillary isoelectric focusing(CIEF). CE offers high resolution and automatedsampling and data collection of HPLC, plus the

2

added advantages of low sample volume require-ments, and low reagent use and cost. Our laboratoryhas recently demonstrated the utility of CIEF for Hbanalysis.(1) Using different CIEF configurations,similar work has been presented by Zhu(2) andThormann(3) and co-workers. The purpose of thepresent experiments was to evaluate CIEF for use inthe routine clinical analysis of hemoglobin variants.

Experimental

Capillary Isoelectric Focusing

The CIEF method was described in detail byNolan(4) in another Beckman Application Infor-mation Bulletin (A-1750). All analyses were con-ducted using a Beckman P/ACE 2200 equippedwith UV detection and Gold software. All sepa-rations were performed in a J &W (Folsom, CA)50-µm i.d. × 27 cm DB-1 coated capillary. Thecapillary temperature was maintained at 20°C.Cathode and anode solutions were 20 mM NaOHand 100 mM H3PO4 in 0.4% methylcellulose, re-spectively. Prior to each assay, the capillary wasfilled with solution containing 0.4% methylcellu-lose and 2% Pharmalyte, pH 6–8. Whole blood (10µL, or enough to give 5–7 mg total Hb/mL), waslysed in 200µL hemolyzing solution (EDTA, 5mmol/L; KCN, 10 mmol/L) then introduced intothe capillary by low-pressure injection (20 sec).Sample constituents were focused for 3 min at 30kV, then eluted past the detector (UV absorbanceat 415 nm) under simultaneous low pressure andvoltage. Normal and abnormal hemoglobin con-trols were obtained from Bio-Rad (Hercules, CA)and Isolab (Akron, OH). Patient samples were ob-tained from Children’s Hospital of New Orleans.Duplicate samples were also sent to a commercialreference laboratory for correlation of assay re-sults. Linear regression was used to calculate Hbisoelectric points (pI) based on the pI and migra-tion times of known Hb in control samples. Hb inunknown samples were identified based on calcu-lated pI. Quantitation was based on peak detectionand integration of UV absorbance at 415 nm, withvalues expressed as percent of total Hb (peak areafor individual Hb divided by total peak area).

ResultsThe CIEF method employed in our present workprovides high resolution of Hb variants differingin pIs by as little as 0.02 pH units with run timesof less than 15 min. The method uses coated capil-

laries in conjunction with run buffers containingampholytes and methylcellulose. In this CIEF con-figuration, the electroosmotic flow isnegligable.(1, 4) After focusing, the protein zonesare eluted from the capillary by pressure mobiliza-tion under high voltage, thereby maintaining theirhigh efficiency separation. Figures 1 through 4 ex-emplify the high resolving power of the currentmethod for Hb analysis in clinical samples.

In Figure 1, the Hb patterns of normal bloodand that of a patient with sickle cell trait are com-pared. Since Hb A represents about 95% of totalHb, an expanded view was selected to highlightthe minor constituents. Note that Hb S was re-solved as a distinct peak from normal Hb. Table 1shows the data of the quantified Hb variants in thetwo samples. Hb A is significantly lower in thepatient’s blood (61.7 vs. 97.3%), while Hb S is ab-sent in the normal sample. In both cases, Hb F waspresent in quantities of less than 0.5%.

Figure 1. Hemoglobin constituents in normal blood,and blood from a patient with sickle cell trait. Note thatHb S was resolved as a distinct peak from normal Hb.

Table 1. Hb Analysis of Blood from a Patientwith Sickle Cell Trait Compared to a Normal

Sample

% Total HbNormal Sickle Cell Trait

Hb A2 2.7 3.1Hb S – 35.2Hb F < 0.5 < 0.5Hb A 97.3 61.7

Figure 2 compares the electropherograms of apatient with hemoglobin S/C disease with a normalsample. Hb C can be identified and is well sepa-rated from other major Hbs. The migration time ofHb C is also different from Hb E, which migrates

3

similarly to Hb A2 (data not shown). As with mostother Hb assay methods, Hb A2 cannot be accu-rately quantified in the presence of of Hb C.

Hb A2

Hb C Hb S Hb A

Hb A1c

Hb F

S/C DiseaseNormal

70.000

0.005

0.010

0.015

0.020

8 9 10 11 12 13 14


Abs

orba

nce

(415

nm

)

Figure 2. Normal and Hb S/C disease. Hb C isidentified and separated from other major Hbconstituents.

Diagnosis of β thalassemia as a cause of anemiais facilitated by accurate quantitation of low levels ofminor constituents Hbs A2 and F. Since these cannotbe accurately analyzed at low levels by conventionalHb electrophoresis, most laboratories use minicolumnion-exchange chromatography to quantitate Hb A2,and alkali denaturation to measure low levels of HbF. Figure 3 compares analysis of Hb variant levels innormal blood to the elevated levels of Hb A2 and HbF present in the blood of a subject with β thalassemiaminor. This information, plus evidence of microcyto-sis, hypochromia, and mild, intermediate, or severeanemia are indicative ofβ thalassemia.

Figure 3. Normal and β thalassemia. Microcytosis,hypochromia, elevated Hb A2, and mild anemia arecharacteristic features of β thalassemia minor.

Finally, Figure 4 shows the capability of CIEFto resolve Hb S and Hb D. These variants—whichdiffer in pI by only 0.026 pH unit—cannot beseparated with alkaline electrophoresis. The upperpanels show the electropherograms of two abnor-mal Hb control samples containing Hb S and HbD, respectively. The bottom panel shows the elec-tropherogram resulting from a 1:1 mixture of thesecontrols.

Figure 4. Hb S and Hb D cannot be separated byalkaline electrophoresis, but are differentiated byCIEF. Upper frame: abnormal Hb control containingHb S. Middle frame: abnormal Hb control containingHb D. Bottom frame: equal-volume mixture of thesetwo controls show separate peaks attributable toHb S and Hb D which differ in pI by only 0.026 pHunit.

ConclusionThe presented CIEF method provides high resolu-tion, short run times, and quantitation for theanalysis of Hb variants in clinical samples. Themethod is sensitive enough to measure lowamounts of normal variants and specific enough toidentify many abnormal variants. With P/ACE, au-tomated sampling and detection can be performed.In contrast to conventional IEF methods based on

BECKMAN


BR94-XXXX-SCL-20 © 1994 Beckman Instruments, Inc. Printed in U.S.A. on recycled paper.

Worldwide Offices: Africa, Middle East, Eastern Europe (Switzerland) (22) 994 07 07. Australia (61) 02 816-5288. Austria (2243)85656-0. Canada (800) 387-6799. China (861) 5051241-2. France (33) 1 43 01 70 00. Germany (49) 89-38871. Hong Kong (852) 8147431. Italy (39) 2-953921. Japan 3-3221-5831. Mexico 525 575 5200, 525 575 3511. Netherlands 02979-85651. Poland 408822, 408833.Singapore (65) 339 3633. South Africa (27) 11-805-2014/5. Spain (1) 358-0051. Sweden (8) 98-5320. Switzerland (22) 994 07 07. Taiwan(886) 02 378-3456. U.K. (0494) 441181. U.S.A. 1-800-742-2345.

slab gels, no staining of the protein zones is re-quired. In addition, the automated CIEF method isvery reproducible (1, 4), and involves minimal(microvolume) sample requirements and reagentcost. CIEF has great potential to replace conven-tional Hb assay methods in clinical research labo-ratories.

References1. Hempe, J. M. Clin. Chem (submitted 1994)

2. Zhu, M., Rodriguez, R., Wehr, T., Siebert, C.J. Chromatogr. 608, 225 (1992)

3. Molteni, S., Frischknecht, H., Thormann, W.Electrophoresis 15, 22 (1994)

4. Nolan, J. A. Application Information BulletinA-1750, Beckman Instruments, Fullerton, CA.(1993)

EVALUACIóN FINAL CURSO DE MéTODOS EN BIOTECNOLOGíA

Documents

Transcript of EVALUACIóN FINAL CURSO DE MéTODOS EN BIOTECNOLOGíA