Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques [3 ed.] 9780849333293, 0849333296

Citation preview

Dedication ———————

This undertaking would not have been possible without the support of my family and colleagues. Dedication is, first and foremost, to family—first, my wife, Lianne, and my children, Miranda, Kate, and Ben—who not only recognize my passion for these academic endeavors but also fully encourage me to pursue them. I appreciate the sacrifices they made en route to completion of this undertaking. Second, this laboratory compendium would not exist without the immeasurable effort from the scientists who authored the chapters—they are leaders in their particular fields, both nationally and internationally—I am grateful that they saw value in contributing here. In recognition of their willingness to contribute to this handbook, even in the midst of their already overextended commitments, this book is also dedicated to them.

Contents Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi

Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xiii

Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xv

Part I Fundamentals and Methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

Chapter 1 Introduction to Capillary Electrophoresis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . James P. Landers

3

Chapter 2 Protein Analysis by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . James M. Hempe

75

Chapter 3 Micellar Electrokinetic Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Shigeru Terabe Chapter 4 Capillary Electrophoresis for Pharmaceutical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 Eamon McEvoy, Alex Marsh, Kevin Altria, Sheila Donegan, and Joe Power Chapter 5 Principles and Practice of Capillary Electrochromatography . . . . . . . . . . . . . . . . . . . . . 183 Myra T. Koesdjojo, Carlos F. Gonzalez, and Vincent T. Remcho Chapter 6 Capillary Electrophoresis of Nucleic Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 Eszter Szántai and András Guttman Chapter 7 Analysis of Carbohydrates by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . 251 Julia Khandurina The Coupling of Capillary Electrophoresis and Mass Spectrometry in Proteomics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 Haleem J. Issaq and Timothy D. Veenstra Chapter 8

Chapter 9 Light-Based Detection Methods for Capillary Electrophoresis. . . . . . . . . . . . . . . . . . . 305 Cory Scanlan, Theodore Lapainis, and Jonathan V. Sweedler Chapter 10 Microfluidic Devices for Electrophoretic Separations: Fabrication and Use . . . . 335 Lindsay A. Legendre, Jerome P. Ferrance, and James P. Landers

Part IIA

Capillary-Based Systems: Core Methods and Technologies . . . . . . . . . . . . . 359

Chapter 11 Kinetic Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361 Maxim V. Berezovski and Sergey N. Krylov

vii

viii

Contents

Chapter 12 DNA Sequencing and Genotyping by Free-Solution Conjugate Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381 Jennifer A. Coyne, Jennifer S. Lin, and Annelise E. Barron Chapter 13 Online Sample Preconcentration for Capillary Electrophoresis . . . . . . . . . . . . . . . . . . 413 Dean S. Burgi and Braden C. Giordano Chapter 14 Capillary Electrophoresis for the Analysis of Single Cells: Sampling, Detection, and Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429 Imee G. Arcibal, Michael F. Santillo, and Andrew G. Ewing Chapter 15 Ultrafast Electrophoretic Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445 Michael G. Roper, Christelle Guillo, and B. Jill Venton Chapter 16 DNA Sequencing by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467 David L. Yang, Rachel Sauvageot, and Stephen L. Pentoney, Jr. Chapter 17 Dynamic Computer Simulation Software for Capillary Electrophoresis . . . . . . . . . 515 Michael C. Breadmore and Wolfgang Thormann Chapter 18 Heat Production and Dissipation in Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . 545 Christopher J. Evenhuis, Rosanne M. Guijt, Miroslav Macka, Philip J. Marriott, and Paul R. Haddad Chapter 19 Isoelectric Focusing in Capillary Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 563 Jiaqi Wu, Tiemin Huang, and Janusz Pawliszyn

Part IIB

Capillary-Based Systems: Specialized Methods and Technologies . . . . . 581

Chapter 20 Subcellular Analysis by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 583 Bobby G. Poe and Edgar A. Arriaga Chapter 21 Chemical Cytometry: Capillary Electrophoresis Analysis at the Level of the Single Cell. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 611 Colin Whitmore, Kimia Sobhani, Ryan Bonn, Danqian Mao, Emily Turner, James Kraly, David Michels, Monica Palcic, Ole Hindsgaul, and Norman J. Dovichi Chapter 22 Glycoprotein Analysis by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 631 Michel Girard, Izaskun Lacunza, Jose Carlos Diez-Masa, and Mercedes de Frutos Chapter 23 Capillary Electrophoresis of Post-Translationally Modified Proteins and Peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 707 Bettina Sarg and Herbert H. Lindner Chapter 24 Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 723 Wm. Hampton Henley and James W. Jorgenson Chapter 25 Separation of DNA for Forensic Applications Using Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 761 Lilliana I. Moreno and Bruce McCord Chapter 26 Clinical Application of CE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 785 Zak K. Shihabi

Contents

ix

Chapter 27 Solid-Phase Microextraction and Solid-Phase Extraction with Capillary Electrophoresis and Related Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 811 Stephen G. Weber Chapter 28 CE-SELEX: Isolating Aptamers Using Capillary Electrophoresis . . . . . . . . . . . . . . . 825 Renee K. Mosing and Michael T. Bowser Chapter 29 Microfluidic Technology as a Platform to Investigate Microcirculation . . . . . . . . . 841 Dana M. Spence Chapter 30 Capillary Electrophoresis Applications for Food Analysis. . . . . . . . . . . . . . . . . . . . . . . . 853 Belinda Vallejo-Cordoba and María Gabriela Vargas Martínez Chapter 31 Separation Strategies for Environmental Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 913 Fernando G. Tonin and Marina F.M. Tavares

Part IIIA Microchip-Based: Core Methods and Technologies . . . . . . . . . . . . . . . . . . . . . . 979 Chapter 32 Cell Manipulation at the Micron Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 981 Thomas M. Keenan and David J. Beebe Chapter 33 Multidimensional Microfluidic Systems for Protein and Peptide Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1001 Don L. DeVoe and Cheng S. Lee Chapter 34 Microchip Immunoassays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1013 Kiichi Sato and Takehiko Kitamori Chapter 35 Solvent Extraction on Chips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1021 Manabu Tokeshi and Takehiko Kitamori Chapter 36 Electrophoretic Microdevices for Clinical Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . 1037 Jerome P. Ferrance Chapter 37 Advances in Microfluidics: Development of a Forensic Integrated DNA Microchip (IDChip). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1065 Katie M. Horsman and James P. Landers Chapter 38 Taylor Dispersion in Sample Preconcentration Methods . . . . . . . . . . . . . . . . . . . . . . . . . 1085 Rajiv Bharadwaj, David E. Huber, Tarun Khurana, and Juan G. Santiago Chapter 39 The Mechanical Behavior of Films and Interfaces in Microfluidic Devices: Implications for Performance and Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1121 Matthew R. Begley and Jennifer Monahan Chapter 40 Practical Fluid Control Strategies for Microfluidic Devices . . . . . . . . . . . . . . . . . . . . . . 1153 Christopher J. Easley and James P. Landers Chapter 41 Low-Cost Technologies for Microfluidic Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 1169 Wendell Karlos Tomazelli Coltro and Emanuel Carrilho Chapter 42 Microfluidic Reactors for Small Molecule and Nanomaterial Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1185 Andrew J. deMello, Christopher J. Cullen, Robin Fortt, and Robert C.R. Wootton

x

Contents

Part IIIB Microchip-Based: Specialized Methods and Technologies . . . . . . . . . . . . . . 1205 Chapter 43 Sample Processing with Integrated Microfluidic Systems . . . . . . . . . . . . . . . . . . . . . . . . 1207 Joan M. Bienvenue and James P. Landers Chapter 44 Cell and Particle Separation and Manipulation Using Acoustic Standing Waves in Microfluidic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1229 Thomas Laurell and Johan Nilsson Chapter 45 Optical Detection Systems for Microchips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1253 James M. Karlinsey and James P. Landers Chapter 46 Microfabricated Electrophoresis Devices for High-Throughput Genetic Analysis: Milestones and Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1277 Charles A. Emrich and Richard A. Mathies Chapter 47 Macroporous Monoliths for Chromatographic Separations in Microchannels. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1297 Frantisek Svec and Timothy B. Stachowiak Chapter 48 Microdialysis and Microchip Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1327 Barbara A. Fogarty, Pradyot Nandi, and Susan M. Lunte Chapter 49 Microfluidic Sample Preparation for Proteomics Analysis Using MALDI-MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1341 Simon Ekström, Johan Nilsson, György Marko-Varga, and Thomas Laurell Chapter 50 Implementing Sample Preconcentration in Microfluidic Devices . . . . . . . . . . . . . . . . 1375 Paul M. van Midwoud and Elisabeth Verpoorte Chapter 51 Using Phase-Changing Sacrificial Materials to Fabricate Microdevices for Chemical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1419 Hernan V. Fuentes and Adam T. Woolley Chapter 52 Materials and Modification Strategies for Electrophoresis Microchips . . . . . . . . . . 1441 Charles S. Henry and Brian M. Dressen Chapter 53 Microfluidic Devices with Mass Spectrometry Detection . . . . . . . . . . . . . . . . . . . . . . . . 1459 Iulia M. Lazar Chapter 54 Nanoscale Self-Assembly of Stationary Phases for Capillary Electrophoresis of DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1507 Kevin D. Dorfman and Jean-Louis Viovy Chapter 55 Nanoscale DNA Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1527 Laili Mahmoudian, Mohamad Reza Mohamadi, Noritada Kaji, Manabu Tokeshi, and Yoshinobu Baba Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1543

Foreword The esteemed Chinese philosopher Confucius placed works such as this in succinct context when he wrote “You cannot open a book without learning something.” Each chapter of this book describes remarkable advancements in one of the most controversial and yet, powerful, analytical tools in separation sciences. Even the most accomplished practitioners in separation science will extract new information on methods and applications that can serve unmet needs in science and medicine. Even after more than two-and-a-half decades of research and development, many scientists debate whether capillary electrophoresis (CE) technology has provided the speed, resolving power, peak capacity, sensitivity, robustness and cost reduction promised by the pioneers of CE during its genesis in the 1980s and 1990s. Thousands of researchers, focusing on some of the most widespread and practical analytical goals ever considered (that utilize microscale volumes of sample), have demonstrated the utility of CE on a wide range of applications across many disciplines, such as forensic science, medical diagnostics, pharmaceutical science, biotechnology, and environmental science. The last 7 years have seen CE technology evolve quickly from the research laboratory into practical applications in many fields, and the total body of international CE literature continues to grow at an ever-increasing rate. The first flush of excitement came with the key role that the CE technology played in completion of the Human Genome Project. The project was successfully completed ahead of schedule and at a fraction of the predicted cost. This monumental breakthrough allowed the placement of CE-based DNA sequencers in the most prestigious institutions providing medical-legal testing and in numerous research centers worldwide working on different DNA projects. In addition, an official compendium for the analysis of erythropoietin is now part of a monograph of the European Pharmacopeia. Consequently, CE is now part of the quality control that precedes the release of commercial batches of this critically important drug. Other major biopharmaceutical companies also have official CEbased quality control methods filed with the FDA. In fact, the growth of widely used biotechnology compounds now rivals the growth of conventional pharmaceutical products and, thus, stricter regulations have been applied to biomolecule development due to the complexity of their physicochemical properties. Each contributor to this book has been actively engaged in CE or microchip research and is a leader in the field. Accordingly, each author contributes various aspects of their experience, specifically the detailed descriptions and awareness of the practical problems inherent in their particular area of research and application. Furthermore, contributors provide sound advice on how to overcome practical problems, thus enabling readers to pursue novel applications equipped with much of the practical knowledge and wisdom possessed by the leading researchers in the field. The chapters in this book indicate the extraordinary breadth and scope of the work carried out in capillary electrophoresis and microchip technology over the last few years. The core section on “Fundamentals and Methodologies” contains 10 chapters focused on CE technology. These chapters provide an introduction to CE in general, followed by in-depth descriptions of the most widely used modes of CE and their associated detection methods. Several important applications related to protein, nucleic acid, and carbohydrate research are presented, as well as an overview of the emerging field of fabrication and use of microchips. The next 20 or so specialized chapters continue to describe specific applications and methodologies that can be addressed with capillary systems, including studies on the analysis of cellular and subcellular entities, food analysis, environmental science,

xi

xii

Foreword

clinical and forensic applications, solid-phase microextraction/online sample preconcentration, and extreme resolution CE. This bandwidth of applications is illustrative of the exploding world of microchip systems that perform chemical and biochemical analysis. It is difficult, if not impossible, to determine whether microfabricated analytical devices were predominantly inspired by capillary systems, or by visionaries who recognized the potential of microfabrication techniques (originally designed for microelectronics) to revolutionize chemical, biomolecular, and cellular analysis. Clearly, inspiration has flowed in both directions, and the revolution is upon us. Techniques to exploit the precision, scalability, and unique microscale phenomena enabled by microchip technology now span much of the routine practice of biochemistry and biology, including cell manipulation, protein separations, forensic science, molecular and nanomaterial synthesis, etc. The second part of this book explores core microsystem technologies, such as immunoassays, flow control, device fabrication and reliability. These core technologies are widely used in many applications, such as those outlined in the specialized chapters that conclude this book, and they span everything from on-chip sample preparation to genetic analysis to devices coupled with mass spectrometry. It would appear that the days of categorizing capillary and microchip electrophoresis as narrow endeavors with limited impact are long gone. The confluence of chemistry, biology, and microengineering undoubtedly will continue to produce new tools for scientific discovery, as well as for forensic applications and clinical diagnostics. These latter applications are poised for a technology revolution that will paradigm shift their respective fields. Microdevices that accept and analyze the components of whole blood—ions, proteins, nucleic acids—using integrated sample preparation domains fluidically interfaced with the analytic process are coming to fruition and will soon become available. Moreover, microfluidics offers the unique capability to control and manipulate (sub)nanoliter volumes of liquids in unique, precise, and reproducible ways that were previously unimaginable in capillary systems. This fact, together with innumerable new developments and insight on cell manipulation and analysis, and recent progress in (inexpensive) polymer microfabrication combine to create exciting opportunities for future analytical microsystems that can be applied as disposable diagnostic tools for site-of-analysis testing. It is our hope that readers will find this book as helpful as the last two editions have been. This work is an outstanding contribution to the development of increasingly powerful analytical tools that will continue to have a positive impact on science and society in the years ahead. Norberto A. Guzman, Ph.D. Senior Research Fellow Bioanalysis, Drug Metabolism, and Drug Toxicity Johnson & Johnson Pharmaceutical R&D Raritan, New Jersey Prof. Dr. Albert van den Berg BIOS Lab-on-a-Chip group MESA+ Institute for Nanotechnology University of Twente The Netherlands Matthew R. Begley, Ph.D. Associate Professor Department of Mechanical Engineering Department of Materials Science & Engineering University of Virginia, Charlottesville, Virginia

Editor James P. Landers is professor of chemistry and professor of mechanical engineering at the University of Virginia, as well as an associate professor of pathology at the University of Virginia Health System. Professor Landers received a Bachelor of Science degree in biochemistry with a minor in biomedicine from the University of Guelph in Ontario, Canada, in 1984. He earned his Ph.D. in biochemistry from the same department in 1988. After a one-year postdoctoral fellowship at the Banting Institute at the University of Toronto School of Medicine, he was awarded a Canadian Medical Research Council Fellowship to study cancer biology and diagnostics under Dr. Thomas Spelsberg, a breast cancer biochemist at the Mayo Clinic. He launched and directed Mayo Clinic’s Clinical Capillary Electrophoresis Facility in the Department of Laboratory Medicine and Pathology developing clinical assays based on capillary electrophoretic technology—some are still on-board at Mayo today. Beginning as an assistant professor of analytical chemistry at the University of Pittsburgh in 1997, he forayed into analytical microfluidic systems with the goal of developing the next-generation molecular diagnostics platform. These efforts were bolstered by a move to the University of Virginia, where access to a dedicated class-100 cleanroom for microchip fabrication allowed his group to rapidly prototype microdevices for separations, DNA purification, and DNA amplification. In addition to editing the first two editions of this book, he has authored more than 175 papers and 25 book chapters on receptor biochemistry, capillary electrophoretic method development, microchip fabrication, and integrated microfluidic systems for application in the clinical and forensic arenas.

xiii

Contributors Kevin Altria Research and Development GlaxoSmithKline Harlow, Essex, United Kingdom Imee G. Arcibal Department of Chemistry Pennsylvania State University University Park, Pennsylvania Edgar A. Arriaga Department of Chemistry University of Minnesota Minneapolis, Minnesota Yoshinobu Baba Department of Applied Chemistry and MEXT Innovative Research Center for Preventive Medical Engineering Nagoya University Nagoya, Japan and Health Technology Research Center National Institute of Advanced Industrial Science and Technology Takamatsu, Japan Annelise E. Barron Department of Bioengineering Stanford University Stanford, California David J. Beebe Department of Biomedical Engineering University of Wisconsin Madison, Wisconsin Matthew R. Begley Department of Mechanical and Aerospace Engineering and Department of Materials Science and Engineering University of Virginia Charlottesville, Virginia

Maxim V. Berezovski Department of Chemistry York University Toronto, Ontario, Canada and Campbell Family Institute for Breast Cancer Research University of Toronto Toronto, Ontario, Canada Rajiv Bharadwaj Microfluidics Group Caliper Life Sciences Mountain View, California Joan M. Bienvenue Department of Defense DNA Registry Armed Forces DNA Identification Laboratory Armed Forces Institute of Pathology Rockville, Maryland Ryan Bonn Department of Chemistry University of Washington Seattle, Washington Michael T. Bowser Department of Chemistry University of Minnesota Minneapolis, Minnesota Michael C. Breadmore Australian Centre for Research on Separation Science School of Chemistry University of Tasmania Hobart, Australia Dean S. Burgi Molecular Diagnostics Affymetrix, Inc. Santa Clara, California Emanuel Carrilho Department of Chemistry and Molecular Physics Institute of Chemistry at São Carlos University of São Paulo São Carlos, Brazil xv

xvi

Contributors

Wendell Karlos Tomazelli Coltro Department of Chemistry and Molecular Physics Institute of Chemistry at São Carlos University of São Paulo São Carlos, Brazil

Christopher J. Easley Department of Molecular Physiology and Biophysics Vanderbilt University Medical Center Nashville, Tennesse

Jennifer A. Coyne Department of Chemical Engineering Stanford University Stanford, California

Simon Ekström Department of Electrical Measurements Division of Nanobiotechnology Lund University Lund, Sweden

Christopher J. Cullen Department of Chemistry Imperial College London London, United Kingdom Mercedes de Frutos Institute of Organic Chemistry Madrid, Spain Andrew J. deMello Department of Chemistry Imperial College London London, United Kingdom Don L. DeVoe Department of Mechanical Engineering University of Maryland College Park, Maryland Jose Carlos Diez-Masa Institute of Organic Chemistry Madrid, Spain Sheila Donegan Department of Chemical and Life Sciences Waterford Institute of Technology Waterford, Ireland Kevin D. Dorfman Department of Chemical Engineering and Material Science University of Minnesota Minneapolis, Minnesota Norman J. Dovichi Department of Chemistry University of Washington Seattle, Washington Brian M. Dressen Department of Chemistry Colorado State University Fort Collins, Colorado

Charles A. Emrich Department of Chemistry and Biophysics Graduate Group University of California Berkeley, California Christopher J. Evenhuis Australian Centre for Research on Separation Science School of Chemistry University of Tasmania Hobart, Australia Andrew G. Ewing Department of Chemistry Pennsylvania State University University Park, Pennsylvania Jerome P. Ferrance Department of Chemistry University of Virginia Charlottesville, Virginia Barbara A. Fogarty Tyndall National Institute Lee Maltings Cork, Ireland Robin Fortt Department of Chemistry Imperial College London London, United Kingdom Hernan V. Fuentes Department of Chemistry and Biochemistry Brigham Young University Provo, Utah Braden C. Giordano Nova Research Inc. Alexandria, Virginia

xvii

Contributors

Michel Girard Centre for Biologics Research Health Canada Ottawa, Ontario, Canada

Katie M. Horsman Department of Chemistry University of Virginia Charlottesville, Virginia

Carlos F. Gonzalez Department of Chemistry Oregon State University Corvallis, Oregon

Tiemin Huang Convergent Bioscience Ltd. Toronto, Ontario, Canada

Rosanne M. Guijt Australian Centre for Research on Separation Science School of Chemistry University of Tasmania Hobart, Australia Christelle Guillo Department of Chemistry and Biochemistry Florida State University Tallahassee, Florida András Guttman Horváth Laboratory of Bioseparation Sciences University of Innsbruck Innsbruck, Austria Paul R. Haddad Australian Centre for Research on Separation Science School of Chemistry University of Tasmania Hobart, Australia James M. Hempe Children’s Hospital Research Institute for Children and Department of Pediatrics Louisiana State University Health Sciences Center New Orleans, Louisiana Wm. Hampton Henley Department of Chemistry The University of North Carolina at Chapel Hill Chapel Hill, North Carolina Charles S. Henry Department of Chemistry Colorado State University Fort Collins, Colorado Ole Hindsgaul Carlsberg Laboratory Valby Copenhagen, Denmark

David E. Huber Microfluidics Department Sandia National Laboratories Livermore, California Haleem J. Issaq Laboratory of Proteomics and Analytical Technologies SAIC Frederick Inc. Frederick, Maryland James W. Jorgenson Department of Chemistry The University of North Carolina at Chapel Hill Chapel Hill, North Carolina Noritada Kaji Department of Applied Chemistry and MEXT Innovative Research Center for Preventive Medical Engineering Nagoya University Nagoya, Japan James M. Karlinsey Department of Chemistry Penn State Berks The Pennsylvania State University Reading, Pennsylvania Thomas M. Keenan Department of Biomedical Engineering University of Wisconsin Madison, Wisconsin Julia Khandurina Anadys Pharmaceuticals San Diego, California Tarun Khurana Mechanical Engineering Stanford University Palo Alto, California

xviii

Takehiko Kitamori Department of Applied Chemistry Nagoya University Aichi, Japan and Micro Chemistry Group Kanagawa Academy of Science and Technology Kanagawa, Japan Myra T. Koesdjojo Department of Chemistry Oregon State University Corvallis, Oregon James Kraly Department of Chemistry University of Washington Seattle, Washington Sergey N. Krylov Department of Chemistry York University Toronto, Ontario, Canada

Contributors

Lindsay A. Legendre Department of Chemistry University of Virginia Charlottesville, Virginia Jennifer S. Lin Department of Chemical and Biological Engineering Northwestern University Evanston, Illinois Herbert H. Lindner Division of Clinical Biochemistry Innsbruck Medical University Innsbruck, Austria Susan M. Lunte Departments of Chemistry and Pharmaceutical Chemistry Ralph N. Adams Institute of Bioanalytical Chemistry University of Kansas Lawrence, Kansas

Izaskun Lacunza Institute of Organic Chemistry Madrid, Spain

Miroslav Macka School of Chemical Sciences Dublin City University Dublin, Ireland

James P. Landers Department of Chemistry University of Virginia Charlottesville, Virginia

Laili Mahmoudian Department of Applied Chemistry Nagoya University Nagoya, Japan

Theodore Lapainis Department of Chemistry University of Illinois at Urbana-Champaign Urbana, Illinois

Danqian Mao Department of Chemistry University of Washington Seattle, Washington

Thomas Laurell Department of Electrical Measurements Division of Nanobiotechnology Lund University Lund, Sweden

György Marko-Varga Department of Analytical Chemistry Lund University Lund, Sweden

Iulia M. Lazar Virginia Bioinformatics Institute and Department of Biological Sciences Virginia Polytechnic Institute and State University Blacksburg, Virginia

Philip J. Marriott Australian Centre for Research on Separation Science School of Applied Sciences RMIT University Melbourne, Australia

Cheng S. Lee Department of Chemistry and Biochemistry University of Maryland College Park, Maryland

Alex Marsh Research and Development GlaxoSmithKline Harlow, Essex, United Kingdom

xix

Contributors

María Gabriela Vargas Martínez Universidad Nacional Autónama de México Depto. de Química Cuautitlán, Edo. México, México Richard A. Mathies Chemistry Department University of California Berkeley, California Bruce McCord International Forensic Research Institute Florida International University Miami, Florida Eamon McEvoy Department of Chemical and Life Sciences Waterford Institute of Technology Waterford, Ireland David Michels Amgen Seattle, Washington Mohamad Reza Mohamadi Department of Applied Chemistry Nagoya University Nagoya, Japan Jennifer Monahan Birck Nanotechnology Center Purdue University Lafayette, Indiana Lilliana I. Moreno International Forensic Research Institute Florida International University Miami, Florida Renee K. Mosing Department of Chemistry University of Minnesota Minneapolis, Minnesota Pradyot Nandi Department of Pharmaceutical Chemistry Ralph N. Adams Institute of Bioanalytical Chemistry University of Kansas Lawrence, Kansas Johan Nilsson Department of Electrical Measurements Division of Nanobiotechnology Lund University Lund, Sweden

Monica Palcic Carlsberg Laboratory Valby Copenhagen, Denmark Janusz Pawliszyn Department of Chemistry University of Waterloo Waterloo, Ontario, Canada Stephen L. Pentoney, Jr. Advanced Technology Center Beckman Coulter, Inc. Fullerton, California Bobby G. Poe Chemistry Department University of Minnesota Minneapolis, Minnesota Joe Power Department of Chemical and Life Sciences Waterford Institute of Technology Waterford, Ireland Vincent T. Remcho Department of Chemistry Oregon State University Corvallis, Oregon Michael G. Roper Department of Chemistry and Biochemistry Florida State University Tallahassee, Florida Juan G. Santiago Department of Mechanical Engineering Stanford University Stanford, California Michael F. Santillo Department of Chemistry Pennsylvania State University University Park, Pennsylvania Bettina Sarg Division of Clinical Biochemistry Innsbruck Medical University Innsbruck, Austria Kiichi Sato Department of Applied Biological Chemistry The University of Tokyo Tokyo, Japan

xx

Contributors

Rachel Sauvageot Advanced Technology Center Beckman Coulter, Inc. Fullerton, California

Wolfgang Thormann Department of Clinical Pharmacology University of Bern Bern, Switzerland

Cory Scanlan Department of Chemistry University of Illinois at Urbana-Champaign Urbana, Illinois

Manabu Tokeshi Department of Applied Chemistry Nagoya University Aichi, Japan and Micro Chemistry Group Kanagawa Academy of Science and Technology Kanagawa, Japan

Zak K. Shihabi Department of Pathology Wake Forest University School of Medicine Winston-Salem, North Carolina Kimia Sobhani Department of Chemistry University of Washington Seattle, Washington

Fernando G. Tonin Institute of Chemistry University of São Paulo São Paulo, Brazil

Dana M. Spence Department of Chemistry Michigan State University East Lansing, Michigan

Emily Turner Department of Chemistry University of Washington Seattle, Washington

Timothy B. Stachowiak Department of Chemical Engineering University of California Berkeley, California

Belinda Vallejo-Cordoba Centro de Investigation en Alimentacion y Desarrolla, A.C. Hermosillo, Sonora, Mexico

Frantisek Svec The Molecular Foundry Lawrence Berkeley National Laboratory Berkeley, California

Paul M. van Midwoud Pharmaceutical Analysis Group Groningen Research Institute of Pharmacy University of Groningen Groningen, the Netherlands

Jonathan V. Sweedler Department of Chemistry University of Illinois at Urbana-Champaign Urbana, Illinois Eszter Szántai Institute of Medical Chemistry, Molecular Biology, and Pathobiology Semmelweis University Budapest, Hungary Marina F.M. Tavares Institute of Chemistry University of São Paulo São Paulo, Brazil Shigeru Terabe University of Hyogo Hyogo, Japan

Timothy D. Veenstra Laboratory of Proteomics and Analytical Technologies SAIC Frederick Inc. Frederick, Maryland B. Jill Venton Department of Chemistry University of Virginia Charlottesville, Virginia Elisabeth Verpoorte Pharmaceutical Analysis Group Groningen Research Institute of Pharmacy University of Groningen Groningen, the Netherlands

xxi

Contributors

Jean-Louis Viovy Laboratoire Physicochimie-Curie Section de Recherche, Institut Curie Paris, France Stephen G. Weber Department of Chemistry University of Pittsburgh Pittsburgh, Pennsylvania Colin Whitmore Department of Chemistry University of Washington Seattle, Washington

Adam T. Woolley Department of Chemistry and Biochemistry Brigham Young University Provo, Utah Robert C.R. Wootton Department of Pharmacy and Chemistry Liverpool John Moores University Liverpool, United Kingdom Jiaqi Wu Convergent Bioscience Ltd. Toronto, Ontario, Canada David L. Yang Advanced Technology Center Beckman Coulter, Inc. Fullerton, California

Part I Fundamentals and Methodologies

to Capillary 1 Introduction Electrophoresis James P. Landers CONTENTS 1.1 1.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.1 Why Electrophoresis in a Capillary? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.2 The Family of CE Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.3 Capillary Zone Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.3.1 Instrumentation and CE Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.3.2 Role of EOF in CE Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.3.3 A Description of the Electrophoretic Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.3.4 The Capillary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Method Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.1 Steps in Designing a Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.2 Sample Parameters to Consider . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.3 Separation Parameters to Consider . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.3.1 Electrode Polarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.3.2 Applied Voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.3.3 Capillary Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.3.4 Capillary Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.3.5 Buffers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 Introduction of Sample into the Capillary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.5 On-Capillary Sample Concentration Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.5.1 Sample Stacking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.5.2 Sample Focusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.5.3 Isotachoporetic Sample Enrichment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.5.4 Online Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.6 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.1 Mobility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.1.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.2 Corrected Peak Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.2.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3 Quantity of Sample Introduced into the Capillary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3.1 Hydrodynamic Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3.1.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3.2 Electrokinetic Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3.2.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4 7 7 8 9 9 10 12 19 22 23 23 25 26 26 27 29 29 40 41 41 42 43 43 43 44 50 50 51 52 52 53 53 53 54 54

3

4

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

A.4 Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.4.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.5 Efficiency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.5.1 Example 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.5.2 Example 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.6 Joule Heating. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.6.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3.1 Ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3.2 Small Molecules: Charged and Neutral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3.3 Peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3.4 Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3.5 Nucleic Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

54 55 55 55 55 56 56 56 59 60 60 63 65 68 69

1.1 INTRODUCTION While the term “electrophoresis” was coined in 1909 by Michaelis,1 it was the pioneering experiments of Tiselius in 19372 that first showed the separation of serum proteins—albumin, and α-, β-, and γ -globulins—by “moving boundary electrophoresis”; this provided the first intimation of the potential use of electrophoretic analysis for biologically active molecules. However, this approach to electrophoresis was limited by the incomplete separation of proteins, the relatively large sample volume needed, and the necessity of relatively low electrical fields due to the convection currents generated by Joule heating, even in the presence of dense sucrose solutions. The historical events that followed in electrophoretic method development were paradigm shifting—for historical detail, the reader is referred to two reviews in the journal Electrophoresis, one by Hjertén3 in the late 1980s and the other by Rilbe4 in the mid-1990s. The 1940s saw major efforts directed at improving anticonvective media for zone electrophoresis, while paper electrophoresis was shown to be applicable to the analysis of a wide variety of molecules.5,6 Gels formed from starch7 and, later, agarose8,9 were developed for the analysis of peptides, proteins, and oligonucleotides. In the 1950s, Kolin10 attempted isoelectric focusing (IEF), but the absence of “ampholytes” led to short pH gradients that had poor stability over time. It was in this period that the pioneering work of Hjertén3 laid the groundwork for the capillary electrophoresis (CE)analysis of diverse analytes, ranging from small molecules (inorganic ions, nucleotides) to proteins and viruses. Much of this work was conducted with a functional CE-like instrument, albeit with 3 mm tubes that Hjertén constructed as early as 1959 (see Figure 1.1). It was not until the late 1960s that Vesterberg11 synthesized ampholines to create stable pH gradients that allowed for effective IEF to be accomplished and take its place among the routine molecular tools for biochemical and biomedical analysis. The functionality and acceptance of starch gels was followed by the introduction of acrylamide as a sieving matrix in 1959 where, for the first time, control over pore size and stability was possible.12 Ornstein13 and Davis14 independently introduced disc gel electrophoresis in 1964, which was followed by the introduction of sodium dodecyl sulfate (SDS) as a denaturing agent for protein separation in 1969. Building on 30 years of development, the 1970s bought the concept of stacking gel together with the use of SDS for the separation of the components in T4 phage,15 which laid the groundwork for the seminal work by two-dimensional electrophoresis [combination of IEF separation followed by polyacrylamide gel electrophoresis (PAGE)] that was described by Dale and Latner16 and Macko and Stegemann.17 Stegemann introduced IEF in polyacrylamide gels followed by SDS–PAGE,18 which was developed

Introduction to Capillary Electrophoresis

5

FIGURE 1.1 Photograph of the CE system designed by Hjertén. On the left are stacked the high voltage supply and electronic components for the detector, topped off by a strip-chart recorder. In the center is the carriage with the capillary and the electrode vessels above the immersion bath, while the cooling reservoir flanks it on the right.

into a fine art by O’Farrell19 in 1975. Interestingly, around this time, Virtanen20 followed the work of Hjertén a decade earlier with the use of smaller internal diameter (0.2 mm) tubes, which eliminated convection problems and simplified instrumental design. In numerous ways, all of this work seeded the later advances that included development of sequencing gels in 1977, agarose gels a short time later, and then pulsed-field gel electrophoresis in 1983. It was in the very same year (1983) that Jorgenson and Lukacs21 defined electrophoresis in micron-scale capillaries. The concurrent development of advanced separation technology in the form of high-performance liquid chromatography (HPLC) made speed, high resolution, quantitative results, and automation a reality. This platform saw widespread adoption in industrial and clinical laboratory settings, filling the need where electrophoresis could not to meet the demands of quantitative analysis, preparative isolation, or automation. These advantages proved HPLC effective for the analysis of small molecules, oligonucleotides, peptides, and small proteins, but with the disadvantage that larger analytes (e.g., structural proteins) were problematic. Moreover, more stringent regulations on waste resulted in increased cost for waste disposal, leading to organic solvent waste generated by HPLC considered as a major contributor to operating costs. Micro LC, in both open-tubular and packed formats,22−24 were shown to have promise,25 but technical difficulties in the design of high-pressure, low-volume solvent delivery systems and operational challenges in handling the severe pressure drop in packed column chromatography deterred commercial development. It was with this historical backdrop of increasing demands for high resolution, quantitative precision of biopharmaceuticals, and control of waste management costs that CE arrived on the analytical scene. The pioneering work of Hjertén3 and Virtanen20 preceded the demonstration that CE was shown to be a viable analytical technique by Mikkers et al.26 and Jorgenson and Lukacs.21 CE demonstrated the potential for producing high-resolution separations of biopolymers, as well as smaller pharmaceutical agents, and used miniscule amounts of both sample and reagents. In 1989, a decade after Mikkers et al.26 carried out their defining CE experiments, CE was ready for primetime

6

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

6000 5000 4000 3000 2000 1000 0 1990

1992

1994

1996

1998

2000

2002

2004

2006

Capillary

FIGURE 1.2 Growth of the CE and microchip literature. The ISI Web of Science database was searched with the subject keywords of “CE” on a year-by-year basis beginning in 1983 through March 2007 and plotted as a function of publication year.

as a result of improvements in the sensitivity of detectors, advances in automation technology, and, most importantly, the widespread availability of high-quality narrow-bore capillary silica tubing. Beckman Instruments introduced the scientific community to CE with the commercial launch of the first fully automated CE instrument in the form of the P/ACE™ 2000. The result of sound engineering is that some of these models can still be found in labs today (ours being one of them). Within a few years, research groups throughout the world had expanded the horizons of CE. Hjertén,3 Mazzeo and Krull,27 and Wehr28 were early in the pursuit of improvements in capillary coatings to prevent analyte adsorption to the capillary surface while techniques, for casting polyacrylamide and agarose gels in capillaries were developed and applied to protein29,30 and DNA31−33 separations. These have now evolved into more sophisticated coatings involving phospholipids,34 lipoproteins,35 cholesterol,36 and polyelectrolytes.37 Terabe et al.38,39 pioneered micellar electrokinetic chromatography (MEKC), with the use of micellar solutions to separate neutral and charged species, while IEF was adapted to the capillary format,40−42 using chemical mobilization methods to move the focused protein zones past the detector. In addition Zhu et al.,43 Bruin et al.,44,45 and Ganzler et al.,46 developed non-gel sieving matrices for size-based separation of biopolymers, which began a flurry of activity defining linear polymers for DNA and proteins analysis by CE.47 Theoretical studies have led to understanding the problems associated with the technique, and the methods required to overcome them. Productization of more effective and sophisticated instrumentation has made room for the development of new detectors48−50 and mass spectrometer interfaces.51 Many of these topics are covered in detail in the chapters that follow in this book. Beginning with four publications in 1983, growth of the CE literature was exponential for roughly the first decade (not surprising with such small numbers), linearized from 1993 to 1998, and then slowed (but did not plateau) over the next 5 years (1999–2003) (Figure 1.2). While these types of plots often seem superfluous, the point is this—despite the slowing of growth in the CE field around the turn of the century, interest in CE has not stalled. In fact, there has been modest resurgence since then. Method development still continues, but the application focus has intensified in the last half decade. Diversity is evident by application of the technique to a wide spectrum of analyses in a variety of disciplines ranging from the detection and quantitation of priority pollutants in environmental samples, to the analysis of the components of a single cell, the screening for abnormal proteins or DNA fragments indicative of disease or specific typing an individual. Consistent with the theme of the handbook, this chapter has a practice-oriented focus, with the presentation of select theoretical aspects of CE limited to the basic principles needed for understanding how molecules separate in an applied voltage, and the factors that affect the separation. Appendix 2 provides the reader with a guide to troubleshooting typical problems that may be encountered with CE.

Introduction to Capillary Electrophoresis

7

1.2 CAPILLARY ELECTROPHORESIS 1.2.1 WHY ELECTROPHORESIS IN A CAPILLARY? As introduced in Section 1.1, electrophoresis has been one of the most widely utilized techniques for the separation and analysis of ionic substances. Almost all modes of electrophoresis utilize some form of solid support to prevent convectional distortion of the analyte bands. This has been in the form of paper (high voltage separation of amino acids and other small organic molecules) or, more commonly, polyacrylamide or agarose gels, which have been used extensively for both protein and deoxyribonucleic acid/ribonucleic acid (DNA/RNA) analysis. In light of this, there is little doubt that gel electrophoresis has been an invaluable analytical tool for modern biochemical research. However, despite the ability to resolve the components of complex systems, traditional forms of electrophoresis suffer from several disadvantages, most of which scientists had endured for lack of an effective alternative. Perhaps the most obvious disadvantage is the speed of separation, which is ultimately limited by Joule heating (the heating of a conducting medium as current flows through it). The relatively poor dissipation of Joule heat in slab systems limits their use to low potential electric fields. Moreover, from a methodological perspective, the entire process is a series of cumbersome, time-consuming tasks, from the casting of the gel, preparation and loading of samples, electrophoretic resolution of the ionic/molecular species, to the final stage where the gel is stained, and the results obtained. Other problems include poor reproducibility, particularly with two-dimensional analysis, analyte-dependent differences in staining, which makes quantitative accuracy difficult to achieve (e.g., glycosylated proteins have different dye-binding properties than their unglycosylated counterparts), and the cumbersome methodology involved in modern gel electrophoresis that makes it virtually impossible for the entire electrophoretic procedure to be automated. The use of capillaries as an electromigration channel for separation of a diverse array of analytes, not only presents a unique approach to separation, but is also associated with several advantages over the standard solid supports used during the evolution of electrophoresis. In particular, the physical characteristics of narrow-bore capillaries make them ideal for electrophoresis. Fused-silica capillaries employed in CE typically have an internal diameter (i.d.) of 20–100 µm (375 µm outside diameter, o.d.), lengths of 20–100 cm, and are externally coated with a polymeric substance, polyimide, which imparts tremendous flexibility to a capillary that would otherwise be very fragile (Figure 1.3). The high surface-to-volume ratio of capillaries with these dimensions allows for very efficient dissipation of Joule heat generated from large applied fields. This is illustrated in Table 1.1, which compares the surface-to-volume ratio of a standard analytical slab gel system with standard capillaries used in CE. From this table, it is clear why the slab gel is limited to fields in the range of ≈15–40 V/cm whereas up to 800 V/cm can be applied to a capillary containing the same type of gel matrix shown early on by Dovichi and coworkers.52 As a result of this ability to dissipate heat, electrophoretic separations can easily be performed at up to 30,000 V with the external capillary environmental thermostatted at ambient temperature. In addition to the ability to dissipate Joule heat efficiently, the use of capillaries for electrophoresis is associated with many advantages. With typical electrophoresis capillaries, the small dimensions yield total column volumes in the microliter range, thus requiring the use of only milliliter quantities of buffer. Moreover, adhering to the chromatographic rule of thumb restricting sample volume to 1–5% of the total capillary volume, sample volumes introduced into the capillary are in the nanoliter range (as low as 0.2 nL). As a result, as little as a 5 µL of sample will suffice for repetitive analysis on some commercial instruments. Considering the small reagent (buffer) and sample requirements, as well as the rapid analysis times associated with the application of high fields (30,000 V), it is clear why CE has been applied to, and has provided the solution to, a diverse number of analytical problems.

8

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a)

(b)

375 µm (od)

~5 µm polyimide coat

20–100 µm (id)

(c)

Static Layer (Stern Layer) (~0.1 nm)

Mobile Layer (Outer Helmholtz Plane)

FIGURE 1.3 Diagram of the capillary, its inner surface and the ionic layer. (a) Shows an end view of a capillary with a 375 µm o.d. and a 50 µm i.d. (b) An illustration of the double-ionic layer formed in bare silica capillaries critical in generating endoosmotic flow. (c) A scanning electron photomicrograph of a 375 µm (o.d.) × 75 µm (i.d.) capillary enlarged 170 times. The polyimide coating which grants the capillary flexibility is visible and the “etch” used to break the capillary is clearly seen on the left.

TABLE 1.1 A Comparison of Surface-to-Volume Ratios for an Analytical Slab Gel and a 57 cm Capillary Having a Varied Internal Diameter Surface Area (mm2 )

Volume (µL)

32,200 35.81 89.53 134.3 179.1 258.1

24,150 0.179 1.119 2.518 4.477 17.907

Slab gel (14 × 11.5 × 0.15 cm) 20 mm i.d. capillary 50 mm i.d. capillary 75 mm i.d. capillary 100 mm i.d. capillary 200 mm i.d. capillary

Surface-to-Volume Ratio 1.3 200 80 53 40 20

1.2.2 THE FAMILY OF CE MODES In much the same way that standard gel electrophoretic techniques diversified, so did CE. This has resulted in a family of specialized modes that collectively constitute “CE.” The main modes of CE that have been developed and are presently being exploited include capillary zone electrophoresis (CZE), MEKC, capillary electrokinetic chromatography capillary (CEC), capillary isoelectric focusing (CIEF), capillary gel electrophoresis (CGE), and capillary isotachophoresis (CITP). Over the last decade or so, this latter mode has been primarily used as an on-capillary preconcentration technique.

Introduction to Capillary Electrophoresis

9

TABLE 1.2 Modes Used for Analysis of Various Classes of Analytes CZE Ions Small molecules Peptides Proteins Carbohydrates

MEKC

CEC

CIEF

CGE

Small molecules Peptides

Small molecules Peptides Proteins Carbohydrates

Peptides Proteins

Nucleic acids

CZE is the most universal of the techniques, having been shown to be useful for the separation of a diverse array of analytes varying in size and character. Although a brief description of CZE is given in this chapter, various aspects of the technique will be discussed in a number of chapters to follow in this handbook. The other CE modes are covered exclusively, and in detail, in separate chapters dedicated specifically to the theoretical and practical aspects of those modes, along with examples illustrating applications amenable to that particular mode. Table 1.2 provides a reference table categorizing the modes used for the analysis of various classes of analytes.

1.2.3 CAPILLARY ZONE ELECTROPHORESIS Capillary zone electrophoresis is not only the simplest form of CE, but also the most commonly utilized. Discussion of this mode permits the presentation of a generic design for the instrumentation for CE. The addition of specialized reagents to the separation buffer readily allows the same instrumentation to be used with the other modes mentioned in the previous section: addition of surfactants with MEKC, ampholines for CIEF and a sieving matrix (linear polymers, entangled matrices) for CGE. The discussion on CZE in the following subsections allows for analysis of some of the basic principles governing analyte separation by this technique.

1.2.3.1 Instrumentation and CE Analysis A diagrammatic representation of a CE instrument is presented in Figure 1.4. The basic components include a high voltage power supply (0–30 kV; possibly 60 kV if you are in Jorgenson’s lab), a polyimide-coated capillary with an internal diameter ≤200 µm, two buffer reservoirs that can accommodate both the capillary and the electrodes connected to the power supply, and a detector. As will be discussed later in this chapter, thermostatting of the capillary is critical to efficient and reproducible separations and, hence, some type of capillary thermostatting system should be used. To perform a capillary zone electrophoretic separation, the capillary is filled with an appropriate separation buffer at the desired pH and sample is introduced at the inlet. Both ends of the capillary and the electrodes from the high voltage power supply are placed into buffer reservoirs and up to 30,000 V applied to the system. The ionic species in the sample plug migrate with an electrophoretic mobility (direction and velocity) determined by their charge and mass, and eventually pass a detector where information is collected and stored by a data acquisition/analysis system. Electrophoretic mobility (µ) of a charged molecular species can be approximated from the Debye–Huckel–Henry theory µ = q/6π ηr

(1.1)

10

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Net flow Capillary length (20–100 cm)

Lamp Optics Detector HighVoltage Power (5–30 kV)

Post-injection

Inlet buffer/ sample injection

OutletBuffer

Post-detection

FIGURE 1.4 General schematic of a CE instrument.

where q is the charge on the particle, η is the viscosity of the buffer, and r is the Stokes’ radius of the particle. The mass of the particle may be related to the Stokes’ radius by M = (4/3)πr 3 V , where V is the particle specific volume of the solute. Although one might infer the direct proportionality of mass and radius of the particle, empirical data suggest modifications of Equation 1.1 to allow for the nonspherical shape, the counterion effects, and nonideal behavior of proteins and biological molecules.53 With CZE, the “normal” polarity is considered to be [inlet—(+), detector—(−) outlet] as shown in Figures 1.4 and 1.5. As electrophoresis ensues, the analytes separate according to their individual electrophoretic mobilities and pass the detector as “analyte zones” (hence, the term capillary zone electrophoresis or CZE). The fact that, under appropriate conditions, all species (net positive, net negative, or neutral) pass the detector indicates that a force other than electrophoretic mobility is involved. If the applied field were the only force acting on the ions, net positively charged (cationic) substances would pass the detector while neutral components would remain static (i.e., at the inlet) and anionic components would be driven away from the detector. It is clear that, if this were the case, CE would be of limited use. Fortuitously, there is another force, “electroosmotic flow” (EOF), driving the movement of all components in the capillary towards the detector when under an applied field (and a normal polarity). EOF plays a principle role in many of the modes of CE and most certainly in CZE. This is discussed briefly in the next section. 1.2.3.2 Role of EOF in CE Analysis Electroosmotic flow was first identified in the late 1800s when Helmoltz54 conducted experiments involving the application of an electrical field to a horizontal glass tube containing an aqueous salt solution. Curious about the ionic character of the inner wall and the movement of ions, he found that the silica imparted a layer of negative charge to the inner surface of the tube, which under an applied electric field, led to the net movement of fluid toward the cathode. More than a century later, this phenomenon still plays the fundamental role in CE analysis. Moreover, the importance of the control of EOF has been realized and has become the focus of several research groups. As a continuation of the pioneering work of Helmholtz, the basic principles governing EOF have been evaluated extensively. As shown by the expanded region of the inner wall of a capillary in Figure 1.3, the ionized silanol groups (SIO) of the capillary wall attract cationic species from the buffer. Obviously, the buffer pH will determine the fraction of the silanol groups that will be ionized; understanding the amorphous nature of silica and the pKa range (4–6) associated with the various types of silanol groups55 is key. The ionic layer that is formed has a positive charge density that

Introduction to Capillary Electrophoresis

11

DETECTOR

(Apply voltage) t=0

N– + +N – N + – + – N

Electro osmotic flow

t=1

(+)

–– N + N

(−)

– + N +

– –

+N

t=2

– N + – N+

(+)

– –

N

(−)

+ N

+

t=n

– – –

(+)

–

Migration time = EOF

N

+ +

N

+

+

N

(−)

_

Intensity

+

N

N

Time (min)

FIGURE 1.5 Mobility of charged and uncharged molecules in an applied field.

decreases exponentially as the distance from the wall increases. The double layer formed closest to the surface is termed the “Inner Helmholtz” or Stern” layer and is essentially static. A more diffuse layer formed distal to the Stern Layer is termed the “Outer Helmholtz Plane” (OHP). Under an applied field, cations in the OHP migrate in the direction of the cathode carrying waters of hydration with them. Because of the cohesive nature of the hydrogen bonding of the waters of hydration to the water molecules of the bulk solution, the entire buffer solution is pulled toward the cathode. This EOF or “bulk flow” acts as a pumping mechanism to propel all molecules (cationic, neutral and anionic) toward the detector with separation ultimately being determined by differences in the electrophoretic migration of the individual analytes. The importance of EOF in capillary electrophoretic analysis is highlighted in Figure 1.5. The buffer entering the capillary inlet behind the sample plug is represented by a “graded shading” for illustrative purposes, and is identical to the buffer preceding the sample. As electrophoretic migration occurs, all analytes are swept towards the detector by bulk flow. Provided that the EOF is adequate but not too strong, the respective electrophoretic mobilities of each of the analytes leads to the formation of discrete zones by the time they pass the detector. If the EOF is low, diffusion of the analyte zones could result in substantial band broadening and, under conditions of very low EOF, some of the analytes may not reach the detector within a reasonable analysis time. As previously mentioned, the EOF is pH dependent and can be quite strong. For example, in 20 mM borate buffer at pH 9.0, the EOF is ≈2 mm/s, which translates to a flow of ≈4 nL/s in a 50 µm i.d. capillary. The inclusion of EOF into the calculation of velocity is essential and results in vi = µapp E = (µep µeo )E

(1.2)

where µep is the mobility due to the applied electric potential and µeo is the mobility due to EOF. From a practical perspective, EOF acts as an electric field-driven pump that may be considered analogous to the mechanical pump used in HPLC. While the simplicity of this pumping system has

12

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

obvious advantages over mechanical pumps (e.g., no moving parts, no mechanical wear) one does not have the same degree of control associated with mechanical systems. It is important to have some idea of the magnitude of EOF under the conditions used for a given separation for two reasons. First, under conditions where EOF is very fast, components of the mixture may not have adequate “on-capillary time” for separation to occur. Second, it is useful to know where neutral compounds migrate in the obtained electropherogram. Knowing this, information about the charge character of the sample components is obtained on the basis of whether they migrate faster (cationic) or slower (anionic) than EOF. The EOF marker is also useful as an internal standard for calculating “relative” migration times for the components of the sample from the apparent migration times. Some compounds that adequately serve as neutral markers include dimethyl formamide (DMF), dimethyl sulfoxide (DMSO), and mesityl oxide, although, because of rapid volatilization, the latter compound is of limited use with samples kept at ambient temperature. Typically, a “marker” is a 0.1% solution in water; a 1-s hydrostatic (0.5 psi) sample injection of the 0.1% solution provides an adequate signal with ultraviolet (UV) detection at either 214 or 200 nm. Fluorescent markers (e.g., BODIPY) can serve the same purpose for LIF detection These markers are not of use with low pH separations where most analytes migrate faster than the EOF, which is extremely low. If a peak is still required as an internal standard for correction of migration time, any compound with a fast cathodic electrophoretic mobility will suffice as a “frontal marker.” We have found that a synthetic peptide containing seven lysine residues and a single tryptophan (K3 WK4 ) functions adequately for this purpose at pH 2.5. A frontal marker may also be useful at higher pH (e.g., when the neutral marker comigrates with a species of interest), where cationic species such as normetanephrine can suffice. It is interesting that, since its inception, the holy grail in CZE has been a mechanism to control EOF. The application of radial field in a manner that controls the magnitude of the “zeta potential” (or the thickness of the double layer) on the capillary wall would, in theory, allow one to simply dial up the desired magnitude of EOF. This mechanism has been illusive for the better part of 15 years, still has not been determined, but clearly still is of interest.56 1.2.3.3 A Description of the Electrophoretic Process In the early 1980s, Jorgenson and Lukacs21 discussed the theory and basis for electrophoretic separations in capillaries. The following brief discussion develops, albeit at a basic level, the theoretical concepts describing the movement of charged molecules in a buffer-filled capillary under an applied electrical field and the shape of the zone during the electrophoretic process. The separation process is discussed in terms of resolution (how adequately two components are separated) and efficiency (how long the separation takes). Finally, we discuss factors that affect the resolution. For more detailed evaluation of these processes, the reader is referred to several excellent discussions on this subject.57−59 1.2.3.3.1 Mobility of an Analyte in a Capillary A charged particle in solution will become mobile when placed in an electric field. The velocity, vi , acquired by the solute under the influence of the applied voltage H, is the product of µapp , the apparent solute mobility, and the applied field E (E = H/L, where L is the length of the field)21 vi = µapp E

(1.3)

µapp is a property of the particle, and proportional to its charge, and inversely proportional to the frictional forces acting upon it in solution. The electrical force can be given by Fe1 = qE

(1.4)

Introduction to Capillary Electrophoresis

13

and the frictional force on a spherical ion is Ffr = − 6π ηrvi .

(1.5)

During the electrophoresis, a steady state is attained, where the two forces are equal, but in opposite directions: qEr = − 6π ηrvi

(1.6)

Solving for velocity (or E) and substituting Equation 1.6 into Equation 1.3 and rearranging for µ, we obtain (1.7)

µapp = q/6π ηr

From this equation, it is evident that the mobility of the analyte is a property of both the charge and size; a small, highly charged particle will have a high mobility, while a large, minimally charged species will have a low mobility. Refer to Appendix 1 for example on calculating mobility. 1.2.3.3.2 Shape of the Analyte Zone Under the initial conditions of electrophoresis, the boundary between the buffer solution and the solute mixture forms a zone of infinitesimal thinness at right angles to the direction of applied current and migration. As migration proceeds, this initially sharp boundary will undergo a progressive deterioration in shape. The most important influence on this process is diffusion, as the initial conditions impart a severe concentration gradient across the zonal boundary. By applying Fick’s second law of diffusion, Weber57 has shown that the variation of solute concentration in the direction of migration is given by the equation

Cx,t =

−(x−vi t) k e 4Di t 1/2 2(Di π t)

2

(1.8)

where C is the solute concentration at a distance x from the initial position after time t, vi is the electrophoretic velocity of the solute, Di is the diffusion coefficient of the solute, and k is a constant. By integration, initially with the boundary conditions at t = 0, at distant x = ∞, C = 0, and a second time with t = 0, x = 0 and dC/dx = 0, we obtain a mathematical description of the concentration profile. As the solute migrates a distance x from the origin after time t, the zone may be described by a Gaussian curve. The characteristics of a Gaussian profile describe the peak maximum and the peak width. The maximum is dependent on initial concentration of solute. The width depends on the length of time from initial conditions (i.e., application of sample in to the system) and the diffusion constant, Di . The width may be given by the distance between the inflection points, in our case, 2/(2Di t)1/2 . From analogy with probability calculations, the width of a Gaussian curve is termed the standard deviation (σ ) σ = (2Di t)1/2

(1.9)

and the square of standard deviation is called the variance σ 2 = 2Di t.

(1.10)

14

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Remembering that t = Ld /vi = Ld Lt /µapp V = Ld /µapp E,

(1.11)

where Ld is the distance from the origin to the detector, and Lt is the total length of the capillary (i.e., the length of the applied electric field), we may substitute into Equation 1.10, to obtain σ 2 = 2Di Ld /(µapp E).

(1.12)

One should remember that the discussion to this point has only dealt with diffusion as a dispersive phenomenon affecting peak shape. Other factors that contribute to what is observed on the electropherogram as band broadening, the practical result of variance, will be discussed later. 1.2.3.3.3 Resolution and Efficiency The simplest way to characterize the separation of two components is to divide the difference in migration distance by the average peak width to obtain resolution (Res) Res = 2(xi2 − xi1 )/(w1 + w2 )

(1.13)

where xi is the migration distance of the analyte i, and the subscript 2 denotes the slower moving component, and w is the width of the peak at the baseline.57 We can readily see that the position of a peak, xt , is determined by the electrophoretic mobility. The peak width, w, is determined by diffusion and other dispersive phenomena. For two neighboring peaks, w1 = w2 , and Res = (xi2 − xi1 )/w2 .

(1.14)

From the equation describing a Gaussian curve, the two peaks touch at baseline when

xi = x2 = 4σ , and Res = 1, or Res = xi /4σ .

(1.15)

Remembering that distance is equal to velocity multiplied by time (xi = vt t) and substituting for σ from Equation 1.9 into Equation 1.15, we obtain Res = ( µapp E)t/4(2Di,avg t)1/2

(1.16)

where µapp is the difference in apparent electrophoretic mobility of the two solutes and Di,avg is the average diffusion of the two solutes. To obtain a measure of efficiency for the process, we use probability theory.58 For a random walk process of length L, made of n steps, the variance is given by σ = l(n)1/2

(1.17)

where l is the length of each step. If each step is independent of any other step, each contributes to the total variance of the process59 2 σtot =




σi2 .

(1.18)

Substituting from Equation 1.17 and rearranging 2 = N. 1/n = L 2 /σtot

(1.19)

Introduction to Capillary Electrophoresis

15

The number of steps in the random process, n, is inversely related to the number of theoretical plates, N, a measure of efficiency for the process. We may substitute for L and σ in Equation 1.19 to express N = (µavg E)2 t 2 /(2Di t)

(1.20)

where µavg is the average mobility of the two solutes. By comparing the expression for Res (Equation 1.16) with the definition of N (Equation 1.20), we obtain an expression relating resolution to the number of theoretical plates Res = (1/4)( µapp /µavg )N 1/2 .

(1.21)

The utility of Equation 1.21 is that it permits one to independently assess the two factors that affect resolution, selectivity, and efficiency. The selectivity is reflected in the mobility of the analyte(s), while the efficiency of the separation process is indicated by N. If Res = 1, then N = 16/( µapp /µavg )2 .

(1.22)

Another expression for N is derived from Equation 1.19, using the width at half-height of a Gaussian peak N = 5.54(L/w1/2 )2 ,

(1.23)

where 5.54 = 8 ln 2, and w1/2 is the peak width at half-height.60 At this point, it is important to note that it is, in fact, misleading to discuss theoretical plates in electrophoresis. The concept is a carry-over from chromatographic theory, where a true partition equilibrium between two phases is the physical basis of separation. In electrophoresis, separation of the components of a mixture is determined by their relative mobilities in the applied electric field, which is a function of their charge, mass and shape. The theoretical plate is merely a convenient concept to describe the analyte peak shape, and to assess the factors that affect separation. Refer to Appendix 1 for examples on calculating resolution and efficiency. 1.2.3.3.4 Source of Variance While N is a useful concept to compare the efficiency of separation among columns, or between laboratories, it is difficult to use to assess the factors that affect that efficiency. This is due to the fact that it refers to the behavior of a single component during the separation process, and is unsuited to describing the separation of two components or the resolving power of a capillary. A more useful parameter is the height equivalent of a theoretical plate (HETP).58 2 HETP = L/N = σtot /L.

(1.24)

HETP might be thought of as the fraction of the capillary occupied by the analyte. It is more practical to measure HETP as an index of separation efficiency, rather than N, as the individual components that contribute to HETP may be individually evaluated and combined to determine an overall value. The variance of multiple dispersive phenomena on the analyte may be summed: 2 2 2 2 σtot + σwall· = σdiff + σT2 + σint

(1.25)

2 should include not only diffusion, but also differA consideration of all the factors influencing σtot ences in mobility or diffusion generated by Joule heating, the reality that the sample is not introduced

16

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

as a thin disk but as a plug of finite dimensions, and interaction of analytes with the capillary wall. Each of these factors will be addressed separately in a simplistic manner. Theoretical derivation will be given only to emphasize the importance of various parameters, and how they contribute to the overall result. References shall direct those desiring more information to the appropriate literature. Two studies have addressed the variance due to a variety of sources. Huang et al.61 investigated small molecules; Jones et al.,62 studied proteins as well as amino acids and a neutral marker. The two studies demonstrate that the band broadening observed in CE is in excess over calculated values of diffusion and analyte interaction. Huang et al.61 and Jones et al.62 attribute the excess variance to sample introductory practices. 1.2.3.3.4.1 Variance Caused by Temperature Temperature control is an issue, because it may be effected by the efficiency with which the capillary is thermostatted, or by choice of buffer ionic species or ionic strength. Thus, from a practical point of view, we discuss the variance caused by temperature to impress upon the practitioner the importance of controlling temperature in obtaining reproducible electropherograms. Current passing through a conducting solution generates heat. A sample way of looking at the problem is to compare the equivalent expressions for the applied potential, H, and heat generation, W H = i/κπ r 2

(1.26)

W = i2 /κ(π r 2 )2

(1.27)

where i is the current through the electrolyte solution, π r 2 the cross-sectional area of the capillary, and κ is the specific conductance of the buffer. By combining Equations 1.26 and 1.27, to take the ratio we obtain H/W = (i/κπr 2 )(κ(πr 2 )2 /i2 ) = π r 2 /I

(1.28)

From Equation 1.28, it can be readily seen that the reduction of heat production can be achieved by reducing the current density in the capillary. This may be accomplished by increasing the crosssectional area of the capillary or by reducing the current. The latter is preferred, since increasing the diameter of the capillary results in a reduction of the surface-to-volume ratio, and leads to less efficient heat dissipation (see Table 1.1). The choices to reduce the sysetm current lie in carrying out the separation at a lower voltage, or reducing the ionic strength of the separation buffer. To develop the quantitative expression for thermal heating, we need to describe the electrophoretic front, and its behavior under a thermal gradient. According to the Poiseuille equation, which describes the parabolic flow due to pressure, vz = Pr 2 /8Lη.

(1.29)

Joule heating of the electrolyte solution creates a similar flow profile, the equivalent expression being vz = v(1 + E 2 Cb Br 2 /4kb T 2 )[1 − (rx /r)2 ]

(1.30)

where E 2 is the rate of heat generation per unit volume, Cb is the buffer concentration, kb is the thermal conductivity of the electrolyte solution. is the equivalent conductanceof the electrolyte

Introduction to Capillary Electrophoresis

17

solution, T is the absolute temperature, rx is the radial position (which varies from 0 at the center of the capillary to r at the capillary wall), and B is a buffer-related viscosity constant.63 The variance caused by dispersion in a parabolic velocity profile is given by σT2



6 4

= 2Di t + r E

Cb2 B2 2 t





2

2

24Di 8kb T − E Cb Br

2

2 

.

(1.31)

For most capillary applications, where radius is small (25–50 µm), and E is kV, E 2 Cb Br 2 is ≪ 8kb T 2 and the second term reduces to σT2 = r 6 E 4 Cb2 B2 2 t/1536Di kb2 T 4 .

(1.32)

The strong dependence of σT2 on the radial dimension of the capillary (r 6 ) and the field strength demonstrate the importance of performing high voltage electrophoretic separations in narrowbore capillaries. It also highlights the necessity of obtaining efficient capillary cooling, to prevent thermal effects from not only affecting sample liability, but also to avoid affecting solute mobility. For small molecules with relatively larger diffusion constants (on the order of 10−5 cm2 /s) Grushka et al.63 have calculated a maximum radius of 65 µm with a 0.1 M buffer solution at 30 kV, and 130 mm with 10 µM buffer for less than 5% increase in dispersion. With larger molecules, having diffusion constants on the order of 10−6 cm2 /s, these dimensions were halved. Jones et al.62 have introduced a novel method to study Joule heating effects and nonideal plug flow contributions to analyte dispersion, with polarity reversal at constant voltage. By recording the variance (band broadening) of the peaks over time and plotting the slope of the obtained line versus applied potential, these authors have developed a measure of nonideal behavior. A comparison of charged analyte behavior with neutral solute variance (which should be unaffected by electrophoretic mobility effects of thermal heating, but affected equally by the temperature-dependent effects on a diffusion and viscosity) allows an estimate of the Joule heating effect. With adequate thermostatting, one need not worry about thermal dispersion under normal operating conditions (see Reference 64), but one needs to be aware of how temperature generated within the capillary can affect the separation and resolution of the analytes. (E 4 )

1.2.3.3.4.2 Variance Caused by Finite Sample Introduction Volume In the mobility section above, we assume the solute was initially present as a think disk that dispersed into a zone or band. In reality, the sample is introduced into the capillary as a cylindrical plug. Sternberg60 derived the commonly used equation describing variance due to sample introduction. 2 2 σint = lint /12

(1.33)

where lint is the length of the sample introduction plug. This formula assumes that the sample is introduced as a rectangular plug, which is a close approximation for CE. As the plug has a finite volume, it also has measurable length. Huang et al.61 and Jones et al.62 have investigated the relative contribution of introduction plug length (by hydrodynamic or electrokinetic methods) to the observed peak width, and concluded that lint /Ld of less than 3% does not lead to excessive band broadening. Only at very small plug lengths (1 mm o.d.) in shapes that are commonly cylindrical, but also rectangular. The following sections describe some pertinent points regarding capillary preparation, conditioning before use, maintenance, and storage respectively. 1.2.3.4.1 Preparation for CE For use in CE, the capillary is cut to the appropriate length with a ceramic knife, or an ampoule file, so that both ends are square and flat. This is of particular importance with the inlet end of the capillary where sample is to be introduced. Closer to the outlet, a “window” is created through removal of the polyimide coating so that online detection is possible. The window should ideally be approximately 0.3 cm in length (no longer than 1.0 cm) and can easily be made by burning off the polyimide with a flame and wiping the surface with an ethanol-soaked lens tissue. Another, more labor-intensive method is to scrape off the polyimide surface, being careful not to scratch the silica, or break the capillary. Several mechanical devices have been devised to accomplish this task.65,66 Commercial instruments are available for carrying this out (e.g., Capital Analytical, UK) Caution must be taken when handling the capillary once the window is created since the window area is extremely fragile. It is important to note that with internally coated capillaries, detector windows cannot be created by burning off the polyimide since the heat required will damage the internal coating. Alternatively, dropwise addition of 103◦ C sulfuric acid33 or hot concentrated KOH67 from a burette or glass pipette have been reported to remove the polyimide coating to create a window without damaging the internal coating. We use an adaptation of the hot sulfuric acid method. A drop of concentrated sulfuric acid is placed on the capillary where the window is desired, and the tip of a hot soldering iron is briefly touched to the drop. This readily removes the polyimide coating without overheating the interior surface, and prevent the necessity of heating a larger volume of acid than would otherwise be needed. The excess acid must be washed off the exterior surface, especially if the capillary is to be placed into a cartridge where the acid would damage the cartridge housing.

20

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

When the capillary is installed, it is useful to determine the “capillary fill time.” This information will be useful for determining the length of time for each rinse step (NaOH, separation buffer or rinse solution), as well as for diagnosing potential capillary problem, such as partial or complete obstruction. One approach for accomplishing this (with ultraviolet, UV detection) is to pressure rinse the capillary with water, zero the detector, and follow with a rise with 100 mM, NaOH. The time required for a maximum change in absorbance to occur is the “fill time” to the detector. Factoring in the length between the detector and the capillary outlet yields the “capillary fill time.” Before using new bare fused-silica capillary for analysis, the capillary should be “preconditioned” by rinsing with 5–10 column volumes of 100 mM NaOH followed by 5–10 column volumes of water, before rinsing with 3–5 column volumes of separation buffer. If the capillary is coated, preconditioning should be performed as per the protocols recommended by the manufacturer or as deemed necessary by the surface coating. 1.2.3.4.2 Conditioning of the Capillary When using a newly installed capillary or changing to a new separation buffer, the capillary must be adequately equilibrated with the separation buffer, a process termed “conditioning.” Equilibration is particularly important when a phosphate-containing buffer is involved. For acceptable reproducibility, a phosphate-containing buffer should be equilibrated in the capillary for a minimum of 4 h before use.68 This process can be slightly accelerated by applying a separation-scale voltage to the system during the equilibration period. When preparing to use a new separation buffer in a capillary that has been equilibrated with a phosphate-containing buffer, extensive washing and equilibration of the capillary is typically required. The capillary should be equilibrated with the desired separation buffer for at least several hours before use and, if possible, conditioned overnight. It is for this reason that “dedication” of capillaries to individual buffer systems may be highly recommended. In this manner, no pre-equilibration time may be required, since the capillary can be stored in the buffer to which it is dedicated. However, users often report reproducibility problem that are suspect of surface chemistry deviation, particularly in the buffer ph range of 4–7. Watzig et al.69 describe preliminary results using x-ray photoelectron spectroscopy to probe the Si-C landscape of fused-silica surfaces. This led the authors to suggest that, in order to yield capillaries that will provide stable migration times, especially in the pH range 4–7, preconditioned for longer than 1 h is required. 1.2.3.4.3 Regeneration of the Capillary Surface Capillary maintenance plays a critical role in attaining reproducible results with CE. As with any untreated silica surface, ionized silanol groups are ideal for interaction with charged analytes, particularly peptides and proteins in neutral/basic pH buffers. Hence, following each separation, the capillary surface must be “regenerated” or “reconditioned”, that is, cleansed of any wall-adsorpted material. This is accomplished by following each run with a 3–5 column-volume rinse with 100 mM NaOH, followed by flushing with 5–8 column volumes of fresh separation buffer. This, of course, should be optimized for the particular buffer system used. However, the users should heed the results of Watzig and coworkers.69 Alternative solutions for cleaning the capillary are 100 mM sodium tetraborate, pH 11, or mM trisodium phosphate, particularly for use with phosphate containing buffers. When using acidic buffers, rinsing with NaOH may be a disadvantage since the drastic changes in pH may induce the requirement for extensive rinsing with the separation buffer for adequate pH re-equilibration of the capillary. Under these conditions, it may be advisable to follow the NaOH trines with a brief rinse with a concentrated separation buffer (10X or more), followed by the normal rinse with the IX separation buffer. Alternatively, depending on the sample, the 10X separation buffer rinse alone may be adequate for regenerating the capillary and, hence, the NaOH rinse may be avoided completely. For peptide separations in phosphate buffer, pH 2.5, 1.0 M phosphoric acid solutions may be used for capillary regeneration. Another good protein cleaning solutions is 1 M HNO3 .

Introduction to Capillary Electrophoresis

21

Occasionally, after extensive use with protein solutions, the capillary will become fouled and will need to be cleaned to remove strongly adsorpted material. This is usually noticed with migration times that are excessive, with peaks distinctly asymmetrical with extensive tailing, or filling times become noticeably longer. A rinse with sodium ethoxide solution (blended from equal volumes of 1 M NaOH and 95% ethanol) for 5 min, followed by extensive (10–20 column volumes) water rinses usually suffices. A partially plugged capillary may be unplugged by filling with sodium ethoxide and allowing the solution to stand for 5 min before flushing with fresh solution, and following with an extensive water rinse before buffer equilibration. It should be noted that the sodium ethoxide will etch a new surface by removing the surface layer of silica, eventually making the capillary brittle. While we have used this procedure in our laboratory, keeping some capillaries in use for up to one year, this solution should NOT be used with coated capillaries, as it will remove the coating. 1.2.3.4.4 Storage Because of the small dimensions of the capillary, plugging is always a potential problem. This may occur as a result of solvent evaporation at the capillary ends, which leads to salt crystal formation. This may be avoided in several ways. Dedicated capillaries to be stored for a short time (less than a week) should be rinsed with 100 mM NaOH, re-equilibrated with separation buffer and stored with the ends immersed in buffer/distilled water, or capped with silicone rubber stoppers. Optionally, the capillary may also be stored dry using the procedure below. If the capillary is to be stored for longer periods of time (i.e., >1 week), it should be washed with 100 mM NaOH, rinsed thoroughly with distilled water, purged with nitrogen or air (by pressure) and stored dry. If a capillary has been used with a detergent-/surfactant-containing solution, all traces of the additive must be removed before storage. If this is not done, the capillary will require longer than normal re-equilibration time upon reuse owing to the presence of the residual additive. To place a stored capillary back into service, one should follow the same procedure outlined for a new capillary. If the capillary has been stored dry, we have found it useful to rehydrate the lumenal surface with a 5 min rinse of distilled water before rinsing with 100 mM NaOH and re-equilibration with buffer. Before storing coated capillaries, they should be thoroughly rinsed with the appropriate buffers or solvents. For example, we have found that it is best to rinse hydrophobically coated capillaries with one-column volume 100 mM NaOH, then extensively with water (5–10 column volumes), and finally with methanol (5–10 column volumes), followed by drying and storage. Capillaries used with physical gels should be rinsed, filled with fresh buffer, and stored with both ends tightly capped. Capillary coating for particular CE applications can be of paramount importance because of the affinity of some analytes, particularly proteins, for interaction with the wall. A number of methodologies have been described in the literature for coating the internal surface of the capillary.70 The pH stability, as well as the effectiveness for preventing adsorption of analyte to the wall (i.e., separation efficiency), varies with the chemical nature of the coating. Some precoated capillaries are commercially-available from several suppliers. 1.2.3.4.5 Basic Aspects of a Typical CE Method As a means of collating the above information into practical form, the following provides what might be considered to be a rudimentary CE method. 1. Mount capillary in the instrument playing close attention to alignment with the aperature/detection optics. Once the new uncoated fused-silica capillary has been conditioned as described above, the capillary is mounted in the instrument and properly equilibrated. 2. Determine the capillary fill time. First fill the capillary with some low-absorbing solution (buffer or dH2 O) and then zero the detector. Using the rinse mode, fill the capillary with a strongly absorbing solution (e.g., 0.1 M NaOH) noting the amount of time to reach

22

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

3. 4.

5.

6.

7.

8.

the detector. Knowing the ratio of the effective length (inlet → detector) and the total length (inlet → outlet), calculate the time required to fill the entire capillary. Conditioning the capillary as described in Section 1.2.3.4.2. Make sure capillary is equilibrated. The capillary should be equilibrated preanalysis with a minimum of a two capillary volume rinse with unelectrophoresed separation buffer, that is, a separate vial of separation buffer that has not been electrophoresed. This assumes that the capillary is well conditioned and equilibrated if new, or well re-equilibrated if it is not the first run. Inject sample. Sample is injected either by hydrostatic pressure or electrokinetically depending on the nature of the analyte and the sample matrix. Ideally, the injection should be followed by a second injection of separation buffer (equivalent of 1–2 s at 0.5 psi) to avoid any loss of sample into the inlet vial during the first few seconds of voltage application. This is particularly important with sample matrices that end to generate significant Joule heat. Apply voltage to the system. With most applications, it is recommended that a low voltage is applied for a short period of time (e.g., 1 kV for 1 min) before application of the higher separation voltage. If the desired resolution is not obtained, one may consider examining voltage ramping. Regenerate the capillary surface. The capillary should be regenerated with a minimum two capillary volume rinse with regenerating solution before re-equilibration with separation buffer. For neutral to high pH separations, 0.1–1.0 M NaOH is effective while 0.1–0.5 M phosphoric acid (or the acid equivalent of your separation buffer) is adequate for low pH separations. This should be followed with a short (one capillary volume) rinse with distilled water. Re-equilibrate the capillary surface. The capillary should be re-equilibrated with at least 8–10 capillary volumes of unelectrophoresed separation buffer. If this is not the first run, this can be shortened to 5–6 capillary volumes since the first step in the method (step 4 above) precedes injection with a 2–3 capillary volume rinse with separation buffer. This is also an appropriate end point following the last run since the capillary remains in separation buffer (and not 0.1 M NaOH) overnight.

It should be clear that these are basic guidelines for setting up a CE method. As emphasized throughout this book, no method is universally applicable to all separations. Each method is likely to require the incorporation of idiosyncratic steps specific to the mode, sample, type of capillary, buffer, pH, etc. used.

1.3 METHOD DEVELOPMENT∗ As with HPLC, there will be no universal CE conditions that will be appropriate for the analysis of all types of samples/analytes. In fact, it has been shown that one set of conditions is not sufficient for the analysis of one class of proteins, for example, glycoproteins,71,72 or even variants of the same proteins from different species.73 One of the first steps in designing a method is to determine, on the basis of the type or “class” of analyte involved, which of the CE modes is best suited to the sample (Table 1.2). Once the appropriate CE mode has been identified, analysis is carried out and separation optimization initiated. Optimal separation of the components of any sample requires a logical approach to sample solubilization and/or dealing with diverse sample matrices as well as identification and utilization of the correct combination of CE operating parameters. Each of these will have distinct effects on the resultant separation and resolution. There is a massive literature on ∗ Based on information given in the p/ACE 2000 Customer Training Manual. With permission.

Introduction to Capillary Electrophoresis

23

CZE method development, not so much in the form of papers that address this per se, but more in the papers that describe a novel separation of a previously unstudied analyte. In these it is likely that fairly extensive details are provided on the developed method.74−80 In addition, several reviews have been written on aspects of this (e.g., References 81 and 82).

1.3.1 STEPS IN DESIGNING A METHOD A general strategy for CE analysis of a sample is diagrammed in Figure 1.6. The first step is an obvious but important one: choosing the optimum wavelength for detection. This may not necessarily be the most sensitive wavelength (i.e., 200 nm for proteins), but more importantly, one that specifically enhances the detectability of the substance of interest and minimizes the absorbance of background components. Choice of a buffer system is of the utmost importance. The buffer needs to be selected with strict attention to not only the pH requirements of the sample from a stability perspective, but also to the wavelength restrictions of the buffer (i.e., λmin for detection; see Table 1.3 in this chapter). If the use of a specific high ionic strength (high conductivity) buffer system is required (e.g., for preservation of bioactivity), the maximum working voltage should be determined by an Ohm’s Law plot (discussed in Section 1.3.3.5 on buffers) to avoid buffer overheating problems. A review of the literature will provide a general idea of the basic parameters for an initial test run. A quick trial should establish the appropriateness of the sample introduction method, initial field strength, and capillary length. Optimization of the separation and resolution should include modification of the field strength, modification of the buffer by either altering pH or using additives to alter analyte mobility or EOF, and changing the length of the capillary as necessary. If peak resolution is poor, one might consider increasing the length of the capillary and/or the frequency of data collection. The remainder of the steps rely largely on the experience and intuition of the operation based on the characteristics of the sample/analyte. Once adequate separation of the sample components is attained, sample introduction should be optimized for a maximum signal-to-noise ratio. This step might include, where possible, altering the sample concentration or the buffer composition (ionic strength, type, pH, etc.). Typical starting conditions for analysis of unknowns (e.g., proteins or some drugs) in our laboratory are 50 µm i.d. × 30 or 40 cm in effective separation length capillary, 100 mM borate, pH 8.3, 25 kV, with a 3 s pressure injection at 0.5 psi. We find 200 nm to be a good wavelength for UV detection of most organic compounds with most buffers; due to the absorbance of conjugated bonds, there is a 5–10-fold enhancement of the signal over that observed at 280 nm with most proteins and peptides. A reasonable starting concentration for protein samples is 0.1–1.0 mg/mL, usually in 10 mM borate, pH 8.3 or 10–20 mM Tris, pH 7.5.

1.3.2 SAMPLE PARAMETERS TO CONSIDER Clearly, analysis of a sample requires that it be solubilized. Although it is not always possible to know the physicochemical properties of the sample to be analyzed, it is important to be familiar with at least some of the physical properties of the specific analyte of interest in the sample. Some physicochemical properties may ultimately have to be determined empirically. Different questions should be posed depending on whether the sample involved is lyophilized or already solubilized, as in a biological fluid. In either case, it is important to have some knowledge of the detectability, purity, and stability of the sample/component of interest. A number of questions need to be addressed. How complex is the sample? Is the substance of interest a major or minor component? What other substances in the sample might interfere with the detection or separation of the analyte of interest? What is its λmax ? Is it thermally stable? If structural information is available, what are the pK values of the ionizable groups? If the sample contains proteins or peptides, what pIs are involved? If the sample is not in solution, there are additional questions that must be considered. Is it soluble in water or low ionic strength buffer at mg/mL concentrations? Does it require extremes of pH for solubility and, if so, is it stable at this pH? Is solubility enhanced by buffer additives such as urea, methanol,

CZE

MEKC

usually 200, 208, 214 or 260 nm

Do Ohm's Law Plot (OLP)

increase if shorter migration times req’d decrease if current too high

pH 2.0 -4.5

phosphate, citrate,formate or acetate at

use additives such asMeOH orAcN • try sample stacking or focusing

•

Maximize sample solubility

use buffer additives • organic solvents (MeOH,AcN) • ion pairing agents (HSA)

• •

Do Ohm's Law Plot (OLP)

increase if shorter migration times req’d decrease if current too high (>1-2 W/m)

keep injection 50% of the total protein in a sample). If the samples can be separated into several major peaks (each major peak should be >20% of the total protein in a sample), the starting point can be calculated as follows: Number of major peaks × 0.1mg/mL In summary, for an unknown sample, the initial cIEF analysis conditions would include 4% pH 3–10 Pharmalyte, 0.5% MC as the additive, 0.08 M H3 PO4 as the anolyte, 0.1 M NaOH as the catholyte, 0.2 mg/mL sample concentration, and 6 min focusing time at 500 V/cm voltage with a 1–2 min prefocusing step.

19.5.2 METHOD DEVELOPMENT FLOW CHART Figure 19.12 illustrates a flow chart for the development of the cIEF method. At the beginning, it is recommended that the initial conditions described in Section 19.5.1 are tested. The second step should then involve adjusting the sample concentration. The third step is the critical step in the method development because a high-resolution method is based on good reproducibility. All methods should be attempted to ensure a reproducible peak pattern. Then, the separation resolution should be enhanced with the use of narrow pH range CAs. The final step in the method development is pI determination and sample peak identification.

19.5.3 pI DETERMINATION AND PEAK IDENTIFICATION Determination of the pI of a protein using cIEF requires two basic conditions: pI markers with accurate pI values and a linear pH gradient created by the chosen CAs. Using pI markers is the only way to characterize a pH gradient within the capillary column created by CAs [44]. As for the second condition, a pI value in a nonlinear pH gradient is difficult to characterize using a limited number of pI markers since the number of pI markers in a given pH region is always limited. In most cases, these pI markers do not distribute evenly within the pH region. Curve fitting in a nonlinear pH gradient using these pI markers involves a large error that is difficult to estimate. The determination of the pI value should be performed in a single CA with good linearity in its pH gradient.

576

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Initial conditions Adjust sample concentration Yes

If Abs < 0.1 or > 0.6 No

1. Additives 2. Different CAs No

Peak pattern reproducible? Yes

1. Narrow range CAs 2. CA mixtures

Need higher resolution?

Yes

No Results

FIGURE 19.12 cIEF method development.

2000 Peak position (pixel)

Pharmalyte 3–10 1500

1000

r2 = 0.9883

500

0 3

5

7

9

pH

FIGURE 19.13 pH gradient profile of 8% Pharmalyte pH 3–10.

Profiling the pH gradient of different CAs is important for this purpose. Figure 19.13 illustrates an example with a pH gradient for Pharmalyte pH 3–10. The overall linearity of the pH gradient is good (r 2 is about .99). However, the gradient bends at two places, around pH 5 and 8. If pI determination is performed around these two regions, multiple pI markers should be used to ensure good accuracy. However, in other pH regions of this CA, the linearity in the pH gradient is well above r 2 = .99. Two pI markers should therefore be adequate to calibrate the pI values. In most practical applications, the determined pI values of a sample are only used to identify component peaks. The true pI value of a peak is not relevant in the applications, as long as the determined pI value of this peak (component) is reproducible in all samples run under identical experimental conditions. It has been proven that peaks (components) in different samples can be identified using the determined pI values within a standard deviation of 0.1 pH units (90% and the nitrogen pressure can be adjusted to obtain the desired efficiency. A phase-contrast microscope or trypan blue staining can be used to visually confirm cell disruption. 20.5.1.2 Differential Centrifugation Following lysis, centrifugation is nearly universally implemented to separate organelle fractions based on their size or density. Centrifugation is popular because it is very versatile; it can be used to fractionate organelles from several liters of tissue homogenates or submilliliter volumes from cell cultures. The simplest centrifugation procedure is differential centrifugation, which produces crudely purified organelle fractions. Differential centrifugation begins by pelleting nuclei at low centrifugal force (i.e., 600–1000 × g for 10 min). The resulting nuclear fraction is always contaminated with organelles that have been trapped in the cytoskeletal network, whole cells, and large membrane debris. The postnuclear supernatant is then transferred to a new microcentrifuge tube and the light-organelle fraction can be sedimented at 16,000 × g for 10 min, and the supernatant will contain the cytosolic components. The light-organelle fraction is now enriched with mitochondria, but also contains significant amounts

Subcellular Analysis by Capillary Electrophoresis

599

of acidic organelles, microsomes, and peroxisomes. For some analyses that use organelle-specific labels, no further purification is necessary. However, if purer organelle fractions are required, the light-organelle fraction can be further fractionated as described below. 20.5.1.3 Density Gradient Centrifugation The mitochondria in the light-organelle fraction can be further purified using common methods such as continuous gradient and discontinuous gradient centrifugation. While continuous gradients can be used (e.g., sucrose gradient), since the osmolarity changes quite drastically through the gradient, discontinuous gradients are preferred to reduce mitochondrial disruption. Furthermore, the mitochondrial band is harder to distinguish when continuous gradients are used, which reduces the recovery. Discontinuous density gradient centrifugation separates organelles based on their density and has been described by Storrie and Madden.84 Unfortunately, the popular medium for density gradient centrifugation, Metrizamide, is no longer produced, and therefore, the existing methods must be redesigned with appropriate new media. We have found success using Histodenz (Sigma, St Louis, MO) following the procedure of Okado-Matsumoto and Fridovich.85 Briefly, a 50% (w/v) solution of Histodenz is prepared by dissolving in 5 mM Tris–HCl, 1 mM EGTA at pH 7.4. Solutions of 34%, 30%, 25%, 23%, and 20% solutions (w/v) of Histodenz are prepared by diluting the stock solution in 250 mM sucrose, 5 mM Tris−HCl, 1 mM EGTA at pH 7.4. The light-organelle fraction is suspended in ∼10 mL of 25% Histodenz and the discontinuous gradient is prepared as shown in Figure 20.15. The samples are then centrifuged at 52,000 × g for 90 min at 4◦ C. After centrifugation is complete, the bands can be removed sequentially including the band with the purified mitochondria located at the 25/30% interface. Finally, the purity and yield of the mitochondrial preparation can be determined using enzymatic marker assays, such as cytochrome c oxidase (mitochondrial marker) and β-galactosidase (lysosomal marker).

20.5.2 CAPILLARY MODIFICATIONS Mitochondria interact strongly with the surface of bare fused-silica capillaries, preventing any accurate measurement of electrophoretic mobility. Since mitochondria and the capillary surface both carry a net negative charge, the adsorption is probably not due to any electrostatic interactions, but rather to hydrophobic intermolecular interactions. To prevent the mitochondria from interacting with the capillary surface, coatings can be applied to the bare silica. Among these coatings, AAP has been reported to be more hydrophilic and hydrolytically stable in comparison with acrylamide, dimethyl acrylamide, and N-acryloylaminoethoxyethanol.86 We have found that capillaries

20% 20/23% 23% Light-organelle fraction

23/25%

25% 25/30% 30%

Purified mitochondria

30/34%

34%

FIGURE 20.15 Density gradient centrifugation. The postnuclear supernatant, suspended in 25% Histodenz (Hz), is layered on top of the 30% and 34% Hz solutions. Then the tube is filled sequentially with the 23% and 20% solutions of Hz. Centrifugation is performed at 52,000×g for 90 min at 4◦ C. The mitochondria will form a band at the 25/30% Hz interface.

600

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

coated with poly(AAP), based on Gelfi’s procedure,87 reduce organelle adsorption, and thereby allow reproducible separations; this coating remains stable for several months. Covalent surface modification with poly(AAP) consist of the following steps: 1. Flush 4–5 m of fused-silica capillary sequentially with 0.1 M KOH for 3 h, 0.1 M HCl for 1 h, MeOH for 1 h with nitrogen pressure (90 psi for all flushes unless stated otherwise) 2. Heat the capillary to 110◦ C and flush with nitrogen at 10 psi for 8 h 3. Place the capillary in a moisture-free, inert environment (glove box) and do not remove until step (9) 4. Flush the capillary with thionyl chloride for 1 h at 65◦ C 5. Seal the capillary ends with a gas chromatography (GC) septum and heat for 8 h at 65◦ C 6. Flush the capillary with freshly distilled tetrahydrofuran (THF) for 30 min at 65◦ C 7. Prepare a 0.25 M vinylmagnesium bromide solution with freshly distilled THF, flush the capillary with this solution for 1 h and heat for 8 h at 70◦ C 8. Flush the capillary with freshly distilled THF for 1 h and water for 30 min 9. Cut the capillary into 50 cm pieces 10. Add 5 µL of freshly prepared 10% ammonium persulfate and 2 µL tetramethylethylenediamine (TEMED) to a monomer solution and immediately flush this solution through the capillaries for 5 min 11. Polymerize the capillaries at room temperature for 1 h 12. Flush the capillaries with water for 1 h Freshly distilled THF must be used in each step to prevent the precipitation of vinylmagnesium bromide in the capillary. To increase the robustness of the capillary modification, steps (4) and (7) may be repeated. The AAP monomer can be synthesized in a single step reaction with >99% yield and >99% purity.86 The polymerization time and concentrations of ammonium persulfate and TEMED must be optimized for each batch of AAP. The success of the procedure can be checked by measuring the electroosmotic flow, which should be 8.5) is a dimer containing in each of the monomer an N-linked glycan of the high-mannose type. Using neutral hydrophilic capillaries and phosphate buffer at pH 2.5 as background electrolyte (BGE), the 15 possible rhBMP-2 glycoforms were separated into nine peaks. The use of the glycan and peptide maps, together with fast atom bombardment mass spectrum of the generated glycopeptides, enabled the authors to identify these peaks as the nine glycoforms of the structure (rhBP-2)2 -(GlcNAc)4 -(Manz ), where z varies from 10 to 18. As all of the glycoforms have the same charge, the authors concluded that the size of a mannose residue was fundamental for the separation and that the increasing hydrodynamic radius of the protein could control the separation. However, in rhBMP-2, the mannose glycans appear to be large enough to cover part of the polypeptide backbone, a situation that would lead to the shielding of the most charged region of the protein and to a reduction of its effective charge. An estimation of the effect of the mass of a mannose on the mobility of rhBMP-2 in terms of differences in charge and frictional coefficient led to the conclusion that it was not possible to assess whether the effect of the mannose residue was from a reduction of the z potential, from an increase of the frictional coefficient, or from a combination of the two. From the above discussion, it can be concluded that both the effective charge and the hydrodynamic radius play a major role in the control of the electrophoretic mobility in the separation of proteins by CZE. Since proteins are flexible molecules, the interplay of both factors has to be taken into consideration to explain their mobility in electrophoresis. For glycoproteins, a full understanding of the factors that control their separation in CZE is a complex task because the size and flexibility of the glycans constitute an additional cause of structural variability.

22.4 METHODS DEVELOPMENT GUIDELINES 22.4.1 SAMPLE PREPARATION As in any other analytical procedure, sample preparation, in those cases where it is needed, is a key step in the separation of glycoforms by CE. The main goals of the sample preparation step are concentration and purification and, when designing the protocol, the peculiarities of each analyte and of the sample matrix must be taken into account. For protein separations by CE, in addition to aspects common to other types of analysis such as filtration, it is often necessary to eliminate compounds that may interfere either by causing a high electrical current, by adsorbing to the capillary walls, or by comigrating with the analytes. Without sample preparation, the analysis of a dilute serum sample by CE with ultraviolet (UV) detection leads to the separation and detection of the major serum proteins, γ-globulin, transferrin (Tf), β-lipoproteins, haptoglobin, α2 -macroglobulin, α1 -antitrypsin (AAT), α1 -lipoproteins, human serum albumin (HSA), pre-albumin, and complements [32], as shown in Figure 22.4. These proteins often mask other important analytes that are usually more clinically relevant than the major proteins. In fact, it has been widely mentioned that global approaches to proteome analysis detect less than 0.1% of the protein species present in a sample, which span a range of 10 orders of magnitude between the most abundant and the less abundant proteins [33]. The following discussion will focus on an example of a glycoprotein to which several sample preparation procedures have been reported, that is, erythropoietin (EPO) to which a lot of attention has been devoted for its analysis by CE. The elimination of low molecular weight (MW) compounds from a sample is widely carried out by using microcentrifugal devices provided with membranes of different MW cutoffs. Clean-up of the rhEPO Biological Reference Preparation (BRP) of the European Pharmacopoeia (EP) was carried out by passing an aqueous solution through a microconcentrator cartridge with a 10 kDa

Glycoprotein Analysis by Capillary Electrophoresis

639

30

Miliabsorbance (200 nm)

25

20

9 15

10

5

3

2 2'

1 0 100

4

5

6 7

150

8 200

10 250

Time (s)

FIGURE 22.4 Electropherogram of a diluted normal control serum. Peaks: 1, neutral marker; 2, γ-globulin; 3, transferrin; 4, β-lipoproteins; 5, haptoglobin; 6, α2 -macroglobulin; 7, α1 -antitrypsin; 8, α1 -lipoproteins; 9, albumin; 10, prealbumin; 2′ , complements. Analytical conditions: fused-silica capillary 25 cm × 25 µm i.d.; buffer: Beckman proprietary buffer of pH 10.0; 10 kV; 22◦ C. (From Chen, F.-T.A., J. Chromatogr., 559, 445, 1991. With permission.)

MW cutoff membrane and followed by two consecutive rinses with 0.2 mL water [34,35]. A more detailed procedure detailing the time and the centrifugal force to be applied as well as an increase in the number of washing steps was subsequently reported [36]. Ways to decrease the preparation time by increasing the centrifugal force [37,38], and to increase the recovery by repeatedly washing the membrane and inverting the filtering device were also described [38]. An intermediate procedure led to an enhanced resolution of rhEPO glycoforms when compared with the analysis performed under identical conditions but without sample pretreatment [30]. As it will be shown later, this type of sample preparation step that eliminates excipients of low MW is especially important when the separation mode is CIEF [36,39]. This same type of procedure for eliminating low MW excipients was also useful to remove mannitol from a standard of human urinary EPO (uEPO) to be analyzed by CZE [40]. When the sample matrix contains high MW components, such as in biological fluids or in pharmaceutical products formulated with protein excipients, strategies other than MW-based sample preparation are often required. Using again rhEPO as an example, glycoform analysis of products containing HSA under the same CZE conditions as those used for samples containing low MW components was unsuccessful. An approach at separation that did not require a sample preparation step was the development of a CZE method involving the selective binding of HSA. This was accomplished on a polyamine-coated capillary by adding nickel ions to the BGE. The metal ions selectively bonded HSA and decreased its mobility leading to separation of the otherwise comigrating rhEPO [41]. Polybrene (PB)-coated capillaries with a buffer containing acetic acid and methanol also allowed the separation of rhEPO from HSA [42]. A different approach relied on performing the immunochromatographic removal of HSA before separation. Figure 22.5 shows the electropherograms of the unresolved peak obtained for an HSA-formulated rhEPO sample before HSA removal and the separated rhEPO bands after the immunochromatographic HSA depletion [43]. Figure 22.6

640

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a)

Absorbance (U.A.)

0.03

0.02

0.01

0.00 0

(b)

10

20 30 Time (min)

40

50

0.02

Absorbance (U.A.)

0.01

0.00

-0.01

-0.02

10

15

20

25

30

Time (min)

FIGURE 22.5 CZE electropherograms of rhEPO alpha formulated with human serum albumin (HSA), (a) without immunodepletion of HSA, (b) after HSA immunodepletion. Separation conditions: Uncoated fusedsilica capillary 87 cm × 50 µm i.d. Separation buffer: 0.01 M Tricine, 0.01 M NaCl, 0.01 M sodium acetate, 7 M urea, and 3.9 mM putrescine, pH 5.5; pressure injection 30 s at 0.5 psi; 25 kV; temperature 35◦ C; detection at 214 nm. (Adapted from Lara-Quintanar, P. et al., J. Chromatogr. A, 1153, 227, 2007. With permission.)

shows the separation of a similar sample on polyamine-coated capillary and nickel ions added to the BGE [41]. Both methods provide separation of glycoforms and while the one performed on coated capillaries does not require a prior sample preparation step, the other method can be performed with bare fused-silica capillaries and provides better glycoform resolution. Sample preparation steps used for other samples are mainly a function of the matrix, the most complicated ones being for the biological fluids. Specific methods used to prepare some of these samples are detailed in Section 22.5.2.

Glycoprotein Analysis by Capillary Electrophoresis

641

Abs at 200 nm

(d)

(c)

(b)

(a) 0.00

10.00

20.00

30.00

40.00

45.00

Time (min)

FIGURE 22.6 CZE electropherograms of (a) human serum albumin, (b) rEPO, (c) HSA-formulated rEPO. Inset (d): enlarged view showing rEPO glycoform separation in formulated product. Separation conditions: Beckman eCAP amine capillary, 200 mM sodium phosphate, pH 4.0, 1 mM nickel chloride, –15 kV (75 µA), UV detection at 200 nm. (From Bietlot, H.P. and Girard, M., J. Chromatogr. A, 759, 177, 1997. With permission.)

22.4.2 SEPARATION OF GLYCOFORMS 22.4.2.1 Solutions to Protein Adsorption Problems Very high separation efficiency can be achieved when the interaction between a protein and the inner wall of the capillary can be eliminated [44]. This is especially important for glycoproteins where this type of unwanted interaction has to be prevented to achieve the high efficiency required for the separation of glycoforms. Two basic strategies can be utilized to overcome this problem. The first one consists in changing the buffer pH and the second one in modifying the inner capillary wall. The first, and simplest, approach is to select a buffer pH at which interactions become small. At a pH below 3, the silanol groups on the silica surface are fully protonated and bear no electric charge, thus minimizing interactions since most proteins have pI values above 3. On the other hand, at pH 8 or higher, silanol groups are fully ionized and the silica surface becomes negatively charged. So, at pH 8, a protein having a pI smaller than 8 will have a net negative charge and will be repelled by the surface, while one having a pI higher than 8 (a basic protein) will be charged positively and may be adsorbed on the capillary walls. In such a case, a buffer with a pH higher than the pI of the

642

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

protein could be used for preventing their adsorption. Although these strategies have been proposed for the separation of proteins since the early days of CE [45,46], they limit the selection of buffer pH as a freely adjustable parameter for optimization. In addition, too large differences between the buffer pH and the protein pI can lead to structural changes resulting in band broadening and low recovery. These limitations have led to other approaches for preventing protein adsorption, which allow a wider selection of buffer pH. Several methods employing capillary wall modifications have been developed and they can be grouped into several categories: dynamic coatings by the addition of a cationic modifier (usually an amine) to the BGE [47,48], permanent coatings by physical adsorption of a cationic modifier (usually a polymer) [47,48], and permanent coating by covalent bonding of a hydrophilic polymeric layer [49]. 22.4.2.1.1 Cationic Buffer Additives The addition to the buffer of cationic substrates that interact with the silanol groups at the surface of the fused-silica capillary represents the simplest method to improve the separation efficiency in glycoprotein analysis. Several effects have been attributed to these additives, which include decreasing the interaction of the analyte with the inner capillary walls, controlling the electroosmotic flow (EOF), increasing the protein solubility, and enhancing the selectivity of the separation. Although additives of several types, including neutral and ionic substances, have been reported [47,48], amine modifiers have been more frequently used for the separation of protein glycoforms. There are two main groups of amines, one comprising diaminoalkanes and the other polyaminoalkanes. In the first group, short alkyl chain diamines have been used at millimolar concentrations for the effective separation of glycoproteins such as ovalbumin (OVA) [50,51], rhEPO [52], recombinant factor VIIa [53], granulocyte colony-stimulating factor (rhGCSF) [54], Tf [55], and human chorionic gonadotropin (hCG) [56]. The diaminoalkanes play a role in the separation by affecting both the resolution and the migration time. The separation of hCG glycoforms in uncoated silica capillaries at different concentrations of 1,3 diaminopropane (DAP) in borate buffer (Figure 22.7) shows that an increase in the diamine concentration results in an increase in resolution with a concomitant increase in migration times [56]. Replacing the terminal diamino groups of diaminoalkanes by quaternary ammonium moieties gave rise to more efficient separations [57]. Although a similar resolution was achieved with both types of additives, lower concentrations of quaternary ammonium salts were necessary and shorter migration times were obtained, an indication of effectiveness of these bis-quaternary ammonium compounds. Quaternary ammonium salts such as hexamethonium chloride and bromide and decamethonium bromide (DcBr), having alkyl chains of 6 and 10 carbon atoms, respectively, have been used for the separation of the isoforms of OVA and hCG [57], and Tf [55]. For quaternary ammonium salts, the longer alkyl chains were shown to be more effective and chloride salts led to shorter analysis time than bromide ones. Polyaminealkanes, such as spermine and spermidine, have been used for the separation of OVA glycoforms [51]. Interestingly, while the efficiency of a polyamine for preventing protein adsorption to the silica walls was shown to increase with the number of amino groups in the chain, this number needed to be higher than three for it to become an effective adsorption inhibitor [58]. 22.4.2.1.2 Capillary Coatings by Physical Adsorption Physical adsorption of polymeric cationic coatings on the inner surface can be easily accomplished by equilibrating the capillary with a solution containing the polymeric additive. In some cases, the polymer remains so tightly bonded to the surface that this does not lose its coating even after performing rinses and, therefore, the polymer does not need to be included in the buffer. Polybrene (PB) has been used to prepare physically adsorbed polymeric coatings [59,60]. PB is a polycationic polymer, composed of quaternary amines (N,N,N ′ ,N ′ -tetramethyl-1,3propylenediamine), which strongly adsorbs to the inner surface of the capillary and reverses the

Glycoprotein Analysis by Capillary Electrophoresis

643

(a) DMF

(b)

DMF

(c)

DMF

(d)

0

DMF

10

20

30

40

50

Time post-injection (min)

FIGURE 22.7 Effect of various concentrations of diaminopropane on the electrophoretic migration and separation of hCG (4 mg/mL). Separation capillary: 100 cm × 50 µm fused silica. Separation buffer: 25 mM borate (pH 8.8). Other separation conditions: 25 kV, 28◦ C. Detection: 200 nm. The diaminopropane concentrations used were: 0 (a), 1.0 (b), 2.5 (c), and 5.0 mM (d). (From, Morbeck, D.E. et al., J. Chromatogr. A, 680, 217, 1994. With permission.)

EOF in acidic buffers. In these buffers, most proteins are positively charged and are repelled from the surface. This strategy has been used for the separation of rhEPO glycoforms [37], however, with a lower resolution than that obtained with the diaminoalkane putrescine (1,4-diaminobutane, DAB). Its main advantage resides in that this coating allows the use of volatile buffers compatible with MS. Glycoforms of ribonuclease B (RNase B) and horseradish peroxidase (HRP) [61] have also been separated using this strategy. Other polymers prepared from 1,3-propylenediamine and having different charge density have also been reported [62]. PB has a limited stability and is degraded after a few separations. Better coating stability was shown for multiple ionic polymer layers. They are prepared from a first layer made up of a cationic PB coating that is then covered with another physically adsorbed layer from an anionic polymer such as dextran sulfate [63] or poly(vinylsulfonate) [64]. In these coatings the EOF is essentially constant at most pH values. However, because of the negative charge at the capillary surface, positively charged proteins could adsorb to the coating and as such cannot be separated in this type of capillaries. The use of a third layer composed of a cationic polymer, usually PB, has overcome this limitation and has improved the stability of the coating [63,65]. Several groups [66–69] have used commercial

644

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

reagents that produce a double ionic layer (e.g., CEofix-CDT kit; Analisis, Namur, Belgium) for the analysis of Tf isoforms in human serum as will be shown in detail in Section 22.5.2. Recently, a new polyacrylamide (PAA) derivative that allows dynamic coating (UltraTrol LN-; Target Discovery, Palo Alto, CA) has been reported for the efficient separation of glycoforms of ATT [70]. A comparison of the ability of PB and UltraTrol LN for the separation of bovine alpha1-acid glycoprotein (AGP), bovine serum fetuin (BSF), and rhEPO sialoforms and their application coupling with MS has been reported [30,71]. Better resolution of sialoforms was obtained with capillaries coated with UltraTrol-LN in acidic volatile buffers used for coupling CE to ESI-TOF MS. 22.4.2.1.3 Capillary Coatings by Covalent Binding The covalent binding of polymers to the inner capillary walls leads to permanent coatings with a thin layer of polymeric material. In a way similar to physically adsorbed polymeric coatings, they minimize protein adsorption by shielding the silanol groups at the surface. In addition, these types of coatings eliminate or fix the EOF over a wide range of buffer pH allowing for a better control on the separation. However, good permanent coatings are more difficult to prepare than dynamically prepared coatings. Typically, preparation is carried out in a two-step process. First, a bifunctional compound, usually a silane derivative, is reacted with the silanol groups at the capillary surface through one of its functional groups, giving rise to a first layer and leaving the second functional group available to be covalently bound to the second layer. The second layer is produced by reacting a monomeric unit and polymerizing it. In some cases, an out-of-column synthesized polymer is bonded directly to the first layer. Several hydrophilic polymers such as linear and cross-linked PAA [72,73], methylcellulose and dextran [74], and polyvinyl alcohol (PVA) [75] to cite only a few have been developed for the separation of proteins. Commercial hydrophilic-coated capillaries have also been used for the separation of recombinant human tissue plasminogen activator (rhtPA) glycoforms [76]. The importance of using coated capillaries for glycoform separations has been shown by Thorne et al. [77] who demonstrated that rhtPA does not migrate in uncoated capillaries, even when εaminocaproic acid was added to the separation buffer. However, when the same buffer was used with capillaries having hydrophilic coatings, rhtPA could be separated into several isoforms. They also obtained better separations using PVA-coated capillaries than with PAA-coated ones. The importance of additives to the separation buffer, even for coated capillaries, should be emphasized here. Both zwitterionic buffers and tensioactive additives have been shown to improve the separation and recovery of rhtPA with PVA-coated capillaries [77]. On DB1 capillaries, a dimethylpolysiloxane coating widely used in gas chromatography (GC), optimal separations of fetuin, rhEPO, and AGP isoforms were obtained with acidic acetate buffers containing 0.4–0.5% (w/v) hydroxypropylmethycellulose (HPMC). In this case, the addition of a neutral polymer to the separation buffer makes the EOF almost negligible and protein interactions with the inner surface of the capillary were prevented [78]. The importance of the elimination of protein adsorption to the surface to obtain good glycoform resolution should also be emphasized. This is well illustrated in the separation of rhBMP-2 mentioned earlier, where the use of a commercial hydrophilic-coated capillary with a phosphate buffer without any additive led to efficiencies in excess of 7 × 105 plates/meter and up to nine isoforms differing only in the number of mannose residues being separated (Figure 22.8) [31]. 22.4.2.2 Factors Affecting the Resolution of Glycoform Peaks Owing to the importance of the capillary on resolution, the previous section was devoted to its coating. Besides preventing or diminishing interactions with the silanol groups on the silica surface, the capillary coating modifies the EOF. In this way, the apparent electrophoretic mobilities of the glycoform peaks are modified with respect to their effective mobilities, thus allowing for the modulation of the separation. In addition, some of the compounds used as dynamic coatings have been

Glycoprotein Analysis by Capillary Electrophoresis

Absorbance (200 nm)

1.0

645

2

3

4

5

0.8 6 0.6

1 7

0.4

8 N = 350 000/50 cm R = 1.12

0.2

9

0.0 22.0

22.5

23.0

23.5

24.0

24.5

25.0

Migration time (min)

FIGURE 22.8 CZE separation of rhBMP-2 in a 50 cm × 50 µm i.d. Bio-Rad coated capillary. Sample was electroinjected for 4–8 s at 6–12 kV in 0.1 M phosphoric buffer (pH 2.5). Detection: 200 nm. Separation temperature 20◦ C. (From, Yim, K. et al., J. Chromatogr. A, 716, 401, 1995. With permission.)

shown to play a role in the separation of glycoforms through interactions with glycoproteins or their complexes, and in this sense they will be considered in this section. In principle, each CE mode should be adequate to perform the separation of intact glycoforms as long as they differ in molecular properties that are distinguished in the particular mode used. When taking into account that the change in mass between two glycoforms is often almost negligible compared with the total mass of the protein and that relative changes in charge, when occurring, are more noticeable, it is not surprising that most separations have been achieved by CZE and CIEF. However, the other modes, MEKC, CGE, ACE, and CITP have also proven to be useful in some instances. The respective advantages and drawbacks of different modes have been discussed for some glycoproteins [36,76,77,79,80]. The following discussion will focus on aspects of the effect of factors affecting glycoform resolution that can be helpful for methods development together with providing examples to illustrate them. Although this section is organized according to separation modes, it does not imply that in some instances more than one mechanism may be involved in the separation achieved. Details can also be found in Section 22.5.

22.4.2.2.1 Capillary Zone Electrophoresis CZE was used for the separations of forms of glycoproteins, namely iron-free Tf and RNase almost two decades ago [79,81]. Different factors are known to influence the CZE separation of glycoforms and, when possible, the influence of individual variables will be considered. However, as it will be mentioned, in several instances an interconnection between the effects of two or more factors exists. The resolution between two peaks can be described [45] as

RS =

µe N 1/4 , 4 (µEOF + µe )

(22.2)

where µe is the difference in the electrophoretic mobility between two peaks, µEOF is the electroosmotic mobility, and µe is the mean electrophoretic mobility of the two peaks. Consequently, a decrease in the µEOF should result in an increase in resolution for other factors remaining constant. An option to selectively modify µEOF is through manipulation of the zeta potential of the capillary

646

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

and ways of achieving low values of µEOF by using dynamically or covalently coated capillaries have been described in the previous section of this chapter. Other options to increase resolution would be to increase the capillary plate number, to decrease the mean electrophoretic mobility, to increase the difference in the electrophoretic mobility of the two peaks, or to match the electroosmotic and the mean electrophoretic mobilities of different signs so the vectorial result is close to zero. The following discussion will briefly present the influence of some factors on these parameters. Buffer additives, which can act as capillary coatings, will also be considered in this section with respect to their effects on resolution by a joint action with other additives.

22.4.2.2.1.1 Buffer pH A variable that is typically changed to optimize differences in effective mobilities between glycoforms is the buffer pH. This frequently used approach derives from the potential improvement in resolution when using a buffer at a pH close to the (pI) of the glycoprotein where both its charge and mobility are lower. In other words, the closer the pH and pI values the higher the chances of increasing charge differences between two glycoforms. An example of this effect was shown for rhtPA, a protein with pIs in the range 4.5–6.0, for which better resolution was obtained at pH 3.6 than at pH 2. [82]. The enhancement of resolution is frequently accompanied by an increase in the analysis time as a consequence of the decrease in apparent mobility. For example, enhanced resolution at the cost of doubling the migration time was obtained by decreasing the pH from 5.5 to 4.5 when analyzing the novel erythropoiesis-stimulating protein (NESP), a protein known to be more acidic than rhEPO [83]. The same effect was observed for AGP, a very acidic protein (pI 1.8–3.8), when the buffer pH was changed from 4.5 to 3.5. In this case, the increase in analysis time was compensated by using a shorter capillary, while maintaining baseline resolution (Figure 22.9) [84]. While the previous examples were carried out on bare fused-silica capillaries, a similar phenomenon has been observed on coated capillaries. Resolution of fetuin (pI 3.2–3.8) glycoforms on DB-1 coated capillaries gradually increased for the pH buffer going from 9.0 to 5.0 [78]. In some instances, the choice of the most appropriate buffer pH precludes the need for buffer additives. This is well illustrated in a study of rhGCSF that showed that although additives, mainly DAB, could be effective in enhancing the separation of glycoforms, the pH control in borate buffer

0.03 (c) 1

Absorbance at 214 nm

0.02

2

3

4

5 6

7 8 3 1 2

(b)

4

5

6

7

8

0.01 (a) 0 0

10

20

30

50 Time (min)

–0.01 –0.02

40

EOF marker

–0.03

FIGURE 22.9 CZE electropherograms of standard human AGP. Conditions: (a) uncoated capillary, 87 cm × 50 µm; separation buffer: 0.01 M Tricine, 0.01 M NaCl, 0.01 M sodium acetate, 7 M urea and 3.9 mM putrescine, pH 5.5; injection 30 s at 0.5 psi; 25 kV; 35◦ C; detection at 214 nm; pH of separation buffer 4.5, other conditions as in (a); (c) capillary length 77 cm, other conditions as in (b). (From Lacunza, I. et al., Electrophoresis, 27, 4205, 2006. With permission.)

Glycoprotein Analysis by Capillary Electrophoresis

647

was enough to obtain baseline resolution of the two glycosylated, the nonglycosylated, and the desialylated forms [54]. 22.4.2.2.1.2 Buffer Ionic Strength The ionic strength of the separation buffer is another factor to be taken into account when trying to optimize glycoform resolution. Generally, an increase in ionic strength increases resolution by decreasing the EOF and the analyte apparent mobility. However, anomalous results have been observed. This is the case for formic acid used as the separation buffer in the CZE-MS analysis of RNase B on PB-coated capillaries. Increased resolution and decreased EOF were observed when the formic acid concentration was increased from 0.1 M to 2.0 M. Interestingly, at 0.5 M formic acid, resolution was worse than for any other concentration although the EOF was lower than at 0.1 M formic acid. This behavior was interpreted as resulting from conformational changes of the protein [61]. 22.4.2.2.1.3 Borate Buffer In view of its ability to resolve peaks not solely based on charge differential borate has become a substance of major importance in the resolution of glycoforms. The formation of complexes between hydroxyl groups and borate ions is at the basis of the separation mechanism, making it possible to resolve peaks of glycoforms with the same charge. Borate complexation is usually favored with cis 1,2-diols over trans 1,2-diols, but some carbohydrates do not follow this rule [85]. For a constant amount of carbohydrate, the complex concentration increases with increasing borate concentration and pH as a result of higher alkaline borate ion concentration, which is known to be the species that complexes with diols [86]. Asialo fetuin analyzed in DB-1 coated capillaries with Tris-borate buffer at pH 8.5 and containing HPMC is an example. The addition of the cellulose polymer, which shields the capillary surface from interactions with the protein, was necessary to achieve resolution; however, excess HPMC hampered the separation, probably due to higher viscosity [78]. Depending on the polymer concentration a sieving effect could also be in play. A comparison of the separation buffers Tris-glycine, ammonium formate, and borate for the analysis of hCG glycoforms also showed an enhancement in resolution when using the borate buffer [56]. However, borate is not necessarily always the buffer of choice and depending on the glycoproteins and the specific separation conditions, other buffers provide better resolution [87]. 22.4.2.2.1.4 Borate Buffer and Diamino Additives The combined action of borate and diamino additives in the separation buffer reported in 1992 by the groups of Taverna and Landers [50,82] has been widely used to achieve separation of glycoforms. In addition to those reported by the two groups, other types of bifunctional alkyl amines or bis-quarternary amines have been reported in separations involving borate-based buffers [56,57]. The necessity for the diamine to exist as a divalent cation was stated in these studies and a mechanism in which the borate-complexed protein interacts with the amino group of the capillary-bound additive was suggested. Such a partition mechanism that would involve CZE and electrochromatography components has been accepted by other authors [88]. 22.4.2.2.1.5 Other Components of the Separation Buffer The use of zwitterions as buffer additives has the advantage of not affecting the conductivity of the separation buffer. These additives are thought to enhance glycoform resolution by suppressing protein–wall interactions, by forming ion pairs with the glycoproteins, and by preventing individual protein molecules from interacting with lysine residues on adjacent molecules resulting in precipitation. For example, EACA [76] and other similar ω-amino acid buffers have proven effective in the separation of rhtPA forms [77]. In this instance, the addition of a nonionic detergent such as Tween 80 increased the protein recovery. Modified celluloses added to a borate buffer in DB-17 coated capillaries allowed the separation of Tf sialoforms. Although the separation mechanism is not clear, it is thought that both charge differences and gel sieving are contributing factors [89]. 22.4.2.2.1.6 CZE Separation of EPO Glycoforms To illustrate the many factors that can be modified to optimize glycoforms resolution it is interesting to follow the different approaches reported

648

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

for the CZE separation of rhEPO, starting with the method reported by Watson and Yao [52]. On a bare fused-silica capillary, no separation was possible for a commercial rhEPO sample at pH values between 5 and 10 using buffers of different ionic strengths. However, at pH 6.2, the use of additives that decreased the EOF allowed some resolution, with the best results achieved with 2.5 mM DAB. Further improvement was achieved by adding 7 M urea to the buffer and the optimized conditions consisted of a separation buffer of 10 mM tricine, 10 mM NaCl, 2.5 mM DAB, 7 M urea, pH 6.2, in a 50 cm bare fused-silica capillary and applying 10 kV. Baseline resolution of six peaks was obtained in 35 min. A modified version of this method was developed and used for an international collaborative study to assess the BRP for rhEPO of the EP, which consisted in a mixture of alphaand beta-epoetin. [34]. The modified method consisted in the addition of 10 mM sodium acetate to the separation buffer, adjusting the pH to 5.5 with 2 M acetic acid, and performing the separation at 30◦ C in a 107 cm capillary. In the initial study, large discrepancies in migration times were found among participants (between 32 and 120 min). These results were possibly attributable to differences of the capillary walls since capillaries from different manufacturers and even from different batches from the same manufacturer have been shown to have very different EOFs. This method was later implemented as an official method for the identification of rhEPO (see Section 22.5.1.1) [35]. A reduction of migration times from 70 to 36 min while keeping enough resolution between peaks and good repeatability has been achieved by decreasing the DAB concentration to 0.025 mM [36]. In contrast, the increased analysis time at higher DAB concentrations could be reduced by applying a higher voltage, from 15.4 to 30 kV [90]. In this instance, baseline resolution of glycoforms have been performed in less than 30 min, and this method could also be applied, although without baseline resolution, to the NESP [38,43]. The development of several CZE methods for the same glycoprotein, using buffers and capillaries other than those previously mentioned, clearly shows the versatility of CE. As another example, the early work by Tran et al. [91] on the influence of factors such as buffer pH and type, and the addition of organic modifiers showed the possibility of obtaining partial resolution of multiple rhEPO peaks with a 100 mM acetate–phosphate buffer at pH 4 in uncoated fused-silica capillaries. A marked effect on resolution was observed when the pre-equilibration time before using the capillary was increased from 4 to 10 h. The shorter pre-equilibration time and the same separation conditions have been used by others [92]. On the other hand, a different method for rhEPO analysis has been performed on C8-coated capillaries using a phosphate buffer [93]. A 300 µm i.d. tube made up of fluorinated ethylene–propylene (FEP) copolymer was used in a hydrodynamically closed separation system, which allowed enhancing the loadability. A buffer consisting on N-(2-hydroxyethyl)piperazin-N ′ 2-(hydroxypropane sulfonic acid) (HEPPSO), 1,3-bis[tris(hydroxymethyl)-methylamino] propane (BTP), and methylhydroxyethylcellulose (MHEC) 30,000, at pH 7.25 allowed to obtain partial resolution of seven bands for rhEPO [94]. CZE under similar conditions with a borate buffer at pH 8.8 containing MHEC was used to monitor the fractionation of rhEPO by preparative CITP [95]. The coupling of the CZE step to detection systems other than UV has required the development of separation conditions compatible to the detection system used. For instance, the presence of primary amines, such as DAB, in buffers needed to be avoided for compatibility with laser-induced fluorescence (LIF) of compounds derivatized with fluorogenic substrates through their amino groups [90]. Baseline resolution of eight peaks in approximately the same time was achieved by substituting DAB by morpholine and tricine by boric acid (to avoid potential traces of primary amines in the tricine buffer) and by adjusting the concentration of other buffer components to compensate for the increase in electrical current. In the same work, modifications were also required to achieve compatibility with MS detection where nonvolatile salts, urea, and amines should be usually avoided. A physically adsorbed polyethylenimine-coated capillary was used to overcome protein adsorption to the capillary walls in the absence of cationic additives and the use of an acetate buffer at pH 5.05 allowed the partial resolution of at least five bands of rhEPO. Other types of coated capillaries have been used for the analysis of EPO by CE-MS as detailed in Section 22.4.3.3 [30,37,42,62,96].

Glycoprotein Analysis by Capillary Electrophoresis

649

22.4.2.2.2 Capillary Isoelectric Focusing Almost two decades ago, CIEF was described by Kilar and Hjerten [79,97] for the separation of Tf glycoforms and it has, in principle, some advantages over other CE modes, namely, (1) the large volume of sample introduced in the capillary would result in an increase in sensitivity compared with nanoliters amounts in other modes; (2) the measurement of an important molecular property, the isoelectric point (pI), is determined; and (3) a large number of experimental factors can be modified to improve resolution. With regard to the first two advantages, experimental considerations made it that their importance is not what it should be. With respect to the sensitivity issue, it should be stated that commercial ampholytes are generally not designed to work in a capillary format with UV detection and they show high absorbance at 214 nm. In fact, they are manufactured for classical IEF slab gels with detection at 280 nm, a wavelength at which the extinction coefficient of proteins is lower than at 214 nm. So, in practice, the sensitivity of CIEF has been shown in some instances to be similar to that obtained for the same glycoprotein by CZE [36]. With respect to the pI determination, the value obtained for a given analyte should be considered cautiously since, in many instances, urea or other additives that affect the pI of a protein are used in addition to the fact that the pH gradient is not always linear [98]. CIEF can be performed in two general ways: a one-step or a two-step process. In the onestep process, the capillary retains a residual EOF that allows protein bands to migrate through the detection window, while in the two-step process the absence of EOF is desirable for better focusing of the protein bands being mobilized to pass the detection point in the next step. However, even in the one-step mode a controlled EOF is necessary to achieve enough focusing before reaching the detector. Both modes have advantages and drawbacks as was shown for the effect of different factors in the separation of AGP [99]. In CIEF, the total time elapsed between sample introduction and detection depends on the focusing time and on the time for a band to reach the detection window during mobilization. These two effects can take place either simultaneously or consecutively depending on whether the process is a one-step or two-steps, respectively. So, the term “migration time” is not strictly correct in CIEF but it is used in this chapter for simplicity. It is important to note that the need for mobilization through the detector window can be eliminated in systems where whole-capillary imaging detection is used [100]. Samples are introduced in the capillary either separated or mixed from ampholytes, the latter being the most frequently used. In addition to the sample and ampholytes, additives can also be introduced in the capillary in the so-called sample mixture, as will be shown in later text. The different types of capillary coatings providing reduced or no EOF have been described in a previous section. A comparison of different capillary coatings for the resolution of rhtPA glycoforms has been performed and, in some cases, capillaries with different degree of EOF reduction have shown similar resolution, a situation that was probably due to the presence of additives controlling the EOF [101]. While similar resolution has been achieved with different coatings, the stability of the coatings in terms of the number of runs performed was evaluated in the analysis of recombinant human antithrombin III (rhATIII) and the recombinant human immunodeficiency virus envelope glycoprotein rgp 160s. Higher stability was observed for dextran-coated [102] and PVA-coated [103] capillaries than for PAA-coated capillaries. Besides the capillary coating, other variables that can be controlled for improving resolution of glycoforms are described in the following text. 22.4.2.2.2.1 Additives for Reducing EOF Besides the use of coated capillaries for controlling the EOF, the addition of substances aimed toward reducing the EOF through an increase in the viscosity is frequently used. This is the case for poly(ethylene oxide) [36,39,77] and HPMC [101,103]. The latter has proven to be effective even when using bare fused-silica capillaries, where it acts as a dynamic coating [104]. 22.4.2.2.2.2 Ampholytes This is one of the variables to which a lot of attention has been paid. Both the nature and pH of ampholytes as well as the combination of several pH ranges have been

0.16 (a) 0.12 0.08 0.04 0

10

–0.04

15

20

25

30

35

40

Time (min)

0.025 (c) 0.02 0.015 0.01 0.005 0 –0.005

Absorbance (280 nm)

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

10

20 Time (min)

30

40

Absorbance (280 nm)

Absorbance (280 nm)

Absorbance (280 nm)

650

0.025 0.02

(b)

0.015 0.01 0.005 0 –0.005

0.025 0.02

10

15

20

25

30

35

40

Time (min)

(d)

0.015 0.01 0.005 0 –0.005 10

15

20

25

30

35

40

Time (min)

FIGURE 22.10 Effect of types of ampholytes on separation of AGP forms. Total ampholyte concentration in the sample mixture: 6.3% (v/v). Types of ampholytes, specified by their pH range (and ratios (v/v)): (a) range 3–10; (b) mixture of ranges 3–10 and 2.5–5 (1:1); (c) mixture of ranges 3–10, 2.5–5, and 2–4 (1:1:1), (d) mixture of ranges 3–10, 2.5–5, 2–4, and 3–5 (1:1:1:1). Other analytical conditions: PVA-coated capillary, 27 cm (effective length 20 cm). Sample mixture in CIEF gel: 1 mg/mL AGP, 5.6 M urea, 20 mM NaCl, and ampholytes. Catholyte: 20 mM NaOH titrated with H3 PO4 to pH 11.85. Anolyte: 91 mM H3 PO4 in CIEF gel. Focusing: 10 min at 20 kV. Mobilization step: 20 kV and 0.5 p.s.i. N2 pressure. Temperature: 20◦ C. Detection: 280 nm. (From Lacunza, I. et al., Electrophoresis, 28, 1204, 2007. With permission.)

considered. In general, better resolution is attained with mixtures of ampholytes. The influence of the commercial source of ampholytes on resolution has been shown, even when only a wide-range ampholyte solution is used [77]. Similarly, for a given total concentration of ampholytes mixed in constant proportions, the resolution was shown to depend on the nature of the ampholytes [105]. Besides providing different resolutions of rhtPA glycoforms, commercial ampholytes have different optical properties that have an impact on detection [106]. Ampholytes from different sources have been shown to be different in nature and in focusing properties [107,108]. As such, mixtures from different manufacturers are generally preferred, since the different nature of the compounds from the different sources may help to achieve better focused zones by reinforcing the pI range of interest. This effect was clearly seen for the separation of rhEPO [36] andAGP glycoforms [109] (see Figure 22.10). Apart from the nature of ampholytes, their proportion and total concentration are other factors to be considered. Usually, higher proportions of an ampholyte or narrow-range ampholytes close to the pI of the glycoprotein provide better resolution [103]. For the amount of total ampholyte, a compromise is generally achieved since higher percentages not only provide better resolution but also lead to higher background absorption and, consequently, decreased sensitivity [39,103,109]. 22.4.2.2.2.3 Agents to Avoid Precipitation Protein precipitation is favored during the focusing process where a protein is concentrated in a narrow zone close to its pI, and associated salts are separated from it. The addition of nonionic surfactants, organic modifiers, or chaotropes to the sample can be used to decrease chances of precipitation [110]. The presence of urea in the separation buffer has been shown to markedly improve the resolution of AGP an EPO glycoforms [36,100,109], being also used to obtain adequate separation of rhtPA glycoforms [77,104]. Triton X-100 was found to be unnecessary when urea was used to separate rhtPA glycoforms [101]. This detergent was used to avoid precipitation for the analysis of rhATIII [102]. However, it may be necessary to resort to a combination of agents. This was the case for recombinant human immunodeficiency virus envelope glycoprotein where several additives including a chaotrope (urea), a

Glycoprotein Analysis by Capillary Electrophoresis

651

Absorbance

0.02

(a)

–0.015

0.01

0.005

0

–0.005

0

0.005

2

4

6

8

10

12

14

16

18

20

16

18

6

(b)

5

0.004

7

4

0.003 0.002 3

0.001 2 0

1

–0.001 –0.002 0

2

4

6

8

10

12

14

Time (min)

FIGURE 22.11 Effect of urea on CIEF separation of rhEPO. (a) Without urea in the sample mixture, and (b) with 7 M urea. Analytical conditions: polyacrylamide-coated capillary 27 cm (effective length 20 cm) × 50 µm i.d. Focusing: 6 min at 25 kV. Sample mixed with CIEF gel and ampholytes. Ampholyte mixture (1:2, v/v) of pH 3–10 and 2.5–5 ranges. Detection: 280 nm. (From Cifuentes, A. et al., J. Chromatogr. A, 830, 453, 1999. With permission.)

zwitterion [3-(cyclohexylamino)-1-propanesulfonic acid, CAPS], and a sugar (saccharose) provided the best results [103]. 22.4.2.2.2.4 pH Extenders In order to avoid the focusing of glycoforms past the detection window, a situation that would prevent their detection, it is necessary to add a substance with an appropriate pI to block that zone of the capillary. It is referred to as a pH extender and the most frequently used is N,N,N ′ ,N ′ -tetramethylenediamine (TEMED). It is usually added to the sample mixture at a concentration proportional to the length of capillary to be blocked. The larger the pH extender concentration is the farther away a glycoform will focus from the detection point and, consequently, the greater the migration time will be, as was observed for rhtPA without any marked

652

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

influence on resolution [105]. The same authors also reported that for TEMED concentrations about 10 times higher, in the range of 3.75–7.5% (v/v), the effect on migration time was not noticeable [101]. Caution should be taken when using TEMED as it has been shown to damage the capillary coating [99,103]. Depending on the pI of the glycoprotein, other less basic extenders, such as alanine, can be used [99]. 22.4.2.2.2.5 Focusing Time and Voltage For the one-step CIEF separation of rhtPAglycoforms, an increase in voltage led, as expected, to a decrease in migration times without any influence on resolution [105]. For the two-step process, the two factors involved on the focusing step (i.e., voltage and time) have an impact on resolution. According to the theory [111], increasing voltages up to 25 kV provided slightly better resolution of rhEPO glycoforms; however, loss of resolution was observed at higher voltage, probably due to an excess of Joule heating [36]. It is usually considered that focusing is finished when the current decreases below 10% of its initial value and stabilizes. However, in practice, focusing times as short as 0.5 min and corresponding to 69% of the initial current have been shown to lead to the same resolution of EPO isoforms as for longer focusing times. This can be understood taking into account that during the pressuremobilization step, the voltage is applied to avoid band broadening resulting in a continuation of focusing [36]. In contrast, focusing times longer than needed may have a detrimental effect not only on the analysis time but also on the quality of the separation. The total depletion of salts due to a long focusing time led to precipitation of monoclonal anti-alpha-1 antitrypsin [100]. In addition, focusing times that are too long may lead to greater anodic drift and can affect the isoform separation of highly acidic proteins such as AGP [109]. Incomplete focusing due to times shorter than needed, however, may lead to double peaks. These examples show that careful optimization of the focusing time is required. 22.4.2.2.2.6 Presence of Salts in the Sample Mixture The presence of salts in the sample mixture remains a controversial point. Some authors have claimed that the ionic strength of the sample should be as low as possible while others have found it to help improve resolution. For the analysis of anti-alpha-1 antitrypsin [100], AGP [109], or rhEPO [36,39], the presence of salt was necessary for the separation of isoforms. This effect is likely related to prevent protein precipitation and denaturation, although other effects such as improved focusing or masking of the silanols of the capillary walls cannot be excluded. 22.4.2.2.2.7 Mobilization Step In the two-step CIEF mode the mobilization step is performed after focusing. Hydrodynamic (more frequently performed by applying pressure rather than by vacuum) and chemical mobilization are the two general ways of driving the focused zones to the detection point. As indicated earlier, when the mobilization is performed by pressure, a voltage is applied simultaneously, usually at the same strength as the voltage used during the focusing step. However, a higher voltage may be needed in some instances as shown for a monoclonal antibody (mAb) [112]. Chemical mobilization with a strong electrolyte has been shown to lead to better resolution than by using a weak one [97]. Performing hydrodynamic and chemical mobilizations together has been shown to be beneficial for the separation of rhEPO [39] and even necessary for the separation of glycoforms of acidic proteins such as AGP [109] (see Figure 22.12). The addition of phosphoric acid to the catholyte, which consists of sodium hydroxide, at the very beginning of the focusing step promotes chemical mobilization that continues when the pressure is applied at the hydrodynamic mobilization step. 22.4.2.2.3 Micellar Electrokinetic Chromatography There are many factors involved in the MEKC separation mode that contribute to the complexity of its mechanism. On the other hand, they provide a large number of variables that can be modulated to optimize the separation. MEKC has been used on a few occasions for the separation of glycoproteins,

Glycoprotein Analysis by Capillary Electrophoresis

653

0,05

Absorbance (280 nm)

0,04 0,03 (b)

0,02 0,01

(a) 0 0 –0,01

5

10

15

20

25

30

Time (min)

FIGURE 22.12 One-step CIEF of AGP. PAA coated-capillary 27 cm (effective length 7 cm) × 50 µm i.d. Voltage: –20 kV. Temperature: 20◦ C, Detection 280 nm. Sample mixture: 5.6 M urea, 1.7% (v/v) TEMED, 9.7 % (v/v) ampholytes in the following distribution of pH ranges: 3–10, 2.5–5, 2–4, 3–5 (2:2:3:3) and 1 mg/mL AGP in CIEF gel. (a) anolyte: 91 mM H3 PO4 in CIEF gel, catholyte: 20 mM NaOH, (b) anolyte: 91 mM H3 PO4 in CIEF gel, catholyte: 20 mM NaOH titrated to pH 11.85 with H3 PO4 .

usually under conditions involving the use of borate or phosphate salts and SDS as the micellar agent. Factors, such as borate/phosphate and SDS concentrations, and the nature and pH of the buffer can be modified to optimize resolution. RNase B was separated into five peaks, each one of them corresponding to one of the Man5 to Man9 oligomannose structures, in a fused-silica capillary with a buffer consisting of sodium phosphate, SDS, and sodium tetraborate [113]. For recombinant human interferon-γ (rhIFN-γ), over 30 species were resolved giving rise to three groups of peaks, PG1, PG2, and PG3, as shown in Figure 22.13 [114]. The increase in the concentrations of the borate buffer and SDS led to a concomitant increase in separation efficiency and migration times. The comparison between phosphate buffers at pH 6.5 and 7.5 and borate buffers at pH 8.5, at a constant SDS concentration and at buffer concentrations chosen to generate similar currents, showed that the best resolution was obtained with borate buffers. At a higher pH (9.5) and a lower borate concentration than that used at pH 8.5, higher current, slower separation, and lower efficiency were observed. The better resolution obtained with the borate buffer pH 8.5 was likely due to the effect of the complexation between borate anions and diols mentioned earlier. It has been hypothesized that the early migrating glycoforms would be those with the larger glycan structures and that this should be taken into account when designing a separation method or when assigning peaks to specific glycoforms. The reasoning behind this hypothesis is that at high borate concentrations, sugar residues would be extensively complexed and that the resulting negative charge, when added to that from sialic acid residues, would cause repulsion between SDS micelles and the borate-complexed glycoprotein. According to the authors, the repulsion intensity would depend on the glycan size. The same study reported the possibility of separating the glycoforms of RNAse B and HRP, while the separation of BSF led to the separation of only one major peak and three minor, broad ones, and no separation at all of AGP glycoforms. According to the authors, the lack of separation for proteins glycosylated at multiple sites such as BSF and AGP could be due to the fact that the heterogeneity of individual oligosaccharides does not give rise to measurable differences in carbohydrate contents. 22.4.2.2.4 Capillary Electrophoresis in SDS In conventional polyacrylamide gel electrophoresis (PAGE) as well as in its capillary counterpart CE-SDS, proteins form stable complexes with SDS and are separated on the basis of their size. The separation of glycoproteins is known, however, to deviate from the linear relationship of migration versus MW observed with nonglycosylated proteins. This anomalous behavior arises from the fact that carbohydrate moieties bind with much lower amounts of SDS than expected for a corresponding

654

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a)

4.5

5.0

5.5

6.0

6.5

Absorbance at 200 nm

(b)

7

8

10

9

(c)

10

12

14

16

18

20

(d) PG1 PG2

PG3

15

20

25

30

35

40

Migration time (min)

FIGURE 22.13 Optimization of the separation of IFN-γ glycoforms by MEKC. Borate/SDS buffer at the following concentrations: (a) 40 mM borate, 10 mM SDS; (b) 40 mM borate, 100 mM SDS; (c) 400 mM borate, 10 mM SDS; (d) 400 mM borate, 100 mM SDS. Fused-silica column 57 cm × 50 µm i.d. Injection 5 s of 1 mg/mL protein in 50 mM borate, 50 mM SDS, pH 8.5. Voltage: 22 kV. In (d), the main groups are designated PG1, PG2, and PG3 in order of migration. (From James, D.C. et al., Anal. Biochem., 222, 315, 1994. With permission.)

peptide of similar MW. Advantages were taken from this fact in order to differentiate glycosylated fetuin from the nonglycosylated protein produced by enzymatic hydrolysis [115]. The release of the carbohydrate moiety gives rise to a decrease in size that translates into an increase in the chargeto-mass ratio since the charge imparted by SDS to the carbohydrate moiety is low compared with that imparted by the interaction of the surfactant with the peptide core. The resulting differences in mobility make it possible to monitor the deglycosylation reaction. Similarly, rhATIIIα was separated from rhATIIIβ by CE-SDS due to the lack of glycosylation at Asp 135 in the latter [116].

Glycoprotein Analysis by Capillary Electrophoresis

2,3,4

655

3,4

(c) (b)

(a)

3,4 (d) 1

1

2

1

14

17

14

17

14

(e)

34 2

2

2 4 1 3

1

17 14

17

14

17

Migration time (min)

FIGURE 22.14 Changes in the migration times of OVA glycoforms with LCA concentration. Capillary: linear PAA-coated 58 cm × 50 µm i.d; Buffer: 100 mM phosphate pH 6.8, (A) not containing LCA, (b–e) containing 0.4, 0.8, 1.4, and 2.0 mg/mL LCA, respectively. Voltage: 20 kV; Detection: 214 nm. (From Uegaki, K. et al., Anal. Biochem., 309, 269, 2002. With permission.)

22.4.2.2.5 Affinity Capillary Electrophoresis Glycoform separations by ACE have been carried out by taking advantage of the known interactions between lectins and carbohydrates. The “partial filling technique” in which only a portion of the capillary is filled with an affinity ligand present in the separation buffer before injecting the sample was shown to be useful for separating AGP glycoforms. The separation of two peaks based on the dependence of the strength of the AGP-concanavalin A (Con A) interaction as a function of the content in biantennary glycans was achieved [117]. Differential interactions between lectins and glycans are useful not only to perform glycoform separations but also to estimate the values of the apparent association constant between the lectin and each of the separated glycoform. PB-coated and PAAcoated capillaries completely filled with a buffer containing the lectin Lens culinaris agglutinin (LCA) were shown to be suitable for separating glycoforms of RNase B and OVA, respectively, and for calculating the affinity constants without a prior separation of the glycoforms outside the capillary. Figure 22.14 shows the separation obtained for OVA glycoforms and the influence of the LCA concentration [118]. 22.4.2.2.6 Capillary Isotachophoresis Few studies on the use of the CITP electrophoretic mode have been reported for the separation of protein glycoforms. The separation of AGP isoforms by CIEF mentioned previously [109] can be considered to involve an isotachophoretic mechanism taking into account the concurrence of focusing and chemical mobilization [119]. Preparative CITP has been used to fractionate rhEPO with a leading electrolyte containing chloride as anion and BTP as counter ion, and a terminating electrolyte containing glycine as anion and BTP as counter ion. Under these conditions, the fractions contained mixtures of glycoforms enriched in the predominant glycoform of each fraction. The process was monitored by CZE [95]. 22.4.2.3 Reproducibility Problems With respect to the issue of reproducibility, the separation of protein glycoforms by CE is faced with the same kind of problems than for any other protein. However, the consequences of a lack of

656

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

adequate precision can have a detrimental impact due to the fact that differences in the charge/mass ratio or pI among isoforms are so small that the different peaks usually migrate very close to one another, making it sometimes impossible to compare samples. For example, peaks of rhBMP-2 separated by CZE on the basis of a single mannose residue between adjacent peaks differ in migration times by only 0.15–0.22 min or 0.01 × 10−4 cm2 V−1 s−1 in their respective mobility [31]. This can give rise to unreliable comparisons of samples since the difference in migration time between two consecutive peaks can be in the same range as the migration time dispersion for a given peak [38]. For instance, this problem made it impossible to correctly compare a reference hCG sample with a sample obtained from the urine of a patient with metastatic choriocarcinoma [56]. Solutions to this type of problems require a two-pronged approach. First, experimental factors affecting reproducibility must be carefully controlled. Once this is achieved, reliable migration parameters must be used for band assignment. Some experimental factors such as buffer preparation are usually well controlled. However, there are other variables to which not enough attention is paid. One of these, which is of key importance for achieving good intra- and interday repeatability, is capillary conditioning. All of the steps performed for the conditioning of brand-new capillaries as well as the washing steps carried out between runs and before the storage of capillaries are decisive in achieving good precision. This may help in explaining the low reproducibility reported for some methods [17], while the same methods when applied to the same analytes showed highly reproducible results in other studies [36,38,90]. Another factor that has a large influence on migration time reproducibility, especially for coated capillaries, is capillary aging. Buffer components that can alter the coating need to be identified and avoided. However, even when using the appropriate buffers, capillaries gradually degrade and they should be discarded if the resolution achieved is no longer within easily verifiable limits or if peak assignments cannot be reliably performed. Once experimental factors are controlled as much as possible there are several ways to minimize the problem of migration time reproducibility through the use of migration parameters. One of the more common approaches is the use of internal standards from which it is possible to calculate the migration time of each isoform relative to that of the internal standard. This approach has been used in the analysis of several glycoprotein isoforms [38,39,109,120]. For example, in the CIEF separation of rhEPO glycoforms, the intraday migration time reproducibility was improved from an RSD value of around 1.8–0.3% when using an internal standard [39]. In another approach and using the CZE analysis of rhEPO as a model, a statistical program was developed in order to choose, among others, the best migration parameter to correctly assign peaks and reliably compare different samples. The program estimated the probability of correct assignment for each migration parameter [38] and was used for the comparison of rhEPO from different samples [38,43] and for the comparison of electrophoretic profiles of AGP purified from healthy donors and cancer patients [84]. In the latter publication, it was shown that the electrophoretic mobility of each AGP peak and the migration time of each peak relative to the migration time of the EOF marker were more effective migration parameters than the migration time of each peak.

22.4.3 DETECTION AND IDENTIFICATION OF FORMS OF GLYCOPROTEINS 22.4.3.1 UV Detection The UV absorbance detector is one of the most widely used systems in CE. Detection takes place in the same separation capillary, avoiding band broadening of the separated compounds. Since the internal diameter of the capillary is very small (25–100 µm), the detection pathlength is very short and the concentration sensitivity is typically 10–20 times poorer than that obtained with the same type of detector in other separation techniques (e.g., HPLC). Glycoproteins, similar to nonglycosylated ones, can be monitored at 280 nm at which wavelength the aromatic residues of the polypeptide chain, tryptophan, tyrosine, and, to a lesser extent, pheny-

Glycoprotein Analysis by Capillary Electrophoresis

657

lalanine give relatively good absorption. However, the detection limit at this wavelength remains in the low mg/mL range and instead, detection is usually accomplished at 200 nm where proteins present 50- to 100-fold greater absorptivity. At this wavelength, sensitivities in the µg/mL range can be obtained. While this level of sensitivity may be enough for the quality control of glycoproteins, it is, in most cases, insufficient for the analysis of these proteins in clinical samples. Strategies based on sample concentration or protein derivatization can sometimes be used in these cases and some aspects have already been discussed in other parts of this chapter and will only be considered briefly in this section. 22.4.3.2 Laser-Induced Fluorescence Detection For glycoproteins present at low sample concentrations (generally lower than 10−6 M), such as in clinical samples, fluorescence detection will likely be one of the detection methods to be improved in the future for analysis of these proteins by CE. It should be recalled that, due to the small size of the detection volume in on-column monitoring (usually 5–10 nL), high optical power (>1 mW) from the light source should reach the detection window of the capillary for intense fluorescence excitation [121]. This is more efficiently achieved using a laser as a source of light than a ultraviolet-visible (UV-VIS) lamp and confirmed by practical experience of the fluorescence detection of proteins and glycoproteins by CE. LIF detection of glycoproteins is not an easy task. Most analytes do not produce fluorescence at the wavelength of the emission of the most frequently used lasers (those with light emission in the 400–600 nm range) and, therefore, native fluorescence cannot be obtained with these light sources. Derivatization with a fluorescent reagent is a useful strategy for LIF detection when native fluorescence is unavailable. Although several fluorogenic compounds with good chemical and spectroscopic characteristics have been described in the literature [122], this approach suffers from several problems when the analyte is a glycoprotein, including the formation of multiple reaction products [123] and the difficulties associated with derivatization of proteins at low concentration (below 10−8 M). In the following discussion, a few approaches reported in the literature to make LIF detection feasible for glycoproteins are presented. Glycoproteins can be directly detected by native fluorescence owing to the fluorescence emission of the tryptophan and tyrosine residues; however, this requires the use of laser with emission in the UV region of the electromagnetic spectrum. This approach has been scarcely used [124] for glycoproteins, probably due to the high cost of such lasers and the expertise required for their maintenance. In the past few years, however, pulsed diode-pumped solid-state (DPSS) lasers with emission in the UV region (266 nm) have become affordable and this has opened the way to a more universal native LIF detection of glycoproteins. The covalent derivatization of a glycoprotein with a fluorescent agent has been shown to be useful in some cases. As mentioned earlier, the main difficulty with this type of approach is the formation of multiple products due to the fact that each lysine residue in the peptide chain can react to a different extent with the derivatizing agent. As a result, protein molecules having a variable number of covalently bonded fluorescent tags are produced, resulting in a variable number of species with similar MW (the fluorescent moiety has a small MW in relation to that of the protein) but with different electric charges. In these cases, analysis by CZE or CIEF may give rise to broad or even multiple peaks for the same protein. On the other hand, when the separation mode is CE-SDS, narrow peaks with good efficiency are obtained and make LIF a more suitable technique for this separation mode. Using this approach, a CE-SDS method with LIF detection has been developed and validated as a quality control procedure for the purity determination of a recombinant mAb [125]. 5-Carboxytetramethylrhodamine succinimidyl ester was used as fluorescent reagent and the optimal conditions for derivatization of the nonreduced and reduced (reacted with dithiothreitol) mAb and

658

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

their respective degradation products were determined. It is worth mentioning that for reduced samples, the proposed method was able to separate the fraction corresponding to the nonglycosylated fragment of the mAb from the glycosylated one. The commercial LIF instrumentation (argon ion laser, λexc 488 nm, λem 560 nm) provided excellent limits of detection (around 10 ng/mL) for the intact antibody and its degradation products. One approach to overcome the formation of multiple derivatives and the difficulties associated with derivatization of proteins at low concentrations was proposed by the group of Dovichi [126,127]. They developed a derivatization procedure for proteins with the fluorogenic reagent, 5′ -furoylquinoline-3-carboxaldehyde (FQ), that provided good sensitivity and minimized the band broadening caused by the variable labeling from the reaction of FQ with lysine residues. The derivatization could be carried out before the sample injection or inside the separation column. For the latter, two successive short plugs, one containing the sample and the other containing the reagent, were injected into the column; both plugs were mixed inside the capillary (due to their differences in electrophoretic mobility) and allowed to react. After derivatization, either inside or outside the capillary, the protein separation took place in a buffer containing an-alkylsulfate tensioactive (e.g., sodium dodecylsulfate or sodium pentylsulfate) at submicellar concentrations, which masked the charge differences introduced in the protein by the multiple labeling. The method provided good results in terms of sensitivity, with an assay detection limit of 10−10 M for OVA (the authors define the assay detection limit as the minimum amount of protein that can be successively derivatized and detected), and band broadening (around 20,000 theoretical plates) (Figure 22.15). Limitations of the technique included variable protein sensitivity due to the different number of lysine residues and the possibility of masking differences in mobility of the protein glycoforms from the added tensioactive [128]. It should be noted that the reported sensitivity was achieved with postcolumn sheath-flow LIF detection; sensitivity using a commercial on-column LIF detector was about 100 times poorer.

Migration time (sec) 140

150

160

170

180

190

200 70 x 10–3

6

60

50 4 40 3

(c)

30

2

(d)

(b) 1

20

Absorbance (au)

Fluorescence signal (v )

5

10

(a) 0

0 180

200

220

240

260

280

Migration time (sec)

FIGURE 22.15 Comparison of ovalbumin analyzed under various conditions. Running buffer: (a–c) 25 mM tricine, pH 8.0; (d) 25 mM tricine +5 mM SDS. (a) FQ-labelled ovalbumin, 15 s reaction, LIF detection; (b) unlabelled ovalbumin, UV adbsorbance detection; (c) FQ-labelled ovalbumin, 10 min reaction. LIF detection. Note that electropherogram B is plotted versus top x-axis and right y-axis. (From Pinto, D.M. et al., Anal. Chem., 69, 3015, 1997. With permission.)

Glycoprotein Analysis by Capillary Electrophoresis

659

Affinity binding is another good alternative for fluorescent detection of glycoproteins. This kind of binding involves several types of interactions between the glycoproteins and the fluorescent reagent (excluding the covalent one) and include electrostatic, van der Waals, hydrophobic, and hydrogenbonding forces. A large choice of fluorescently labeled biomolecules can be used in this approach. The preparation of a homogeneously labeled biomolecule is in some cases a labor-intensive task involving chromatographic purification of the probe. Bornemann et al. [129] have used affinity derivatization with LIF monitoring to improve the selective detection of rhEPO. They first prepared a monomeric antigen-binding fragment (Fab) from the mAb 5F12 that bonded to a conformationally independent epitope of the N-terminal region of human EPO. This fragment was then labeled with the fluorescent dye, Alexa Fluor 488, yielding a mixture of labeled products (doubly and singly labeled and unlabeled) from which a homogeneous fraction was prepared by HPLC. This fraction was used as affinity reagent in CIEF with LIF detection (argon ion laser λexc 488 nm, λem 520 nm) for the detection of less than 100 pmol of rhEPO. Although this detection limit is far from the sensitivity necessary for the determination of this hormone in serum or urine, the authors believe that further refinement of the labeled Fab fragment and decrease of sample volume should permit the detection of smaller amounts of EPO. Postcolumn derivatization with a fluorescent reagent is another alternative. However, this approach requires the selection of a fast derivatization reaction to achieve good sensitivity and a careful design of the postcolumn reactor that minimizes the loss in resolution achieved in the separation capillary. To this end, affinity interactions of some biomolecules, such as antibodies, with other glycoproteins are especially useful. That has been the case reported by Kelly and Lee [130] who have developed a CE method with online postcapillary affinity LIF detection for the monitoring and quantification of microheterogeneities of a mouse antihuman follicle-stimulating hormone mAb in samples containing culture medium without further purification. The fluorogenic reagent, consisting in fragment B (BF) of protein A conjugated with fluorescein, was mixed with the effluent of the CE column using a homemade postcolumn capillary reactor without causing large band broadening. Affinity binding between the antibody variants and the BF–fluorescein occurred in the reaction capillary and the complexes were monitored by the LIF detector (argon ion laser, λexc 488 nm, λem 515 nm) placed in the same capillary. By binding the BF–fluorescein fragment to the antibody, the emission of the fluorescein moiety was enhanced, so that the mAb peaks were detected with a low fluorescence background. The detection of up to five peaks was reported and the authors speculated that this could be due to the presence of different isoforms of the antibodies differing in the sialic acid content of their oligosaccharides.

22.4.3.3 Mass Spectrometry Recent progress made in the development of MS ionization techniques has enabled its application to the analysis of intact glycoproteins. The most widely used ionization modes are ESI and matrix-assisted laser desorption ionization (MALDI). When considering the gains in sensitivity when compared with conventional UV detection, the structural information that can be derived from its use and the possibility of increasing the selectivity in the selected ion monitoring (SIM) mode, it is not surprising that MS has become a widely used detection system for coupling to CE. In this section, only those systems where the online coupling of CE to mass spectrometry (CE-MS), and for which MS is used as the detection (and information) system for the CE separation of glycoforms will be considered. Owing to the high heterogeneity and structural complexity of glycoproteins, a prior separation step is generally necessary to obtain qualitatively and quantitatively useful results. Otherwise, only broad signals are obtained. Despite the compatibility of CE with ESI-MS and the relatively easy coupling of the two techniques, appropriate separation buffers, usually consisting of volatile substances,

660

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

must be employed. This is one of the more difficult issues to the coupling of CE to MS where buffer electrolytes compatibility with MS involves choosing buffers that do not inhibit ionization and that, at the same time, provide reasonable glycoform separation [30]. Formic acid was used as the separation buffer in a pioneering work aimed at the characterization of glycoforms [61]. Analysis required constructing a CE–MS interface that could be adapted to a commercial CE cartridge. Under these separation conditions, the CE-MS system provided analysis of intact RNase B and HRP, a more complex glycoprotein in which the microheterogeneity arises from glycosylation at two residues and variability in the carbohydrate composition at each site. Figure 22.16 shows the CE-UV and the CE-MS electropherograms obtained for RNase B. Although volatile buffers are usually employed, successful analyses of intact glycoproteins have also been achieved by using low concentrations of nonvolatile compounds in separation buffers with low ionic strength [131]. Using PAA-coated capillaries and a separation buffer consisting of 50 mM β-alanine adjusted to pH 3.5 with acetic acid, two high mannose-containing proteins, RNAse B and rhBMP-2, were analyzed by CE-MS employing a home-built CE instrument. Reporter ions for carbohydrates generated by in-source fragmentation for the glycoforms separated by CE were used to provide information about the degree of glycosylation of the different isoforms. The presence of urea in the separation buffer was found to be needed to accomplish the separation of antithrombin III (ATIII) glycoforms from a commercial sample purified from plasma [132]. In order to make the CZE separation step compatible with the ESI-MS ion trap, special precautions about the decoupling of the needle were needed to minimize the amount of urea introduced into the spray chamber. Although the resolution was lower than that obtained by CZE-UV, it was enough to allow deconvolution of the multiply charged peaks, even for minor isoforms. It was also possible to distinguish isoforms differing in fucosylation for which small differences in mass for species having similar pIs would not make it possible to be differentiated on the basis of electrophoretic mobilities. The precision of the molecular mass determination was better than 1%. Assignment of particular glycan structures to isoforms was aided by the known canonical structures of the main isoforms of alpha- and beta-ATIII. Successful characterization of protein glycoforms has been reported using MS-friendly conditions on capillaries with dynamic coatings that generate low-EOF [70]. An acidic BGE containing an organic additive was used (1 M acetic acid/20% methanol at pH 2.4). Under these conditions, partial separation of glycoforms was attained for several glycoproteins including AGP and BSF [71]. By using ionene-coated capillaries and an acetate buffer at pH 4.8, baseline resolution of the three main isoforms of rhEPO and a comparison of rhEPO and uEPO were obtained by CE-ESI-MS [62]. The development of CZE-ESI-MS methods using an orthogonal accelerated time-of-flight (TOF) mass spectrometer has allowed distinguishing glycosylation differences between a reference preparation of rhEPO and a commercially available product sold for research purposes [42]. Membrane concentrators were used to obtain sample concentrations at the required level followed by separation on PB-coated capillaries (the amine coating effectively reverses the EOF). The concentration step provided the additional advantage of removing small MW excipients (e.g., salts, polysorbate), which are known to interfere in the ionization step. Despite the lack of complete resolution in the CE separation step, the MS dimension provided additional resolution and enabled the identification of intact glycoforms varying in the number of HexHexNAc residues, sialic acids, and even the identification of fucosylated, acetylated, and oxidized glycoforms. A total of about 135 isoforms were distinguished in the rhEPO BRP of the EP. A possible composition of the recombinant protein was speculated taking into account the information provided by the CE-MS system and that obtained by other methods. The same method allowed the characterization of the two pharmaceutical products, epoetin-α and epoetin-β [96]. Ultimately, the two products could be distinguished on the basis of the presence of two additional basic (i.e., glycoforms containing less sialic acids) glycoforms in epoetin-β.

Glycoprotein Analysis by Capillary Electrophoresis

RNase B glycoforms

A 0.05

Absorbance at 200 nm

661

Man5

0.04 Man6

0.03

Man7

0.02

Man8 Man9

0.01

RNase A

0.00 10

(a) Rel. int. (%)

B

15

25

20

100

Man

Man5

75 Asn34

Man6

50

GlcNAC

Man7 Man8 Man9

25

RNase A

0 25

30

35

40

45

Time (min)

Rel. int. (%)

(b)

100

9+

Mr:14901

75 8+ 10+

50 25

11+

0

Rel. int. (%)

(c)

100

9+

Mr:15063

75 50 10+ 25

8+

11+

0

Rel. int. (%)

(d)

100

9+

Mr:13680

75 8+ 50 10+ 25 0 1400

1600

1800

2000

m/z

FIGURE 22.16 CZE of RNase B. (A) Fused-silica capillary 80 cm × 50 µm i.d. coated with a solution of polybrene and ethylenglycol; Buffer containing 2M formic acid; injection of 6 pmol of RNase B; UV detection. (B) Fused-silica capillary 110 cm × 50 µm i.d. coated with a solution of polybrene and ethylenglycol; Buffer containing 2 M formic acid; injection of 6 pmol of RNase B; ESMS detection. Sheath flow 5 µL/min of an aqueous solution of 0.2% formic acid and methanol. (a) Total ion electropherogram for the full mass scan acquisition (m/z 1300–2000). (b) Extracted mass spectra for peaks migrating at 35.2 min, (c) at 35.6 min, and (d) at 40.8 min. The calculated molecular mass is shown on the right corner of each spectrum. (From Kelly, J.F. et al., J. Chromatogr. A, 720, 409, 1996. With permission.)

662

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

A comparison of rhEPO by CZE-ESI-MS using capillaries that provided reversal of the EOF (PB coating) or suppressed EOF (UltraTrol Dynamic PreCoat LN) showed, as indicated in a previous section of this chapter, that increased resolution was obtained with the suppressed EOF capillaries (see Figure 22.3) [30]. The improved separation allowed detection of low-level glycoforms. The observed mass differences were attributed to differences in the content of sialic acids, hexoses, and N-acetylhexosamines. In addition, the high mass resolution of the system allowed the detection of comigrating substances differing by 42 and 16 Da corresponding to acetylated and oxidized variants, respectively. Similarly to the results obtained for rhEPO, better glycoform resolution was obtained for bovine fetuin and bovine alpha 1-acid glycoprotein standards on capillaries with suppressed EOF [71]. Despite a level of sensitivity about 100 times lower than that obtained by CE-UV, inductively coupled plasma mass spectrometry (ICP-MS) coupled to CZE has been shown to separate and distinguish between isoforms containing 3, 4, and 5 sialic acid residues in standard Tf samples [133]. For more complete details on the use of CE-MS for glycoproteins some recent reviews have also appeared [134–138]. 22.4.3.4 Other Detection Modes Conductivity detection has been used in a CITP instrument for the fractionation of glycoforms of rhEPO mentioned previously in this chapter [95]. The preparative CITP instrument is composed of two columns made out of a fluorinated ethylene copolymer, one acting as a separation column and the other one as a trapping column. Both columns are provided with their respective conductivity detectors. That of the first one served, in addition to monitoring the CITP separation, as a proper timing in the switching of the columns. The other one in the second column provided the isotacopherograms used in the control of the fraction collector placed at the end of this second column, and equipped with a collection valve. Although the sensitivity of conductivity detectors was poor for the monitoring of proteins, in this case sensitivity was not an issue since 100 µg of rhEPO were injected in each run.

22.5 PRACTICAL APPLICATIONS 22.5.1 CHARACTERIZATION AND QUALITY ASSESSMENT OF BIOLOGICALS AND BIOPHARMACEUTICALS 22.5.1.1 General Considerations The development of physicochemical methods of analysis based on high-resolution techniques such as CE has had a significant impact on fields of activities dealing with complex biomolecules such as glycoproteins. While undoubtedly variable in magnitude, this impact has been considerable in the development of foods, pharmaceuticals, vaccines, blood and blood products, diagnostics, and for regulatory authorities. In many of these areas, the establishment of more sensitive, precise, and selective methods of analysis has particular relevance since manufacturers are required to provide scientific evidence with regard to major tenets that guide the development of regulatory requirements in most jurisdictions: that products entering the market must be safe and of high quality. More specifically, the assessment of a product through the verification of its identity, the quantitative evaluation of its purity, its stability, or the consistency of its manufacturing is of prime importance. For example, considerations regarding the decision to market a pharmaceutical protein resides not only on the demonstration of its biological activity but also on the demonstration that the manufacturing process can provide a product consistently, at the same level of quality and purity as that used for clinical trials. While the application of these specifications does not affect all areas universally

Glycoprotein Analysis by Capillary Electrophoresis

663

(e.g., purity evaluation may not be a significant factor for a diagnostic product), they nevertheless are now generally well adhered to an industry and applied at various stages of the manufacturing process. Data generated are then incorporated in the preparation of documentation to be submitted to regulatory authorities for evaluation. In certain instances, harmonized guidelines relating to test procedures have been developed. This is the case for biotechnological and biological products where an international initiative, the International Conference on Harmonization (ICH), has enshrined these principles in a series of guidelines of test procedures for the quality evaluation (Q5A–Q5E) and specifications (Q6B) of these products [139]. Traditionally, complex biomolecules such as glycoproteins have been assessed using conventional gel electrophoresis methods. However, these methods are laborious to carry out, cannot be automated, and do not consistently provide quantitative results. In contrast, CE has been demonstrated to provide rapid, automated, and quantitative results. In addition, it is now often used as a complementary technique to the more established HPLC in several fields of activities. This progress has led to the recognition by pharmacopeial authorities in Japan, the United States, and Europe of CE for the generation of data relating to product identification, assay or tests for related protein impurities, and to its use by industry in various other aspects of the development process including drug substance and drug product characterization, in-process monitoring, and stability determination. It is also one of the recommended physicochemical techniques of the ICH Q6B guideline on specifications for biotechnological/biological products mentioned earlier. Underlying this recognition is the necessity to devise validated methods based on widely accepted analytical parameters. Typically, method validation involves the determination of one or more analytical parameters (e.g., accuracy, precision, linearity range, limit of detection, limit of quantitation, robustness, specificity) to demonstrate that the method performs according to its intended purpose. The degree with which a method requires validation varies according to the procedure being implemented as shown in Table 22.1 derived from the ICH guidelines Q2(R1). For example, while an identification test requires only that specificity be demonstrated, the quantitative evaluation of impurities necessitates that all characteristics be determined with the exception of the limit of detection. The coming of age of CE as a recognized analytical method has been marked by its introduction into pharmacopeial monographs. Given its improved separation and quantification capabilities, a CZE method for rhEPO glycoform separation was found suitable to replace conventional IEF as an identification test and, as previously mentioned, has been incorporated into the EP monograph for rhEPO concentrated solution (i.e., the bulk substance solution before formulation) [35]. Using

TABLE 22.1 ICH Q2(R1) Recommendations for Method Validation of Analytical Test Procedures Type of Analytical Procedure Characteristics Accuracy Precision Repeatability Interm. precision Specificity Detection limit Quantitation limit Linearity Range

Identification

Testing for Impurities

Assay (Content/Potency)

−

Quantitation/Limit +/−

+

− − + − − − −

+/− +/− +/+ −/+ +/− +/− +/−

+ + + − − + +

Source: http://www.ich.org/LOB/media/MEDIA417.pdf, Accessed November 2006.

664

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

conditions adapted from the method of Watson and Yao [52], rhEPO glycoforms are separated on a bare fused silica with a BGE consisting of 0.01 M tricine, 0.01 M sodium chloride, 0.01 M sodium acetate, 7 M urea, and 2.5 mM putrescine at pH 5.55. The method system suitability specifies several required criteria including the separation of the BRP into a number of well-resolved peaks [34] as shown in Figure 22.17. For this particular preparation, the rhEPO BRP batch 1, eight glycoforms are separated with nearly baseline resolution and the order of migration follows the increasing total number of terminal sialic acid residues present on the carbohydrate chains. In other words, glycoforms with the lowest total number of sialic acids migrate first and those with the highest number migrate last. It should be noted here that the BRP was a mixture prepared from known quantities of two commercialized rhEPO forms, rhEPO-α and rhEPO-β. Characterization of these two products had shown that they had different glycoform profiles and distribution, with rhEPO-β containing higher amounts of the more basic glycoforms (i.e., those with lower total amounts of sialic acid residues). The use of a mixture of products for the BRP allowed the preparation of a unified monograph appropriate to analyze both forms of rhEPO that have essentially identical biological activities. Specifications on the distribution of glycoforms as a percentage of the total content in the test solution are provided and constitute an integral part of the identity test (Table 22.2). (At present, a new standard, rhEPO BRP batch 2, has been established with a slightly different electropherogram to that shown in Figure 22.17). Somatropin, the recombinant version of the 22 kDa form of human growth hormone, is another important biopharmaceutical for which a CZE method has recently been adopted for inclusion into the EP somatropin monographs. Although not a glycoprotein, it is mentioned here to exemplify the increased use of CE for the specifications of proteins. In this case, the test is for the quantitative assessment of charged variants [140]. It is expected that CE will become a regularly used technique in pharmacopeial monographs of biopharmaceuticals in view of its usefulness for the characterization of these structurally complex molecules. In addition, the patent expiration date for several biopharmaceuticals has either passed or is fast approaching and this will undoubtedly lead to several manufacturers wishing to enter this growing market to produce similar products.

5 6

4 7

3

1

0

20 1. isoform 1 2. isoform 2

2

40 3. isoform 3 4. isoform 4

8

60 5. isoform 5 6. isoform 6

min

7. isoform 7 8. isoform 8

FIGURE 22.17 Electropherogram of the European Pharmacopoeia erythropoietin biological reference preparation batch 1. (From Bristow, A. and Charton, E., Pharmaeuropa, 11, 290, 1999. With permission.)

Glycoprotein Analysis by Capillary Electrophoresis

665

TABLE 22.2 Ranges of Glycoforms Content Specified in the EP Monograph for Erythropoietin Isoform

Content (Percent)a

1 2 3 4 5 6 7 8

0–15 0–15 5–20 10–35 15–40 10–35 0–20 0–15

a Ranges for the current EP standard, rhEPO BRP batch

2 have been modified (see Behr-Gross, M.-E., Daas, A. and Bristow, A., Pharmeuropa Bio, 2004, 1, 23, 2004). Source: Erythropoietin concentrated solution, 01/2005:1316, European Pharmacopoeia 5th edition, published by EDQM, June 2004.

From a strategic perspective, product characterization is paramount to the development of identity, purity, or assay procedures since it involves the determination of many of the structural and physicochemical properties of the product under study. Consequently, product characterization is, by necessity, first in the series of test procedures to be developed. For glycoproteins, in addition to the more basic structural properties inherent to proteins such as amino acid sequence and higher-order structural elements (i.e., secondary, tertiary, and quaternary structures), in-depth characterization of carbohydrate-mediated properties are required. In the following sections, an overview of the applications of CE with respect to the guiding principles and analytical parameters described earlier will be presented. For more in-depth discussions on specific aspects or individual glycoproteins, excellent reviews cited in the Section 22.2 can be read.

22.5.1.2 Product Characterization and Identity Testing As previously mentioned, glycoproteins exist as mixtures of closely related species that differ in their glycosylation patterns. These differences are often the result of both compositional and sequence variations of the glycan chains. Moreover, the biological activity of glycoproteins is frequently linked to the presence of these carbohydrates and, consequently, the characterization of glycoprotein microheterogeneity represents one of the more challenging tasks for the characterization and identity testing. In both cases, high specificity is required since it is expected that the method will differentiate between the active ingredient, variants, impurities, and contaminants. Identity tests typically involve the determination of one or more intrinsic properties of a molecule, whether physicochemical, biological, and/or immunological and are often carried out by comparison with a reference standard. Apart from performing a simple identity test through coanalysis of the substrate and the reference standard, qualitative and quantitative information derived from CE experiments with respect to specific structural characteristics such as carbohydrate-mediated glycoform profiles, pI, or MW determination can be useful for product characterization and identity testing.

666

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

22.5.1.2.1 Recombinant Sialoglycoproteins Several important biopharmaceuticals that received market authorization around the world in the past decades are recombinant glycoproteins that include such successful products as EPOs (rhEPO-α and -β), darbepoetin-α (also called NESP), rhtPA, interferons (rhIFN-β1, -γ), and colony-stimulating factors (rhGCSF, recombinant human granulocyte macrophage colony-stimulating factor, rhGMCSF). They are used for the management of diseases not otherwise well treated with small molecule pharmaceuticals. Unequivocally, it is advances in molecular biology and recombinant DNA technology that are largely responsible for the availability of these molecules in large quantities, thus allowing biopharmaceutical companies to introduce them in the markets. Despite these advances, protein glycosylation remains a crucial factor that requires a thorough assessment especially when considering that the glycosylation of recombinant proteins is inherently heterogeneous and rarely identical to that found in humans, a situation that has the potential to result in problems linked to immunogenicity. The ubiquitous occurrence of sialic acid residues on glycan chains of recombinant glycoproteins has provided an ideal tool for analysis by CE by taking advantage of the presence of the ionizable carboxylic acid group. As indicated in previous sections, both CZE and CIEF have been shown to be particularly sensitive to differences in net charge and, as such, have been widely applicable to characterize sialic acid-mediated glycoprotein profiles. However, as it will become evident in the remainder of this section, the biggest limitation to the application of CE methods is undoubtedly the complex nature of sialoglycoproteins, especially when more than one glycosylation site is present in the molecule. In such cases, the number of possible combinations and permutations of individual carbohydrate residues make it difficult to even anticipate separating all glycoforms. Studies of relatively simple sialoglycoproteins, that is, containing a single glycosylation site, were among the first examples of the use of CE for characterization of intact glycoproteins. RhGCSF stimulates progenitor cell proliferation, differentiation, and activation of neutrophils and features a single glycosylation site at Thr-133. It was separated into two glycoforms under basic (pH 7–9) [54] or acidic (pH 2.5) conditions [141]. In the latter case, rhGCSF was analyzed in a preparation containing HSA and the addition of HPMC prevented adsorption to the capillary walls and facilitated detection at low pH. In both cases, the order of migration occurred in a predictable fashion according to increasing number of sialic acid residues. Both methods could distinguish between the nonglycosylated and glycosylated products. A highly efficient CZE method capable of resolving in excess of 30 glycoforms of a simple sialoglycoprotein, a recombinant 24 kDa glycoprotein containing a single glycosylation site, has been reported by Berkowitz et al. [142]. Using a bare fused-silica capillary with triethanolamine/phosphoric acid, pH 2.5 as BGE, the 30 glycoforms were separated on a 27 cm capillary in less than 15 min (Figure 22.18). Lengthening the capillary to 77 cm resolved over 60 peaks in less than 30 min. Partial validation of the method was carried out by examining day-to-day reproducibility and sample matrix effect. For the latter, it became apparent that increased resolution was obtained for samples dissolved in low ionic buffer and low pH. While the effect of low sample ionic strength compared with the BGE is well known and is due to stacking effects, the low sample pH effect was explained in terms of the initial movement of the protein, which if injected from a sample at high pH was toward the anode and, consequently, against the stacking effect since the protein has a relatively high pI (6.5 < pI < 8.0). Using a kinetic study of the enzymatic removal of terminal sialic acids in conjunction with fractionation of the glycoprotein by anion exchange, they were able to identify several glycoforms according to their antennarity and sialic acid content as shown by the shaded areas in Figure 22.18. Although not all of the peaks were identified, this approach allowed to demonstrate that parts of the observed microheterogeneity were due to the presence of large amounts of deamidated variants (i.e., >30%). In addition to the qualitative assessment of the product, the results when taken collectively allowed to derive quantitative expressions for determination of the content in deamidated forms, nonglycosylated forms and monoglycosylated forms, biantennary forms, and

Glycoprotein Analysis by Capillary Electrophoresis

BiNA2 (Area #3)

(a)

(b)

0.14

0.14

0.12

0.12

0.12

0.10

0.10

0.10

0.10

0.08

0.08

0.08

0.08

0.06

0.06

0.06

0.06

0.04

0.04

0.02

BiNA1 (Area #2)

0.04

BiNA0 (Area #1)

0.12

A.U.

0.14

0.14

A.U.

667

0.04 Non-and monosialylated Glycans (Area #4)

0.00 7.5

10.0 Min

0.02

0.02

0.00

0.00

0.02 0.00 7.5

10.0 Min

FIGURE 22.18 CZE separation of 24 kDa sialoglycoproteins with peak region assignments (shaded areas). In (a): BiNA0, biantennary oligosaccharides with no terminal sialic acid residues; BiNA1, biantennary oligosaccharides with one sialic acid terminal residue; and BiNA2, biantennary oligosaccharides with two terminal sialic acid residues. In (b): nonglycosylated protein and monosialylated glycoforms (area #4). Separation conditions: fused-silica capillary, 50 µm × 27 cm capillary, 100 mM triethanolamine–phosphoric acid, pH 2.5, 10 kV, 25 ◦ C and UV detection at 214 nm. (From Berkowitz, S.A. et al., J. Chromatogr. A, 1079, 254, 2005. With permission.)

finally, sialylation level of biantennary forms. The method was also adequate to monitor the effects of changes in culture conditions such as host strains or growth conditions. The analysis of complex sialoglycoproteins, that is, containing multiple glycosylation sites has been the subject of many reports. One such example is rhtPA, an important biopharmaceutical used in the treatment of myocardial infarction, which contains three N-glycosylated sites. Among several articles published on CE separation conditions, Thorne et al. [77] reported the separation of rhtPA into no less than eight glycoforms by CIEF. The method showed high reproducibility and precision for migration times and good total recovery of rhtPA from the capillary was demonstrated. CZE conditions using amines as additives to the BGE were also used to separate numerous major and minor glycoforms. In addition, they demonstrated the use of CE-SDS as a potential purity test for rhtPA Type I and Type II, which differ in the level of glycosylation site-occupancy. RhEPO is another complex sialoglycoprotein, with three N-linked and one O-linked glycosylation sites, that has been extensively characterized by CZE and cIEF as described in previous sections. The comparison of rhEPO and uEPO has been reported on uncoated capillaries [40] and on capillaries dynamically coated with ionene [62]. Although increases in resolution and sensitivity are desirable, and the samples of uEPO corresponded to purified standards, these works showed the potential of discriminating between exogenous and endogenous EPO. CE was also used to fractionate rhEPO glycoforms as briefly discussed previously. Madajova et al. [95] developed a preparative CITP method, operating in a discontinuous mode, which enabled the separation of 100 µg of rhEPO into six fractions in 30 min. The number of collected fractions corresponded approximately to the number of glycoforms detected in the original product by CZE. Subsequent analysis of the collected fractions by CZE showed that several glycoforms had been substantially enriched by 10- to 100-fold in some cases. Such an approach may provide a desirable way to access specific glycoform levels in relation to their biological activity or for the evaluation of low glycoprotein levels in biological matrices.

668

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Recombinant human deoxyribonuclease (rhDNAse) is a complex sialoglycoprotein whose heterogeneity is due in part to the presence of phosphorylated mannose residues. CE studies of this acidic (3.5 < pI < 4.5) phosphosialoglycoprotein that contains two glycosylation sites have been reported [143,144]. CZE methods were developed under both acidic (pH 4.8) and basic (pH 8.0) conditions. The addition of calcium ions significantly improved resolution at both pH as a consequence of interactions with calcium-binding sites that stabilize the protein and make it more resistant to proteases. The pH-dependent calcium binding provided a means to distinguish between acidic and neutral glycoforms that were identified by a two-dimensional (2D) investigation of neuraminidase and alkaline phosphatase–protein digestions. The acidic pH resolved acidic charge heterogeneity and the basic pH discriminated neutral heterogeneity. 22.5.1.2.2 Naturally Occurring Sialoglycoproteins The need for characterization and identity determination of naturally occurring glycoproteins has been one of the driving forces behind the development of CE methods. Naturally occurring glycoproteins are highly heterogeneous and sometimes present a formidable challenge. Only a few examples of approaches to their characterization will be discussed in this section. ATIII is a glycoprotein that plays a key role in the inhibition of blood coagulation. Despite the availability of a recombinant version, plasma-derived ATIII has continued to be the subject of recent characterization studies by CE. Two isoform groups isolated from plasma-derived ATIII that differ in the number of glycosylation sites, ATIII-α (glycosylated at four asparagines) and ATIII-β (glycosylated at three asparagines), were studied by a combination of techniques that included gel IEF, 2D-gel electrophoresis, CIEF, and CZE [145]. The two isoform groups were separated by CZE as shown in Figure 22.19. With a lower glycosylation site-occupancy and, consequently less sialic acid residues, the major component of ATIII-β (peak A in Figure 22.19a and d) predictably migrated before the ATIII-α components. The quantitative analysis indicated a content of about 70% ATIII-α main isoform and about 6.6% of ATIII-β in good agreement with published data. The pI values of ATIII determined by CIEF using an internal calibration were in fair agreement with values of the main isoforms measured by 2D-gel electrophoresis. The CZE method provided a definite advantage in terms of quantification and accuracy. In another report, the same group further characterized the separated isoforms by online coupling with MS [132]. The commercially available glycoprotein, OVA, has been extensively used as a model protein for approaches to glycoform separations by CE since the very early years of the technique. Recent studies on OVA have focused on novel separation conditions that included the effect of new additives to the BGE and new dynamic coatings [87,146,147] as well as conditions suitable for online coupling to MS [148]. As indicated in a previous section, an innovative separation approach based on the freesolution (i.e., without immobilization) selective affinity of OVA glycoforms to a common protein, LCA, has been reported [118]. The separation mechanism was explained in terms of the increased affinity of glycoforms having high-mannose type N-glycans for LCAand, consequently, these species were more strongly retarded than other glycoforms. A similar type of approach was recently used to separate AGP glycoforms without biantennary glycans from those having one or more biantennary glycans [117]. Once again, the separation mechanism appeared to involve greater affinity of the ligand for high-mannose species. 22.5.1.2.3 Glycoproteins without Sialic Acids CE separation of glycoforms of proteins where no sialic acid residues are present has been reported. RhBMP-2 is a protein with high-mannose glycan chains that has been unexpectedly separated under CZE conditions [31,131]. Glycoforms were separated on the basis of the number of mannoses a given rhBMP-2 molecule possesses (see Figure 22.8). This type of approach may prove useful for the qualitative and quantitative monitoring of fermentation processes for glycoproteins without substantial charge differences.

Glycoprotein Analysis by Capillary Electrophoresis

(a)

214 nm

669

(b)

C

214 nm 15

mAU

C

ATIII

15 10

10

D A

5

B

mAU

E

B

D

E

5

0 24 (c)

26

28

30 32 Time (min)

214 nm

34

26 28 Time (min)

24

(d)

A

20

0

36

30

214 nm C A

15 15 10

mAU

mAU

10

5

5

0

0 21

22

23

24 Time (min)

25

26

27

20

22

24

26

28

Time (min)

FIGURE 22.19 CZE of (a) ATIII, (b) ATIII-α, (c) ATIII-β, and (d) ATIII-(α + β). Separation conditions: PVA-coated capillary 50 µm i.d. × 64.5 cm (56 cm effective length), 1 M acetic acid containing 4 M urea, 0.1 M sodium chloride added to the aqueous sample solution, 120 kV, UV at 214 nm. (From Kremser, L., et al., Electrophoresis, 24, 4282, 2003. With permission.)

Cellobiohydrolase I (CBH-I) is a phosphoglycoprotein that contains a catalytic domain with four N-glycosylation sites and where phosphate groups form phosphodiester links between the anomeric carbon and the C6 of two consecutive Man residues (i.e., Man1-P-6Man). CIEF was used to separate the phospho-isoforms of the catalytic domain of CBH-I into four glycoform species according to the number of phosphate groups, that is, 0, 1, 2, and 3 groups [149]. Since glycosylation of CBH-I is dependent on the strain and growth medium, the method may be useful for the analysis of cellulase glycosylation from different strains, different organisms, mutated organisms, or cellulases expressed in other organisms. 22.5.1.2.4 Monoclonal Antibodies A number of therapeutic mAbs produced by recombinant DNA technology have been marketed in recent years. Most mAbs exhibit some heterogeneity due to PTM and degradation occurring during the manufacturing process and shelf life of the product. As such, they have been the subject of several studies using CE-based methods and reviewed [150]. Charge heterogeneity is a parameter commonly monitored for mAbs. Common sources of charge heterogeneity include sialylation, deamidation, and C-terminal lysine cleavage. Sialylation can vary widely between mAbs, from being insignificant in some cases to being the major source of charge heterogeneity in other cases. It follows that the sialic acid-mediated heterogeneity is commonly determined from the characterization of the charge heterogeneity before and after enzymatic treatment with sialidase.

670

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Both CIEF and CZE-based methods have been used to study mAbs. A combination of cation exchange chromatography and CIEF was used to study human tumor necrosis factor (hTNF) mAb [151]. Of the four peaks separated by cIEF, three corresponded to C-terminal lysine variants and the remaining peak to sialic acid-mediated charge differences. Similarly, CIEF was used to determine the sialic acidmediated heterogeneity of a IgG1 mAb that contained N-linked glycosylation in the Fab region at Asn56, with approximately half of the biantennary glycans being sialylated [152]. Using an approach based on the well-known formation of complexes between borate and diols, the carbohydrate heterogeneity of a mAb was demonstrated by CZE [153]. Three partially resolved peaks were observed after optimization of the conditions to 150 mM borate, pH 9.4, on a fused-silica capillary. The method was further used to monitor batch-to-batch consistency and for stability testing. In addition to providing molecular mass information, CE-SDS has been used for the detection and quantification of low levels of unglycosylated heavy chain (HC) in mAbs. The optimization and validation of generic and quantitative CE-SDS procedures with LIF detection was reported [125,154]. An alkylation step was incorporated to decrease thermally induced fragmentation of nonreduced, labeled mAb samples. The unglycosylated variant was detected under reducing conditions where light chain (LC) and HC are separated (Figure 22.20). A collaborative study report involving multiple laboratories in different organizations has demonstrated the robustness of CE-SDS without labeling and using UV detection for assessing mAbs and the unglycosylated variant [155].

HC 100

LC

90

NGHC

80

RFU

70 Free dye

60

Incomplete reduction

50

With IAM

40 30

With IAA

20 10 No alkylation

0 0

2

4

6

8

10

12

14

Migration time (min)

FIGURE 22.20 CE-SDS separations of reduced labeled mAb samples in the presence of different alkylating agents: NGHC, nonglycosylated heavy chain; LC, light chain; HC, heavy chain; IAA, iodoacetic acid; IAM, iodoacetamide. Separation conditions: uncoated fused-silica capillary 50 µm i.d. × 30 cm (effective length 19.4 cm), both anode and cathode buffers were the Bio-Rad SDS running buffer, samples were injected at a constant electric field of 417 V/cm for 15 s and electrophoresed at 625 V/cm (21.2 µA) at 20◦ C, detection was performed with LIF using a 3.5-mW argon ion laser, 488 nm excitation, 560 ± 20 nm emission. (From Salas-Solano, O. et al., Anal. Chem., 78, 6583, 2006. With permission.)

Glycoprotein Analysis by Capillary Electrophoresis

671

22.5.1.3 In-Process Monitoring and Product Consistency Studies of recombinant DNA technology processes have clearly demonstrated that PTM such as glycosylation are not constant throughout culture and that many factors must be examined to obtain and maintain the required glycosylation. They include choice of host cell, genetic engineering of glycan processing, or control of bioprocess parameters such as culture environment, method of cell culture, and culture time [156]. Furthermore, even after a process has been well established, changes may be required from time to time, a situation that may also have an impact on PTM. Examination of pre- and postchange products is then required to ascertain the absence of unwanted changes to the product quality [157]. It follows that monitoring of glycosylation at all stages of a bioprocess has become a necessity to ensure product quality and batch-to-batch product consistency. Practical considerations to the application of an analytical technique for in-process monitoring and product consistency assessment range from the necessity of the method to have high accuracy and sensitivity, to provide high reproducibility and be relatively simple in its application, to be robust enough to sustain minor changes in sample matrix, to be specific and quantitative, and to permit fast development and analysis time. CE has become widely applicable at various stages of the manufacturing process due to its already mentioned advantages regarding the flexibility provided by the various separation modes that can be applied, the low sample volume required, and its capability for automation and fast analysis. 22.5.1.3.1 In-Process Monitoring RhIFN-γ is, as mentioned in a previous section of this chapter, a sialoglycoprotein with two N-linked glycosylation sites at Asn25 and Asn97, which has been shown to be heterogeneous in the number of species resulting from variable glycosylation site-occupancy (nonglycosylated: 0N; monoglycosylated at Asn25: 1N; and diglycosylated at Asn25 and Asn97: 2N) as well as from variability in the number of terminal sialic acids [114,134,158]. In combination with CIEF, a method based on MEKC conditions discussed previously (see Section 4.2.2.3 and Figure 22.13) was used to monitor rhIFN-γ glycosylation during perfused fluidized-bed production [159]. Glycosylation site-occupancy was monitored by MEKC. A plot of the relative amounts of each of 0N, 1N, and 2N provided evidence of constant levels of glycosylation site-occupancy over the course of the process. In addition, the sialic acid mediated isoform content was monitored by CIEF using a commercial kit over a pI range of 3–10. No less than 11 major variants were detected between pI values of 3.4 and 6.4 (Figure 22.21b). After enzymatic desialylation of this mixture, a decrease in the number of variants and, as expected, a major shift to higher pI values was obtained (Figure 22.21a), indicating that most of the observed heterogeneity resulted from sialic acid variability. The remaining heterogeneity was ascribed to C-terminal truncated variants. The quantitative analysis of sialylated variants indicated a sharp decrease in the mean pI after 200 h of perfusion culture (Figure 22.21c). This suggested that sialylation is increased during the perfusion process to reach a constant level afterward. On the other hand, examination by MEKC of glycosylation site-occupancy for a stirred-tank batch culture showed that the process did not provide constant levels, but gradually declined throughout the fermentation process. They also demonstrated that this decline was at the expense of the 2N variant and that there were generally higher proportions of the doubly glycosylated (2N) rhIFN-γ glycoform and lower proportions of the nonglycosylated (0N) variant in the perfused fluidized-bed process compared with stirred-tank culture, that is, the rhIFN-γ protein was more heavily glycosylated during perfusion culture. Similar MEKC conditions were used to monitor rhEPO glycosylation during continuous culture of Chinese hamster ovary (CHO) cells in a fluidized-bed bioreactor [160]. While individual glycoforms were not separated under these conditions, clear differences in both peak shapes and migration times were observed when compared with unglycosylated EPO, suggesting that adequate glycosylation was obtained. Analysis of the product by 2D-electrophoresis provided further evidence that the required isoforms had been obtained.

672

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Absorbance at 280 nm

0.018

(a)

9,6

(b)

2,9 4,2

9,5

0.012

8,2 9,3

0.006

3,8 3,3

5,7 6,4

4,5

3,9 3,73,5 3,4

0.000 10

15

20 25 30 Migration time (min)

35

5.2

40

(c)

Weighted mean pI

5.1 5.0 4.9 4.8 4.7 4.6 100

200

300 Culture time (h)

400

500

FIGURE 22.21 Monitoring of rh IFN-γ microheterogeneity by capillary isoelectric focusing. (a) rhIFN-γ desialylated with neuraminidase from A. ureafaciens. (b) Typical electropherogram of immunoaffinity-purified rhIFN-γ from perfusion culture of CHO cells. (c) The weighted mean pI of rhIFN-γ variants secreted by CHO cells during perfusion culture. (From Goldman, M. H. et al., Biotech. Bioeng. 60, 596, 1998. With permission.)

The influence of changing conditions for the fermentation process of rhATIII using a combination of liquid chromatography and CZE has been reported [161]. While the CZE method only partially resolved glycoforms, it provided quantification within 2 min. The use of a 2D–HPLC/CZE design provided additional sensitivity and resolution and quantitative results in 5 min. Although not developed specifically for glycoproteins, a rapid CE-SDS method for the in-process monitoring of fermentation, hydridoma cell cultivation, and purification has been reported [162]. It was carried out in a total analysis time of less than 5 min. 22.5.1.3.2 Product Consistency The ability to detect differences of product quality is a key component of the assessment of any manufacturing process. As seen in the preceding section, biological processes are complex and many factors influence the quality of the resulting product. In addition, manufacturing processes are rarely static and more often than not involve continual refinement to improve yields, to obtain higher quality products, or to implement new technologies. In these cases, manufacturing changes are implemented and the significance of these changes on the product consistency needs to be assessed. In addition to examples already described in previous sections, a CIEF method was used to compare the isoform profiles of IgG2a obtained under different conditions of osmolality and CO2 partial pressure (pCO2 ) in growth media [163]. Significant increases in the pIs of the major peaks as well as in the number of peaks were observed with increases in osmolality and pCO2 (Figure 22.22). The change in pI was not a consequence of differences in sialic acid content (this particular mAb

Glycoprotein Analysis by Capillary Electrophoresis

673

435 mOsm/kg

375 mOsm/kg

8.6

6.4

320 mOsm/kg pI std

FIGURE 22.22 Representative CIEF separations showing the effect of osmolality on the pI distribution of IgG2a produced in serum-containing medium. The pI standard curve is shown at the bottom of the figure for comparison. A shift in the main pI peaks toward higher pI values with increasing osmolality, as well as an increase in the number of peaks, especially at higher pI values are observed. (From Schmelzer, A. E. and Miller, W.M., Biotechnol. Prog., 18, 346, 2002. With permission.)

was not sialylated) and was due to an increase in the level of galactose incorporation, a situation that indicated that hyperglycosylation was occurring, possibly arising from galactosylation at another Asn site.

22.5.1.4 Product Comparability and Analysis of Finished Products The ability to compare products derived from different manufacturing processes or sources is an important aspect of the regulatory process, for the setting of standards, the evaluation of potential problems related to one specific product or the assessment of differences linked to product characteristics such as biological activity. For example, currently throughout the world, there are biopharmaceuticals for which more than one manufacturer have received marketing authorization and this situation is likely to expand as the expiration dates of patents for several products are fast approaching. In addition, some biopharmaceuticals are being used in situations other than those for which they were marketed, such as their use as performance-enhancing substances by athletes. In many instances, products can be obtained through a number of sources other than conventional pharmacies and may, in some instances, be produced by unauthorized manufacturers. In this context, it has become essential to develop generic methods for the evaluation of multisource products. 22.5.1.4.1 Product Comparability There are several sources of commercially available OVA, and CZE can be used to distinguish their glycoform profiles [146]. Two batches of turkey egg ovalbumin (tOVA) showed qualitatively similar glycoform profiles with some differences in the levels of a number of glycoforms (Figure 22.23a and b). On the other hand, the glycoform profile of chicken egg ovalbumin (cOVA) (Figure 22.23c)

674

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

6

(a)

A200 nm × 1000

5 4 3 2 1 0 6

(b)

A200 nm × 1000

5 4 3 2 1 0 6

(c)

A200 nm × 1000

5 4 3 2 1 0 10

15

20

25

30 Time (min)

35

40

45

10

FIGURE 22.23 Electropherograms of (a) and (b): two batches of turkey ovalbumin and (c): chicken ovalbumin. Separation conditions: fused-silica capillary 72 cm (50 cm effective length), 100 mM H3 BO3 /NaOH and 1.8 mM putrescine, pH 8.6, 20 kV, detection at 200 nm, temperature controlled at 30 ◦ C. (From Che, F.Y. et al., J. Chromatogr. A, 849, 599, 1999. With permission.)

Glycoprotein Analysis by Capillary Electrophoresis

675

(a) Sample 1 Free PSA

1.0

0

3

Relative absorbance (A214)

Relative absorbance (A214)

2.0

2.0

6

9

12 15 18 Time (min)

(c)

21

Sample 3 Free PSA

1.0

0

3

6

9

12

15

Time (min)

18

21

3.0

24

(b) Sample 2 Free PSA

1.0

0

24

3

Relative absorbance (A214)

Relative absorbance (A214)

revealed marked qualitative differences when compared with tOVA with respect to migration times. These differences were a result of both polypeptide and glycan chains structural variations. Caseinoglycomacropeptide (CGMP) is a small glycoprotein (MW of approx. 7000 Da) derived from bovine κ-casein that has been shown to have a variety of biological activities such as inhibition of gastric secretion, depression of platelet aggregation, growth promoting effect on bifidobacteria, inhibition of oral Actinomyces adhesion to red blood cell membranes, inhibition of adhesion of oral Streptococci to saliva-coated hydroxyapatite beads, inhibition of adhesion of Streptococcus sanguis to human buccal epithelial cells, and inhibition of cholera toxin binding to its receptor. A CZE method that had been previously reported for the separation of CGMP glycoforms [164] and for stability monitoring in cosmetic lotions [165] was also used for the assessment of sialylation levels of commercial batches of CGMP [166]. Results indicated that the quantitative assessment using CZE were in good agreement with those obtained by enzymatic release of sialic acids or colorimetric methods. The CZE method was fully validated for the usual analytical criteria (e.g., linearity, precision, accuracy, LOD, and LOQ). This approach provided a fast, reliable, precise, and cost-effective way to compare products. Another example of product comparability was reported for prostate-specific antigen (PSA), a single-chain glycoprotein that is used as a biomarker for prostate-related diseases [167]. PSA has one known PTM, a sialylated biantennary N-linked glycan chain attached to Asn45. It is commercially available from different sources. The glycoform profiles of seven free PSA (fPSA) samples from several manufacturers, of which two were specialized, enzymatically active PSA (EA-PSA) and noncomplexing PSA (NC-PSA), were assessed by CZE. Results indicated that PSA samples could be classified into three distinct groups according to glycoform profiles. Figure 22.24 shows

2.0

6

9

12 15 18 Time (min)

(d)

21

24

Sample 4 Enzymatically active

1.0

0

3

6

9

12 15

18

21

24

Time (min)

FIGURE 22.24 Electropherograms of protein specific antigen (PSA) group I. (a) Sample 1 (f PSA); (b) Sample 2 (f PSA); (c) Sample 3 (f PSA); (d) Sample 4 (EA-PSA). Separation conditions: 50 µm i.d. × 60 cm fused-silica capillary, 20 mM sodium borate (pH 8.0)/5 mM diaminopropane or 20 mM sodium phosphate (pH 7.0 or 8.0), 25 kV, detection at 214 nm. (From Donohue, M.J. et al., Anal. Biochem., 339, 318, 2005. With permission.)

676

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

electropherograms of products from group I that contained 4–7 peaks with one peak dominating and representing 51–70% of the total fPSA in the sample. This study demonstrated that proprietary purification procedures have a major influence on the composition and yield of the number of structural forms observed in a PSA sample. 22.5.1.4.2 Glycoprotein Drug Products Methods capable of analyzing the active ingredient in a formulated (i.e., finished) drug product are useful to measure product integrity and/or its stability as well as for comparison to other products. However, the approach to their analysis entails careful consideration of a number of specific issues to drug products such as the low amounts of active ingredient, the presence of large amounts of excipients, or the potential of modifying the product through manipulation. (Glyco)protein drug products generally contain low amounts of the active ingredient since the therapeutic effect can usually be achieved at low concentrations. For example, it is not unusual to find active ingredient at low microgram quantities in a dosage form. In turn, formulations at low protein content usually require the addition of large amounts of excipients to enhance product stability and prevent nonspecific adsorption to the container. Commonly used excipients vary widely in terms of their chemical nature and include inorganic salts, amino acids, sugars, and surfactants such as polysorbate or other proteins such as HSA. Typically, isotonic salt preparations are produced as most of these products are injectables and, consequently, high salt concentrations are present. Furthermore, many of these excipients may be present simultaneously, leading to complex mixtures. They have also been shown in some cases to interfere with traditional assay methodologies such as UV or high-pressure liquid chromatography (HPLC). The choice of CE as a separation technique to be applied to the analysis of protein drug products resides in its ability to provide sufficient selectivity to separate the active ingredient from interfering species. Studies aiming the direct analysis of finished products with no or minimal prefractionation or manipulations in order to minimize potential loss of product or the generation of product artifacts have been reported. As mentioned previously, a CZE method using an amine-coated capillary to separate rhEPO glycoforms in drug products containing large amounts of HSA and inorganic salts has been reported [41]. The addition of nickel ions to the high ionic strength BGE (200 mM phosphate, pH 4.0) selectively altered the mobility of HSA and allowed the separation of the otherwise comigrating proteins (Figure 22.6). Dosage forms containing HSA/rhEPO at a ratio of 250:1 w/w could be analyzed under these conditions. The method was validated and provided satisfactory results for active ingredient assay, quantification of glycoforms, and comparison of products from different manufacturers. The usefulness of the method was further extended to assess manufacturing changes and to compare rhEPO products formulated with different excipients as shown in Figure 22.25 for different rhEPO-α and rhEPO-β products formulated with HSA or polysorbate [168]. As described in a previous section, a similar comparison of finished rhEPO products was carried out following the immunochromatographic removal of HSA from formulations (see Section 4.1) and enabled the comparison of glycoform profiles of products from different manufacturers and from the BRP of the EP formulated with excipients of low-MW and with HSA [43]. The biophysical properties of two commercially available rhEPO-α products, Epogen and Eprex, both of which are produced under similar conditions by different manufacturers, were studied [169]. Among many techniques used in this study, the analysis by CZE showed that the products had similar glycoform profiles in terms of isoform type and distribution, an indication that the products had similar sialic acid contents. CZE was also used for the analysis of two preparations of rhGCSF, Granocyte, a glycosylated product formulated with large amounts of HSA, and Gran300, a nonglycosylated product [170]. The two products were compared with a bulk rhGCSF preparation that had been obtained from Escherichia coli (nonglycosylated). Analysis with 50 mM Tricine, 20 mM NaCl, 2.5 mM DAB, pH 8.0, as BGE showed the separation of two glycoforms in Granocyte that corresponded to monosialylated and disialylated glycoforms. As expected, the nonglycosylated product, Gran300, showed

Glycoprotein Analysis by Capillary Electrophoresis

677

0.004

0.004

0.003

0.003

0.002

0.002

0.001

0.001

0.000

0.000

−0.001

−0.001

−0.002

−0.002

−0.003

−0.003 8

10

12

14

16

18

20

22

24

26

28

A.U.

A.U.

0.005

30

Min

FIGURE 22.25 Typical electropherograms of rhEPO drug products (from top to bottom): rhEPO-α formulated with HSA (product I), rhEPO-α formulated with polysorbate 80 (product II), rhEPO-α formulated with HSA (product III), and rhEPO-β formulated with polysorbate 20 (product IV). Separation conditions: Beckman eCAP amine capillary, 200 mM sodium phosphate/1 mM nickel chloride pH 4.0, –15 kV (75 µA), UV detection at 200 nm. (From Girard, M. et al., Presented at the 7th International symposium on capillary electrophoresis in the biotechnology and pharmaceutical industries, Montreal, August 2005.)

a single peak with a shorter migration time than that of the two glycosylated counterparts and that corresponded to that of the bulk rhGCSF from E. coli. 22.5.1.5 Determination of Biological Activity/Potency The determination of the biological activity of biopharmaceuticals by physicochemical methods has long been a desirable goal of manufacturers, regulatory authorities, and policy makers alike, driven in part by the ability to replace costly and often imprecise biological assay testing using animals. There are currently a few examples of biopharmaceuticals for which this has occurred. Recombinant insulin and recombinant somatropin (human growth hormone) are two such products, both of which can be described as relatively simple, well-characterized proteins. For both of these two nonglycosylated proteins, physicochemical tests have replaced animal-based bioassays for batch release of bulk products (for somatropin see [140]). In-depth physicochemical characterizations as well as the establishment of a direct correlation with the in vivo bioassay were carried out. For more complex products such as glycoproteins, difficulties in obtaining in-depth characterization of the glycosylation-mediated microheterogeneity have limited the development of this type of approach. However, in a few cases, reports have shown that there exists a correlation between glycoform profiles and biological activity. This is the case for recombinant follicle stimulating hormone (rhFSH) where the CZE-derived glycoform profile has been shown to correlate with the in vivo bioassay in a study directed at predicting the biological potency of several preparations [171]. Using CZE conditions consisting of a fused-silica capillary and 100 mM borate, 5 mM DAP at pH 8.9 [172], where rhFSH isoforms migrated as a broad peak between two internal standards, sucrose and the dipeptide Lys-Asp, four highly purified rhFSH preparations differing in their isoform composition and biological potencies were analyzed (Figure 22.26a). The inclusion of internal standards enabled the accurate determination of the median migration time (tm) of the rhFSH peak in each of the four

678

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a) Neutral marker

A 200 nm (AU)

FSH A

FSH C

Iys-asp FSH

Sucrose 20

0

40

60

0

A 200 nm (AU)

FSH B

0

20

40

60

20

40

60

FSH D

20 60 40 Migration time (min)

0

Migration time (min) Anode

(b)

3.50 4.15 4.55 5.20

pl markers 2.5–6.5

FSH D

FSH C

pl markers 3–10

FSH B

FSH A

pl markers 2.5–6.5

Cathode

FIGURE 22.26 (a) Electropherograms of rhFSH preparations A–D with different isoform compositions and biological potencies. Separation conditions: fused-silica capillary i.d.: 20 µm × 77 cm (effective length 70.6 cm), 100 mM borate, 5 mM diaminopropane (DAP) at pH 8.9, 20 kV (approx. 3.5 µA), 28 ◦ C, detection at 200 nm. (b) Isoelectric focusing electropherograms of preparations A–D. (From Storring, P.L. et al., Biologicals, 30, 217, 2002. With permission.)

preparations; tm correlated directly with the biological potency on the basis that an increase in the content of highly sialylated (the more acidic) isoforms led to a corresponding increase in biological activity. The same four preparations were also analyzed by gel IEF (Figure 22.26b) and the predicted biological potencies were derived and compared with those obtained by CZE. In both cases it was found that the methods were sufficiently accurate, precise, and robust to predict the bioactivity of batches of rhFSH when used in conjunction with a standard preparation.

22.5.2 CE OF ISOFORMS OF INTACT GLYCOPROTEINS IN THE CLINICAL FIELD As mentioned in the introduction to this chapter, changes in the glycosylation of proteins have been widely related to pathophysiological changes in an individual [1–4]. For instance, a common

Glycoprotein Analysis by Capillary Electrophoresis

679

phenotypic alteration in malignant cells is the transformation of their glycosylation [173]. CE is an attractive technique for the study of glycoproteins with a clinical interest, due to its well-known quantitative results and speed of analysis [21]. However, not many works dealing with this subject have been published. There are many challenges in the application of CE to glycoproteins of clinical interest. One of the problems is that, as for many other important molecules, these glycoproteins often exist at low concentrations in biological matrices where other proteins are usually present in large amounts and interfere in the analysis. Thus, a sample preparation step, sometimes including concentration, is required. Moreover, increasing the detection sensitivity of CE for the analysis of glycoproteins is one of the challenges reviewed in this chapter. Another pitfall of the technique is that it is not easy to develop a single method for the analysis of most of the proteins as it is the case for other analytical techniques such as SDS–PAGE. On the other hand, CE presents the advantage of its high resolving power that enables the study of slight modifications in protein composition such as for glycoproteins. In addition, its small sample volume requirements make it appropriate for the analysis of clinical samples. This section is devoted to a discussion of CE reports on the separation of isoforms of glycoproteins with a clinical interest. 22.5.2.1 Transferrin The isoforms of Tf, one of the major proteins in serum, have been widely analyzed by CE as alcohol abuse markers. Tf is a glycoprotein synthesized mainly in the liver, which consists, in a single polypeptide chain, of 679 amino acid residues and two N-linked complex-type oligosaccharide chains. It has an important role in iron metabolism and it also acts as a growth factor [174,175]. Human Tf presents several heterogeneities due to its iron content, its genetic polymorphism, and its carbohydrate moieties. With respect to iron content, a Tf molecule can be apotransferrin, monoferric transferrin, or diferric transferrin [174,175]. The influence of this heterogeneity on CE separations of glycoforms can be easily overcome by iron saturation of the sample. On the other hand, more than 38 genetic Tf variants, attributable to substitutions at one or more amino acids, have been described, although only four of them show a prevalence of ≥1% [176]. The common type of Tf is called C-type and most Caucasian individuals express this allele [177]. With respect to carbohydrate heterogeneity, each carbohydrate chain can be bi-, tri-, or tetra-antennary and each antenna possesses a terminal sialic acid residue. Therefore, nine sialoforms (from zero to eight sialic acid residues per molecule) may be present in serum, resulting in pI variations of 0.1 pH unit/residue. The most abundant Tfform in normal sera is tetrasialo-Tf and it has a pI of 5.4. The other forms have higher and lower pI values [178]. Variations in the heterogeneity of Tf are induced by several pathological and physiological conditions, such as rheumatoid arthritis (RA), congenital disorders of glycosylation (CDG), or pregnancy [175,179], even though variations in the Tf pattern due to alcohol abuse has attracted the highest interest [176,178,180]. The first to report abnormal Tf heterogeneity in cerebrospinal fluid (CSF) associated with alcoholic cerebellar degeneration were Stibler and Kjellin [181] and further studies determined the Tf abnormality in serum of alcohol abusers [182,183]. The variation in Tf heterogeneity in sera from alcohol abusers corresponds to increased amounts of asialo-, monosialo-, and disialotransferrrin (generically called carbohydrate-deficient transferrin or CDT). At first, the increased CDT was only attributed to changes in the number of sialic acid residues. However, it was later learnt that the difference was more complex and included changes in some of the neutral carbohydrates [178]. CDT has become the most specific marker for chronic alcohol abuse [66,177,178,184]. The first studies dealing with the quantitative determination of CDT in sera from alcohol abusers were performed by conventional IEF with immunological detection. This method can easily detect genetic variants and has the power to resolve individual isoforms. However, it is laborious and time consuming [178,184]. Since then, a number of CDT methods of analysis, such as anion-exchange chromatography, chromatofocusing, HPLC, and CZE have been developed [176,184].

680

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

At the present time, there are several commercial kits for the quantification of CDT based on the fractionation of CDT and non-CDT serum variants by anion-exchange chromatography with immunochemical detection [176]. The main characteristic of almost all of these micro columns is that they do not separate CDT sialoforms and they are quantified as a whole. Therefore, the CDT forms may contain an undetermined amount of trisialo-Tf, which is under debate as to whether it should be considered a CDT form, and that introduces an error in the determination [185,186]. Another drawback of the anion-exchange chromatographic methods is that they fail to detect genetic variants, which may give rise to incorrect determination of CDT or even false-positive results when sera from individuals with genetic Tf variants are analyzed [177,186]. Some authors have claimed that the separation of individual Tf sialoforms is a more accurate procedure than CDT quantification as a whole [68,69,185,186]. One of these commercial tests was compared with HPLC and CZE methods to conclude that the data obtained needed to be systematically confirmed by HPLC or CZE [187]. In spite of these drawbacks, these kits are still convenient for routine analysis due to their speed. HPLC and CZE methods have been developed to resolve all of the Tf sialoforms present in serum. An HPLC method based on anion-exchange chromatography with direct detection at 420 nm was developed by Jeppsson et al. [188] to individually separate Tf sialoforms in about 16 min. The serum required to be saturated with iron and lipoproteins to be precipitated. Several authors followed the method with minor changes and improvements [186,189,190]. 22.5.2.1.1 CZE of Carbohydrate-Deficient Transferrin (CDT) In general, the CZE identification of Tf sialoforms in the electropherograms has been performed in three ways: first, using anti-Tf to perform immunosubtraction and then comparing the electropherograms of the Tf sample before and after being immunosubtracted; second, the amounts of the different sialoforms can be monitored by CZE analysis after progressive desialylation of Tf with neuraminidase treatment of the sample; and third, by identification from the migration times of each sialoform. With very few exceptions, Tf is iron-saturated before performing CE to eliminate the iron-binding-mediated heterogeneity. 22.5.2.1.1.1 Sample Preparation Commercially available Tf standards were first separated by CIEF and CZE by Kilar and Hjerten [79,97] and Oda and Landers [55]. Iourin et al. [120] purified Tf from sera of one donor and of two CDG type I syndrome by consecutive precipitation of sera with rivanol and ammonium sulfate for their analysis by CZE. On the other hand, immunopurification of Tf from sera was used as sample preparation before CE analysis [89,191]. While these were pioneering attempts at the use of CZE to study CDT as a disease marker, the laborious and timeconsuming sample preparation steps required were not conducive to their use in routine settings. Sample preparation of Tf from serum became easier after the development of a procedure that involved only iron saturation and dilution of the serum [192]. This sample preparation method has since then been adopted by most authors with only slight modifications [66–68,177,185,193–201]. In another approach, interfering proteins, such as immunoglobulins, can be eliminated from serum with protein A to enhance CDT detection [202]. Another interesting approach to sample preparation is the direct injection of serum into the capillary, carrying out complexation of Tf with iron during the electrophoretic separation [203]. 22.5.2.1.1.2 CZE Separation With respect to CZE performance, many approaches have been published for the separation of Tf sialoforms. Two groups of methods can be distinguished. In the first one, studies were conducted with a commercial buffer system (CEofix-CDT kit; Analisis, Namur, Belgium), as already mentioned in a previous section, widely accepted and used, sometimes with slight modifications to the protocol proposed by the manufacturer [66– 69,177,185,195,197–199,203]. This commercial reagent kit offers the advantage of interlaboratory standardization. In the second group of studies, methods developed in-house were used

Glycoprotein Analysis by Capillary Electrophoresis

681

[55,79,89,97,120,185,191,192,196,200–202]. To the best of our knowledge, all of the reported methods were carried out at basic pH (around 8.5) so that Tf is negatively charged to prevent protein adsorption to the capillary wall. The commercial buffer system is based on the dynamic double coating of an uncoated, bare fused-silica capillary. A first buffer (the so-called initiator) is used to coat the capillary wall with a polycation and a second buffer containing a polyanion adds the second layer. The separation buffer is usually borate-based. Between injections, the capillary is rinsed with NaOH so that the coating layers are eliminated after each analysis. This procedure ensures a constant EOF and increases the negative charges on the capillary wall for preventing protein adsorption and increasing the speed of analysis (usually between 6 and 13 min, depending on the modifications of the method) [185,203]. Although in one of the first studies with this commercial system, disialo- and trisialo-Tf could not be resolved [203], modifications of the initial method, which included among others offline iron-saturation and dilution of serum, increased capillary length, and injection of an SDS solution before the injection of the sample, led to baseline resolution of tri, tetra, penta, and hexasialo-Tf in nonalcoholic donors and asialo, di, tri, tetra, penta, and hexasialo-Tf in alcoholic patients [177]. It appears that the SDS plug before sample injection is performed to keep β-lipoprotein peaks out of the area of the electropherogram of interest. SDS is now incorporated in the FeCl3 solution used for iron saturation of the commercial reagent set, so it is not necessary to add it separately if the complete reagent kit is used [69]. In the second group of CZE separations, many different approaches have been published. The main goal has been to avoid Tf adsorption to the capillary walls. This problem has been overcome by working with covalently coated capillaries, in some cases with hydroxyethylcellulose in the running buffer [79,89,191,202], or silica capillaries dynamically coated with DAB [185,196,200,202], spermine [185], DcBr [55], or diethylentriamine (DETA) [201]. In general, the methods performed in uncoated capillaries with dynamic coatings are slower than the ones performed with the commercial buffers. On the other hand, the methods using covalently coated capillaries are as fast (usually in less than 10 min of analysis time) as the methods developed with the commercial buffers. Lanz et al. [185] compared two different methods in fused-silica capillaries with alkaline borate buffers and different dynamic coatings (DAB and spermine, respectively) and with the doublecoating method based on commercially available buffers. The latter method provided faster and more reproducible results. In addition, Tf isoforms peaks were more intense when analyzed with the latter method. The same group [66] studied the precision of a method based on the double-coating buffer kit over a 20-day period. They found out that the method is highly reproducible both in migration times and peak areas of Tf isoforms. Martello et al. [198] ran an interlaboratory comparison of a method based on the dynamic double coating of the capillary wall performed with commercially available reagents. Results from both laboratories showed high correlation. 22.5.2.1.2 Clinical Applications In general, the study by CZE of CDT as an illness marker is performed by qualitative or quantitative comparison of CDT between two different groups of participants. To the best of our knowledge, CDT reference limits for direct classification of samples are not yet routinely used, although some attempts have been performed. In that sense, not even a common CDT calculation is performed. For example, area of disialo-Tf as a percentage of tetrasialo-Tf [185,192,201,204], area % of disialo-Tf in relation to the sum of all Tf-isoforms [191,202], and CDT isoforms (asialo-, monosialo-, and disialo-Tf) as area % of total Tf isoforms [177,194,203] have been used as CDT measurements. Lanz and Thormann [205] worked on the establishment of normal CDT reference limits using a CZE method based on a dynamic double coating of the capillary with commercially available reagent kits. They found that the reference CDT intervals (in the case of CDT, only the upper reference limit is important) for a group of 54 individuals with no or moderate alcohol consumption was dependent on the complete or incomplete separation of disialo- and trisialo-Tf, on the integration approach

682

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

and on the applied voltage. The upper reference limit obtained by Lanz et al. [197] was used in a subsequent study where more than 600 samples were analyzed with a CZE method established as a routine method. Later on, the reference intervals found by those authors [205] were revised for a modified version of the commercially available reagent kits [67]. The same samples from the 54 individuals were re-analyzed with the new reagents and the CDT reference limits were found to be comparable, although slightly smaller, to the old ones so that the same upper reference limit was valid for the study. The establishment of reliable reference CDT intervals is critical in the sensitivity (number of false negatives) and specificity (number of false positives) of the method. 22.5.2.1.2.1 Alcohol Abuse Alcohol abuse is an important public health problem, and selfreporting of alcoholism or alcohol abuse is not reliable. Furthermore, diagnosis on the basis of clinical symptoms is not easy [197]. Therefore, reliable alcohol markers are required for the diagnosis of alcoholism and CDT is one of the better markers of alcohol abuse. Prasad et al. [191] based their CDT calculation of Tf separated by CZE on the ratio of the area of disialo-Tf (they did not detect asialo-Tf) to the area of total Tf in the sample. On the basis of this index and with a control population of social drinkers, they established a CDT cutoff value. This cutoff value was exceeded by the majority of alcohol abusers studied. This CDT measurement was compared with other alcohol abuse markers (e.g., aspartate amino transferase, alanine amino transferase, and others) and it was found to be the most specific one. In another study, Tagliaro et al. [192] found that by comparing the Tf profile of a control group of 30 participants against a group of 13 alcoholics, the latter showed a significant increase in disialoand trisialo-Tf (expressed as percentages of the tetrasialo-Tf peak). They did not detect asialo-Tf in alcoholic samples. These results are controversial since several studies have claimed that not only trisialo-Tf is not an alcohol abuse marker [68,195] but also that it interferes with the correct CDT quantification [204]. On the other hand, there are also favorable opinions to include trisialo-Tf in the calculation of CDT or, at least, to establish a debate [177,206,207]. Crivellente et al. [193] improved on their published method [192] and tested it with real samples. Although the presence of asialo-Tf was suspected to be present in the sera of alcoholic individuals, no confirmation was performed. An increase in disialo- and trisialo-Tf (expressed as percentage of the tetrasialo-Tf peak) compared with healthy participants happened in alcoholic individuals. Another study using CZE for the analysis of CDT as a marker of alcohol abuse was carried out by Giordano et al. [202]. They did not detect asialo-Tf in sera from alcoholic people, but they detected a significant increase in the disialo-Tf (related to total Tf) in those samples in comparison with control (nonalcoholic subjects) samples. In the same direction, and by establishing the Tf index as % area of disialo-Tf in relation to tetrasialo-Tf, Lanz et al. [185] found that a good classification of sera could be performed on the basis of this index increasing in sera of alcoholic individuals. Interesting CZE studies on CDT isoforms were reported by Legros et al. [68,69]. They found that asialo-Tf was missing in teetotalers and was present in 92% of alcohol abusers, and that disialo-Tf was increased in alcohol abusers. Figure 22.27 shows the comparison of the Tf profiles of an alcoholic and a teetotaler. Under these conditions, C-reactive protein (CRP) comigrated with monosialo-Tf in the alcoholics electropherogram. Samples before and after Tf immunosubtraction are also compared in the figure. They proposed to use the presence of asialo-Tf as alcohol intake marker because it showed the highest sensitivity and specificity when compared with other CDT isoforms or combination of CDT isoforms. The study focused on clearly different groups (teetotalers and alcohol abusers). In their subsequent work [69], they included moderate drinkers in another group. They found that asialo-Tf was able to discriminate between moderate drinkers and alcohol abusers better than other CDT measurements tested. 22.5.2.1.2.2 Congenital Disorders of Glycosylation Congenital disorders of glycosylation (CDG) (previously known as carbohydrate-deficient glycoprotein syndromes) are inherited

Glycoprotein Analysis by Capillary Electrophoresis

683

P4

P5

P2 P0

P3

1/crp

P6

AA

Anti-Tf P5

P2

P3

P6 TT

Anti-Tf

5.5

6.0 Time (min)

FIGURE 22.27 Comparison between CZE Tf electropherograms of an alcohol abuser (AA) and a teetotaler (TT). In both cases, anti-Tf antibody was added after the first electrophoretic run. P0: asialo-Tf. 1/CRP: comigration of monosialo-Tf and CRP; P2: disialo-Tf; P3: trisialo-Tf; P4: tetrasialo-Tf; P5: pentasialo-Tf; P6: hexasialo-Tf. Analytical conditions: fused-silica capillary 57 cm × 50 µm i.d; buffer: reagents from the CEofix CDT kit. Capillary was first rinsed with a solution of polycation dissolved in malic acid, pH 4.8; then rinsed again with a polyanion dissolved in Tris-borate, pH 8.5. Capillary was rinsed with the same buffer for 0.5 min under low pressure. After 3 s low-pressure injection of SDS, the iron-saturated sample was injected by low pressure for 2 s; Voltage: 28 kV; Temperature: 40◦ C; Detection: 214 nm. (From Legros, F.J. et al., Clin. Chem., 48, 2177, 2002. With permission.)

conditions that are usually recognized from glycosylation changes of serum proteins [2]. The hypoglycosylation of different proteins and sometimes of other glycoconjugates leads to several symptoms, such as mental retardation. IEF measurement of Tf is the main diagnostic test used for the detection of CDG because these disorders appear to have an influence on its sialylation (and consequently on its pI) [199,208]. There are several studies that have been reported in which the CZE of Tf has been used to study CDG. An increase in disialo- and asialo-Tf in CDG patients has been reported by several authors [89,120,196,203]. Carchon et al. [199] demonstrated that patients with abnormal IEF results and with confirmed CDG could be identified by CZE. However, they warned about the possibility of finding compounds migrating in the Tf-region and when this is suspected in a sample, CZE analysis with immunosubtraction should be performed.

684

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

22.5.2.1.2.3 Genetic Variants In addition to the common C-type Tf, other variants such as B and D have been reported. As stated earlier, genetic polymorphism is limited to Caucasians. Most individuals present CC phenotypes (homozygous for Tf with the C gene variant) and, rarely, CB phenotypes. West African, African American, and indigenous American populations, for example, present higher frequency of the D allele [177]. Certain genetic variants are not detected by anion-exchange chromatography-based methods and may result in false positives for alcohol abuse [177,178]. In that sense, CZE is a promising tool for the determination of genetic variants due to its high resolving power. Several workers have published Tf CZE patterns of different genetic variants [66,67,177,197,203]. In general, these patterns are characterized by having two major peaks instead of one (corresponding to the tetrasialo-Tf of both genetic variants). Wuyts et al. [177] found that their CZE method was able to separate (in an alcoholic carrying a CD phenotype) D-asialo-, D-disialo-, D-trisialo-, D-tetrasialo-, and D-pentasialo-Tf in addition to Ctetrasialo-, C-pentasialo-, and C-hexasialo-Tf. In spite of this high resolution power, some of the D-Tf peaks comigrated with C-CDT forms. To solve this problem for the CDT calculation for alcohol consumption diagnostic purposes, the authors proposed a CDT calculation adjusted to CD-phenotypes in order to avoid false positives. Figure 22.28 shows the comparison of Tf patterns from a nonalcohol consumer homozygous (CC), an alcoholic homozygous (CC), a nonalcohol consumer heterozygous (CD), and an alcoholic heterozygous (CD). On the other hand, Lanz et al. [67] analyzed a CD-phenotype alcoholic individual. They detected two peaks for asialo-Tf and two peaks for disialo-Tf. Since C-disialo-Tf was not baseline resolved, the CDT calculation was done by taking twice the area of D-disialo-Tf because both peaks had the same height. The CDT value clearly exceeded the upper CDT reference limit established for normal nonalcoholic-CC phenotypes. 22.5.2.1.2.4 Transferrin and Cancer A study has been recently performed in which the Tf CZE profile of a group of cancer patients who consumed alcohol moderately and a group of cancer patients who were alcoholics has been studied [195]. Their profiles were compared with a group of participants without cancer and that were teetotalers, moderate drinkers, and alcohol abusers. While asialo-Tf was present in 95% of alcohol abusers (with and without cancer), trisialo-Tf was higher in cancer patients. The CZE peak corresponding to this trisialo-Tf was not completely eliminated by Tf immunosubtraction. Some tests on the remaining peak suggested that it might be a polysaccharide. 22.5.2.1.2.5 Interferences in CDT Determination Usually, Tf analysis by CZE is performed after the injection of diluted, iron-saturated serum, which is an extremely complex media. Therefore, possible interferences in the determination of Tf should be taken into account. Generally, interferences are found after Tf immunosubtraction of the sample and re-analysis. Sometimes, immunosubtraction of the interference is also performed. Legros et al. [68] detected comigration of CRP with monosialo-Tf, so this peak was not taken into account in the CDT calculations. CRP was also identified and migrated before disialo-Tf in some samples analyzed by Lanz et al. [66,197]. In some patients with hepatic disorders, a broad interference, assumed to be due to high levels of immunoglobulin A, was found under the Tf peaks [197]. Wuyts et al. [203] added antihemoglobin and anti-C3c (degradation products of complement-C) antibodies to the running buffer to avoid interferences from these compounds. As described before, SDS was injected into the capillary before the sample for methods performed with commercial reagent buffers to avoid β-lipoprotein peaks comigrating with Tf [68,69,177,195]. This modification was later adopted by the manufacturer [69]. Moreover, a compound comigrating with trisialo-Tf (probably, a polysaccharide) in cancer patients was described [195]. Recently, the commercial reagents for the double-coating CZE method have been reformulated to avoid undesired interferences, and CRP comigration is no longer interfering, neither are some paraproteins that interfered in some samples [67].

Glycoprotein Analysis by Capillary Electrophoresis

685

4

CC nonalcoholic 5 6

3

4 CC alcoholic

5 2

3

6

0

CC nonalcoholic

D4

D5

C4

C5 C6

D3 D4 CC alcoholic

C4

D5 D0 D2

8.0

D3

C5 C6

8.4 8.8 9.2 Migration time (min)

FIGURE 22.28 Comparison between CZE Tf electropherograms of a healthy nonconsuming (CC nonalcoholic) and an alcohol-consuming (CC alcoholic) carrier of the homozygote CC-Tf, and a nonconsuming (CD nonalcoholic) and an alcohol-consuming (CD alcoholic) carrier of the heterozygous CD-Tf. C0 : asialo-C, C2 : disialo-C, C3 : trisialo-C, C4 : tetrasialo-C, C5 : pentasialo-C, C6 : hexasialo-C-Tf. D2 : disialo-D, D3 : trisialo-D, D4 : tetrasialo-D, D5 : pentasialo-D-Tf. Analytical conditions: fused-silica capillary 67 cm × 50 µm i.d; buffer: reagents from the CEofix CDT kit. Capillary was first rinsed with the initiator solution, then rinsed again with the buffer solution containing a polyanion. A plug of SDS was injected during 5 s before the sample. The iron-saturated sample was injected for 1 s. Voltage: 28 kV; detection: 214 nm. (From Wutys, B. et al., Clin. Chem. Lab. Med., 39, 937, 2001. With permission.)

686

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

22.5.2.1.3 Concluding Remarks The CZE analysis of Tf seems to be an alternative to CDT measurements kits, which are prone to inaccurate diagnostics. Normal CDT ranges are method-dependent and, consequently, results must be interpreted by taking into account method-specific cutoff values [176,184]. In any case, several authors support the application of CZE for CDT determination in the clinical laboratory [191] and claim that it represents an alternative to HPLC and that it should be taken into account as a reference method for CDT [67]. In comparison to HPLC, CZE methods have the advantages of easier sample preparation, faster analysis times, higher isoforms resolution, and faster column reconditioning [197]. 22.5.2.2 Alpha-1-Acid Glycoprotein AGP or orosomucoid is synthesized mainly in the liver. It is an acute-phase protein with an unclear function, although it is usually accepted as a natural immunomodulatory and anti-inflammatory agent. It is a 41–43 kDa protein with a pI of 2.8–3.8 with 5 N-complex type glycans. Terminal sialic acid residues are usually present and contribute to the acidity of the protein [209]. Since it is expressed by various alleles at two loci, it exhibits genetic polymorphism [210]. Changes in the glycosylation of AGP, as well as in the distribution of genetic AGP forms, have been related to several pathophysiological states, such as cancer [210–213], RA [214,215], and other types of inflammation [216–218]. There are several studies that have described the separation of AGP forms by CE [78,84,87,100,109,117,219,220]. However, only a few authors have carried out CE analysis of AGP samples from patients in order to study if the technique is suitable for distinguishing the changes described earlier. Kinoshita et al. [78] analyzed AGP purified from rat sera from normal and inflammation states. They separated electrokinetically injected AGP into several nonbaseline resolved peaks in a DB-1 capillary with an acetate buffer (pH 4.1) containing HPMC. They found qualitative differences between the two samples. The sample from rat sera in the inflammation state had more AGP bands moving slower than the AGP sample from sera in normal state. The authors speculated about those bands having decreased sialic acid content. Later on, Kakehi et al. [219] proposed an interesting approach to the analysis of AGP by CZE. They performed electrokinetic injection of desalted serum in a DB-1 coated capillary with a running buffer at pH 4.5. They assumed that almost all serum proteins have pIs above that pH except for AGP, which is negatively charged at that pH. Therefore, a selective electro-injection of AGP was performed. They obtained nonbaseline resolution of 10 AGP forms. They analyzed two serum samples from patients who had acquired methicillin-resistant Staphylococcus aureus during hospitalization and one serum sample from a healthy volunteer. However, they did not find a relationship between the abundance of each glycoform and the clinical data of patients. A completely different approach, based on lectin-affinity CE was used by Bergstrom et al. [117] to separate AGP fractions. As mentioned in a previous section, using Con A as affinity ligand, they separated AGP into two peaks according to the biantennary content of the glycoforms. They applied this methodology to two AGP samples from patients with severe RA and one sample from a healthy donor. They found that the RA samples showed a decrease in the relative peak area of the biantennary peak compared with normal AGP. In another study, a CZE method capable of baseline separation of up to 11 AGP bands was developed [84]. This method, together with a statistical program that allowed to correctly compare AGP bands from different samples, was used to compare three AGP samples from one healthy donor and two pools of sera from cancer patients. Results showed that samples from cancer patients had one extra AGP band with a higher charge to mass ratio, while the AGP profile from the healthy donor presented one extra band with lower charge to mass ratio. A quantitative study of the samples demonstrated that there were peak area differences for the AGP forms. The cancer patient samples

Glycoprotein Analysis by Capillary Electrophoresis

687

presented higher proportion of forms with higher charge to mass ratio and lower proportion of forms with lower charge to mass ratio when compared with AGP from the healthy donor (Figure 22.29). The same authors also developed a CIEF method to separate AGP isoforms and applied it to the analysis of AGP from sera from ovary cancer patients [109]. The studies discussed earlier represent the beginning of research efforts directed toward the utility of CE analysis of AGP for clinical uses. In all cases, AGP samples from people and animals with different pathophysiological conditions were analyzed as a proof of concept for the methods developed. Studies of large populations are needed to make clinical conclusions. 22.5.2.3 Other Glycoproteins hCG is a glycohormone of 38,000 Da that consists of two noncovalently bound subunits. Both subunits (α and β, respectively) contain two N-glycans and the β-subunit also contains four O-glycosylated chains. Its role is to avoid the disintegration of the corpus luteum in the ovary. The more acidic hCG glycoforms have increased biological activity. In addition, the glycosylation pattern of this hormone seems to change in patients with trophoblastic disease [56]. As indicated previously, this glycoprotein (in its native heterodimer form) was separated into eight peaks in an uncoated capillary with a borate buffer containing DAP [56]. The method was used to analyze a reference hCG, two samples purified from crude urinary hCG, and one hCG sample from the urine of a patient with metastatic choriocarcinoma. The four hCG electropherograms had the same number of forms, but the relative concentration of each one appeared to vary. However, the hCG sample coming from the cancer patient migrated in a way that did not allow an accurate comparison with the other samples. Those peaks could be aligned with either the previous or subsequent peak, resulting in a different quantitative result. Hiroaka et al. [221] developed a one-step CIEF method for the analysis of CSF proteins with molecular masses between 10,000 and 50,000 Da. They found differences in the distribution of lipocalin-type prostaglandin D synthase isoforms (a sialic acid-containing glycoprotein) in the CSF of patients with certain neurological disorders. It was speculated that the four different peaks attributable to this protein were due to different numbers of sialic acid residues. PSA is a glycoprotein that is used as a biomarker for prostate cancer. Its release into blood is increased during the development of prostate cancer. It has a MW of 28,430 Da and contains one N-glycosylation site. A fraction of circulating PSA is bound to plasma proteins. The measurement of unbound PSA increases the diagnosis specificity of total blood PSA [222]. However, these measurements do not take into account the glycosylation pattern of PSA. It was recently stated that when comparing PSA oligosaccharides from healthy donors with that from a prostate tumor cell line, the PSA glycosylation pattern may be different [223]. In that case, the use of CZE for the comparative analysis of PSA glycosylation patterns in samples from healthy persons and prostate cancer patients would contribute to the study of PSA glycosylation as prostate marker. Donohue et al. [167] obtained a separation of PSA forms from several manufacturers in 12 min in an uncoated capillary with a borate buffer containing DAP using UV detection. The commercial PSA samples had been purified from human seminal fluid. As mentioned earlier in this chapter, an interesting finding of this work was that the purification procedure influences the PSA pattern.

22.6 CONCLUDING REMARKS AND FUTURE PROSPECTS As demonstrated in this chapter, CE has clearly become a widely applicable tool for the analysis of glycoforms of glycoproteins, without which the determination of some significant qualitative and quantitative information would not have been possible. This is aptly exemplified from some of the earliest work in the field on Tf, work which has culminated in the use of CE in clinical settings, to a more recent application of CE-MS from which over 130 glycoforms of rhEPO were

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Area percentage of each AGP band

Absorbance (214 nm)

Absorbance (214 nm)

Absorbance (214 nm)

688

G

0.04 (a)

F

H

E

I

D

-0.01

15

0.04

J

C

AB

20 t (min)

25

(b) G

F D

BC -0.01

0.04

K

15

H I J

E

K L

20 t (min)

(c)

25

G

H

I

F D BC -0.01

15

30

(d)

J

E

K

L

20 t (min)

25

25

20 AGP pool from ovary carcinoma patients 15

AGP pool from lymphoma patients AGP from a healthy donor

10

5

0 A

B

Lower

C

D

E

F

G

H

Charge/mass ratio

I

J

K

L AGP bands

Higher

FIGURE 22.29 (a, b, and c): Electropherograms of AGP purified from serum of a healthy donor, a pool of sera from ovary cancer patients, and a pool of sera from lymphoma patients, respectively. A statistical program was used to compare the three electropherograms, showing the different AGP profile between the healthy donor and the cancer patients. Analytical conditions: fused-silica capillary 77 cm × 50 µm i.d; buffer: 0.01 M Tricine, 0.01 M NaCl, 0.01 M sodium acetate, 7 M urea and 3.9 mM putrescine, pH 4.5; voltage: 25 kV, 35◦ C. (d) Mean percentage areas and standard deviation (indicated as “I” and calculated with standard AGP, n = 6) of AGP peaks in real samples (2 injections each). Names of the AGP peaks according to electropherograms (a), (b), and (c). Discontinuous lines are used to link the real values for the sake of clarity of the representation. (Adapted from Lacunza, I. et al., Electrophoresis, 27, 4205, 2006. With permission.)

Glycoprotein Analysis by Capillary Electrophoresis

689

identified. Undoubtedly, efforts at understanding the separation mechanism of glycoforms have made it possible. The applicability of CE at most stages of biopharmaceutical manufacturing processes as well as its usefulness for comparing products from different sources, whether of recombinant or natural origins, has led most manufacturers to endorse this technique. CE methods have also appeared in pharmacopeial monographs to replace the more laborious, conventional gel methods. A list of several proteins for which glycoforms have been separated and the CE method employed in each case can be seen in Table 22.3. One particular aspect of CE analysis that requires improvement to make its practical application wider is sample preparation. There is little sense in developing simple and fast CE methods if

TABLE 22.3 Separation of Glycoprotein Forms by Capillary Electrophoresis Protein

Separation Mode

24 kDa glycoprotein, recombinant Alpha-1-acid glycoprotein, human (AGP) (orosomucoid)

CZE

Berkowitz, S.A. et al., J. Chromatogr. A, 1079, 254, 2005

ACE CIEF

CZE

Bergstrom, M. et al., J. Chromatogr. B, 809, 323, 2004 Wu, J. et al., J. Chromatogr. A, 817, 163, 1998 Lacunza, I. and de Frutos, M., PACE Setter, 10, 5, 2006 Lacunza, I. et al., Electrophoresis, 28, 1204, 2007 James, D.C. et al., Anal. Biochem., 222, 315, 1994 Kubo, K., J. Chromatogr. B, 697, 217, 1997 Kinoshita, M. et al., J. Chromatogr. A, 866, 261, 2000 Pacakova, V. et al., Electrophoresis, 22, 459, 2001 Kakehi, K. et al., Anal. Chem., 73, 2640, 2001 Sei, K. et al., J. Chromatogr. A, 958, 273, 2002 Lacunza, I. et al., Electrophoresis, 27, 4205, 2006 Che, F.-Y. et al., Electrophoresis, 20, 2930, 1999

CZE

Kinoshita, M. et al., J. Chromatogr. A, 866, 261, 2000 Balaguer, E. and Neususs, C., Anal. Chem., 78, 5384, 2006 Kinoshita, M. et al., J. Chromatogr. A, 866, 261, 2000

CZE

Kinoshita, M. et al., J. Chromatogr. A, 866, 261, 2000

CZE

Kubo, K., J. Chromatogr. B, 697, 217, 1997 Chang, W.W.P. et al., Electrophoresis, 26, 2179, 2005 Kremser, L. et al., Electrophoresis, 24, 4282, 2003 Demelbauer, U.M. et al., Electrophoresis, 25, 2026, 2004 Kremser, L. et al., Electrophoresis, 24, 4282, 2003 Reif, O.-W. and Freitag, R., J. Chromatogr. A, 680, 383, 1994 Reif, O.-W. and Freitag, R., J. Chromatogr. A, 680, 383, 1994 Buchacher, A. et al, J. Chromatogr. A, 802, 355, 1998 Buchacher, A. et al., J. Chromatogr. A, 802, 355, 1998 Bateman, K.P. et al., Methods Mol. Biol., 213, 219, 2003 Hiraoka, A. et al., Electrophoresis, 22, 3433, 2001 Yim, K. et al., J. Chromatogr. A, 716, 401, 1995 Yeung, B. et al., Anal. Chem., 69, 2510, 1997 Cherkaoui, S. et al., J. Chromatogr. A, 790, 195, 1997 Cherkaoui, S. et al, Chromatographia, 50, 311, 1999 Daali, Y. et al., J. Pharm. Biomed. Anal., 24, 849, 2001 Wu, S.-L. et al., J. Chromatogr., 516, 115, 1990

MEKC CZE

Alpha-1-acid glycoprotein, bovine (bovine AGP)

Alpha-1-acid glycoprotein, rat (rat AGP) Alpha-1-acid glycoprotein, sheep (sheep AGP) Alpha-1-antitrypsin (AAT) Antithrombin III, human (ATIII)

CIEF CZE

Antithrombin III, recombinant human (rhATIII)

CZE CIEF

Antithrombin III β Avidin Beta-trace proteins (BTP) Bone morphogenic protein-2, recombinant human (rhBMP-2) Caseinoglycomacropeptide

CE-SDS CZE CIEF CZE

CD4, recombinant (rCD4)

CZE

CZE

Reference

690

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

TABLE 22.3 (Continued) Protein

Separation Mode

Cellobiohydrolase I (CBH-I) Cell adhesion molecule (CAM) Chorionic gonadotropin, human (hCG) Darbepoetin-α (novel erythropoiesis stimulating protein, NESP) Desmodus salivary plasminogen activator (DSPAα1) DNAse, recombinant human (rhDNAse) Erythropoietin, recombinant human (rhEPO)

CIEF CITP CZE CZE

CZE CZE CZE

CIEF

Erythropoietin, urinary (uEPO)

CITP CZE

Factor VIIa, recombinant human

CZE

Factor IX, human Fetuin, bovine

CIEF CZE

Fetuin

CE-SDS MEKC

Reference Sandra, K. et al., J. Chromatogr. A, 1058, 263, 2004 Josic, D. et al., J. Chromatogr., 516, 89, 1990 Morbeck, D.E. et al., J. Chromatogr. A, 680, 217, 1994 Oda, R.P. et al, J. Chromatogr. A, 680, 85, 1994 Lacunza, I. et al., Electrophoresis, 25, 1569, 2004 Sanz-Nebot, V. et al., Electrophoresis, 26, 1451, 2005 Apfel, A. et al., J. Chromatogr. A, 717, 41, 1995 Chakel, J.A. et al., J. Chromatogr. B, 689, 215, 1997 Felten, C. et al., J. Chromatogr. A, 853, 295, 1999 Quan, C.P. et al., Chromatographia Suppl., 53, S39, 2001 Tran, A.D. et al., J. Chromatogr., 542, 459, 1991 Watson, E. and Yao, F., Anal. Biochem., 210, 389, 1993 Nieto, O. et al., Anal. Commun., 33, 425, 1996 Bietlot, H.P. and Girard, M., J. Chromatogr. A, 759, 177, 1997 Kaniansky, D. et al., J. Chromatogr. A, 772, 103, 1997 Zhou, G.-H. et al., Electrophoresis, 19, 2348, 1998 Bristow, A. and Charton, E., Pharmaeuropa, 11, 290, 1999 Che, F.-Y. et al., Electrophoresis, 20, 2930, 1999 Kinoshita, M. et al., J. Chromatogr. A, 866, 261, 2000 Lopez-Soto-Yarritu, P. et al., J. Sep. Sci., 25, 1112, 2002 Sanz-Nebot, V. et al., Anal. Chem. 75, 5220, 2003 European Pharmacopoeia 5th edition, published by EDQM. June 2004 Lacunza, I. et al., Electrophoresis, 25, 1569, 2004 Neususs, C. et al., Electrophoresis, 26, 1442, 2005 Madajova, V. et al., Electrophoresis, 26, 2664, 2005 Yu, B. et al., J. Sep. Sci., 28, 2390, 2005 Balaguer, E. and Neususs, C., Chromatographia, 64, 351, 2006 Balaguer, E. and Neususs, C., Anal. Chem., 78, 5384, 2006 Balaguer, E. et al., Electrophoresis, 27, 2638, 2006 Kamoda, S. and Kakehi, K., Electrophoresis, 27, 2495, 2006 Lara-Quintanar, P. et al., J. Chromatogr. A, 1153, 227, 2007 Kubach, J. and Grimm, R., J. Chromatogr. A, 737, 281, 1996 Cifuentes, A. et al., J. Chromatogr. A, 830, 453, 1999 Lopez-Soto-Yarritu, P. et al., J. Chromatogr. A, 968, 221, 2002 Madajova, V. et al., Electrophoresis, 26, 2664, 2005 de Frutos, M. et al., Electrophoresis, 24, 678, 2003 Yu, B. et al., J. Sep. Sci., 28, 2390, 2005 Klausen, N.K. and Kornfelt, T., J. Chromatogr. A, 718, 195, 1995 Buchacher, A. et al., J. Chromatogr. A, 802, 355, 1998 Kinoshita, M. et al., J. Chromatogr. A, 866, 261, 2000 Balaguer, E. and Neususs, C., Anal. Chem., 78, 5384, 2006 Werner, W.E. et al., Anal. Biochem., 212, 253, 1993 James, D.C. et al., Anal. Biochem. 222, 315, 1994 Continued

Glycoprotein Analysis by Capillary Electrophoresis

691

TABLE 22.3 (Continued) Protein

Separation Mode

FG basic chimeric glycoprotein, recombinant Follicle stimulating hormone, recombinant human (rhFSH)

CZE

Tsuji, K. and Little, R.J., J. Chromatogr., 594, 317, 1992

CZE

Mulders, J.W.M. et al., 3rd WCBP meeting, Washington, DC, 1999 Storring, P.L. et al., Biologicals, 30, 217, 2002

Granulocyte colony-stimulating factor, recombinant human (rhG-CSF) Hepatitis C virus highly glycosylated protein, recombinant Hirudin, novel O-glycosylated P6 (leech) HIV gp120, recombinant

CZE

MEKC

Watson, E. and Yao, F., J. Chromatogr., 630, 442, 1993 Somerville, L. E. et al., J. Chromatogr. B, 732, 81, 1999 Zhou, G.-H. et al., J. Pharm. Biomed. Anal., 35, 425, 2004 Kundu, S. et al., J. Cap. Electrophor., 3, 301, 1996

CZE

Steiner, V. et al., Biochemistry, 31, 2294, 1992

MEKC CIEF MEKC CZE

Jones, D.H. et al., Vaccine, 13, 991, 1995 Tran, N.T. et al., J. Chromatogr. A, 866, 121, 2000 James, D.C. et al., Anal. Biochem., 222, 315, 1994 Kelly, J.F. et al., J. Chromatogr. A, 720, 409, 1996

Interferon-γ, recombinant human (rhIFN-γ)

MEKC

Interferon-ω Interleukin-2 (IL-2)

MEKC CZE

James, D.C. et al., Anal. Biochem. 222, 315, 1994 James, D. C. et al., Prot. Sci., 5, 331, 1996 Goldman, M.H. et al., Biotech. Bioeng., 60, 597, 1998 Hooker, A.D. and James, D.C., Mol. Biotech., 14, 241, 2000 Kopp, K. et al., Arzneim. Forsch./Drug Res., 46, 1191, 1996 Knuver-Hopf, J. and Mohr, H., J. Chromatogr. A, 717, 71, 1995

Lipocalin-type prostaglandin D synthase Monoclonal anti-alpha-1 antitrypsin Monoclonal antibody HER2, recombinant humanized (rhuMAbHER2) Monoclonal antibody antihuman follicle stimulating hormone, mouse Monoclonal antibodies anti-HIV gp-41 Monoclonal antibody, unspecified (mAb) Monoclonal antibody, unspecified (mAb) Monoclonal antibody, unspecified (mAb) Monoclonal antibody, unspecified (mAb)

CIEF

Hiroaka, A. et al., Electrophoresis, 22, 3433, 2001

CIEF

Wu, J. et al., J. Chromatogr. A, 817, 163, 1998

CIEF CE-SDS CZE

Hunt, G. et al., J. Chromatogr. A, 744, 295, 1996 Harris, R.J., J. Chromatogr. A, 705, 129, 1995 Hunt, G. et al., J. Chromatogr. A, 744, 295, 1996 Kelly, J.A. and Lee, C.S., J. Chromatogr. A, 790, 207, 1997

CZE

Wenisch, E., et al., J. Chromatogr., 516, 13, 1990

CZE

Compton, B.J. J. Chromatogr., 559, 357, 1991

CIEF

Kubach, J. and Grimm, R., J. Chromatogr. A, 737, 281, 1996

CIEF

Schwer, C., Electrophoresis, 16, 2121, 1995

CE-SDS

Salas-Solano, O. et al., Anal.Chem., 78, 6583, 2006

Horseradish peroxidase (HRP)

Reference

Continued

692

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

TABLE 22.3 (Continued) Protein

Separation Mode

Ovalbumin (OVA)

CZE

Ovalbumin (turkey)

ACE CZE

Pepsin Placental alkaline phosphatase Pollen allergens Prostate-specific antigen (PSA) Proteinase A (S. cerevisiae)

CZE CZE CZE CZE CZE

Ribonuclease B (RNase B)

MEKC

Bullock, J.A. and Yuan, L.C., J. Microcol. Sep., 3, 241, 1991 Landers, J.P. et al., Anal. Biochem., 205, 115, 1992 Taverna, M. et al., Electrophoresis, 13, 359, 1992 Bullock, J., J. Chromatogr., 633, 235, 1993 Kelly, J.F. et al., Beckman discovery Series 2, 1993 Thibault, P. et al., HPCE’94, San Diego, CA 1994, Abstract P-323, p. 126 Oda, R.P. et al, J. Chromatogr. A, 680, 85, 1994 Legaz, M.E. and Pedrosa, M.M., J. Chromatogr. A, 719, 159, 1996 Oda, R.P. and Landers, J.P., Mol. Biotechnol. 5, 165, 1996 Chakel, J.A. et al., J. Chromatogr. B, 689, 215, 1997 Chen, Y., J. Chromatogr. A, 768, 39, 1997 Pinto, D.M. et al., Anal. Chem., 69, 3015, 1997 Che, F.Y., et al., J. Chromatogr. A, 849, 599, 1999 Kubo K. and Hattori, A., Electrophoresis, 22, 3389, 2001 Pacakova, V. et al., Electrophoresis, 22, 459, 2001 Catai, J.R. et al., J. Chromatogr. A, 1083, 185, 2005 Uegaki, K. et al., Anal. Biochem., 309, 269, 2002 Che, F.Y., et al., J. Chromatogr. A, 849, 599, 1999

ACE

Landers, J.P. et al, Anal. Biochem., 205, 115, 1992 Eriksson, H.J.C. et al., J. Chromatogr. B, 755, 311, 2001 Pacakova, V. et al., Electrophoresis, 22, 459, 2001 Donohue, M. J. et al., Anal. Biochem., 339, 318, 2005 Pedersen, J. and Biedermann, K., Biotechnol. Appl. Biochem. 18, 377, 1993 Rudd, P.M. et al., Glycoconj. J., 9, 86, 1992 James, D.C. et al., Anal. Biochem., 222, 315, 1994 Rudd, P.M. et al., Biochemistry, 33, 17, 1994 Grossman, P.D. et al., Anal. Chem., 61, 1186, 1989 Kelly, J.F. et al., J. Chromatogr. A, 720, 409, 1996 Yeung, B. et al., Anal. Chem., 69, 2510, 1997 Bateman, K.P. et al., Meth. Mol. Biol., 213, 219, 2003 Uegaki, K. et al., Anal. Biochem., 309, 269, 2002

CZE

Kim, J.W. et al., Clin. Chem., 39, 689, 1993

CGE

Tsuji, K., J. Chromatogr. A, 652, 139, 1993

CZE

Wenisch, E. et al., J. Chromatogr., 516, 13, 1990

CZE

Wu, S.-L. et al., J. Chromatogr., 516, 115, 1990 Yim, K., J. Chromatogr., 559, 401, 1991 Taverna, M. et al., Electrophoresis, 13, 359, 1992 Thorne, J. M. et al., J. Chromatogr. A, 744, 155, 1996 Yim, K., J. Chromatogr., 559, 401, 1991 Taverna, M. et al., Electrophoresis, 13, 359, 1992 Moorhouse, K.G. et al., J. Chromatogr. A, 717, 61, 1995 Moorhouse, K.G. et al., Electrophoresis, 17, 423, 1996 Chen, A.B. et al., J. Chromatogr. A, 744, 279, 1996

CZE

Serum proteins (albumin, globulins, IgG), human Somatropin, recombinant bovine (rbSt) Superoxide dismutase, recombinant Tissue plasminogen activator, recombinant human (rhtPA)

Reference

CIEF

Glycoprotein Analysis by Capillary Electrophoresis

693

TABLE 22.3 (Continued) Protein

Transferrin, human (Tf)

Separation Mode

CGE-SDS CZE

CIEF

Trypsin inhibitor, human urinary (UTI) Tumor necrosis factor receptor fusion protein, recombinant human (rhTNFR:Fc)

Reference Thorne, J.M. et al., J. Chromatogr. A, 744, 155, 1996 Kubach, J. and Grimm, R., J. Chromatogr. A, 737, 281, 1996 Thorne, J.M et al., J. Chromatogr. A, 744, 155, 1996 Kilar, F. and Hjerten, S., J. Chromatogr., 480, 351, 1989 Bergmann, J. et al., Pharmazie, 51, 644, 1996 Oda, R.P. and Landers, J.P., Electrophoresis, 17, 431, 1996 Iourin, O. et al., Glycoconj. J., 13, 1031, 1996 Kubo, K., J. Chromatogr. B, 697, 217, 1997 Prasad, R. et al., Electrophoresis, 18, 1814, 1997 Oda, R.P. et al., Electrophoresis, 18, 1819, 1997 Tagliaro, F. et al., Electrophoresis, 19, 3033, 1998 Tagliaro, F. et al., J. Cap. Electrophor Microchip Tech., 5, 137, 1999 Beisler, A.T. et al., Anal. Biochem., 285, 143, 2000 Giordano, B.C. et al., J. Chromatogr. B, 742, 79, 2000 Trout, A.L. et al., Electrophoresis, 21, 2376, 2000 Crivellente, F. et al., J. Chromatogr. B, 739, 81, 2000 Wuyts, B. et al., Clin. Chem. Lab. Med., 39, 937, 2001 Wuyts, B. et al., Clin. Chem., 47, 247, 2001 Lanz, C. et al., J. Chromatogr. A, 979, 43, 2002 Legros, F.J. et al., Clin. Chem., 48, 2177, 2002 Ramdani, B. et al., Clin. Chem., 49, 1854, 2003 Sanz-Nebot, V. et al., J. Chromatogr. B, 798, 1, 2003 Legros, F.J. et al., Clin. Chem., 49, 440, 2003 Wuyts, B. and Delanghe, J.R., Clin. Chem. Lab. Med., 41, 739, 2003 Lanz, C. et al., J. Chromatogr. A, 1013, 131, 2003 Lanz, C. and Thormann, W., Electrophoresis, 24, 4272, 2003 Ramdani, B. et al., Clin. Chem., 49, 1854, 2003 Fermo, I. et al., Electrophoresis, 25, 469, 2004 Lanz, C. et al., Electrophoresis, 25, 2309, 2004 Martello, S. et al., Forensic Sci. Internat., 141, 153, 2004 Carchon, H.A. et al., Clin. Chem., 50, 101, 2004 Bortolotti, F. et al., Clin. Chem., 51, 2368, 2005 Chang, W.W.P. et al., Electrophoresis, 26, 2179, 2005 Joneli, J. et al., J. Chromatogr. A, 1130, 272, 2006 Kilar, F. and Hjerten, S., Electrophoresis, 10, 23, 1989 Kilar, F. and Hjerten, S., J. Chromatogr., 480, 351, 1989 Molteni, S. and Thormann, W., J. Chromatogr., 638, 187, 1993

CZE

Che, F.-Y. et al., Electrophoresis 20, 2930, 1999

CIEF

Jochheim, C. et al., Chromatographia Suppl., 53, S59, 2001

they first require time-consuming and labor-intensive steps for obtaining adequate samples. Additional efforts along the lines of work described in this chapter for developing CE methods that do not require a prior purification step, or developing fast and simple purification methods will be required. In combination to more general purification methods, specific ones could be included

694

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

to reach this goal. The use of affinity-based systems using lectins, antibodies, or aptamers may provide elegant solutions in this regard. The online combination of an affinity purification step with the CE separation has been demonstrated recently and appears promising even in cases where high throughput is achieved by using multicapillary systems, such as the one described for other analytes [224]. One of the major challenges to be faced in future years will be that of subsequent entry biologics (Canada)/follow-on biologics (USA)/biosimilar medicines (EU), and the ability to establish parameters that will allow the comparison of complex glycoproteins whether it is at the glycoform or isoform levels. Work in this area has been reported in recent years but it is anticipated that with the expiration of patents for several biopharmaceuticals currently on the market, many of them will be produced under a wide variety of manufacturing conditions. This will be even more challenging as it has been shown in this chapter that many parameters can affect the glycosylation profiles. Consequently, the refinement of CE methods will be required in order to improve peak resolution as well as to detect minor glycoforms. The development of multidimensional, multimodal CE, that is, the online coupling of two or more CE separation modes has been reported several years ago. However, despite its potential for the analysis of complex mixtures such as glycoproteins, its application to practical problems has been limited and it is expected that more applications will be reported. The field of microchip CE (MCE) has progressed tremendously over the past few years and its application to the analysis of glycoproteins has only been recently reported [225] and will undoubtedly continue to find applications in the coming years. The majority of the work on CE of glycoforms carried out up to now is related to methods development and the successful demonstration of proof of concept for specific applications. This should lead as a logical next step to its application to a larger number of practical situations, especially in the clinical and diagnostic fields. This will require the methods to be applied to large, well-controlled populations to achieve well-defined conclusions about their validity. However, the application of CE techniques (including MCE) to the clinical field for specific applications such as the analysis of glycoforms as disease markers may be somewhat restrained (if not completely prevented) if more sensitive detection methods are not available. In this sense, newer detection schemes, perhaps involving the development of new fluorescent labeling agents for LIF detection of glycoproteins are required.

ACKNOWLEDGMENTS Financial support from Spanish Ministry of Education and Science (Projects TIC2003-01906, HH04-33, CTQ2006-05214, DEP2006-56207-CO3-01), Fundacion Ramon Areces (Project Biosensors), Fundacion Domingo Martinez (project Microorganisms), and Comunidad de Madrid (Project S2006/GEN-0247) is acknowledged.

ABBREVIATIONS ACE AGP Ara Asn AAT ATIII BGE

Affinity capillary electrophoresis Alpha-1-acid glycoprotein, orosomucoid Arabinose Asparagine Alpha-1-antitrypsin Antithrombin III Background electrolyte

Glycoprotein Analysis by Capillary Electrophoresis

BRP BSF BTP CAPS CBH-I CDG CDT CE CE-MS CE-SDS CGMP CHO CIEF CITP Con A CRP CSF CZE DAB DAP DcBr EACA EOF EP EPO ESI ESI-TOF MS Fuc Gal GalNAc GlcNAc Glu HC hCG HEPPSO HPLC HPMC HRP HSA hTNF Hyl Hyp IEF LC LCA LIF mAb MALDI Man

Biological Reference Preparation Bovine serum fetuin 1,3-Bis[tris(hydroxymethyl)-methylamino] propane 3-(Cyclohexylamino)-1-propanesulfonic acid Cellobiohydrolase I Congenital disorders of glycosylation Carbohydrate-deficient transferrin Capillary electrophoresis Online coupling of CE to mass spectrometry Capillary electrophoresis in sodium dodecylsulfate Caseinoglycomacropeptide Chinese hamster ovary Capillary isoelectric focusing Capillary isotacophoresis Concanavalin A C-reactive protein Cerebrospinal fluid Capillary zone electrophoresis 1,4-Diaminobutane, putrescine 1,3-Diaminopropane Decamethonium bromide ε-Aminocaproic acid Electroosmotic flow European Pharmacopoeia Erythropoietin Electrospray ionization Electrospray ionization time of flight mass spectrometry Fucose Galactose N-Acetylgalactosamine N-Acetylglucosamine Glucose Heavy chain Human chorionic gonadotropin N-(2-hydroxyethyl)piperazin-N ′ -2-(hydroxypropane sulfonic acid) High performance liquid chromatography Hydroxypropylmethylcellulose Horseradish peroxidase Human serum albumin Human tumour necrosis factor Hydroxylysine Hydroxyproline Isoelectric focusing Light chain Lens culinaris agglutinin Laser-induced fluorescence Monoclonal antibody Matrix-assisted laser desorption ionization Mannose

695

696

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

MEKC MHEC MS NESP NeuAc OVA PAA PAGE PB pI PSA PTM PVA RA rhATIII rhBMP-2 rhDNAse rhEPO rhFSH rhGCSF rhGMCSF rhIFN-γ rhtPA RNase SDS SDS–PAGE Ser SIM TEMED Tf Thr TOF uEPO UV Xyl

Micellar electrokinetic chromatography Methylhydroxyethylcellulose Mass spectrometry Novel erythropoiesis-stimulating protein N-Acetylneuraminic acid Ovalbumin Polyacrylamide Polyacrylamide gel electrophoresis Polybrene Isoelectric point Prostate specific antigen Postranslational modification Polyvinyl alcohol Rheumatoid arthritis Recombinant human anti-thrombin III Recombinant human bone morphogenic protein 2 Recombinant human deoxyribonuclease Recombinant human erythropoietin Recombinant human follicle-stimulating hormone Recombinant human granulocyte colony-stimulating factor Recombinant human granulocyte macrophage colony-stimulating factor Recombinant human interferon-γ Recombinant human tissue plasminogen activator Ribonuclease Sodium dodecylsulfate Sodium dodecylsulfate polyacrylamide gel electrophoresis Serine Selected ion monitoring N,N,N ′ ,N ′ -Tetramethylene diamine Transferrin Threonine Time-of-flight Human urinary EPO Ultraviolet Xylose

REFERENCES 1. Dwek, R.A., Glycobiology: Toward understanding the function of sugars, Chem. Rev., 96, 683, 1996. 2. Taylor, M.E. and Drickamer, K., Introduction to Glycobiolgy, Oxford University Press, New York, 2004, chap. 12. 3. Brockhausen, I. and Kuhns, W., Glycoproteins and Human Disease, Springer-Verlag, Heidelberg, Germany, 1997, chap. 2. 4. Durand, G. and Seta, N., Protein glycosylation and diseases: Blood and urinary oligosaccharides as markers for diagnosis and therapeutic monitoring, Clin. Chem., 46, 795, 2000. 5. Walsh, G., Pharmaceutical biotechnology products approved within the European Union, Eur. J. Pharm. Biopharm., 55, 3, 2003. 6. Walsh, G., Second-generation biopharmaceuticals, Eur. J. Pharm. Biopharm., 58, 185, 2004. 7. Jenkins, N. and Curling, M.A., Glycosylation of recombinant proteins: Problems and prospects, Enzyme Microb. Technol., 16, 354, 1994.

Glycoprotein Analysis by Capillary Electrophoresis

697

8. Jung, E. and Williams, K.L., The production of recombinant glycoproteins with special reference to simple eukaryotes including Dictyostelium discoideum, Biotechnol. Appl. Biochem., 25, 3, 1997. 9. Varki, A. et al., Essentials of Glycobiology, Cold Spring Harbor Laboratory Press, New York, 1999, chap. 2. 10. Taylor, M.E. and Drickamer, K., Introduction to Glycobiolgy, Oxford University Press, New York, 2004, chap. 2. 11. Strege, M.A. and Lagu, A.L., Capillary Electrophoresis of Proteins and Peptides, Meth. Mol. Biol., Humana Press, Inc., NJ, 2004, Vol. 276. 12. Prichett, T. and Robey, F.A., Capillary electrophoresis of proteins, in Handbook of Capillary Electrophoresis, Landers, J.P., Ed., CRC Press, Boca Raton, FL, 1997, chap. 9. 13. Kakehi, K., Kinoshita, M., and Nakano, M., Analysis of glycoproteins and the oligosaccharides thereof by high-performance capillary electrophoresis—Significance in regulatory studies on biopharmaceutical products, Biomed. Chromatogr., 16, 103, 2002. 14. Pantazaki, A., Taverna, M., and Vidal-Madjar, C., Recent advances in the capillary electrophoresis of recombinant glycoproteins, Anal. Chim. Acta, 383, 137, 1999. 15. Taverna, M. et al., Electrophoretic methods for process monitoring and the quality assessment of recombinant glycopropteins, Electrophoresis, 19, 2572, 1999. 16. Taverna, M., Tran, N.T., and Ferrier, D., Separation of protein glycoforms by capillary electrophoresis, Meth. Mol. Biol. 213, 163, 2003. 17. Kamoda, S. and Kakehi, K., Capillary electrophoresis for the analysis of glycoprotein pharmaceuticals, Electrophoresis, 27, 2495, 2006. 18. Watzig, H., Degenhardt, M., and Kunkel, A., Strategies for capillary electrophoresis: Method development and validation for pharmaceutical and biological applications, Electrophoresis, 19, 2695, 1998. 19. Chen, A.B. and Canova-Davis, E., Capillary electrophoresis in the development of recombinant protein biopharmaceuticals, Chromatographia Suppl., 53, S7, 2001. 20. Patrick, J.S. and Lagu, A.L., Review applications of capillary electrophoresis to the analysis of biotechnology-derived therapeutic proteins, Electrophoresis, 22, 4179, 2001. 21. Mechref, Y. and Novotny, M.V., Structural investigations of glycoconjugates at high sensitivity, Chem. Rev., 102, 321, 2002. 22. Little, M.J., Paquette, D.M., and Roos, P.K., Electrophoresis of pharmaceutical proteins: Status quo, Electrophoresis, 27, 2477, 2006. 23. Dolnik, V., Capillary electrophoresis of proteins 2003–2005, Electrophoresis, 27, 126, 2006. 24. Huck, C.W., Bakry, R., and Bonn, G.K., Progress in capillary electrophoresis of biomarkers and metabolites between 2002 and 2005, Electrophoresis, 27, 111, 2006. 25. Kishino, S. and Miyazaki, K., Separation methods for glycoprotein analysis and preparation, J. Chromatogr. A, 699, 371, 1997. 26. El Rassi, Z., Recent developments in capillary electrophoresis of carbohydrate species, Electrophoresis, 18, 2400, 1997. 27. Grossman, P.D., Free-solution capillary electrophoresis, in Capillary Electrophoresis. Theory and Practice, Grossman, P.D. and Colburn J.C., Eds., Academic Press, New York, 1992, chap. 2. 28. Kalman, F. et al., Charge and size effect in capillary zone electrophoresis of nuclease A and variants, Electrophoresis, 16, 595, 1995. 29. Gitlin, I., Carbeck, J.D., and Whitesides, G.M., Why are proteins charged? Network of charge-charge interaction in proteins measured by charge ladders and capillary electrophoresis, Angew. Chem. Int. Ed., 45, 3022, 2006. 30. Balaguer, E. et al., Glycoform characterization of erythropoietin combining glycan and intact protein analysis by capillary electrophoresis-electrospray-time-of-flight mass spectrometry, Electrophoresis, 27, 2638, 2006. 31. Yim, K., Abrams, J., and Hsu, A., Capillary zone electrophoretic resolution of recombinant human bone morphogenic protein 2 glycoforms. An investigation into the separation mechanisms for an exquisite separation, J. Chromatogr. A, 716, 401, 1995. 32. Chen, F.-T.A., Rapid protein analysis by capillary electrophoresis, J. Chromatogr., 559, 445, 1991.

698

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

33. van der Greef, J., Stroobant, P., and van der Heijden, R., The role of analytical sciences in medical systems biology, Curr. Opin. Chem. Biol., 8, 559, 2004. 34. Bristow, A. and Charton, E., Assessment of the suitability of a capillary zone electrophoresis method for determining isoform distribution of erythropoietin Pharmaeuropa, 11, 290, 1999. 35. Erythropoietin concentrated solution 01/2005:1316, European Pharmacopoeia 5th edition, published by EDQM. June 2004. 36. Cifuentes, A. et al., Capillary isoelectric focusing of erythropoietin glycoforms and its comparison with flat-bed isoelectric focusing and capillary zone electrophoresis, J. Chromatogr. A, 830, 453, 1999. 37. San-Nebot, V. et al., Separation of recombinant human erythropoietin glycoforms by capillary electrophoresis using volatile electrolytes. Assessment of mass spectrometry for the characterization of erythropoietin glycoforms, Anal. Chem., 75, 5220, 2003. 38. Lacunza, I. et al., Selection of migration parameters for a highly reliable assignment of bands of isoforms of erythropoietin separated by capillary electrophoresis, Electrophoresis, 25, 1569, 2004. 39. Lopez-Soto-Yarritu, P. et al., Improved capillary isoelectric focusing method for recombinant erythropoietin analysis, J. Chromatogr. A, 968, 221, 2002. 40. de Frutos, M., Cifuentes, A., and Diez-Masa, J.C., Differences in capillary electrophoresis profiles of urinary and recombinant erythropoietin, Electrophoresis, 24, 678, 2003. 41. Bietlot, H.P. and Girard, M., Analysis of recombinant human erythropoietin in drug formulations by high performance capillary electrophoresis, J. Chromatogr. A, 759, 177, 1997. 42. Neususs, C., Demelbauer, U., and Pelzing, M., Glycoform characterization of intact erythropoietin by capillary electrophoresis-electrospray-time of flight-mass spectrometry, Electrophoresis, 26, 1442, 2005. 43. Lara-Quintanar, P. et al., Immunochromatographic removal of albumin in erythropoietin biopharmaceutical formulations for its analysis by capillary electrophoresis, J. Chromatogr. A, 1153, 227, 2007. 44. Hjerten, S., Zone broadening in electrophoresis with special reference to the high-performance electrophoresis in capillaries: An interplay between theory and practice, Electrophoresis, 11, 665, 1990. 45. Lauer, H.H. and McManigill, D., Capillary zone electrophoresis of proteins in untreated fused silica tubing, Anal. Chem., 58, 166, 1986. 46. McCormick, R.A., Capillary zone electrophoretic separation of peptides and proteins using low pH buffers in modified silica capillaries, Anal. Chem., 6, 2322, 1986. 47. Corradini, D., Buffer additives other than the surfactant sodium dodecyl sulfate for protein separation by capillary electrophoresis, J. Chromatogr. B, 699, 21, 1997. 48. Righetti, P.G. et al., The state of the art of dynamic coatings, Electrophoresis, 22, 603, 2001. −O and 49. Ahmadzadeh, H., Dovichi, N.J., and Krylov, S. Capillary coating for protein separation on Si− −C covalent bond formation for capillary electrophoresis with laser-induced fluorescence detection, Si− in Capillary Electrophoresis of Proteins and Peptides, Strege, M.A. and Lagu, A.L., Eds., Humana Press, Totowa, NJ, 2004, chap. 2. 50. Landers, J.P. et al., High performance capillary electrophoresis of glycoproteins: The use of modifiers of electroosmotic flow for analysis of microheterogeneity, Anal. Biochem., 205, 115, 1992. 51. Legaz, M.E. and Pedrosa, M.M., Effect of polyamines on the separation of ovalbumin glycoforms by capillary electrophoresis, J. Chromatogr. A, 719, 159, 1996. 52. Watson, E. and Yao, F., Capillary electrophoretic separation of human recombinant erythropoietin (r-HuEPO) glycoforms, Anal. Biochem., 210, 389, 1993. 53. Klausen, N.K. and Kornfelt, T., Analysis of the glycoforms of human recombinant factor VIIa by capillary electrophoresis and high-performance liquid chromatography, J. Chromatogr. A, 718, 195, 1995. 54. Watson, E. and Yao, F., Capillary electrophoretic separation of recombinant granulocyte-colonystimulating factor glycoforms, J. Chromatogr., 630, 442, 1993. 55. Oda, R.P. and Landers, J.P., Effect of cationic buffer additives on the capillary electrophoretic separation of serum transferrin from different species, Electrophoresis, 17, 431, 1996. 56. Morbeck, D.E., Madden, B.J., and McCormick, D.J., Analysis of the microheterogeneity of the glycoprotein chorionic gonadotropin with high-performance capillary electrophoresis, J. Chromatogr. A, 680, 217, 1994. 57. Oda, R.P. et al., α, ω-Bis quaternary ammonium alkanes as effective buffer additives for enhanced capillary electrophoretic separation of glycoproteins, J. Chromatogr. A, 680, 85, 1994.

Glycoprotein Analysis by Capillary Electrophoresis

699

58. Verzola, B., Gelfi, C., and Righetti, P.G., Protein adsorption to the bare silica wall in capillary electrophoresis. Quantitative study on the chemical composition of the background electrolyte for minimizing the phenomenon, J. Chromatogr. A, 868, 85, 2000. 59. Wicktorowicz, J.E. and Colburn, J.E., Separation of cationic proteins via charge reversal in capillary electrophoresis, Electrophoresis, 11, 769, 1990. 60. Cordova, E., Gao, J., and Whitesides G.M., Noncovalent polycationic coating for capillaries in capillary electrophoresis of proteins, Anal. Chem., 69, 1370, 1997. 61. Kelly, J.F. et al., Development of electrophoretic conditions for the characterization of protein glycoforms by capillary electrophoresis-electrospray mass spectrometry, J. Chromatogr. A, 720, 409, 1996. 62. Yu, B. et al., Ionene-dynamically coated capillary for analysis of urinary and recombinant human erythropoietin by capillary electrophoresis and online electrospray ionization mass spectrometry, J. Sep. Sci., 28, 2390, 2005. 63. Katayama, H., Ishihama, Y., and Asakawa, N., Stable capillary coating with successive multiple ionic polymer layers, Anal. Chem., 70, 2254, 1998. 64. Bendahl, L., Hansen, S.H., and Gammelgaard, B., Capillaries modified by noncovalently anionic polymer adsorption for capillary zone electrophoresis, micellar electrokinetic capillary chromatography and capillary electrophoresis, Electrophoresis, 22, 2565, 2001. 65. Graul, T.W. and Schlenoff, J.B., Capillaries modified by polyelectrolyte multilayers for electrophoretic separations, Anal. Chem., 71, 4007, 1999. 66. Lanz, C., Marti, U., and Thormann, W., Capillary zone electrophoresis with a dynamic double coating for analysis of carbohydrate-deficient transferrin in human serum. Precision performance and pattern recognition, J. Chromatogr. A, 1013, 131, 2003. 67. Joneli, J., Lanz, C., and Thormann, W., Capillary zone electrophoresis determination of carbohydratedeficient transferrin using the new Ceofix reagents under high-resolution conditions, J. Chromatogr. A, 1130, 272, 2006. 68. Legros, F.J. et al., Carbohydrate-deficient transferrin isoforms measured by capillary zone electrophoresis for detection of alcohol abuse, Clin. Chem., 48, 2177, 2002. 69. Legros, F.J. et al., Use of capillary zone electrophoresis for differentiating excessive from moderate alcohol consumption, Clin. Chem., 49, 440, 2003. 70. Chang, W.W.P. et al., Rapid separation of protein isoforms by capillary zone electrophoresis with new dynamic coating, Electrophoresis, 26, 2179, 2005. 71. Balaguer, E. and Neususs, C., Glycoprotein characterization combining intact protein and glycan analysis by capillary electrophoresis-electrospray ionization-mass spectrometry, Anal. Chem., 78, 5384, 2006. 72. Hjerten, S., High-performance electrophoresis. Elimination of the electroendosmosis and solute adsorption, J. Chromatogr., 347, 191, 1985. 73. Schmalzing, D. et al., Characterization and performance of a neutral hydrophilic coating for the capillary electrophoretic separation of biopolymers, J. Chromatogr. A., 652, 149, 1993. 74. Hjerten, S. and Kubo, K., A new type of pH- and detergent stable coating for elimination of electroendosmosis and adsorption in (capillary) electrophoresis, Electrophoresis, 14, 390, 1993. 75. Gilges, M., Kleemiss, M.H., and Schomburg, G., Capillary zone electrophoresis separation of basic and acidic proteins using poly(vinyl alcohol) coatings in fused silica capillaries, Anal. Chem., 66, 2038, 1994. 76. Yim, K., Fractionation of the human recombinant tissue plasminogen activator (rtPA) glycoforms by high-performance capillary zone electrophoresis and capillary isoelectric focusing, J. Chromatogr., 559, 401, 1991. 77. Thorne, J.M. et al., Examination of capillary zone electrophoresis, capillary isoelectric focusing and sodium dodecyl sulphate capillary electrophoresis for the analysis of recombinant tissue plasminogen activator, J. Chromatogr. A, 744, 155, 1996. 78. Kinoshita, M. et al., Comparative studies on the analysis of glycosylation heterogeneity of sialic acid-containing glycoproteins using capillary electrophoresis, J. Chromatogr. A, 866, 261, 2000. 79. Kilar, F. and Hjerten, S., Separation of the human transferrin isoforms by carrier-free high-performance zone electrophoresis and isoelectric focusing, J. Chromatogr. A, 480, 351, 1989.

700

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

80. Hunt, G., Moorhouse, K.G., and Chen, A.B., Capillary isoelectric focusing and sodium dodecyl sulfatecapillary gel electrophoresis of recombinant humanized monoclonal antibody HER2, J. Chromatogr. A, 744, 295, 1996. 81. Grossman, P.D. et al., Application of free-solution capillary electrophoresis to the analytical scale separation of proteins and peptides, Anal. Chem., 61, 1186, 1989. 82. Taverna, M. et al., Analysis of carbohydrate-mediated heterogeneity and characterization of N-linked oligosaccharies of glycoproteins by high performance capillary electrophoresis, Electrophoresis, 13, 359, 1992. 83. Sanz-Nebot, V. et al., Capillary electrophoresis and matrix-assisted laser desorption/ionization-time of flight-mass spectrometry for analysis of the novel erythropoiesis-stimulating protein (NESP), Electrophoresis, 26, 1451, 2005. 84. Lacunza, I. et al., CZE of human alpha-1-acid glycoprotein for qualitative and quantitative comparison of samples from different pathological conditions, Electrophoresis, 27, 4205, 2006. 85. Honda, S. et al., Simultaneous determination of reducing monosaccharides by capillary zone electrophoresis as the borate complexes of N-2 pyridylglycamines, Anal. Biochem., 176, 72, 1989. 86. Hoffsteter-Kuhn, S. et al., Influence of borate complexation on the electrophoretic behavior of carbohydrates in capillary electrophoresis, Anal. Chem., 63, 1541, 1991. 87. Pacakova, V. et al., Effects of electrolyte modification and capillary coating on separation of glycoproteins isoforms by capillary electrophoresis, Electrophoresis, 22, 459, 2001. 88. Chen, Y., Critical conditions for separating the microheterogeneous components of glycoproteins by capillary electrophoresis, J. Chromatogr. A, 768, 39, 1997. 89. Oda, R.P. et al., Capillary electrophoresis-based separation of transferrin sialoforms in patients with carbohydrate-deficient glycoprotein syndrome, Electrophoresis, 18, 1819, 1997. 90. Lopez-Soto-Yarritu, P. et al., Comparison of different capillary electrophoresis methods for analysis of recombinant erythropoietin glycoforms, J. Sep. Sci., 25, 1112, 2002. 91. Tran, A.D. et al., Separation of carbohydrate-mediated microheterogeneity of recombinant human erythropoietin by free solution capillary electrophoresis, J. Chromatogr., 542, 459, 1991. 92. Zhou, G.-H. et al., Application of capillary electrophoresis, liquid chromatography, electrospray-mass spectrometry and matrix-assisted laser desorption/ionization-time of flight-mass spectrometry to the characterization of recombinant human erythropoietin, Electrophoresis, 19, 2348, 1998. 93. Nieto, O., Hernandez, P., and Hernandez, L., Capillary zone electrophoresis of human recombinant erythropoietin using C8 coated columns without additives in the running buffer, Anal. Commun., 33, 425, 1996. 94. Kaniansky, D. et al., Capillary zone electrophoresis in a hydrodynamically closed separation system with enhanced sample loadability, J. Chromatogr. A, 772, 103, 1997. 95. Madajova, V. et al., Fractionation of glycoforms of recombinant human erythropoietin by preparative capillary isotachophoresis, Electrophoresis, 26, 2664, 2005. 96. Balaguer, E. and Neususs, C., Intact glycoform characterization of erythropoietin-α and erythropoietin-β by CZE-ESI-TOF-MS, Chromatographia, 64, 351, 2006. 97. Kilar, F. and Hjerten, S., Fast and high resolution analysis of human serum transferrin by high performance isoelectric focusing in capillaries, Electrophoresis, 10, 23, 1989. 98. Righetti, P.G., Determination of the isoelectric point of proteins by capillary isoelectric focusing, J. Chromatogr. A, 1037, 491, 2004. 99. Lacunza, I. and de Frutos, M., Different capillary isoelectric focusing approaches for analysis of α-1-acid glycoprotein, PACE Setter, 10, 5, 2006. 100. Wu, J., Li, S.-C., and Watson, A., Optimizing separation conditions for proteins and peptides using imaged capillary isoelectric focusing, J. Chromatogr. A, 817, 163, 1998. 101. Moorhouse, K.G. et al., Rapid one-step capillary isoelectric focusing method to monitor charged glycoforms of recombinant human tissue-type plasminogen activator, J. Chromatogr. A, 717, 61, 1995. 102. Reif, O.-W. and Freitag, R., Control of the cultivation process of antithrombin III and its characterization by capillary electrophoresis, J. Chromatogr. A, 680, 383, 1994. 103. Tran, N.T. et al., One-step capillary isoelectric focusing for the separation of the recombinant human immunodeficiency virus envelope glycoprotein glycoforms, J. Chromatogr. A, 866, 121, 2000. 104. Kubach, J. and Grimm, R., Non-native capillary isoelectric focusing for the analysis of the microheterogeneity of glycoproteins, J. Chromatogr. A, 737, 281, 1996.

Glycoprotein Analysis by Capillary Electrophoresis

701

105. Moorhouse, K.G., Rickel, C.A., and Chen, A.B., Electrophoretic separation of recombinant tissuetype plasminogen activator glycoforms: Validation issues for capillary isoelectric focusing methods, Electrophoresis, 17, 423, 1996. 106. Chen, A.B. et al., Comparison of ampholytes used for slab gel and capillary isoelectric focusing of recombinant tissue-type plasminogen activator glycoforms, J. Chromatogr. A, 744, 279, 1996. 107. Sebastiano, R. et al., Mass distribution and focusing properties of carrier ampholytes for isoelectric focusing: I. Novel and unexpected results, Electrophoresis, 27, 3919, 2006. 108. Simo, C. et al., Mass distribution, polydispersity and focusing properties of carrier ampholytes for IEF II: pH 4-6 intervals, Electrophoresis, 27, 4849, 2006. 109. Lacunza, I. et al., CIEF with hydrodynamic and chemical mobilization for the separation of forms of alpha-1-acid glycoprotein, Electrophoresis, 28, 1204, 2007. 110. Wehr, T., Rodriguez-Diaz, R., and Zhu, M., Recent advances in capillary isoelectric focusing, Chromatographia Suppl., 53, S45, 2001. 111. Righetti, P.G., Isoelectric Focusing: Theory, Methodology, and Applications, Elsevier Biomedical, Amsterdam, 1989, 1. 112. Schwer, C., Capillary isoelectric focusing: A routine method for protein analysis?, Electrophoresis, 16, 2121, 1995. 113. Rudd, P.M. et al., Separation and analysis of the glycoform populations of ribonuclease B using capillary electrophoresis, Glycoconj. J., 9, 86, 1992. 114. James, D.C. et al., High-resolution separation of recombinant human interferon-γ glycoforms by micellar electrokinetic capillary chromatography, Anal. Biochem., 222, 315, 1994. 115. Werner, W.E. et al., Size-dependent separation of proteins denatured in SDS by capillary electrophoresis using a replaceable sieving matrix, Anal. Biochem., 212, 253, 1993. 116. Buchacher, A. et al., High-performance capillary electrophoresis for in-process control in the production of antithrombin III and human clotting factor IX, J. Chromatogr. A, 802, 355, 1998. 117. Bergstrom, M. et al., Lectin affinity capillary electrophoresis in glycoform analysis applying the partial filling technique, J. Chromatogr. B, 809, 323, 2004. 118. Uegaki, K. et al., Simultaneous estimation of the association constants of glycoprotein glycoforms to a common protein by capillary electrophoresis, Anal. Biochem., 309, 269, 2002. 119. Thormann, W. and Mosher, R.A., High-resolution computer simulation of the dynamics of isoelectric focusing using carrier ampholytes: Focusing with concurrent electrophoretic mobilation is an isotachophoretic process, Electrophoresis, 27, 968, 2006. 120. Iourin, O. et al., The identification of abnormal glycoforms of serum transferrin in carbohydrate deficient glycoprotein syndrome type I by capillary zone electrophoresis, Glycoconj. J., 13, 1031, 1996. 121. Thibault, P. and Dovichi, N., General instrumentation and detection systems including mass spectrometry, in Capillary Electrophoresis. Theory and Practice, Camilleri, P., Ed., CRC Press, Boca Raton, FL, 1998, chap. 2. 122. http://probes.invitrogen.com/handbook/sections/0900.html Accessed January 2007. 123. Craig, D.B. and Dovichi, N.J., Multiple labeling of proteins, Anal. Chem., 70, 2493, 1998. 124. Tseng, W.-L., Lin, Y.-W., and Chang, H.-T., Improved separation of microheterogeinities and isoforms of proteins by capillary electrophoresis using segmental filling with SDS and PEO in the background electrolyte, Anal. Chem., 74, 4828, 2002. 125. Salas-Solano, O. et al., Optimization and validation of a quantitative capillary electrophoresis sodium dodecyl sulfate method for quality control and stability monitoring of monoclonal antibodies, Anal.Chem.,78, 6583, 2006. 126. Pinto, D.M. et al., Picomolar assay of native proteins by capillary electrophoresis precolumn. Submicellar separation, and laser-induced fluorescence detection, Anal. Chem., 69, 3015, 1997. 127. Lee, I.H. et al., Picomolar analysis of proteins using electrophoretically mediated microanalysis and capillary electrophoresis with laser-induced fluorescence detection, Anal. Chem., 70, 4546, 1998. 128. Wehr, T., Rodriguez-Diaz, R., and Zhu, M., Capillary Electrophoresis of Proteins, Marcell Dekker, New York, 1999, chap. 3.

702

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

129. Bornemann, C. et al., Fluorescence-labelled antigen-binding fragments (Fab) from monoclonal antibody 5F12 detect human erythropoietin in immunoaffinity capillary electrophoresis, Anal. Bioanal. Chem., 376, 1074, 2003. 130. Kelly, J.A. and Lee, C.S., Online postcapillary affinity detection of immunoglobulin G subclass and monoclonal antibodies variants for capillary electrophoresis, J. Chromatogr. A, 790, 207, 1997. 131. Yeung, B., Porter, T.J., and Vath, J.E., Direct isoform analysis of high-mannose-containing glycoproteins by online capillary electrophoresis electrospray mass spectrometry, Anal. Chem., 69, 2510, 1997. 132. Demelbauer, U.M. et al., Characterization of glyco isoforms in plasma-derived human antithrombin by online capillary zone electrophoresis-electrospray ionization-quadrupole ion trap-mass spectrometry of the intact glycoproteins, Electrophoresis, 25, 2026, 2004. 133. Arizaga-Rodriguez, S. et al., Detection of transferrin isoforms in human serum: Comparison of UV and ICP-MS detection after CZE and HPLC separations, Anal. Bioanal. Chem., 383, 390, 2005. 134. Hooker, A.D. and James, D.C., Analysis of glycoprotein heterogeneity by capillary electrophoresis and mass spectrometry, Mol. Biotech., 14, 241, 2000. 135. Morelle, W. and Michalski, J.C., The mass spectrometric analysis of glycoproteins and their glycan structures, Curr. Anal. Chem., 1, 29, 2005. 136. Bateman, K.P. et al., Glycoprotein analysis by capillary zone electrophoresis–electrospray mass spectrometry, Meth. Mol. Biol., 213, 219, 2003. 137. Monton, M.R.N. and Terabe, S., Recent developments in capillary electrophoresis—Mass spectrometry of proteins and peptides, Anal. Sci., 21, 5, 2005. 138. Servais, A.-C., Crommen, J., and Fillet, M., Capillary electrophoresis-mass spectrometry, an attractive tool for drug bioanalysis and biomarker discovery, Electrophoresis, 27, 2616, 2006. 139. ICH Guidelines, http://www.ich.org/cache/compo/276-254-1.html, accessed November 2006. 140. Somatropin, 07/2007:0951, European Pharmacopoeia 5th edition, published by EDQM, April 2007. 141. Somerville, L.E., Douglas, A.J., and Irvine, A.E., Discrimination of granulocyte colony-stimulating factor isoforms by high-performance capillary electrophoresis, J. Chromatogr. B, 732, 81, 1999. 142. Berkowitz, S.A. et al., Rapid quantitative capillary zone electrophoresis method for monitoring the micro-heterogeneity of an intact recombinant glycoprotein, J. Chromatogr. A, 1079, 254, 2005. 143. Felten, C. et al., Use of acidic and basic pH and calcium ion addition in the capillary zone electrophoretic characterization of recombinant human deoxyribonuclease, a complex phosphoglycoprotein, J. Chromatogr. A, 853, 295, 1999. 144. Quan, C.P., Canova-Davis, E., and Chen, A.B., Effects of the solution environment on the resolution of recombinant human deoxyribonuclease variants in capillary zone electrophoresis, Chromatographia Suppl., 53, S39, 2001. 145. Kremser, L. et al., Characterization of antithrombin III from human plasma by two-dimensional gel electrophoresis and capillary electrophoretic methods. Electrophoresis, 24, 4282, 2003. 146. Che, F.Y. et al., Comparative study on the distribution of ovalbumin glycoforms by capillary electrophoresis, J. Chromatogr. A, 849, 599, 1999. 147. Kubo, K. and Hattori, A., Assessment of the capillary zone electrophoretic behaviour of proteins in the presence of electroosmotic modifiers: Protein-polyamine interaction studied using a polyacrylamide-coated capillary, Electrophoresis, 22, 3389, 2001. 148. Catai, J.R. et al., Noncovalently bilayer-coated capillaries for efficient and reproducible analysis of proteins by capillary electrophoresis, J. Chromatogr. A, 1083, 185, 2005. 149. Sandra, K. et al., Combining gel and capillary electrophoresis, nano-LC and mass spectrometry for the elucidation of posttranslational modifications of Trichoderma reesei cellobiohydrolase I., J. Chromatogr. A, 1058, 263, 2004. 150. Krull, I.S. et al., HPCE methods for the identification and quantitation of antibodies, their conjugates and complexes, J. Pharm. Biomed. Anal., 16, 377, 1997. 151. Santora, L.C., Krull, I.S., and Grant, K., Characterization of recombinant human monoclonal tissue necrosis factor-α antibody using cation-exchange HPLC and capillary isoelectric focusing, Anal. Biochem., 275, 98, 1999. 152. Harmon, B.J., Imaged capillary isoelectric focusing for the characterization and monitoring of charge heterogeneity of recombinant monoclonal antibodies, in Proceedings of the 5th Symposium on CE in the Biotechnology and Pharmaceutical Industries, San Francisco, CA, 2003.

Glycoprotein Analysis by Capillary Electrophoresis

703

153. Hoffstetter-Kuhn, S., Alt, G., and Kuhn, R. Profiling of oligosaccharide-mediated microheterogeneity of a monoclonal antibody by capillary electrophoresis, Electrophoresis, 17, 418, 1996. 154. Ma, S. in State of the Art Analytical Methods for the Characterization of Biological Products and Assessment of Comparability, Mire-Sluis, A.R., Ed., Karger, Basel, 2005, pp. 49–68. 155. Nunnally, B. et al., Aseries of collaborations between various pharmaceutical companies and regulatory authorities concerning the analysis of biomolecules using capillary electrophoresis, Chromatographia, 64, 359, 2006. 156. Jenkins, N., Parekh, R.B., and James, D.C., Getting the glycosylation right—Implications for the biotechnology industry, Nat. Biotechnol., 14, 975, 1996. 157. Chirino, A.J. and Mire-Sluis, A., Characterizing biological products and assessing comparability following manufacturing changes, Nat. Biotechnol., 22, 1383, 2004. 158. James, D.C. et al., Posttranslational processing of recombinant human interferon-γ in animal expression systems, Prot. Sci., 5, 331, 1996. 159. Goldman, M.H. et al., Monitoring recombinant human interferon-gamma N-glycosylation during perfused fluidized-bed and stirred-tank batch culture of CHO cells, Biotech. Bioeng., 60, 596, 1998. 160. Wang, M.D. et al., Erythropoietin production from CHO cells grown by continuous culture in a fluidized-bed bioreactor, Biotech. Bioeng., 77, 194, 2002. 161. Freitag, R., et al., Production of recombinant h-AT III with mammalian cell cultures using capillary electrophoresis for product monitoring, Cytotechnology, 21, 205, 1996. 162. Klyushnichenko, V., Capillary electrophoresis in the analysis and monitoring of biotechnological processes, Meth. Mol. Biol., 276, 77, 2004. 163. Schmelzer, A.E. and Miller, W.M., Hyperosmotic stress and elevated pCO2 alter monoclonal antibody charge distribution and monosaccharide content, Biotechnol. Prog., 18, 346, 2002. 164. Cherkaoui, S. et al., Development of a capillary zone electrophoresis method for caseinoglycomacropeptide determination, J. Chromatogr. A, 790, 195, 1997. 165. Cherkaoui, S. et al., Determination of the stability of caseinoglycomacropeptide in a cosmetic lotion by use of capillary zone electrophoresis with a coated capillary, Chromatographia, 50, 311, 1999. 166. Daali, Y., Cherkaoui, S., and Veuthey, J.L., Capillary electrophoresis and high-performance anion exchange chromatography for monitoring caseinoglycomacropeptide sialylation, J. Pharm. Biomed. Anal., 24, 849, 2001. 167. Donohue, M.J. et al., Capillary electrophoresis for the investigation of prostate-specific antigen heterogeneity, Anal. Biochem., 339, 318, 2005. 168. Girard, M., Boucher, S., and Kane, A., Analysis of erythropoietin preparations obtained from various sources worldwide by capillary zone electrophoresis, in Proceedings of the 7th International symposium on capillary electrophoresis in the biotechnology and pharmaceutical industries, Montreal, 2005. 169. Deechongkit, S. et al., Biophysical comparability of the same protein from different manufacturers: a case study using epoetin alfa from Epogen® and Eprex®, J. Pharm. Sci., 95, 1931, 2006. 170. Zhou, G.-H. et al., Characterization of recombinant human granulocyte colony stimulating factor (rHuG-CSF) by capillary zone electrophoresis, capillary isoelectric focusing electrophoresis and electrospray ionization mass spectrometry, J. Pharm. Biomed. Anal., 35, 425, 2004. 171. Storring, P.L. et al., Physicochemical Methods for predicting the biological potency of recombinant follicle stimulating hormone: An international collaborative study of isoelectric focusing and capillary zone electrophoresis, Biologicals, 30, 217, 2002. 172. Mulders, J.W.M., Wijn, H., and Theunissen, F. Development and validation of a capillary zone electrophoresis method which can be used to predict the in-vivo bioactivity of recombinant follicle stimulating hormone, in Proceedings of the 3rd Symposium on the Analysis of Well Characterized Biotechnology Pharmaceuticals (WCBP), Washington, DC, 1999. 173. Kim, Y.J. and Varki, A., Perspectives on the significance of altered glycosylation of glycoproteins in cancer, Glycoconj. J., 14, 569, 1997. 174. de Jong, G. and van Eijk, H.G., Microheterogeneity of human serum transferrin: A biological phenomenon studied by isoelectric focusing in immobilized pH gradients, Electrophoresis, 9, 589, 1998. 175. de Jong, G., van Dijk, J.P., and van Eijk, H.G., The biology of transferrin, Clin. Chim. Acta, 190, 1, 1990.

704

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

176. Arndt, T., Carbohydrate-deficient transferrin as a marker of chronic alcohol abuse: A critical review of preanalysis, analysis, and interpretation, Clin. Chem., 47, 13, 2001. 177. Wuyts, B. et al., Carbohydrate-deficient transferrin and chronic alcohol ingestion in subjects with transferrin CD-variants, Clin. Chem. Lab. Med., 39, 937, 2001. 178. Stibler, H., Carbohydrate-deficient transferrin in serum: A new marker of potentially harmful alcohol consumption reviewed, Clin. Chem., 37, 2029, 1991. 179. Keir, G., Winchester, B.G., and Clayton, P., Carbohydrate-deficient glycoprotein syndromes: inborn errors of protein glycosylation, Ann. Clin. Biochem., 36, 20, 1999. 180. Whitfield, J.B., Transferrin isoform analysis for the diagnosis and management of hazardous or dependent drinking, Clin. Chem., 48, 2095, 2002. 181. Stibler, H. and Kjellin, KG., Isoelectric-focusing and electrophoresis of CSF proteins in tremor of different origins—Study of 38 cases with cerebellar syndrome of chronic-alcoholism, Parkinson’s-disease and essential tremor, J. Neurol. Sci., 30, 269, 1976. 182. Stibler, H. et al., Abnormal microheterogeneity of transferrin in serum and cerebrospinal-fluid in alcoholism, Acta Med. Scand., 204, 49, 1978. 183. Stibler, H., Borg, S., and Allgulander, C., Clinical significance of abnormal heterogeneity of transferrin in relation to alcohol-consumption, Acta Med. Scand., 206, 275, 1979. 184. Wuyts, B. and Delanghe, J.R., The analysis of carbohydrate-deficient transferrin, marker of chronic alcoholism, using capillary electrophoresis, Clin. Chem. Lab. Med., 41, 739, 2003. 185. Lanz, C. et al., Evaluation and optimization of capillary zone electrophoresis with different dynamic capillary coatings for the determination of carbohydrate-deficient transferrin in human serum, J. Chromatogr. A, 979, 43, 2002. 186. Helander, A. et al., Interference of transferrin isoform types with carbohydrate-deficient transferrin quantification in the identification of alcohol abuse, Clin. Chem., 47, 1225, 2001. 187. Bortolotti, F. et al., Analysis of carbohydrate-deficient transferrin: Comparative evaluation of turbidimetric immunoassay, capillary zone electrophoresis, and HPLC, Clin. Chem., 51, 2368, 2005. 188. Jeppsson, J.-O., Kristensson, H., and Fimiani, C., Carbohydrate-deficient transferrin quantified by HPLC to determine heavy consumption of alcohol, Clin. Chem., 39, 2115, 1993. 189. Turpeinen, U. et al., Comparison of HPLC and small column (CDTect) methods for disialotransferrin, Clin. Chem., 47, 1782, 2001. 190. Helander, A., Husa, A., and Jeppsson, J.-O., Improved HPLC method for carbohydrate-deficient transferrin in serum, Clin. Chem., 49, 1881, 2003. 191. Prasad, R. et al., Analysis of carbohydrate deficient transferrin by capillary zone electrophoresis, Electrophoresis, 18, 1814, 1997. 192. Tagliaro, F. et al., Optimized determination of carbohydrate-deficient transferrin isoforms in serum by capillary zone electrophoresis, Electrophoresis, 19, 3033, 1998. 193. Crivellente, F. et al., Improved method for carbohydrate-deficient transferrin determination in human serum by capillary zone electrophoresis, J. Chromatogr. B, 739, 81, 2000. 194. Trout, A.L. et al., Direct capillary electrophoretic detection of carbohydrate-deficient transferrin in neat serum, Electrophoresis, 21, 2376, 2000. 195. Ramdani, B. et al., Analyte comigrating with trisialotransferrin during capillary zone electrophoresis of sera from patients with cancer, Clin. Chem., 49, 1854, 2003. 196. Sanz-Nebot, V. et al., Characterization of human transferrin glycoforms by capillary electrophoresis and electrospray ionization mass spectrometry, J. Chromatogr. B, 798, 1, 2003. 197. Lanz, C. et al., Improved capillary electrophoresis method for the determination of carbohydratedeficient transferrin in patient sera, Electrophoresis, 25, 2309, 2004. 198. Martello, S. et al., Determination of carbohydrate deficient transferrin (CDT) with capillary electrophoresis: an inter laboratory comparison, Forensic Sci. Internat., 141, 153, 2004. 199. Carchon, H.A. et al., Diagnosis of congenital disorders of glycosylation by capillary zone electrophoresis of serum transferrin, Clin. Chem., 50, 101, 2004. 200. Tagliaro, F. et al., Carbohydrate-deficient transferrin determination revisited with capillary electrophoresis: A new biochemical marker of chronic alcohol abuse, J. Cap. Electrophor. Microchip Tech., 5, 137, 1999. 201. Fermo, I. et al., Capillary zone electrophoresis for determination of carbohydrate-deficient transferrin in huan serum, Electrophoresis, 25, 469, 2004.

Glycoprotein Analysis by Capillary Electrophoresis

705

202. Giordano, B.C. et al., Dynamically-coated capillaries allow for capillary electrophoretic resolution of transferrin sialoforms via direct analysis of human serum, J. Chromatogr. B, 742, 79, 2000. 203. Wuyts, B. et al., Determination of carbohydrate-deficient transferrin using capillary zone electrophoresis, Clin. Chem., 47, 247, 2001. 204. Tagliaro, F. et al., Caveats in carbohydrate-deficient transferrin determination, Clin. Chem., 48, 208, 2001. 205. Lanz, C. and Thormann, W., Capillary zone electrophoresis with a dynamic double coating for analysis of carbohydrate-deficient transferrin in human serum: Impact of resolution between disialo- and trisialotransferrin on reference limits, Electrophoresis, 24, 4272, 2003. 206. Vitala, K., Lähdesmäki, K., and Niemela, O., Comparison of the Axis % CDT TIA and the CDTect method as laboratory tests of alcohol abuse, Clin. Chem., 44, 1209, 1998. 207. Delanghe, J.R., Wuyts, B., and De Buyzere, M.L., Caveats in carbohydrate-deficient transferring determination–Response, Clin. Chem., 48, 209, 2002. 208. Marquardt, T. and Dencke, J., Congenital disorders of glycosylation: Review of their molecular bases, clinical presentations and specific therapies, Eur. J. Pediatr., 162, 359, 2003. 209. Fournier, T., Medjoubi-N, N., and Porquet, D., Alpha-1-acid glycoprotein, Biochim. Biophys. Acta, 1482, 157, 2000. 210. Duche, J.-C. et al., Expression of the genetic variants of human alpha-1-acid glycoprotein in cancer, Clin. Chem., 33, 197, 2000. 211. Mackiewicz, A. and Mackiewicz, K., Glycoforms of serum α1-acid glycoprotein as markers of inflammation and cancer, Glycoconj. J., 12, 241, 1995. 212. Hashimoto, S. et al., α1 -Glycoprotein fucosylation as a marker of carcinoma progression and prognosis, Cancer, 101, 2825, 2004. 213. Kremmer, T. et al., Liquid chromatographic and mass spectrometric analysis of human serum acid alpha-1-glycoprotein, Biomed. Chromatogr., 18, 323, 2004. 214. Hrycaj, P. et al., Micoheterogeneity of alpha 1 acid glycoprotein in rheumatoid arthritis: Dependent on disease duration?, Ann. Rheum. Dis., 52, 138, 1993. 215. Elliott, M.A. et al., Investigation into the Concavalin A reactivity, fucosylation and oligosaccharide microheterogeneity of α1 -acid glycoprotein expressed in the sera of patients with rheumatoid arthritis, J. Chromatogr. B, 688, 229, 1997. 216. Iijima, S. et al., Changes of α1 -glycoprotein microheterogeneity in acute inflammation stages analyzed by isoelectric focusing using serum obtained postoperatively, Electrophoresis, 21, 753, 2000. 217. Higai, K. et al., Glycosylation of site-specific glycans of α1 -acid glycoprotein and alterations in acute ad chronic inflammation, Biochim. Biophys. Acta, 1725, 128, 2005. 218. Higai, K. et al., Altered glycosylation of α1 -acid glycoprotein in patients with inflammation and diabetes mellitus, Clin. Chim. Acta, 329, 117, 2003. 219. Kakehi, K. et al., Capillary electrophoresis of sialic acid-containing glycoprotein. Effect of the heterogeneity of carbohydrate chains on glycoform separation using an α1 -acid glycoprotein as a model, Anal. Chem., 73, 2640, 2001. 220. Sei, K. et al., Collection of α1 -acid glycoprotein molecular species by capillary electrophoresis and the analysis of their molecular masses and carbohydrate chains. Basic studies on the analysis of glycoprotein glycoforms, J. Chromatogr. A, 958, 273, 2002. 221. Hiroaka, A. et al., Charge microheterogeneity of the β-trace proteins (lipocalin-type prostaglandin D synthase) in the cerebrospinal fluid of patients with neurological disorders analyzed by capillary isoelectric focusing, Electrophoresis, 22, 3433, 2001. 222. Lilja, H., Biology of prostate-specific antigen, Urology, 62, 27, 2003. 223. Peracaula, R. et al., Altered glycosylation pattern allows the distinction between prostate-specific antigen (PSA) from normal and tumor origins, Glycobiology, 13, 457, 2003. 224. Guzman, N.A., Immunoaffinity capillary electrophoresis applications of clinical and pharmaceutical relevance, Anal. Bioanal. Chem., 378, 37, 2004. 225. Mao, X., Chu, I.K., and Lin B., A sheath-flow nanoelectrospray interface of microchip electrophoresis MS for glycoprotein and glycopeptide analysis, Electrophoresis, 27, 5059, 2006.

Electrophoresis of 23 Capillary Post-Translationally Modified Proteins and Peptides Bettina Sarg and Herbert H. Lindner CONTENTS 23.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2 Theoretical Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.1 Proteins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.1.1 Phosphorylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.1.2 Acetylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.1.3 Deamidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.1.4 Methylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.2 Peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.2.1 Phosphorylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.2.2 Deamidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.2.3 Farnesylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.4 Method Development Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.4.1 Buffers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.4.2 Capillary Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.4.3 Method Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

707 708 711 711 711 713 714 715 716 716 717 717 717 718 718 718 718 719 719

23.1 INTRODUCTION A large number of chemical modifications can occur in individual amino acids that can fundamentally affect their physicochemical and functional properties. Over 200 distinct covalent modifications have been reported, with phosphorylation, glycosylation, acetylation, methylation, and ADP-ribosylation being the most common. Some amino acids can be converted into other amino acids, for example, asparagine in aspartic acid or glutamine in glutamic acid by deamidation. Knowledge of these modifications is extremely important because they may alter physical and chemical properties, folding, conformation distribution, stability, activity, and consequently, function of the proteins. Phosphorylation, principally on serine, threonine, or tyrosine residues, is one of the most important and abundant post-translational modifications (PTMs), with more than 30% of proteins being modified by the covalent attachment of one or more phosphate groups. It plays a critical role in the regulation of various cellular processes including cell cycle, growth, apoptosis, and transmitting extracellular signals to the nucleus. In fact, protein phosphorylation is 707

708

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

probably the single most common intracellular signal transduction event. Owing to its biochemical importance, various analytical techniques for the detection and analysis of protein phosphorylation have been described.1 A widely used approach for the detection of phosphorylation involves metabolic labeling with [32 P]orthophosphate followed by one- or two-dimensional polyacrylamide gel electrophoresis. Detection of phosphorylation sites is based on radiography combined with enzymatic digestions of the separated proteins and two-dimensional phosphopeptide mapping. Highperformance capillary electrophoresis (HPCE) offers substantial method inherent advantages over such protocols, for example, a more rapid and accurate analysis with less sample consumption, while avoiding hazardous radioactive labeling. In vivo methylation of the side chains of specific arginines, histidines, and lysines in proteins is a common phenomenon in nature involving numerous classes of proteins in both prokaryotic and eukaryotic cells.2,3 Methylation has been most well studied in histones, with distinct lysine residues mono-, di-, or tri-methylated playing a major role in the regulation of gene expression, DNA replication, and repair. Methylated amino acids have often been determined in protein and tissue hydrolysates using amino acids analyzers and through cells radiolabeled with [methyl-3 H]methionine. Acetylation of proteins occurs in two different ways. The N-terminal acetylation is one of the most common protein modification reactions in eukaryotes and it is estimated that up to about 85% of all mammalian proteins are affected. It is an irreversible process occurring cotranslationally, unlike the reversible side chain acetylation of internal lysine residues, most famously for histones and transcription factors that affect selective gene transcription and chromatin structure. Despite many hypotheses about the role of N-terminal acetylation, its biological significance is still unclear. Protein glycosylation is a PTM of eukaryotic proteins and does not occur in prokaryotes. The carbohydrate groups are highly variable with significant effects on protein folding, stability, and activity. The importance of HPCE for the analysis of glycosylated proteins is described in Chapter 22. Nonenzymatic deamidation of peptides or proteins represents an important degradation reaction occurring in vitro in the course of isolation or storage and in vivo during development and/or aging of cells.4,5 Deamidation is a hydrolytic reaction resulting not only in the introduction of negative charges but also in a change in the primary structure of proteins or peptides. Deamidation is a common PTM resulting in the conversion of an asparagine residue to a mixture of isoaspartate and aspartate. Deamidation of glutamine residues can occur but does so at a much lower rate. Detecting and separating deamidated forms from the parent molecules are still problematic aspects of protein analysis. HPCE continues to become more widely used for the detection, separation, and quantification of peptides and proteins, since the use of capillaries greatly reduces sample volume and analysis time compared to conventional gel electrophoresis. This offers perfect qualification for the analysis of PTMs, but surprisingly only a small number of HPCE applications have been developed during the past years. Since these modifications represent phenomena that occur in vivo often in very small amounts, they can be easily missed. This may explain why in certain cases recombinantly expressed proteins have an altered or absent activity compared with the naturally expressed proteins. Taking this into account, more sensitive methods of sample detection have been developed, for example, special detection cell constructions for ultraviolet (UV) adsorption and laser-induced fluorescence (LIF), and particularly the introduction of mass spectrometry (MS) brought tremendous progress in online and offline characterization not only of modified peptides but also of modified proteins. In this chapter, we try to provide information about capillary electrophoresis (CE) methods developed for the separation of proteins and peptides with various PTMs.

23.2 THEORETICAL ASPECTS More or less all modification reactions occurring on proteins may alter their overall charge either by introducing or eliminating charges, thereby changing to some extent the m/z ratio and the pI of

Capillary Electrophoresis of Post-Translationally Modified Proteins and Peptides

709

the parent proteins. For this reason, capillary zone electrophoresis (CZE) and isoelectric focusing (IEF) are CE modes that are especially well suited to separate post-translationally modified forms. However, the ease of use makes CZE the preferentially used method in the laboratory and only very few applications employing other CE modes can be found in the literature. Depending on the pI of the proteins and the pH of the buffers used for the separation in CZE proteins are either positively or negatively charged (except when pI = pH, where the overall charge of the proteins is zero). In the course of phosphorylation, for example, negatively charged phosphate groups are bound to uncharged amino acids such as Ser, Thr, or Tyr. Also, glycosylation has an impact on protein charge if one or more charged sialic acid molecules are attached (detailed in Chapter 22). In both cases, modification causes either a decrease in the overall positive charge or at appropriate buffer pH conditions an increase in the negative charge of the modified forms. Acetylation can take place either on the N-terminal amino group of the first amino acid of a protein or at the ε-amino group of lysines. As basic amino groups, forming cationic ammonium ions under usual buffer conditions, become neutral due to amide formation one positive charge for each acetyl group bound will be removed in course of this modification reaction. This charge difference enables the separation of distinctly acetylated proteins from each other and from their unacetylated form. Another modification of lysine, its mono-, di-, and, tri-methylation, has been the focus of great attention in histones due to its biological importance in gene regulation. As expected, the impact of methylation on charge is not as pronounced as of acetylation; however, electron donor effects of the alkyl group slightly increase the charge density of the nitrogen of the amino group. These differences can be sufficiently high for a CE separation provided the molecular mass of the protein is low. An essential and in terms of its biological importance as well as of its frequency often underestimated modification reaction of proteins and peptides occurring under both physiological and laboratory conditions involves the nonenzymatic spontaneous deamidation of asparagine and glutamine residues to aspartate and glutamate, respectively. The deamidation of Asn, which takes place much more frequently than that of Gln, follows a rather complex mechanism via a cyclic imide intermediate. End products of its hydrolytic cleavage are the formation of isoaspartic acid as main reaction product and aspartic acid. In certain cases also truncation of the protein backbone can occur.4 Under buffer conditions where proteins are cations (pH < pI) the presence of both isoaspartic and aspartic acid reduces the overall positive charge of the proteins compared to the Asn containing protein. Moreover, due to minor differences in the pKa values of iso-Asp and Asp (pKa isoAsp < pKa Asp) a separation of these two isomeric protein forms can be achieved if pH of the separation buffer is adjusted thoroughly. Certain protein classes are known to be multiply modified. Histone proteins are most probably the best investigated representatives in this respect. Under certain biological conditions, they are known to be acetylated, phosphorylated, and methylated, for example, in a definite manner at the same time. Several examples of successful CE separations of even multiply modified proteins are reported in the literature.6,7 Problems in CE occurring with the analysis of PTMs are usually not related to the presence of the modifying residues. They are much more associated with problems generally occurring when proteins are analyzed by capillary electrophoresis. Proteins stick to many different surfaces including metals, plastics, and also glass, which are the most commonly used materials for capillaries in CE. Depending on the extent of interaction between proteins and the glass surface of the capillary peak tailing, loss of resolution, reduced sensitivity, or even total adsorption can occur. The primary causes for these detrimental effects are ionic interactions between cationic protein regions and the negatively charged silanol groups of the capillary surface. Moreover, also hydrophobic interactions may contribute to these adsorption phenomena. A variety of strategies were developed in the past to overcome these troublesome effects. Among these, a few are very simple to accomplish, for example, avoiding very low buffer concentrations. Increasing the buffer concentration decreases protein–wall interactions by reducing the effective

710

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

H1.5

0.030

A 200

0.020

0.010

0.000 20

22

24

26

28

30

Time (min)

FIGURE 23.1 CZE separation of phosphorylated histone H1.5 from human tumor cells (CCRF-CEM) in a 100 mM phosphate buffer (pH 2.0). Other conditions: voltage 12 kV, temperature 25◦ C, injection 2 s, detection 200 nm, untreated capillary (50 cm × 75 µm i.d.).

surface charge and, depending on buffer type also ion-pairing effects come into play. Another approach to limit adsorption is simply by working at pH extremes: At low pH (∼ 2) silanol groups are protonated, their dissociation is significantly reduced and in this way the interaction of the under these conditions cationic proteins with the now more or less uncharged surface minimized. Conversely, at high pH (∼10) both the silica surface and the proteins will be deprotonated and negatively charged. In this case, adsorption is remarkably reduced as a result of a charge repulsion effect. However, while these approaches may be very successful strategies for the separation of modified forms of peptides and some small proteins ( 10) and also hydrophobic properties, histone proteins particularly strongly interfere with the inner surface of the silica wall. From Figure 23.1, it is evident that even at this low pH essential electrostatic protein– wall interactions still occur, which are responsible for the poor resolution, broad peaks, and low sensitivity. In contrast to separations performed at low pH, working at high pH is not always feasible for the following reasons: 1. Artificial protein modification reactions can be induced, for example, deamidation of proteins. 2. At very high pH (>11) dissolution of the silica becomes an issue. For this reason, distinct coating procedures, often combined with low pH buffers, have been found to be a prerequisite for the successful separation of proteins as well as of their PTMs. Two basic strategies have been applied in CE to limit protein adsorption. 1. Permanent modification by covalently bonded or physically adhered phases. Basically, hydrophilic polymers, polar functional groups, or even positively charged residues, for

Capillary Electrophoresis of Post-Translationally Modified Proteins and Peptides

711

example, amino groups are bound to the silica surface.8 In general, life time of the permanent coating and batch to batch reproducibility may still be a problem. Therefore, it is not so surprising that most applications for the separation of protein modifications, which can be found in the literature, are based on an alternative approach to suppress undesired protein–surface interactions. 2. It is the concept of dynamically coating the inner wall of the capillary by the addition of suitable additives to the running buffer. These additives themselves show high affinity to the silica surface and act like a shield to prevent positively charged proteins from coming into close contact with the capillary wall. Various compounds have been used for this purpose; for instance, neutral and cationic hydrophilic polymers, cationic hydrophobic polymers, diamines, and so on.8 Outstanding advantages of this approach are stability of the coating and simplicity of handling. Since the coating agent is in the buffer, the coating layer of the capillary surface is continuously regenerated and no permanent stability is required. Moreover, cleaning of the capillary can be performed easily by rinsing with 0.1 M sodium hydroxide and water. However, potential disadvantages of the dynamic coating approach are effects of the surfactants, for example, influencing protein structure, incompatibility with MS analysis, range of pH applicable can be limited, time for equilibrium needed to obtain reproducible surface coating, surface properties may influence quality and reproducibility of the coating. As a consequence, both methods have advantages and shortcomings and no method is clearly superior. For this reason, depending on protein primary structure and type of modification method development is still an issue; however, many promising recipes for their successful separation are available in the meantime.

23.3 APPLICATIONS Many proteins such as histones, ribonucleic acid (RNA) polymerase II, tubulin, myelin basic protein, p53, and tyrosine kinases are post-translationally modified at multiple sites. Among them histones are the most extensively studied group as their hydrophilic N- and C-terminal tail domains are subjected to a great variety of PTMs including phosphorylation, acetylation, methylation, deamidation, ubiquitination, and ADP-ribosylation. Distinct combinations of covalent histone modifications including lysine acetylation, lysine and arginine methylation, and serine phosphorylation form the basis of the histone code hypothesis.9–11 This hypothesis proposes that a pre-existing modification affects subsequent modifications on histone tails and that these modifications generate unique surfaces for the binding of various proteins or protein complexes responsible for higher-order chromatin organization and gene activation and inactivation. Owing to their biological importance, it is not surprising that a variety of methods were developed particularly for the CE separation of histone modifications, more than for any other protein family. This chapter summarizes practical applications of HPCE on the analysis of various modified histones, other proteins and, moreover, also some modified peptide separations are described.

23.3.1 PROTEINS 23.3.1.1 Phosphorylation A set of different microscale techniques of CE exists for analyzing phosphorylated proteins and peptides. The effectiveness of a buffer system containing hydroxypropylmethyl cellulose (HPMC) as dynamic coating agent in combination with low pH preventing undesired interactions of positively charged histone proteins with the silica surface could be first demonstrated by Lindner et al.12–14 A complex mixture of rat testis H1 histones consisting of eight microsequence variants and, in addition,

712

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

H1.5

0.040

p2 p1

A200

0.030

0.020

p0 p3

p4

p5

0.010

0.000 16

17

18

19

Time (min)

FIGURE 23.2 CE separation of phosphorylated histone H1.5 from human tumor cells (CCRF-CEM) in a 100 mM phosphate buffer (pH 2.0) and in the presence of 0.02% HPMC. Other conditions: voltage 12 kV, temperature 25◦ C, injection 2 s, detection 200 nm, untreated capillary (50 cm × 75 µm i.d.). Designations p0–p5 = non-, mono-, di-, tri-, tetra-, and pentaphosphorylated histone H1.5.

various phosphorylated forms, was separated using an uncoated capillary and a sodium phosphate buffer (pH 2.0) and 0.03% HPMC.Astriking example for the resolving power of this separation buffer system is depicted in Figure 23.2, illustrating the CE separation of the same multiphosphorylated histone H1.5 sample already shown in Figure 23.1. Applying the same CE conditions as described in Figure 23.1, except that 0.02% HPMC was added to the phosphate buffer, all five phosphorylated forms are now clearly separated from each other and from the unphosphorylated parent protein. The nonphosphorylated protein migrates fastest, followed by the distinctly phosphorylated forms, because binding of the negatively charged phosphate groups decrease the overall positive charge of the histone molecule thus diminishing the electrophoretic mobility of the phosphorylated protein species. Recently, Yoon et al.15 developed a CE-LIF method to determine phosphorylation levels and to follow the translocation of the green fluorescent protein-extracellular signal regulated protein kinase 2 (GFP-ERK2) from the cytoplasm to the nucleus. CE conditions applied were an untreated capillary and a 100 mM CAPS buffer (pH 11.0) containing 2 M betaine. LIF detection was performed with a 5 mW air-cooled argon ion laser (excitation, 488 nm/emission, 520 nm). Phosphorylated and nonphosphorylated GFP-ERK2 were not separated simultaneously, but in consecutive runs. Significant differences in migration time allowed a clear assignment of the modified and unmodified protein. Owing to separation conditions employed (high pH, high salt buffer), the GFP-ERK2 proteins are negatively charged and, therefore, no dynamic coating agent was added. Phosphorylation causes a further increase in the negative charge; thereby, phosphorylated GFP-ERK2 migrates slower. Compared to conventional Western blotting, the CE method allows the analysis of sample volumes as low as a few nanoliters and does not require a separate sample purification step and radiolabeling. On the basis of their differences in isoelectric point (pI), phosphorylated proteins can be separated in capillary isoelectric focusing (CIEF). Wei et al.16 employed a CIEF method using a capillary covalently coated with linear polyacrylamide and a pH gradient from 4 to 6.5 for the resolution of mono- and diphosphorylated ovalbumins. Proteins were detected by their UV absorbance at 280 nm. Additional ovalbumin variants within each of the mono- and diphosphoovalbumins, differing in their amount of glycosylation, were further analyzed by online CIEF-electrospray ionization (ESI)-MS.

Capillary Electrophoresis of Post-Translationally Modified Proteins and Peptides

713

In a similar manner, non-, mono-, and diphosphorylated myosin light chain using CIEF, either with UV or LIF detection, was separated by Shiraishi et al.17 Neutral coated capillaries (eCAP; Beckman Coulter) and HPMC in the ampholyte solution reduced electroosmotic flow (EOF) and protein adsorption. A detection limit of ∼1 pg fluorescently labeled myosin light chain/capillary was achieved. 23.3.1.2 Acetylation Histone H4 with a calculated pI of 11.9 is one of the most basic proteins known. Under certain biological conditions, it can be reversibly acetylated in its N-terminal region at four lysine residues. CE separation of these different acetylated forms was not possible applying chemically coated capillaries. However, when HPMC was used as a dynamic coating agent, the histone H4 sample isolated from whole histones by reversed phase chromatography (RPC) was clearly resolved into the non-, mono-, di-, tri-, and tetra-acetylated forms within about 22 min.18 Applying a special buffer system consisting of 500 mM formic acid/LiOH/10 mM urea (pH 2.0) with 0.02% HPMC, an ultrafast CE separation (4 min) of the distinctly acetylated H4 proteins could be achieved (shown in Figure 23.3). As acetylation of lysines decreases the positive charge of histone H4, the unacetylated protein migrates fastest. Using HPMC as buffer additive another core histone, subfraction H2A, which is a single peak in RPC, could be further separated by CE into five peaks consisting of non-and monoacetylated H2A.2a and H2A.2b, respectively, and even a third subtype H2A.3 was resolved.19 Wiktorowicz and Colburn20 separated core histone H4 into its different acetylated forms using a commercially available cationic surfactant (MicroCoat, ABI). Coating of the silica surface was performed by rinsing the capillary with the reagent followed by a wash with running buffer. Like other cationic surfactants such as cetyl trimethylammonium bromide (CTAB), the coating reagent is primarily bound to the silica surface by ionic interaction. In a second step, the surface-bound neutralized surfactant binds additional reagent cations by virtue of hydrophobic interactions between alkyl side groups. In this way, a remarkably stable bilayer is formed and the charge of the capillary wall is reversed from negative to positive. Consequently, the positively charged histone molecules 0.008

H4 ac0

0.006

A200

0.004 ac1

0.002

ac2 ac3

0.000

ac4 –0.002 3.5

3.7

3.9

4.1

4.3

4.5

Time (min)

FIGURE 23.3 Electropherogram of multiacetylated histone H4 from mouse tumor cells (NIH) in 500 mM formic acid/LiOH/10 mM urea buffer (pH 2.0) containing 0.02% HPMC. Conditions: voltage 30 kV, injection 1 s. Other conditions are as in Figure 23.2. Designations ac0–ac4 = non-, mono-, di-, tri-, and tetra-acetylated histone H4.

714

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

are repelled from the capillary surface and under the influence of an electric field, a reversal of the EOF takes place. It should be mentioned, however, that strictly speaking this particular approach of charge reversal is a special case of the dynamic coating concept, as no coating reagent is included in the running buffer. For this reason, periodic replenishment of the coating by flushing the capillary with the surfactant solution is required. Besides histones, very little is described about the CE analysis of other acetylated proteins. Some articles report on the separation of N-terminally acetylated and unacetylated forms of metallothioneins (MTs). MTs are structurally unusual proteins due to their small size (6–7 kDa), high cysteine and metal content, and remarkable kinetic lability. Purified rat MT-2 protein was separated by MEKC using an electrolyte of 100 mM sodium borate buffer containing 75 mM SDS at pH 8.4 and showed two peaks.21 Identification of both peaks was performed with online CE-MS using 100 mM formic acid and 2% methanol. The two peaks differed by a mass equivalent to that of a single acetyl group. Another group established a CE-ESI-time of flight (TOF)-MS method to characterize rabbit liver MT isoforms and was successful in separating several N-acetylated and non-N-acetylated forms using 100 mM acetic acid:100 mM formic acid (pH 2.3).22 CIEF has been applied to the separation and quantification of the three main hemoglobin components of umbilical cord blood (fetal, acetylated fetal, and adult hemoglobins).23 CIEF was performed with a poly (acryloylaminoethoxy-ethonal) [poly(AAEE)] coated capillary and a carrier ampholyte consisting of 5% Ampholine pH 6–8, supplemented with 0.5% TEMED. An interesting CE application has been the use of so-called charge ladders of proteins for measuring the role of charge in protein stability, protein–ligand binding, and ultrafiltration.24–26 A protein charge ladder is a collection of derivatives of a protein by converting its charged groups into electrically neutral ones. Charge ladders can be easily generated in vitro by the treatment of a variety of model proteins with acetic anhydride. CE has been shown to be an effective tool in the analysis of such charge ladders, as it is used to separate the proteins that constitute the charge ladder into individual “rungs,” each rung contains derivatives with the same number of modified groups. Cordova et al.27 explored the CE behavior of protein charge ladders obtained by acetylation of lysozyme and carbonic anhydrase II using noncovalent polycationic coated capillaries. Two of them, polyethylenimine and Polybrene, were very effective in preventing the adsorption of positively charged proteins. Conditions used were fused-silica capillary (50 µm i.d. × 38 cm) coated with the polymer by flushing the capillary with a 7.5% (w/v) polymer solution, prepared in 25 mM Tris– 192 mM Gly buffer (pH 8.3), for 15 min. The running buffer was 25 mM Tris–192 mM Gly buffer (pH 8.3) in the absence of polymer. Separations were obtained within 5 min. Carbeck et al.28 showed CE-ESI MS to be a useful tool for the study of charge ladders of lysozyme, carbonic anhydrase II, and bovine pancreatic trypsin inhibitor and for the examination of the relationship between the properties of proteins in the solution phase and in the gas phase. An example for a CE separation of such a charge ladder is shown in Figure 23.6 (Section 23.4.3). Separation of multiply modified proteins place high demands on the method applied. An example for a successful separation of a protein containing differently phosphorylated and acetylated forms in a single run using the HPMC-based CZE method is shown in Figure 23.4.

23.3.1.3 Deamidation The analysis of recombinant human growth factor (rhGH), one of the first biotechnologically produced proteins in Escherichia coli, by CE with UV absorbance and MS detection using bilayer-coated capillaries was demonstrated very recently.29 The authors present an improved CE method using capillaries noncovalently coated with polybrene and poly(vinyl sulfonic acid) and a background electrolyte of 400 mM Tris phosphate (pH 8.5) to achieve efficient separations of intact rhGH and degradation products like deamidated and oxidated forms.

Capillary Electrophoresis of Post-Translationally Modified Proteins and Peptides

H5

0.050

715

ac0p0

0.040

A200

0.030 ac1p0 ac0 p1 p2

0.020 0.010

ac1p1 0.000 20

21

22

23

24

Time (min)

FIGURE 23.4 Electropherogram of histone H5 from chicken erythrocytes. Conditions were the same as in Figure 23.2. Designations ac0p0 = nonacetylated nonphosphorylated H5; ac0p1 and ac0p2 = nonacetylated mono- and diphosphorylated H5; ac1p0 and ac1p1 = monoacetylated non-and monophosphorylated histone H5.

A novel method for the determination of glutamine deamidation in a long repetitive protein polymer via bioconjugate capillary electrophoresis was published recently by Won et al.30 By conjugating a monodisperse, fluorescently labeled DNA oligomer to long polydisperse nearly electrostatically neutral protein polymers, protein polymers differing in degree of deamidation were successfully separated. For protein polymers with increasing extents of deamidation, the electromotive force of DNA + polypeptide conjugate molecules increases due to the introduced negative charge of deamidated glutamic acid residues. CE analysis reveals increasing differences in the electrophoretic mobilities of conjugate molecules, which qualitatively shows the degree of deamidation. CE of protein–DNA conjugates was performed using a capillary coated with POP-5 polymer and a 50 mM Tris, 50 mM TAPS, 2 mM EDTA buffer containing 7 M urea (pH 8.4) added with 3% POP-5 solution. Histone H10 is known to be deamidated in the course of aging.5 Intact H10 contains asparagine in position 3, while depending on the age of the organ deamidated forms of H10 containing either aspartic acid or isoaspartic acid accumulate. The buffer system containing HPMC enables the separation of intact histone H10 fragment (residues 1–52) as well as its deamidated forms (Figure 23.5).

23.3.1.4 Methylation Unlike protein modification by acetylation or phosphorylation, methylation does not greatly influence the charge of individual amino acids, thus making the electrophoretic separation of distinct methylated proteins from each other and from the unmethylated parent protein a problematic part. Separation of acetylated histone H4 has already been described by Lindner et al.18 using CZE with HPMC as dynamic coating reagent. The same method enables the simultaneous separation of mono-, di-, and trimethylated H4, including their distinct acetylated forms.6 Changes in the modification pattern of H4 from normal tissues, cancer cell lines, and primary tumors were found suggesting that a global loss of monoacetylation and trimethylation of histone H4 is a common hallmark of human cancer.

716

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

H10 (1–51)

Asn

A200

Asp

isoAsp

17

19

21 Time (min)

23

25

FIGURE 23.5 CE separation of deamidated proteins. Histone H10 from rat liver was digested with chymotrypsin and the N-terminal fragment obtained by RPC analyzed by CE. Conditions were the same as in Figure 23.2. Designation Asn, Asp, isoAsp = asparagine, aspartic acid, and isoaspartic acid containing fragments.

23.3.2 PEPTIDES 23.3.2.1 Phosphorylation A CZE procedure for detection and assay of protein kinase and phosphatase activities in complex biological mixtures has been developed by Dawson et al.31 The phosphorylated and dephosphorylated forms of several peptides were resolved using an uncoated capillary and a 150 mM phosphoric acid buffer at pH 2.0 or pH 5.0. Furthermore, the CZE-based assay was capable of resolving a peptide phosphorylated on different sites and permitted the quantification of each peptide. Since this application, much research has been done on protein kinase assays based on CE. For the same purpose, Gamble et al.32 tested both neutral and cationic coated and uncoated capillaries and found that the best conditions for the separation of phosphopeptide isomers is formic acid or phosphate buffer with pH ranging between 5.5 and 6.5 and a PVA-coated capillary. Synthetic peptides were incubated with various protein kinases and phosphatases and the phosphorylation and dephosphorylation was monitored using CZE. These methods enabled the assay of several protein kinases and phosphatases and the determination of the sites of phosphorylation. Improved CE-based methods to measure kinase activation in single cells have been described recently.33 Phosphorylated and nonphosphorylated forms of peptide substrates for protein kinase C and calcium-calmodulin activated kinase were separated by CZE using a polydimethylacrylamide (PDMA) coated capillary and buffers containing different concentrations of betaine (0–1 M). The separation system is compatible with a living cell and, therefore, adaptable to the laser micropipet system, a strategy to measure the activation of enzymes in single mammalian cells. Synthetic peptides containing phosphorylated tyrosine have been shown to be easily separated by a CZE method using a linear polyacrylamide-coated capillary and a 300 mM borate buffer (pH 8.5).34 Another separation mode of HPCE, namely MEKC, has been used to separate a mixture of phosphopeptide isomers of the insulin receptor peptide.35 The resolution in coated and uncoated capillaries was compared using a 50 mM phosphate buffer with 25 mM SDS (pH 6.1). An efficient separation of diphosphorylated isomers could be obtained by using polyacrylamide-coated capillaries.

Capillary Electrophoresis of Post-Translationally Modified Proteins and Peptides

717

Mass spectrometry of phosphopeptides has become a powerful tool for phosphorylation site identification. However, proteolytic digests examined by MS are often likely to fail to detect phosphopeptides because the ionization of phosphorylated peptides in positive ion mode MS is generally less efficient compared with the ionization of their nonphosphorylated counterparts resulting in ion suppression effects. A further problem is that phosphopeptides may not be retained by RP chromatography because they are too small and/or hydrophilic to bind to the C18 stationary phase. Therefore, capillary electrophoresis coupled to MS is a powerful method to enhance the detection of phosphoproteins and phosphopeptides due to their efficient separation by CE. Sandra et al.36 developed a CE-MS technique to the characterization of the N-glycans from the glycoprotein cellobiohydrolase I (CBH I) with special interest for the phosphorylated species. CBH I shows phosphorylated glycans that are only present when the organism is grown under minimal conditions and could be related to stress response. The glycans were labeled with the negatively charged tag 8-aminopyrene-1,3,6-trisulfonate (APTS) by reductive amination and separated using an untreated capillary and a 25 mM ammonium acetate buffer (pH 4.55), allowing the simultaneous analysis of uncharged and charged glycans and, moreover, the differentiation of phosphorylated isomers. Detection of phosphorylated proteins and peptides can be enhanced by selectively isolating these species. Online immobilized metal affinity chromatography (IMAC)-CE-ESI-MS is such a powerful analytical tool. The IMAC resin retains and preconcentrates phosphorylated proteins and peptides, CE separates the phosphorylated species and MS/MS identifies the components and their phosphorylation sites. Cao and Stults37,38 applied this method to the analysis of phosphorylated angiotensin II and tryptic digests of α- and β-casein (CE conditions: buffer, 0.1% acetic acid/10% methanol; uncoated capillary). Beta-casein is a well-characterized protein with five phosphorylation sites and is widely used as a standard for protein phosphorylation studies. 23.3.2.2 Deamidation The effectiveness of CZE for the separation of deamidated peptides was shown by Ganzler et al.39 Excellent resolution of peptides containing all three forms, Asn, Asp, and isoAsp, could be achieved by using a 20 mM sodium citrate buffer (pH 2.5). 23.3.2.3 Farnesylation Protein farnesylation, catalyzed by protein farnesyltransferase, plays important roles in the membrane association and protein–protein interaction of a number of eukaryotic proteins. The enzyme transfers a 15-carbon farnesyl moiety from farnesyl diphosphate (FPP) to the sulfhydryl group of cysteine. The activity of the enzyme was measured by CE with LIF detection, which is a powerful alternative to classical methods involving radiolabeled FPP.40 LIF detection was performed with an argon ion laser (excitation, 488 nm/emission, 520 nm). A fluorescently labeled pentapeptide that was used as substrate was clearly separated from its farnesylated form under the four CE buffer conditions investigated (e.g., 25 mM borax, 25 mM SDS, pH 9.3, uncoated capillary). The method will be of great value in studies of inhibitors of protein farnesyltransferase in vitro.

23.4 METHOD DEVELOPMENT GUIDELINES Many factors may influence the quality of CE separations and contribute to the resolution of protein modifications. Generally, optimization strategies, precautions, and CE conditions applied to peptides are also applicable to the separation of modified proteins. However, optimum parameters for protein separations must still be determined empirically to some extent. In addition to well-known factors influencing resolution in CE like capillary surface and coating, buffer pH, also effects of different

718

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

buffer cations and anions, buffer concentration, additives, and chaotropic agents can be utilized to enhance resolution of protein mixtures.

23.4.1 BUFFERS The running buffer selection is substantial to the success of any CE separation. Factors being most important are the UV absorbance at low wavelengths, good pH stability, and conductivity. The highest sensitivity detection is achieved at low UV wavelengths (190–220 nm) where peptide bonds have an absorbance maximum. Good sensitivity at these wavelengths requires buffer systems with high-UV transparency. The most frequently used buffers are sodium phosphate, sodium citrate, Tris, and sodium borate. Sodium formate shows optimal compatibility to mass spectrometric analysis and is therefore often used in CE coupled online to MS. The buffer pH should be stable during a series of analyses. For reproducible analysis times the buffer should be exchanged or remixed every 5–10 runs. Cellulose-based buffers should be freshly prepared at least once a week, depending on the number of runs, and should not(!) be stored in the refrigerator. The quality and purity of all buffer components should be of the highest grade and the buffers should be filtered through 0.45 µm pore size filters before use. A starting buffer, which gives superior results in many cases, is a 100 mM sodium phosphate buffer (pH 2.5) with the addition of 0.02% HPMC. Owing to the excellent UV transparency of the separation buffer, proteins can be detected very sensitively at 200 nm. In case this system does not provide satisfying results, more complex separation conditions must be taken into account according to specific properties of the proteins (e.g., pI, hydrophobicity, etc.).

23.4.2 CAPILLARY TREATMENT Using an untreated capillary every 10–15 injections the capillary should be rinsed with water, 0.1 M NaOH, water, 0.5 M H2 SO4 , and finally with the running buffer. Washing should be done for 2 min with each solvent. The same wash, but without running buffer, should be applied at the end of sample analyses. After flushing the capillary with air, it can be stored. There are many different static wall coatings described in the literature, recently reviewed by Horvath and Dolnik.8 Static coated capillaries should be treated according to the manufacturer’s guidelines. Solvents such as NaOH and H2 SO4 must not be used.

23.4.3 METHOD EVALUATION The evaluation of the applicability of a certain CE method for the separation of modified proteins can easily be performed by analyzing acetylated lysozyme generated by the treatment of the protein with acetic anhydride (described in Section 23.3.1.2). An aqueous solution of lysozyme (0.1 mM, 0.5 mL) adjusted to pH 10 by using 0.1 M NaOH is incubated with 4.5 µL acetic anhydride (100 mM in dioxane) for 5 min at room temperature according to Cordova et al.27 An example using the suggested phosphate buffer with HPMC is shown in Figure 23.6.

23.5 CONCLUDING REMARKS The analysis of post-translationally modified proteins is due to their biological significance a field of ever increasing importance. For many years, different kinds of PAGE have been the most commonly used tool for studying modified proteins with all the shortcomings of this technique generally known. The advent of capillary electrophoresis provided a promising new tool for their separation. HPCE has made significant advances in recent years and is establishing itself as a rapid, reproducible, sensitive, and highly resolving method. Hyphenated techniques, for example, the combination with MS, will

Capillary Electrophoresis of Post-Translationally Modified Proteins and Peptides

A200

0.20

719

(a)

0.10

0.00

A200

0.02

(b)

0.01

0.00 10

15

20

25

30

Time (min)

FIGURE 23.6 CE analysis of the lysozyme charge ladder. (a) Untreated lysozyme and (b) charge ladder from acetylation of lysozyme. Conditions were the same as in Figure 23.2.

be able to provide detailed structural information like modification status and sites and will further increase the range of potential CE applications in this field.

ACKNOWLEDGMENTS This work, as part of the European Science Foundation EUROCORES Programme EuroDYNA, was partly supported by funds from the Austrian Science Foundation (project I23-B03) and by the Jubilee Fund of the Austrian National Bank, grant no. 9319.

REFERENCES 1. Yan, J. X., N. H. Packer, A. A. Gooley, and K. L. Williams, Protein phosphorylation: Technologies for the identification of phosphoamino acids, J. Chromatogr. A, 808, 23–41, 1998. 2. Paik, W. K. and S. Kim, Protein Methylation, John Wiley & Sons, New York, 1980. 3. Wold, F., In vivo chemical modification of proteins (post-translational modification), Annu. Rev. Biochem., 50, 783–814, 1981. 4. Lindner, H. and W. Helliger, Age-dependent deamidation of asparagine residues in proteins, Exp. Gerontol., 36, 1551–1563, 2001. 5. Lindner, H., B. Sarg, B. Hoertnagl, and W. Helliger, The microheterogeneity of the mammalian H1(0) histone. Evidence for an age-dependent deamidation, J. Biol. Chem., 273, 13324–13330, 1998. 6. Fraga, M. F., E. Ballestar, A. Villar-Garea, M. Boix-Chornet, J. Espada, G. Schotta, T. Bonaldi, et al., Loss of acetylation at Lys16 and trimethylation at Lys20 of histone H4 is a common hallmark of human cancer, Nat. Genet., 37, 391–400, 2005. 7. Sarg, B., W. Helliger, B. Hoertnagl, B. Puschendorf, and H. Lindner, The N-terminally acetylated form of mammalian histone H1(0), but not that of avian histone H5, increases with age, ABB, 372, 333–339, 1999.

720

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

8. Horvath, J. and V. Dolnik, Polymer wall coatings for capillary electrophoresis, Electrophoresis, 22, 644–655, 2001. 9. Strahl, B. D. and C. D. Allis, The language of covalent histone modifications, Nature, 403, 41–45, 2000. 10. Turner, B. M., Histone acetylation and an epigenetic code, Bioessays, 22, 836–845, 2000. 11. Jenuwein, T. and C. D. Allis, Translating the histone code, Science, 293, 1074–1080, 2001. 12. Lindner, H., W. Helliger, A. Dirschlmayer, H. Talasz, M. Wurm, B. Sarg, M. Jaquemar, and B. Puschendorf, Separation of phosphorylated histone H1 variants by high- performance capillary electrophoresis, J. Chromatogr., 608, 211–216, 1992. 13. Lindner, H., M. Wurm, A. Dirschlmayer, B. Sarg, and W. Helliger, Application of high-performance capillary electrophoresis to the analysis of H1 histones, Electrophoresis, 14, 480–485, 1993. 14. Lindner, H., W. Helliger, B. Sarg, and C. Meraner, Effect of buffer composition on the migration order and separation of histone H1 subtypes, Electrophoresis, 16, 604–610, 1995. 15. Yoon, S., K. Y. Han, H. S. Nam, L. V. Nga, and Y. S. Yoo, Determination of protein phosphorylation and the translocation of green fluorescence protein-extracellular signal-regulated kinase 2 by capillary electrophoresis using laser induced fluorescence detection, J. Chromatogr. A, 1056, 237–242, 2004. 16. Wei, J., L. Yang, A. K. Harrata, and C. S. Lee, High resolution analysis of protein phosphorylation using capillary isoelectric focusing—electrospray ionization—mass spectrometry, Electrophoresis, 19, 2356–2360, 1998. 17. Shiraishi, M., R. D. Loutzenhiser, and M. P. Walsh, A highly sensitive method for quantification of myosin light chain phosphorylation by capillary isoelectric focusing with laser-induced fluorescence detection, Electrophoresis, 26, 571–580, 2005. 18. Lindner, H., W. Helliger, A. Dirschlmayer, M. Jaquemar, and B. Puschendorf, High-performance capillary electrophoresis of core histones and their acetylated modified derivatives, Biochem. J., 283 (Pt 2), 467–471, 1992. 19. Lindner, H., B. Sarg, C. Meraner, and W. Helliger, Separation of acetylated core histones by hydrophilic-interaction liquid chromatography, J. Chromatogr. A, 743, 137–144, 1996. 20. Wiktorowicz, J. E. and J. C. Colburn, Separation of cationic proteins via charge reversal in capillary electrophoresis, Electrophoresis, 11, 769–773, 1990. 21. Beattie, J. H., A. M. Wood, and G. J. Duncan, Rat metallothionein-2 contains N(alpha)-acetylated and unacetylated isoforms, Electrophoresis, 20, 1613–1618, 1999. 22. Andon, B., J. Barbosa, and V. Sanz-Nebot, Separation and characterization of rabbit liver apothioneins by capillary electrophoresis coupled to electrospray ionization time-of-flight mass spectrometry, Electrophoresis, 27, 3661–3670, 2006. 23. Conti, M., C. Gelfi, and P. G. Righetti, Screening of umbilical cord blood hemoglobins by isoelectric focusing in capillaries, Electrophoresis, 16, 1485–1491, 1995. 24. Negin, R. S. and J. D. Carbeck, Measurement of electrostatic interactions in protein folding with the use of protein charge ladders, J. Am. Chem. Soc., 124, 2911–2916, 2002. 25. Gao, J., M. Mammen, and G. M. Whitesides, Evaluating electrostatic contributions to binding with the use of protein charge ladders, Science, 272, 535–537, 1996. 26. Menon, M. K. and A. L. Zydney, Effect of ion binding on protein transport through ultrafiltration membranes, Biotechnol. Bioeng., 63, 298–307, 1999. 27. Cordova, E., J. Gao, and G. M. Whitesides, Noncovalent polycationic coatings for capillaries in capillary electrophoresis of proteins, Anal. Chem., 69, 1370–1379, 1997. 28. Carbeck, J. D., J. C. Severs, J. Gao, Q. Wu, R. D. Smith, and G. M. Whitesides, Correlation between the charge of proteins in solution and in the gas phase investigated by protein charge ladders, capillary electrophoresis, and electrospray ionization mass spectrometry, J. Phys. Chem. B, 102, 10596–10601, 1998. 29. Catai, J. R., T. J. Sastre, P. M. Jongen, G. J. de Jong, and G. W. Somsen, Analysis of recombinant human growth hormone by capillary electrophoresis with bilayer-coated capillaries using UV and MS detection, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci., 852, 160–166, 2007. 30. Won, J. I., R. J. Meagher, and A. E. Barron, Characterization of glutamine deamidation in a long, repetitive protein polymer via bioconjugate capillary electrophoresis, Biomacromolecules, 5, 618–627, 2004.

Capillary Electrophoresis of Post-Translationally Modified Proteins and Peptides

721

31. Dawson, J. F., M. P. Boland, and C. F. Holmes, A capillary electrophoresis-based assay for protein kinases and protein phosphatases using peptide substrates, Anal. Biochem., 220, 340–345, 1994. 32. Gamble, T. N., C. Ramachandran, and K. P. Bateman, Phosphopeptide isomer separation using capillary zone electrophoresis for the study of protein kinases and phosphatases, Anal. Chem., 71, 3469–3476, 1999. 33. Li, H., H. Y. Wu, Y. Wang, C. E. Sims, and N. L. Allbritton, Improved capillary electrophoresis conditions for the separation of kinase substrates by the laser micropipet system, J. Chromatogr. B Biomed. Sci. Appl., 757, 79–88, 2001. 34. Bonewald, L. F., L. Bibbs, S. A. Kates, A. Khatri, J. S. McMurray, K. F. Medzihradszky, and S. T. Weintraub, Study on the synthesis and characterization of peptides containing phosphorylated tyrosine, J. Pept. Res., 53, 161–169, 1999. 35. Tadey, T. and W. C. Purdy, Capillary electrophoretic resolution of phosphorylated peptide isomers using micellar solutions and coated capillaries, Electrophoresis, 16, 574–579, 1995. 36. Sandra, K., J. Van Beeumen, I. Stals, P. Sandra, M. Claeyssens, and B. Devreese, Characterization of cellobiohydrolase I N-glycans and differentiation of their phosphorylated isomers by capillary electrophoresis-Q-Trap mass spectrometry, Anal. Chem., 76, 5878–5886, 2004. 37. Cao, P. and J. T. Stults, Mapping the phosphorylation sites of proteins using on-line immobilized metal affinity chromatography/capillary electrophoresis/electrospray ionization multiple stage tandem mass spectrometry, Rapid Commun. Mass Spectrom., 14, 1600–1606, 2000. 38. Cao, P. and J. T. Stults, Phosphopeptide analysis by on-line immobilized metal-ion affinity chromatography-capillary electrophoresis-electrospray ionization mass spectrometry, J. Chromatogr. A, 853, 225–235, 1999. 39. Ganzler, K., N. W. Warne, and W. S. Hancock, Analysis of rDNA-derived proteins and their posttranslational modifications, in Capillary Electrophoresis in Analytical Biotechnology, P. G. Rhigetti (Ed.) CRC press, Boca Raton, FL, pp. 185–238, 1996. 40. Berezovski, M., W. P. Li, C. D. Poulter, and S. N. Krylov, Measuring the activity of farnesyltransferase by capillary electrophoresis with laser-induced fluorescence detection, Electrophoresis, 23, 3398–3403, 2002.

Resolution in 24 Extreme Capillary Electrophoresis: UHVCE, FCCE, and SCCE Wm. Hampton Henley and James W. Jorgenson CONTENTS 24.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2 Ultrahigh Voltage Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.1 Background and Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.2 Practical Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.2.1 High Voltage Capillary Shielding System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.2.2 Power Supply and Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.2.3 HV Insulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.2.4 Modes of Operation and Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.2.5 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.2.6 Methods Development Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.2.7 Buffer Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.2.8 Capillary Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.2.9 Sample Separation Strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3 Flow Counterbalanced Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.2 Theoretical Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.3 Practical Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.3.1 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.3.2 Capillary Packing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.3.3 Modes of Operation and Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.4 Methods Development Guidelines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.4.1 Handling Extended Separation Times. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.4.2 Band Broadening Prevention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.4.3 Determining the Best Separation Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4 Synchronous Cyclic Capillary Electrophoresis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.1.1 Microchip SCCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.1.2 Capillary SCCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.2 Theoretical Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.2.1 Absolute versus “Effective” Voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.2.2 Band Broadening Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.3 Practical Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.3.1 Power Supply and Voltage Switching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.3.2 Connection of Fused-Silica Capillaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.3.3 Modes of Operation and Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

724 725 725 728 728 728 730 730 731 732 733 735 735 736 736 736 738 738 740 741 744 744 745 746 746 746 747 747 748 748 748 749 749 749 750 723

724

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

24.4.3.4 Methods of Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.3.5 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.4 Methods Development Guidelines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.4.1 Handling Extended Separation Times. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.4.2 Band Broadening Prevention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.4.3 Sample Separation Strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

753 753 753 753 754 754 755 755 755

24.1 INTRODUCTION The majority of difficult separations can be divided into two main classes: complex mixtures and challenging pairs (Figure 24.1). Complex mixtures, such as proteins and protein digests, demonstrate “random” or broad spectrum behavior with respect to their properties, whereas “well-behaved” analytes such as nucleic acids display predictable behavior determined mainly by the size of the molecule. The analysis of these complex samples requires a separation system with high peak capacity and the ability to cover a broad spectrum of physical properties. Challenging pairs, such as enantiomers and isotopomers, display properties that are so similar in nature that a separation technique must have extremely high efficiency in order to resolve the analytes. The high efficiency can come at the expense of the ability to analyze a broad spectrum of different analyte properties because the species of interest are confined to closely migrating bands. Conventional approaches to increasing the resolution for difficult separations in capillary electrophoresis (CE) usually require extensive development of buffers and/or capillary coatings particular to an individual separation. The improvement in resolution is usually the result of exploiting a specific intrinsic property of the analytes. A universal approach to achieving broad spectrum high resolution separations can be achieved using CE with an extremely high applied potential. Separation efficiency increases linearly with applied voltage, and resolution improves with the square root of applied voltage. Ultrahigh voltage CE (UHVCE) uses existing buffer systems and separation conditions but gives the analytes more time to separate at a given electric field strength. The separation of challenging pairs that are not resolved using traditional methods is often achieved by developing novel pseudo-stationary phases for enantiomers or by utilizing mass spectrometry for analyzing isotopomers. Resolution of these tough pairs can also be achieved using

Difficult separations

Complex mixtures

“Random”

Peptides/digests

Proteins

Challenging pairs

“Well behaved”

Enantiomers

Isotopomers

Nucleic acids

Molecules

Small ions

FIGURE 24.1 Dividing difficult separations into several categories can help in the development of separation strategies.

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

725

extended separation times under high electric field strength. Because the difficult pairs exist as closely migrating analyte bands, techniques such as flow counterbalanced CE (FCCE) and synchronous cyclic CE (SCCE) can be used. These methods use conventional separation voltages in relatively short capillaries but use procedures to increase the duration in a high electric field.

24.2 ULTRAHIGH VOLTAGE CAPILLARY ELECTROPHORESIS 24.2.1 BACKGROUND AND THEORY Early work with CE demonstrated how higher applied voltages offered vast improvements for electrophoretic separations.1–5 Separation efficiency shows a linear dependence on applied voltage when analytes are detected very close to the end of the capillary: (µe + µEOF ) V N∼ , = 2D

(24.1)

where N is the efficiency in theoretical plates, µe is the electrophoretic mobility of the analyte, µEOF is the mobility of the electroosmotic flow (EOF), V is the applied voltage, and D is the diffusion coefficient of the analyte.6 Resolution between analytes with electrophoretic mobilities µ1 and µ2 show a dependence on the square root of the applied voltage:7 Rs = 0.177(µ1 − µ2 )




V D(µe + µEOF )

1/2

.

(24.2)

In addition, the speed of a separation can be greatly increased with higher applied voltages. Migration time for an analyte (tm ) displays an inverse linear relationship with applied voltage for the same length capillary (L) and separation distance (l): tm =

lL . V (µe + µEOF )

(24.3)

The flat electrokinetic flow profile suggests that as long as Joule heating can be prevented, there is no theoretical limit to the improvements to be gained by increasing the applied voltage. In practice, several factors combined to limit the applied voltage used in early instrumentation to approximately 30 kV. High electric field strengths can lead to resistive or Joule heating of the buffer solution within the capillary. A detailed discussion of the effects of Joule heating with voltages up to 60 kV has been described by Palonen et al.8 In the simplest of terms, the power (P) running through the capillary is given by P = iV =

V2 , R

(24.4)

where R is the resistance of the capillary. Since it is desirable to have the greatest possible voltage, the resistance of the capillary must be large enough to keep the power below about 1 W per meter of capillary length. Initial experiments conducted at voltages in excess of 30 kV resulted in broken capillaries. A quick calculation of the radial electric field at the silica/buffer interface (Figure 24.2) using the following equation explains why this occurs:9 Eradial =

Vi − Vo , ln(b/rc )Krc

(24.5)

726

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

E

E Longitudinal field

Radial field

FIGURE 24.2 Longitudinal electric fields are used to separate analytes in CE. Radial fields also exist through the capillary wall, and at high voltages, they can destroy the capillary via dielectric breakdown.

where Vi is the voltage inside the capillary, Vo is the voltage of a nearby conductor distance b from the center of the capillary, rc is the length of the inner radius of the capillary, and K is the dielectric constant of the capillary wall (∼3.9 for fused silica). For a capillary 50 µm in diameter, 1 m from the nearest grounded conductor, with 100 kV applied, the radial electric field strength is about 100 million V/m. Radial electric fields greater than the dielectric strength of fused silica (∼15 million V/m) damage the fused silica wall and cause defects in the structure leading to fragmentation. Early efforts using a van de Graff generator10 at ∼200 kV and a thick dielectric coating such as a sleeve of Teflon® tubing around the capillary were unsuccessful. In another attempt, a plastic tube surrounding the capillary was filled with isopropyl alcohol.11 In theory, both the capillary and the tube filled with alcohol can be treated as long resistors, and therefore the potential will drop linearly down the length of the capillary. The potential drop in the outer tube should match the potential drop in the capillary and therefore reduce the radial field through the capillary wall. Unfortunately, the capillaries still degraded in a very short time. The first capillary shielding system that could prevent dielectric breakdown of the fused-silica capillary was demonstrated by Hutterer and Jorgenson.12–15 The capillary was placed inside a series of metal rings charged to potentials closely matching those inside the capillary (Figure 24.3). The electric potential from the rings counteracts the radial field from the capillary and shields the capillary from damage. The high voltage direct current (HVDC) was generated using a 26-stage Cockcroft– Walton voltage multiplier16 with 2.4 nF HV capacitors. The voltages needed to charge the rings were supplied by different stages of the multiplier. The entire instrument was submerged in a tank containing transformer oil to provide electrical insulation. This instrument showed the expected improvements in efficiency and resolution for samples of peptides via CE,7,12 hyaluronic acid via capillary gel electrophoresis (CGE),12,14 and oligosaccharide mixtures via micellular electrokinetic capillary chromatography.12,15 Application of up to 120 kV was possible, and no degradation of the capillary was observed. Figure 24.4 shows electropherograms of a cytochrome c digest separated on the same capillary using 28 kV and 120 kV.12 It is easy to see the improvements in resolution and the decrease in separation time when higher voltages are applied. The current state-of-the-art instrumentation uses larger capacitors (10 nF) to create a 100-fold Cockcroft–Walton voltage multiplier capable of reaching potentials as high as 410 kV.6 A total of 25 shielding rings keep the potential difference between the capillary and shielding system at less than 20 kV. The benchtop instrument runs in air, using plastic dielectric material instead of the tank of transformer oil. The HV polarity can easily be reversed, allowing both positive and negative HV separations. The temperature of the instrument can be controlled to within a few degrees over a range of ambient temperature up to 60◦ C. It can be easily moved and coupled to different detection systems including laser-induced fluorescence (LIF) detection, UV absorbance detection, and electrospray ionization-mass spectrometry (ESI-MS). Separations of peptides, proteins, and nucleic acids using the instrument with these detectors are described in Sections 24.2.2.4, 24.3.3.3, and 24.4.3.3.

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

727

80 kV

80 kV

60 kV

40 kV

20 kV

0 kV

FIGURE 24.3 Charged metal shielding rings can be used to reduce the radial electric field strength. Closely matching the potential charge on the ring to that within the capillary prevents dielectric breakdown and allows operation at very high voltages. (Reprinted from Henley, W. H., PhD Dissertation, University of North Carolina, Chapel Hill, 2005. Copyright 2005. With permission.) (a) 0.6 28 kV

mAbs

0.4

0.2

0.0 220

230

240

250

260

270 280 Time (min)

290

300

310

320

330

(b) 1.0

mAbs

0.8

120 kV

0.6 0.4 0.2 0.0 55

60

65

70

75 Time (min)

80

85

90

95

FIGURE 24.4 Separations of cytochrome c tryptic digest using the same capillary at conventional (top) and UHV conditions. (Example: Reprinted from Hutterer, K. M., Doctorial Thesis. University of North Carolina, Chapel Hill, 2000. Copyright 2000. With permission.)

728

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

24.2.2 PRACTICAL IMPLEMENTATION 24.2.2.1 High Voltage Capillary Shielding System The first step in implementing UHVCE is to devise a capillary shielding system adequate for the desired separation voltage. The radial field that causes dielectric breakdown of the capillary must be dealt with in some manner. The dielectric breakdown strength of silica is approximately 15 MV/m. Equation 24.4 can be used to show that small diameter capillaries have a large radial electric field at relatively modest voltages. One approach is to simply use a large i.d. capillary with a thick wall. This may work for relatively low voltages, but not without band broadening and heat dissipation problems inherent to large bore capillaries. Another approach uses a conductive, but high resistance, material to surround the capillary in a uniform coating. High voltage is applied to the capillary and to the surrounding material. If the resistance of the material is equal along its length, the potential drop should mirror that of the capillary and thus neutralize the radial electric field through the wall of the capillary. Attempts using isopropyl alcohol in a tube surrounding the capillary resulted in failure after a relatively short time.11 Nevertheless, such a system should be possible. The most suitable method found so far is the use of charged metal rings to surround the capillary and neutralize the radial field (Figure 24.3). Briefly, the capillary is surrounded by a series of metal rings charged to a potential that roughly matches those inside the capillary. Long lengths of capillary can be equally spaced within the rings by winding them onto a suitable spool. This method has been successfully used without any noticeable degradation of the capillary. 24.2.2.2 Power Supply and Development The development of the power supply for UHVCE requires several considerations. Voltages of the magnitude used for UHVCE can easily arc through the air and “crawl” along insulating surfaces for several feet. Safety is paramount, and a grounded metal shell or Faraday cage should be used to protect the operator from electrocution and to protect sensitive nearby equipment from electromagnetic pulses caused by stray electric arcs. The supply must generate high voltage with sufficient current to drive the separation and charge the shielding system. Loss of charge from corona and stray capacitance must be prevented if the full potential of the power supply is to be realized. The generation of high voltage can be accomplished in many ways. Perhaps the simplest method is the van de Graff generator.10 Briefly, charge is pumped from one roller to anther roller held inside a metal sphere via a belt of dielectric material. The van de Graff generator is a constant current source, and therefore the voltage is limited by the loss of charge to the surrounding air and resistive load. Submerging the generator in transformer oil can dramatically reduce charge loss and improve the efficiency of the generator. Since the voltage is generated in a single step, charging metal shield rings to different potentials to shield the capillary may not be practical. If a single commercial high voltage power supply is used, a string of high voltage resistors set up to divide the voltage will be needed to bias the shielding rings. Such a system would be most practical for voltages up to about 100 kV, and even then require careful insulation of the resistors and high voltage wires. The Cockcroft-Walton voltage multiplier has been used to generate the high voltage for the two UHVCE systems developed thus far.6,7,12,14 The instrumentation has already been thoroughly detailed elsewhere, and so will only briefly be discussed here. The multiplier is a convenient and inexpensive solution to several problems surrounding UHVCE. It produces successively higher d.c. voltages as the multiplier chain is extended, providing multiple voltage sources to bias the shielding rings. The electronics are simple and can be easily contained within modestly sized shielding rings so that dielectric stresses on the electronic components themselves can be minimized (Figure 24.5). The efficiency of the multiplier decreases as the number of stages is increased. This can be mediated somewhat by using large capacitors in the first stages and decreasing the capacitance of each successive stage. Using fast recovery diodes can prevent excessive losses from reverse current leakage. The

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

(a)

729

Female banana jack connectors (at bottom) + Capacitor

C1

– C2

Diode Capillary spool Male banana jack connectors (sticking up and out of top)

C3

C4

Aluminum ring

(b)

(c)

FIGURE 24.5 (a) A schematic representation of the electronics in the individual modules that make up the Cockcroft–Walton voltage multiplier and shielding rings. The capillary spool can be seen in the photo (b), as well as the connections between modules (c). (Reprinted from Henley, W. H., PhD Dissertation, University of North Carolina, Chapel Hill, 2005. Copyright 2005. With permission.)

730

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

a.c. input frequency should be carefully selected for maximum efficiency of the Cockcroft-Walton generator.

24.2.2.3 HV Insulation When the electric field at a metal surface becomes greater in magnitude than the dielectric strength of the surrounding material, ionization of that material will occur. This phenomenon is called corona when it occurs in air. It is undesirable because it is a source of potential voltage loss and can lead to an electric arc. The strength of the electric field is dependent on the radius of curvature of the metal shielding ring. For every 30 kV of potential, the radius of curvature must increase 1 cm to prevent corona. Potentials of 500 kV require a radius of curvature of over 33 cm. Such large diameters can be avoided if material of sufficient dielectric strength and thickness is used to surround a more modest diameter. The first UHVCE instrument was surrounded by a tank of transformer oil. The oil proved difficult to work in. It absorbed moisture from the atmosphere and became slightly conductive, reducing the maximum attainable voltage. The second UHVCE instrument6 used thick coatings of plastics for dielectric insulation (Figure 24.5). All metal surfaces were painted with vinyl coatings. The electronic components were potted in thermoplastic polyethylene commonly available as “hot melt” glue. This material is inexpensive, has good dielectric properties, can be applied in thick coatings, and can be easily reworked with a scalpel and heat. Insulation between the shielding rings was provided by polyethylene caps. The insulation from the surrounding air was provided by an acrylic tube with many layers of polyester film. The polyester film had an exceptionally large dielectric strength (∼3 × 106 V/cm). Every effort was taken to prevent an ionizable path to ground. The high voltage end of the instrument housing was completely sealed with polyester/styrene casting resin except for the tiny injection port. The injection port was plugged with a polyetheretherketone (PEEK) injection rod (Figure 24.6) covered in silicone vacuum grease to exclude air from the seal.

24.2.2.4 Modes of Operation and Examples Open tube UHVCZE can be used for the separation of small molecules, peptides, proteins, and protein digests.6,7,12 Increasing the time that the analytes migrate will improve resolution and efficiency. One method of achieving this is to increase the length of the capillary, but when the maximum attainable voltage is limited, the electric field strength will be relatively low in long capillaries. UHVCE can improve these separations by maintaining a high field strength in a long capillary. Figure 24.7 shows two separations of a model peptide sample in a 566-cm-long capillary. The sample was first separated at 28 kV (close to the maximum conventional CE applied voltage) resulting in an electric field strength of only 50 V/cm. Leu-enk had an efficiency of 1 million theoretical plates and required approximately 328 min to migrate to the detector. Using the UHVCE power supply at 330 kV, the electric field strength was 580 V/cm. Leu-enk showed a 10-fold efficiency improvement at higher voltage, resulting in an efficiency of 10 million plates. Migration time was reduced to 28 min. Resolution between the analytes and impurity peaks can also been seen. In open tube UHVCE, it is important to reduce analyte adsorption as much as possible. Analytes that adsorb to the fused-silica surface do not tend to show much improvement as the length of capillary and applied voltage are increased. Figure 24.8 shows a model protein separation that demonstrated a small amount of adsorption in a 50 cm long separation. Using UHV in a 501 cm capillary, resolution and efficiency are actually reduced, and migration time is proportionally longer. Capillaries coated with aminopropyl species have been used to reduce the adsorption of peptides and proteins. UHVCE separations of peptides and proteins using these coating have been reported

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

Terminal module

(a)

Capillary

UHMWPE injection port

731

Alligator clip Buffer vial w/ electrode

0 Magnet Sample

Electrode connection Acrylic tube wrapped in polyester film

Cast polyester and styrene resin cap

PEEK injection rod External HVDC

(b)

0

FIGURE 24.6 Schematic diagram showing the operation of the injection port for sampling at low voltage. (Reprinted from Henley, W. H., PhD Dissertation, University of North Carolina, Chapel Hill, 2005. Copyright 2005. With permission.)

with UV absorbance and ESI-MS.6 Large improvements were seen in resolution and efficiency using UHVCE compared with shorter capillaries run at the same field strength at lower voltages. If the capillary is filled with a suitable sieving matrix, polymers such as nucleic acids6 and hyaluronic acid12,14 can be separated using UHV capillary gel electrophoresis (UHVCGE). Figure 24.9 shows four separations of MegaBASE™ 4 Color Sequencing Standard using different lengths of 75µm i.d., 360µm o.d. acrylamide-coated capillary filled with MegaBACE™ Long Read Matrix (Amersham Biosciences). The applied voltage ranges from −9 kV in the 60 cm capillary to –83 kV in the 5.5 m capillary to maintain a constant field strength of 150 V/cm. The time has been rescaled for the longer separations by a factor of X (given in the legend) so that the resolution improvements at higher applied voltages can be easily seen. Neutral analytes have been separated using UHV micellular electrokinetic chromatography (UHVMEKC).12,15 This was demonstrated by the separation of oligosaccharide mixtures in a surfactant-containing buffer. Higher resolution was achieved using UHV than standard running voltages, and plate counts over 1 million plates were seen.

24.2.2.5 Limitations Despite the advantages offered by UHVCE, several problems still limit its application. The lack of available instrumentation and unfamiliarity with voltages in excess of 30 kV keep many potential applications from being explored. The dangers involved with working at such high voltages are more perceived than real, and safety can be assured by using a grounded Faraday cage with interlocks designed to protect the operator from high voltage discharges. Also, the stored energy in most high voltage systems is usually very small, much less than 50 J for the current UHVCE system.

732

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a)

GlyAsn

Trp

1500 28 kV 1000 LIF signal

Leu-enk ~1 million plates

500

AlaGln

GlyLeuTyr

0 325

330

335

340

Min 3500

GlyLeuTyr

GlyAsn

Trp

(b) 3000 Leu-enk ~10 million plates

LIF signal

2500

330 kV AlaGln

2000 1500 1000 500

27.0

27.5

28.0

28.5 Min

29.0

29.5

30.0

FIGURE 24.7 Separations of a model peptide sample run on the same capillary at conventional and UHV. The 10-fold increase in electric field strength improves efficiency and reduces separation time 10-fold. Resolution increase can be clearly seen by the closely migrating impurity peaks. (Reprinted from Henley, W. H., PhD Dissertation, University of North Carolina, Chapel Hill, 2005. Copyright 2005. With permission.)

24.2.2.6 Methods Development Guidelines Part of the advantage of UHVCE is that it allows the use of known separation systems that work well for similar analytes but are unable to completely resolve the analytes of interest using conventional voltages and capillary lengths. Several factors relating to the buffer system and capillary should be taken into consideration before attempting to use a separation system for UHVCE.

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

2500

Ribonuclease A 470 kplates

Lysozyme 415 kplates UVabs@200 nm x0.002 AU

733

2000

Trpsinogen 270 kplates

1500

L = 501 cm 253 kV beta-lactoglobulin A and B 490, 240, 360 kplates

1000

500 4.4

4.6

4.8

5.0

5.2

5.4

5.6

Min Lysozyme 610 kplates

UVabs@200 nm x0.002 AU

1000

Ribonuclease A 320 kplates

800

Trpsinogen/ beta-lactoglobulin A and B (unresolved) 220 kplates

600 L = 501 cm 253 kV

400 200 0 –200

60

65

70

75

Min

FIGURE 24.8 Separation improvements are not realized in longer capillaries at higher voltages if analyte adsorption occurs. Shown here is a separation of proteins in short (a) and long (b) capillaries under the same electric field strength. Efficiency and resolution are actually reduced at UHV conditions due to analyte adsorption over the long capillary wall. (Reprinted from Henley, W. H., PhD Dissertation, University of North Carolina, Chapel Hill, 2005. Copyright 2005. With permission.)

24.2.2.7 Buffer Selection When selecting a buffer for UHVCE, the conductivity of the buffer must be as low as possible so that the electric field strength can be high without unacceptable Joule heating. The maximum field strength for a buffer/capillary system can easily be found by generating a current to field strength

734

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Note resolution improvements

5500

L = 60/ I = 50 cm (9kV) X = 100%

LIF signal

5000

L = 120/ I = 110 cm (18kV) X = 52%

4500 L = 184/ I = 174 cm (27.6kV) X = 31%

4000

L = 554/ I = 544 cm (83kV) X = 48%

40.0

40.5

41.0

41.5

42.0

42.5

Rescaled time (minutes*x)

FIGURE 24.9 Separations of DNAsequencing standard rescaled in the time domain by factor X for comparison of resolution improvements at higher applied voltages. Longer lengths (L) of capillary were used to maintain a constant 150 V/cm field strength. (Reprinted from Henley, W. H., PhD Dissertation, University of North Carolina, Chapel Hill, 2005. Copyright 2005. With permission.)

plot using a conventional power supply in a short piece of capillary. The onset of appreciable Joule heating can be found at the field strength where the plot begins to show nonlinear behavior. High concentrations of salt are used in some buffers to prevent adsorption of analyte molecules (peptides and proteins17 in particular) to the capillary wall. Zwitterionic molecules contain an equal number of positive and negative charges at a pH equal to their isoelectric point. They can be used to raise the ionic strength of the buffer without increasing the conductivity greatly. In UHVCE applications, they can sometimes be substituted for more conductive buffers to allow the use of higher field strengths while maintaining sufficient ionic strength to reduce analyte adsorption. It is important to note that the pI of the zwitterionic molecule may be several pH units different from their pKa values. In cases where this occurs, they cannot be relied upon to buffer the solution effectively. High concentrations of ions with small radii of hydration such as H3 O+ , K + , and OH– should be avoided. These small ions are rapidly transported within the linear electric field of the capillary, resulting in high electric currents. For this reason, ions with lower mobility such as Li+ and CH3 COO– should be substituted wherever possible.18 The buffer capacity can become important in separations of extended duration. For example, a CGE separation requiring 30 min on a 30 cm capillary will require 300 minutes on a 300 cm capillary if the same electric field strength is used. The 10-fold increase in analysis time results in 10-fold the amount of electrolysis products that must be dealt with by the buffer. The moles of OH– at the cathode can be calculated from the electrophoretic current (i):

it

1 = molOH− , F

(24.6)

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

735

where t is time in seconds and F is Faraday’s constant. A significant pH shift can cause problems with reproducibility. Large capacity buffer vials can be used if the instrumentation has enough available space. Buffer systems that do not adequately address problems of analyte adsorption will not be compatible with UHVCE in long capillaries.6 Buffers that contain ions with significantly different mobilities from the analyte ions will also result in electrodispersion that will be exacerbated by long separation distances. Both of these problems can be avoided by examining the shape of the analyte bands in short capillaries at lower voltages. Peaks that show fronting or tailing indicate potential problems, whereas highly Gaussian peaks indicate compatibility with UHVCE. 24.2.2.8 Capillary Selection The length of the capillary should be determined by the desired degree of improvement over a known, lower voltage, separation. For example, if twice the resolution of a known system using 30 kV in a 50 cm capillary is desired, then 120 kV should be used in a capillary 2 m long. This separation will also require fourfold the time. Joule heating can be avoided by simply reducing the power consumption of the capillary. Reducing the capillary diameter by half decreases the current fourfold. For the separation described above, a 50 cm length of capillary with half the diameter could be used at 120 kV without excessive Joule heating. In this case, the separation would require fourfold less time. The inner diameter of the capillary cannot be decreased too greatly if UV absorbance is used for detection. Below ∼50 µm, the pathlength is too short for sensitive detection. ESI-MS also requires a certain flow rate for optimal detection that may not be achievable in very narrow capillaries. LIF detection is well suited for use with narrow capillaries, but not every analyte can be labeled with fluorescent tags. 24.2.2.9 Sample Separation Strategies Unlike techniques that are limited to a narrow range of analyte properties, UHVCE is well suited for the analysis of complex samples containing analytes with a broad spectrum of properties (Figure 24.1). These samples include, among others, cell lysates, protein digests, or DNA sequencing preparations. The UHVCE analysis of samples should ideally start with a fairly concentrated sample containing low amounts of salt. High sample ionic strength can result in electrodispersion of the sample when used with low conductivity buffers. Dilute samples can be problematic if poor detection limits require excessively large sample injections.19 This may or may not be solved using sample stacking techniques. The best solution found thus far for the preventing peptide and protein samples from adsorbing to the capillary wall in UHVCE is electrostatic repulsion. This can be accomplished in bare silica capillaries by using very basic pH buffers where the deprotonated silanol groups and the majority of peptides and proteins are negatively charged. This method is not directly compatible with positive mode ESI-MS, but may work for UV absorbance detection or LIF.17 Various coatings that impart a positive charge to the surface can be used for low pH separations. Covalently bound coatings such as aminopropyltrimethoxysilane have been used with good success for UV detection and ESI-MS of peptides.20,21 Other coating schemes based on the same principles have been used to varying degrees of success, and diisopropylaminopropylethyoxysilane has shown remarkable stability to degradation from hydrolysis.6,22 Cationic polymers such as PolyE-32323,24 have been used to great effect for CE/ESI-MS for proteins and peptides. These coating are easily applied to long capillaries and stable in buffers with organic modifiers. The analysis or sequencing of DNA in CE and its advantages over standard slab gels has been well documented.25–40 Preliminary study of UHVGCE for DNA separations indicates that it may be best suited for high-resolution separations of relatively short sequences of DNA.6

736

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Polysaccharides and other complex carbohydrates can be difficult to analyze for several reasons. The lack of charged moieties requires the use of a sieving gel or pseudo-stationary phase. Sieving gels can be quite useful for highly branched structures found in some of these analytes. UHVCGE or UHVMEKC has been shown to be useful for high-resolution analysis of these samples.12,14,15

24.3 FLOW COUNTERBALANCED CAPILLARY ELECTROPHORESIS 24.3.1 BACKGROUND The effectiveness of capillary electrophoretic separations can be greatly affected by the magnitude and direction of the bulk fluid flow. As pointed out by Jorgenson and Lukacs,4 analyte bands that migrate in the direction opposite to the EOF spend a greater amount of time in the electric field and thus are more affected than analytes that migrate with the EOF. Direct manipulation of the EOF or application of hydrodynamic (pressure driven) flow can be used to retard the migration of analytes to improve efficiency and resolution. Resolution improvement by the direct control of the EOF through the manipulation of radial electric fields has been shown at low pH and low ionic strength.41 Cheng et al.42 noted the potential of a counter flow to improve the resolution of analytes in a paper investigating the effect of a constant applied pressure from a sheath flow cuvette on separations of amino acids during CE separations. Pressure driven flows have also been used for the preconcentration of samples for capillary isotachophoresis with long preconcentration times in short capillaries.43–49 Reduction of the EOF using surfactants and polymers to increase the viscosity of background electrolyte near the wall has also been demonstrated for resolution improvement.50–55 Lucy and McDonald51 used this technique to resolve the major isotopes of chlorine using CE. Several preparative scale purification techniques using a counter flow have been reported for enantiomers and other closely related species.56–63 Another study used a counter flow to prevent buffer components that improve selectivity but increase the UV absorbance background signal from reaching the detector.64 A practical system for analytical scale high-resolution separations was first demonstrated by Culbertson and Jorgenson.65 Narrow diameter capillaries and LIF detection were used to resolve analytes with extremely small differences in mobility such as peptide isomers. Misleveling of the buffer reservoirs, using a siphoning action to provide a counter flow in large diameter capillaries, was also reported with modest resolution improvements.66 Band broadening caused by the counter flow limited high-resolution FCCE to narrow bore capillaries and LIF detection for over a decade. Recently, the use of large diameter capillaries packed with bare silica particles to flatten the parabolic flow profile of the counter flow has been demonstrated.67 Packed capillary FCCE should not be confused with capillary electrokinetic chromatography (CEC). Packing material used in FCCE serves no chromatographic purpose and does not act as a stationary phase. The purpose of the packing material is simply to prevent band broadening in large bore capillaries so that UV absorbance detection can be used to detect unlabeled analytes.

24.3.2 THEORETICAL ASPECTS The power of FCCE lies in the ability to maintain analyte bands within a high electric field strength for a long period of time. A counter flow is applied that pushes the bulk solution in the direction opposite to the analyte’s migration, extending the duration of the separation. For a given length (L) of capillary with applied voltage (V ), the electric field strength (E) is simply

E=

V . L

(24.7)

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

737

The time (tm ) required for an analyte band to migrate a distance equal to the length of the capillary (L) can be determined from the following equation: tm =

L , E(µe + µEOF )

(24.8)

where µe is the analyte’s electrophoretic mobility and µEOF is the mobility of the EOF. Suppose that a counter flow is applied opposite to the direction of the analyte’s net migration so that the analyte band now requires fourfold the time to transverse the same length of capillary. The electric field strength, length of capillary, and mobilities of the analyte band and EOF remain unchanged, but now the effective distance (Leff ) traveled by the analyte band is equal to 4L. Neglecting any band broadening, an equivalent separation without a counter flow would require a capillary four times as long. In order to obtain the same electric field strength, four times the voltage would have to be applied. This can be thought of as the effective voltage (Veff ) that can be determined from the known electric field strength: E=

V Veff = . L Leff

(24.9)

Since the resolution between two analyte bands scales with the square root of the applied voltage, fourfold the effective voltage would generate twofold the resolution. For example, consider two analyte bands that transverse a 30 cm capillary in an average time of 30 s with 30 kV applied. If the resolution between the bands is 0.5, then a capillary 120 cm long with 120 kV applied would be needed to obtain a resolution of 1.0. The same resolution can be obtained using the 30 cm capillary with 30 kV applied if a counter flow is applied so that the analytes transverse the same distance in 120 s. EOF is generated at the wall of the capillary, resulting in a flat flow profile that limits band broadening to that caused by longitudinal diffusion. This broadening can be expressed as a variance  2 σB using Einstein’s equation: σB2 = 2Dt,

(24.10)

where D is the diffusion coefficient of the analyte and t is time. The application pressure results in a hydrodynamic flow. Flow velocity is greatest in the center of the capillary due to drag at the capillary wall and a parabolic flow profile is the result. Analyte bands diffuse laterally across this flow profile and broaden to a degree proportional to its magnitude (Taylor dispersion). Culbertson and Jorgenson 65 derived an equation for the degree of band broadening for FCCE in an open tube. In addition, the maximum allowable diameter of the capillary was determined from the following expression: σc2 =

2 t dc2 νpa

96D

≤ 2Dt = σB2 ,

(24.11)

where the broadening of the analyte bands from hydrodynamic flow is expressed as a variance, σC2 , which should be less than or equal to the band broadening caused by longitudinal diffusion alone. The diameter of the open capillary (dC ) can be calculated from the average velocity of the hydrodynamic flow (vpa ) required to counterbalance the electrokinetic migration of the analyte band and the diffusion coefficient of the analytes (D). For the analytes investigated, open tubular capillaries 5–10 µm were deemed acceptable.65,68 Narrow bore capillaries have a limited optical pathlength and LIF detection is usually required for sensitive detection. Band broadening from hydrodynamic flow in packed capillaries is independent of capillary diameter, allowing the use of capillaries with a pathlength large enough for UV absorption detection.

738

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

A similar equation for band broadening in packed capillary FCCE can be derived, as can a similar method of calculating the required particle diameter (dp ) to minimize band broadening.6,67 Briefly, 2 ), which is the sum of the contributions of the total broadening is expressed as a variance (σtotal longitudinal diffusion and hydrodynamic flow: 2 σtotal

=

σB2

+ σC2

= 2Dt +

2 t dp2 νpa b

10D

,

(24.12)

where tb is the time during which the counter flow is applied. Using the same strategy as expression 11, the maximum particle diameter can be calculated. The efficiency for FCCE in packed capillaries can be found using the following equation: N=

10D(tνek )2 , 2 t 20D2 t + dp2 νpa b

(24.13)

where vek is the average electrokinetic velocity for the analyte bands and N is the efficiency in theoretical plates.6,67 High efficiencies can be most easily obtained by using small diameter particles and long separation times.

24.3.3 PRACTICAL IMPLEMENTATION 24.3.3.1 Instrumentation The instrumentation required to perform FCCE is relatively simple. The power supply can be any commercially available HVDC power source, but an ideal power supply would have dual polarity, 0–30 kV at several 100 µA, with safety interlocks and computer control capabilities. The maximum voltage needed is determined by the desired electric field strength and length of the capillary. In order to allow for some margin of error in control of the counter flow, the minimum length of capillary should be at least three to four times the physical width of the desired separation window. These margins allow for the analyte bands to be maintained within the inner half of the capillary so that they are not pushed out on either side of the capillary, with detection occurring in the center of the capillary. The source of the counter flow, or pumping system, is determined by the velocity of the analyte band and the backpressure required to reverse or stop its forward migration. Where the analyte bands migrate slowly in an open tube, the pumping system may be as simple as a misleveling of the buffer vials. Rapidly migrating analyte bands in narrow open capillaries may require a compressed gas driven system such as that shown in Figure 24.10.65,68 Packed capillaries require much greater pressures and so an high-performance liquid chromatography (HPLC) pump or syringe pump capable of generating the desired fluid velocity must be used (Figure 24.11).6,67 If an HPLC pump is used, some consideration should be given regarding the volumetric flow rate needed to maintain a stable pressure. Typically, flow rates for HPLC columns range in the mL or µL per minute range. The flow rate for FCCE will be much smaller, typically nL/min. The use of a “splitter” capillary to divert the majority of the flow volume back to the pump’s buffer reservoir can save many liters of running buffer. This may be especially important for buffers containing expensive reagents. Syringe pumps have limited capacity and the flow rate will therefore determine the maximum separation time. One notable exception to this is the Haskel International Inc. (Burbank, CA) DSHF-300 Air Driven Fluid Pump. This is a syringe-style hydraulic pump capable of producing up to 50,000 psi of liquid pressure via a 300-fold amplification of applied gas pressure. The advantages of this pump include its stable applied pressure over a volumetric flow rate ranging from no flow at all to a few mL per minute. It has a low stoke volume of a few mL, but it automatically recycles and refills itself at the end of a piston stoke, maintaining a fairly constant pressure with built in check valves.

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

739

High voltage power supply LIF detection system

Pressure reservoir

Cutaway view Interlock box He gas pressure regulator

Microcomputer

Electronically actuated pneumatic valve

FIGURE 24.10 Open tube FCCE using a gas pressure driven pumping system. (Reprinted from Culbertson, C. T. and Jorgenson, J. W., Analytical Chemistry 1994, 66, 955–962. Copyright 1994 American Chemical Society. With permission.)

Syringe clean out port

High voltage power supply

Manual valve UV detector

Capillary

PEEK tubing

Gas in (low pressure)

Waste

Capillary

316SS T-fitting Pneumatically actuated valve

PEEK tubing coil 100 k
current monitoring resistor Buffer reservoir

Buffer out (p = 300 X Gas pressure)

Ground

Haskel DSHF-300 air driven fluid pump

Waste Buffer reservoir

FIGURE 24.11 Packed capillary FCCE using high pressure from a hydraulic pump controlled by a pneumatically actuated valve. (Reprinted from Henley, W. H., et al., Analytical Chemistry 2005, 77, 7024–7031. Copyright 2005 American Chemical Society. With permission.)

Ideally, the velocity of the counter flow should be exactly equal and opposite to the velocity of the analyte band so that band broadening from pressure driven flow is minimized. The practical implementation of such a system requires a detector capable of imaging the entire width of the separation window such as a linear diode array or charge-coupled device (CCD) camera.

740

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Single point detectors using UV absorbance or LIF detection require that the analyte bands pass though the detection point in order for data to be collected. Therefore, analytes are first allowed to migrate past the detection point electrokinetically. Then a pressure great enough to push them backwards past the detector is applied via a computer-controlled valve. Once the analytes have been sufficiently pushed back, the pressure is released and the cycle is repeated. The electrophoretic voltage is maintained during the entire course of the separation, so the analytes are continuously separating electrokinetically, even while being pushed backwards by the counter flow. An alternative scheme using constant pressure with intermittently applied voltage would only resolve the analytes while the voltage was applied. Valves for controlling the back pressure must actuate quickly, be rated to the desired pressure, and ideally be computer controlled. As long as the pressure remains constant, dead volume is usually not a major concern. If the valve is constructed from metal, it must be electronically isolated from the electrophoretic system or else electrolysis processes may corrode and destroy the valve, in addition to ruining the separation. Several suitable valves are manufactured by Valco Instrument Co. Inc. (VICI) (Houston, TX). When making the connection between the capillary and pumping system, it is very important to consider the flow of electrophoretic current and the effect it may have on components of the system. Figure 24.12 shows two ways of connecting a packed separation capillary to the highpressure pumping system. A stainless steel Swagelok® compression fitting forms a union between the fused-silica capillary and a piece of PEEK tubing. The configuration in Figure 24.12a results in electrolysis products entering the capillary and spoiling the separation. Even though the metal fitting is electrically floating, it is not isolated from the electrophoretic current. Current flows as ions through the capillary and through the PEEK tubing, but the highly conductive metal fitting offers a low resistance path for electrons to travel. The result is that electrolytic processes occur at the surface of the metal near the exit of the capillary and the entrance to the PEEK tubing. An easy way to avoid this is to simply push the capillary into the PEEK tubing so that the metal fitting is truly electrically isolated from the electrophoretic current (Figure 24.12b). 24.3.3.2 Capillary Packing Packing the capillary with nonporous bare silica can greatly reduce the amount of band broadening observed at similar counter flow linear velocities. Figure 24.13 shows the temporal variance of an Electrolysis occursat both metal/buffer junctions

Capillary

Plastic ferrule

PEEK tubing connected to grounded Tfitting

PEEK tubing connected to grounded Tfitting

Capillary

Steel ferrule

Stainless steel Swagelok ® 1/16”union

Plastic ferrule

Steel ferrule

Stainless steel Swagelok ® 1/16”union

FIGURE 24.12 In order to prevent electrolysis products from entering the capillary at the connection of the capillary to high-pressure counter flow system, metal surfaces must be floating and isolated from the path of the electrophoretic current. (Reprinted from Henley, W. H., PhD Dissertation, University of North Carolina, Chapel Hill, 2005. Copyright 2005. With permission.)

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

741

Temporal variance (sec2)

100 50 µm OTC 50 µm PBC

80 60 40 20 0 0.0

0.1 0.2 0.3 Counter flow linear velocity (cm/sec)

0.4

FIGURE 24.13 A comparison between the band broadening seen in a pack capillary and open tubular capillary under similar counter flow linear velocities shows much less broadening in the packed capillary. (Reprinted from Henley, W. H., et al., Analytical Chemistry 2005, 77, 7024–7031. Copyright 2005 American Chemical Society. With permission.)

R-(−)-epinephrine peak using a 50 µm open tubular capillary (OTC) and a 75 µm capillary packed with 4.1 µm nonporous bare silica particles. The diameters of the capillaries were chosen so that they would have approximately the same dead volume and electrophoretic current. It is easy to see from the plot of temporal variance versus counter flow linear velocity that much less broadening is seen in the packed capillary as compared with the large diameter open tube. The preparation of the packed capillary for FCCE using UV absorbance detection is not difficult, but it differs in several aspects from the methods used to prepare capillary HPLC or UHPLC columns.69 For chromatography, capillaries are typically run with mobile phases flowing in the same direction as the capillary was packed. For FCCE, the alternative application and release of pressure flowing opposite to the flow of the EOF causes large voids to develop within the packed bed. Consolidation of the capillary can be achieved via sonication under hydrodynamic flow.70,71 Briefly, dry silica particles are tapped into one end of an OTC. Heat from an electric arc is used to sinter them together and to the capillary wall. Pressure (up to 50,000 psi) from packing bomb is used to force a ∼10 mg/mL slurry of packing material into the open end of the capillary. After the bed is packed, an inlet frit is sintered at the head of the bed using a heated tungsten wire. The pressure is released, and the capillary is cut to length. The capillary, except for the fritted ends, is placed into a water bath and sonicated for ∼5 min. Pressure is then applied to the capillary in the direction opposite to that in which it was originally packed. A ∼5–10% reduction in the length of the packed bed should be observed. A new frit is then sintered using a heated tungsten wire while the capillary is still under pressure. Sonication can be repeated with pressure applied from the opposite end if further consolidation is desired. Figure 24.14 shows how the consolidation of the packing material reduces band broadening caused by the counter flow. The same capillary was used for measurements of an R-(–)-epinephrine peak using 3.1 µm nonporous bare silica packing.6,67

24.3.3.3 Modes of Operation and Examples Open tube FCCE can be performed in capillaries of relatively large diameter if the analyte bands migrate slowly and only a modest improvement in resolution is desired. Misleveling of buffer vials can provide a simple way of improving the resolution for analyte bands that are almost, but not quite resolved under standard conditions.66 Band broadening will occur but usually not to a great extent over short durations.

742

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

350

Temporal variance (sec2)

300

Before sonication After sonication one side of window After sonication both side of window

250 200 150 100 50 0 0

5

10

15

20

25

30

35

40

45

50

Migration time (min)

FIGURE 24.14 Reducing the interparticle porosity of the capillary packing can decrease the amount of band broadening caused by the counter flow. Data taken from migration of R-(−)-epinephrine using 3.1 µm nonporous bare silica packing. (Reprinted from Henley, W. H., et al., Analytical Chemistry 2005, 77, 7024–7031. Copyright 2005 American Chemical Society. With permission.)

Open tube FCCE in narrow capillaries using zone electrophoresis is useful for resolving analytes that can be tagged with fluorescent labels for LIF detection. Culbertson72 demonstrated resolution of the peptide fragment YAGAVVNDL and YAGAVVNDI (Figure 24.15), which only have an electrokinetic mobility difference of 1.0033. Open tube FCCE using micellular electrokinetic chromatography (MEKC) can be used to obtain high resolution for charged or neutral analytes. High-resolution separations of TRITC-labeled phenylalanine with different numbers of hydrogen atoms replaced with deuterium atom showed resolution between analytes with differences of a single neutron (0.16% mass difference). The relative levels of deuterium substitution could also be determined from the peak areas. The ability of FCCE with MEKC to separate these isotopomers is believed to be based on slight differences in hydrophobicity caused by the influence of the heavier deuterium atom.68 Packed capillary FCCE can be used with larger diameter capillaries with sufficient pathlength for sensitive UV/Vis absorption detection. UV detection allows analysis of nonfluorescent analytes and sample impurities. Most organic and some inorganic compounds absorb UV or visible light to some degree, and therefore packed capillary FCCE can be used for a much larger range of analytes than open tubular FCCE. Separations of racemic samples of several basic and acidic pharmaceuticals using packed capillary FCCE have been reported previously.6,67 Figure 24.16 shows the separation of the enantiomers of fenoprofen using β-cyclodextrin as the chiral pseudo-stationary phase. The top portion of the figure (a) shows the entire separation. The analyte peak is allowed to electrokinetically migrate pass the detector (forward pass) and then pushed backwards under hydrodynamic flow (reverse pass). Seven cycles with forward passes (F1–F7) and six reverse passes (R1–R6) can be seen in Figure 24.16a. Figure 24.16b shows the first electrokinetic pass in front of the detector. This part of the electropherogram represents the resolution (0.34) and efficiency (21,000 plates) that would result from traditional CE under these separation conditions. Figure 24.16c shows the final pass in front of the detector, where the analyte bands have broadened by a small degree, but are approaching baseline resolution with greatly increased separation efficiency (320,000 plates). Table 24.1 gives a list of acidic and basic pharmaceutical compounds separated with packed capillary FCCE. Electrokinetic only migration times, total number of passes, along with resolution and efficiency improvements are given.

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

(a)

743

(b)

Time (min) (c)

Time (min) (d)

Time (min)

Time (min) Velocity ratio = 1.0033

FIGURE 24.15 Open tube FCCE with LIF detection showing the resolution of YAGAVVNDL and YAGAVVNDI. (Reprinted from Culbertson, C.T., p. 264, 1996, University of North Carolina, Chapel Hill, NC. Copyright 1996. With permission.)

UVabs signal (AUFs x 106)

(a)

Time (min) (c)

UVabs signal (AUFs x 106)

UVabs signal (AUFs x 106)

(b)

Time (min)

Time (min)

FIGURE 24.16 Packed capillary FCCE electropherogram showing the separation of fenoprofen enantiomers using β-cyclodextrin as a chiral selector. (Reprinted from Henley, W. H., et al., Analytical Chemistry 2005, 77, 7024–7031. Copyright 2005 American Chemical Society. With permission.)

744

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

TABLE 24.1 Racemic Mixtures Resolved Using FCCE in Packed Capillaries Using β-Cyclodextrin Degree of Resolution Efficiency of Migration Time Total Run Total No. of 1st Pass/Last Pass at 1st Pass/at (min) for 1st Time (min) Forward Passes (103 Plates) Last Pass Forward Pass Epinephrinea 40.9 329 8 0.42/1.6 31/213 Norepinephrinea 66.7 275 7 0.36/1.02 14/77 Synephrinea 117 138 2 0.988/1.02 15/25 Norphenylephrine HCla 62.1 147 4 1.13/1.76 35/68 Phenylpropanolaminea 71 195 5 0.21/0.77 9/41 Octopaminea 61.07 403 10 ND/0.55 8/45 Chlorpheniramine maleatea 32.8 118 3 2.12/3.35 90/240 Doxylaminea 23.5 202 6 1.59/3.38 120/820 Ibuprofenb 34.1 259 9 0.496/1.08 44/550 Ketoprofenb 39.8 688 22 ND/0.88 10/340 Fenoprofenb 35.7 242 7 0.34/0.88 21/380 Compound

ND = none detected. a Basic chiral analytes were separated in 50 mM phosphoric acid, 10 mM β-cyclodextrin 0.5% triethylamine. bAcidic compounds were analyzed using a 100 mM MES, 10 mM β-cyclodextrin buffer system.

Source: Reprinted from Henley, W. H., et al., Analytical Chemistry 2005, 77, 7024–7031. Copyright 2005 American Chemical Society. With permission.

Several isotopomer separations have been reported for packed capillary FCCE. Deuteriumsubstituted phenylalanine showing no resolution during the first forward pass was almost baseline resolved after an 11 h separation time.67 Figure 24.17 shows a packed capillary FCCE separation of bromine-79 and bromine-81 using UV absorbance detection. After 6 h of run time (Figure 24.17b), a distinct notch is seen in the bromide ion peak. The separation was maintained for over 60 forward passes requiring over 1100 min, and clear separation of the isotopes can easily be seen.6,67 Before attempting MEKC in packed capillaries, an examination of the properties of the micelles under high pressures should be considered. Unpublished work suggests that at the high pressures required for reversing the analyte bands, the properties of the micelles may change, resulting in an unstable system.

24.3.4 METHODS DEVELOPMENT GUIDELINES 24.3.4.1 Handling Extended Separation Times The high resolution and efficiency obtained in FCCE separations comes at the expense of extended separation time. Although the buffer within the capillary is replenished by the counter flow,60 care should be paid to ensure sufficient buffer capacity for reservoirs with fixed volume. Measurements of the electrophoretic current can be used to calculate the expected pH change in each buffer reservoir for a given separation time. In addition, steps should be taken to prevent electrolysis products from entering the capillary. Figure 24.11 shows a long coil of PEEK tubing connected between the capillary and the grounded electrode. This coil has sufficient length and volume to prevent gas bubbles or UV absorbing electrolysis products from entering the separation capillary for many hours under typical counter flow rates. Automation of the valve control using a computer program is not only useful for ensuring that the analyte band stays within the capillary but also lets the operator leave the instrument running

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

745

(b)

(a)

200 600 150 400

100 50

UV abs signal (AUFs x 106)

200

0 0

–50 6

7

8

9

360

10

(c)

362

364

(d) 100

200

80 150

60 40

100

20 0

50

–20

0

–40 738

740

742

744

1112

1113

1114

1115

1116

Time (min)

FIGURE 24.17 A separation of bromine isotopes using packed capillary FCCE. (Reprinted from Henley, W. H., et al., Analytical Chemistry 2005, 77, 7024–7031. Copyright 2005 American Chemical Society. With permission.)

autonomously for hours at a time. There are many different ways to automate the control of the system, but a few basic principles should be followed for reliable operation. After injection of the sample, the electrokinetic migration velocity of the analyte band can be determined by the time it takes to travel the distance to the detector. After allowing the band to travel some distance beyond the detector, the counter flow can be applied (while maintaining the applied voltage). The time required for the analyte band to be pushed back in front of the detector can be used to calculate the analyte band velocity under hydrodynamic flow. These two parameters can be used to calculate the position of the analytes at any given point in the FCCE separation cycle. The progress of the separation should be checked every few hours to correct any timing errors. Adjustments will have to be made for long separations due to variations in pump pressure, EOF velocity, and buffer viscosity caused by temperature changes or buffer depletion. 24.3.4.2 Band Broadening Prevention One of the main advantages of FCCE is that published buffer conditions that may not completely resolve certain analytes in a traditional CE system can be used to fully resolve the analytes by giving them more time to separate. This can save method development time when analyzing a variety of

746

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

compounds with similar characteristics. The main limitation on the timescale of an FCCE experiment is the degree to which band broadening reduces the analyte signal strength, and therefore the counter flow velocity should be kept as low as possible. Adsorption of the analyte to the capillary wall or bare silica packing material must also be prevented to avoid signal loss. Electrostatic repulsion works well for negatively charged species in basic buffer with bare capillaries. Amine or other basic analytes can be analyzed at low pH in bare capillaries if a competitive inhibitor such as triethylamine is added to the running buffer. The use of dynamic coatings such as polymers or surfactants in the running buffer to reduce adsorption of analytes is probably best suited to open tubular FCCE as increased viscosity may result in high back pressures in packed capillaries. Permanent, covalently bound coating should be used only after their long-term stability has been confirmed, and analyte band migration time should be closely monitored. Electrodispersion can be a major source of band broadening in FCCE due to the long separation times. Close matching of the electrophoretic mobilities of the analyte bands and running buffer can reduce electrodispersion greatly. Gaussian-shaped peaks usually indicate a compatible sample/buffer whereas fronted or tailed peaks indicate potential problems. Furthermore, the ionic strength of the running buffer should be closely matched to that of the sample. In addition, in packed capillary FCCE, care must be taken to ensure that the packing material is fully consolidated. Voids in the packed bed will dramatically increase the observed broadening as seen in Figure 24.14. 24.3.4.3 Determining the Best Separation Strategy The first thing to consider when designing an FCCE separation strategy for a particular sample is the desired method of detection. Both open tubular FCCE and packed capillary FCCE have their advantages and drawbacks. Open tube FCCE uses a slightly less complicated instrumental setup (Figure 24.10) than packed capillary FCCE (Figure 24.11). While not difficult, the equipment and skills needed to pack capillaries with bare silica particles can represent a considerable investment. The high sensitivity of LIF detection makes open tube FCCE the method of choice for analytes that can be easily labeled with fluorescent tags such as peptides and protein fragments. FCCE in packed capillaries can be used for many more types of analytes. Direct or indirect UV/Vis absorbance detection can be used to detect everything from monoatomic ions like Br– to basic and acidic pharmaceuticals. Pseudo-stationary phases can be used to improve the resolving power of either type of FCCE. MEKC works well in narrow diameter open tubes for separating neutral analytes based on their hydrophobicity differences. Chiral selectors such as cyclodextrins can be used with either method. Other methods of detection such as noncontact conductivity and modes such as CEC have yet to be explored but may prove useful in the future.

24.4 SYNCHRONOUS CYCLIC CAPILLARY ELECTROPHORESIS 24.4.1 BACKGROUND Synchronous cyclic CE, as the name implies, uses a series of electrodes placed at regular intervals around a closed loop to “chase” analyte bands around a continuous separation channel. Voltage switching is synchronized to maintain analytes of a particular mobility within the separation channel for a long period of time.73 A similar concept had been applied in chromatography using two chromatographic columns and timed valve changes to continuously separate analyte bands.74 The problems connecting separation channels on the extremely small scale of capillary electrophoretic systems prohibited the implementation of SCCE until several years later.75

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

747

24.4.1.1 Microchip SCCE Burggraf et al.75–77 reported the first practical SCCE system using microfluidic technology. Photolithographic techniques can be used to cheaply manufacture microfluidic structures with channels one to several hundred microns in dimension. The interconnection of these channels can be used to create networks with negligible dead volumes. The first pattern used for SCCE was four overlapping, 2-cm channels arranged in a square. An additional channel intersecting the middle of a main channel was used to perform electrokinetic injections. Several other designs have since been reported by Manz’s group.78,79 Figure 24.18 shows the layout for an SCCE that was used to separate fluorescently label amino acids and human urine.79

24.4.1.2 Capillary SCCE The use of polyimide coated fused-silica capillaries for SCCE with low dead volume connections was first reported by Zhao et al. in 1999.80,81 Fused-silica capillaries offered several advantages over the microchip-base SCCE systems. The capillaries provided longer separation distances for the analyte bands. While the geometry of the connections on a microfluidic device cannot be easily changed during an experiment, the use of adjustable gaps between the capillaries provides very low dead volume. In addition, conventional UV-Vis absorption detection was feasible for SCCE for the first time using capillaries 50 µm in diameter. A basic schematic representation of the instrumental setup is shown in Figure 24.19.80 Another device has been reported, the “electrophoretron” that used two capillaries with different polarity surface treatments. The capillaries were connected in a loop and a single power supply was used to apply a fixed potential across the two capillaries. The different polarity surface treatments cause the EOF to flow in a continuous loop around the capillaries. Injected analytes would continuously travel around the loop under the influence of the EOF.82 Such a device may prove useful for CEC separations of neutral analytes. However, any separation of charged species in one capillary will be almost negated in the second capillary due to the change in the polarity of the electric field.73

6

7 8

5 Sample Volume defined injection scheme 4

9

2

3 SW 2.0 cm 5.0 cm

FIGURE 24.18 Microchip-base SCCE system used with MEKC to separate FITC-labeled amino acids and human urine components. (Reprinted from von Heeren, F., et al., Analytical Chemistry 1996, 68, 2044–2053. Copyright 1996 American Chemical Society. With permission.)

748

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

UV power supply

3 UV detector

2

4 1 Capillary

Cross-flow cell

Polycarbonate boards

FIGURE 24.19 Schematic representation of capillary-based SCCE system used to resolve isotopomers and enantiomers. (Reprinted from Zhao, J., et al., Journal of Microcolumn Separations 1999, 11, 431–437. Copyright 1999 John Wiley & Sons, Inc. With permission.)

24.4.2 THEORETICAL ASPECTS 24.4.2.1 Absolute versus “Effective” Voltage The electric field strength is determined by the applied potential difference divided by the distance over which it is applied. SCCE is similar to FCCE in that the effective separation voltage can be much higher than the actual, physically applied voltages. For a single cycle, the effective voltage is twice that of the applied voltage. For an analyte band transversing n cycles, the effective voltage (Veff ) is Veff = 2nVapplied ,

(24.14)

where Vapplied is the applied voltage. Using Equation 24.2, an equation for the expected resolution between two analyte bands for a given applied voltage can be derived: Rs = 0.177(µ1 − µ2 )




2nVapplied D((µe + µEOF

1/2

.

(24.15)

24.4.2.2 Band Broadening Mechanisms There are several sources of band broadening in SCCE to consider. Longitudinal diffusion occurs in all CE separations. As seen in Equation 24.10, it is mainly governed by separation time and the analyte’s diffusion coefficient. Turns in the separation channel create another source of band broadening.83–87 As the analyte bands migrate around a turn, the portion of the band on the inside of the turn travels a shorter distance than the portion of the band on the outside of the turn. In addition, the electric field is concentrated near the wall of the inside of the turn, increasing the rate of band broadening. Taylor dispersion then

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

749

smoothes out the distorted band shape. The dispersion created by a turn in the separation channel can be described as a variance 2 σturn =

(l)2 (2θw (1 − exp (−tD /tt )))2 = . X X

(24.16)

where l is the difference in length traveled around the inside and outside of the turn, θ is the angle subtended in radians, w is the width at the top of the separation channel, tD is transverse diffusion equilibrium time, tt is the turn transit time, and X is a constant ∼ =12 when tD is large.83 The smaller radius of curvature found in microchip-based channels exacerbate this phenomenon, which is typically not a concern in capillary-based SCCE. Broadening at the connections (corners on microchip-based SCCE) between the separation channels is another important source of band broadening and sample loss in SCCE. The total broadening 2 can be expressed as a variance, σtotal 2 = σtotal



σi2 +

i=sides



2 2 σj2 = σmigr.dist + nσcorner ,

(24.17)

j=corners

2 is the variance from the total where n is the number of corners the band travels through, σmigr.dist 2 migration distance, and σcorner is the broadening caused each time a band travels through a corner or connection of channels.75

24.4.3 PRACTICAL IMPLEMENTATION 24.4.3.1 Power Supply and Voltage Switching Power supplies for microchip-based SCCE and fused-silica capillary-based SCCE mainly differ in the number of power supplies and in the magnitude of the applied voltage. The small size found in most microchip-based SCCE systems limits the separation channel length and so relatively small potential differences are needed to generate high electric field strengths (∼2 kV/cm). In contrast, capillary-based SCCE use approximately 50 cm lengths of capillary, and so power supplies generating up to 30 kV and a few hundred microamperes are typically used. The configuration of the separation channels determines the number of HV electrodes needed. The actual number of HV supplies needed can be reduced if relays are used to switch the potential between different electrodes. The fused-silica capillary-based SCCE instrumentation reported used a single 30 kV HV supply with four HVDC relays.80 The first chip-based SCCE system used four HV power supplies at 2.5 kV and eight HV relays to control the voltage switching.75,76 Later work with chip-based SCCE used six 10 kV HV supplies and thirteen HV relays.79 Different geometries with shorter capillary channels were also studied. A triangular arrangement resulted in the fewest HV supplies needed.78 24.4.3.2 Connection of Fused-Silica Capillaries While the connection of the separation channels used in microchip-based SCCE can be manufactured with almost no dead volume using photolithographic techniques, connection of fused-silica capillaries is challenging. Zhao et al.80,81 report the use of a controllable gap using a cross flow connection and a tight fitting Teflon sleeve (Figure 24.20). The gap between the capillaries offers several advantages over other methods. In order to allow the flow of electrophoretic current, the gap is opened, which also allows fresh buffer to flow into the capillary network and old buffer to be removed. While the analyte bands migrate through the gap, the gap is tightly closed, reducing the dead volume dramatically. This system can be implemented with solenoid, piezoelectric, hydraulic, or even manual actuation. Careful preparation of the capillaries is required to ensure a tight, low loss

750

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

to HV

Electrode Buffer PEEK tube

Clamp

Teflon sleeve

Capillary Cross-flow cell

Lever Pull solenoid

Pull spring

Micropositioners Buffer

FIGURE 24.20 Schematic diagram detailing the connection of the capillaries and the controlled gap for capillary-based SCCE. (Reprinted from Zhao, J., et al., Journal of Microcolumn Separations 1999, 11, 431–437. Copyright 1999 John Wiley & Sons, Inc. With permission.)

fit. Capillaries were cut to precisely equal lengths so that E fields and flow velocities would be equal in each length of the four capillaries. The ends of the capillaries were polished flat using a jeweler’s lathe and fine grit sandpaper to ensure a perfectly flat mating surface.80

24.4.3.3 Modes of Operation and Examples Capillary zone electrophoresis with microchip-based SCCE has been used to separate fluorescent dye from degradation products88 and to separate fluorescein isothiocyanate (FITC) labeled amino acids.78 Capillary zone electrophoresis (CZE) in fused-silica capillaries has been used to separate the racemic mixtures of (α-hydroxybenzyl)methyltrimethylammonium and (2-hydroxy1-phenyl)ethyltrimethylammonium with β-cyclodextrin as the chiral pseudo-stationary phase.81 l-Phenylalanine and l-phenylalanine-ring-D5 (the hydrogens in the aromatic ring were substituted with deuterium)81 and another separation of the closely related amino acids phenylalanine and tyrosine have been resolved using CZE with capillary-based SCCE.80 MEKC has been implemented in both microchip-based SCCE separations and capillary-based SCCE separations. Figure 24.21a shows a microchip-based SCCE separation using MEKC of FITClabeled amino acids.79 The same chip was also used to resolve FITC-labeled components of human urine using MEKC in addition to performing an MEKC-based immunoassay of serum theophylline levels.79 MEKC using capillary-based SCCE has resulted in separations with as many as 100 million theoretical plates in 15 h of run time. A separation of l-phenylalanine and l-phenylalanine-ring-D5 is

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

(a)

asn phe arg gln phe asn ser gly

gly

ser

gln

Fluorescence

1 1/4 cycles

60

50

*

1/4 cycle

80

40

80

70

asn phe ser

asn ser

2 1/4 cycles

1 1/4 cycles

0

751

3 1/4 cycles

120 Time (s)

200

160

(b) gly ser asn phe

arg gly ser asn phe Fluorescence

*

1 1/4 cycles

60

65

70

75

80

ser phe gly asn

1/4 cycle 0

1 1/4 cycles 40

asn

2 1/4 cycles

80 120 Time (s)

85

3 1/4 cycles 160

200

FIGURE 24.21 (a) MEKC separation of FITC-labeled amino acids using microchip-based SCCE. (b) CGE separation of FITC-labeled amino acids using the same microchip-based SCCE device. (Reprinted from von Heeren, F., et al., Analytical Chemistry 1996, 68, 2044–2053. Copyright 1996 American Chemical Society. With permission.)

752

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

5 (a) All cycles abs., mAU

4 3 2 1 0 0

200

400

600

800

Time (min)

abs., mAU

4

(b) Cycle #5

3 2 1 0 120

121

122

123

124

125

126

127

720

721

722

Time (min) 1.4 1.2

(c) Cycle #26

abs., mAU

1.0 0.8 0.6 0.4 0.2 0.0 715

716

717

718

719

Time (min)

FIGURE 24.22 MEKC separation of phenylalanine and phenylalanine-ring-D5 using capillary-based SCCE. (Reprinted from Zhao, J. and Jorgenson, J. W., Journal of Microcolumn Separations 1999, 11, 439–449. Copyright 1999 John Wiley & Sons, Inc. With permission.)

shown in Figure 24.22. These two analytes have only a 0.4% mobility difference and were separated in only 14 h.81 CGE separations have been reported using microchip-based SCCE. Figure 24.21b shows an SCCE with CGE separation of FITC-labeled amino acids.79 SCCE with CGE has also been used with another microchip design for the separation of a PhiX174/Hae III DNA ladder. Separations with difference synchronization times were used to isolate different DNA fragments.78

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

753

24.4.3.4 Methods of Detection Laser-induced fluorescence detection is a logical choice for high-resolution SCCE when performed on microfluidic chip-based systems. High laser power can result in photobleaching, decreasing the analyte signal every time it passes through the beam.75,76,79 UV absorbance can easily be done on capillary-based systems using capillaries at least 50–75 µm in diameter. It is a universal method of detection that can be used with high sensitivity with most analytes. The short pathlength provide by the shallow channels make UV absorbance detection difficult to implement on microfluidic systems. In addition, microchip-based systems fabricated from glass instead of fused silica absorb significant amounts of UV, further complicating its implementation. In future experiments, detection methods may include direct conductivity measurements (microchip-based SCCE) and contactless conductivity methods (capillary-based SCCE). 24.4.3.5 Limitations Theoretically, the ability of SCCE to separate analytes is limited only by longitudinal diffusion. In practice, broadening from the dead volume connecting the separation capillaries and adsorption of the analytes to the capillary wall ultimately limit the efficiency of the separation. Microchip-based SCCE also loses efficiency from broadening around tight turns. SCCE, similar to FCCE, is a method in which each cycle reduces the range of mobilities that can be analyzed. The result is that the peak capacity of the separation will be reduced during each subsequent cycle. This effect is most notable when short channels and low voltages are used.78 The switching of potentials in SCCE causes analytes of the selected mobility to continuously travel around the separation loop in the same direction. Analytes that migrate at higher mobilities are lost at the corners and flushed into the waste reservoirs. Analytes of lower mobility that do not make it into the next separation channel before the potentials are switched will travel in the reverse direction. This can cause data analysis problems for complex samples when these backward migrating peaks pass in front of the detector again. This has been observed in several microchipbased separations including the SCCE separation of fluorescein isothiocyanate and its degradation products,88 the MEKC and CGE with SCCE separations of FITC-labeled amino acids and human urine,79 and SCCE of double-stranded DNA in a sieving matrix.78 While this phenomenon is not much of a problem with relatively simple samples, data analysis may be confusing and complicated for complex samples such as cell lysates.

24.4.4 METHODS DEVELOPMENT GUIDELINES 24.4.4.1 Handling Extended Separation Times The selection of the buffer components is very important in SCCE. Just as in FCCE and UHVCE, careful attention must be paid toward the amount of analyte adsorption to the channel walls. At extended separation times, losses from adsorption will add to reduce the signal to noise value greatly. Although the separation channels are flushed with fresh buffer during the SCCE separation, buffering capacity must be fairly high for long separation times if the reservoir size is small. This is a bigger problem with microchip-based SCCE where typical reservoir volumes are less than 1 mL. Capillary-based SCCE typically uses 20 mL buffer reservoirs and buffering capacity problems are usually not observed. Similar to FCCE, automation of SCCE voltage switching can greatly improve the reliability and usability of the SCCE system. Zhao et al.80 report the use of software that determined the rising and falling signal of the detected analyte above a predetermined threshold value during each

754

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

cycle and recalculated switching times to ensure that the analyte bands stayed within the separation channels. Longer separation channels have a few advantages over shorter channels when separation times are long. Longer channels allow longer separation distances without as much broadening and sample loss associated with channel connections and corners. In addition, longer separation channels maintain a higher peak capacity during extended run times, and the analyte bands can broaden considerably without being cut off at the edges.

24.4.4.2 Band Broadening Prevention Band broadening prevention strategies for SCCE are similar to those used in UHVCE and FCCE. Known formulas can be used for buffer components if little peak tailing is seen under standard CE separations. Electrostatic repulsion from channel walls that share the same electrical charge as the analytes can reduce broadening greatly. The electrophoretic mobilities of the buffer ions should closely match those of the sample ions to prevent broadening from electrodispersion. Similarly, the ionic strength of the buffer should be greater than or equal to that of the sample. Several strategies for reducing dead volume at the channel connections have been reported. The use of T-shaped junctions at the middle of the separation channels (Figure 24.19) has been employed in an effort to reduce losses at the corner connections of the separation channels. Experimental observation indicated that broadening was more significant than that seen in previous designs with reservoir connections at the corners.79 Recently, Manz et al.78 have reported the use of side channels that are much shallower (∼1/8th depth) than the separation channel in an effort to reduce losses at the connections of the channels. The use of narrower rounded turns for microchip-based SCCE is reported for reducing band broadening at the corners.78 This approach has been shown to reduce the broadening seen for turns in other microfluidic chip designs.83,86,87,89,90 Other, more exotic approaches to reducing dispersion at the turns of channels on microfluidic devices have been reported. One technique using a pulsed UV laser to modify the surface of plastic chips at the turn to increase the EOF by up to 4%. This technique was shown to reduce band broadening at the turns.91

24.4.4.3 Sample Separation Strategies Amino acids and peptides are easily labeled with fluorescent dyes. CZE, CGE, or MEKC in capillary or microchip-based SCCE can typically resolve these analytes. Microchip-based SCCE offers higher sensitivity for very small samples, and higher electric field strengths can be used without Joule heating problems. Capillary-based SCCE offers a much higher peak capacity and the use of UV absorbance detection of unlabeled analytes. Isotopomers are best separated using capillary-based SCCE due to the extremely long separation times required. The components of the buffer must be carefully selected so that their ionic strength and ion mobility are very close to that of the analytes and adsorption to the capillary wall is prevented. MEKC can be used in some cases to separate isotopomers more quickly than CZE, most likely due to small changes in hydrophobicity caused by deuterium substitution. Resolution of the enantiomers in a racemic mixture will require the use of a chiral selector such as a cyclodextrin. Fluorescently labeled analytes can be detected at very low concentration using microchip-based SCCE, and unlabeled analytes can be easily detected with UV absorption using capillary-based SCCE. The long separation times available with capillary-based SCCE allows the use of less than optimal chiral resolving agents, whereas a more selective chiral selector is needed in microchip-based SCCE.

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

755

24.5 CONCLUDING REMARKS The three techniques presented in this chapter demonstrate different approaches to improving the resolution and efficiency of CE separations. All three methods work by increasing the time and distance that the analyte bands electrophoretically migrate under a high electric field strength, but the different ways in which each method accomplishes this goal determine their useful application. UHVCE uses the direct application of extremely high voltages on long capillaries, and it allows the examination of analytes that span the entire range of mobilities found within the sample. Samples containing complex mixtures of proteins, peptides, nucleic acids, and many others display a wide range of mobilities and UHVCE is well suited to their analysis. FCCE and SCCE use conventional voltages on relatively short capillaries, but they maintain analytes with a narrow range of mobilities within the capillaries for very long periods of time. Because high-resolution analysis is limited to such a narrow range of mobilities, FCCE and SCCE are best suited for the analysis of difficult to resolve pairs such as closely related peptides, enantiomers, and isotopomers. The ultimate utility of SCCE and FCCE depends on the ability of the experimenter to eliminate sources of band broadening. The amount of broadening observed in SCCE depends mainly on the design of the instrumentation. Band broadening in FCCE is inherent to the magnitude of the counter flow required to reverse the migration of the analyte bands. The analysis of extremely complex samples such as cell lysates can be simplified through a combination of these methods. Preliminary high-efficiency, high-resolution analysis with UHVCE can be used to elucidate areas of interest within the sample. After the mobilities of the analytes of interest have been determined, FCCE or SCCE can then be used for much more detailed analysis of particular bands. The analysis of complicated biological samples remains an area where the implementation of these techniques can provide new insight.

ACKNOWLEDGMENT National Science Foundation Predoctoral Fellowship Program.

REFERENCES 1. Jorgenson, J. W. and Lukacs, K. D., Capillary zone electrophoresis. Journal of Chromatography Library 1985, 30 (Microcolumn Sep.), 121–131. 2. Jorgenson, J. W. and Lukacs, K. D., Capillary zone electrophoresis. Science (Washington, DC, United States) 1983, 222, 266–272. 3. Jorgenson, J. W. and Lukacs, K. D., Free-zone electrophoresis in glass capillaries. Clinical Chemistry (Washington, DC, United States) 1981, 27, 1551–1553. 4. Jorgenson, J. W. and Lukacs, K. D., Zone electrophoresis in open-tubular glass capillaries. Analytical Chemistry 1981, 53, 1298–1302. 5. Lukacs, K. Theory, instrumentation, and applications of capillary zone electrophoresis. PhD Dissertation, University of North Carolina, Chapel Hill, 1983. 6. Henley, W. H. Two distinct approaches for increasing the resolution and efficiency of capillary electrophoretic separations. PhD Dissertation, University of North Carolina, Chapel Hill, 2005. 7. Hutterer, K. M. and Jorgenson, J. W., Ultrahigh-voltage capillary zone electrophoresis. Analytical Chemistry 1999, 71, 1293–1297. 8. Palonen, S., Jussila, M., Porras, S. P., Hyotylainen, T., and Riekkola, M. L., Extremely high electric field strengths in non-aqueous capillary electrophoresis. Journal of Chromatography A 2001, 916, 89–99. 9. Halliday, D., Resnick. R., and Krane, K. S., Physics. John Wiley & Sons, Inc., New York, 1992. 10. Van de Graaff, R. J., Compton, K. T., and Van Atta, L. C., The electrostatic production of high voltage for nuclear investigations. Physical Review 1933, 43, 149–157.

756

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

11. Kraft, E. M. S., Development of capillary zone electrophoresis at 100,000 to 150,000 volts. M.S. Thesis, University of North Carolina, Chapel Hill, NC, 1996. 12. Hutterer, K. M., Ultra high voltage capillary electrophoresis. Doctorial Thesis, University of North Carolina, Chapel Hill, NC, 2000. 13. Hutterer, K. M. and Dolnik, V., Capillary electrophoresis of proteins 2001–2003. Electrophoresis 2003, 24, 3998–4012. 14. Hutterer, K. M. and Jorgenson, J. W., Separation of hyaluronic acid by ultrahigh-voltage capillary gel electrophoresis. Electrophoresis 2005, 26, 2027–2033. 15. Hutterer, K. M., Birrell, H., Camilleri, P., and Jorgenson, J. W., High resolution of oligosaccharide mixtures by ultrahigh voltage micellar electrokinetic capillary chromatography. Journal of Chromatography B 2000, 745, 365–372. 16. Verma, R., Shyam, A., and Nair, L., Development of battery powered 100 kV dc power supply. Review of Scientific Instruments 2006, 77 (10, Pt. 1), 106104/1–106104/4. 17. Gordon, M. J., Lee, K. J., Arias, A. A., and Zare, R. N., Protocol for resolving protein mixtures in capillary zone electrophoresis. Analytical Chemistry 1991, 63, 69–72. 18. Duso, A. B. and Chen, D. D. Y., Proton and hydroxide ion mobility in capillary electrophoresis. Analytical Chemistry 2002, 74, 2938–2942. 19. Huang, X., Coleman, W. F., and Zare, R. N., Analysis of factors causing peak broadening in capillary zone electrophoresis. Journal of Chromatography 1989, 480, 95–110. 20. Moseley, M. A., Deterding, L. J., Tomer, K. B., and Jorgenson, J. W., Determination of bioactive peptides using capillary zone electrophoresis/mass spectrometry. Analytical Chemistry 1991, 63, 109–114. 21. Moseley, M. A., Jorgenson, J. W., Shabanowitz, J., Hunt, D. F., and Tomer, K. B., Optimization of capillary zone electrophoresis/electrospray ionization parameters for the mass spectrometry and tandem mass spectrometry analysis of peptides. Journal of the American Society for Mass Spectrometry 1992, 3, 289–300. 22. Kirkland, J. J., Glajch, J. L., and Farlee, R. D., Synthesis and characterization of highly stable bonded phases for high-performance liquid chromatography column packings. Analytical Chemistry 1989, 61, 2–11. 23. Hardenborg, E., Zuberovic, A., Ullsten, S., Soderberg, L., Heldin, E., and Markides, K. E., Novel polyamine coating providing non-covalent deactivation and reversed electroosmotic flow of fusedsilica capillaries for capillary electrophoresis. Journal of Chromatography A 2003, 1003, 217–221. 24. Ullsten, S., Zuberovic, A., Wetterhall, M., Hardenborg, E., Markides, K. E., and Bergguist, J., A polyamine coating for enhanced capillary electrophoresis-electrospray ionization-mass spectrometry of proteins and peptides. Electrophoresis 2004, 25, 2090–2099. 25. Albarghouthi, M. N. and Barron, A. E., Polymeric matrices for DNA sequencing by capillary electrophoresis. Electrophoresis 2000, 21, 4096–4111. 26. Albarghouthi, M. N., Buchholz, B. A., Doherty, E. A. S., Bogdan, F. M., Zhou, H., and Barron, A. E., Impact of polymer hydrophobicity on the properties and performance of DNA sequencing matrices for capillary electrophoresis. Electrophoresis 2001, 22, 737–747. 27. Albarghouthi, M. N., Buchholz, B. A., Stein, T. M., and Barron, A. E., PolyDuramide: A novel, hydrophilic, self-coating polymer matrix for DNA sequencing by capillary electrophoresis. Abstracts of Papers, 222nd ACS National Meeting, Chicago, IL, United States, August 26–30, 2001, ANYL-119. 28. Buchholz, B. A., Doherty, E. A. S., Albarghouthi, M. N., Bogdan, F. M., Zahn, J. M., and Barron, A. E., Microchannel DNA sequencing matrices with a thermally controlled “viscosity Switch”. Analytical Chemistry 2001, 73, 157–164. 29. Carrilho, E., Ruiz-Martinez, M. C., Berka, J., Smirnov, I., Goetzinger, W., Miller, A. W., Brady, D., and Karger, B. L., Rapid DNA sequencing of more than 1000 bases per run by capillary electrophoresis using replaceable linear polyacrylamide solutions. Analytical Chemistry 1996, 68, 3305–3313. 30. Dolnik, V., DNA sequencing by capillary electrophoresis (review). Journal of Biochemical and Biophysical Methods 1999, 41, 103–119. 31. Dolnik, V., Gurske, W. A., and Padua, A., Galactomannans as a sieving matrix in capillary electrophoresis. Electrophoresis 2001, 22, 707–719. 32. Heller, C., Separation of double-stranded and single-stranded DNA in polymer solutions. Part 2. Separation, peak width, and resolution. Electrophoresis 1999, 20, 1978–1986.

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

757

33. Salas-Solano, O., Carrilho, E., Kotler, L., Miller, A. W., Goetzinger, W., Sosic, Z., and Karger, B. L., Routine DNA sequencing of 1000 bases in less than one hour by capillary electrophoresis with replaceable linear polyacrylamide solutions. Analytical Chemistry 1998, 70, 3996–4003. 34. Slater, G. W., Desruisseaux, C., Hubert, S. J., Mercier, J.-F., Labrie, J., Boileau, J., Tessier, F., and Pepin, M. P., Theory of DNA electrophoresis: A look at some current challenges. Electrophoresis 2000, 21, 3873–3887. 35. Song, L., Liang, D., Chen, Z., Fang, D., and Chu, B., DNA sequencing by capillary electrophoresis using mixtures of polyacrylamide and poly(N,N-dimethylacrylamide). Journal of Chromatography A 2001, 915, 231–239. 36. Song, L., Liang, D., Kielescawa, J., Liang, J., Tjoe, E., Fang, D., and Chu, B., DNA sequencing by capillary electrophoresis using copolymers of acrylamide and N,N-dimethyl-acrylamide. Electrophoresis 2001, 22, 729–736. 37. Wei, W. and Yeung, E. S., Improvements in DNA sequencing by capillary electrophoresis at elevated temperature using poly(ethylene oxide) as a sieving matrix. Journal of Chromatography B: Biomedical Sciences and Applications 2000, 745, 221–230. 38. Zhou, H., Miller, A. W., Sosic, Z., Buchholz, B., Barron, A. E., Kotler, L., and Karger, B. L., DNA Sequencing up to 1300 bases in two hours by capillary electrophoresis with mixed replaceable linear polyacrylamide solutions. Analytical Chemistry 2000, 72, 1045–1052. 39. Huber, C. G. and Oberacher, H., Analysis of nucleic acids by on-line liquid chromatography-mass spectrometry. Mass Spectrometry Reviews 2002, 20, 310–343. 40. Shibata, K., Itoh, M., Aizawa, K., Nagaoka, S., Sasaki, N., Carninci, P., Konno, H., et al., RIKEN integrated sequence analysis (RISA) system—384-format sequencing pipeline with 384 multicapillary sequencer. Genome Research 2000, 10, 1757–1771. 41. Culbertson, C. T. and Jorgenson, J. W., Increasing the resolving power of capillary electrophoresis through electroosmotic flow control using radial fields. Journal of Microcolumn Separations 1999, 11, 167–174. 42. Cheng, Y. F., Wu, S., Chen, D. Y., and Dovichi, N. J., Interaction of capillary zone electrophoresis with a sheath flow cuvette detector. Analytical Chemistry 1990, 62, 496–503. 43. Williams, B. A. and Vigh, G., The use of hydrodynamic counterflow to improve the resolution and detection of the minor component in the capillary electrophoretic analysis of enantiomers. Enantiomer 1996, 1, 183–191. 44. Bier, M., Twitty, G. E., and Sloan, J. E., Recycling isoelectric focusing and isotachophoresis. Journal of Chromatography 1989, 470, 369–376. 45. Everaerts, F. M., Vacik, J., Verheggen, T. P. E. M., and Zuska, J., Displacement electrophoresis. Experiments with counterflow of electrolyte. Journal of Chromatography 1970, 49, 262–268. 46. Everaerts, F. M., Vacik, J., Verheggen, T. P. E. M., and Zuska, J., Isotachophoresis. Experiments with electrolyte counterflow. Journal of Chromatography 1971, 60, 397–405. 47. Everaerts, F. M., Verheggen, T. P. E. M., and Van de Venne, J. L. M., Isotachophoretic experiments with a counter flow of electrolyte. Journal of Chromatography 1976, 123, 139–148. 48. Reinhoud, N. J., Tjaden, U. R., and van der Greef, J., Automated isotachophoretic analyte focusing for capillary zone electrophoresis in a single capillary using hydrodynamic back-pressure programming. Journal of Chromatography 1993, 641, 155–162. 49. Reinhoud, N. J., Tjaden, U. R., and van der Greef, J., Strategy for setting up single-capillary isotachophoresis-zone electrophoresis. Journal of Chromatography A 1993, 653, 303–312. 50. Clark, S. L. and Remcho, V. T., Electrochromatographic retention studies on a flavin-binding RNA aptamer sorbent. Analytical Chemistry 2003, 75, 5692–5696. 51. Lucy, C. A. and McDonald, T. L., Separation of chloride isotopes by capillary electrophoresis based on the isotope effect on ion mobility. Analytical Chemistry 1995, 67, 1074–1078. 52. Soga, T., Inoue, Y., and Ross, G. A., Analysis for halides, oxyhalides and metal oxoacids by capillary electrophoresis with suppressed electroosmotic flow. Journal of Chromatography A 1995, 718, 421–428. 53. Terabe, S., Yashima, T., Tanaka, N., and Araki, M., Separation of oxygen isotopic benzoic acids by capillary zone electrophoresis based on isotope effects on the dissociation of the carboxyl group. Analytical Chemistry 1988, 60, 1673–1677.

758

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

54. Yeung, K. K. C. and Lucy, C. A., Isotopic separation of [14N]- and [15N]aniline by capillary electrophoresis using surfactant-controlled reversed electroosmotic flow. Analytical Chemistry 1998, 70, 3286–3290. 55. Yeung, K. K. C. and Lucy, C. A., Ultrahigh-resolution capillary electrophoretic separation with indirect ultraviolet detection: Isotopic separation of [14N]-and [15N]ammonium. Electrophoresis 1999, 20, 2554–2559. 56. Brewer, A. K., Madorsky, S. L., and Westhaver, J. W., The concentration of K39 and K41 by balanced ion migration in a counterflowing electrolyte. Science (Washington, DC, United States) 1946, 104, 156–157. 57. Chankvetadze, B., Burjanadze, N., Bergenthal, D., and Blaschke, G., Potential of flow-counterbalanced capillary electrophoresis for analytical and micropreparative separations. Electrophoresis 1999, 20, 2680–2685. 58. Ivory, C. F., Preparative free-flow electrofocusing in a vortex-stabilized annulus. Electrophoresis 2004, 25, 360–374. 59. McLaren, D. G. and Chen, D. D. Y., A quantitative study of continuous flow-counterbalanced capillary electrophoresis for sample purification. Electrophoresis 2003, 24, 2887–2895. 60. McLaren, D. G. and Chen, D. D. Y., Continuous electrophoretic purification of individual analytes from multicomponent mixtures. Analytical Chemistry 2004, 76, 2298–2305. 61. Thome, B. and Ivory, C. F., Continuous fractionation of enantiomer pairs in free solution using an electrophoretic analog of simulated moving bed chromatography. Journal of Chromatography, A 2002, 953, 263–277. 62. Thome, B. M. and Ivory, C. F., Development of a segmented model for a continuous electrophoretic moving bed enantiomer separation. Biotechnology Progress 2003, 19, 1703–1712. 63. Tracy, N. I. and Ivory, C. F., Preparative isoelectric focusing of proteins using binary buffers in a vortex-stabilized, free-flow apparatus. Electrophoresis 2004, 25, 1748–1757. 64. Kutter, J. and Welsch, T., The effect of electroosmotic and hydrodynamic flow profile superposition on band broadening in capillary electrophoresis. Journal of High Resolution Chromatography 1995, 18, 741–744. 65. Culbertson, C. T. and Jorgenson, J. W., Flow counterbalanced capillary electrophoresis. Analytical Chemistry 1994, 66, 955–962. 66. Iwata, T. and Kuroshu, Y., Enhancement of the sensitivity and resolution by flow-counterbalanced capillary electrophoresis. Analytical Sciences 1995, 11, 131–133. 67. Henley, W. H., Wilburn, R. T., Crouch, A. M., and Jorgenson, J. W., Flow counterbalanced capillary electrophoresis using packed capillary columns: Resolution of enantiomers and isotopomers. Analytical Chemistry 2005, 77, 7024–7031. 68. Culbertson, C. T. and Jorgenson, J. W., Separation of fluorescently derivatized deuterated isotopomers of phenylalanine using micellar electrokinetic chromatography and flow counterbalanced micellar electrokinetic chromatography. Journal of Microcolumn Separations 1999, 11, 175–183. 69. MacNair, J. E., Lewis, K. C., and Jorgenson, J. W., Ultrahigh-pressure reversed-phase liquid chromatography in packed capillary columns. Analytical Chemistry 1997, 69, 983–989. 70. Shalliker, R. A., Broyles, B. S., and Guiochon, G., Evaluation of the secondary consolidation of columns for liquid chromatography by ultrasonic irradiation. Journal of Chromatography, A 2000, 878, 153–163. 71. Tong, D., Bartle, K. D., Clifford, A. A., and Edge, A. M., Theoretical studies of the preparation of packed capillary columns for chromatography. Journal of Microcolumn Separations 1995, 7, 265–278. 72. Culbertson, C. T. New methods for increasing the resolving power and lowering the UV absorbance detection limits in capillary electrophoresis. PhD Thesis, University of North Carolina, Chapel Hill, 1996. 73. Eijkel, J. C. T., van den Berg, A., and Manz, A., Cyclic electrophoretic and chromatographic separation methods. Electrophoresis 2004, 25, 243–252. 74. Ramsteiner, K. A., Systematic approach to column switching. Journal of Chromatography 1988, 456, 3–20.

Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE

759

75. Burggraf, N., Manz, A., De Rooij, N. F., and Widmer, H. M., Synchronized cyclic capillary electrophoresis: A novel concept for high-performance separations using low voltages. Analytical Methods & Instrumentation 1993, 1, 55–59. 76. Burggraf, N., Manz, A., Effenhauser, C. S., Verpoorte, E., de Rooij, N. F., and Widmer, H. M., Synchronized cyclic capillary electrophoresis—A novel approach to ion separations in solution. Journal of High Resolution Chromatography 1993, 16, 594–596. 77. Burggraf, N., Manz, A., Verpoorte, E., Effenhauser, C. S., Widmer, H. M., and de Rooij, N. F., A novel approach to ion separations in solution: Synchronized cyclic capillary electrophoresis (SCCE). Sensors and Actuators, B: Chemical 1994, 20, 103–110. 78. Manz, A., Bousse, L., Chow, A., Metha, T. B., Kopf-Sill, A., and Parce, J. W., Synchronized cyclic ′ capillary electrophoresis using channels arranged in a triangle and low voltages. Fresenius Journal of Analytical Chemistry 2001, 371, 195–201. 79. von Heeren, F., Verpoorte, E., Manz, A., and Thormann, W., Micellar electrokinetic chromatography separations and analyses of biological samples on a cyclic planar microstructure. Analytical Chemistry 1996, 68, 2044–2053. 80. Zhao, J., Hooker, T., and Jorgenson, J. W., Synchronous cyclic capillary electrophoresis using conventional capillaries: System design and preliminary results. Journal of Microcolumn Separations 1999, 11, 431–437. 81. Zhao, J. and Jorgenson, J. W., Application of synchronous cyclic capillary electrophoresis: Isotopic and chiral separations. Journal of Microcolumn Separations 1999, 11, 439–449. 82. Choi, J. G., Kim, M., Dadoo, R., and Zare, R. N., Electrophoretron: A new method for enhancing resolution in electrokinetic separations. Journal of Chromatography, A 2001, 924, 53–58. 83. Culbertson, C. T., Jacobson, S. C., and Ramsey, J. M., Dispersion sources for compact geometries on microchips. Analytical Chemistry 1998, 70, 3781–3789. 84. Fu, L.-M., Yang, R.-J., and Lee, G.-B., Analysis of geometry effects on band spreading of microchip electrophoresis. Electrophoresis 2002, 23, 602–612. 85. Griffiths, S. K. and Nilson, R. H., Design and analysis of folded channels for chip-based separations. Analytical Chemistry 2002, 74, 2960–2967. 86. Wang, Y., Lin, Q., and Mukherjee, T., System-oriented dispersion models of general-shaped electrophoresis microchannels. Lab on a Chip 2004, 4, 453–463. 87. Yao, Z.-H., Yoder, G. L., Culbertson, C. T., and Ramsey, J. M., Numerical simulation of dispersion generated by a 180 Deg turn in a microchannel. Chinese Physics (Beijing, China) 2002, 11, 226–232. 88. Manz, A., Verpoorte, E., Effenhauser, C. S., Burggraf, N., Raymond, D. E., and Widmer, H. M., Planar chip technology for capillary electrophoresis. Fresenius’ Journal of Analytical Chemistry 1994, 348, 567–571. 89. Culbertson, C. T., Jacobson, S. C., and Ramsey, J. M., Microchip devices for high-efficiency separations. Analytical Chemistry 2000, 72, 5814–5819. 90. Ramsey, J. D., Jacobson, S. C., Culbertson, C. T., and Ramsey, J. M., High-efficiency, two-dimensional separations of protein digests on microfluidic devices. Analytical Chemistry 2003, 75, 3758–3764. 91. Johnson, T. J., Ross, D., Gaitan, M., and Locascio, L. E., Laser modification of preformed polymer microchannels: Application to reduce band broadening around turns subject to electrokinetic flow. Analytical Chemistry 2001, 73, 3656–3661.

of DNA for 25 Separation Forensic Applications Using Capillary Electrophoresis Lilliana I. Moreno and Bruce McCord CONTENTS 25.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3 Theoretical Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3.1 The Capillary and the Sieving Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3.2 Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3.3 Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3.4 Size Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4 Practical Applications in Forensic Biology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4.1 STRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4.2 Mini STRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4.3 Mitochondrial DNA Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4.4 Y-STRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4.5 Single Nucleotide Polymorphisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4.6 Mutation Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4.7 Nonhuman DNA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4.7.1 Animal DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4.7.2 Botanical DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4.7.3 Microbial DNA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

761 762 762 762 763 764 764 764 765 768 770 771 773 775 775 775 776 777 779 779 779

25.1 INTRODUCTION Forensic science is defined as the application of science to the law. It is the goal of the forensic scientist to identify and compare physical evidence and use the resulting observations to aid in solving criminal or civil matters. From its inception, forensic scientists have recognized the potential of capillary electrophoresis (CE) as a useful tool to assist in the analysis of a wide array of trace evidence samples [1,2]. This is particularly true for applications in forensic biology. In this discipline, biological fluids (blood, hair, semen) left behind at a crime scene are probed to establish the essential facts of the crime. Sample analysis is performed via extraction and analysis of the genetic material within these samples. The key to the procedure is targeting specific locations in the genome containing polymorphisms (different allelic forms) that permit differentiation between individuals. The statistical probability 761

762

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

of an individual inheriting any given polymorphism can then be used to assess the evidence and compare it with known samples from victims and suspects. In the early years of DNA analysis, the separation and sizing of DNA was accomplished by slab gel electrophoresis. However, since 1998 CE has gradually taken over slab gel techniques in this field because of its ease of automation, minimal sample consumption (which is of utmost importance in the field), and its high-throughput capabilities.

25.2 BACKGROUND Although seldom used for genotyping in current forensic applications, slab gels set the precedent for the development of CE. Slab gel systems had an advantage in forensic analysis in that multiple samples could be run simultaneously, allowing easy comparisons and rapid analysis. The ability to compare different samples on the same gel was particularly useful in early procedures involving restriction fragment length polymorphisms (RFLPs) and multilocus probes. With the RFLP techniques, DNA fragment sizes ranged up to 20,000 bases and results were based on the presence or absence of bands within different size ranges. Separations took place on large format agarose gels and genetic loci were detected using Southern blotting with radioactive probes. As a result, a complete analysis could take several weeks because of the time involved in repetitive DNA transfer, hybridization, and exposure to radioactive film [3]. With the advent of the polymerase chain reaction (PCR), there was a complete paradigm shift. The PCR permits a few copies of a DNA sequence to be amplified to millions of copies [4], thus increasing sensitivity and eliminating the need for radioactive probes. DNA fragment sizes were reduced to a few hundred bases, and the total time for an analysis dropped to a single day. With PCR, sample quantities were sufficient for detection via simple silver staining processes; however due to the smaller size of the alleles, higher resolution acrylamide gels were required. As the exceptional capabilities of the PCR in DNA typing became apparent, law enforcement personnel began to take advantage of the expanded speed and sensitivity that the new technique had to offer. Consequently, an increasing number of evidence samples began to be submitted to laboratories for testing. Therein arose a new concern. How were laboratories going to be able to handle the bulk of samples received in a timely manner? An analytical technique was needed that could provide as good or better results than those obtained via slab gel systems in a standardized and automated fashion. This technique was to be CE. The principle governing both slab gel and CE techniques is the same: the separation of a series of analytes based on size selective sieving through a gel or polymer network under the influence of an electric field. The slab gel matrix permits a size-dependent separation of fragments since larger DNA fragments contain an invariant charge-to-size ratio due to their sugar-phosphate backbone. However, slab gels require manual operations such as the preparation of the gel and the loading of samples. Automated CE systems eliminated these tedious tasks and substituted replaceable, entangled polymer solutions for the rigid cross-linked acrylamide gels. Also, the improved heat dissipation of the fusedsilica capillary permits much higher voltages to be used, reducing the run time of the analysis [5]. Laser-induced fluorescence detection also improved throughput by allowing multiwavelength analysis [6]. Soon, the forensic science community became interested in this robust technology and began to adopt it for a variety of analyses. In particular, CE systems can be used for DNA quantification, genotyping, sequencing, and mutation detection.

25.3 THEORETICAL ASPECTS 25.3.1 THE CAPILLARY AND THE SIEVING MATRIX In 1988, it was shown that DNA single nucleotide separation could be achieved using cross-linked polyacrylamide gel filled capillaries coupled with UV/Vis detection [7]. These chemical gels provided

Separation of DNA for Forensic Applications Using Capillary Electrophoresis

763

size selective separation of DNA molecules based on their ability to migrate through transient pores created in the polymer matrix. Unfortunately, cross-linked gel filled capillaries had a limited lifetime due to the ready formation of voids in the gel during electrophoresis [8]. There were also concerns with carryover from one run to the next, an important point when precious evidentiary samples were being run. The solution to this problem was the development of CE systems with entangled polymer buffers such as polyacrylamide, hydroxyethylcellulose, and polyethylene oxide [9]. With proper control of polymer molecular weight and concentration, these buffers provided equivalent separation to cross-linked gels with the added advantage that they permitted refilling and reuse of the capillary. This property allowed the same capillary to be used up to 100 times before replacement. An additional problem with these novel sieving polymers was the effect of the electroosmotic flow (EOF) on migration time reproducibility. For optimum results, the polymer network must not only separate the DNA fragments but also eliminate wall effects such as adsorption and osmotic flow. This is an important issue as the primary function of the method is to precisely determine the size of the DNA fragments. Variations in migration time due to slight differences in EOF can negatively affect estimation of fragment size. For this reason, linear polydimethyl acrylamide “POP” became the polymer of choice for these separations because of its relatively low viscosity and its capability to eliminate EOF [10,11]. Another critical issue in the development of the separation was the choice of buffer composition. The optimal buffer should produce low and stable currents and be able to separate DNA under both denaturing—where the formation of secondary structure of the sample is halted by denaturing agents such as heat, formamide, or urea [12,13]—or nondenaturing conditions—where secondary structure formation is desired to study subtle differences in DNA sequences [14,15]. For the majority of genotyping applications such as sequencing and fragment length determination, denaturing conditions are preferred as resolution is superior and there is a predictable relationship between fragment size and migration time. Buffers such as trishydroxymethylaminomethane (Tris) and N-tris[hydroxymethyl]methyl-3-aminopropanesulfonic acid (TAPS) are commonly used as they produce low currents and can buffer at physiological pH. A commonly used buffer for sequencing and genotyping applications consists of 100 mM TAPS at pH 8.0 with 5% pyrolidinone and 8 M urea as denaturants [11]. Four to six percent polydimethyl acrylamide at a controlled molecular weight is added to sieve the DNA and eliminate EOF. Uncoated fused-silica capillaries of 50 µm are typically used as this diameter provides a good compromise between resolution, sensitivity, and resistance to clogging.

25.3.2 INJECTION For most applications in forensic DNA analysis, electrokinetic injection is used. This injection mode provides improved peak shape and intensity as a result of field amplified sample stacking. The stacking process results when the DNA sample is prepared in low ionic strength solutions. The application of an electric field accelerates the DNA to the sample/buffer interface where it concentrates because of the drop in field strength at that point. Because this process is highly dependent on the ionic strength of the sample solution, great care must be taken to keep salt concentrations low. Unfortunately PCR amplifications occur at relatively high-salt concentrations (70–100 mM KCl) and various buffers are often added to the stabilize DNA for long-term storage. Thus, the sample must be diluted or dialyzed before injection as smaller buffer ions have higher electrophoretic mobility and interfere with the injection of the DNA. Sample injection is an important issue in forensic analysis as there should be a semi-quantitative relationship between peak height and sample concentration. This relationship helps the analyst assess the quality of the sample and its preparation. It also helps define the relative level of different contributors to a mixture. The removal of salts through dialysis or the unintentional addition of higher quantities of salt can produce peak intensities that are less representative of sample quantity and affect the aforementioned relationship.

764

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

In a typical injection, 1–2 µL of a completed PCR is diluted in 10–20 µL of formamide. Water can also be used but the formamide produces complete denaturation of the sample without further processing. High purity, low conductance formamide must be used to avoid loss of peak intensity or other injection artifacts.

25.3.3 DETECTION Laser-induced fluorescence is the most common method of detection for DNA analyses due to its high sensitivity and multiplex capabilities. Typically argon-ion lasers are used with excitation wavelengths of 488 and 514 nm. Detection occurs using a charge-coupled device (CCD) that collects the fluorescence emission produced by the various dyes bound to the DNA molecules. There are three basic methods for detection of the DNA using these dyes: intercalation, amplification with dye-labeled PCR primers, or incorporation of dye-labeled bases into the DNA sequence during replication. For single channel detection of native DNA, intercalating dyes produce excellent results [16]. These dyes may simply be added to the CE buffer and the DNA is labeled as it moves through the gel toward the detector. For genotyping of denatured DNA, dye-labeled primers are used [17] while dye-labeled bases are used in DNA sequencing [18]. The specific dyes used for genotyping and DNA sequencing are designed to simultaneously absorb at a single laser wavelength but emit at a variety of different wavelengths. Rhodamine and fluorescein derivatives are commonly used. Using these dyes, multiple loci can be amplified and genotyped without interfering with each other by simply labeling each set of PCR reactants with a different dye. Current commercial systems have the capability to detect as many as five different dyes simultaneously on a single CE capillary. Special software is used to eliminate problems with dye overlap by applying virtual filters and various calibration procedures [17].

25.3.4 SIZE ESTIMATION As mentioned previously, the primary function of the separation in genetic analysis is the estimation of fragment size. Multichannel fluorescence systems typically reserve one channel for use as an internal standard. The internal standard consists of multiple peaks throughout the size range of the analysis and is used by the computer to permit a precise estimate of unknown fragments in the other dye channels. Computer algorithms can then be used to produce a size estimation of the unknown DNA fragments in the other dye lanes. The samples can be further processed by comparison of these data with an external standard that consists of all known mutations. Using both sets of standards, it is possible to produce size estimates with a precision of better than 0.17 base pairs (bp) making the system capable of distinguishing single base differences at sizes up to 350 bp [19]. It is important to note at this point that the size estimates produced by these techniques can be influenced by temperature or sequence effects, and thus careful control of temperature and denaturant concentration is important in order to maintain precision [20,21].

25.4 PRACTICAL APPLICATIONS IN FORENSIC BIOLOGY The major application of CE in forensic biology is in the detection and analysis of short tandem repeats (STRs). STR markers are preferred because of the powerful statistical result that is possible with these markers and the large databases that exist for convicted offenders’ profiles. Other related applications include the analysis of haploid markers in the Y chromosome and in mitochondrial DNA (mtDNA). Nonhuman DNA testing can also be performed depending on the circumstances of the case. The techniques involved include genotyping, DNA sequencing, and mutation detection.

Separation of DNA for Forensic Applications Using Capillary Electrophoresis

765

25.4.1 STRs The analysis of STR loci in DNA is the most common method for the determination of human identity and can indisputably distinguish between two or more unrelated individuals if sufficient loci can be detected [22]. STR loci occur in noncoding regions of the human genome and consist of short segments of DNA 2–7 bp in length such as AATG, which are repeated consecutively multiple times. The number of repeats at a given locus can vary between individuals and there is a statistical probability that a given individual will have a set number of repeats at a particular STR locus. It has been estimated that over 100,000 STR loci exist in the human genome and research continues in an effort to determine their exact function [23]. Among this large number of STRs, the forensic community in the United States has established a set of 13 loci (Table 25.1) [24,25] that can be used to develop a genetic profile for the identification of individuals in criminal casework. To process the results from each analysis, large database known as Combined DNA Index System (CODIS) has been set up. This database stores profiles from convicted offenders and unsolved casework. Similar databases have been set up in Europe, Japan, and other countries. The information in these databases can be used to detect and apprehend serial offenders by permitting rapid exchange of information between crime laboratories [26–28]. STRs can be targeted for PCR amplification by preparing primer sequences that bind to more conserved sequences flanking the variable STR regions. The chemistry and protocols necessary for identifying each set of STRs are included in one or a combination of kits provided by companies such as Applied Biosystems and Promega. These kits (Table 25.2) have been subjected to strict validation processes to ensure the quality of the data [19,29–31]. The kits permit multiplex PCR amplification of up to 16 loci (including the sex determination gene) simultaneously from a single sample (Figure 25.1). The different loci included in the kit contain multiple alleles and are separated by size and dye label. The repeat motifs are 4–5 bases in length and the motif sequence can repeat itself up to 51 times (Table 25.1). To help define the size and migration time of each known allele an external standard known as an allelic ladder is run subsequent to each set of samples and used to more precisely define the identity of each peak (Figure 25.2).

TABLE 25.1 Thirteen STR Markers Commonly Used for Forensic DNA Analyses in the United States Marker (Locus)

Repeat Motif

CSF1PO FGA THO1 TPOX vWA D3S1358 D5S818 D7S820 D8S1179 D13S317 D16S539 D18S51 D21S11

TAGA CTTT TCAT GAAT [TCTG][TCTA] [TCTG][TCTA] AGAT GATA [TCTA][TCTG] TATC GATA AGAA [TCTA][TCTG]

Allele Range 6–16 15–51.2 3–14 6–13 10–24 9–20 7–16 6–15 8–19 5–15 5–15 7–27 24–38

Source: Adapted from Butler, J.M., Forensic DNA Typing: Biology, Technology, and Genetics of STR Markers, 2nd ed., Academic Press, San Diego, CA, 2005. With permission.

766

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

TABLE 25.2 Commercially Available Human STR Amplification Kits Kit Name

Target Loci

Discrimination Power

(A) Applied Biosystems, Foster City, CA AmpFLSTR® Profiler® D3S1358, vWA, FGA, THO1, TPOX, CSF1PO, D5S818, D13S317, D7S820, and Amelogenin AmpFL® SEFiler™ D2S1338, D3S1358, D8S1179, D16S539, D18S51, D19S433, D21S11, SE-33, FGA, vWA, and Amelogenin AmpFL® Cofiler® CSF1PO, D16S539, THO1, Amelogenin, TPOX, D3S1358, and D7S820 AmpFL Profiler Plus D3S1358, D5S818, D7S820, D8S1179, D13S317, D18S51, D21S11, FGA, vWA, and Amelogenin AmpFL® SGM Plus® D2S1338, D3S1358, D8S1179, D16S539, D18S51, D19S433, D21S11, THO1, FGA, vWA AmpFL® Identifiler® CSF1PO, D3S1358, D5S818, D7S820, D8S1179, D13S317, D16S539, D18S51, D21S11, vWA, FGA, THO1, TPOX, D2S1338, D19S433, and Amelogenin AmpFL® Green I THO1, TPOX, Amelogenin, and CSF1PO AmpFL® Blue D3S1358, vWA, FGA

1:3.6 × 109

1:8.4 × 105 1:9.6 × 1010 1:3.3 × 1012 1:2.1 × 1017

1:410 1:5000

(B) Promega Corporation, Madison, WI PowerPlex® 16 System Penta E, D18S51, D21S11, THO1, D3S1358, FGA, TPOX, D8S1179, vWA, Amelogenin, Penta D, CSF1PO, D16S539, D7S820, D13S317, and D5S818

1:1.8 × 1017

Source: Applied Biosystems, Foster City, CA (www.appliedbiosystems.com) and Promega Corporation, Madison, WI (www.promega.com).

D8S1179

D3S138

THO1

vWA

D19S20

AMEL

D5S818

D7S820

D21S11

D13S317

D16S539

TPOX

CSF1PO

D2S1338

D18S21

FGA

FIGURE 25.1 AmpFlSTR Identifiler™16-plex amplification results. The electropherogram shows 15 STR loci as well as the amelogenin (sex determining) marker. Individual STR loci are separated by amplicon size and dye label. Each row represents a different dye marker: 1, the 6-FAM labeled loci; 2, VIC labeled loci; 3, NED labeled loci; 4, PET labeled loci; an internal standard labeled in rox is also run simultaneously but is not shown. (Courtesy of Ada Nuñez, Florida International University.)

Separation of DNA for Forensic Applications Using Capillary Electrophoresis

767

100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 * Ladder.fsa 2 Blue ladder

D21S11

D8S1179

D7S820

CSF1PO

3000 2000 1000

* Ladder.fsa

2 Green Ladder

THO1

D3S138

D13S317

D2S1338

D16S539

2000 1000

* Ladder.fsa

2 Yellow ladder

D19S20

vWA

TPOX

D18S21

4000 3000 2000 1000

* Ladder.fsa

AMEL

2 Red

ladder

D5S818

2000

FGA

1500 1000 500

FIGURE 25.2 AmpFlSTR™Identifiler kit allelic ladder. A separation of the most common alleles for each STR loci. The STR analysis software utilizes this ladder as an external standard to assign the correct alleles to evidence samples such as is illustrated in Figure 25.1. Notice the presence of two base variant alleles (from the normal four base STR repeat) at the D21, D18, D19, and FGA loci. There is also a one base variant allele located in the THO1 locus. (Courtesy of Ada Nuñez, Florida International University.)

One of the major issues in the separation of these large multiplex sets by CE is the potential presence of variant alleles that differ from the standard 4 base repeat unit by 1 or 2 bp. Single base resolution and high precision are necessary over the range of fragment lengths up to 350 bp in order to reliably detect these variant alleles and distinguish them from artifacts, spikes, and noise. Another issue that is unique to this application is the role of the PCR in defining system sensitivity. Amplification of DNA concentrations less than 100 pg (about 17 cells) can produce stochastic intensity fluctuations leading to peak imbalance and occasional loss of signal (allele dropout). Low level mixtures may also be present in the electropherogram further complicating the interpretation of the data [32,33]. The ability to produce clear and unambiguous electropherograms is critical in criminal casework since DNA evidence may be the only information tying the suspect to the crime. Loss of peak intensity can complicate data interpretation. As a result, rigid rules have been developed for interpretation of signals below an intensity threshold that can be defined by the ability to reproducibly amplify DNA and detect it at a level above the system’s limit of quantification [34]. For mixtures such as those often encountered in sexual assault cases, special techniques have been developed for isolation of male DNA (sperm heads) from female epithelial cells [35]. These procedures aid the interpretation of the data by removing most of the female contribution to the profile. This process, known as differential extraction, can be performed by selective digestion of the epithelial cells followed by isolation of undigested sperm heads through centrifugation. The sperm fraction is further digested and extracted using dithiothreitol (DTT). However in certain situations such as the case with vasectomized males or multiple assailants, isolation of individual profiles is more difficult. In these cases careful analysis of the results may be necessary to reach

768

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

FIGURE 25.3 Mixed male/female DNA Profile. The amplification of a 1:1 mixture of male and female DNA using the Identifiler™STR multiplex. Note the presence of 2, 3, or 4 alleles at each locus indicating the presence of 2-, 1-, or 0-shared alleles. Also, notice the 3:1 ratio of X to Y sex typing alleles in the third panel due to the mixture of XX and XY alleles. Analysis performed on an ABI310 Genetic Analyzer. (Courtesy of Stephanie King and George Duncan, Broward County Sheriff’s Office Crime Laboratory.)

a conclusion. Figure 25.3 illustrates the electropherogram of a mixed profile. Note that the sample can be determined to be a mixture of male and female DNA based on the relative peak areas of the Amelogenin sex marker. In this case there is a 3/1 ratio of X- to Y-chromosome, suggesting a 1/1 mixture of male (XY) to female (XX) DNA. Once the separation is complete, the data are analyzed using specialized software that assigns an allele repeat number to each peak based on the alleles identified by the previously run allelic ladder. Statistical analyses can be performed using relevant population frequencies to determine the overall probability of another individual having an identical DNA profile [36]. Because the different genetic loci used in determining the profile are inherited independently, the individual probabilities of having a particular set of alleles at one locus can be multiplied together for all loci producing a random match probability for the 13 CODIS loci of 1×10−15 [24]. For all intents and purposes, the data from a single source profile such as that shown in Figure 25.1 can be used to provide identification of a suspect [37]. Analysis of more complex samples such as those containing multiple contributors (Figure 25.3) or related individuals is less straightforward and requires in-depth statistical analysis [38–41].

25.4.2 MINI STRs Large STR multiplex sets are used for most forensic evidence samples. However, biological fluids or materials left behind in a crime scene may be degraded because of exposure to a variety of harsh

Separation of DNA for Forensic Applications Using Capillary Electrophoresis

769

(a)

(b)

FIGURE 25.4 (See color insert following page 810.) A comparison between a 9947 control DNA sample (a) and a degraded DNA sample (b) extracted from a recovered bone fragment. There is an evident loss of signal at the larger size loci in sample b (arrow) due to sample degradation. The sample is extracted and amplified DNA prepared using the Promega Powerplex STR multiplex kit and consists of 16 separate genetic loci labeled with three different dyes and separated via capillary gel electrophoresis using the ABI 310 genetic analyzer. (Courtesy of Kerry Opel, International Forensic Research Institute, Florida International University.)

conditions and/or due to the presence of various contaminants that have been mixed with the sample [42,43]. In both cases, the result may be a partial DNA profile (Figure 25.4) in which some alleles are missing from the profile or are present below a laboratory’s interpretational threshold [44,45]. Characteristically, the larger alleles lose intensity due to increasing decomposition of the template into smaller fragments. The resulting partial electropherograms are far less definitive than a full profile since fewer loci are available for statistical analysis. This leads to an increase in the probability of finding a random unrelated individual in the general population with a matching profile. In these situations, one approach has been to perform further testing using mtDNA sequencing. Since there are multiple copies of mitochondrion in each cell, the likelihood of obtaining a result is greatly increased when compared to nuclear DNA. However, mtDNA analysis involves difficult and expensive analytical procedures, and because it is inherited solely through the female line, maternally related individuals will all have the same profile and statistical results are much less conclusive. In recent years, investigators driven by this issue have developed a viable alternative—the use of mini STRs. Mini STRs are reduced size STR amplicons that can be obtained by redesigning the amplification primers in such a way that they bind closer to the STR repeat regions [46–48]. To keep the fragment size as short as possible only one or two STR loci can be used in each dye lane. These shorter amplicons can be detected when the original template is too fragmented to properly amplify (Figure 25.5). Many of these STR markers targeted by modified primers are the same as those already established for the CODIS database [46]. Concordance studies between the mini STR and the commercially available STR primers have been performed and some discrepancies (0.2%) have been observed in the allele calls. These, however, can be explained by the fact that the primers from the mini STRs bind at different locations than those of the original STRs, thus insertions/deletions can affect primer

770

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

FIGURE 25.5 Mini STR amplification. The electropherogram illustrates the analysis of an extracted and amplified bone sample. The electropherogram consists of four panels showing three STR loci and an internal lane standard run simultaneously and detected with four different dyes. Unlike the larger multiplexes illustrated in previous figures, only one locus appears in each dye lane in an effort to keep the amplified products as short as possible. (Courtesy of Kerry Opel, International Forensic Research Institute, Florida International University.)

binding resulting in an apparent discrepancy in allele size [49]. This problem is being addressed through alternative primer designs, and researchers are continuing to identify new mini STRs to increase the information content of STR assays [50].

25.4.3 MITOCHONDRIAL DNA ANALYSIS In situations in which minimal DNA can be recovered from a sample due to severe degradation or lack of recoverable DNA, mtDNA can be exploited due to its relatively small size and presence in multiple copies within the cell [51]. The mitochondrion is a self-replicating organelle that is able to synthesize its own DNA [52]. The DNA is circular and contains 16,569 bp [53]. It contains both coding and noncoding regions, and it is in the noncoding hypervariable regions, HVI and HVII, where information relevant to forensic information is found [18,54]. These hypervariable regions have high-mutation rates that enable them to be a useful tool for human genetic analysis [55]. Owing to its small size, mtDNA does not contain STRs or other repetitive elements and instead analysis relies on variations in sequences [51,56]. Mutation detection takes place through the analysis of single nucleotide polymorphisms (SNPs); deletions, additions, and substitutions of various nucleotides in HVI and HVII. The treatment the samples receive before being loaded into the CE system is also different from that of the STRs. Because these samples are generally present at low copy number and are highly degraded, extensive use of control samples with strict isolation protocols is necessary in order to avoid cross contamination and maintain data reliability. The samples are first amplified and the resultant native, unlabeled PCR products are quantified using CGE or microfluidic CGE with fluorescent intercalating dyes to determine input levels and

Separation of DNA for Forensic Applications Using Capillary Electrophoresis

771

FIGURE 25.6 Quantitative analysis of PCR-amplified mtDNA by microfluidic CE. Determination of the overall quality and quantity of the DNA before sequencing is important to assure high-quality results. The figure illustrates the detection of an mtDNA amplicon from the HVI region located between two internal standards used for sizing. The analysis is performed on an extracted blood stain by an Agilent 2100 bioanalyzer using fluorescent detection with an intercalating dye. (Courtesy of the DNA Analysis Unit II at the FBI Laboratory, Quantico, VA.)

sample quality for the subsequent sequencing reaction [16,57] (Figure 25.6). Sequencing reactions are then performed using the PCR template, and products are separated using denaturing CGE. The results are compared to a reference sequence to catalog the specific point mutations that are present [58] (Figure 25.7). The statistical analyses of these data are not as definitive as that obtained from STR typing since mitochondria are transmitted directly from mothers to their progeny and there is no admixing or shuffling of genetic information such as occurs in meiosis. Thus mtDNA analysis cannot produce the high-statistical certainty of identification produced in STR typing and instead is most useful as a means for maternal lineage determination [59,60]. However, in situations in which the only evidence available is badly degraded or where only a few cells containing DNA are present (such as the situation in which a single shed hair is recovered), mtDNA may be the only way useful genetic information can be obtained [51,61–63].

25.4.4 Y-STRs As with mtDNA, which provides maternal lineage, the Y-chromosome can be used to provide paternal lineage. The first Y-STR was discovered in 1992 [64] Since then, a number of additional Y-STRs have been validated for use in the forensic field [65,66] and several commercial multiplex amplification kits are now available (Table 25.3). Initially used for paternity testing and rape case scenarios [67,68], Y-STRs are now being used to aid in missing persons investigations [69], genealogical research, and evolutionary studies [70–72]. Y-chromosome markers are particularly useful in the detection of small amounts of male DNA in the presence of an overwhelming abundance of female DNA. This type of STR loci are most useful when trying to isolate the male fraction from a DNA mixture (Figure 25.8) [73,74]. .

772

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

FIGURE 25.7 (See color insert following page 810.) Mitochondrial DNA profile of an HVII sequence from a hair shaft. A repetitive analysis of the same hair sample is used to compare and align the sequences. The nuclear DNA in hair is often badly degraded and difficult to amplify. mtDNA provides an alternative procedure that can provide a DNA haplotype for forensic DNA profiling based on point mutations detected in the sequence when compared to a reference sample. Analysis performed on an ABI 310 Genetic Analyzer. (Courtesy of the DNA Analysis Unit II at the FBI Laboratory, Quantico, VA.)

TABLE 25.3 Validated Y-STR Loci for Forensic Casework Locus DYS393 DYS392 DYS391 DYS389I DYS389II Y-GATA A7.2 DYS438 DYS385 DYS19 DYS425 DYS388 DYS390 DYS439 DYS434 DYS437 Y-GATA C.4 Y-GATA A7.1 Y-GATA H.4

Repeat AGAT TAT TCTA (TCTG)(TCTA) (TCTG)(TCTA) TAGA TTTTC GAAA TAGA TGT ATT (TCTA)(TCTG) AGAT ATCT TCTA TATC ATAG TAGA

Size 108–132 236–263 275–295 239–263 353–385 174–190 203–233 252–300 242–254 104–110 119–131 200–251 242–258 106–116 188–192 250–269 104–112 130–143

Source: Adapted from Daniels, D.L., Hall, A.M., and Ballantyne J., J Forensic Sci, 43, 668, 2004. With permission.

Separation of DNA for Forensic Applications Using Capillary Electrophoresis

sample file

773

sample name

FIGURE 25.8 A Y-STR profile of the mixture in Figure 25.3 using the ABI Y Filer™Y chromosomal STR multiplex. Note that the DY385 locus has two alleles due to a duplication of the sequence on the Y chromosome. Analysis performed using an ABI 310 Genetic Analyzer. (Courtesy of Stephanie King and George Duncan, Broward County Sheriff’s Office Crime Laboratory.)

An example of such a scenario might be a fingernail scraping from a female victim or the detection of a mixed DNA profile from the handle of an automobile. The profile depicted in Figure 25.9a results from an extract taken from a knife handle that includes DNA from the male suspect along with two female profiles. Use of the Y profile (Figure 25.9b) permits isolation of just male DNA, although it should be noted that all of the suspect’s male relatives from his paternal lineage would also have the same Y-haplotype. Statistical analysis of the result involves performing an estimate of the frequency of a given profile in a database. Frequency estimates for Y profiles are much less specific than those obtained with autosomal STRs as individual allele frequencies cannot be multiplied together. Nevertheless, the data provide important information regarding the potential placement of an individual at a crime scene.

25.4.5 SINGLE NUCLEOTIDE POLYMORPHISMS The various types of point mutations detected via mtDNA sequencing can also be targeted in nuclear DNA. SNPs are particularly useful in ethnicity testing and in the analysis of highly degraded or compromised samples [57,75,76]. In the human genome SNPs occur every 1000–2000 bp, thereby

774

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a)

(b)

FIGURE 25.9 A comparison of (a) a mixed DNA profile (two females and one male contributor) followed by (b) the Y-STR profile of the same sample. The sample was collected with a single surface swab (moistened with DI water) from a knife handle (homicide case). Amplification was performed using PowerPlex®Y with 1.2 ng and Profiler Plus™with 0.8 ng of template DNA. “The male component DNA obtained from the knife handle and the suspect has the same Y-haplotype; therefore, the suspect could not be eliminated as the source of the male DNA in the mixture. The results also indicate the presence of only one male contributor in the DNA extract. It should be noted that all of the suspect’s male relatives from his paternal lineage would also have the same Y-haplotype.” (Courtesy of DNA Unit—Orange County Sheriff-Coroner Department, CA.)

accounting for about 90% of genetic variation [75,77,78]. Unlike tandem repeats, SNPs are found both in the coding and noncoding regions of the DNA molecule. They can play an important role in the field of forensics as they have, in contrast to STRs, lower mutation rates and may eventually be linked to physical features such as hair color, stature, and skin shade [79,80]. Unfortunately, SNP

Separation of DNA for Forensic Applications Using Capillary Electrophoresis

775

systems have limited numbers of alleles and therefore many more loci are required to perform a full genetic profile than is the case with STRs. In fact, about 50 SNPs would require analysis to achieve the same level of statistical uniqueness developed with 13 STRs [78]. In addition, SNPs cannot easily be used in mixture studies as fragment sizes overlap making it difficult to isolate multiple contributors to a profile. SNPs can be detected using a single base primer extension assay in which a special primer is designed to target a known SNP location [81]. A polymerase is then added and used to probe the site of the polymorphism (the next unincorporated base pair) through the addition of a terminal fluorescently labeled dideoxynucleotide (ddNTP). The SNP assay can be multiplexed by simultaneously labeling multiple numbers of these loci and using changes in primer length to permit all locations to be detected simultaneously. Like mtDNA sequencing and mini STRs analysis, SNPs can be valuable in the recovery of information from degraded DNA. Figure 25.10 illustrates the comparison of results from the standard STR typing of a degraded DNA sample followed by SNP typing of that same DNA extract. The SNP amplicons are much shorter and provide additional genetic information to assist in the identification of the partial STR profile.

25.4.6 MUTATION DETECTION Although more commonly used in oncology and detection of genetic diseases [82–84], mutation detection techniques can provide useful information in forensic analysis. Sequence polymorphisms resulting from one or more base pair changes in PCR-amplified DNA fragments can be exploited through differences in melting points, heteroduplex formation, or conformational variations. A variety of different CE procedures exist to detect these differences, including single-stranded conformational polymorphism (SSCP), heteroduplex polymorphism, and constant denaturant CE. Using these techniques, DNA fragments of the same or similar length that would otherwise comigrate can be differentiated based on mobility differences resulting from the effect of temperature, denaturant concentration, or rehybridization effects. These types of mobility assays have been used in paternity disputes [85] and in ABO allele discrimination [86]. These techniques are relatively inexpensive to perform and can be highly sensitive to slight differences in sequence [87] making them highly useful in the detection of previously unknown mutational events.

25.4.7 NONHUMAN DNA Nonhuman DNA can provide crucial information in a variety of crime scenarios [88]. Items found at a crime scene, such as animal hair, plant material, fungal spores, and soils, may all contain recoverable DNA. Genetic markers such as STRs have been isolated from these types of materials and a number of different loci have been validated for use in criminal casework [89,90]. For other items such as microbial DNA, soil and plant materials, sequence information may not exist, and various exploratory procedures must be used to recover genetic information. 25.4.7.1 Animal DNA Feline and canine STR analyses have been used in forensic casework [91–94] and specific kits have been developed to determine the genetic fingerprint of both animals [95,96]. Canine genetic material has also been analyzed via CE using mtDNA sequencing [97]. As with human samples, canine DNA can help link a suspected animal to crimes such as bite attacks or mauling. Similarly, it can also be used to exclude the animal from implication in attacks or unlawful deaths [98]. More commonly, however, dog hair and other animal hair has been found associated with victims at crime scenes [94,99]. Figure 25.11 illustrates an electropherogram of a suspect dog hair that might be used to associate an animal with the victim of an abduction. Genetic analysis of this material provides further information on the circumstances of the crime and can connect the victim to a particular

776

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

FIGURE 25.10 (See color insert following page 810.) The recovery of information from degraded DNA using the Identifiler™STR multiplex and an SNP multiplex, both analyzed using the ABI 310 genetic analyzer. The top electropherogram depicts a degraded sample amplified via STRs. The bottom profile depicts a SNP multiplex on the same sample. The STR profile is blank for many larger alleles such as D7 CSF, D16, D2, TPOX, D18, and FGA. The smaller sized SNP fragments permit recovery of additional genetic information from the sample. (From Vallone, P.M., Decker, A.E, and Butler, J.M., The evaluation of Autosomal SNP assays, Chemical Science and Technology Laboratory, National Institute of Standards and Technology, Div.831.)

location or suspect. Although more commonly used for lineage analysis, other genetic material such as equine and bovine DNA could also be used in criminal casework [100]. 25.4.7.2 Botanical DNA Plant DNA analysis has been increasing over the past couple of years because of its ability to pinpoint the origin of drug-related plant material [101,102]. Plant STRs are not yet well characterized; however, other molecular methods can be combined with capillary gel electrophoresis to establish the identity of plant species [103]. These techniques include amplified fragment length polymorphism (AFLP), random amplification of polymorphic DNA (RAPD), and other similar random primer annealing techniques used for mutation detection. AFLP is a molecular tool in which the DNA is cut by a combination of restriction enzymes for subsequent amplification utilizing relatively nonspecific primers that are tagged with a fluorescent dye [104,105]. These primers will anneal only to the fraction of the cut fragments that contain a short complimentary sequence, leaving behind those that do not. Since DNA sequences vary, the cut

Separation of DNA for Forensic Applications Using Capillary Electrophoresis

777

FIGURE 25.11 Canine STR analysis. Ten canine markers were successfully amplified and analyzed after extracting DNA from a single dog hair. (Courtesy of Lilliana Moreno, Forensic DNA Profiling Facility, Florida International University.)

sites produced by the selected restriction enzymes will vary between the different plant materials submitted for analysis. The result is a set of DNA fragments of different sizes and intensities that can be related to a particular plant cultivar. The specific fragments can be isolated and detected by capillary gel electrophoresis. Various computer software are then used to define and catalog the differences between samples. AFLP has proven to be a particularly valuable tool for establishing plant identification and for tracing plant material back to its original source [106]. Figure 25.12 illustrates the analysis of a seized marijuana sample using AFLP with multichannel fluorescent CE detection.

25.4.7.3 Microbial DNA Recent events involving the use of microbes as bioterrorists’ weapons have intensified research in the developing field of microbial forensics. This new field couples law enforcement efforts with existing procedures to identify patterns in disease outbreaks, determine the pathogen involved, control its spread, and trace the microorganism back to its source [107]. A variety of techniques are currently being used as tools for the classification and identification of microorganisms and these procedures can also be used to assist investigators in the detection of criminal acts. While most of the reported incidents to date involve hoaxes with nonpathogenic material, the consequences of such an attack require that laboratories be properly prepared to meet the threat. In addition, these same techniques can also be used to examine trace evidence such as soils left behind following a criminal act [108,109]. A number of molecular techniques for microbial detection and analysis are listed below. Terminal restriction fragment length polymorphism (TRFLP) uses end-labeled primers that will bind to specific primer sequences used when amplifying DNA by PCR [110,111]. After digestion of the PCR products with restriction enzymes, the samples are loaded into the CE system for separation.

778

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

FIGURE 25.12 Genetic analysis of a marijuana sample using AFLP. The sample consists of amplified DNA fragments characteristic of the plant overlaid on top of an internal lane standard used to size the individual fragments. Analysis performed on an ABI 310 genetic analyzer. (Courtesy of Ira Lurie, Yin Shen, and Bruce McCord.)

FIGURE 25.13 Soil microbial DNA profile. The 16S rRNA gene hypervariable region V1_V2 was amplified using ALH technique. The different DNA fragments identified in the figure depict the different microbes present in the soil. These data can then be used to aid in the identification of soils associated with a crime scene and to help answer important questions such as has a body been moved from its current location. Analysis performed using an ABI 310 genetic analyzer. (Courtesy of Lilliana Moreno, Forensic DNA Profiling Facility, Florida International University.)

Separation of DNA for Forensic Applications Using Capillary Electrophoresis

779

This method has been extensively used in conjunction with the small ribosomal subunit (16S rRNA) to establish differences between microbial entities [112,113] and was the technique first used in the analysis of microbial communities for forensic purposes [108]. Subsequent researchers in microbial forensics examined a method known as amplicon length heterogeneity (ALH), which is commonly used in microbial ecology. ALH bases its profiles on differences in length within select hypervariable domains of the 16S rRNA genes. DNA extracts from different microbial communities are amplified using universal primers that bind outside of the variable region and are capable of hybridizing to the majority of target organisms. The output is a pattern of peaks that provide, just like STRs, a unique pattern that is used to differentiate microbial communities (Figure 25.13). Unlike TRFLP, this highly reproducible method does not require restriction endonucleases but is based on the natural variation in sequence lengths of specific regions within the gene [114,115]. Because of its ubiquitous nature, soil can be a valuable evidence in crime scene investigations; however, it is seldom used because the physical analysis is complicated and requires experts in the field of geology. A novel approach utilizing ALH-PCR to discriminate between soils for forensic purposes has been developed and tested and provides a promising foundation for the future of the application [109].

25.5 CONCLUSIONS Capillary electrophoresis technology has become an indispensable tool for forensic scientists in the biology field since it is able to provide valuable information to aid in the process of law enforcement. The primary application of the technique is in the qualitative analysis of STRs. Isolation of STR mixtures is also possible using relative peak heights. Other applications of CE include quantitative analysis of PCR products, mtDNA sequencing, and mutation detection for the analysis of plant and bacterial DNA. Based on the performance of the methods illustrated above, it is reasonable to expect future researchers and practitioners to continue working to exploit the capabilities of this robust scientific technique and its application to criminal investigations.

ACKNOWLEDGMENTS The authors would like to thank John Butler, Dee Mills, Alice Isenberg, Kate Theisen, Pete Vallone, Ed Buse, Ada Nuñez, Kerry Opel, Ira Lurie, Stephanie King, Yin Shen, and George Duncan for their contributions to this work. Major funding from the National Institute of Justice is also gratefully acknowledged. Points of view in the document are those of the authors and do not necessarily represent the official view of the U.S. Department of Justice.

REFERENCES 1. Tagliaro, F. and Smith, F., Forensic capillary electrophoresis, Trends Anal Chem, 15, 513, 1996. 2. Northrop, D., McCord, B., and Butler, J., Forensic applications of capillary electrophoresis, J Capillary Electrophor, 1, 158, 1994. 3. Barron, A., Sunada, W., and Blanch, H., Capillary electrophoresis of DNA in uncrosslinked polymer solutions: Evidence for a new mechanism of DNA separation, Biotechnol Bioeng, 52, 259, 1996. 4. Arnheim, N., Polymerase chain reaction strategy, Annu Rev Biochem, 61, 131, 1992. 5. Luckey, J. and Smith, L., Advances in DNA sequencing technology, Proc SPIE, 1891, 21, 1993. 6. Fregeau, C. and Fourney, R., DNA typing with fluorescently tagged short tandem repeats: A sensitive and accurate approach to human identification, BioTechniques, 15, 100, 1993. 7. Cohen, A., Najarian, D., Paulus, A., et al., Rapid separation and purification of oligonucleotides by high-performance capillary gel electrophoresis, Proc Natl Acad Sci USA, 85, 9660, 1988. 8. Kasper, T., Melera, M., Gozel, P., et al., Separation and detection of DNA by capillary electrophoresis, J Chromatogr, 458, 303, 1988.

780

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

9. Albarghouthi, M., Buchholz, B., Huiberts, P., et al., Poly-N-hydroxyethylacrylamide (polyDuramide): A novel, hydrophilic, self-coating polymer matrix for DNA sequencing by capillary electrophoresis, Electrophoresis, 23, 1429, 2002. 10. Tseng, W. and Chang, H., A new strategy for optimizing sensitivity, speed and resolution in capillary electrophoretic separation of DNA, Electrophoresis, 22, 763, 2001. 11. Mandabhushi, R., Separation of 4-color DNA sequencing extension products in noncovalently coated capillaries using low viscosity polymer solutions, Electrophoresis, 19, 224, 1998. 12. Rocheleau, M., Grey, R., Chen, D., et al., Formamide modified polyacrylamide gels for DNA sequencing by capillary gel electrophoresis, Electrophoresis, 13, 484, 1992. 13. Guttman, A. and Cooke, N., Denaturing capillary gel electrophoresis, Am Biotechnol Lab, 9, 10, 1991. 14. Brining, S., Chen, N., Yi, D., et al., Gel electrophoretic distinction between toxic and nontoxic forms of beta-amyloid (1-40), Electrophoresis, 20, 1398, 1999. 15. Ozawa, S., Sugano, K., Sonehere, T., et al., High resolution for single-strand conformation polymorphism analysis by capillary electrophoresis, Anal Chem, 76, 6122, 2004. 16. Butler, J., McCord, B., Jung, J., et al., Quantitation of polymerase chain reaction products by capillary electrophoresis using laser fluorescence, J Chromatogr B, 658, 271, 1994. 17. Butler, J., Buel, E., Crivellente, F., et al., Forensic DNA typing by capillary electrophoresis using the ABI Prism 310 and 3100 genetic analyzers for STR analysis, Electrophoresis, 25, 1397, 2004. 18. Wilson, M., Stoneking, M., Holland, M., et al., Guidelines for the use of mitochondrial DNA sequencing in forensic science, Crime Lab Digest, 20, 68, 1993. 19. LaFountain, M., Schwartz, M., Svete, P., et al., TWGDAM validation of the AmpFlSTR profiler Plus and AmpFlSTR Cofiler STR multiplex systems using capillary electrophoresis, J Forensic Sci, 46, 1191, 2001. 20. Hartzell, B., Graham, K., and McCord, B., Response of short tandem repeat systems to temperature and sizing methods, Forensic Sci Int, 133, 228, 2003. 21. Nock, T., Dove, J., McCord, B., et al., Temperature and pH studies of short tandem repeat systems using capillary electrophoresis at elevated pH, Electrophoresis, 22, 755, 2001. 22. Budowle, B., Chakraborty, R., Carmody, G., et al., Source attribution of a forensic DNA profile, Forensic Sci Comm, 2, 1, 2000. 23. Moxon, E. and Wills, C., DNA microsatellites: Agents of evolution? Sci Am, 280, 94, 1999. 24. Butler, J., Forensic DNA Typing: Biology, Technology, and Genetics of STR Markers, 2nd edn. Academic Press, San Diego, CA, 2005. 25. www.fbi.gov, accessed 12/2006. 26. Budowle, B., Moretti, T. R., Niezgoda, S. J., et al., CODIS and PCR-based short tandem repeat loci: Law enforcement tools., Forensic Science Research and Training Center, FBI Academy, 73–77. 27. Ruitberg, C. M., Reeder, D. J., and Butler, J. M., STRBase: A short tandem repeat DNA database for the human identity testing community, Nucleic Acids Res, 29, 320, 2001. 28. http://www.FBi.gov/hg/lab/codis/program.htm, accessed 8/21/07. 29. Cotton, E., Allsop, R., Guest, J., et al., Validation of the AMPFlSTR SGM plus system for use in forensic casework, Forensic Sci Int, 112, 151, 2000. 30. Krenke, B., Tereba, A., Anderson, S., et al., Validation of a 16-locus fluorescent multiplex system, J Forensic Sci, 47, 773, 2002. 31. Hobson, D., Smerick, J., Brown, A., et al., A summary of the validation of the STR analysis by capillary electrophoresis: development of interpretation guidelines for the Profiler Plus and Cofiler systems for use in forensic casework, in Proceedings from the 10th International Symposium on Human Identification, Promega Corp, Coronoado Springs (1999). 32. Gill, P., Sparkes, R., Pinchin, R., et al., Interpreting simple STR mixtures using allele peak areas, Forensic Sci Int, 91, 41, 1997. 33. Ladd, C., Lee, H., Yang, N., et al., Interpretation of complex forensic DNA mixtures, Croat Med J, 42, 244, 2001. 34. Gill, P., Whitaker, J., Flaxman, C., et al., An investigation of the rigor of interpretation rules for STRs derived from less than 100 pg of DNA, Forensic Sci Int, 112, 17, 2000. 35. Gill, P., Jeffreys, A., and Werrett, D., Forensic application of DNA “fingerprints,” Nature, 318, 577, 1985.

Separation of DNA for Forensic Applications Using Capillary Electrophoresis

781

36. Weir, B., The second national research council report on forensic DNA evidence, Am J Hum Genet, 59, 497, 1996. 37. Sajantila, A. and Budowle, B., Identification of individuals with DNA testing, Ann Med, 23, 637, 1991. 38. Fung, W. and Hu, Y., Interpreting DNA mixtures based on the NRC-II recommendation 4.1, Forensic Sci Comm, 2, 1–11, 2000. 39. Weir, B., Triggs, C., Starling, L., et al., Interpreting DNA mixtures, J Forensic Sci, 42, 213, 1997. 40. Clayton, T., Whitaker, J., Sparkes, R., et al., Analysis and interpretation of mixed forensic stains using DNA STR profiling, Forensic Sci Int, 91, 55, 1998. 41. Evett, I., Gill, P., and Lambert, J., Taking account of peak areas when interpreting mixed DNA profiles, J Forensic Sci, 43, 62, 1998. 42. Ludes, B., Pfitzinger, H., and Mangin, P., DNA fingerprinting from tissues after variable postmortem periods, J Forensic Sci, 38, 686, 1993. 43. Smith, S. and Morin, P., Optimal storage conditions for highly dilute DNA samples: A role for trehalose as a preserving agent, J Forensic Sci, 50, 1101, 2005. 44. Whitaker, J., Clayton, T., Urquhart, A., et al., Short tandem repeat typing of bodies from a mass disaster: High success rate and characteristic amplification patterns in highly degraded samples, BioTechniques, 18, 670, 1995. 45. Anderson, T., Ross, J., Roby, R., et al., A validation study for the extraction and analysis of DNA from human nail material and its application to forensic casework, J Forensic Sci, 44, 1053, 1999. 46. Butler, J. M., Shen, Y., and McCord, B. R., The development of reduced size STR amplicons as tools for analysis of degraded DNA, J Forensic Sci, 48, 1054, 2003. 47. Wiegand, P. and Kleiber, M., Less is more—Length reduction of STR amplicons using redesigned primers, Int J Legal Med, 114, 285, 2001. 48. Tsukada, K., Takayanagi, K., Asamura, H., et al., Multiplex short tandem repeat typing in degraded samples using newly designed primers for the TH01, TPOX, CSF1PO, and vWA loci, Legal Med, 4, 239, 2002. 49. Drabek, J., Chung, D., Butler, J., et al., Concordance study between miniplex assays and a commercial STR typing kit, J Forensic Sci, 49, 859, 2004. 50. Coble, M. and Butler, J., Characterization of new miniSTR loci to aid analysis of degraded DNA, J Forensic Sci, 50, 43, 2005. 51. Bender, K., Schneider, P., and Rittner, C., Application of mtDNA sequence analysis in forensic casework for the identification of human remains, Forensic Sci Int, 113, 103, 2000. 52. Saccone, C., Gissi, C., Lanave, C., et al., Evolution of the mitochondrial genetic system: An overview, Gene, 261, 153, 2000. 53. Anderson, S., Bankier, A., Barrell, B., et al., Sequence and organization of the human mitochondrial genome, Nature, 290, 457, 1981. 54. Holland, M. and Parsons, T., Mitochondrial DNA sequence analysis—Validation and use for forensic casework, Forensic Sci Rev, 11, 21, 1999. 55. Jazin, E., Soodyall, H., Jalonen, P., et al., Mitochondrial mutation rate revisited: Hot spots and polymorphisms, Nat Genet, 18, 109, 1998. 56. Wilson, M., DiZinno, J., Polanskey, D., et al., Validation of mitochondrial DNAsequencing for forensic casework analysis, Int J Legal Med, 108, 68, 1995. 57. Kline, M., Vallone, P., Redman, J., et al., Mitochondrial DNA typing screens with control region and coding region SNPs, J Forensic Sci, 50, 377, 2005. 58. Isenberg, A. and Moore, J., Mitochondrial DNA analysis at the FBI laboratory, Forensic Sci Comm, 1, 1999. 58. Cann, R., Stoneking, M., and Wilson, A., Mitochondrial DNA and human evolution, Nature, 325, 31–36 1987. 59. Baasner, A., Schafer, C., Juge, A., et al., Polymorphic sites in human mitochondrial DNA control region sequences: Population data and maternal inheritance, Forensic Sci Int, 98, 169, 1998. 60. Budowle, B., Allard, M., Wilson, M., et al., Forensics and mitochondrial DNA: Applications, debates, and foundations, Annu Rev Genom Hum G, 4, 119, 2003. 61. Stone, A., Starrs, J., and Stoneking, M., Mitochondrial DNA analysis of the presumptive remains of Jesse James, J Forensic Sci, 46, 173, 2001.

782

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

62. vonWurmb-Schwark, N., Harbeck, M., Wiesbrock, U., et al., Extraction and amplification of nuclear and mitochondrial DNA from ancient and artificially aged bones, Legal Med, 5, 169, 2003. 63. Roewer, L. and Epplen, J., Rapid and sensitive typing of forensic stains by PCR amplification of polymorphic simple repeat sequences in case work, Forensic Sci Int, 53, 163, 1992. 64. Daniels, D., Hall, A., and Ballantyne, J., SWGDAM developmental validation of a 19-locus Y-STR system for forensic casework, J Forensic Sci, 49, 1, 2004. 65. Hall, A. and Ballantyne, J., The development of an 18-locus Y-STR system for forensic casework, Anal Bioanal Chem, 376, 1234, 2003. 66. Kayser, M. and Sajantila, A., Mutations at Y-STR loci: Implications for paternity testing and forensic analysis, Forensic Sci Int, 118, 116, 2001. 67. Corach, D., Filgueira, R., Marino, M., et al., Routine Y-STR typing in forensic casework, Forensic Sci Int, 118, 131, 2001. 68. Huffine, E., Crews, J., Kennedy, B., et al., Mass identification of persons missing from the break-up of the former Yugoslavia: Structure, function, and role of the International Commission of Missing Persons, Croat Med J, 42, 271, 2001. 69. Butler, J., Recent developments in Y-short tandem repeat and Y-single nucleotide polymorphism analysis, Forensic Sci Rev, 15, 91, 2003. 70. Hedman, M., Pimenoff, V., Lukka, M., et al., Analysis of 16 Y STR loci in the Finnish population reveals a local reduction in the diversity of male lineages, Forensic Sci Int, 142, 37, 2004. 71. Silva, D., Carvalho, E., Costa, G., et al., Y-chromosome genetic variation in Rio De Janeiro population, Am J Hum Biol, 18, 829, 2006. 72. Cerri, N., Ricci, U., Sani, I., et al., Mixed stains from sexual assault cases: Autosomal or Y-chromosome short tandem repeats? Croat Med J, 44, 289, 2003. 73. Hanson, E. and Ballantyne, J., A highly discriminating 21 locus Y-STR “megaplex” system designed to augment the minimal haplotype loci for forensic casework, J Forensic Sci, 49, 40, 2004. 74. Budowle, B., SNP typing strategies, Forensic Sci Int, 146, 139, 2004. 75. Vallone, P., Decker, A., and Butler, J., Allele frequencies for 70 autosomal SNP loci with US Caucasian, African American, and Hispanic samples, Forensic Sci Int, 149, 279, 2005. 76. Stoneking, M., Single nucleotide polymorphisms: From the evolutionary past, Nature, 409, 821, 2001. 77. Venter, J. C., Adams, M. D., Myers, E. W., et al., The sequence of the human genome, Science, 291, 1304, 2001. 78. Amorim, A. and Pereira, L., Pros and cons in the use of SNPs in forensic kinship investigation: A comparative analysis with STRs, Forensic Sci Int, 150, 17, 2005. 79. Payseur, B. and Cutter, A., Integrating patterns of polymorphism at SNPs and STRs, Trends Genet, 22, 424, 2006. 80. Syvanen, A., Accessing genetic variation: Genotyping single nucleotide polymorphisms, Nat Rev Genet, 2, 930, 2001. 81. Tian, H., Jaquins-Gerstl, A., Munro, N., et al., Single-strand conformation polymorphism analysis by capillary and microchip electrophoresis: A fast, simple method for detection of common mutations in BRCA1 and BRCA2, Genomics, 63, 25, 2000. 82. Atha, D., Kasprzak, W., O’Connell, C., et al., Prediction of DNA single-strand conformation polymorphism: Analysis by capillary electrophoresis and computerized DNA modeling, Nucleic Acids Res, 29, 4643, 2001. 83. Orita, M., Iwahana, H., Kanazawa, H., et al., Detection of polymorphisms of human DNA by gel electrophoresis as single-strand conformation polymorphisms, Proc Natl Acad Sci USA, 86, 2766, 1989. 84. Fukuda, M. and Tamaki, Y., Subtyping of D11S488 STR alleles by single-strand conformation polymorphism (SSCP) analysis in two cases of disputed parentage, Nihon Hoigaku Zasshi, 52, 42, 1998. 85. Tsai, L., Kao, L., Chang, J., et al., Rapid identification of the ABO genotypes by their single-strand conformation polymorphism, Electrophoresis, 21, 537, 2000. 86. Sunnucks, P., Wilson, A., Beheregaray, L., et al., SSCP is not so difficult: The application and utility of single-stranded conformation polymorphism in evolutionary biology and molecular ecology, Mol Ecol, 9, 1699, 2000.

Separation of DNA for Forensic Applications Using Capillary Electrophoresis

783

87. Day, A., Nonhuman DNA testing increases DNA’s power to identify and convict criminals, Silent Witness, 6, 2001. 88. Hellmann, A., Rohleder, U., Eichmann, C., et al., A proposal for standardization in forensic canine DNA typing: Allele nomenclature of six canine-specific STR loci, J Forensic Sci, 51, 274–281, 2006. 89. Eichmann, C., Berger, B., and Parson, W., Aproposed nomenclature for 15 canine-specific polymorphic STR loci for forensic purposes, Int J Legal Med, 118, 249, 2004. 90. Sensabaugh, G. and Kaye, D., Non-human DNA evidence, Jurimetrics J, 38, 1, 1998. 91. Eichmann, C., Berger, B., Steinlechner, M., et al., Estimating the probability of identity in a random dog population using 15 highly polymorphic canine STR markers, Forensic Sci Int, 151, 37, 2005. 92. Halverson, J. and Basten, C., A PCR multiplex and database for forensic DNA identification of dogs, J Forensic Sci, 50, 352, 2005. 93. State of Washington v. Kenneth J. Leuluaialii and Geroge J. Tuilefano, in No. 43507-8-I consolidated with No. 437-1-I, Court of Appeals of the State of Washington, 2003. 94. Butler, J., David, V., O’Brien, S., et al., The MeowPlex: A new DNA test using tetranucleotide STR markers for the domestic cat, Profiles DNA, 5, 7, 2002. 95. www.appliedbiosystems.com, accessed 12/2006. 96. State of California v. David Alan Westerfield, in Case No. SCD165805, Superior Court of California, County of San Diego Central Division, 2002. 97. Braunner, P., Reshef, A., and Gorski, A., DNA profiling of trace evidence—Mitigating evidence in a dog biting case, J Forensic Sci, 46, 1232, 2001. 98. Menotti-Raymond, M., David, V., and O’Brien, S., Pet cat hair implicates murder suspect, Nature, 386, 774, 1997. 99. Giovambattista, G., Ripoli, M., Liron, J., et al., DNA typing in a cattle stealing case, J Forensic Sci, 46, 1484, 2001. 100. Gilmore, S., Peakall, R., and Robertson, J., Short tandem repeat (STR) DNA markers are hypervariable and informative in Cannabis sativa: Implications for forensic investigations, Forensic Sci Int, 131, 65, 2003. 101. Hsieh, H., Hou, R., Tsai, L., et al., A highly polymorphic STR locus in Cannabis sativa, Forensic Sci Int, 131, 53, 2003. 102. Yoon, C., Forensic science: Botanical witness for the prosecution, Science, 260, 894, 1993. 103. Vos, P., Hogers, R., Bleeker, M., et al., AFLP: A new technique for DNA fingerprinting, Nucleic Acids Res, 23, 4407, 1995. 104. Coyle, H., Ladd, C., Palmbach, T., et al., The green revolution: Botanical contributions to forensics and drug enforcement, Croat Med J, 42, 340, 2001. 105. Bless, C., Palmeter, H., and M.M., Identification of acer rubrum using amplified fragment length polymorphism, J Forensic Sci, 51, 31, 2006. 106. Budowle, B., Genetics and attribution issues that confront the microbial forensics field, Forensic Sci Int, 146, 185, 2004. 107. Horswell, J., Cordiner, S., Maas, E., et al., Forensic comparison of soils by bacterial community DNA profiling, J Forensic Sci, 47, 350, 2002. 108. Moreno, L., Mills, D., Entry, J., et al., Microbial metagenome profiling using amplicon length heterogeneity-polymerase chain reaction proves more effective than elemental analysis in discriminating soil specimens, J Forensic Sci, 51, 2006. 109. Marsh, T., Terminal restriction fragment length polymorphism (T-RFLP): An emerging method for characterizing diversity among homologous populations of amplification products, Curr Opin Microbiol, 2, 323–327, 1999. 110. Clement, B., Kehl, L., DeBord, K., et al., Terminal restriction fragment patterns (TRFPs), a rapid, PCR-based method for the comparison of complex bacterial communities, J Microbiol Meth, 31, 135, 1998. 111. Chan, O. C., Yang, X., Fu, Y., et al., 16S rRNA gene analyses of bacterial community structures in the soils of evergreen broad-leaved forests in south-west China, FEMS Microbiol Ecol, 58, 247, 2006. 112. Kvist, T., Ahring, B., Lasken, R., et al., Specific single-cell isolation and genomic amplification of uncultured microorganisms, Appl Microbiol Biotechnol, 74, 926–935, 2006.

784

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

113. Mills, D., Fitzgerald, K., Litchfield, C., et al., A comparison of DNA profiling techniques for monitoring nutrient impact on microbial community composition during bioremediation of petroleum-contaminated soils, J Microbiol Meth, 54, 57, 2003. 114. Dunbar, J., Ticknor, L., and Kuske, C., Assessment of microbial diversity on four southwestern United States soils by 16S rRNA gene terminal restriction fragment analysis, Appl Environ Microbiol, 66, 2943, 2000.

26 Clinical Application of CE Zak K. Shihabi CONTENTS 26.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.1.1 Special Aspects of CE in Clinical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.1.2 Advantages of CE in the Clinical Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.1.3 Limitations of CE in the Clinical Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2 Practical Aspects of CE in Clinical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2.2 Sample Matrix Effects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2.3 High Salt Content in the Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2.4 High Protein Content in the Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2.5 Wide Variability in Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2.6 Precision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2.7 Coated versus Noncoated Capillary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2.8 Stacking of Compounds of Clinical Interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2.9 Compounds Difficult to Analyze by CE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2.10 Compounds Suited for Analysis by CE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.3 Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.3.1 Serum Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.3.2 Immunofixation (Immunosubtraction). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.3.3 Cryoglobulins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.3.4 Urinary Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.3.5 Cerebrospinal Fluid Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.4 Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.5 Hemoglobin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.5.1 Hemoglobin Variant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.5.2 Hemoglobin A1C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.5.3 Globin Chains of Hemoglobin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.6 Peptides and Polypeptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.7 CE and Proteomics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.7.1 Nucleic Acids (DNA). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.8 Small Molecule Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.8.1 Drug Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.8.2 Endogenous Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.8.3 Amino Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.8.4 Ion Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.9 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

786 786 787 787 787 787 787 788 788 788 789 789 789 790 790 791 791 791 792 793 793 793 797 797 797 799 799 799 801 801 801 801 801 802 805 805

785

786

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

26.1 INTRODUCTION The movement of soil colloidal particles was the first description of electrophoresis as early as 1809. However, Arne Tiselius (∼1937) was the first to construct a successful instrument useful for the separation of serum protein by electrophoresis using the boundary separation principle. Because of the clinical significance of this type of separation, many improvements and refinements followed, such as utilizing paper, cellulose acetate, gel, and more recently capillaries in order to speed up and better separate (into distinct zones) the different proteins. The electric current can be utilized in the clinical applications to accomplish not just separation but other tasks: 1. It can be utilized to move fluids (through the electroosmotic flow, EOF). This feature is very useful for analysis in the microchip and in capillary electrochromatography (CEC) where the additions of pumps to move and mix fluids are not feasible. 2. It can be used to concentrate dilute compounds directly during the electrophoretic step (through stacking) as it will be discussed later. This feature is very important for analysis of clinical compounds present in biological fluids at very low concentration. 3. It can be used to separate, quantify, and identify different compounds in clinical samples using different electrophoretic principles. Capillary electrophoresis (CE) is a general analytical technique for separation and quantification of a wide variety of molecules including those of clinical interest, utilizing narrow bore capillaries under high voltage. It separates various compounds not only based on charge but also based on size, hydrophobicity, and stereospecificity as discussed in Chapters 1, 2, 19. The flexibility of this technique stems from its ability to incorporate easily in the separation buffer different additives, that can interact with some of the analytes relative to others to alter their velocity so as to achieve the desired separation. These additives give the CE great ability to separate numerous clinical compounds in the same instrument using nonexpensive capillaries. Thus, most of the clinical tests can be adapted to CE; however in practice, some are better suited than others for analysis by this method. The clinical/biological field and pharmaceutical industry are the main areas that are benefiting most from the application of the CE. Most of the clinical tests can be analyzed by CE as well by other techniques, such as high-performance liquid chromatography (HPLC) or slab gel electrophoresis (SGE). However, the CE offers certain advantages for clinical analyses notably a high plate number, a characteristic that is useful when dealing with complex samples containing numerous compounds such as serum or urine. The CE also offers rapid analysis time and a low cost per test with full automation. These characteristics were the driving forces for adopting the CE for completing the human genome sequence project. However, the CE in clinical analysis requires more thoughtful planning to achieve a good separation compared with other methods. For example, many steroids cannot be separated in free zone electrophoresis. Many of these are neutral or weakly charged compounds and migrate with EOF. They separate much better by the addition of a micellar compound, such as sodium dodecyl sulfate (SDS) to achieve separation by micellar electrokinetic capillary chromatography (MEKC). The CE has some challenges such as poor sensitivity of detection, problems with the sample matrix, and adsorption to the capillary walls as it will be discussed later here and Chapter 13. However, with careful planning these obstacles can be resolved. Many applications can fall under the umbrella of clinical applications; however, tests that have routine applications in the clinical field will be discussed more in detail in this chapter.

26.1.1 SPECIAL ASPECTS OF CE IN CLINICAL ANALYSIS The clinical field is a very vast one encompassing molecules of different physical and chemical characteristics. The sizes vary from small ions such as Na and K to very large ones such as protein and DNA. It encompasses compounds that are very abundant in concentration such as albumin

Clinical Application of CE

787

(∼30,000 mg/L) to compounds very low in concentration such as prostate-specific antigen (∼4 µg/L). It encompasses compounds with different solubility and hydrophobicity. Thus, the analysis of these different compounds requires different strategies and represents different degrees of difficulties. Some analyses work well and easily with capillary zone electrophoresis (CZE) (e.g., serum proteins), size separation (e.g., DNA) while others are difficult to adapt to any form of CE. The lower the concentration is the more difficult the analysis becomes.

26.1.2 ADVANTAGES OF CE IN THE CLINICAL FIELD The main advantage of the CE in clinical analysis is the flexibility of the separation so that the same instrument can be utilized to separate numerous biological compounds. It is easy to add different additives to the separation buffer to induce a change in the velocity of some of the compounds leading to a better separation. The principle of the separation can easily be changed from free solutions to hydrophobicity separation (MEKC), size separation, chiral, isofocusing to suit a particular group of compounds as described under Chapters 1, 2. The second important advantage is the high plate number in CE generated in the capillary due to the flat profile, compared with the laminar flow in HPLC, without dealing with the high pressure or the high cost of the HPLC packed column. A third feature is the ability to perform separations without the need to use large volumes of organic solvents. Organic solvents, which are used often in HPLC, are becoming more expensive to purchase and more difficult, under many state laws, to store and dispose of. This eventually may compel shifting of many separations from HPLC to the CE.

26.1.3 LIMITATIONS OF CE IN THE CLINICAL FIELD On the other hand, CE suffers from a few problems. Unlike HPLC, CE is greatly affected by sample matrix (i.e., salts and proteins) [1–4], which are very high in biological samples. Another major problem in clinical analysis is the suboptimal detection sensitivity. The third problem is sample interaction with the capillary walls. To utilize the CE successfully for practical separation of clinical samples it is important to understand how these factors affect the CE and the different maneuvers needed to overcome them.

26.2 PRACTICAL ASPECTS OF CE IN CLINICAL ANALYSIS 26.2.1 BACKGROUND Many applications fall under the umbrella of clinical applications. For the sake of simplicity, these can be divided into two general areas. One is the research, which is performed occasionally for gaining basic scientific information. The other area is the routine analysis for patient care and diagnosis such as detection of monoclonal gammopathies to detect specific tumors. This test is performed often. In research, the need is for the flexibility and versatility with good separation with emphasis on detection of compounds at low concentration. In routine work, the emphasis is more on good reproducibility (precision) with minimum amount of analytical steps. The different forms of CE can fulfill both goals. For example, the ability to add different additives to the buffer allows for great selectivity and versatility, for example, from separation based on charge to separation based on hydrophobicity, or to size. On the other hand, the MEKC form of the CE offers ability to simply inject the sample directly on the capillary without any treatment.

26.2.2 SAMPLE MATRIX EFFECTS Many compounds can be analyzed easily from pure standard solutions, tablets, or from samples with clean matrix. However, the analysis of the same compounds from biological sources poses many

788

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

more difficulties. This is because of the effects of both the high salt and high protein content of the biological samples on the separation in CE. Although the sample, in most instances, constitutes a very small portion of the overall volume in the capillary once injected (99%)

SPME Silica impregnated with polydimethylsiloxane

LLE cyclohexane (48–90%) Air sample collected on polyurethane foams

LIF, bubble cell, 325 nm, 2.5 mW HeCd laser 0.9–21.7 µg L−1

UV, 280, 333 nm 10−10 to 10−9 mol L−1

LIF Nd-YAG, 266 nm

25 mmol L−1 SB-β-CD, 20 mmol L−1 M-β-CD, 50 mmol L−1 borate

0.2 mol L−1 PFOS− , 50% DMSO, 0.1 mol L−1 H3 PO4

30 mmol L−1 borate, 60 mmol L−1 SDS, 12.5 mmol L−1 γ-CD (pH 9.0)

125

52

43

38

36

UV, 254 nm

UV, 214 nm

27

15

References

UV, 214 nm 3–25 ng mL−1

UV, 280 nm 10 mg L−1

Detection (LOD)

20% THF, 0.00625 mol L−1 OSUA

50% ACN (pH 9.9) 0.1 mol L−1 phosphate, 0.1 mol L−1 borate, 50 mmol L−1 STDC, 30% acetone 50 mmol L−1 NH4Ac, 100 mmol L−1 THA+ in 100% MeOH

8.5 mmol L−1 borate, 85 mmol L−1 SDS,

Optimal Electrolyte

ASE, accelerated solvent extraction; HPGPC, high-performance gel permeation chromatography; OSUA, oligomers of sodium undecylenic acid; SFE, supercritical fluid extraction; SDβCD, sulfobutyl ether β-cyclodextrin; MβCD, methyl β-cyclodextrin; PFOS− , perfluorooctanic sulfate; DMSO, dimethyl sulfoxide; THA+ , tetrahexylammonium; SPME, solid-phase microextraction; STDC, sodium taurodeoxycholate; LIF, laser-induced fluorescence; LLE, liquid–liquid extraction.

EKC

CE Mode

7 PAH homologs in 10 min

Compounds

Preconcentration Procedure (Recovery)

TABLE 31.2 Selected Applications of Electroseparation Methods for PAH in Environmental Samples

Separation Strategies for Environmental Analysis 929

930

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

3 6 1 4

2

11

7 5

9

8

6

8

10

12

14

16

18

20

22

24

Retention time (min)

FIGURE 31.2 EKC separation of PAHs. Electrolyte: 10 mmol L−1 H3 PO4 and 70 mmol L−1 sodium n-tetradecylsulfate in 75:25 methanol:H2 O. Other conditions: −20 kV, direct UV-detection at 254 nm. Peak labels: benzo[a]perylene (1), perylene (2), benzo[a]anthracene (3), pyrene (4), 9-methylanthracene (5), anthracene (6), fluorene (7), naphthalene (8), and benzophenone (9). (Modified from J. Li and J.S. Fritz, Electrophoresis, 20, 84–91, 1999. With permission.)

simultaneously (chlorophenol congeners, for instance), EKC methodologies should be selected as a means to increasing selectivity. Nevertheless, CZE has, by far, received more attention for real sample applications. Literature proposed CZE methods for phenols and derivatives using test mixtures based on aqueous buffered systems (phosphate–borate153 and borate154 ), volatile electrolytes (ammonium hydrogencarbonate,155 diethylmalonic acid/dimethylamine in isopropanol156 and l-cysteic acid, 3-amino-1-propanesulfonic acid, aminomethanesulfonic acid, and diethylmalonic acid157 ), and nonaqueous media (ammonium acetate in ACN/acetic acid in MeOH;158 acetate, bromide, chloride, and malonate in ACN; and diprotic acids/tetrabutylammonium hydroxide159 and maleate in MeOH160 ). Environmental assessment using CZE methodologies includes pressurized hot water extraction of phenols from sea sand and soil,161 chlorophenols162 and priority phenols11 in wastewater, nitrophenols163 and iodophenol164 in river waters, and methoxyphenols in biomass burning.86 Details of these methodologies are compiled in Table 31.3. The impact of different surfactants (SDS,22 DOSS,165 CTAB166 and hexadimethrine bromide, HDB,167 bile salts29,30 ), nonionic168 and mixed micelles,22,169 and additives (neutral170 and anionic CDs,44 tetraalkylammonium salts,171 organic solvents29,165 ) in EKC separations has been demonstrated with phenol test mixtures. In addition, phenols have been chosen to introduce the applicability of more exotic EKC secondary phases such as SDS modified by bovine serum albumin,172 watersoluble calixarene,173 starburst dendrimers,174 cationic replaceable polymeric phases,175 ionenes,176 amphiphilic block copolymers,177 polyelectrolye complexes,178 and liposome-coated capillaries.179 The separation of phenols of environmental interest180 as well as the sources and transformations of chlorophenols in the natural environment181 have been revised. Examples of the investigation of phenols by EKC methodologies in aquatic systems,39,41,81 soil,44 and gas phase182 are compiled in Table 31.3. Figure 31.3 illustrates the electromigration separation of phenols by both CZE and EKC modes.

31.4.4 AMINES Amines are organic bases that are usually present in biological materials (biogenic amines), processed foods and beverages (of concern are the nitrosamines in fried bacon, smoked/cured meat and fish,

CZE

EKC

EKC

EKC

EKC

CZE

11 Priority phenols in 12 min

10 Nitrophenols in 20 min

25 Phenols in 20 min

19 Chlorophenols in ca. 35 min

12 Substituted methoxy phenols and aromatic acids in 9.5 min

CE Mode

26 Priority phenols in 40 min

Compounds

AEROSOL Biomass burning Aerosol

WATER Rain water Tap water Process water SOIL Certified soil Reference standard WATER River water

WATER River water Sea water

WATER Waste water

Matrix

Berner type impactor filter (140–420 nm, 50% cut off)

SPE PSDVB 81–116%

—

ITP spiked sample

SPE PSDVB spiked sample 63.3–113.4% SPE spiked sample 74.2–106%

Preconcentration Procedure (Recovery)

41

UV, 254 nm 19–80 µg L−1

ELECTROCHEMICAL graphite-epoxy electrode versus Ag/AgCl 0.07–0.2 µg L−1 ESI-IT-MS (1) 0.1–1.0 µmol L−1 (2) 0.3–0.7 µmol L−1

50 mmol L−1 ACES, 22 mmol L−1 SDS (pH 6.1)

(1) 20 mmol L−1 NH4Ac, 10% MeOH (pH 9.1) (2) 1 mol L−1 NH4 OH (pH 11)

UV, 214 nm 0.05–0.33 mg L−1

50 mmol L−1 phosphate, 1 mmol L−1 SB-β-CD (pH 7.5)

Continued

86

81

44

39

UV, 214 nm 28.1–215 nmol L−1

50 mmol L−1 phosphate, 25 mmol L−1 tetraborate-phosphate, 7.5 mmol L−1 DBTHS (pH 7) 50 mmol L−1 glycine, 0.2% m-HEC, 2.5% PVP (pH 9.1)

N-methylformamide-ACN mixture (non-aqueous)

11

References

UV, 280 nm 28–629 µg L−1

Detection (LOD)

150 mmol L−1 NH4Ac in 75:25

Optimal Electrolyte

TABLE 31.3 Selected Applications of Electroseparation Methods for Phenols in Environmental Samples

Separation Strategies for Environmental Analysis 931

CZE

o-Nitrophenol, m-nitrophenol, p-nitrophenol in 11 min 2-Iodophenol, 4-iodophenol, 2,4,6-triiodophenol in 6.6 min

EKC

AIR

WATER Waste water WATER River water WATER River water

SOIL Soil and sea sand mixture Sea sand

Matrix

Loop-supported liquid film

Spiked sample 95.6–100.9% SPME CAR-PDMS ca. 100%

PHWE 77.7–105.2% (soil sample) 77.7–98.4% (sea sand sample) —

Preconcentration Procedure (Recovery)

UV, 214 nm — UV, 191 nm 10.2–40.6 µmol L−1 ICP-MS 2.4–3.9 µg L−1 (CE-ICP-MS) 0.03–0.04 µg L−1 (SPME-CE-ICP-MS) UV, 205 nm 3.5–17 µg L−1

10 mmol L−1 phosphate buffer, 40% acetone (pH 8.23) 20 mmol L−1 borate-carbonate, 10% MeOH (pH 9.4) 20 mmol L−1 CAPS (pH 11.0)

25 mmol L−1 , borate 10 mmol L−1 phosphate, 10 mmol L−1 SDS (8.86)

UV, 220 nm 270–410 µg L−1

Detection (LOD)

30 mmol L−1 CHES (pH 9.7)

Optimal Electrolyte

182

164

163

162

161

References

DBTHS, disodium 5,13-bis(dodecyloxymethyl-4,7,11,14-tetraoxa-1,17-heptadecanedisulfonate; ITP, isotachophoresis; m-HEC, methyl-hydroxyethylcellulose; PVP, polyvinylpirrolidone; PSDVB, polystyrene-divinylbenzene copolymer; SB-β-CD, sulfobutylether-β-cyclodextrin; CAR-PDMS, carboxen-poly(dimethylsiloxane); ESI-IT-MS, electrospray ionization ion trap-mass spectrometry; CHES, 2-(N-cyclohexylamino)-ethanesulphonic acid; CAPS, 3-(cyclohexylamino)-1-propanesulfonic acid; PHWE, pressurized hot water extraction; ICP-MS, inductively coupled plasma-mass spectrometry; SPME, solid-phase microextraction; SPE, solid-phase extraction.

12 Chloro and nitrophenols in 10 min

CZE

17 Chlorophenols in 15 min

CZE

CZE

CE Mode

Phenol, 3-methylphenol, 4-chloro-3-methylphenol, 3,4-dichlorophenol in 6 min

Compounds

TABLE 31.3 (Continued)

932 Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Separation Strategies for Environmental Analysis

933

(a) 5 3

5

4 7

mAU

9

1

3 2

4 2

EOF

6

8

10

11 1 0 2

3

4

5

Migration time (min) (b) 1

3 5 2 4 6 9 8 7

10

11 5

6.0

7.0

8.0

9.0

10.0

11.0

12.0

Time (min)

FIGURE 31.3 (a) CZE and (b) EKC separation of EPA phenols. Electrolytes: (a) 20 mmol L−1 NaH2 BO3 at pH 10.00; (b) 20 mmol L−1 phosphate, 8% 2-butanol, 0.001% HDB, pH 11.95. Other conditions: (a) +20 kV; direct UV-detection at 254 nm; (b) +25 kV; direct UV-detection at 210 nm. Peak labels: (a) 2,4-dinitrophenol (1), 2-methyl-4,6-dinitrophenol (2), pentachlorophenol (3), 2,4,6-trichlorophenol (4), 4-nitrophenol (5), 2-nitrophenol (6), 2,4-dichlorophenol (7), 2-chlorophenol (8), 4-chloro-3-methylphenol (9), phenol (10), and 2,4-dimethylphenol (11). (b) 2-nitrophenol (1), 2-chlorophenol (2), 2,4,6-trichlorophenol (3), phenol (4), 4-nitrophenol (5), 2,4-dinitrophenol (6), 4-chloro-3-methylphenol (7), 2,4-dichlorophenol (8), 2-methyl4,6-dinitrophenol (9), 2,4-dimethylphenol (10), and pentachlorophenol (11). (Modified from (a) I. Canals et al., Anal. Chim. Acta, 458, 355–366, 2002 and (b) P. Kubáˇn et al., J. Chromatogr. A, 912, 163–170, 2001. With permission.)

canned sausage, etc.), as well as environmental samples. Several amines are strongly toxic and suspected carcinogens. Major sources of amines in the environment include effluents or byproducts of several chemical industry sectors such as oil refining, synthetic polymers, adhesives, rubber tyre manufacturing, leather tanning, pharmaceuticals, pesticide production, and explosives.

31.4.4.1 Aliphatic Amines The analysis of small aliphatic amines by electromigration methods faces a few challenges. Although readily protonated at low pH electrolytes, the lack of chromophore precludes direct UV-detection in both CZE and EKC mode unless derivatizing schemes are devised or alternative detectors are selected. A further complication of the EKC mode derives from the hydrophilicity and polar characteristic of aliphatic amines, which usually impair strong interaction with commonly used micellized surfactants, demanding alternative secondary phases or more elaborate electrolytes.

934

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Indirect UV-detection is a feasible option and particularly interesting for the CZE determination of primary to quarternary amines without derivatization. Imidazole, N-ethylbenzylamine, and benzyltriethylammonium chloride have been proposed as electrolyte chromophoric constituents.183 The indirect UV-detection of aliphatic amines in ambient air has been performed by CZE in imidazolebased buffer modified by ethanol and EDTA.184 By replacing imidazole with ammonium, the electrolyte became applicable to MS detection (these details and other CZE methodologies are reported in Table 31.4). EKC separations of aliphatic amines include electrolyte systems composed of several surfactants (SDS,185 cholate,186 Brij 3572 ) modified by certain additives (urea,185 neutral CDs,185 organic solvents75,186 ), mixed CDs,187 and more unusual secondary phases [resorcarene-octacarboxylic acid,188 calixarene,47 poly(sodium 4-styrenesulfonate), PSSS189 ]. Derivatization is performed when UV (o-phthaldialdehyde, OPA,185,187 as derivatizing agents) or LIF detection [3-(2-furoyl)quinoline2-carboxaldehyde,186 5-(4,6-dichloro-s-triazin-2-ylamino)fluorescein (DTAF)72 as labeling agents] is designed. Examples of the EKC determination of aliphatic amines in water compartments75 and atmospheric aerosol samples71 are compiled in Table 31.4. 31.4.4.2 Aromatic Amines Direct UV spectrophotometric detection of aromatic amines from azo dye reduction in textile industry wastewaters190 and electrophoretic methods for biogenic and aromatic amines191 have been the subject of recent reviews. In addition, environmental applications of mutagenic heterocyclic amines have recently been revised.192 CZE analysis of aromatic amines include the use of simple buffers at varying pH (acetate,80 phosphate,10,193 and borate70 ) with direct UV,10 electrochemical,80 or fluorescence70,80,193 detection. EKC methodologies for aromatic amines are based on electrolytes containing SDS,64,91,194,195 mixtures of SDS and nonionic surfactants,196 and bile salts29 as well as modifiers (tetraalkylammonium salts,195 urea,64 organic solvents,29,64,91,194,196 and CDs197 ). Combinations of anionic soluble polymers198 or crown ether199 and CDs have also been reported. Despite being chromophoric, alternative detectors for aromatic amines have been proposed (CE-ESI-MS,91 electrochemical,80 and fluorescence80 detection). Examples of EKC methodologies for aromatic amines in water21,35,37,80 are compiled in Table 31.4. Figure 31.4 illustrates the electromigration separation of aliphatic and aromatic amines.

31.4.5 CARBONYLS Low-molecular mass carbonyls are among the most abundant and ubiquitous volatile organic compounds in the atmosphere. They are produced from industrial activity and incomplete combustion of fossil fuels and biomass. Many aldehydes are also emitted indoors (plastic, foam insulation, lacquers, etc.). As a source of free radicals, aldehydes play an important role in the ozone formation, in urban smog events, as well as in the photochemistry of the unpolluted troposphere. Aldehydes are recognized irritants of the eye and respiratory tract, and often, carcinogenic and mutagenic characteristics are also attributed to them. To implement CZE and EKC methodologies for the determination of carbonyl compounds, structure-related issues must be addressed. Aldehydes and ketones of environmental importance are essentially neutral molecules, interact poorly with micellar phases, and present no chromophoric moieties. Two methodological approaches have been presented: either generation of charged adducts or generation of large UV-absorbing or fluorescence derivatives. In the former, the charged adducts can be detected directly (if aromatic) or indirectly (if aliphatic). The separation of anionic aldehydebisulfite adducts is a fine example of the first approach.25 The second approach comprises the derivatization of the molecules to generate either neutral or charged UV-absorbing or fluorescent

2-Toluidine, 4-toluidine, 1-naphthylamine, 2-naphthylamine, 2-aminobiphenyl, 4-aminobiphenyl, 2-methoxy-5-methylaniline, 4-methoxy-2-methylaniline, 2-chloroaniline, 4-chloroaniline, 2,4,5-trimethylaniline, 2,4,6-trimethylaniline, 2-anisidine, 4-anisidine, 2,4-xylidine, 2,6-xylidine in 50 min

Aromatic amines 21 Aromatic amines in 25 min

Methylamine, ethylamine, propylamine, butylamine, dimethylamine, diethylamine, cadaverine, putrescin, spermidine in 9.5 min

Aliphatic amines Methylamine, dimethylamine, diethylamine, dipropylamine, piperidine, pyrrolidine, morpholine in 10 min Diaminopropane, putrescine, cadaverine, diaminohexane in 7.5 min

Compounds

WATER Waste water

EKC

AIR Metal working Fluid aerosols and ambient air

CZE

WATER SOIL

WATER Lake water

EKC

CZE

AIR Particulate aerosol

Matrix

EKC

CE Mode

LLE/SPE methylene chloride/ 20:80 (w/w) keto derivatized– underivatized PSDVB 47–97% SPE/LLE (cation exchange resin; tertiary butyl methyl ether) 78.4–94.4%

75

Chemiluminescence, postcolumn reagents: 3 mmol L−1 K3 Fe(CN)6 , 0.8 mol L−1 NaOH. (3.5–12) × 10−8 mol L−1 UV, indirect detection, 214 nm 2 µg mL−1

21 UV, 214 nm 0.263–9.525 µg mL−1 50 mmol L−1 phosphate—adjusted to pH 8.5 with 100 mmol L−1 borate, 300 mmol L−1 SDS, 10 mmol L−1 cholic acid, 10 mmol L−1 Tween 80

Continued

10 UV, 280 nm 0.06–1.8 mg L−1

184

71

References

LIF, 488 nm Ar laser, 520 nm emission 10−9 mol L−1

Detection (LOD)

50 mmol L−1 NaH2 PO4 , 7 mmol L−1 1,3-diaminopropane, (pH 2.35 with H3 PO4 )

20 mmol L−1 imidazole, 4 mmol L−1 EDTA, 25% ethanol (pH 2.5 with acetic acid)

Quartz filter — SPE poly(acrylatemethacrylate) copolymer 95–99%

ABEI-DSC derivatization 92.8–106.8%

20 mmol L−1 borate, 20% acetone, 5 mmol L−1 DM-β-CD 10 mmol L−1 borate, 100 mmol L−1 H2 O2 , 80 mmol L−1 SDS (pH 9.3)

Optimal Electrolyte

FITC derivatization

Preconcentration Procedure (Recovery)

TABLE 31.4 Selected Applications of Electroseparation Methods for Amines in Environmental Samples

Separation Strategies for Environmental Analysis 935

WATER lake water

WATER Tap water Lake water

WATER Lake water

WATER Surface water near textile and leather industries

EKC

CZE

EKC

Matrix

EKC

CE Mode

UV, 214 nm LIF, 266 nm solid-state UV laser, 310 nm emission 5.7 × 10−8 to 4.9 × 10−7 mol L−1

LIF, 488 nm Ar laser, 520 nm emission 10−10 mol L−1

Multiphoton-excited fluorescence detection, diode laser, 808 nm 1.25–2.60 µmol L−1 Fluorescence, mercury–xenon lamp, 495 nm emission 1 µg L−1

400 mmol L−1 borate, 40 mmol L−1 OG (pH 9.0)

15 mmol L−1 borate (pH 9.5)

5 mmol L−1 tetraborate, 4.5 mmol L−1 boric acid, 20 mmol L−1 SDS (pH 9)

FITC derivatization

FITC derivatization

Fluorescamine derivatization

Detection (LOD)

35 DOSS, 8 mmol L−1 borate, 40% ACN (pH 8.5)

mmol L−1

Optimal Electrolyte

FMOC derivatization (anilines)

Preconcentration Procedure (Recovery)

80

70

37

35

References

ABEI, N-(4-aminobutyl)-N-ethylisoluminol); DSC, N,N-disuccinimidyl carbonate; FITC, fluorescein isothiocyanate; OG, n-octylglucopyranoside; FMOC, 9-fluoroenylmethyl chloroformate; DOSS, dioctyl sulfosuccinate; EDTA, ethylenediaminetetraacetic acid; PSDVB, polystyrene-divinylbenzene copolymer; DMβCD, di-methyl β-cyclodextrin; SPE, solid-phase extraction; LLE, liquid–liquid extraction; LIF, laser-induced fluorescence.

1,3-Phenylenediamine, 2-methoxy aniline, 4-ethoxyaniline, 4,4′ -diaminobiphenyl, 2-methylaniline, 2,4- dimethylaniline, 2-ethylaniline, 2,6-dimethylaniline in 4 min

4-Chloroaniline, 4-bromoaniline, 3,4-dichloroaniline, 3-chloroaniline, 3-chloro-4-methylaniline, 4-isopropylaniline, aniline, 3-methylaniline, acetophenone, propiophenone, butyrophenone, valerophenone, hexanophenone, heptanophenone in 26 min 4-Chloroaniline, 4-bromoaniline, 3,4-dichloroaniline, 3-chloroaniline, 3-chloro-4-methylaniline, 4-isopropylaniline, 3-methylaniline, aniline in 15 min p-Toluidine, aniline, p-ethoxyaniline, p-phenylenediamine in 4 min

Compounds

TABLE 31.4 (Continued)

936 Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Separation Strategies for Environmental Analysis

(a)

4 5

937

6 78

1

2

12 9 10 11

3

14 1617 13

4

5

6

18

19

15

7

8

9

10

Migration time (min)

(b) 12

78

3 5 6

10 9

4

1112 13

25

18 14 20

16 mAU

15 17

21

19

–2.5 0

10

20

30

Time (min)

FIGURE 31.4 Separation of (a) aliphatic and (b) aromatic amines. Electrolytes: (a) 20 mmol L−1 imidazole-acetate at pH 3.5, 4 mmol L−1 EDTA and 25% ethanol; (b) 50 mmol L−1 phosphate buffer at pH 2.35, 7 mmol L−1 1,3-diaminopropane. Other conditions: (a) +20 kV; indirect UV-detection at 214 nm; (b) +30 kV applied voltage; direct UV-detection at 280 nm. Peak labels: (a): methylamine (1), dimethylamine (2), Na+ (3), ethylamine (4), cadaverine (5), putrescine (6), propylamine (7), methanolamine (8), morpholine (9), diethylamine (10), butylamine (11), isopropanolamine (12), piperazine (13), 2-(2-aminoethoxy)-ethanol (14), 2-amino-1-butanol (15), diethanolamine + methyldiethanolamine (16), amino-2-methyl-1-propanol (17), amino-2-ethyl-1,3-propandiol (18), and spermidine (19); (b): pyridine (1), p-phenylenediamine (2), benzidine (3), o-toluidine (4), aniline (5), N, N-dimethylaniline (6), p-anisidine (7), p-chloroaniline (8), m-chloroaniline (9), ethylaniline (10), α-naphthylamine (11), diethylaniline (12), N-(1-naphthyl)ethylenediamine (13), 4-aminophenazone (14), o-chloroaniline (15), 3,4dichloroaniline (16), 3,3′ -dichlorobenzidine (17), 2-methyl-3-nitroaniline (18), 2,4-dichloroaniline (19), 2,3-dichloroaniline (20), and 2,5-dichloroaniline (21). (Modified from (a) A. Fekete et al., Electrophoresis, 27, 1237–1247, 2006 and (b) A. Cavallaro et al., J. Chromatogr. A, 709, 361–366, 1995. With permission.)

derivatives, usually assessed by EKC methods. Several derivatizing reagents have been proposed in environmental applications: DNPH,25,68,200,201 3-methyl-2-benzothiazoline hydrazinone (MBTH),202 5-(dimethylamino)-naphthalene-1-sulfohydrazide (dansylhydrazine, DNSH),68,203 and 4-hydrazinobenzoic acid (HBA).204 Table 31.5 compiles the details of the electromigration methodologies applied to carbonyls in the environment. Figure 31.5 illustrates the electromigration separation of carbonyl compounds as bisulfite adducts and DNSH derivatives.

31.4.6 SMALL IONS AND ORGANOMETALLIC COMPOUNDS Aspects of the determination of small ions and organometallic compounds by CE methodologies as well as elemental speciation have been extensively covered in the literature by excellent reviews

AIR Vehicular emission

AIR Vehicular emission

EKC

EKC

EKC

EKC

EKC

Formaldehyde, acetaldehyde, propionaldehyde, acrolein, benzaldehyde (bisulfite derivatives) in ca. 7.5 min 5 DNPH derivatives in 14 min

3 DNPH derivatives in 8 min

4 DNPH derivatives in 20 min

5 MBTH derivatives in 10 min

25 UV, 360 nm 0.2–2.0 mg L−1

UV, 214 nm 2.0 mg L−1

UV, 360 nm 0.05 mg L−1 (formaldehyde) 0.08 mg L−1 (acetaldehyde) UV, 216 nm 0.54–4.0 µg L−1

20 mmol L−1 borate, 50 mmol L−1 SDS, 15 mmol L−1 β-CD 0.02 mol L−1 borate, 0.05 mol L−1 SDS (pH 9)

0.02 mol L−1 borate-phosphate, 0.05 mol L−1 SDS (pH 9)

IMPINGER 0.05% MBTH stacking with salt

IMPINGER 0.05 g DNPH in 50 mL 2 mol L−1 HCl LLE in chloroform IMPINGER 1.56 g DNPH in 500 mL 2 mol L−1 HCl LLE in CS2 Spiked sample 97–102%

25 UV, indirect detection, 254 nm 10–40 µg L−1

10 mmol L−1 3,5 dinitrobenzoic acid, 0.2 mmol L−1 CTAB (pH 4.5)

20 mmol L−1 tetraborate, 50 mmol L−1 SDS (pH 9.3)

0.040 mol L−1 tetraborate (pH 9.3)

202

201

200

204

203

UV, 214 nm 1.1–9.5 µg L−1 LIF, He/Cd 325 nm excitation, 520 emission 0.29–5.3 µg L−1 UV, 290 nm 2.7–8.8 ng L−1

20 mmol L−1 phosphate buffer (pH 7.02)

C18 catridges impregnated with dansylhydrazine/ trichloroacetic acid/MeOH C18 catridges impregnated with HBA IMPINGER bisulfite derivatization

68

UV, 218 nm Z-shaped flow cell, 170–300 nmol L−1

References

5 phosphate, 10 mmol L−1 tetraborate, 20% ACN (pH 8.0)

Detection (LOD)

Dansylhydrazine derivatization

mmol L−1

Optimal Electrolyte

Obs LODs refer to each single aldehyde, not the derivative; DNPH, 2,4-dinitrophenylhydrazine; DNSH, 5-dimethylaminonaphthalene-1-sulfohydrazide; HBA, 4-hydrazinobenzoic acid; MBTH, 3-methyl-2-benzothiazoline hydrazinone; LLE, liquid–liquid extraction; CTAB, cetyltrimethylammonium bromide.

AIR Indoors

WATER River water

GAS Stack gas from an organic plant

AIR Indoors

AIR Indoors Outdoors

CZE

CZE

WATER Rain water

Matrix

CZE

CE Mode

4 HBA derivatives in 6 min

Acetone, acetaldehyde, propionaldehyde, benzaldehyde, formaldehyde, methylglyoxal in 10.5 min 9 DNSH derivatives including isomers and impurities in 9 min

Compounds

Preconcentration Procedure (Recovery)

TABLE 31.5 Selected Applications of Electroseparation Methods for Carbonyl Compounds in Environmental Samples

938 Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Separation Strategies for Environmental Analysis

(a)

939

(b)

R

0.030 S

#

1

R

1 2 34

Standards A.U

Standards

0.015 53 4 2

4a

5 **

0.000

6 7 89

36

9

0.030 S

2

R

1

2

0. 4.5

6.0

1

Indoor air sample A.U

Vehicle emission sample

0.005 1

0.015

*

7.5

2 0.000

0

2

4

6 Time (min)

8

10

3

6

9

Time (min)

FIGURE 31.5 Separation of carbonyl compounds as (a) bisulfite adducts and (b) DNSH derivatives. Electrolytes: (a) 10 mmol L−1 3,5-dinitrobenzoic acid at pH 4.5 containing 0.2 mmol L−1 CTAB; (b) 20 mmol L−1 phosphate buffer at pH 7.02. Other conditions: (a) −10 kV; indirect UV detection at 254 nm; (b) +20 kV applied voltage; direct UV detection at 214 nm. Peak labels: (a): formaldehyde (1), acetaldehyde (2), propionaldehyde (3), acrolein (4), and benzaldehyde (5), impurities (∗ ) system peak (S), excess reagent (R); (b): formaldehyde (1), acetaldehyde (2 and 7), propionaldehyde (3 and 6), acrolein (4, 8, and 9), and acetone (5), impurities (#), excess reagent (R). (Modified from (a) E.A. Pereira et al., J. AOAC Int., 82, 1562–1570, 1999 and (b) E.A. Pereira et al., J. Chromatogr. A, 979, 409–416, 2002. With permission.)

and key articles. A comprehensive review on applications of CE to the analysis of inorganic species including organometallic compounds in diverse environmental matrices (drinking, mineral, surface, and ground waters; rainwater; snow; seawater; brine and waste waters; aerosol; and others) has been compiled.205 The determination of inorganic ions in environmental aquatic samples of high salinity, particularly seawater, has been revised.206

31.4.6.1 Inorganic Cations The CZE determination of alkali, alkaline earth, and transition metal cations may be conducted under indirect UV-detection in electrolytes composed of cationic chromophores (protonated imidazole, aromatic amines, pyridine derivatives at moderately low pH electrolytes) and weak complexing agents (α-hydroxyisobutyric acid, HIBA, or other mono- and dicarboxylic acids), eventually modified by solvents. The simultaneous separation of 16 metal ions for inspection of river samples is an example of this approach.207 Direct UV-detection at low wavelengths (190 nm) using nonabsorbing electrolytes (pH 10 borate buffer) is a possibility for absorbing cations such as ammonium, as determined in river and sewage samples.208 Alternatively, UV-absorbing chelating (EDTA)62 or ion-pair forming agents (o-phenanthroline)209–211 can be selected for the determination of alkaline earth and transition metal ions. Figure 31.6 illustrates the electromigration separation of small cations.

2.00

6.00

10.00

2

3

5

4

6

11 10 8 7

9

16

12.00

2

14.00

5

6

7

16.00

18.00

Time (min)

13

Time (min)

20.00

22.00

24.00

16

8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00

River water contaminated by copper industry

1

4.00

1

12 15 13

14

−0.05

0.05

0.15

0.25

0.35

−0.01

0

0.01

0.02

0.03

0.04

0.05

(b)

7.0

7.0

7.5

Fe (II)

7.5

8.0

Cu (II)

Cu (II)

8.0

Zn (II)

Zn (II)

Fe (II)

Time (min)

8.5

Mn (II)

Time (min)

8.5

Mn (II)

9.0

9.5

9.5

F3 F2 F1

NIST SRM 1648 urban particulate matter at three stage sequential extraction

9.0

Cd (II)

Standards

10.0

10.0

FIGURE 31.6 Separation of metal ions. Electrolytes: (a) 10 mmol L−1 4-aminopyridine and 6.5 mmol L−1 HIBA at pH 4.5; (b) 200 mmol L−1 ammonium acetate at pH 5.5, 0.5 mmol L−1 1,10-phenanthroline, 10 mmol L−1 hydroxylamine, and 20% acetone. Other conditions: (a) +25 kV; indirect UV detection at 214 nm; (b) +14 kV applied voltage; direct UV-detection at 226 nm. Peak labels: (a): potassium (1), sodium (2), barium (3), strontium (4), calcium (5), magnesium (6), manganese (7), lithium (8), iron (9), cobalt (10), cadmium (11), niquel (12), zinc (13), lead (14), chromium (15), and copper (16). (Modified from (a) N. Shakulashvili et al., J. Chromatogr. A, 895, 205–212, 2000 and (b) E. Dabek-Zlotorzynska et al., Anal. Bioanal. Chem., 372, 467–472, 2002. With permission.)

0.00

Standards

(a)

Absorbance Absorbance

940 Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Separation Strategies for Environmental Analysis

941

31.4.6.2 Inorganic Anions Similar to the inorganic cations, the determination of inorganic anions and those derived from lowmolecular mass carboxylic acids are commonly approached by indirect UV-detection. However, here the electrolyte chromophore must be anionic (chromate,212,213 benzoate derivatives,214 naphthalene derivatives215,216 among others217,218 ) and an EOF reverser (long-chain tetraalkylammonium quaternary salts) is usually employed. It is worth mentioning that many methodologies for inorganic anions and even metal and nonmetal elements determined as oxoanions or anionic complexes that are reported in the literature as CZE separations are, in fact, EKC separations. This is due to the electrolyte ionic strength that forces micellization of the cationic surfactant used as flow reverser.219 Direct UV-detection at low wavelengths using nonabsorbing electrolytes (phosphate, borate, chloride, formate/chloride buffers modified by solvents, and cationic surfactants) is also a possibility for UV-absorbing anions such as iodide,56,220,221 iodate,222 bromide,223 bromate,224 nitrate,225 and nitrite226 as well as mixtures of anions9,66,227 and aromatic acids.228 Figure 31.7 illustrates the electromigration separation of small anions. 31.4.6.3 Simultaneous Detection of Cations and Anions An interesting strategy based on dual-opposite end sample injection that accomplishes the simultaneous separation of small cations and anions within a single run was demonstrated in the analysis (a)

(b) 11

Standards 18 20

Absorbance (µV)

15000 67

10000 3 1

5000

9

8 10

2

6

13 1517 12 16

12 8 21 23 24 25

14

1

26

2

Standards

8

3

7

5

9

10

4

4

27

0

5 28

4 19 22

0

29

Rain water

12 6

8

−5000

4

3

4

5

t (min)

6

7

3 2

8 Size-classified rain sample

14

0 6

Slag plate

12 8

20000 Absorbance (µV)

mAU

2

4

4

15000

12

0 16

10000

7

17 20

23

6 1

5000

4

12 8

?

6

Sludge extract

8

3

4

0 2

3

4

5

6

Time (min)

7

8

0 1

2

3

4

Time (min)

FIGURE 31.7 Separation of inorganic and organic anions. Electrolytes: (a) 7.5 mmol L−1 salicylic acid, 15 mmol L−1 Tris, 400 µmol L−1 DoTAH, 1050 µmol L−1 Ca2+ , and 600 µmol L−1 Ba2+ ; (b) 50 mmol L−1 tetraborate at pH 9.3, containing 5% methanol. Other conditions: (a) −28 kV; indirect UV-detection at 232 nm; (b) −23 kV applied voltage; direct UV-detection at 200 nm. Peak labels: (a): chloride (1), nitrate (2), sulfate (3), formate (4), fumarate (5), malonate (6), succinate (7), maleate (8), glutarate (9), methanesulfonate (10), carbonate/adipate (11), malate (12), pimelate (13), acetate (14), suberate (15), oxalate (16), azelate (17), sebacate (18), glyoxylate (19), propionate (20), methacrylate (21), lactate (22), butyrate (23), hydroxyiso-butyrate (24), valerate (25), capronate (26), enanthate (27), caprylate (28), and pelargate (29); (b): − bromide (1), iodide (2), nitrite (3), thiosulfate (4), chromate (5), nitrate (6), Fe(CN)4− 6 (7), SCN (8), 2− MoO2− 4 (9), and WO4 (10). (Modified from (a) A. Mainka et al., Chromatographia, 45, 158–162, 1997 and (b) M.I. Turnes-Carou et al., J. Chromatogr. A, 918, 411–421, 2001. With permission.)

942

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

of environmental water samples with indirect fluorescence73 and contactless conductometric78 detection. 31.4.6.4 Speciation and Organometallic Compounds Speciation of chromium (histidine/acetic acid;229 acetate buffer/EDTA230 ), mercury (borate/MeOH),8 tin (pyridine/CTAB),231 lead (SDS/β-CD),232 arsenic (borate,233 phosphate or phosphate/borate,54 phosphate/TTAB,234 phosphate and phosphate/TTAB235 ), sulfur (phosphate/TTAB/ACN),236 and selenium (histidine/acetic acid,237 SDS/β-CD232 ) by CE methods have been reported. Other examples include the analysis of beryllium (as an acetylacetone complex) in digested airborne dust238 and the determination of Fe(II)-, Cu(I)-, Ni(lI)-, Pd(II)-, and Pt(II)-cyano complexes and nitrate from leaching solutions of automobile catalytic converters.239 Metals including As, Co, Fe, Mn, Ni, and V are found at concentrations of 1–1000 ppm in crude oils. About 27–100% of the total metals in crude oil occur as organometallic porphyrin complexes. These petroporphyrins, derived from heme or chlorophyll, usually contain Ni(II) or V(IV)O and are distinguished by various ring types (Etio, DPEP, Rhodo analogs) and different substituents (H, n-alkyl). Petroporphyrin separations are of interest in geochemical sciences, environmental monitoring, and process control. Separation of petroporphyrin models [Ni(II) and V(II)O Etio I and Octaethyl type porphyrins] was approached by MEKC in electrolytes of varied composition: 15.8–22.5 mmol L−1 borate buffer containing 15–42 mmol L−1 SDS and 0–30% acetone, methylethylketone, acetonitrile, or methanol in a 6–30 min window.240 Several examples of the aforementioned methodologies and their application in environment compartments are compiled in Table 31.6.

31.4.7 EXPLOSIVES AND WARFARE RESIDUES 31.4.7.1 Explosives A SB-β-CD-assisted EKC method for the determination of cyclic nitramine explosives and related degradation intermediates and the 14 EPA listed explosives (borate/SDS electrolyte) has been described.241 A volatile electrolyte composed of SB-β-CD modified ammonium acetate buffer was selected for the EKC-MS detection of nitroaromatic and cyclic nitramine compounds in soil and marine sediment,83 as detailed in Table 31.7. The use of phosphate/SDS electrolytes was reported in the separation of the 14 listed nitramine and nitroaromatic explosives for the analysis of extracts of high explosives such as C-4, tetrytol, and detonating cord.242 The simultaneous detection of small cations (ammonium, sodium, potassium, calcium, magnesium, and strontium) and anions (bromide, chloride, nitrite, nitrate, sulfate, perchlorate, thiocyanate, and chlorate) from low explosives in postblast residue using an elaborate electrolyte composed of a cationic chromophore and modifiers (imidazole/HIBA/18-crown-6 ether/ACN), an anionic chromophore (1,3,6-naphthalenesulfonic acid) and flow reversal agent (tetramethylammonium hydroxide) has been presented.243 31.4.7.2 Warfare Residues A CZE method (borate buffer at pH 9.5) with LIF detection (He-Cd laser, 325 nm excitation; 450 nm emission) that simultaneously examines thiols (OPA derivatives) and cyanide (derivatized by OPAtaurine) aiming at monitoring for the potential contamination of drinking water supplies by chemical warfare (CW) nerve agents was described.244 Detection limits of 9.3 µg L−1 for cyanide and from 1.8 to 89 µg L−1 for the thiols were obtained. Table 31.7 compiles the details of CZE245,247 and EKC247,248 methodologies for inspection of contaminated water and soil for alkylphosphonic acids, the breakdown products of organophosphorus

CZE

CZE

Fe2+ , Cu2+ , Zn2+ , Mn2+ , Cd2+ in 9 min

CZE

CZE

CZE

CE Mode

K+ , Na+ , Ba2+ , Sr2+ , Ca2+ , Mg2+ , Mn2+ , Li+ , Fe2+ , Co2+ , Cd2+ , Ni2+ , Zn2+ , Pb2+ , Cr3+ , Cu2+ in 24 min

Lactate, butyrate, salicylate, propionate, phosphate, formate, citrate, acetate, Ba2+ , Ca2+ , Mg2+ , Ni2+ , Cu2+ in ca. 7 min + 2+ + 2+ + NH+ 4 , K , Ca , Na , Mg , Li (internal standard) in ca. 3.8 min

Small ions 2− 2− − Br− , I− , NO− 2 , S2 O3 , CrO4 , NO3 , 2− 2− 4− − Fe(CN)6 , SCN , MoO4 , WO4 in 3.5 min

Compounds

AIR airborne particulate

WATER river water

WATER rain water

WATER rain water river water drinking water industrial residual water WATER pond water

Matrix

Sequential extraction 78–87%

—

—

—

—

Preconcentration Procedure (Recovery)

20 mmol L−1 MES, 20 mmol L−1 His, 0.2 mmol L−1 18-crown-6 (pH 6.2) 10 mmol L−1 4-aminopyridine, 6.5 mmol L−1 HIBA (pH 4.5 adjusted with sulfuric acid) 200 mmol L−1 NH4Ac (pH 5.5), 0.5 mmol L−1 1,10-phenanthroline, 10 mmol L−1 hydroxylamine hydrochloride, acetone 20%

5 mmol L−1 borate buffer, 5 µmol L−1 fluorescein (pH 9.2)

50 mmol L−1 tetraborate 5% MeOH (pH 9.3)

Optimal Electrolyte

UV, 226 nm 29–142 µg L−1

UV, indirect detection, 214 nm 92–454 µg L−1

73

Fluorescence, indirect detection, (LED-induced fluorescence) 3.7–14.6 µmol L−1 Contactless conductometric detection 9.4–24 µg L−1

Continued

210

207

77

9

References

UV, 200 nm 0.02–0.1 mg L−1

Detection (LOD)

TABLE 31.6 Selected Applications of Electroseparation Methods for Small Ions and Organometallic Compounds in Environmental Samples

Separation Strategies for Environmental Analysis 943

CZE

CZE

CZE

CZE

CZE

CZE

CZE

C6–C12 perfluorinated carboxylic acids in 20 min

Bromide in 6 min

Potassium in ca. 11 min

Bromate in 4 min

Nitrite, nitrate in ca. 7 min

Ammonium in ca. 8 min

CE Mode

Zn2+ , Cu2+ , Co2+ , Fe2+ , Cd2+ in ca. 7 min

Compounds

TABLE 31.6 (Continued)

WATER River water SEWAGE

AIR Respirable, fine, and Coarse air Particulate WATER Lake water River water Tap water WATER Ground water Surface water WATER Ground water Surface water WATER Tap water River water WATER Sea water

Matrix

223

223

UV, 200 nm 0.1 mg L−1 UV, indirect detection, 214 nm 0.5 mg L−1 UV, 193 nm 1 µg L−1

5 mmol L−1 formic acid, 42 mmol L−1 NaCl (pH 3.5) 50 mmol L−1 18-crown-6 10 mmol L−1 imidazole (pH 4.5) 10 mmol L−1 phosphate, 10 mmol L−1 Na2 SO4 . (pH 3.2) SE: artificial sea water with pH adjusted to 3.0 using phosphate buffer TE: 600 mmol L−1 acetate (pre-rinse with 0.1 mmol L−1 DDAB) 20 mmol L−1 borate (pH 10 adjusted with 1 mol L−1 sodium hydroxide)

Spiked sample 61.6–102% Spiked sample 73.4–97.5%

—

tITP

226

228

UV, 210 nm 2.7–3.0 µg L−1 as nitrogen

UV, 190 nm 0.24 mg L−1 (as nitrogen)

224

214 UV, indirect detection, 270 nm 0.6–2.4 mg L−1

50 mmol L−1 TRIS, 7 mmol L−1 2,4-DNBA, 50% MeOH (pH 9.0)

—

Spiked sample 100–110%

211

References

UV, 265 nm 0.5–3 µg L−1

Detection (LOD)

30 mmol L−1 hydroxylamine hydrochloride, 0.1 mmol L−1 1,10-phenanthroline, 1% MeOH (pH 3.7)

Optimal Electrolyte

Andersen sampler, filters sonication

Preconcentration Procedure (Recovery)

944 Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Oxalate, glycolate, malonate, pyruvate, formate, suberate, malate, acetate, succinate, glyoxylate, phthalate, lactate, methanesulfonate, propionate, glutarate, benzoate in 8.5 min Chloride, nitrate, sulfate, formate, fumarate, malonate, succinate, maleate, glutarate, methanesulfonate, carbonate/adipate, malate, pimelate, acetate, suberate, oxalate, azelate, sebacate, glyoxylate, propionate, methacrylate, lactate, butyrate, hydroxy-iso-butyrate, valerate, capronate, enanthate, caprylate, pelargate in 9.5 min

WATER Rain water Surface water Drainage water WATER Snow

EKC

Ammonium, potassium, sodium, calcium, magnesium, manganese, lithium, chloride, nitrite, nitrate, sulfate, fluoride, phosphate in 5.5 min − 2− − F− , Cl− , NO− 2 , SO4 , NO3 , HCO3 in ca. 0.7 min AEROSOL Atmospheric aerosol GAS Vehicle emission WATER Size classified raindrops

EKC

EKC

EKC

WATER Rain water

WATER Sea water

EKC

EKC

WATER Lake water

CZE

2− − NO− 3 , Cl , SO4 , formate, acetate in ca. 3.7 min

3,4,5,-Trimethoxybenzoic acid, 4-hydroxyphenylacetic acid, salicylic acid, ferulic acid, p-coumaric acid, vanillic acid, 4-hydroxybenzoic acid in 11 min Iodide in ca. 35 min

47-mm PTFE filters (aerosol) KOH-coated quartz fiber (vehicle emission) freezing in liquid nitrogen (Guttalgor method)

Spiked sample 91–104%

7.5 mmol L−1 salicylic acid, 15 mmol L−1 Tris, 400 µmol L−1 DoTAH, 1050 µmol L−1 Ca2+ , 600 µmol L−1 Ba2+ (pH 6.2)

6.0 mmol L−1 chromate, 2.5 mmol L−1 CTAB, ACN 3.6%, (pH 9.5) 4 mmol L−1 NDC, 14.5 mmol L−1 Bis–Tris, 0.2 mmol L−1 TTAB (pH 6.2)

20 mmol L−1 MES/His 11.5 mmol L−1 18-crown-6, 10 µmol L−1 CTAB (pH 6.0)

—

—

UV, 232 nm Z-shaped flow cell, 90–600 nmol L−1

Continued

217

215

213

78

77

56

UV, 226 nm 0.6 µg L−1

SE: 0.5 mol L−1 NaCl, 25 mmol L−1 CTAC (pH 2.4) TE: 500 mmol L−1 MES (pH 6.0) 20 mmol L−1 MES, 20 mmol L−1 His, 0.2 mmol L−1 CTAB (pH 6.2)

Spiked sample tITP 99.6–100.4% Contactless conductometric detection 48–115 µg L−1 (inorganic ions) Contactless conductometric detection 10–250 µg L−1 UV, indirect detection, 254 nm 0.03–0.2 mg L−1 UV, indirect detection, 214 nm 3–10 µg L−1

228

UV, 214 nm

13 mmol L−1 tetraborate (pH 9.7)

SPE C18 58–108%

Separation Strategies for Environmental Analysis 945

Arsenite, arsenate, dimethylarsinate, monomethylarsonate in ca. 19 min

Organometallic Methylmercury, ethylmercury, phenylmercury, mercury (II), as cysteine complexes in 14 min

CZE

CZE

WATER Sea water

EKC

Nitrate, nitrite, bromide, iodide in ca. 5.8 min

Environmental reference material

WATER Lake water River water

WATER Sea water

WATER Sea water

WATER Tap water River water

WATER Sea water

EKC

EKC

in 2.5 min

Nitrate in ca. 7.6 min

F−

EKC

ClO− 4,

Iodide, iodate in ca. 33 min

SO2− 4 ,

Matrix

EKC

NO− 3,

CE Mode

Iodide in ca. 33 min

Cl− ,

Compounds

TABLE 31.6 (Continued)

LV-FASI

Spiked sample 86.6–111%

—

—

—

tITP

—

Preconcentration Procedure (Recovery)

20 mmol L−1 phosphate, 10 mmol L−1 borate (pH 9.28)

100 mmol L−1 boric acid, 12% MeOH (pH 9.1)

0.25 CTAC, MeOH 30%, 5 mmol L−1 chromate (pH 8.0) SE: 0.5 mol L−1 NaCl, 25 mmol L−1 CTAC (pH 2.4 adjusted with HCl) TE: 0.5 mol L−1 MES (pH 6.0 adjusted with NaOH) SE: 0.5 mol L−1 NaCl , 12.5 mmol L−1 CTAC (pH 2.4 adjusted with HCl) TE: 0.5 mol L−1 MES (pH 6.5) 0.1 mol L−1 sodium phosphate, 0.15 mol L−1 DDAPS (pH 6.2) (1) 10 mmol L−1 DDAPS, 10 mmol L−1 phosphate (pH 8.0) (2) 30 mmol L−1 DDAPS, 10 mmol L−1 phosphate (pH 8.0)

mmol L−1

Optimal Electrolyte

222

225

227

UV, 210 nm, 226 nm 0.23–10 µg L−1 (tITP)

UV, 210 nm 35 µg L−1 UV, 214 nm 5–11 µmol L−1

54

8

221

UV, 226 nm 0.2 µg L−1

Atomic fluorescence spectrometry 6.8–16.5 µg L−1 (as Hg) UV, 200 nm 24–93 µg L−1

218

References

UV, indirect detection, 254 nm 0.09–0.23 mg L−1

Detection (LOD)

946 Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

EKC

Fe(II)-, Cu(I)-, Ni(II)-, Pd(II)-, and Pt(II)-cyano complexes, nitrate in 20 min

AIR Digested airborne Dust (model sample) CATALYTIC CONVERTER RESIDUE leaching solutions

20 mmol L−1 phosphate, 100 mmol L−1 NaCI, 3 mmol L−1 NaCN, 1.2 mmol L−1 TBAB, 40 µmol L−1 TTAB (pH 11)

alkaline NaCN solutions

—

FASI

LLE chloroform 80–104% FASI

235

238

239

UV, 295 nm 1 mg L−1

UV, 208 nm 11–60 µg L−1

234

UV, 185 nm 1 µmol L−1

UV, 185 nm 0.1–0.5 mg L−1

237

Contactless conductometric detection 7.5–190 µg L−1 UV, 210 nm 26–67 mg L−1

8.75 mmol L−1 His (pH 4.0 adjusted with acetic acid) 50 mmol L−1 SDS, 5 mmol L−1 β-CD (pH 6.0) 10 mmol L−1 phosphate buffer, 0.35 mmol L−1 TTAB (pH 9.0) 20 mmol L−1 phosphate, 0.75 mmol L−1 TTAB (pH 9.0) 0.1 mol L−1 TRIS-nitrate, 50 mmol L−1 SDS (pH 7.8)

Sonication with water 90–101%

232

230

Chemiluminescence (luminol) 1–8 pmol L−1

20 mmol L−1 acetate buffer, 1 mmol L−1 EDTA (pH 4.7)

Spiked sample 98–103%

229

Contactless conductometric detection 10–39 µg L−1

4.5 mmol L−1 His (pH 3.40 adjusted with acetic acid)

—

FASI, field-amplified sample injection; SPE, solid-phase extraction; LV-FASI, large-volume field amplified stacking injection; TBAB, tetrabutylammonium bromide; TTAB, tetradecylytrimethylammonium bromide; CTAB, cetyltrimethylammonium bromide; NDC, 2,6-naphthalenedicarboxylic acid; Bis Tris, 2,2-Bis(hydroxymethyl)-2,29,20-nitrilotriethanol; DoTAH, dodecyltrimethylammoniumhydroxide; CTAC, cetyltrimethylammonium chloride; 2,4-DNBA, 2,4-dinitrobenzoic acid; SE, separation electrolyte; TE, terminating electrolyte; tITP, transient isotachophoresis; DDAB, dilauryldimethylammonium bromide; DDAPS, 3-(n,N-dimethyldodecylammonium)propane sulfonate; HIBA, α-hydroxyisobutyric acid; MES, 2-[morphine]ethanesulphonic acid; LLE, liquid–liquid extraction.

EKC

EKC

Be, as diacetylacetonate-beryllium in 15 min

EKC

Trimethyllead, triethyllead, diphenylselenide, phenylselenyl in 20 min Arsenite, arsenate, dimethyl arsenic acid in 4 min

WATER Ground water

WATER Ground water

CZE

Se(IV), Se(VI) in ca. 8 min

EKC

WATER Drainage water

CZE

Cr(III), Cr(IV) in ca. 4.5 min

2− AsO2− 2 , AsO4 , dimethylarsinic acid in 2 min

WATER Rinse water from the galvanic industry WATER Tap water Well water River water Waste water SOIL

CZE

Cr(III), Cr(IV) in 3 min

Separation Strategies for Environmental Analysis 947

SOIL

SOIL Contaminated soil SEDIMENT Marine sediment WATER Surface water Ground water SOIL WATER Tap water Ground water Artificial sea water SOIL

CZE

EKC

EKC

EKC

WATER SOIL

Matrix

CZE

CE Mode

SPE On-Guard-Ba On-Guard-Ag On-Guard-H 90–110%

—

—

PHWE 150 bar, 100◦ C SPE cation exchange cartridge

SPE Bond-Elut SCX

247

248

ESI-IT-MS 0.025–0.5 mg L−1

Conductivity

UV, indirect detection, 210 nm 1–2 µg L−1 (water samples) 25–50 µg L−1 (aqueous leachates of soil)

10 mmol L−1 SB-β-CD, 10 mmol L−1 NH4Ac (pH 6.9)

30 mmol L−1 L-His, 30 mmol L−1 MES, 0.7 mmol L−1 TTAOH, 0.03% wt Triton X-100 200 mmol L−1 boric acid, 10 mmol L−1 phenylphosphonic acid, 0.03% wt Triton X-100, 0.35 mmol L−1 DDAOH (pH 4.0)

83

246

ESI-IT-MS 5 µg L−1

15 mmol L−1 NH4Ac (pH 8.8 adjusted by ammonia solution)

245

Referencess

FPD, (P) mode, 0.1–0.5 µg L−1

Detection (LOD)

50 mmol L−1 NH4Ac (pH 9.0 adjusted by ammonia solution)

Optimal Electrolyte

TTAOH, tetradecyltrimethylammonium hydroxide; ESI-IT-MS, electrospray ionization ion trap mass spectrometry; SB-β-CD, sulfobutylether-β-cyclodextrin; PHWE, pressurized hot water extraction; FPD, flame photometric detector; SPE, solid-phase extraction; MES, 2-[morphine]ethanesulphonic acid; DDAOH, didodecyldimethylammonium hydroxide.

Methylphosphonic acid and its monoacid/ monoalkyl esters (ethyl, isopropyl, and pinacolyl) in 3 min

Methylphosphonic acid, ethylphosphonic acid, propylphosphonic acid, N -butylphosphonic acid, phenylphosphonic acid, ethyl methylphosphonic acid, ethyl methylthiophosphonic acid, isopropyl methylphosphonic acid, pinacolyl methylphosphonic acid, dimethyl phenylphosphonate in 16 min Methylphosphonic acid, ethylphosphonic acid, propylphosphonic acid, phenylphosphonic acid, isopropylphosphonic acid, methyl ethylphosphonic acid, ethyl methylphosphonic acid, ethyl ethylphosphonic acid, methyl propylphosphonic acid, propyl methylphosphonic acid in 15 min 2,4,6-Trinitrotoluene, 1,3,5-trinitrobenzene, hexahydro-1,3,5-trinitro-1,3,5-triazine, octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine, 2,4,6,8,10,12-hexanitro-2,4,6,8,10,12hexaazaisowurtzitane in 16 min Methylphosphonic acid and its monoacid/ monoalkyl esters (ethyl, isopropyl, and pinacolyl) in 10 min

Compounds

Preconcentration Procedure (Recovery)

TABLE 31.7 Selected Applications of Electroseparation Methods for Explosives and Warfare Residues in Environmental Samples

948 Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Separation Strategies for Environmental Analysis

949

2,3 6

14

4 6

4 1

7

Sudon III 8

PMPA

CH3CN

FPD response (mV)

8

MPA

EPA PhPA

Standards

1 2

9 2

EMPTA nBPA nPrPA

13 10

iPrMPA EMPA

(b) DMPhP

(a)

Water sample

10 5

0

Blank solution

11,12

4

6 Time (min)

8

10

0

4

8

12

16

Migration time (min)

FIGURE 31.8 Separation of explosives (a) and warfare residues (b). Electrolytes: (a) 12 mmol L−1 borate at pH 9 and 50 mmol L−1 SDS; (b) 50 mmol L−1 ammonium acetate at pH 9.0. Other conditions: +30 kV, direct UV-detection; (b) +30 kV; make up: 0.5% formic acid; flame photometric detector. Peak legends: (a): octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (1); hexahydro-1,3,5-trinitro-1,3,5-triazine (2); 1,3,5-trinitrobenzene (3); 1,3-dinitrobenzene (4); nitrobenzene (5); 2,4,6-dinitrotoluene (6); N-2,4,6-tetranitroN-methylaniline (7); 2,4-dinitrotoluene (8); 2,6-dinitrotoluene (9); 2-nitrotoluene (10); 3-nitrotoluene (11); 4-nitrotoluene (12); 2-amino-4,6-dinitrotoluene (13); and 4-amino-2,6-dinitrotoluene (14). (b): methylphosphonic acid (MPA), ethylphosphonic acid (EPA), n-propylphosphonic acid (nPrPA), n-butylphosphonic acid (nBPA), phenylphosphonic acid (PhPA), ethylmethylphosphonic acid (EMPA), ehtylmethylthiophosphonic acid (EMPTA), isopropylmethylphosphonic acid (iPrMPA), pinacolylmethylphosphonic acid (PMPA), and dimethylphenylphosphonate (DMPhP); (1) and (2) are suspected contaminants of water sample. (Modified from (a) C.A. Groom et al., J. Chromatogr. A, 999, 17–22, 2003 and (b) E.W.J. Hooijschuur et al., J. Chromatogr. A, 928, 187–199, 2001. With permission.)

nerve agents. The unusual detection system for CE (flame photometric detector)245 with an ability to detect alkylphosphonic acids a fold below the required 1 µg mL−1 level is worth mentioning here. Figure 31.8 illustrates the electromigration separation of EPA explosives and warfare residues.

31.4.8 AROMATIC SULFONATES Aromatic sulfonates and their amino- and hydroxy-derivatives are produced on a large scale in the chemical industry. Although the acute toxicity and the risk of bioaccumulation appear to be small, they became persistent and widespread environmental pollutants due to their high mobility in the aquatic compartments and limited biodegradability. The anionic charge of sulfonates within the entire practical pH range of electromigration separations and the UV-absorbing capability of the aromatic moiety make CZE-based methodologies of aromatic sulfonates feasible. However, improved selectivity and resolution of positional isomers have been sought by means of interactions with CDs,249 nonionic,250 and cationic250 surfactants as well as mixed micelles.251 An overview of the electrophoretic methods for the determination of benzene- and naphthalenesulfonates in water samples has been presented in the literature.252 Details of a CZE methodology for the separation of 14 aromatic sulfonates in river water253 and a CZE-MS evaluation of 22 compounds in influent and effluent samples of a wastewater treatment plant254 are organized in Table 31.8. With the idea that phytoremediation could be used as an alternative means of wastewater treatment of recalcitrant compounds, a simple borate buffer-based

CZE

CZE

EKC

EKC

EKC

Anthraquinone-1-sulphonic acid, anthraquinone-2-sulphonic acid, anthraquinone-1,5-disulphonic acid, anthraquinone-2,6-disulphonic acid, anthraquinone-1,8-disulphonic acid in 7.5 min

21 Naphthalenesulfonate derivatives in 28 min

21 Naphthalenesulfonate derivatives in 30 min

56 Aromatic organic acids were evaluated; 2,4-dichlorophenoxyacetic acid, 2,4,5-trichlorophenoxyacetic acid, orange II, trypan blue, 2,4,5-trichlorophenoxypropionic acid, 3-nitrobenzoic acid, 4-chlorobenzensulfonic acid (recovery studies) in 20 min

CZE

CE Mode

13 Aromatic sulfonates of a wide range of different structure in ca. 20 min

Aromatic sulfonates 14 Aromatic sulfonates of a wide range of different structure in 9.5 min

Compounds

WATER Deionized water SOIL

WATER Hydroponic media PLANT EXTRACTS Plant cultivated in spiked hydroponic media WATER River water WATER Industrial effluent

WATER River water Contaminated seepage water WATER Waste water treatment

Matrix

Soxhlet (soil), SPE cation exchange 38–101% (water) 26.5–94.4% (soil)

Sample was diluted 1/1000 and filtered

SPE

Aqueous extraction (plant samples) direct injection (water samples)

SPE C18 LiChrolut EN 72–132% SPE C18 LiChrolut EN >50%

Preconcentration Procedure (Recovery)

50 mmol L−1 borate, 100 mmol L−1 SDS (pH 8.7) 25 mmol L−1 tetraborate, 75 mmol L−1 Brij 35, 5 mmol L−1 octylamine (pH 9.0) 50 mmol L−1 borate, 100 mmol L−1 cholate (pH 8.3)

20 mmol L−1 tetraborate (pH 9.3)

5 mmol L−1 NH4Ac (pH 10.5)

25 mmol L−1 sodium borate (pH 9.3)

Optimal Electrolyte

UV, 214 nm

UV, 230 nm 20 µg L−1 UV, 220 nm 0.5–3.2 mg L−1

UV, 210 nm fluorescence, Xe lamp 0.2 mg L−1 (CE) 0.1 µg L−1 (SPE-CE) ESI-MS 0.1–0.4 mg L−1 (CE-MS) 0.1 µg L−1 (SPE-CE-MS) UV, 254 nm

Detection (LOD)

TABLE 31.8 Selected Applications of Electroseparation Methods for Aromatic Sulfonates, Dyes, and Surfactants in Environmental Samples

28

19

17

255

254

253

References

950 Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

WATER Waste water

WATER Spiked water SOIL

EKC

EKC

Sulfonated azo dyes: Acid Blue 113, Acid Red 73, Acid Red 13, Mordant Yellow 8, Acid Red 1, Acid Red 14, Acid Red 9, Acid Yellow 23 in 16 min Monosulfonated dyes: cresol red, acid blue, acid orange, tropaeolin, nuclear fast red, orange II, acid red in 27 min SPE ion-pair (water) C18 (soil) 42.5–107% (water) 17.9–105% (soil)

SPE Isolute ENV 54–81%

SPE C18 81–121% Charcoal adsorption

LLE, MeOH/NaOH (44.6–96.2%) SPE, C18 and SAX (33.1–75.7%)

SPE (C18 and SAX) 94–98% (marine water) SPE (Isolute ENV+) 77–93% (wastewater)

100 mmol L−1 cholate, 10% acetone (pH 8.35)

9.5 mmol L−1 NH4Ac, 0.1% Brij 35 (pH 9)

20 mmol L−1 borate (pH 9.2)

5 mmol L−1 NH4Ac (pH 9.0)

10 mmol L−1 phosphate, 40 mmol L−1 SDS, 30% ACN (pH 6.8)

50 mmol L−1 NH4Ac, 30% isopropanol (pH 5.6)

UV, 214 nm

ESI-MS single quadrupole 23–42 µg L−1 LIF, 514.5 nm Ar/Kr ion laser 1 nmol L−1 UV, 214 nm 19–230 mg L−1

UV, 214 nm 1 µg L−1 (SPE-CE-UV) ESI-MS (qualitative confirmation) UV, 200 nm 4 mg L−1

SPE, solid-phase extraction; ESI- MS, electrospray ionization mass spectrometry; LIF, laser-induced fluorescence; LLE, liquid–liquid extraction; LAS, linear alkylbenzenesulfonates.

WATER Ground water

WATER Waste water

WATER Industrial Waste water SEWAGE SLUDGE

WATER Coastal marine water Waste water Treatment plants

CZE

CZE

EKC

CZE

Eosin, fluorescein (internal standard) in 5.3 min

Dyes 5 Vinylsulfone and chlorotriazine reactive dyes in 34 min

19 LAS isomers (commercial formulation) in 45 min

Surfactants 4 LAS (C10, C11, C12, C13) in 45 min

29

20

259

84

18

87

Separation Strategies for Environmental Analysis 951

952

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

0.014

20+21 18 4

0.012

Absorption (relative intensity)

15 0.01

4

12+16+11

6+10

1 0.008

19 22

8

3

5

0.006

23

I I

0.004 EOF 0.002

0 2

4

6

10

8

12

14

Time (min)

FIGURE 31.9 Separation of aromatic sulfonates. Electrolyte: 12 mmol L−1 ammonium acetate at pH 10. Other conditions: +20 kV; direct UV detection at 214 nm. Peak legend: 2-amino-1,5-NDS (1); 1,3,6-NTS (2); 1,3-BDS (3); 1,5-NDS (4); 2,6-NDS (5); 1-OH-3,6-NDS (6); 1-amino-5-NS (7); BS (8); 1-amino-4-NS (9); 2-OH-3,6-NDS (10); 1-OH-6-amino-3-NS (11); 3-nitro-BS (12); 1-amino-6-NS (13); 4-methyl-BS (14); 1-OH4-NS (15); 4-chloro-BS (16); 2-amino-1-NS (17); 1-amino-7-NS (18); 4-chloro-3-nitro-BS (19); 1-NS (20); 2-NS (21); diphenylamine-4-sulfonate (22); and 8-NH2 -1-OH-3,6-NDS (23). Abbreviations: NS, naphthalenesulfonate; NDS, naphthalenedisulfonate; NTS, naphthalenetrisulfonate; BS, benzenesulfonate; BDS, benzenedisulfonate; and OH, hydroxy. (Modified from R. Loos et al., J. Mass Spectrom., 35, 1197–1206, 2000. With permission.)

CZE methodology was used to monitor sulfonated anthraquinones uptake in hydrophonic media, in which plants, bacteria, and algae had grown for 6 weeks.255 EKC methodologies for aromatic sulfonates in industrial effluents,19 mono- and dinaphthalenesulfonates as well as hydroxy- and amino-derivatives in river water,17 and phenoxyacetic acids in spiked water and soils28 are compiled in Table 31.8. Figure 31.9 illustrates the electromigration separation of aromatic sulfonates.

31.4.9 SURFACTANTS Anionic surfactants such as alkanesulfonates, alkyl sulfates, and LAS are water-soluble, surfaceactive materials that are consumed in large quantities in industrial and commercial formulations. Thus, their disposal may impact water reservoirs. LAS surfactants, which are used commercially, are quite complex mixtures containing several homologs and positional isomers. The degradation rates and toxicity of LAS depend on the alkyl chain length and the position of the phenyl ring. Since LAS are chromophoric-charged compounds, they are amenable to CZE18,87 determination. EKC mode is preferred when homologue discrimination or isomeric distribution is requested31,32 or mixture of surfactant classes are screened.33,34 LAS were determined in influent and effluent samples of wastewater treatment plants and coastal waters receiving untreated domestic effluents using a volatile buffer-based CZE method (UV and

Separation Strategies for Environmental Analysis

953

6 15

Absorbance at 224 nm (mAU)

10 5 78 10

9 11

12

14 + 15

13 16

5 1 2 4 3

17 1819 20

0 4

8

12 Time (min)

16

20

FIGURE 31.10 Separation of LAS surfactants. Electrolyte: 10 mmol L−1 sodium phosphate at pH 6.8, 40 mmol L−1 SDS and 30% acetonitrile. Other conditions: +15 kV applied voltage. Peak legend: positional isomers: 5-C10 (1), 4-C10 (2), 3-C10 (3), 2-C10 (4), 6-C11 (5), 5-C11 (6), 4-C11 (7), 3-C11 (8), 2-C11 (9), 6-C12 (10), 5-C12 (11), 4-C12 (12), 3-C12 (13), 2-C12 (14), 7-C13 (15), 6-C13 (16), 5-C13 (17), 4-C13 (18), 3-C13 (19), and 2-C13 (20); first number represents the attachment point of the phenyl ring to the carbon chain. (From J.M. Herrero-Martínez et al., Electrophoresis, 24, 681–686, 2003. With permission.)

MS detection).87 Phosphate and borate buffers containing acetonitrile were used to discriminate LAS homologues in sewage sludge, whereas unmodified buffers were applied to wastewater samples.18 Isomeric separation was only possible using electrolytes with high contents of SDS and acetonitrile as organic modifier. Table 31.8 compiles the details of the aforementioned methodologies. Figure 31.10 illustrates the electromigration separation of LAS surfactants.

31.4.10 DYES Synthetic dyes, including azo compounds, are widely used as coloring agents in a variety of products such as textiles, paper, leather, gasoline, and foodstuffs. Synthetic dyes persist even after conventional water treatment procedures due to their hydrophilic character and high solubility (they usually bear carboxylic or sulfonic acid groups in their structure) and, therefore, can be distributed in the environment from urban and industrial wastewater. CZE separation of synthetic dyes has been approached by simple (borate and citrate)256 and volatile buffers (ammonium acetate)257 modified by solvents as well as nonaqueous systems (ammonium acetate/acetic acid in MeOH).258 Environmental applications of CZE methodologies include the analysis of spent dyebaths and wastewater samples257 and the monitoring of groundwater migration, where eosin was used as a fluorescent tracer (details in Table 31.8).259 EKC separations of synthetic dyes include buffered SDS systems260 and polymeric electrolytes,261 modified by organic solvents.260 Examples of the determination of azo dyes, monoand disulfonated compounds in water samples,20 as well as synthetic dyes in spiked water and soil

954

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

matrices29 have been compiled in Table 31.8. Figure 31.11 illustrates the electromigration separations of synthetic dyes.

31.4.11 ENDOCRINE DISRUPTORS AND PHARMACEUTICALS Recently, it has been established that certain synthetic organic chemicals affect the reproductive health of higher organisms by contributing to infertility in various ways and even increasing the rate of cancer in reproductive organs. These chemicals have been termed “environmental estrogens” due to their disrupting effects on the endocrine system of hormone production and transmission. In addition to organochlorine insecticides, PCBs and dioxins (covered in Section 31.4.2), phenolic compounds such as bisphenol A—a widely used substance that is polymerized industrially into polycarbonate and nonylphenol—a product of the breakdown of large surfactants used in spermicides, some plastics and plasticizers such as phthalic acid esters, as well as pharmaceuticals such as synthetic estrogens belong to the list of chemicals suspected to have endocrine-disrupting effects. 31.4.11.1 Phenolic Compounds The CZE separation of halogenated phenolic and bisphenolic compounds from 25 potentially interfering phenolic derivatives for inspection in water, sludge, and sediments has been attempted262–264 using aqueous (borate buffer at pH 9.4) and nonaqueous (borate buffer in MeOH) systems (details in Table 31.9). MEKC separation and online concentration of bisphenolAand alkyl phenols has been approached in a series of experiments using SDS and other alkyl chain anionic surfactants, bile salts, and TTAB in organic solvent and CD-modified buffers.59,265,266 The EKC determination of several target endocrine disruptors in water267,268 and biosolids (sewage sludge and treated sludge)269 has been detailed in Table 31.9. The potential of MEEKC for the separation of priority endocrine-disrupting compounds in wastewater samples has been investigated.45 31.4.11.2 Phthalate Esters Excessive use of phthalate esters in industrial applications, mainly as plasticizers, is the main cause of their persistent presence in consumer goods and in the environment. The economic and social interest in the control of phthalate esters and the availability of analytical methodologies for areas such as environmental and food analyses have been discussed.270 31.4.11.3 Pharmaceutical Residues The occurrence and fate of pharmaceuticals from various prescription classes and related metabolites and medicinal products for veterinary use in aquatic environments as well as their removal is one of the emerging issues in environmental chemistry.271 More than 80 different prescribed pharmaceuticals, metabolites, and veterinary drugs have been detected in aquatic environment (e.g., sewage influent and effluent samples, surface and groundwater, and even drinking water), as reported by studies carried out in several European countries, United States, Canada, and Brazil.271 Tetracyclines,272 nonsteroidal anti-inflammatory,58,273 antidepressants,274 and mixtures of acidic drugs82,275 as well as veterinary drug276 residues have been determined by CZE (citric acid/citrate, borate, borate in MeOH, phosphate/MeOH, ammonium formate/formic acid/ACN, ammonium acetate, ammonium acetate/acetic acid/MeOH) and EKC (SDS/pH 2.5 phosphate buffer/ACN) methodologies in water samples from influent and effluent of sewage treatment plants as well as wastewater, river, surface and groundwater as compiled in Table 31.9. Estrogens have been receiving increased attention due to the already mentioned possible interference with the reproductive

Separation Strategies for Environmental Analysis

(a)

955

100

Effluent m/z 343.5 5 (OH)

%

(b)

0 100

Effluent m/z 264.7 4 (OH/OH)

%

0 (c)

100

Effluent m/z 270.7

4 (OH/CI)

%

0 (d)

100

Effluent m/z 501

2 (OH)

%

(e)

0 100

2 (OH)

1 (OH)

3 (OH/OH) NSA

Ref. mixture RIC 4 (VS/CI) 5 (OH)

5 (VS)

%

4 (OH/OH) 4 (OH/CI)

0 10

15

20

25 Time (min)

30

35

40

FIGURE 31.11 Identification of synthetic dyes in a sewage extract effluent (mass traces a–d) and a reference mixture (e) by CE-MS. Electrolyte: 5 mmol L−1 ammonium acetate with 40% acetonitrile. Other conditions: +30 kV applied voltage; sheath liquid: 80:20 isopropanol:water. Abbreviations: OH, hydroxy form; Cl, chloride form; VS, vinylsulfone; NSA, naphthalene sulfonic acid; RIC, reconstructed ion chromatogram. (From T. Poiger et al. J. Chromatogr. A, 886, 271–282, 2000. With permission.)

17β-Estradiol, diethylstilbestrol, ethynylestradiol, octylphenol, nonylphenol, bisphenol A in 15 min

4-Bromo-3-methylphenol, 2-bromo-4-methylphenol, pentabromophenol, 2,4,6-tribromophenol, 2,4-dibromophenol, 2-bromophenol, 2,6-dibromophenol, tetrabromobisphenol A, tetrachlorobisphenol A in ca. 8 min 2,6-Dibromophenol, 2,4,6-tribromophenol, tetrabromobisphenol A, pentabromophenol, tetrachlorobisphenol A in 28 min Octylphenol, nonylphenol in 10 min

Endocrine disruptors 2,4,6-Tribromophenol, pentabromophenol, tetrabromobisphenol A, tetrachlorobisphenol A, 2,6-Dibromophenol in 28 min

Compounds

WATER Waste water treatment effluents and sludges

EKC

WATER River water

SEWAGE SEDIMENT River sediment Marine sediment

CZE

EKC

WATER Waste water

WATER River water Waste water

Matrix

CZE

CZE

CE Mode

Spiked sample analytes solubilized in 10% ACN:90% buffer

SPE C18 spiked sample 25.6–50.8%

MSPD-LVSEP Florisil 75.7–106.4%

SPE-LVSEP Polystyrene– divinylbenzene 95.5–105.8% (river water) 73–94% (Waste water) SPE PS–DVB

Preconcentration Procedure (Recovery)

45

267

UV, 210 nm, 230 nm 2.7–5.3 µg L−1

UV, 214 nm 50 mg L−1 (tested)

UV, 214 nm low mg L−1 range

20 mmol L−1 tetraborate in MeOH (nonaqueous) (pH 9.4) 25 mmol L−1 phosphate, 200 mmol L−1 SDS, 900 mmol L−1 butanol, 80 mmol L−1 heptane, 20% propanol (pH 2) 100 mmol L−1 phosphate, 12.5% ACN, 25 mmol L−1 SDS, 1 mmol L−1 HP-β-CD (pH 1.8)

264

263

UV, 210 nm

20 mmol L−1 sodium tetraborate (pH 9.6)

262

References

UV, 210 nm, 230 nm 0.4–1.7 µg L−1

Detection (LOD)

20 mmol L−1 tetraborate in MeOH (non-aqueous) (pH 9.4)

Optimal Electrolyte

TABLE 31.9 Selected Applications of Electroseparation Methods for Endocrine Disruptors and Pharmaceuticals in Environmental Samples and Miscellaneous 956 Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

CZE

WATER Surface water SEWAGE

WATER River water SEWAGE

CZE

Clofibric acid, naproxen, bezafibrate in 15 min

WATER Bottled water

CZE

Acetylsalicylic acid, ibuprofen, fenoprofen, naproxen, diclofenac, ketoprofen in 26 min Fluoxetine, venlafaxine, citalopram, sertraline, paroxetine, clomipramine, trazodone in 17 min

WATER Surface water

CZE

WATER River water

WATER Spiked water

EKC

CZE

WATER River water

EKC

Tetracycline, oxytetracycline, doxycycline in 17.5 min

Pharmaceutical residues Clofibric acid, naproxen, bezafibrate, diclofenac, ibuprofen, mefenamic acid in ca. 20 min

Estriol, phenol, trichlorophenol, bisphenol A, pentachlorophenol, butylphenol, estrone, β-estradiol, diethylstilbestrol, ethinylestradiol, nonylphenol in 25 min Estrone, β-estradiol, and ethynylestradiol in 10 min

LLE/SPE hexane/MTBE (50:50)/Bondesil ODS 43.1–98% Online SPE STRATA-X 94–106% SPE-LVSEP LiChrolut RP-18 64–95% SPE hydrophilic divinylbenzene 85–99% SPE LiChrolut RP-18 70–89% (surface water) LLE/SPE hexane/MTBE (50:50)/LiChrolut RP-18 50–58% (sewage)

SPE C18; sweeping >96%

Spiked sample

Continued

275

274 ESI-TOF-MS 13–53 µg L−1 (SPE-CE-MS) ESI-MS single quadrupole 100 ng L−1

273

UV, 214 nm 2–9 ng L−1

20 mmol L−1 NH4Ac, 60% MeOH (pH 4.5 adjusted with acetic acid)

272

UV, 260 nm 1.6–2 µg L−1

50 mmol L−1 citric acid (pH 2.5 adjusted with HCl 1 mol L−1 ) 30 mmol L−1 tetraborate, 70% MeOH (pH 9.8) 1.5 mol L−1 formic acid, 50 mmol L−1 ammonium formate, 15% ACN

82

278

ESI-MS single quadrupole 25–59 µg L−1

UV, 214 nm 0.16–0.30 nmol L−1

30 mmol L−1 phosphoric acid, 80 mmol L−1 SDS, 20% MeOH

268

20 mmol L−1 NH4Ac (pH 5.1 adjusted with acetic acid)

UV, 200 nm 2.0–7.4 mg L−1

20 mmol L−1 CAPS, 25 mmol L−1 SDS, 15% ACN (pH 11.5)

Separation Strategies for Environmental Analysis 957

EKC

Pollen allergens and organic pollutants (nonspecified); 20 compounds in 13 min

Centrifugation, freezing–thawing, filtering Centrifugation, freezing–thawing, filtering LLE acetone followed by acidic and alkaline aqueous solutions

SPE-SRMM, SRMM-ASEI, FESI-RMM LiChrolut RP-18 70–100%

SPE-LVSS Oasis HLB 75.6–100.3%

Preconcentration Procedure (Recovery)

UV, 230, 240, 278 nm 0.89–2.77 mg L−1 UV, 230, 240, 278 nm 0.73–2.81 mg L−1 UV, 206 nm

25 mmol L−1 sodium tetraborate, 100 mmol L−1 SDS (pH 9.3) 20 mmol L−1 TRIS, 5 mmol L−1 H3 PO4 , 50 mmol L−1 SDS (pH 8.7)

UV, 214 nm 0.07–0.23 µg L−1 (SRMM) 0.050–0.195 µg L−1 (SRMM-ASEI) 0.7–1.6 µg L−1 (FESI-RMM)

25 mmol L−1 phosphate, 75 mmol L−1 SDS, 40% ACN (pH 2.5 adjusted with HCl)

25 mmol L−1 sodium tetraborate (pH 9.3)

UV, 265 nm 2.59–22.95 µg L−1

Detection (LOD)

45 mmol L−1 sodium phosphate, 10% MeOH (pH 7.3)

Optimal Electrolyte

288

286

286

58

276

References

HP-β-CD, hydroxypropyl-β-cyclodextrin; CAPS, cyclohexylamino-1-propanesulfonic acid; LVSS, large-volume sample stacking; SRMM, stacking with reverse migrating micelles; SRMM– ASEI, stacking with reverse migrating micelles–anion selective exhaustive injection; FESI–RMM, field-enhanced sample injection with reverse migrating micelles; LVSEP, large-volume sample stacking using the electroosmotic flow pump; ESI-TOF-MS, electrospray ionization time of flight mass spectrometry; ESI-MS, electrospray ionization mass spectrometry; MSPD, matrix solid-phase dispersion; PS–DVB, polystyrene–divinylbenzene; LLE, liquid–liquid extraction; SPE, solid-phase extraction.

EKC

WATER Water bloom Crude extracts WATER Water bloom Crude extracts AIR Airborne dust samples

WATER Mineral water

EKC

CZE

WATER Ground water

Matrix

CZE

CE Mode

Anatoxin-a, microcystin-LR, cylindrospermopsin in 9 min

Miscellaneous Anatoxin-a, microcystin-LR, cylindrospermopsin in 3 min

Sulfapyridine, sulfamethazine, sulfamerazine, sulfamether, sulfadiazine, sulfadimethoxine, sulfamethoxazole, sulfachlorpyridazine, sulfamethizole in 18 min Diclofenac, ibuprofen, fenoprofen, naproxen, ketoprofen in 35 min

Compounds

TABLE 31.9 (Continued)

958 Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Separation Strategies for Environmental Analysis

959

system of man and animals. A few applications have been reported (Table 31.9).277,278 Figure 31.12 illustrates the electromigration separation of endocrine disruptors and pharmaceuticals.

31.4.12 MISCELLANEOUS 31.4.12.1 Humic Substances CZE of natural organic matter279 and separation methods for humic substances280 have been subjects of recent reviews. CZE and EKC separations (borate/TRIS/EDTA modified by CDs/SDS) and (a) 40 35

5

30 4

AU

25 1

20

3

15 2

10 5

River water

0

Blank

5 0

5

40

10

15

20

25

35 30 AU

30 5

4

3

25

1

20 15

Effluent waste water

10

2

5 0

Blank

−5

5

35

10

15

20

25

30 4

AU

25 20 2

Influent waste water

10 5

5

3

1

15

30

0 Blank

5 0 0

5

10

15

20

25

30

Migration time (min)

FIGURE 31.12 Separation of (a) endocrine disruptors and (b) veterinary drug residues. Electrolytes: (a) 20 mmol L−1 sodium tetraborate at pH 9.4; (b) 45 mmol L−1 sodium phosphate at pH 7.3 with 10% methanol. Other conditions: (a) −30 kV; direct UV-detection at 210 nm; (b) +25 kV; direct UV-detection at 265 nm. Peak legends: (a): 2,4,6-tribromophenol (1), pentabromophenol (2), 2,6-dibromophenol (3), tetrabromobisphenol A (4), and tetrachlorobisphenol A (5). (b): Sulfapyridine (SPD), sulfamethazine (SMZ), sulfamerazine (SMR), sulfamether (SMT), sulfadiazine (SDZ), sulfadimethoxine (SDM), sulfamethoxazole (SMX), sulfachloropyridazine (SCP), sulfamethizole (SMI), and p-aminobenzoic acid (PABA). (Modified from (a) E. Blanco et al., J. Chromatogr. A, 1071, 205–211, 2005 and (b) J.J. Soto-Chinchilla et al., Electrophoresis, 27, 4360–4368, 2006. With permission.)

960

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(b) 20 17.5 Intensity (mAU)

PABA

SDZ SMX SDM SMI

SMR SMT

15

SCP SMZ

12.5

SPD

10 7.5 5 2.5

Spiked ground water

0 2

4

6

8

10 Time (min)

12

14

16

18

FIGURE 31.12 (Continued)

characterization (TOF-MS) of humic acids isolated from Antarctica soil281 as well as interactions of organic matter and environmental pollutants282,283 were reported (details in Table 31.9).

31.4.12.2 Algal Toxins Methods for determining microcystins and microcystin-producing cyanobacteria have been reviewed.284 Details of CZE and EKC analyses of cyanobacterial toxins in environmental samples285–287 are compiled in Table 31.9. Figure 31.13 presents the separation of cyanobacterial toxins.

31.4.12.3 Other Applications CZE and MEKC methods were used to monitor extraction procedures involving pollen allergens and organic pollutants from dust samples collected before, during, and after pollen seasons at different locations (car-traffic tunnel in Prague and a metro station in Paris) using air-filtration devices.288 Water and acetic acid extracts were analyzed by CZE using acetic acid as background electrolyte whereas water and alkaline water–SDS-buffer extracts were analyzed by MEKC in Tris-phosphate– SDS electrolytes. Significant differences were found in the profiles of dust extracts from different origins. Drinking water and humidity condensate samples collected from U.S. Space Shuttle and the Russian Mir Space Stations are analyzed routinely at the NASA–Johnson Space Center as a means of verifying water quality and monitoring the environment of the spacecraft.289 Anions and cations were determined by ion chromatography whereas carboxylates and amines were determined by CE (phthalate/TTAB for carboxylates and imidazole/HIBA for amines). Results showed that Shuttle water is of distilled quality whereas Mir-recovered water contains various levels of minerals. Organic ions were rarely detected in potable water samples but were present in humidity condensates. Table 31.9 compiles the details of the previously cited methodologies.

Separation Strategies for Environmental Analysis

961

(a)

(b)

20 a

YA a

15 10

YR

10

mAU

mAU

LR

b

5

YR

YA

7

LR

b

4

1

0 2

2.5

3

3.5

4

4.5

2

Time (min)

4

6

8

10

12

Time (min)

FIGURE 31.13 (a) CZE and (b) MEKC separation of cyanobacterial toxins. Electrolytes: (a) 25 mmol L−1 borate at pH 9.3; (b) 25 mmol L−1 borate at pH 9.3 containing 75 mmol L−1 SDS. Other conditions: (a) and (b) +25 kV; direct UV-detection at 238. Peak legends: microcystin YA (YA), microcystin YR (YR), and microcystin LR (LR); a: water bloom sample; and b: spiked sample. (Modified from G. Vasas et al. J. Biochem. Biophys. Methods, 66, 87–97, 2006. With permission.)

31.5 A METHOD DEVELOPMENT GUIDE FOR ENVIRONMENTAL ANALYSIS Figure 31.14 depicts a simplified diagram that can be useful during the first stages of method development for environmental analysis. In CE, method development relies strongly on the knowledge of the analyte structural features. For neutral compounds (for instance, many classes of pesticides, PAH, PCB, dioxins, derivatized carbonyls, etc.), the obvious choice is the selection of an EKC protocol. As a first attempt, MEKC in borate/SDS electrolytes is usually recommended. Depending on the result of the exploratory run, several additives can be further tested as a means to modulating the solute–micelle interaction. Other EKC modes are also feasible options, especially when SDS fails in promoting separation, which is the case of either highly hydrophilic or highly hydrophobic compounds. Occasionally, EKC has been the technique of choice for certain classes of ionizable pollutants, such as phenol derivatives and amines, simply because the separation of a large number of positional isomers and conformers is attempted and extra selectivity is required. Method development procedures for charged compounds, either organic pollutants with ionizable functional groups within the CE operational pH range (phenols, amines, carboxylic acids, many dyes, pharmaceuticals, etc.) or explicitly ionic compounds such as small ions, sulfonates, LAS surfactants, and so forth, will follow two distinct routes in the diagram of Figure 31.14, depending on the compound’s UV-absorbing characteristics (only UV-absorbance detectors were considered here because they are practically a part of all commercially available CE equipments). If the pollutant is charged and exhibits UV-absorbing properties, the CZE mode is readily recommended. For basic pollutants, moderately low to low pH buffers are indicated and the analyte migrates coelectroosmotically as a cation (protonated species) whereas for acidic pollutants, high pH buffers will promote the analyte dissociation and it migrates counterelectroosmotically as an anion. In both cases, buffer pH and concentration are the variables to optimize before the addition of any modifiers is considered. If the pollutant is charged but does not present a chromophoric group, either a cationic (separation of cation) or an anionic (separation of anions) chromophore must be added to the electrolyte and

962

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Nature of pollutant NEUTRAL

Chiral selectors

MEKC

Cyclodextrin Crown ether Other

UV absorbing

UV non-absorbing

CZE

EXPLORATORY RUN (borate/SDS) Hydrophobic cpds Hydrophilic cpds

Low pH BASIC

CATIONIC

EXPLORATORY RUN (proper chromophore)

ACID

Organic solvents

Additives

Urea Neutral surfactants

EXPLORATORY RUN (buffered systems)

Bile salts

Additives Organic solvents

OTHER EKC MODES

Anionic

Cationic surfactants

ANIONIC

High pH

Amines Other

Complexing agent

Cationic

EKC

CHARGED (ionic or ionizable)

CHIRAL

EOF reverser

FIGURE 31.14 A method development guide for environmental analysis.

indirect detection is performed. For the cations, complexing agents will further improve resolution of similar mobility species. In the case of nonabsorbing anions, a judicious choice of both co- and counterions is recommended, the co-ion due to peak symmetry considerations while the counterion for signal enhancement. Furthermore, an EOF modifier (alkyl quaternary ammonium salt) is usually selected to expedite the separation. Finally, in the electrolyte design during chiral pollutants method development, a chiral selector must always be considered. Cyclodextrins are among the most commonly used chiral additives and can be employed in both EKC and CZE methodologies, depending on the nature of the pollutant.

31.6 CONCLUDING REMARKS On the basis of the diversity of applications presented in this chapter, it becomes clear that CE has found a niche among separation techniques for environmental analysis. Future directions will likely include refinement of preconcentration strategies and further detector improvements to achieve the desirable low-concentration limits of detection. In addition, with the consolidation of CE-MS technology, more robust methods are likely to emerge with enhanced sensitivity and superior selectivity, improving further the acceptance of CE in the routine determination of pollutants in real samples.

ACKNOWLEDGMENTS The authors wish to acknowledge the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and the Fundação de Amparo à Pesquisa do Estado de São Paulo (Fapesp) of Brazil for financial support (Fapesp 04/08503-2; 04/08931-4) and fellowships (CNPq 306068/2003-6).

Separation Strategies for Environmental Analysis

963

REFERENCES 1. E. Dabek-Zlotorzynska and V. Celo, Recent advances in capillary electrophoresis and capillary electrochromatography of pollutants, Electrophoresis, 27, 304–322, 2006. 2. E. Dabek-Zlotorzynska, H. Chen and L. Ding, Recent advances in capillary electrophoresis and capillary electrochromatography of pollutants, Electrophoresis, 24, 4128–4149, 2003. 3. I. Ali, V.K. Gupta and H.Y. Aboul-Enein, Chiral resolution of some environmental pollutants by capillary electrophoresis, Electrophoresis, 24, 1360–1374, 2003. 4. D. Martínez, M.J. Cugat, F. Borrul and M. Calull, Solid-phase extraction coupling to capillary electrophoresis with emphasis on environmental analysis, J. Chromatogr. A, 902, 65–89, 2000. 5. M.C. Bruzzoniti, C. Sarzanini and E. Mentasti, Preconcentration of contaminants in water analysis, J. Chromatogr. A, 902, 289–309, 2000. 6. D. Martínez, F. Borrull and M. Calull, Strategies for determining environmental pollutants at trace levels in aqueous samples by capillary electrophoresis, Trends Anal. Chem., 18, 282–291, 1999. 7. G.W. Sovocool, W.C. Brumley and J.R. Donnelly, Capillary electrophoresis and capillary electrochromatography of organic pollutants, Electrophoresis, 20, 3297–3310, 1999. 8. X.-P. Yan, X.-B. Yin, D.-Q. Jiang and X.-W. He, Speciation of mercury by hydrostatically modified electroosmotic flow capillary electrophoresis coupled with volatile species generation atomic fluorescence spectrometry, Anal. Chem., 75, 1726–1732, 2003. 9. M.I. Turnes-Carou, P. López-Mahía, S. Muniategui-Lorenzo, E. Fernández-Fernández and D. Prada-Rodríguez, Capillary zone electrophoresis for the determination of light-absorbing anions in environmental samples, J. Chromatogr. A, 918, 411–421, 2001. 10. A. Cavallaro, V. Piangerelli, F. Nerini, S. Cavalli and C. Reschiotto, Selective determination of aromatic amines in water samples by capillary zone electrophoresis and solid-phase extraction, J. Chromatogr. A, 709, 361–366, 1995. 11. S. Morales and R. Cela, Highly selective and efficient determination of US Environmental Protection Agency priority phenols employing solid-phase extraction and non-aqueous capillary electrophoresis, J. Chromatogr. A, 896, 95–104, 2000. 12. M. Molina, D. Peréz-Bendito and M. Silva, Multi-residue analysis of N-methylcarbamate pesticides and their hydrolytic metabolites in environmental waters by use of solid-phase extraction and micellar electrokinetic chromatography, Electrophoresis, 20, 3439–3449, 1999. 13. A. Segura-Carretero, C. Cruces-Blanco, S. Pérez Duran and A. Fernández Gutiérrez, Determination of imidacloprid and its metabolite 6-chloronicotinic acid in greenhouse air by application of micellar electrokinetic capillary chromatography with solid-phase extraction, J. Chromatogr. A, 1003, 189–195, 2003. 14. C.-E. Lin, C.-C. Hsueh, T.-Z. Wang, T.-C. Chiu and Y.-C. Chen, Migration behavior and separation of s-triazines in micellar electrokinetic capillary chromatography using a cationic surfactant, J. Chromatogr. A, 835, 197–207, 1999. 15. O. Brüggemann and R. Freitag, Determination of polycyclic aromatic hydrocarbons in soil samples by micellar electrokinetic capillary chromatography with photodiode-array detection, J. Chromatogr. A, 717, 309–324, 1995. 16. S. Frías, M.J. Sánchez and M.A. Rodríguez, Determination of triazine compounds in ground water samples by micellar electrokinetic capillary chromatography, Anal. Chim. Acta, 503, 271–278, 2004. 17. S.J. Kok, E.H.M. Koster, C. Gooijer, N.H. Velthorst and U.A.Th. Brinkman, Separation of twenty-one naphthalene sulfonates by means of capillary electrophoresis, J. High Resol. Chromatogr., 19, 99–104, 1996. 18. K. Heinig, C. Vogt and G. Werner, Determination of linear alkylbenzenesulfonates in industrial and environmental samples by capillary electrophoresis, Analyst, 123, 349–353, 1998. 19. S. Angelino, A. Bianco Prevot, M.G. Genaro and E. Pramauro, Ion-interaction high-performance chromatography and micellar electrokinetic capillary chromatography: Two complementary techniques for the separation of aromatic sulfonated compounds, J. Chromatogr. A, 845, 257–271, 1999. 20. E.R. Cunha and M.F. Alpendurada, Comparison of HPLC and micellar electrokinetic chromatography in determination of sulfonated azo dyes in waste water, J. Liq. Chromatogr. Rel. Technol., 25, 1835–1854, 2002. 21. R.J.G. Jeevan, M. Bhaskar, R. Chandrasekar and G. Radhakrishan, Analysis of arylamine isomers by micellar electrokinetic chromatography, Electrophoresis, 23, 584–590, 2002.

964

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

22. M. Wang, D. Wu, Q. Yao and X. Shen, Separation and selectivity in micellar electrokinetic chromatography using sodium dodecyl sulfate micelles or Tween 20-modified mixed micelles, Anal. Chim. Acta, 519, 73–78, 2004. 23. L. Song, Q. Ou, W. Yu and G. Li, Separation of 6 phenylureas and chlorsulfuron standards by micellar, mixed micellar and microemulsion electrokinetic chromatography, J. Chromatogr. A, 699, 371–382, 1995. 24. M.H. Chen and W.H. Ding, Separation and migration behavior of positional and structural naphthalenesulfonate isomers by cyclodextrin-mediated capillary electrophoresis, J. Chromatogr. A, 1033, 167–172, 2004. 25. E.A. Pereira, M.F.M. Tavares and A.A. Cardoso, Alternative methodologies for the determination of aldehydes by capillary electrophoresis, J. AOAC Int., 82, 1562–1570, 1999. 26. M.-I. Acedo-Valenzuela, T. Galeano-Días, N. Mora-Díez and A. Silva-Rodríguez, Determination of neutral and cationic herbicides in water by micellar electrokinetic capillary chromatography, Anal. Chim. Acta, 519, 65–71, 2004. 27. E. Dabek-Zlotorzynska and E.P.C. Lai, Separation of polynuclear aromatic hydrocarbons by micellar electrokinetic capillary chromatography using sodium taurodeoxycholate modified with organic solvents, J. Cap. Elec., 3, 31–35, 1996. 28. W.C. Brumley and C.M. Brownrigg, Electrophoretic behavior of aromatic-containing organic acids and the determination of selected compounds in water and soil by capillary electrophoresis, J. Chromatogr., 646, 377–389, 1993. 29. W.C. Brumley, C.M. Brownrigg and A.H. Grange, Capillary liquid-chromatography-mass spectrometry and micellar electrokinetic chromatography as complementary techniques in environmental analysis, J. Chromatogr. A, 680, 635–644, 1994. 30. J.M. Bermúdez-Saldaña, M.A. García, M.J. Medina-Hernández and M.L. Marina, Micellar electrokinetic chromatography with bile salts for predicting ecotoxicity of aromatic compounds, J. Chromatogr. A, 1052, 171–180, 2004. 31. J.M. Herrero-Martínez, E.F. Simó-Alfonso and G. Ramis-Ramos, Separation of homologues and isomers of linear alkylbenzenesulfonates by capillary electrophoresis with sodium dodecyl sulfate, carboxylic acids and bile salts, Electrophoresis, 24, 681–686, 2003. 32. J.M. Herrero-Martínez, E.F. Simó-Alfonso and G. Ramis-Ramos, Separation and determination of homologues of linear alkylbenzenesulfonates by nonaqueous capillary zone electrophoresis using alkylammonium salts in ethanol, Electrophoresis, 22, 2017–2024, 2001. 33. V. Bernabé-Zafón, S. Ortega-Gadea, E.F. Simó-Alfonso and G. Ramis-Ramos, Characterization and quantitation of mixtures of alkyl ether sulfates and carboxylic acids by capillary electrophoresis with indirect photometric detection, Electrophoresis, 24, 2805–2813, 2003. 34. V. Bernabé-Zafón, J.R. Torres-Lapasió, S. Ortega-Gadea, E.F. Simó-Alfonso and G. Ramis-Ramos, Capillary electrophoresis enhanced by automatic two-way background correction using cubic smoothing splines and multivariate data analysis applied to the characterisation of mixtures of surfactants, J. Chromatogr. A, 1065, 301–313, 2005. 35. W. Wall, K. Chan and Z. El Rassi, Electrically driven microseparation methods for pesticides and metabolites. VI. Surfactant-mediated electrokinetic capillary chromatography of aniline pesticidic metabolites derivatized with 9-fluoroenylmethyl chloroformate and their detection by laser-induced fluorescence, Electrophoresis, 22, 2320–2326, 2001. 36. J.T. Koch, B. Beam, K.S. Phillips and J.F. Wheeler, Hydrophobic interaction electrokinetic chromatography for the separation of polycyclic aromatic hydrocarbons using non-aqueous matrices, J. Chromatogr. A, 914, 223–231, 2001. 37. W. Wall and Z. El Rassi, Electrically driven microseparation methods for pesticides and metabolites. V. Micellar electrokinetic capillary chromatography of aniline pesticidic metabolites derivatized with fluorescein isothiocyanate and their detection in real water at low levels by laser-induced fluorescence, Electrophoresis, 22, 2312–2319, 2001. 38. T.W. Moy, P.L. Ferguson, A.H. Grange, W.H. Matchett, V.A. Kelliber, W.C. Brumley, J. Glassman and J.Q. Farley, Development of separation systems for polynuclear aromatic hydrocarbon environmental contaminants using micellar electrokinetic chromatography with molecular micelles and free zone electrophoresis, Electrophoresis, 19, 2090–2094, 1998.

Separation Strategies for Environmental Analysis

965

39. H. Harino, S. Tsunoi, T. Sato and M. Tanaka, Applicability of micellar electrokinetic chromatography with a double-chain surfactant having two sulfonate groups to the determination of pollutant phenols in water, Anal. Sci., 16, 1349–1351, 2000. 40. S.K. Wiedmer, J. Hautala, J.M. Holopainen, P.K.J. Kinnunen, and M.-L. Riekkola, Study on liposomes by capillary electrophoresis, Electrophoresis, 22, 1305–1313, 2001. 41. D. Kaniansky, E. Krˇcmová, V. Madajová, M. Masár, J. Marák and F.I. Onuska, Determination of nitrophenols by capillary zone electrophoresis in a hydrodynamically closed separation compartment, J. Chromatogr. A, 772, 327–337, 1997. 42. C. García-Ruiz, Y. Martín-Biosca, A.L. Crego and M.L. Marina, Rapid enantiomeric separation of polychlorinated biphenyls by electrokinetic chromatography using mixtures of neutral and charged cyclodextrin derivatives, J. Chromatogr. A, 910, 157–164, 2001. 43. R.S. Brown, J.H.T. Luong, O.H.J. Szolar, A. Halasz and J. Hawari, Cyclodextrin-modified capillary electrophoresis: Determination of polycyclic aromatic hydrocarbons in contaminated soils, Anal. Chem., 68, 287–292, 1996. 44. C.A. Groom and J.H.T. Luong, Sulfobutylether-beta-cyclodextrin-mediated capillary electrophoresis for separation of chlorinated and substituted phenols, Electrophoresis, 18, 1166–1172, 1997. 45. B. Fogarty, E. Dempsey and F. Regan, Potential of microemulsion electrokinetic chromatography for the separation of priority endocrine disrupting compounds, J. Chromatogr. A, 1014, 129–139, 2003. 46. D. Shobat and E. Grushka, Use of calixarenes to modify selectivities in capillary electrophoresis, Anal. Chem., 66, 747–750, 1994. 47. L. Arce, A. Segura-Carretero, A. Rios, C. Cruces, A. Fernández and M. Valcárcel, Use of calixarene compounds as selectivity modifiers in capillary electrophoresis separations, J. Chromatogr. A, 816, 243–249, 1998. 48. K. Bächmann, B. Göttlicher, I. Haag, K.I. Han, W. Hensel and A. Mainka, Capillary electrokinetic chromatography with a suspension of chromatographic particles, J. Chromatogr. A, 688, 283–292, 1994. 49. B. Göttlicher and K. Bächmann, Application of particles as pseudo-stationary phases in electrokinetic chromatography, J. Chromatogr. A, 780, 63–73, 1997. 50. S. Terabe and T. Isemura, Ion-exchange electrokinetic chromatography with polymer ions for the separation of isomeric ions having identical electrophoretic mobilities, Anal. Chem., 62, 650–652, 1990. 51. J.S. Fritz, M.C. Breadmore, E.F. Hilder and P.R. Haddad, Use of ionic polymers as stationary and pseudo-stationary phases in the separation of ions by capillary electrophoresis and capillary electrochromatography, J. Chromatogr. A, 942, 11–32, 2002. 52. Y. Takagai and S. Igarashi, Homogeneous liquid-liquid extraction and micellar electrokinetic chromatography using sweeping effect concentration system for determination of trace amounts of several polycyclic aromatic hydrocarbons, Anal. Bioanal. Chem., 373, 87–92, 2002. 53. E. Turiel, P. Fernández, C. Pérez-Conde and C. Cámara, Trace-level determination of triazines and several degradation products in environmental waters by disk solid-phase extraction and micellar electrokinetic chromatography, J. Chromatogr. A, 872, 299–307, 2000. 54. P. Zhang, G. Xu, J. Xiong, Y. Zheng, Q. Yang and F. Wei, Determination of arsenic species by capillary electrophoresis with large-volume field-amplified stacking injection, Electrophoresis, 22, 3567–3572, 2001. 55. Y. He and H.-K. Lee, Combination of solid-phase extraction and field-amplified concentration for trace analysis of organonitrogen pesticides by micellar electrokinetic chromatography, Electrophoresis, 18, 2036–2041, 1997. 56. T. Hirokawa, T. Ichihara, K. Ito and A.R. Timerbaev, Trace ion analysis of seawater by capillary electrophoresis: Determination of iodide using transient isotachophoresis preconcentration, Electrophoresis, 24, 2328–2334, 2003. 57. K. Otsuka, H. Hayashibara, S. Yamauchi, J.P. Quirino and S. Terabe, Highly sensitive micellar electrokinetic chromatographic analysis of dioxin-related compounds using on-line concentration, J. Chromatogr. A, 853, 413–420, 1999.

966

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

58. A. Macià, F. Borrull, M. Calull and C. Aguilar, Different sample stacking strategies to analyze some nonsteroidal anti-inflammatory drugs by micellar electrokinetic capillary chromatography in mineral waters, J. Chromatogr. A, 1117, 234–245, 2006. 59. S. Takeda, A. Omura, K. Chayama, H. Tsuji, K. Fukushi, M. Yamane, S.-I. Wakida, S. Tsubota and S. Terabe, Separation and on-line concentration of bisphenol A and alkylphenols by micellar electrokinetic chromatography with anionic surfactant, J. Chromatogr. A, 1014, 103–107, 2003. 60. J.P. Quirino, N. Inoue and S. Terabe, Reversed migration micellar electrokinetic chromatography with off-line and on-line concentration analysis of phenylurea herbicides, J. Chromatogr. A, 892, 187–194, 2000. 61. C.L. Silva, E.C. Lima and M.F.M. Tavares, Investigation of preconcentration strategies for the trace analysis of multiresidue pesticides in real samples by capillary electrophoresis, J. Chromatogr. A, 1014, 109–116, 2003. 62. K. Isoo and S. Terabe, Analysis of metal ions by sweeping via dynamic complexation and cationselective exhaustive injection in capillary electrophoresis, Anal. Chem., 75, 6789–6798, 2003. 63. O. Núñez, J.B. Kim, E. Moyano, M.T. Galcerán and S. Terabe, Analysis of the herbicides paraquat, diquat and dibenzoquat in drinking water by micellar electrokinetic chromatography using sweeping and cation selective exhaustive injection, J. Chromatogr. A, 961, 65–75, 2002. 64. J.P. Quirino, Y. Iwai, K. Otsuka and S. Terabe, Determination of environmentally relevant aromatic amines in the ppt levels by cation selective exhaustive injection-sweeping-micellar electrokinetic chromatography, Electrophoresis, 21, 2899–2903, 2000. 65. P. Hinsmann, L. Arce, A. Ríos and M. Valcárcel, Determination of pesticides in water by automation online solid-phase extraction-capillary electrophoresis, J. Chromatogr. A, 866, 137–146, 2000. 66. M.C. Breadmore, A.S. Palmer, M. Curran, M. Macka, N. Avdalovic and P.R. Haddad, On-column ion-exchange of inorganic anions in open tubular capillary electrochromatography with elution using transient-isotachophoresis gradients. 3. Implementation and method development, Anal. Chem., 74, 2112–2118, 2002. 67. J.F. Huertas-Pérez, M. del Olmo Iruela, A.M. García-Campaña, A. González-Casado and A. SánchezNavarro, Determination of the herbicide metribuzin and its major conversion products in soil by micellar electrokinetic chromatography, J. Chromatogr. A, 1102, 280–286, 2006. 68. A. Mainka and K. Bächmann, UV detection of derivatized carbonyl compounds in rain samples in capillary electrophoresis using sample stacking and a Z-shaped flow cell, J. Chromatogr. A, 767, 241–247, 1997. 69. J. Kuijt, C. García-Ruiz, G.J. Stroomberg, M.L. Marina, F. Ariese, U.A.Th. Brinkman and C. Gooijer, Laser-induced fluorescence detection at 266 nm in capillary electrophoresis. Polycyclic aromatic hydrocarbon metabolites in biota, J. Chromatogr. A, 907, 291–299, 2001. 70. S. Chen, Y. Xu, Y. Bi, W. Du and B.-F. Liu, Analysis of environmental pollutants metabolized from pesticides using capillary electrophoresis with multiphoton-excited fluorescence detection, Talanta, 70, 63–67, 2006. 71. E. Dabek-Zlotorzynska and W. Maruszak, Determination of dimethylamine and other lowmolecular mass amines using capillary electrophoresis with laser-induced fluorescence detection, J. Chromatogr. B, 714, 77–85, 1998. 72. M. Molina and M. Silva, In-capillary derivatization and analysis of amino acids, amino phosphonic acid-herbicides and biogenic amines by capillary electrophoresis with laser-induced fluorescence detection, Electrophoresis, 23, 2333–2340, 2002. 73. S.-J. Chen, M.-J. Chen and H.-T. Chang, Light-emitting diode-based indirect fluorescence detection for simultaneous determination of anions and cations in capillary electrophoresis, J. Chromatogr. A, 1017, 215–224, 2003. 74. S.Y. Chang and C.-H. Liao, Analysis of glyphosate, glufosinate and aminomethylphosphonic acid by capillary electrophoresis with indirect fluorescence detection, J. Chromatogr. A, 959, 309–315, 2002. 75. Y.-M. Liu and J.-K. Cheng, Separation of biogenic amines by micellar electrokinetic chromatography with on-line chemiluminescence detection, J. Chromatogr. A, 1003, 211–216, 2003. 76. F.R. Rocha, J.A.F. da Silva, C.L. do Lago, A. Fornaro and I.G.R. Gutz, Wet deposition and related atmospheric chemistry in the Sao Paulo metropolis, Brazil; Part 1. Major inorganic ions in rainwater

Separation Strategies for Environmental Analysis

77.

78.

79.

80.

81.

82.

83.

84. 85. 86.

87.

88.

89.

90.

91.

92. 93.

94. 95.

967

as evaluated by capillary electrophoresis with contactless conductivity detection, Atmos. Environ., 37, 105–115, 2003. A. Fornaro and I.G.R. Gutz, Wet deposition and related atmospheric chemistry in the Sao Paulo metropolis, Brazil; Part 2. Contribution of formic and acetic acids, Atmos. Environ., 37, 117–128, 2003. P. Kubáˇn, B. Karlberg, P. Kubáˇn and V. Kubáˇn, Application of a contactless conductometric detector for the simultaneous determination of small anions and cations by capillary electrophoresis with dual-opposite end injection, J. Chromatogr. A, 964, 227–241, 2002. M. Chicharro, A. Zapardiel, E. Bermejo and A. Sánchez, Simultaneous UV and electrochemical determination of the herbicide asulam in tap water samples by micellar electrokinetic capillary chromatography, Anal. Chim. Acta, 469, 243–252, 2002. A. Asthana, D. Bose, A. Durgbanshi, S.K. Sanghi and W.Th. Kok, Determination of aromatic amines in water samples by capillary electrophoresis with electrochemical and fluorescence detection, J. Chromatogr. A, 895, 197–203, 2000. M. van Bruijnsvoort, S.K. Sanghi, H. Poppe and W.Th. Kok, Determination of chlorophenols by micellar electrokinetic chromatography with electrochemical detection, J. Chromatogr. A, 757, 203–223, 1997. W. Ahrer, E. Scherwenk and W. Buchberger, Determination of drug residues in water by the combination of liquid chromatography or capillary electrophoresis with electrospray mass spectrometry, J. Chromatogr. A, 910, 69–78, 2001. C.A. Groom, A. Halasz, L. Paquet, S. Thiboutot, G. Ampleman and J. Hawari, Detection of nitroaromatic and cyclic nitramine compounds by cyclodextrin assisted capillary electrophoresis quadrupole ion trap mass spectrometry, J. Chromatogr. A, 1072, 73–82, 2005. T. Poiger, S.D. Richardson and G.L. Baughman, Identification of reactive dyes in spent dyebaths and wastewater by capillary-electrophoresis-mass spectrometry, J. Chromatogr. A, 886, 271–282, 2000. O. Núñez, E. Moyano and M.T. Galceran, Capillary electrophoresis-mass spectrometry for the analysis of quaternary ammonium herbicides, J. Chromatogr. A, 974, 243–255, 2002. Y. Iinuma and H. Herrmann, Method development for the analysis of particle phase substituted methoxy phenols and aromatic acids from biomass burning using capillary electrophoresis/electrospray ionization mass spectrometry (CE/ESI/MS), J. Chromatogr. A, 1018, 105–115, 2003. J. Riu, P. Eichhorn, J.A. Guerrero, Th.P. Knepper and D. Barceló, Determination of linear alkylbenzenesulfonates in wastewater treatment plants and coastal waters by automated solid-phase extraction followed by capillary electrophoresis-UV detection and confirmation by capillary electrophoresis-mass spectrometry, J. Chromatogr. A, 889, 221–229, 2000. M. Molina, S.K. Wiedmer, M. Jussila, M. Silva and M.L. Riekkola, Use of partial filling technique and reverse migrating micelles in the study of N-methylcarbamate pesticides by micellar electrokinetic chromatography-electrospray ionization mass spectrometry, J. Chromatogr. A, 24, 191–202, 2001. F. Menzinger, P. Schmitt-Kopplin, M. Frommberger, D. Freitag and A. Kettrup, Partial-filling micellar electrokinetic chromatography and nonaqueous capillary electrophoresis for the analysis of selected agrochemicals, Fresenius J. Anal. Chem., 371, 25–34, 2001. Y.-R. Chen, M.-C. Tseng, Y.-Z. Chang and G.-R. Her, A low-flow CE/electrospray ionization MS interface for capillary zone electrophoresis, large-volume sample stacking, and micellar electrokinetic chromatography, Anal. Chem., 75, 503–508, 2003. K. Isoo, K. Otsuka and S. Terabe, Application of sweeping to micellar electrokinetic chromatographyatmospheric pressure chemical ionization-mass spectrometric analysis of environmental pollutants, Electrophoresis, 22, 3426–3432, 2001. O.H.J. Szolar, Environmental and pharmaceutical analysis of dithiocarbamates, Anal. Chim. Acta, 582, 191–200, 2007. L. Gámiz-Gracia, A.M. García-Campaña, J.J. Soto-Chinchilla, J.F. Huertas-Pérez and A. GonzálezCasado, Analysis of pesticides by chemiluminescence detection in the liquid phase, Trends Anal. Chem., 24, 927–942, 2005. V. Andreu and Y. Picó, Determination of pesticides and their degradation products in soil: Critical review and comparison of methods, Trends Anal. Chem., 23, 772–789, 2004. J. Hernández-Borges, S. Frías-García, A. Cifuentes and M.A. Rodrígues-Delgado, Pesticide analysis by capillary electrophoresis, J. Sep. Sci., 27, 947–963, 2004.

968

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

96. Y. Picó, R. Rodriguez and J. Mañes, Capillary electrophoresis for the determination of pesticide residues, Trends Anal. Chem., 22, 133–151, 2003. 97. Ph. Schmitt, T. Poiger, R. Simon, D. Freitag, A. Kettrup and A.W. Garrison, Simultaneous determination of ionization constants and isoelectric points of 12 hydroxy-s-triazines by capillary zone electrophoresis and capillary isoelectric focusing, Anal. Chem., 69, 2559–2566, 1997. 98. M. Chicharro, A. Zapardiel, E. Bermejo and M. Moreno, Determination of 3-amino-1,2,4-triazole (amitrole) in environmental waters by capillary electrophoresis, Talanta, 59, 37–45, 2003. 99. R. Loos and R. Niessner, Analysis of atrazine, terbutylazine and their N-dealkylated chloro and hydroxyl metabolites by solid-phase extraction and gas chromatography mass spectrometry and capillary electrophoresis-ultraviolet detection, J. Chromatogr. A, 835, 217–229, 1999. 100. H. Stutz, K. Pittertschatscher and H. Malissa, Jr., Capillary zone electrophoresis determination of hydroxymetabolites of atrazine in potable water using solid-phase extraction with amberchrom resins, Mikrochim. Acta, 128, 107–117, 1998. 101. J. You, M. Kaljurand and J.A. Koropchak, Direct determination of glyphosate in environmental waters using capillary electrophoresis with electrospray condensation nucleation light scattering detection, Intern. J. Environ. Anal. Chem., 83, 797–806, 2003. 102. J. Hernández-Borges, F.J. García-Montelongo, A. Cifuentes and M.A. Rodríguez-Delgado, Determination of herbicides in mineral and stagnant waters at ng/L levels using capillary electrophoresis and UV detection combined with solid-phase extraction and sample stacking, J. Chromatogr. A, 1070, 171–177, 2005. 103. R.G. Wuilloud, M. Shah, S.S. Kannamkumarath and J.C. Altamirano, The potential of inductively coupled plasma-mass spectrometric detection for capillary electrophoretic analysis of pesticides, Electrophoresis, 26, 1598–1605, 2005. 104. X. Song and W.L. Budde, Determination of chlorinated acid herbicides and related compounds in water by capillary electrophoresis-electrospray negative ion mass spectrometry, J. Chromatogr. A, 829, 327–340, 1998. 105. R. Carabías-Martínez, E. Rodríguez-Gonzalo, J. Dominguez-Álvarez and J. Hernández-Méndez, Determination of triazine herbicides in natural waters by solid-phase extraction and non-aqueous capillary zone electrophoresis, J. Chromatogr. A, 869, 451–461, 2000. 106. N.V. Komarova and L.A. Kartsova, Factors responsible for the electrophoretic behavior of carboxylic acid and triazine derivatives under conditions of capillary zone electrophoresis and micellar electrokinetic chromatography, J. Anal. Chem., 59, 662–668, 2004. 107. M. Chicharro, A. Zapardiel, E. Bermejo and A. Sanchez, Multiresidue analysis of pesticides in environmental waters by capillary electrophoresis using simultaneous UV and electrochemical detection, Electroanalysis, 16, 311–318, 2004. 108. R. Carabías-Martínez, E. Rodríguez-Gonzalo, P. Revilla-Ruiz and J. Domínguez-Alvarez, Solid-phase extraction and sample stacking-micellar electrokinetic capillary chromatography for the determination of multiresidue of herbicides and metabolites, J. Chromatogr. A, 990, 291–302, 2003. 109. C.W. Henry III, S.A. Shamsi and I.M. Warner, Separation of natural pyrethrum extracts using micellar electrokinetic chromatography, J. Chromatogr. A, 863, 89–103, 1999. 110. C.-E. Lin, Y.-C. Liu, T.-Y. Yang, T.-Z. Wang and C.-C. Yang, On-line concentration of s-triazine herbicides in micellar electrokinetic chromatography using a cationic surfactant, J. Chromatogr. A, 916, 239–245, 2001. 111. Y.S. Wu, H.K. Lee and S.F.Y. Li, A fluorescence detection scheme for capillary electrophoresis of N-methylcarbamates with on-column thermal decomposition and derivatization, Anal. Chem., 72, 1441–1447, 2000. 112. A. Karcher and Z. El Rassi, Electrically driven microseparation methods for pesticides and metabolites: IV. Effects of the nature of fluorescent labels on the enantioseparation of pesticides and their degradation products by capillary zone electrophoresis with UV and laser-induced fluorescence detection, Electrophoresis, 21, 2043–2050, 2000. 113. O. Núñez, E. Moyano, L. Puignou and M.T. Galcerán, Sample stacking with matrix removal for the determination of paraquat, diquat and difenzoquat in water by capillary electrophoresis, J. Chromatogr. A, 912, 353–361, 2001. 114. G. Dinelli, A. Vicari and A. Bonetti, Separation of sulfonylurea metabolites in water by capillary electrophoresis, J. Chromatogr. A, 700, 195–200, 1995.

Separation Strategies for Environmental Analysis

969

115. C. Klein, R.J. Schneider, M.T. Meyer and D.S. Aga, Enantiomeric separation of metolachlor and its metabolites using LC-MS and CZE, Chemosphere, 62, 1591–1599, 2006. 116. J. Hernández-Borges, F.J. García-Montelongo, A. Cifuentes and M.A. Rodríguez-Delgado, Analysis of triazolopyrimidine herbicides in soils using field-enhanced sample injection-coelectroosmotic capillary electrophoresis combined with solid-phase extraction, J. Chromatogr. A, 1100, 236–242, 2005. 117. N.V. Komarova and L.A. Kartsova, Determination of herbicides of the chlorophenoxycarboxylic acid type in natural and drinking water by capillary zone electrophoresis, Russian J. Appl. Chem., 76, 238–243, 2003. 118. M. Corbera, M. Hidalgo, V. Salvado and P.P. Wieczorek, Determination of glyphosate and aminomethylphosphonic acid in natural water using capillary electrophoresis combined with enrichment step, Anal. Chim. Acta, 540, 3–7, 2005. 119. H. Prosen, M. Guˇcek and L. Zupanˇciˇc-Kralj, Optimization of liquid chromatography and micellar electrokinetic chromatography for the determination of atrazine and its first degradation products in humic waters without preparation, Chromatographia, 60(Suppl. 60), S107–S112, 2004. 120. D.T. Kubilius and R.J. Bushway, Determination of hexazinone and its metabolites in groundwater by capillary electrophoresis, J. Chromatogr. A, 793, 349–355, 1998. 121. M.B. Barroso, L.N. Honda and G. Morovjan, Multiresidue analysis of phenylurea herbicides in environmental and drinking water by capillary electrophoresis, J. High Resol. Chromatogr., 22, 171–176, 1999. 122. J.L. Jarman, W.J. Jones, L.A. Howell and A.W. Garrison, Application of capillary electrophoresis to study the enantioselective transformation of five chiral pesticides in aerobic soil slurries, J. Agric. Food Chem., 53, 6175–6182, 2005. 123. R. Rodríguez, Y. Picó, G. Font and J. Mañes, Analysis of post-harvest fungicides by micellar electrokinetic chromatography, J. Chromatogr. A, 924, 387–396, 2001. 124. M. Chicharro, A. Zapardiel, E. Bermejo, J.A. Perez and A. Moreno, Multiresidue analysis of S-triazines herbicides in environmental waters by micellar electrokinetic capillary chromatography, J. Liq. Chromatogr. Rel. Technol., 24, 461–478, 2001. 125. X. Xu and R.J. Hurtubise, Separation of polycyclic aromatic hydrocarbon metabolites by gammacyclodextrin-modified micellar electrokinetic chromatography, J. Liq. Chromatogr. Rel. Technol., 23, 1657–1670, 2000. 126. J. Li and J.S. Fritz, Nonaqueous media for separation of nonionic organic compounds by capillary electrophoresis, Electrophoresis, 20, 84–91, 1999. 127. G. Kavran and F.B. Erim, Separation of polycyclic aromatic hydrocarbons with sodium dodecylbenzenesulfonate in electrokinetic chromatography, J. Chromatogr. A, 949, 301–305, 2002. 128. G.K. Belin, E.B. Erim and F. Gulacar, Capillary electrokinetic separation of polycyclic aromatic hydrocarbons using cetylpyridinium bromide, Polycyclic Arom. Compds., 24, 343–352, 2004. 129. H. Ren, X. Li, M. Qi, C. Stathakis and N.J. Dovichi, Capillary electrophoretic separation of polynuclear aromatic hydrocarbons using sodium cholate in mixed aqueous-organic buffers, J. Chromatogr. A, 817, 307–311, 1998. 130. J.H.T. Luong, The combined effect of acetonitrile and urea on the separation of polycyclic aromatic hydrocarbons using dioctyl sulfosuccinate in electrokinetic chromatography, Electrophoresis, 19, 1461–1467, 1998. 131. Y. Shi and J.S. Fritz, Capillary zone electrophoresis of neutral organic molecules in organic–aqueous solution, J. High Resol. Chromatogr., 17, 713–718, 1994. 132. J.L. Miller, M.G. Khaledi and D. Shea, Separation of hydrophobic solutes by nonaqueous capillary electrophoresis through dipolar and charge-transfer interactions with pyrylium salts, J. Microcolumn Sep., 10, 681–685, 1998. 133. A.-L. Nguyen and J.H.T. Luong, Separation and determination of polycyclic aromatic hydrocarbons by solid-phase microextraction/cyclodextrin-modified capillary electrophoresis, Anal. Chem., 69, 1726–1731, 1997. 134. C.P. Palmer, M.Y. Khaledi and H.M. McNair, A monomolecular pseudostationary phase for micellar electrokinetic capillary chromatography, J. High Resol. Chromatogr., 15, 756–762, 1992. 135. D. Norton and S.A. Shamsi, Separation of methylated isomers of benzo[a]pyrene using micellar electrokinetic chromatography, Anal. Chim. Acta, 496, 165–176, 2003.

970

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

136. C. Akbay, S.A. Shamsi and I.M. Warner, Separation of monomethyl-benz[a]antracene isomers using cyclodextrin-modified electrokinetic chromatography, J. Chromatogr. A, 910, 147–155, 2001. 137. C. Fujimoto, Y. Fujise and S. Kawaguchi, Macromolecular surfactant as a pseudo-stationary phase in micellar electrokinetic capillary chromatography, J. Chromatogr. A, 871, 415–425, 2000. 138. D.S. Peterson and C.P. Palmer, An anionic siloxane polymer as a pseudostationary phase for electrokinetic chromatography, Electrophoresis, 21, 3174–3180, 2000. 139. N. Tanaka, T. Fukutome, K. Hosoya, K. Kimata and T. Anaki, Polymer-supported pseudo-stationary phase for electrokinetic chromatography. Electrokinetic chromatography in a full range of methanol– water mixtures with alkylated starburst dendrimers, J. Chromatogr. A, 716, 57–67, 1995. 140. J.M. Treubig Jr. and P.R. Brown, Novel approach to the analysis and use of fullerenes in capillary electrophoresis, J. Chromatogr. A, 873, 257–267, 2000. 141. B. Göttlicher and K. Bachmann, Investigation of the separation efficiency of hydrophobic compounds in suspension electrokinetic chromatography, J. Chromatogr. A, 768, 320–324, 1997. 142. I. Benito, J.M. Saz, M.L. Marina, J. Jiménez-Barbero, M.J. González and J.C. Díez-Masa, Micellar electrokinetic capillary chromatographic separation of polychlorinated biphenyl congeners, J. Chromatogr. A, 778, 77–85, 1997. 143. M.L. Marina, I. Benito, J.C. Díez-Masa and M.J. González, Separation of chiral polychlorinated biphenyls by micellar electrokinetic chromatography using β- and γ-cyclodextrin mixtures in the separation buffer, J. Chromatogr. A, 752, 265–270, 1996. 144. W.C. Lin, F.C. Chang and C.H. Kuei, Separation of atropisomeric polychlorinated biphenyls by cyclodextrin modified micellar electrokinetic chromatography, J. Microcolumn Sep., 11, 231–238, 1999. 145. A.L. Crego, M.A. García and M.L. Marina, Enantiomeric separation of chiral polychlorinated biphenyls by micellar electrokinetic chromatography using mixtures of bile salts and sodium dodecyl sulphate with and without gamma-cyclodextrin in the separation buffer, J. Microcolumn. Sep., 12, 33–40, 2000. 146. C. García-Ruiz, A.L. Crego and M.L. Marina, Comparison of charged cyclodextrin derivatives for the chiral separation of atropisomeric polychlorinated biphenyls by capillary electrophoresis, Electrophoresis, 24, 2657–2664, 2003. 147. S.H. Edwards and S.A. Shamsi, Micellar electrokinetic chromatography of polychlorinated biphenyl congeners using a polymeric surfactant as the pseudostationary phase, J. Chromatogr. A, 903, 227–236, 2000. 148. S.H. Edwards and S.A. Shamsi, Chiral separation of polychlorinated biphenyls using a combination of hydroxypropyl-gamma-cyclodextrin and a polymeric chiral surfactant, Electrophoresis, 23, 1320–1327, 2002. 149. A. Hilmi, J.H.T. Luong and A.-L. Nguyen, Applicability of micellar electrokinetic chromatography to kinetic studies of photocatalytic oxidation of dibenzo-p-dioxin, Chemosphere, 36, 3113–3117, 1998. 150. J. Grainger, P.C. McClure, A. Liu, et al., Isomer identification of chlorinated dibenzo-p-dioxins by orthogonal spectroscopic and chromatographic techniques, Chemosphere, 32, 13–23, 1996. 151. S. Terabe, Y. Miyashita, O. Shibata, E.R. Barnhart, L.R. Alexander, D.G. Patterson, B.L. Karger, K. Hosoya and N. Tanaka, Separation of highly hydrophobic compounds by cyclodextrin-modified micellar electrokinetic chromatography, J. Chromatogr. A, 516, 23–31, 1990. 152. J.T. Koch, B. Beam, K.S. Phillips and J.F. Wheeler, Hydrophobic interaction electrokinetic chromatography for the separation of polycyclic aromatic hydrocarbons using nonaqueous matrices, J. Chromatogr. A, 914, 223–231, 2001. 153. C.-E. Lin, C.-C. Hsueh, W.-C. Lin and C.-C. Chang, Migration behavior and separation of trichlorophenols by capillary zone electrophoresis, J. Chromatogr. A, 746, 295–299, 1996. 154. I. Canals, E. Bosch and M. Roses, Prediction of the separation of phenols by capillary zone electrophoresis, Anal. Chim. Acta, 458, 355–366, 2002. 155. S. Takeda, Y. Tanaka, M. Yamane, Z. Siroma, S.-I. Wakida, K. Otsuka and S. Terabe, Ionization of dichlorophenols for their analysis by capillary electrophoresis-mass spectrometry, J. Chromatogr. A, 924, 415–420, 2001. 156. O. Jáuregui, E. Moyano and M.T. Galceran, Capillary electrophoresis-electrospray ion-trap mass spectrometry for the separation of chlorophenols, J. Chromatogr. A, 896, 125–133, 2000. 157. O. Jáuregui, L. Puignou and M.T. Galceran, New carrier electrolytes for the separation of chlorophenols by capillary electrophoresis, Electrophoresis, 21, 611–618, 2000.

Separation Strategies for Environmental Analysis

971

158. S. Morales and R. Cela, Capillary electrophoresis and sample stacking in non-aqueous media for the analysis of priority pollutant phenols, J. Chromatogr. A, 846, 401–411, 1999. 159. S.P. Porras, R. Kuldvee, S. Palonen and M.-L. Riekkola, Capillary electrophoresis of methylsubstituted phenols in acetonitrile, J. Chromatogr. A, 990, 35–44, 2003. 160. S.P. Porras, R. Kuldvee, M. Jussila, S. Palonen and M.-L. Riekkola, Capillary zone electrophoresis of phenol and methylphenols at high pH in methanol, J. Sep. Sci., 26, 857–862, 2003. 161. J. Kronholm, P. Revilla-Ruiz, S.P. Porras, K. Hartonen, R. Carabías-Martínez and M.-L. Riekkola, Comparison of gas chromatography-mass spectrometry and capillary electrophoresis in analysis of phenolic compounds extracted from solid matrices with pressurized hot water, J. Chromatogr. A, 1022, 9, 2004. 162. S. Fu, F. Li, S.G. Chu and X.B. Xu, The determination of chlorophenols in waste water by capillary zone electrophoresis with an organic modifier, Chromatographia, 56, 69–72, 2002. 163. X. Guo, Z. Wang and S. Zhou, The separation and determination of nitrophenol isomers by highperformance capillary zone electrophoresis, Talanta, 64, 135–139, 2004. 164. S.S. Kannamkumarath, R.G. Wuilloud, S. Jayasinghe and J.A. Caruso, Fast speciation analysis of iodophenol compounds in river waters by capillary electrophoresis-inductively coupled plasmamass spectrometry with off-line solid-phase microextraction, Electrophoresis, 25, 1843–1851, 2004. 165. W. Wall and Z. El Rassi, Electrically driven microseparation methods for pesticides and metabolites. Part VII: Capillary electrophoresis and electrochromatography of underivatized phenol pesticidic metabolites. Preconcentration and laser-induced fluorescence detection of dilute samples, J. Sep. Sci., 25, 1231–1244, 2002. 166. A. Zemann and D. Volgger, Separation of priority pollutant phenols with coelectroosmotic capillary electrophoresis, Anal. Chem., 69, 3243–3250, 1997. 167. S.M. Masselter and A.J. Zemann, Influence of organic solvents in coelectroosmotic capillary electrophoresis of phenols, Anal. Chem., 67, 1047–1053, 1995. 168. M.R.N. Monton, J.P. Quirino, K. Otsuka and S. Terabe, Separation and on-line preconcentration by sweeping of charged analytes in electrokinetic chromatography with nonionic micelles, J. Chromatogr. A, 939, 99–108, 2001. 169. M.R.N. Monton, K. Otsuka and S. Terabe, On-line sample preconcentration in micellar electrokinetic chromatography by sweeping with anionic-zwitterionic mixed micelles, J. Chromatogr. A, 985, 435–445, 2003. 170. G. Jeevan, M. Bhaskar, R. Chandrasekar and G. Radhakrishnan, Separation of harmful chlorophenols by cyclodextrin-assisted capillary electrokinetic chromatography, J. Sep. Sci., 25, 1143–1146, 2002. 171. P. Kuban, M. Berg, C. García and B. Karlberg, On-line flow sample stacking in a flow injection analysis-capillary electrophoresis system: 2000-fold enhancement of detection sensitivity for priority phenol pollutants, J. Chromatogr. A, 912, 163–170, 2001. 172. L. Gaillon, S. Cozette, J. Lelievre and R. Gaboriaud, New pseudo-stationary phases for electrokinetic capillary chromatography. Complexes between bovine serum albumin and sodium dodecyl sulfate, J. Chromatogr. A, 876, 169–182, 2000. 173. T. Zhao, X.B. Hu, J.K. Cheng and X.R. Lu, P-sulfonic calix[4]arene as running buffer additive in electrokinetic chromatography, J. Liq. Chromatogr. Rel. Technol., 21, 3111–3124, 1998. 174. A.L. Gray and J.T. Hsu, Novel sulfonic acid-modified Starburst dendrimer used as a pseudostationary phase in electrokinetic chromatography, J. Chromatogr. A, 824, 119–124, 1998. 175. B. Maichel, B. Potocek, B. Gaš and E. Kenndler, Capillary electrokinetic chromatography with polyethyleneimine as replaceable cationic pseudostationary phase. Influence of methanol and acetonitrile on separation selectivity, J. Chromatogr. A, 853, 121–129, 1999. 176. A.V. Pirogov and O.A. Shpigun, Application of water-soluble polymers as modifiers in electrophoretic analysis of phenols, Electrophoresis, 24, 2099–2105, 2003. 177. T. Nakamura, A. Ohki, M. Mishiro, O. Tsuyashima and S. Maeda, Aggregate of amphiphilic block copolymer as a pseudo-stationary phase in capillary electrophoresis, Anal. Sci., 15, 879–883, 1999. 178. A.V. Pirogov, A.V. Shpak and O.A. Shpigun, Application of polyelectrolyte complexes as novel pseudostationary phases in MEKC, Anal. Bioanal. Chem., 375, 1199–1203, 2003.

972

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

179. J.T. Hautala, S.K. Wiedmer and M.-L. Riekkola, Anionic liposomes in capillary electrophoresis: Effect of calcium on 1-palmitoyl-2-oleyl-sn-glycero-3-phosphatidylcholine/phosphatidylserine-coating in silica capillaries, Anal. Bioanal. Chem., 378, 1769–1776, 2004. 180. A.L. Crego and M.L. Marina, Capillary zone electrophoresis versus micellar electrokinetic chromatography in the separation of phenols of environmental interest, J. Liq. Chrom. Rel. Technol., 20, 1–20, 1997. 181. M. Czaplicka, Sources and transformations of chlorophenols in the natural environment, Sci. Total Environ., 322, 21–39, 2004. 182. S. Kar and P.K. Dasgupta, Measurement of phenols on a loop-supported liquid film by micellar electrokinetic chromatography and direct UV detection, J. Chromatogr. A, 379, 379–387, 1996. 183. W.H. Matchett and W.C. Brumley, Preconcentration of aliphatic amines from water determined by capillary electrophoresis with indirect UV detection, J. Liq. Chromatogr. Rel. Technol., 20, 79–100, 1997. 184. A. Fekete, M. Frommberger, G. Ping, M.R. Lahaniatis, J. Lintelman, J. Fekete, I. Gebefugi, A.K. Malik, A. Kettrup and P. Schmitt-Kopplin, Development of a capillary electrophoretic method for the analysis of low-molecular-weight amines from metal working fluid aerosols and ambient air, Electrophoresis, 27, 1237–1247, 2006. 185. W. Maruszak, Analysis of aliphatic amines by micellar electrokinetic chromatography, J. High Resol. Chromatogr., 22, 126–128, 1999. 186. X. Liu, L.-X. Yang and Y.-T. Lu, Determination of biogenic amines by 3-(2-furoyl)quinoline2-carboxaldehyde and capillary electrophoresis with laser-induced fluorescence detection, J. Chromatogr. A, 998, 213–219, 2003. 187. K.B. Male and J.H.T. Luong, Derivatization, stabilization and detection of biogenic amines by cyclodextrin-modified capillary electrophoresis-laser-induced fluorescence detection, J. Chromatogr. A, 926, 309–317, 2001. 188. A. Bazzanella, H. Mörbel, K. Bächmann, R. Milbradt, V. Böhmer and W. Vogt, Highly efficient separation of amines by electrokinetic chromatography using resorcarene-octacarboxylic acids as pseudostationary phases, J. Chromatogr. A, 792, 143–149, 1997. 189. B.A. Musial, M.N. Martin and N.D. Danielson, Effect of an anionic polymer on the separation of cationic molecules by capillary electrophoresis with conductivity detection, J. Sep. Sci., 25, 311–318, 2002. 190. H.M. Pinheiro, E. Toraud and O. Thomas, Aromatic amines from azo dye reduction: Status review with emphasis on direct UV spectrophotometric detection in textile industry wastewaters, Dyes Pigments, 61, 121–139, 2004. 191. S. Oguri, Electromigration methods for amino acids, biogenic amines and aromatic amines, J. Chromatogr. B, 747, 1–19, 2000. 192. H. Kataoka, Methods for the determination of mutagenic heterocyclic amines and their applications in environmental analysis, J. Chromatogr. A, 774, 121–142, 1997. 193. S.-A. Leung and A.J. de Mello, Electrophoretic analysis of amines using reversed-phase, reversedpolarity, head-column field-amplified sample stacking and laser-induced fluorescence detection, J. Chromatogr. A, 979, 171–178, 2002. 194. G. Taibi, M.R. Schiavo, P.C. Rindina, R. Muratore and C.M.A. Nicotra, Micellar electrokinetic chromatography of polyamines and monoacetylpolyamines, J. Chromatogr. A, 921, 323–329, 2001. 195. C.M. Knapp and J.J. Breen, Effects of tetraalkylammonium salts on the micellar electrokinetic chromatography of aniline and substituted anilines, J. Chromatogr. A, 799, 289–299, 1998. 196. C.-E. Lin, C.-C. Chen, H.-W. Chen, H.-C. Huang, C.-H. Lin and Y.-C. Liu, Optimization of separation and migration behavior of chloropyridines in micellar electrokinetic chromatography, J. Chromatogr. A, 910, 165–171, 2001. 197. S.D. Mendonsa and R.J. Hurtubise, Capillary electrophoresis (CE) methods for the separation of carcinogenic heterocyclic aromatic amines, J. Liq. Chrom. Rel. Technol., 22, 1027–1040, 1999. 198. P. Zakaria, M. Macka and P.R. Haddad, Mixed-mode electrokinetic chromatography of aromatic bases with two pseudo-stationary phases and pH control, J. Chromatogr. A, 997, 207–218, 2003. 199. D.W. Armstrong, L.W. Chang and S.S.C. Chang, Mechanism of capillary electrophoresis enantioseparations using a combination of an achiral crown ether plus cyclodextrins, J. Chromatogr. A, 793, 115–134, 1998.

Separation Strategies for Environmental Analysis

973

200. S.-L. Zhao, T.-Y. Dai, Z. Liu, F.-S. Wei, H.-F. Zou and X.-B. Xu, Determination of lower aliphatic carbonyl compounds in stack gas as their 2,4-dinitrophenylhydrazones by micellar electrokinetic chromatography, Chemosphere, 35, 2131–2136, 1997. 201. S. Takeda, S. Wakida, M. Yamane and K. Higashi, Analysis of lower aliphatic-aldehydes in water by micellar electrokinetic chromatography with derivatization to 2,4-dinitrophenylhydrazones, Electrophoresis, 15, 1332–1334, 1994. 202. E.A. Pereira, A.A. Cardoso and M.F.M. Tavares, Determination of low-aliphatic aldehydes indoors by micellar electrokinetic chromatography using sample dissolution manipulation for signal enhancement, Electrophoresis, 24, 700–706, 2003. 203. E.A. Pereira, E. Carrilho and M.F.M. Tavares, Laser-induced fluorescence and UV detection of derivatized aldehydes in air sample using capillary electrophoresis, J. Chromatogr. A, 979, 409–416, 2002. 204. E.A. Pereira, M.O.O. Rezende and M.F.M. Tavares, Analysis of low molecular weight aldehydes in air samples by capillary electrophoresis after derivatization with 4-hydrazinobenzoic acid, J. Sep. Sci., 27, 28–32, 2004. 205. K. Fukushi, S. Takeda, K. Chayama and S.-I. Wakida, Application of capillary electrophoresis to the analysis of inorganic ions in environmental samples, J. Chromatogr. A, 834, 349–362, 1999. 206. A.R. Timerbaev and K. Fukushi, Analysis of seawater and different highly saline natural waters by capillary zone electrophoresis, Marine Chem., 82, 221–238, 2003. 207. N. Shakulashvili, T. Faller and H. Engelhardt, Simultaneous determination of alkali, alkaline earth and transition metal ions by capillary electrophoresis with indirect UV detection, J. Chromatogr. A, 895, 205–212, 2000. 208. K. Fukushi, H. Ito, K. Kimura, K. Yokota, K. Saito, K. Chayama, S. Takeda and S.-I. Wakida, Determination of ammonium in river water and sewage samples by capillary zone electrophoresis with direct UV detection, J. Chromatogr. A, 1106, 61–66, 2006. 209. E. Dabek-Zlotorzynska, M. Kelly, H. Chen and C.L. Chakrabarti, Application of capillary electrophoresis combined with a modified BCR sequential extraction for estimating of distribution of selected trace metals in PM2.5 fractions of urban air particulate matter, Chemosphere, 58, 1365–1376, 2005. 210. E. Dabek-Zlotorzynska and R. Aranda-Rodriguez, Development and validation of capillary electrophoresis for the determination of selected metal ions in airborne particulate matter after sequential extraction, Anal. Bioanal. Chem., 372, 467–472, 2002. 211. Y.-S. Fung and H.-S. Tung, Capillary electrophoresis for trace metal ion analysis on environmental studies, Electrophoresis, 20, 1832–1841, 1999. 212. A. Gáspár and E. Dudás, Application of internal universal calibration for determination of fully dissociated species in capillary electrophoresis using indirect UV detection and electrokinetic injection, J. Chromatogr. A, 1110, 254–260, 2006. 213. E. Yang and Z. Zhang, Simultaneous and fast detection of anions in snow using short tube by capillary zone electrophoresis, Intern. J. Environ. Anal. Chem., 82, 353–360, 2002. 214. L. Wójcik, K. Korczak, B. Szostek and M. Trojanowicz, Separation and determination of perfluorinated carboxylic acids using capillary zone electrophoresis with indirect photometric detection, J. Chromatogr. A, 1128, 290–299, 2006. 215. E. Dabek-Zlotorzynska, M. Piechowski, M. McGrath and E.P.C. Lai, Determination of lowmolecular-mass carboxylic acids in atmospheric aerosol and vehicle emission samples by capillary electrophoresis, J. Chromatogr. A, 910, 331–345, 2001. 216. E. Dabek-Zlotorzynska, R. Aranda-Rodrigues and L. Graham, Capillary electrophoresis determinative and GC-MS confirmatory method for water-soluble organic acids in airborne particulate matter and vehicle emission, J. Sep. Sci., 28, 1520–1528, 2005. 217. A. Mainka, P. Ebert, M. Kibler, T. Prokop, B. Tenberken and K. Bächmann, Development of new methods for the analysis of carboxylic acids and carbonyl compounds in size classified raindrops by CE for application in modeling atmospheric processes, Chromatographia, 45, 158–162, 1997. 218. A.G. Diress and C.A. Lucy, Study of the selectivity of inorganic anions in hydro-organic solvents using indirect capillary electrophoresis, J. Chromatogr. A, 1085, 155–163, 2005. 219. J.E. Melanson, N.E. Baryla and C.A. Lucy, Dynamic capillary coatings for electroosmotic flow control in capillary electrophoresis, Trends Anal. Chem., 20, 365–374, 2001.

974

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

220. V. Martínez, N. García, I. Antigüedad, R.M. Alonso and R.M. Jiménez, Capillary electrophoresis as a useful tool for the analysis of chemical tracers applied to hydrological systems, J. Chromatogr. A, 1032, 237–242, 2004. 221. K. Ito, T. Ichibara, H. Zhuo, K. Kumamoto, A.R. Timerbaev and T. Hirokawa, Determination of trace iodide in seawater by capillary electrophoresis following transient isotachophoretic preconcentration. Comparison with ion chromatography, Anal. Chim. Acta, 497, 67–74, 2003. 222. Z. Huang, K. Ito, A.R. Timerbaev and T. Hirokawa, Speciation studies by capillary electrophoresis— Simultaneous determination of iodide and iodate in seawater, Anal. Bioanal. Chem., 378, 1836–1841, 2004. 223. S. Rovio, M. Mäntynen and H. Sirén, Determination of bromide and potassium in saline groundwaters by capillary electrophoresis without prior dilution, Appl. Geochem., 19, 1331–1337, 2004. 224. T. Takayanagi, M. Ishida, J. Mbuna, R. Driouich and S. Motomizu, Determination of bromate ion in drinking water by capillary zone electrophoresis with direct photometric detection, J. Chromatogr. A, 1128, 298–302, 2006. 225. C. Tu and H.K. Lee, Determination of nitrate in seawater by capillary zone electrophoresis with chloride-induced sample self-stacking, J. Chromatogr. A, 966, 205–212, 2002. 226. K. Fukushi, Y. Nakayama and J.-I. Tsujimoto, Highly sensitive capillary zone electrophoresis with artificial seawater as the background electrolyte and transient isotachophoresis as the on-line concentration procedure for simultaneous determination of nitrite and nitrate in seawater, J. Chromatogr. A, 1005, 197–205, 2003. 227. M.A. Woodland and C.A. Lucy, Altering the selectivity of inorganic anion separations using electrostatic capillary electrophoresis, Analyst, 126, 28–32, 2001. 228. Y. Deng, X. Fan, A. Delgado, C. Nolan, K. Furton, Y. Zuo and R.D. Jones, Separation and determination of aromatic acids in natural water with preconcentration by capillary zone electrophoresis, J. Chromatogr. A, 817, 145–152, 1998. 229. P. Kubáˇn, P. Kubáˇn and V. Kubáˇn, Speciation of chromium (III) and chromium (VI) by capillary electrophoresis with contactless conductometric detection and dual opposite end injection, Electrophoresis, 24, 1397–1403, 2003. 230. W.-P. Yang, Z.-J. Zhang and W. Deng, Simultaneous, sensitive and selective online chemiluminescence determination of Cr(III) and Cr(VI) by capillary electrophoresis, Anal. Chim. Acta, 485, 169–177, 2003. 231. F. Han, J.L. Fasching and P.R. Brown, Speciation of organotin compounds by capillary electrophoresis using indirect ultraviolet absorbance detection, J. Chromatogr. A, 669, 103–112, 1995. 232. C.L. Ng, H.K. Lee and S.F.Y Li, Determination of organolead and organoselenium compounds by micellar electrokinetic chromatography, J. Chromatogr. A, 652, 547–553, 1993. 233. G. Forte, M. D’Amato and S. Caroli, Capillary electrophoresis speciation analysis of various arsenical compounds, Microchem. J., 79, 15–19, 2005. 234. K.F. Akter, Z. Chen, L. Smith, D. Davey and R. Naidu, Speciation of arsenic in ground water samples: A comparative study of CE-UV, HG-AAS and LC-ICP-MS, Talanta, 68, 406–415, 2005. 235. Z.L. Chen, J.-M. Lin and R. Naidu, Separation of arsenic species by capillary electrophoresis with sample-stacking techniques, Anal. Bioanal. Chem., 375, 679–684, 2003. 236. Z. Chen and R. Naidu, Separation of sulfur species in water by co-electroosmotic capillary electrophoresis with direct and indirect UV detection, Intern. J. Environ. Anal. Chem., 83, 749–759, 2003. 237. P. Kubáˇn, P. Kubáˇn and V. Kubáˇn, Rapid speciation of Se(IV) and Se(VI) by flow injection-capillary electrophoresis system with contactless conductivity detection, Anal. Bioanal. Chem., 378, 378–382, 2004. 238. M. Takaya, Development of an analytical method for beryllium in airborne dust by micellar electrokinetic chromatography, J. Chromatogr. A, 850, 363–368, 1999. 239. M. Aguillar, A. Farran and V. Martí, Capillary electrophoretic determination of cyanide leaching solutions from automobile catalytic converters, J. Chromatogr. A, 778, 397–402, 1997. 240. J.M. Robert and C.D. Spinks, Separation of metallated petroporphyrin models using micellar electrokinetic capillary chromatography, J. Liq. Chrom. Rel. Technol., 20, 2979–2995, 1997.

Separation Strategies for Environmental Analysis

975

241. C.A. Groom, A. Halasz, L. Paquet, P. D’Cruz and J. Hawari, Cuclodextrin-assisted capillary electrophoresis for determination of the cyclic nitramine explosives RDX, HMS and CL-20. Comparison with high-performance liquid chromatography, J. Chromatogr. A, 999, 17–22, 2003. 242. S.A. Oehrle, Analysis of nitramine and nitroaromatic explosives by capillary electrophoresis, J. Chromatogr. A, 745, 233–237, 1996. 243. K.G. Hopper, H. LeChair and B.R. McCord, A novel method for analysis of explosives residue by simultaneous detection of anions and cations via capillary electrophoresis, Talanta, 67, 304–312, 2005. 244. C.L. Copper and G.E. Collins, Separation of thiol and cyanide hydrolysis products of chemical warfare agents by capillary electrophoresis, Electrophoresis, 25, 897–902, 2004. 245. E.W.J. Hooijschuur, A.E. Kientz and U.A.Th. Brinkman, Application of microcolumn liquid chromatography and capillary electrophoresis with flame photometric detection for the screening of degradation products of chemical warfare agents in water and soil, J. Chromatogr. A, 928, 187–199, 2001. 246. M. Lagarrigue, A. Bossée, A. Bégos, A. Varenne, P. Gareil and B. Bellier, Separation and identification of isomeric acidic degradation products of organophosphorus chemical warfare agents by capillary electrophoresis-ion trap mass spectrometry, J. Chromatogr. A, 1137, 110–118, 2006. 247. T.E. Rosso and P.C. Bossle, Capillary ion electrophoresis screening of nerve agent degradation products in environmental samples using conductivity detection, J. Chromatogr. A, 814, 125–134, 1998. 248. A.-E.F. Nassar, S.V. Lucas and L.D. Hoffland, Determination of chemical warfare agent degradation products at low-part-per-billion levels in aqueous samples and sub-part-per-million levels in soils using capillary electrophoresis, Anal. Chem., 71, 1285–1292, 1999. 249. M.-H. Chen and W.-H. Ding, Separation and migration of positional isomers and structural naphthalenesulfonate isomers by cyclodextrin-mediated capillary electrophoresis, J. Chromatogr. A, 1033, 167–172, 2004. 250. M.J. Cugat, F. Borrull and M. Calull, Separation of aromatic sulfonate compounds by coelectroosmotic micellar electrokinetic chromatography, Chromatographia, 50, 229–234, 1999. 251. M.J. Cugat, F. Borrull and M. Calull, Comparative study of capillary zone electrophoresis and micellar electrokinetic chromatography for the separation of twelve aromatic sulphonate compounds, Chromatographia, 46, 204–208, 1997. 252. M.J. Cugat, F. Borrull and M. Calull, An overview of electrophoretic methods for the determination of benzene- and naphthalenesulfonates in water samples, Trends Anal. Chem., 20, 487–499, 2001. 253. R. Loos and R. Niessner, Analysis of aromatic sulfonates in water by solid-phase extraction and capillary electrophoresis, J. Chromatogr. A, 822, 291–303, 1998. 254. R. Loos, J. Riu, M.C. Alonso and D. Barceló, Analysis of polar hydrophilic aromatic sulfonates in waste water treatment plants by CE/MS and LC/MS, J. Mass Spectrom., 35, 1197–1206, 2000. 255. S. Aubert and J.-P. Schwitzguébel, Capillary electrophoretic separation of sulphonated anthraquinones in a variety of matrices, Chromatographia, 56, 693–697, 2002. 256. L. Farry, S.A. Oxspring, W.F. Smyth and R. Marchant, A study of the effects of injection mode, on-capillary stacking and off-line concentration on the capillary electrophoresis limits of detection for four structural types of industrial dyes, Anal. Chim. Acta, 349, 221–229, 1997. 257. T. Poiger, S.D. Richardson and G.L. Baughman, Analysis of anionic metallized azo and formazan dyes by capillary electrophoresis-mass spectrometry, J. Chromatogr. A, 886, 259–270, 2000. 258. A.R. Fakhari, M.C. Breadmore, M. Macka and P.R. Haddad, Non-aqueous capillary electrophoresis with red light emitting diode absorbance detection for the analysis of basic dyes, Anal. Chim. Acta, 580, 188–193, 2006. 259. W. Brumley and J.F. Farley, Determining eosin as a groundwater migration tracer by capillary electrophoresis/laser-induced fluorescence using a multiwavelength laser, Electrophoresis, 24, 2335–2339, 2003. 260. S.M. Burkinshaw, D. Hinks and D.M. Lewis, Capillary zone electrophoresis in the analysis of dyes and other compounds employed in the dye-manufacturing and dye-using industries, J. Chromatogr. A, 640, 413–417, 1993. 261. P. Blatny, C.-H. Fisher, A. Rizzi and E. Kenndler, Linear polymers applied as pseudo-phases in capillary zone electrophoresis of azo compounds used as textile dyes, J. Chromatogr. A, 717, 157–166, 1995.

976

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

262. E. Blanco, M.C. Casais, M.C. Mejuto and R. Cela, Analysis of tetrabromobisphenol A and other phenolic compounds in water samples by non-aqueous capillary electrophoresis coupled to photodiode array ultraviolet detection, J. Chromatogr. A, 1071, 205–211, 2005. 263. E. Blanco, M.C. Casais, M.C. Mejuto and R. Cela, Comparative study of aqueous and non-aqueous capillary electrophoresis in the separation of halogenated phenolic and bisphenolic compounds in water samples, J. Chromatogr. A, 1068, 189–199, 2005. 264. E. Blanco, M.C. Casais, M.C. Mejuto and R. Cela, Approaches for the simultaneous extraction of tetrabromobisphenol A, tetrachlorobisphenol A and related phenolic compounds from sewage sludge and sediment samples based on matrix solid-phase dispersion, J. Chromatogr. A, 78, 2772–2778, 2006. 265. S. Takeda, A. Omura, K. Chayama, H. Tsuji, K. Fukushi, M. Yamane, S.-I. Wakida, S. Tsubota and S. Terabe, Separation and on-line concentration of bisphenol A and alkylphenols by micellar electrokinetic chromatography with cationic surfactant, J. Chromatogr. A, 979, 425–429, 2002. 266. S. Takeda, S. Ilida, K. Chayama, H. Tsuji, K. Fukushi and S. Wakida, Separation of bisphenol A and three alkylphenols by micellar electrokinetic chromatography, J. Chromatogr. A, 895, 213–218, 2000. 267. F. Regan, A. Moran, B. Fogarty and E. Dempsey, Novel modes of capillary electrophoresis for the determination of endocrine disrupting chemicals, J. Chromatogr. A, 1014, 141–152, 2003. 268. B. Fogarty, F. Regan and E. Dempsey, Separation of two groups of oestrogen mimicking compounds using micellar electrokinetic chromatography, J. Chromatogr. A, 895, 237–246, 2000. 269. F. Regan, A. Moran, B. Fogarty and E. Dempsey, Development of comparative methods using gas chromatography-mass spectrometry and capillary electrophoresis for determination of endocrine disrupting chemicals in bio-solids, J. Chromatogr. B, 770, 243–253, 2002. 270. A. Goméz-Hens and M.P. Aguilar-Caballos, Social and economic interest in the control of phthalic acid esters, Trends Anal. Chem., 22, 847–857, 2003. 271. T. Heberer, Occurrence, fate and removal of pharmaceutical residues in the aquatic environment: A review of recent research data, Toxicol. Lett., 131, 5–17, 2002. 272. L. Nozal, L. Arce, B.M. Simonet, A. Rios and M. Valcárcel, Rapid determination of trace levels of tetracyclines in surface water using a continuous flow manifold coupled to a capillary electrophoresis system, Anal. Chim. Acta, 517, 89–94, 2004. 273. A. Macià, F. Borrull, C. Aguilar and M. Calull, Improving sensitivity by large-volume sample stacking using the electroosmotic flow pump to analyze some nonsteroidal anti-inflammatory drugs by capillary electrophoresis in water samples, Electrophoresis, 24, 2779–2787, 2003. 274. M. Himmelsbach, W. Buchberger and C.W. Klampfl, Determination of antidepressants in surface and waste water samples by capillary electrophoresis with electrospray ionization mass spectrometric detection after preconcentration using off-line solid-phase extraction, Electrophoresis, 27, 1220–1226, 2006. 275. A. Macià, F. Borrull, M. Calull and C. Aguilar, Determination of some acidic drugs in surface and sewage treatment plant waters by capillary electrophoresis-electrospray ionization-mass spectrometry, Electrophoresis, 25, 3441–3449, 2004. 276. J.J. Soto-Chincilla, A.M. García-Campaña, L. Gámiz-Gracia and C. Cruces-Blanco, Application of capillary zone electrophoresis with large-volume sample stacking to the sensitive determination of sulfonamides in meat and ground water, Electrophoresis, 27, 4360–4368, 2006. 277. J.-B. Kim, K. Otsuka and S. Terabe, On-line sample concentration in micellar electrokinetic chromatography with cationic micelles in a coated capillary, J. Chromatogr. A, 912, 343–352, 2001. 278. H. Harino, S. Tsunoi, T. Sato and M. Tanaka, Applicability of micellar electrokinetic chromatography to the analysis of estrogens in water, Fresenius J. Anal. Chem., 369, 546–547, 2001. 279. P. Schmitt-Kopplin and J. Junkers, Capillary zone electrophoresis of natural organic matter, J. Chromatogr. A, 998, 1–20, 2003. 280. P. Janoš, Separation methods in the chemistry of humic substances, J. Chromatogr. A, 983, 1–18, 2003. 281. D. Gajdošová, K. Novotná, P. Prošek and J. Havel, Separation and characterization of humic acids from Antarctica by capillary electrophoresis and matrix-assisted laser desorption ionization time-of-flight mass spectrometry. Inclusion complexes of humic acids with cyclodextrins, J. Chromatogr. A, 1014, 117–127, 2003. 282. M.L. Pacheco, E.M. Pena-Méndez and J. Havel, Supramolecular interaction of humic acids with organic and inorganic xenobiotics studied by capillary electrophoresis, Chemosphere, 51, 95–108, 2003.

Separation Strategies for Environmental Analysis

977

283. H. Prosen, S. Fingler, L. Zupanˇciˇc-Kralj and V. Drevenkar, Partitioning of selected environmental pollutants into organic matter as determined by solid-phase microextraction, Chemosphere, 66, 1580–1589, 2007. 284. L.N. Sangolkar, S.S. Maske and T. Chakrabarti, Methods for determining microcystins (peptide hepatotoxins) and microcystin-producing cyanobacteria, Water Res., 40, 3485–3496, 2006. 285. G. Vasas, D. Szydlowska, A. Gáspár, M. Welker, M. Trojanowicz and G. Borbély, Determination of microcystins in environmental samples using capillary electrophoresis, J. Biochem. Biophys. Methods, 66, 87–97, 2006. 286. G. Vasas, A. Gáspár, C. Páger, G. Surányi, C. Máthé, M.M. Hamvas and G. Borbely, Analysis of cyanobacterial toxins (anatoxin-a, cylindrospermopsin, microcystin-LR) by capillary electrophoresis, Electrophoresis, 25, 108–115, 2004. 287. A. Gago-Martínez, J.M. Leão, N. Piñeiro, E. Carballal, E. Vaquero, M. Nogueiras and J.A. RodríguezVáquez, An application of capillary electrophoresis for the analysis of algal toxins from the aquatic environment, Intern. J. Environ. Anal. Chem., 83, 443–456, 2003. 288. P. Sázelová, V. Kašiška, S. Koval, A. Prusík and G. Peltre, Evaluation of the efficiency of extraction of ultraviolet-absorbing pollen allergens and organic pollutants from airborne dust samples by capillary electrophoresis, J. Chromatogr. B, 770, 303–311, 2002. 289. D. Horta, P.D. Mudgett, L. Ding, M. Drybread, J.R. Schultz and R.L. Sauer, Analysis of water from the Space Shuttle and Mir Space Station by ion chromatography and capillary electrophoresis, J. Chromatogr. A, 804, 295–304, 1998.

Part IIIA Microchip-Based: Core Methods and Technologies

Manipulation at the 32 Cell Micron Scale Thomas M. Keenan and David J. Beebe CONTENTS 32.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.2 Microfabrication Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.3 Controlling Cell Position in 3D Cultures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.3.1 Physical Entrapment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.3.2 Photopatterning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.3.3 Dielectrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.3.4 Inkjet Printing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.4 Engineering the Cellular Microenvironment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.4.1 Mechanical Microenvironment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.4.1.1 Static Manipulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.4.1.2 Dynamic Manipulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.4.2 Chemical Microenvironment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.4.2.1 Gradient Generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.4.2.2 Subcellular Chemical Compartmentalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.4.3 Electrical Microenvironment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

981 982 982 983 983 984 986 987 988 988 989 991 992 993 994 997 998

32.1 INTRODUCTION The advent of microfabrication technology and its more recent application to the field of biology has greatly enhanced our ability to engineer in vitro cell culture environments. Cells and their sensing elements are several hundred nanometers to tens of micrometers (microns) in size, and thus interact within a microscale environment, or “microenvironment.” Unlike conventional cell culture methods, microfabrication-based methods allow manipulation of the physical, chemical, and electrical properties of the cellular microenvironment at the microscale. As a result of this enhanced functionality, microengineered culture environments are becoming more widely used in the biological research community for a variety of applications, including investigating fundamental questions in biology, enriching specific cell types from mixed populations, and examining the response of cells to novel chemical compounds. Microengineered culture environments impose unique requirements on analysis tools for detecting and characterizing changes in cell behavior or physiology in response to specific environmental perturbations. Optimizing existing analysis tools for use with microengineered culture environments would greatly enhance the utility of microfabrication-based methods for biological studies. In this chapter, we provide a broad overview of the most common and recent microfabrication-based methods for controlling the position of cells within a culture environment, and the properties of the cellular microenvironment. 981

982

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Photolithography Negative photoresist

Soft lithography

Silicon wafer

Photoresist features PDMS prepolymer

UV light Mask Photoresist features

Microchannels

FIGURE 32.1 Photolithography and soft lithography. Photolithography selective polymerizes a photoactive polymer coating using UV light filtered by a patterned mask to create microscale features of defined dimensions. Soft lithography utilizes the photolithographically defined substrates to mold polymer replicas of the microscale features. (Reprinted and adapted from Li, N., et al., Crit Rev Biomed Eng, 31, 423. Copyright 2003. With permission from Begell House, Inc.)

32.2 MICROFABRICATION TOOLS Many microfabrication tools have been developed over the past 50 years to create more intricate and advanced microelectronics. The ability of these tools to precisely control the size and shape of polymers, metals, and other materials at scales as small as 1 µm, make them extremely adept at creating unique cell culture microenvironments. Two processes called “photolithography” and “soft lithography” are the most commonly used for biological studies.1,2 Photolithography begins by coating a substrate with a thin layer of photosensitive polymer called a “photoresist” (Figure 32.1). The thickness of the photoresist layer is controlled by spinning the substrate at a defined speed and length of time. Thinner layers are produced by spinning the wafer faster and longer. The coated substrate is then placed in contact with a thin sheet of glass or plastic that has user-defined opaque and transparent regions, called a “mask.” Ultraviolet (UV) light is then shone through the mask onto the coated substrate causing exposed regions to degrade (positive photoresist) or polymerize (negative photoresist). The exposed substrate is then placed in a chemical solution called “developer,” which removes the degraded or unpolymerized photoresist leaving user-defined features patterned on the substrate. Soft lithography uses these patterned substrates, called “masters,” to mold elastomeric polymer replicas. The most commonly used elastomer is poly(dimethyl siloxane) (PDMS), due to its biocompatibility, transparency, and high-fidelity replica molding properties.

32.3 CONTROLLING CELL POSITION IN 3D CULTURES The ability to control where a cell is positioned relative to its physical environment, other cells, or regions with particular chemical identities is useful for studying cell exploration and migration in the presence of various physical and chemical cues, investigating specific cell–cell or cell–substrate interactions, characterizing autocrine or paracrine signaling, or for automating cell analysis. There is an extensive body of literature describing methods to control the position of cells on two-dimensional (2D) substrates.1,3,4 For some applications 2D culture systems are ideal; however, for many others they are not. Although invaluable in developing our current understanding of many biological phenomena, 2D culture systems are artificial environments not representative of in vivo culture

Cell Manipulation at the Micron Scale

(a)

2D Culture

983

(b)

3D Culture

FIGURE 32.2 Comparison of 2D vs. 3D culture architectures. (a) 2D cultures consist of patterned or random cells (shaded spheres) on a rigid or semirigid substrate. 2D cultures can offer cells a richer environment with cell organization in three dimensions and incorporation of extracellular matrix components (black lines) with entrained growth factors and signaling molecules. (b) Scanning electron micrograph (SEM) of the deep-sea gulper shark retina. (Photo courtesy of Jill Olin, Hofstra University. With permission.)

environments except in a few distinct cases (e.g., endothelial or epithelial cells). Nearly all cells found in vivo are completely surrounded by an intricate and highly organized arrangement of extracellular matrix proteins, other cells, and a rich milieu of soluble biomolecules distributed throughout the extracellular fluid. Cell–cell contact and the juxtaposition of different cells or cell types relative to one another are essential components of normal biological function. Cells communicate in a variety of ways including integrin and gap junction signaling, or by secreting signaling proteins into the extracellular environment that act on the cell itself (i.e., autocrine signaling) or on neighboring cells (i.e., paracrine signaling). Two-dimensional cell cultures cannot fully replicate these intricate and complex environments (Figure 32.2) and as a result may be limited in the insight and information they can provide about normal and abnormal cell physiology. Three-dimensional (3D) culture architectures better simulate complex in vivo environments and may be able to provide more comprehensive and relevant information about cell responses to specific environmental factors. Because of the thorough description of 2D patterning methods in existing literature and the advantages of 3D culture systems, we will focus our discussion on efforts to control cell position in 3D culture architectures.

32.3.1 PHYSICAL ENTRAPMENT The simplest method of patterning cells in a 3D culture system is to create voids on the surface of a biocompatible gel (Figure 32.3). Voids can be created by degrading select regions of the gel,5 embossing the gel with patterns defined by a microfabricated stamp,6 or by molding the gel on a microfabricated surface.7,8 Cell patterning is achieved by seeding cells at random and rinsing away those that have not fallen into the voids created in the gel. The voids can subsequently be refilled with new gel to provide full encapsulation. Physical entrapment methods can be used with virtually any soft or moldable polymer; however, its reliance on conventional photolithography allows cell patterning in only two dimensions.

32.3.2 PHOTOPATTERNING Cells can also be patterned in 3D culture constructs using a process called “photopatterning.” Photopatterning (Figure 32.4) utilizes photoactive hydrogels that become cross-linked or polymerized when exposed to light.9 Cells are mixed with prepolymerized gel solution and exposed to UV light through a mask, similar to photolithography. Regions exposed to UV light polymerize and trap the cells contained within. The unpolymerized solution can be rinsed away, replaced with acellular gel solution, and flood exposed with UV light to provide a blank background. Alternatively, solutions containing other cell types can be photopatterned relative to the first pattern to create a wide variety of coculture architectures. Photopatterning is a very effective means of patterning cells in 3D cultures,

984

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Degradation

Embossing

Molding

Hydrogel Substrate

Seed cells

Rinse

Encapsulation

FIGURE 32.3 Physical entrapment cell patterning. Cells can be patterned in 3D hydrogel cultures by creating voids using laser (L) degradation, embossing, and molding. Cells are seeded in the voids and those that do not can be rinsed away. Addition of more hydrogel allows full 3D encapsulation.

even though cell position can only be controlled in two dimensions. However, its use of UV light to initiate polymerization may have adverse effects on cell viability and function. UV light can damage deoxyribonucleic acid (DNA) and creates polymer or cross linker radicals that may directly damage cells.

32.3.3 DIELECTROPHORESIS A phenomenon called dielectrophoresis has been used to pattern a variety of cell types on 2D substrates10,11 and more recently in 3D culture constructs.9 Unlike electrophoresis where charged species move in an applied electric field due to Coulombic forces (F = qE), dielectrophoresis capitalizes on the ability of a cell to become polarized when placed in an electric field. Dielectrophoresis is most often used in conjunction with alternating current (AC) electric fields since AC fields eliminate electrophoretic movement, and have less physiological impact on cells than direct current (DC) fields.11 When a cell is placed in an AC field, the magnitude and polarity of the induced dipole depend on the frequency of the applied field and the conductivities of the cell and the surrounding medium, described by the equation p(r) = 4πεm R3




εc − εm ε c + 2ε m



,

(32.1)

Cell Manipulation at the Micron Scale

(a)

985

(b)

(c) UV

(d)

(e)

FIGURE 32.4 Photopatterning. (a) Cells mixed with prepolymerized hydrogel solutions can be selectively illuminated with UV light through the use of an applied photomask (shown in black). (b) Regions exposed to UV become polymerized and entrap the cells. (c) The unpolymerized hydrogel can be rinsed away, replaced with acellular prepolymerized hydrogel solution, and flood exposed with UV to provide a blank background. (d) Photopatterning has been used to pattern 3T3 fibroblasts in PEG hydrogels against a blank background. (e) Aligned with other photopatterned cell types for coculture experiments. (Adapted from Albrecht, D. R., et al., Lab Chip, 5, 111, 2005. With permission from The Royal Society of Chemistry.)

where p(r) is the induced dipole, R is the radius of the cell, ε is the complex permittivity√(ε = ε + σ/(jω)) of the cell (c) or medium (m), σ is the conductivity of the cell or medium, j = − 1, ω is the angular frequency of the applied electric field, and E(r) is the applied electric field. If cells are placed in a nonuniform AC field, the spatial field gradient imparts different amounts of force on each half of the cell dipole (Figure 32.5a). When the complex permittivity of the cell is greater than that of the surrounding medium, the cell forms a positive dipole that causes it to move toward the electrode with higher field strength in what is known as “positive dielectrophoresis” (pDEP). If cell permittivity is less than that of the surrounding medium, a negative dipole is formed resulting in repulsion away from the electrode of higher field strength, or “negative dielectrophoresis” (nDEP). By engineering the shape and position of the electrodes, spatial field gradients can be created that localize cells to precise, user-defined locations (Figure 32.5b). The magnitude and direction of cell movement can be tuned by adjusting the conductivity of the surrounding medium or the frequency of the applied AC field. Since dielectrophoresis relies on the geometry of the electrodes to create the nonuniform electric fields, the method has been limited to the creation of 2D cell patterns within 3D culture constructs (Figure 32.5c).9 True 3D cell patterning would require more complex electrode geometries than those employed to date. Although dielectrophoresis has been used to levitate cells12–14 and could be used for 3D patterning, the method has not been utilized solely or in combination with any other methods to pattern cells in all three dimensions within a 3D culture construct. Because many biological hydrogels (e.g., collagen, fibrin, Matrigel™) have conductivities similar enough to cells to make dielectrophoretic cell patterning difficult, synthetic hydrogels such as poly(ethylene glycol) (PEG) are most commonly used.9 PEG hydrogels are biocompatible, but are devoid of the numerous ligands and other supportive signals that cells normally receive from the extracellular environment. Although there are numerous chemistries that can be incorporated into PEG

986

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a)

F V

(b)

pDEP

nDEP

(c)

FIGURE 32.5 Dielectrophoresis. (a) Nonuniform magnetic fields place different amounts of force on cells with induced dipoles, causing them to move toward (pDEP) or away from (nDEP) high field densities. (b) Dielectrophoresis has been used to pattern 3T3 fibroblasts by controlling the shape and position of the electrodes. (c) Dielectrophoresis has also been used to pattern 3T3 fibroblasts in 3D PEG hydrogels. (Adapted fromAlbrecht, D. R., et al., Lab Chip, 5, 111, 2005. With permission from The Royal Society of Chemistry.)

hydrogels,15 the inability to fully recapitulate the natural environment may confound experimental observations. Other confounding variables may be contributed by localized heating from the applied electric field, which could damage cells or activate intracellular signaling cascades. The electric field can also perturb the electrical state of the cell membrane and the function of membrane-embedded ion channels.

32.3.4 INKJET PRINTING Inkjet printing of biological solutions has been used to pattern a variety of biological molecules on 2D substrates16 and could prove to be an efficient method for patterning cells in engineered 3D culture architectures. Biological inkjet printers are adapted conventional inkjet printers17,18 that deposit one or more biological solutions in user-defined patterns under the control of a computer. The computer controls the location of fluid ejection by rastering the printer head over the substrate or by moving the substrate relative to a stationary printer head. The solutions are ejected from the microfluidic channels and nozzles that constitute the inkjet printer heads using either heat-based or piezoelectric fluid displacement. Heat-based fluid displacement, also known as “bubble jet” printing, utilizes a heating element to vaporize a portion of the ink and create a bubble that ejects fluid from the nozzle and onto the substrate (Figure 32.6a). When the bubble collapses it creates a vacuum that pulls more ink into the printer head. Piezoelectric fluid displacement uses a piezoelectric transducer to physically displace the fluid within the microchannel when a voltage is applied, and eject the fluid out of the nozzle and onto the substrate (Figure 32.6b). Addition of a z-stepper motor allows 3D

Cell Manipulation at the Micron Scale

(a)

987

(b) Piezo Transducer

Resistive heater Bubble

(c)

FIGURE 32.6 Inkjet printing. (a) Bubble jet printer heads use resistive heating elements to vaporize the ink and create a gas bubble that displaces fluid out of the nozzle. (b) Piezoelectric printer heads use a piezoelectric transducer to mechanically displace fluid out of the nozzle. (c) Bovine aortic endothelial cells were inkjet printed on an alginate-coated scaffold to create tubes 4 mm in diameter. (Reprinted from Varghese, D., et al., J Thorac Cardiovasc Surg, 129, 2, 470. Copyright 2005. With permission from American Association for Thoracic Surgery.)

cell culture constructs to be created using a layer-by-layer process. This method has been shown to effectively create 3D cultures of neuronal cell lines in fibrin gels19 and a variety of cells in alginate gels18 (Figure 32.6c). Because of the use of microfluidic channels in inkjet printer heads, care must be taken to choose microchannel dimensions that do not shear damage cells or other biological molecules within the printing solution. Minimization of shear effects often comes at the expense of drop size resolution, which limits how close regions of different composition can be placed relative to one another. In addition, the heating of biological ink when using bubble jet biological inkjet printers may damage cells or cause denaturation of proteins.

32.4 ENGINEERING THE CELLULAR MICROENVIRONMENT The ability to control the physical, chemical, and electrical properties of the in vitro cell culture environment will perhaps be the single greatest contribution of microfabrication methods to the field of biology. Engineering the cellular microenvironment can be used to create more physiologically relevant culture conditions, study biological phenomena (e.g., growth, development, response to damage, death), or to direct the behavior of cells toward specific industrial outcomes (e.g., tissue engineering, chemical manufacturing). There are three ways in which the cellular microenvironment can be altered—mechanically, chemically, or electromagnetically. Although most methods intend to alter the cellular microenvironment in only one of these ways, it is often difficult to avoid secondary perturbations to the cellular environment via one or both of the other modalities. Here, we review the major methods for altering the microenvironment using microfabrication technology.

988

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

32.4.1 MECHANICAL MICROENVIRONMENT 32.4.1.1 Static Manipulations Mechanical manipulations of the cell microenvironment can be either static or dynamic. Static manipulations consist of modifications to the physical environment with which the cell interacts. Altering the topography of the cell culture substrate is useful for understanding how cell growth and proliferation is influenced by topographical or morphological cues, and how cells explore or migrate in specific physical environments. Many studies have shown that topographical cues alone can alter cell behavior.20–22 The recent development of nanofabrication methods23–27 has enabled the creation of substrate topographies similar in size to those cells encounter in vivo, such as the protein fibers that comprise the extracellular matrix (Figure 32.7). A variety of different microfabrication methods have been used to create complex topographies on cell culture substrates. Many of these studies20 require expensive fabrication tools and complex methods and will not be discussed further here. The most common method of adding topography to cell culture substrates is by using photolithography to either add features to the substrate or mask the substrate for selective chemical or plasma etching. Alternatively, cell culture substrates can be molded28–30 or embossed31 directly from photolithographically defined substrates. Photolithography and soft lithography have been used to generate a wide range of cell culture substrate topographies28,32 including replicas of cells33 (Figure 32.8). More complex 3D topographies can be generated using 3D patterning techniques.34 The simplest method of generating 3D topographies is employing multiple iterations of photolithography (Figure 32.9a). Unfortunately, multilayer photolithography requires precise alignment of each subsequent layer and may result in poor feature resolution and fidelity. Generating multiple level features in a single photolithography step can be accomplished using gray-scale photolithography (Figure 32.9b). Gray-scale photolithography uses positive photoresists and masks printed with different gray levels35 (i.e., transparency) or masks made of microfluidic channels filled with dye.36 Although gray-scale photolithography is effective at making features of different heights, it cannot make more complex features such as overhangs, closed loops, or hollow objects. These types of complex 3D structures require methods such as multiphoton polymerization.37–39 Multiphoton polymerization uses long wavelength lasers, where each photon has half the energy necessary to activate the photoinitiator in a photoactive polymer. When two beams are timed so that two long wavelength photons arrive at a single photoinitiator molecule at the same time, the photoinitiator absorbs both photons and is excited to its active state causing localized polymerization. By moving the focal point of the two incident laser beams, complex 3D structures can be constructed (Figure 32.9c). These features can subsequently be combined with soft lithography to form more intricate 3D structures than those that can be produced with standard photolithography (Figure 32.9d–e).39

(a)

(b)

FIGURE 32.7 Natural versus synthetic substrates. (a) SEM of the basement membrane of the rhesus macaque urothelium. (From Figure 2 in Abrams, G. A., et al., Urol Res, 31, 341, 2003. Copyright 2003. With permission from Springer Science and Business Media.) (b) Electrospun poly (lactide-co-glycolide) (PLGA) fibers (scale bar = 10 µm). (Adapted from Li et al., J Biomed Mat Res, 60, 4, 613, 2002. With permission.)

Cell Manipulation at the Micron Scale

989

(a)

(b)

(c)

(d)

FIGURE 32.8 Photo-/soft lithography generated topographies. (a) Bone marrow derived connective tissue progenitor cells growing on PDMS channels molded from photolithographically defined masters. (From Figure 5b in Mata, A., et al., Biomed Microdev, 4, 4, 267, 2002. Copyright 2002. With permission from Springer Science and Business Media.) (b) Neurons growing on PDMS ridges with neurite bridge (arrow) spanning the gap. (Reprinted from Goldner, J. S., et al., Biomaterials, 27, 460. Copyright 2006. With permission from Elsevier.) (c) Fixed rat Schwann cells were used to create (d) PDMS replicas that were subsequently found to guide neurite extension when seeded with dorsal root ganglion neurons (not shown). (Reprinted from Bruder, J. M., et al., Langmuir, 22, 20, 8266. Copyright 2006. With permission from American Chemical Society.)

(a)

(b)

(c)

(d)

Photoresist (e)

FIGURE 32.9 3D topographies. (a) A three-level master made with multiple iterations of photolithography. Three mesas, with the center mesa taller than the other two, are connected by ridges 1 × 1 µm in cross section (inset). (b) Multiple level features fabricated with gray-scale photolithography and microfluidic masks. (Adapted from Chen, C., et al. Proc Natl Acad Sci USA, 100, 1499, 2003. With permission.) (c) A 10-µm long, 7-µm tall bull made with two-photon photopolymerization in urethane acrylate polymer. (Reprinted from Kawata, S., et al., Nature, 412, 6848, 697, Copyright 2001. With permission from Macmillan Publishers Ltd.) (d–e) 3D PDMS structures molded from masters fabricated with multiphoton photopolymerization. (Adapted from LaFratta, C. N., et al., Proc Natl Acad Sci USA, 103, 23, 8589, 2006. With permission.)

32.4.1.2 Dynamic Manipulations Dynamic manipulations of the mechanical microenvironment consist of active elements that impose mechanical force on cells. Many cell types experience mechanical forces in vivo. Chondrocytes and osteoblasts within skeletal and load bearing connective tissues endure large compressive forces

990

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a)

(b)

(c)

(d)

FIGURE 32.10 Deformable substrates. (a) Force (black arrows) can be used to bend the cell culture substrate and mechanically strain attached cells. (b) The cell culture substrate can also be stretched, but due to the Poisson effect this results in accessory stresses and strains perpendicular to the direction of applied force (gray arrows), preventing mechanical loads from being truly uniaxial. (c–d) Microfabrication technology enhances the ways in which cells can be mechanically strained. Cells could be cultured on thin PDMS membranes that are patterned with a variety of unique topographies, and then mechanically strained when the membrane is deflected by pressurizing underlying microfluidic channels. (Adapted from Hoffman, J. M., et al., Adv Mater, 16, 23, 2201, 2004. With permission.)

during locomotion. Osteoblasts and endothelial cells experience large shear forces generated by fluid flow within the osteon and vasculature, respectively. The ability to culture cells in environments where different forms of mechanical loadings can be imposed provides invaluable insight into cell physiology. A variety of methods have been developed to apply mechanical forces to cells.40 One very common method is to culture cells on a deformable substrate that can be bent (Figure 32.10a) or stretched (Figure 32.10b). Transducers provide controlled loading with user-defined strains and strain rates, or cyclic loading with user-controlled frequency and amplitude. A recent advance in microfabrication technology now allows cells to be grown on microtopographies while being dynamically loaded. The method developed by Folch and coworkers41 employs thin PDMS membranes less than 20 µm thick that are patterned with complex substrate topographies (Figure 32.10c–d). The membranes are bonded to microfluidic channels connected to pressure or vacuum sources, allowing the membranes to be deflected dynamically. The resulting concave or convex surfaces mechanically strain cells interacting with the unique substrate topography. Fluidic shear forces are commonly generated in vitro using cone and plate viscometers.42 A cone and plate viscometer (Figure 32.11a) consists of a cone that rotates in close proximity to a flat substrate. Constant shear stress is created in the region between the cone and the substrate by engineering the taper of the cone to offset the influence of the increasing tangential velocity that

Cell Manipulation at the Micron Scale

991

(a)

(b) L


z

Po



PL

z V =
r =
z/(tan )  = –(dV/dz)  = –(
/tan )

y

x V=

(Po–PL) 2L

2

( H4

–z 2

)

 = –(dVdz) =–

(PL–Po) z L

FIGURE 32.11 Generating shear forces. (a) Cone and plate viscometers create constant shear stress throughout the volume between the cone and substrate. The shear stress (τ ) is independent of the radius (r) or distance from the cell culture surface (z), and dependent on the angular velocity (ω) of the cone and the angle of the taper (θ). (b) Microfluidic channels have at least one dimension less than 1 mm creating a laminar flow regime restriction that creates steady velocity profiles and shear stresses along the direction of fluid flow. The well-characterized flow regime allows mathematical calculation of both the velocity profile (V ) and the shear stress distribution across the channel (white graph perpendicular to fluid flow direction) using only the microchannel dimensions (L and H) and the applied pressure gradient (PO and PL ). The negative sign convention indicates that shear forces push down on the cells.

accompanies increasing radius on the shear rate. Cells cultured on the flat substrate are exposed to constant shear forces and can be examined for behavioral or physiological effects. Microfabrication technology provides a complementary method to expose cells to specific shear stress environments. Straight microfluidic channels (Figure 32.11b) with highly precise dimensions and controlled fluid flow rates have been used to study the effects of shear stress on cells.43,44 Microfluidic channels have one or more dimensions less than 1 mm, which forces flowing liquids to be confined to laminar flow regimes.45 Laminar fluid flow is characterized by a parabolic velocity profile that does not change downstream of a short flow stabilization region near the entrance of the microchannel. The steady-state parabolic velocity profile produces constant shear stresses in the direction of fluid flow. The velocity profile can be calculated from the fluid channel dimensions and the applied pressure gradient, and be used to calculate the shear stress throughout the microchannel. Because the velocity profiles near the junction of two microchannel walls can be more complex to predict, cells are usually cultured near the middle of microchannels that are much wider than they are tall.

32.4.2 CHEMICAL MICROENVIRONMENT Chemical manipulation of the cellular microenvironment encompasses chemical modification of the physical components with which the cell comes into contact (i.e., substrate, extracellular matrix, etc.), and the distribution of soluble chemical species in the culture medium in which the cells are grown. Numerous methods have been developed to modify the chemistry of cell culture substrates.1,46 Here we focus our discussion on different microfluidic methods for controlling the distribution of soluble chemical species in the cellular microenvironment. For this application, microfluidic devices offer unique advantages due to their restriction to laminar fluid flow.45 In addition to having a constant, parabolic velocity profile, laminar fluid flow is also characterized by more linear streamlines that follow the contours of the microchannel. The fluid does not flow in circuitous, disorganized paths as it does in turbulent flow. The lack of eddies and convective mixing results in a system where chemical species can only mix via diffusion. This limitation allows creation of predictable chemical gradients, which is not possible in conventional cell culture systems. The lack of convective mixing also allows tight control over the location of different flowing streams within the microchannel,

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

992

which has proven useful for exposing specific regions of cells to defined chemical environments. Here, we will discuss both gradient generators and subcellular fluidic compartmentalization using microfluidic channels.

32.4.2.1 Gradient Generators Gradients of diffusible biological molecules play critical roles in many biological phenomena including development, cancer, inflammation, and wound healing. A wide variety of methods have been developed to expose cells to soluble chemical gradients.47–52 Unfortunately, all of these methods offer little temporal or spatial control over the gradient, and in many cases provide no way to quantify the gradient that each cell perceives. A device we will refer to as the “Premixer Device” (Figure 32.12a)53 has been used to expose a variety of cells to biochemical gradients.54–56 The device generates steady-state, user-controlled gradients using constant fluid flow. The device splits and recombines inlet fluids in an upstream microfluidic mixer to generate a variety of different gradient shapes (Figure 32.12b). The use of transparent PDMS to fabricate the device and the closed channel geometry allows direct quantification of the gradient to which cells are exposed using a conventional wide field fluorescence microscope. The only disadvantage of the device is that it exposes cells to fluid flow and shear forces that can destroy autocrine or paracrine signaling, activate intracellular signaling cascades, and alter cell migration.57

(a)

Coverslip PDMS Inlets

PDMS Gradient across channel Flow

(b)

Normalized intensity (%)

Cell inlet

Outlet

Neutrophils

Coverslip

100 80 60 40 20 0

0

100

200

300

0

100

200

300

400

500 0

100

200

300

Channel width (M)

FIGURE 32.12 Premixer device. (a) An upstream microfluidic mixer splits and recombines fluids to expose cells to user-defined gradients under constant fluid flow. (b) The premixer can be designed to generate a wide variety of gradient shapes. (Adapted from Jeon, N. L., et al., Nat Biotechnol, 20, 8, 826. Copyright 2002. With permission from Macmillan Publishers Ltd.)

Cell Manipulation at the Micron Scale

993

(a)

(b)

(c)

(d) Normalized intensity

1 0.5 0 0 Po 150 300 si res tion a erv cro 450 0 oir, ss m

150

300

450

ss acro m ition Pos rvoir,  rese

FIGURE 32.13 Microfluidic multi-injector. (a) 3D schematic representation of the device shows two smalldiameter channels under the control of microfluidic valves that pneumatically eject fluid into a cell culture reservoir to form soluble molecule gradients. (b) Top view of the device in operation forming gradients of fluorescein isothiocyanate (FITC)-conjugated dextran. (c) 3D plot of the fluorescence intensity within the cell culture reservoir. (Adapted from Chung, B. G., et al., Lab Chip, 6, 6, 764, 2006. Reproduced with permission from The Royal Society of Chemistry.)

Another recently developed microfluidic device replicates a method developed by Gunderson and Barrett49 in a microfluidic platform. In the Microfluidic Multi-injector58 (Figure 32.13a), fluids are ejected pneumatically out of a pair of small orifices into a larger cell culture reservoir, under the control of independent microfluidic valves. Once the fluid is ejected into the reservoir, it forms a diffusive gradient (Figure 32.13b–c). Although the device generates gradients without appreciable fluid flow in the cell culture reservoir, as evidenced by a symmetric gradient, it cannot generate steady-state gradients or the complex gradients of the Premixer Device. A microfluidic generator called the “Microjets Device” (Figure 32.14a) generates stable gradients on open surfaces without exposing cells to appreciable fluid flow.59 The open architecture allows efficient nutrient and gas exchange, and the lack of flow minimizes shear forces and facilitates autocrine and paracrine signaling. Like the previously described device, the Microjets Device generates gradients by pneumatically ejecting fluids out of small orifices, which minimize convective flow and allow a diffusive gradient to form in the cell culture area. However, the Microjets Device uses two opposed arrays of small orifices delivering different concentrations of biochemical factor to generate steady-state chemical gradients. In addition, dynamic changes to the slope and position of the gradient can be accomplished by adjusting the air pressures driving the fluid out of each orifice array (Figure 32.14b–c). The Microjets Device is not capable of generating the complex gradients possible with the Premixer Device nor can the gradient be visualized with a conventional wide field fluorescence microscope. The architecture of the device requires confocal microscopy to evaluate the gradient to which the cells are exposed. 32.4.2.2 Subcellular Chemical Compartmentalization The laminar flow restriction and small dimensions of microfluidic devices have been exploited to expose different portions of a single cell to different chemical environments. A method developed by Whitesides and coworkers60,61 utilized hydrodynamic focusing62 on a microfluidic platform to selectively label portions of a cell with different dyes (Figure 32.15). Laminar fluid flow confined the dye solution to particular regions of the channel despite being in contact with streams containing

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

C=0 P = PL

C=0 P = PR

0.50

8

0

50

100

150

0 200

Position (µm)

1.00

PL PR

16

1.2 0.8 0.8 1.2 0.4 1.6

8

0.50

0.00 0

50 100 150 Position (µm)

C/Cº /m × 1000

0 psi +0.4 psi +0.7 psi

C/Cº

(c)

16 P

C/Cº

(b) 1.00

0.00

Open-surface gradient

Microjet arrays

(a)

C/Cº /m × 1000

994

0 200

FIGURE 32.14 Microjets device. (a) 3D schematic representation of the device showing opposed arrays of small orifices (i.e., small microfluidic channels) creating gradients in an open cell culture area. (b) Plot of the concentration profile (solid lines, left axis) and profile derivatives (dots, right axis) for three different pressure conditions illustrating how increasing the driving pressure delivered to the respective orifice arrays by the same magnitude causes an increase in gradient slope with little change in the gradient position. (c) Equal magnitude driving pressure offsets cause the gradient to shift position within the cell culture reservoir but does not affect the gradient shape. (Reprinted from Thomas M., et al., Appl Phys Lett, 89, 114103, 2006. Copyright 2006. With permission from American Institute of Physics.)

no dye. Cells within the microchannel were thus labeled only in the specific subcellular regions exposed to a particular dye. Microfluidic hydrodynamic focusing has since been used to examine the dynamics of epidermal growth factor (EGF) receptor signaling63 and the effects of focal application of agrin in stimulating in vitro neuromuscular junction formation.64,65 Microfluidic channels can also be used to physically compartmentalize portions of a cell. Jeon and coworkers66 replicated a Campenot67 chamber using microfluidic channels to provide subcellular isolation of central nervous system growth cones from their cell bodies. Neurons seeded in a central microchannel were allowed to extend axons through narrow restrictions to another large microchannel (Figure 32.16). The restrictions prevent appreciable fluid connections from being established between the two microchannels allowing the growth cones to be exposed to a different chemical environment than the cell bodies.

32.4.3 ELECTRICAL MICROENVIRONMENT Several important cell types are sensitive to the electrical characteristics of the cellular microenvironment. Neurons, cardiac myocytes, and retinal cells all generate and can be stimulated with electrical impulses. Because of the importance of these cell types, various methods have been devised to record or stimulate electrical activity within cells cultured in vitro. Traditionally, the electrical activity of cells has been recorded or stimulated by simply placing electrodes in the

Cell Manipulation at the Micron Scale

995

(b)

(a) PDMS Outlet

Inlets

Coverglass

(c)

(d) Cell

FIGURE 32.15 Hydrodynamic focusing. (a) A 3D schematic representation of the device shows how a fluid stream containing dye was confined by adjacent flow streams to expose only certain subcellular regions of the cell to the dye. (b) A bovine capillary endothelial cell labeled on one side with Mitotracker Green, and (c) on the other side with FM Mitotracker Red CM-H2 XRos. (d) Phase micrograph of the cell overlayed with fluorescence images from (b) and (c). (Adapted from Takayama, S., et al., Nature, 411, 1016, Copyright 2001. With permission from Macmillan Publishers Ltd.) (a)

Axonal side

Somal side

(b)

20 mm (c)

3 mm

100 µm

3 µm

FIGURE 32.16 Physical compartmentalization. (a) 3D schematic representation and 2D cross section of the device shows how neurons loaded in the somal side of the device (black) are unable to cross through the narrow constriction, unlike their axons, which can extend into the axonal side (white). (b) Texas red dextran injected in the axonal side (right) does not cross into the somal side showing fluidic isolation between the compartments. (c) When green cell tracker dye is loaded into the axonal side it retrograde labels the axons and somas of the neurons. (Adapted from Taylor, A. M., et al., Nat Methods, 2, 8, 599. Copyright 2005. With permission from Macmillan Publishers Ltd.)

culture medium, or alternatively, using a process known as “patch clamping” (Figure 32.17a). Patch clamping uses glass capillaries that are heated and then pulled at a defined speed to form a drawn tip approximately 1 µm in diameter. The drawn capillary, or micropipette, is then filled with a conductive electrolyte solution and a metal electrode, and brought to the surface of a cell using a micromanipulator. In this configuration the micropipette can be used to record the extracellular electrical activity or stimulate cells. To measure single cell transmembrane ion currents, a small hole is ripped in the cell membrane by placing the micropipette tip in contact with the cell membrane and

996

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a)

(b) +

+

V1

V2

–

–

FIGURE 32.17 Micropipette electrical recording/stimulation. (a) Pulled glass pipettes are used to measure the electrical activity of cells across the cell membrane (V1 ) by accessing the intracellular fluidic environment, or extracellularly (V2 ) by simply placing the micropipette near the cell. (b) A major limitation of micropipettebased electrical recording/stimulation is the inability to record from more than a few cells at a time due to space limitations of the micropipette set up. (Adapted Fitzsimonds, R. M., et al., Nature, 388, 439. Copyright 1997. With permission from Macmillan Publishers Ltd.)

gently pulling suction in the capillary. The cell membrane is pulled into the micropipette forming a high-resistance seal, and ruptured within the micropipette to provide fluidic, and thus electrical, access to the intracellular environment. Although the quality of the recordings or stimulations is excellent using micropipettes, the method suffers significantly from its lack of scalability. It is very difficult to record or stimulate electrical activity in more than just a few cells due to space limitations (Figure 32.17b). Trying to comprehensively characterize the electrical activity of cells within a tissue section or large dissociated culture is not possible using individually placed micropipettes. It is also difficult to record from or stimulate cells for more than a few hours due to biofouling or clogging of the micropipette tip by proteins and cell components, and the instability of any seal formed between the micropipette and the cell membrane. Microelectrode arrays (MEA) offer a solution to both the scalability and long-term recording limitations of micropipette electrodes. MEAs (Figure 32.18) are made by depositing a conductive metal on a cell culture substrate, patterning the surface with photolithography, and using the photoresist pattern as a mask for subsequent metal etching. After selective metal etching, the photoresist is stripped and the resulting metal electrodes are connected to a multichannel recorder or stimulator. Dissociated cells or tissue sections are then seeded on the MEA and can be recorded from or stimulated electrically. Micron-scale control over the size, shape, and position of the metal electrodes allows many electrodes to be placed on the cell culture substrate in a variety of configurations. Because MEAs are integrated into the cell culture substrate, they offer the ability to record from cell cultures for long periods of time. MEAs have been used primarily to record electrical activity in tissues and cultured cells,68 although they can also be used to electrically stimulate cells69,70 or other functions such as electroporate dissociated cell cultures with precise spatiotemporal control.71 MEAs are extremely useful for providing a comprehensive electrical profile of an entire cell culture, or for increasing experimental throughput by simultaneously recording from or stimulating multiple cells.68 Although the quality of the electrical recordings of MEAs is not as good as patch clamping, it is comparable with or better than other extracellular recording methods. Effective recording or stimulation of electrically active cells using MEAs requires close contact between the cell and the electrode. Large cell-electrode distances will make action potentials difficult to detect and electrical activity difficult to stimulate due to decreasing field strength with distance from the cell or electrode, respectively. To

Cell Manipulation at the Micron Scale

997

(a) Tissue with signal sources

Tissue/substrate contact

MEA Recording (b)

Stimulation (c)

200 µm

FIGURE 32.18 Microelectrode arrays. (a) MEAs can be used to record or stimulate electrical activity in dissociated cell cultures, engineered 3D cultures, or whole tissue explants. (b) MEAs are constructed using metal deposition and photolithography resulting in intricate, user-defined electrode configurations. (c) Brain tissue explant cultured on a MEA after 7 days in vitro. (Adapted from Stett, A., et al., Anal Bioanal Chem, 377, 486, 2003. Copyright 2003. With permission from Springer Science and Business Media.)

encourage close contact between cells and electrodes, MEAs are often patterned with adhesive proteins such as poly(lysine) and/or extracellular matrix proteins, or adhesion-promoting self-assembled monolayers using microstamping or other patterning techniques.1 MEAs have an added advantage in that they are compatible with and can be incorporated into many other methods discussed in this chapter to address more complex questions. For example, microfluidic channels constructed around an MEA72 would allow the electrical activity of cells or whole tissue sections to be recorded or stimulated while the cells are exposed to unique chemical, mechanical, or thermal73 environments. Cells could be exposed to chemical gradients, regulated shear forces, or disparate fluidic environments for many days while under constant electrical monitoring or stimulation.

32.5 SUMMARY Microfabrication technology has provided a plethora of tools and methods to engineer the position and microenvironment of cells in vitro. The unprecedented level of control over the mechanical, chemical, and electrical nature of the cellular microenvironment allows investigation of questions not addressable with conventional tools and methods. The unique insight into normal and abnormal cell behavior afforded by microfabricated tools and methods may one day lead to cures for injuries and diseases, and the ability to direct cell growth and behavior for tissue engineering or industrial applications. Although microfabrication technology has greatly enhanced the type of complex environments in which cells can be grown and studied, there is a large need for the development of complementary analysis methods to detect and characterize changes in cell response. Existing analysis methods, some of which are covered in the other chapters of this book, are often not adaptable to or optimized for use with microfabricated cell culture tools or environments. The development of novel methods that integrate engineered cell culture environments with complementary analysis tools would greatly accelerate our understanding of normal and abnormal cell behavior.

998

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

REFERENCES 1. Folch, A. and Toner, M., Microengineering of cellular interactions, Annu Rev Biomed Eng, 2, 227, 2000. 2. Li, N., Tourovskaia, A., and Folch, A., Biology on a chip: Microfabrication for studying the behavior of cultured cells, Crit Rev Biomed Eng, 31, 423, 2003. 3. Falconnet, D., Csucs, G., Grandin, H. M., and Textor, M., Surface engineering approaches to micropattern surfaces for cell-based assays, Biomaterials, 27, 3044, 2006. 4. Yap, F. L. and Zhang, Y., Protein and cell micropatterning and its integration with micro/nanoparticles assembly, Biosens Bioelectron, 22, 775, 2007. 5. Liu, Y., Sun, S., Singha, S., Cho, M. R., and Gordon, R. J., 3D femtosecond laser patterning of collagen for directed cell attachment, Biomaterials, 26, 4597, 2005. 6. Suh, K. Y., Seong, J., Khademhosseini, A., Laibinis, P. E., and Langer, R., A simple soft lithographic route to fabrication of poly(ethylene glycol) microstructures for protein and cell patterning, Biomaterials, 25, 557, 2004. 7. Tsang, V. L. and Bhatia, S. N., Three-dimensional tissue fabrication, Adv Drug Deliv Rev, 56, 1635, 2004. 8. Zhu, A., Chen, R., and Chan-Park, M. B., Patterning of a random copolymer of poly[lactide-coglycotide-co-(epsilon-caprolactone)] by UV embossing for tissue engineering, Macromol Biosci, 6, 51, 2006. 9. Albrecht, D. R., Tsang, V. L., Sah, R. L., and Bhatia, S. N., Photo- and electropatterning of hydrogelencapsulated living cell arrays, Lab Chip, 5, 111, 2005. 10. Gray, D. S., Tan, J. L., Voldman, J., and Chen, C. S., Dielectrophoretic registration of living cells to a microelectrode array, Biosens Bioelectron, 19, 1765, 2004. 11. Voldman, J., Electrical forces for microscale cell manipulation, Ann Rev Biomed Eng, 8, 425, 2006. 12. Kaler, K. V. I. S. and Jones, T. B., Dielectrophoretic spectra of single cells determined by feedbackcontrolled levitation, Biophys J, 57, 173, 1990. 13. Fuhr, G., Arnold, W. M., Hagedorn, R., Muller, T., Benecke, W., Wagner, B., and Zimmermann, U., Levitation, holding, and rotation of cells within traps made by high-frequency fields, Biochim Biophys Acta, 1108, 215, 1992. 14. Hakoda, M., Hachisu, T., Wakizaka, Y., Mii, S., and Kitajima, N., Development of a method to analyze single cell activity by using dielectrophoretic levitation, Biotechnol Prog, 21, 1748, 2005. 15. Roberts, M. J., Bentley, M. D., and Harris, J. M., Chemistry for peptide and protein PEGylation, Adv Drug Deliv Rev, 54, 459, 2002. 16. Boland, T., Xu, T., Damon, B., and Cui, X., Application of inkjet printing to tissue engineering, Biotechnol J, 1, 910, 2006. 17. Pardo, L., Wilson, W. C., and Boland, T. J., Characterization of patterned self-assembled monolayers and protein arrays generated by the ink-jet method, Langmuir, 19, 1462, 2003. 18. Varghese, D., Deshpande, M., Xu, T., Kesari, P., Ohri, S., and Boland, T., Advances in tissue engineering: Cell printing, J Thorac Cardiovasc Surg, 129, 470, 2005. 19. Xu, T., Gregory, C. A., Molnar, P., Cui, X., Jalota, S., Bhaduri, S. B., and Boland, T., Viability and electrophysiology of neural cell structures generated by the inkjet printing method, Biomaterials, 27, 3580, 2006. 20. Flemming, R. G., Murphy, C. J., Abrams, G. A., Goodman, S. L., and Nealey, P. F., Effects of synthetic micro- and nano-structured surfaces on cell behavior, Biomaterials, 20, 573, 1999. 21. Stevens, M. M. and George, J. H., Exploring and engineering the cell surface interface, Science, 310, 1135, 2005. 22. Abrams, G. A., Murphy, C. J., Wang, Z. Y., Nealey, P. F., and Bjorling, D. E., Ultrastructural basement membrane topography of the bladder epithelium, Urol Res, 31, 341, 2003. 23. Chen, Y. and Pepin, A., Nanofabrication: Conventional and nonconventional methods, Electrophoresis, 22, 187, 2001. 24. Murugan, R. and Ramakrishna, S., Nano-featured scaffolds for tissue engineering: Areview of spinning methodologies, Tissue Eng, 12, 435, 2006. 25. Norman, J. and Desai, T., Methods for fabrication of nanoscale topography for tissue engineering scaffolds, Ann Biomed Eng, 34, 89, 2006. 26. Pham, Q. P., Sharma, U., and Mikos, A. G., Electrospinning of polymeric nanofibers for tissue engineering applications: A review, Tissue Eng, 12, 1197, 2006.

Cell Manipulation at the Micron Scale

999

27. Li, W. J., Laurencin, C. T., Caterson, E. J., Tuan, R. S., and Ko, F. K., Electrospun nanofibrous structure: A novel scaffold for tissue engineering, J Biomed Mat Res, 60, 613, 2002. 28. Mata, A., Boehm, C., Fleischman, A. J., Muschler, G., and Roy, S., Analysis of connective tissue progenitor cell behavior on polydimethylsiloxane smooth and channel micro-textures, Biomed Microdev, 4, 267, 2002. 29. Martin, C. R. and Aksay, I. A., Microchannel molding: A soft lithography-inspired approach to micrometer-scale patterning, J Mat Res, 20, 1995, 2005. 30. Li, N. Z. and Folch, A., Integration of topographical and biochemical cues by axons during growth on microfabricated 3-D substrates, Exp Cell Res, 311, 307, 2005. 31. Narasimhan, J. and Papautsky, I., Polymer embossing tools for rapid prototyping of plastic microfluidic devices, J Micromech Microeng, 14, 96, 2004. 32. Goldner, J. S., Bruder, J. M., Li, G., Gazzola, D., and Hoffman-Kim, D., Neurite bridging across micropatterned grooves, Biomaterials, 27, 460, 2006. 33. Bruder, J. M., Monu, N. C., Harrison, M. W., and Hoffman-Kim, D., Fabrication of polymeric replicas of cell surfaces with nanoscale resolution, Langmuir, 22, 8266, 2006. 34. Geissler, M. and Xia, Y. N., Patterning: Principles and some new developments, Adv Mat, 16, 1249, 2004. 35. Henke, W., Hoppe, W., Quenzer, H. J., Staudtfischbach, P., and Wagner, B., Simulation and process design of gray-tone lithography for the fabrication of arbitrarily-shaped surfaces, Jpn J Appl Phys Part 1–Regular Papers Short Notes & Review Papers, 33, 6809, 1994. 36. Chen, C., Hirdes, D., and Folch, A., Gray-scale photolithography using microfluidic photomasks, Proc Natl Acad Sci USA, 100, 1499, 2003. 37. Kawata, S., Sun, H. B., Tanaka, T., and Takada, K., Finer features for functional microdevices— Micromachines can be created with higher resolution using two-photon absorption, Nature, 412, 697, 2001. 38. Baldacchini, T., LaFratta, C. N., Farrer, R. A., Teich, M. C., Saleh, B. E. A., Naughton, M. J., and Fourkas, J. T., Acrylic-based resin with favorable properties for three-dimensional two-photon polymerization, J Appl Phys, 95, 6072, 2004. 39. LaFratta, C. N., Lit, L. J., and Fourkas, J. T., Soft-lithographic replication of 3D microstructures with closed loops, Proc Natl Acad Sci USA, 103, 8589, 2006. 40. Brown, T. D., Techniques for mechanical stimulation of cells in vitro: A review, J Biomech, 33, 3, 2000. 41. Hoffman, J. M., Shao, J., Hsu, C. H., and Folch, A., Elastomeric Molds with tunable microtopography, Adv Mat, 16, 2201, 2004. 42. Duerrschmidt, N., Stielow, C., Muller, G., Pagano, P. J., and Morawietz, H., NO-mediated regulation of NAD(P)H oxidase by laminar shear stress in human endothelial cells, J Physiol, 576 (Pt 2), 557, 2006. 43. Frame, M. D., Chapman, G. B., Makino, Y., and Sarelius, I. H., Shear stress gradient over endothelial cells in a curved microchannel system, Biorheology, 35, 245, 1998. 44. Bransky, A., Korin, N., Nemirovski, Y., and Dinnar, U., An automated cell analysis sensing system based on a microfabricated rheoscope for the study of red blood cells physiology, Biosens Bioelectr, 22, 165, 2006. 45. Bird, R. B., Stewart, W. E., and Lightfoot, E. N., Transport Phenomena, John Wiley & Sons, Inc., 1960. 46. Shin, H., Jo, S., and Mikos, A. G., Biomimetic materials for tissue engineering, Biomaterials, 24, 4353, 2003. 47. Boyden, S., The chemotactic effect of mixtures of antibody and antigen on polymorphonuclear leucocytes, J Exp Med, 115, 453, 1962. 48. Zigmond, S. H., Ability of polymorphonuclear leukocytes to orient in gradients of chemotactic factors, J Cell Biol, 75 (Pt 1), 606, 1977. 49. Gunderson, R. W. and Barrett, J. N., Neuronal chemotaxis: Chick dorsal-root axons turn toward high concentrations of nerve growth factor, Science, 206, 1079, 1979. 50. Zicha, D., Dunn, G. A., and Brown, A. F., A new direct-viewing chemotaxis chamber, J Cell Sci, 99 (Pt 4), 769, 1991. 51. Foxman, E. F., Campbell, J. J., and Butcher, E. C., Multistep navigation and the combinatorial control of leukocyte chemotaxis, J Cell Biol, 139, 1349, 1997.

1000

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

52. Chen, H., He, Z., Bagri, A., and Tessier-Lavigne, M., Semaphorin−neuropilin interactions underlying sympathetic axon responses to class III semaphorins, Neuron, 21, 1283, 1998. 53. Jeon, N. L., Dertinger, S. K. W., Chiu, D. T., Choi, I. S., Stroock, A. D., and Whitesides, G. M., Generation of solution and surface gradients using microfluidic systems, Langmuir, 16, 8311, 2000. 54. Jeon, N. L., Baskaran, H., Dertinger, S. K. W., Whitesides, G. M., Van de Water, L., and Toner, M., Neutrophil chemotaxis in linear and complex gradients of interleukin-8 formed in a microfabricated device, Nat Biotechnol, 20, 826, 2002. 55. Chung, B. G., Flanagan, L. A., Rhee, S. W., Schwartz, P. H., Lee, A. P., Monuki, E. S., and Jeon, N. L., Human neural stem cell growth and differentiation in a gradient-generating microfluidic device, Lab Chip, 5, 401, 2005. 56. Saadi, W., Wang, S. J., Lin, F., and Jeon, N. L., Aparallel-gradient microfluidic chamber for quantitative analysis of breast cancer cell chemotaxis, Biomed Microdev, 8, 109, 2006. 57. Walker, G. M., Sai, J., Richmond, A., Stremler, M., Chung, C. Y., and Wikswo, J. P., Effects of flow and diffusion on chemotaxis studies in a microfabricated gradient generator, Lab Chip, 5, 611, 2005. 58. Chung, B. G., Lin, F., and Jeon, N. L., A microfluidic multi-injector for gradient generation, Lab Chip, 6, 764, 2006. 59. Keenan, T. M., Hsu, C. H., and Folch, A., Microfluidic “jets” for generating steady-state gradients of soluble molecules on open surfaces, Appl Phys Lett, 89, 114103, 2006. 60. Takayama, S., Ostuni, E., LeDuc, P., Naruse, K., Ingber, D. E., and Whitesides, G. M., Subcellular positioning of small molecules, Nature, 411, 1016, 2001. 61. Takayama, S., Ostuni, E., LeDuc, P., Naruse, K., Ingber, D. E., and Whitesides, G. M., Selective chemical treatment of cellular microdomains using multiple laminar streams, Chem Biol, 10, 123, 2003. 62. Spielman, L. and Goren, S. L., Improving resolution in Coulter counting by hydrodynamic focusing, J Colloid Interface Sci, 26, 175, 1968. 63. Sawano, A., Takayama, S., Matsuda, M., and Miyawaki, A., Lateral propagation of EGF signaling after local stimulation is dependent on receptor density, Dev Cell, 3, 245, 2002. 64. Tourovskaia, A., Figueroa-Masot, X., and Folch, A., Differentiation-on-a-chip: Amicrofluidic platform for long-term cell culture studies, Lab Chip, 5, 14, 2005. 65. Tourovskaia, A., Kosar, T. F., and Folch, A., Local induction of acetylcholine receptor clustering in myotube cultures using microfluidic application of agrin, Biophys J, 90, 2192, 2006. 66. Taylor, A. M., Blurton-Jones, M., Rhee, S. W., Cribbs, D. H., Cotman, C. W., and Jeon, N. L., A microfluidic culture platform for CNS axonal injury, regeneration and transport, Nat Methods, 2, 599, 2005. 67. Campenot, R. B., Local control of neurite development by nerve growth factor, Proc Natl Acad Sci USA, 74, 4516, 1977. 68. Stett, A., Egert, U., Guenther, E., Hofmann, F., Meyer, T., Nisch, W., and Haemmerle, H., Biological application of microelectrode arrays in drug discovery and basic research, Anal Bioanal Chem, 377, 486, 2003. 69. Jimbo, Y., Kasai, N., Torimitsu, K., Tateno, T., and Robinson, H. P. C., A system for MEA-based multisite stimulation, IEEE Trans Biomed Eng, 50, 241, 2003. 70. Shenai, M. B., Putchakayala, K. G., Hessler, J.A., Orr, B. G., Banaszak Holl, M. M., and Baker, J. R., Jr., A novel MEA/AFM platform for measurement of real-time, nanometric morphological alterations of electrically stimulated neuroblastoma cells, IEEE Trans Nanobiosci, 3, 111, 2004. 71. Jain, T. and Muthuswamy, J., Microsystem for transfection of exogenous molecules with spatiotemporal control into adherent cells, Biosens Bioelectron, 22, 863, 2007. 72. Pearce, T. M., Williams, J. J., Kruzel, S. P., Gidden, M. J., and Williams, J. C., Dynamic control of extracellular environment in in vitro neural recording systems, IEEE Trans Neural Syst Rehabil Eng, 13, 207, 2005. 73. Pearce, T. M., Wilson, J. A., Oakes, S. G., Chiu, S. Y., and Williams, J. C., Integrated microelectrode array and microfluidics for temperature clamp of sensory neurons in culture, Lab Chip, 5, 97, 2005.

Microfluidic 33 Multidimensional Systems for Protein and Peptide Separations Don L. DeVoe and Cheng S. Lee CONTENTS 33.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.2 Microfluidic Platforms for Multidimensional Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.2.1 Time-Multiplexed Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.2.2 Spatially Multiplexed Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.2.3 Interfacing Multidimensional Separations with MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.3 Methods Development Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.3.1 Fabrication of Spatially Multiplexed Microfluidic Chips . . . . . . . . . . . . . . . . . . . . . . . . . 33.3.2 Electrical and Hydrodynamic Crosstalk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.3.3 Sample Dispersion and Loss at Channel Intersections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.3.4 Inhibiting Bulk Flow in Multidimensional Microchannel Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.4 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1001 1002 1002 1003 1005 1008 1008 1008 1009 1010 1011 1012

33.1 INTRODUCTION Since the concept of micro total analysis systems (µTAS) was first proposed,1 the field has advanced rapidly, with ongoing developments promising to profoundly revolutionize modern bioanalytical platforms and methodologies. Whether termed µTAS, lab-on-a-chip, or microfluidics, the technologies that define the field offer important innovations capable of transforming the ways in which bioanalytical techniques are performed. For example, reduced size and power requirements can enable improved portability and higher levels of integration, with lower per-unit cost for disposable applications. Low volume fluid control enabled by microfluidics allows smaller dead volumes and reduced sample consumption, while many pumping methods including capillary action and electroosmotic flow scale favorably in these systems, enabling precise and valveless flow control at the microscale. Similarly, thermal time constants tend to be extremely small due to large surface area to volume ratios inherent in these systems, reducing the onset of significant Joule heating during electrokinetic separations and thus allowing higher separation voltages and shorter analysis times with equivalent or better separation resolution for complex mixtures in an integrated format. A significant advantage of microfluidics is that lithographic and replication-based fabrication techniques readily lend themselves to the formation of complex systems, providing a path for effective manufacturing of highly parallel analytical tools. Specifically, we consider here the promise that microfluidics technology holds for realizing robust multidimensional separations of proteins and peptides in an integrated platform. Assuming the separation techniques used in a given two-dimensional 1001

1002

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(2-D) system are fully orthogonal, that is, the two separation techniques are based on independent physicochemical properties of analytes, the overall peak capacity is the product of the peak capacities of the individual one-dimensional (1-D) methods.2 As a result, multidimensional systems are of great interest for the analysis of complex mixtures.3 Nowhere is the demand for effective analysis of complex matrices greater than in the field of proteomics. A major challenge faced in proteomics analysis is the vast number of proteins present in a typical biological sample, together with the large variation of protein relative abundances (typically greater than 6 orders of magnitude,4 and higher than 10 orders of magnitude for the case of blood plasma5 ), which ultimately dictate the need for high resolution separations to be performed on protein or peptide samples before detection by mass spectrometry (MS). Thus, multidimensional separation technologies capable of reducing dynamic range and enhancing detection sensitivity without substantially sacrificing analytical throughput are highly desirable for many proteomic applications. While intact protein analysis via 2-D polyacrylamide gel electrophoresis (2-D PAGE)6 and bottom-up peptide analysis combining multidimensional chromatography with tandem MS analysis7,8 remain the workhorse technologies for most modern proteomic studies, ongoing improvements in microfluidics for sample preparation and separations in proteomics are beginning to reveal their potential impact on the field. In this chapter, microfluidic platforms for protein and peptide analysis employing multidimensional electrophoretic separations and combined electrophoretic and chromatographic separations are discussed.

33.2 MICROFLUIDIC PLATFORMS FOR MULTIDIMENSIONAL SEPARATIONS 33.2.1 TIME-MULTIPLEXED SEPARATIONS The majority of reported multidimensional microfluidic systems employ serial separations that are time multiplexed. In this approach, fractions from the first separation dimension are sampled sequentially, with each fraction separated in series within the second dimension while the first-dimension separation continues. In one of the first demonstrations of this approach, Rocklin et al.9 performed 2-D separations of peptide mixtures in a microfluidic device using micellar electrokinetic chromatography (MEKC) and capillary zone electrophoresis (CZE) as the first and second dimensions, respectively. A schematic representation of their platform, fabricated in a glass substrate, is shown in Figure 33.1a. The MEKC separation was performed within a 65-mm-long serpentine channel located between the upper and lower channel intersection points shown in the figure, with a 10-mm-long CZE separation performed past the lower channel intersection. With a time scale for CZE on the order of a few seconds, substantially faster than the MEKC separation, large numbers of fractions could be sampled from the first dimension and separated by CZE with an overall analysis time under 10 min. Because MEKC is a transient electrokinetic separation, only a small portion (∼10%) of the analyte in the first dimension could be sampled into the CZE microchannel, since no fractions could be sampled while the second dimension separation was being performed. Despite this sample loss issue, a revised chip design was later reported for peptide separations with a peak capacity greater than 4000, such as the comparative separation of tryptic digest from bovine and human hemoglobin shown in Figure 33.1b revealing distinct peptide variations between these samples.10 Similar approaches based on time multiplexing have been employed by a number of researchers. For example, Gottschlich et al.11 used a spiral shaped glass channel coated with a C18 stationary phase for performing chromatographic separation of trypsin-digested peptides. By providing a cross interface, peptides eluted from an MEKC9 or reversed-phase11 chromatography channel were sampled and rapidly separated by CZE in a short glass microchannel. Herr et al.12 similarly coupled isoelectric focusing (IEF) with serial CZE for 2-D separations of model proteins, a concept that

Multidimensional Microfluidic Systems for Protein and Peptide Separations

(a)

(b) 700

S

1003

(A) Bovine hemoglobin

600

W1

B1

W2

MEKC elution time (s)

500 400 300 700

(B) Human hemoglobin

600 500 400

B2

W 300 0.7 0.8 0.9 1.0 1.1 1.2 CE migration time (s)

FIGURE 33.1 Time-multiplexed MEKC/CZE peptide separation chip schematic (From Rocklin, R. D. et al., Anal. Chem., 2000, 72, 5244–5249.) and (b) pseudo-gel views for chip-based separations of human and bovine hemoglobin. (From Ramsey, J. D. et al., Anal. Chem., 2003, 75, 3758–3764.)

was later implemented in a poly(dimethylsiloxane) (PDMS)-based microfluidic system using elastomeric valving to isolate the separation dimensions.13 In addition, Shadpour and Soper14 developed a microfluidic chip combining sodium dodecyl sulfate (SDS) gel electrophoresis with MEKC, again with multiple fractions sampled serially from the gel electrophoresis separation for time-multiplexed fractionation by MEKC in the second dimension. In each of the preceding examples, the separations were time multiplexed, with rapid second dimension separations chosen to enable repeated sampling of analyte eluted from a slower first dimension separation. Such time-multiplexed separations have also been employed in multidimensional capillary separations, for example, through the combination of a first dimension chromatographic separation together with a second dimension separation based on CZE,15 with low dead-volume switching valves used to couple the separation dimensions. For many applications, multidimensional microfluidic systems employing time-multiplexed separations may not offer compelling advantages over their capillary analogs, but rather introduce additional complexities, costs, and performance limitations, which may be avoided through the use of more traditional capillary technologies. On the other hand, unique benefits exist for applications that can leverage the small footprints and integrated nature of these microfluidic platforms, particularly when coupled with on-chip sample preparation.

33.2.2 SPATIALLY MULTIPLEXED SEPARATIONS In contrast to time multiplexing, high-throughput multidimensional separations may also be performed using spatial multiplexing, with an array of second dimension microchannels used to simultaneously separate analyte sampled as multiple fractions from the entirety of a first dimension separation channel. Using this approach relaxes the requirement for the second dimension separation to be substantially faster than the first dimension, opening the door to a wider range of potential separation modes. Furthermore, spatial multiplexing enables complete sampling of the first

1004

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

dimension separation, even for transient separation modes, thereby preventing sample loss due to a duty cycle mismatch between the dimensions. The use of spatial multiplexing also leverages a significant advantages of microfluidics, namely the ability to combine large numbers of channels in a single chip without the need for complex fluidic interconnects between the channels. The concept of spatially-multiplexed microfluidic separations was first embodied in a microfabricated quartz device proposed by Becker et al.16 which contained a single channel for a first dimension and an array of 500 parallel channels with submicron dimensions positioned orthogonally to the first dimension. While no separations were actually performed in this device, the concept was later extended and validated by several groups. For example, Chen et al.17 described a microfluidic 2-D capillary electrophoresis platform based on a reconfigurable system of six individual PDMS layers. The chip consisted of a 2.5-cm-long microchannel for performing IEF, with an intersecting array of parallel 6-cm-long microchannels for SDS–PAGE. This six-layer PDMS microfluidic device required the alignment, bonding, removal, realignment, and rebonding of various combinations of the six layers to perform a full 2-D protein separation. A fully integrated system developed by Li et al.18 combined IEF with SDS gel electrophoresis of intact proteins in a single polycarbonate chip containing 10 parallel second dimension microchannels (Figure 33.2). By combining both separation dimensions in a single rigid polymer chip, dispersion of analyte during the transfer between the dimensions was minimized. Furthermore, different separation media were employed in each of the dimensions, allowing sample mobility to be independently tailored for rapid free-solution IEF separations and high resolution SDS-capillary gel electrophoresis (CGE) separations. By using a replaceable polyethylene oxide (PEO) gel in the SDS–CGE channels, the interface between the separation media also provided for sample stacking during the transfer of focused proteins bands into the second dimension. Using this system, a comprehensive 2-D protein separation was completed in less than 10 min, with the majority of time consumed in the required SDS–protein complexation reaction. A peak capacity of ∼170 in the second dimension of size-based separation was estimated from a measured bandwidth of 150 µm over a 2.5 cm channel length. Because the separation mechanisms in IEF and SDS–CGE were completely orthogonal, the overall peak capacity was estimated to be 1700 (10 fractions from IEF × 170 from SDS–CGE). Improvements in peak capacity were proposed through the use of longer channels during the sizebased separation, and by increasing the density of microchannels in the array to increasing the number of IEF fractions analyzed during the size-based separation. Owing to the use of parallel

Reservoir D

Reservoir A

Reservoir B

Reservoir C

FIGURE 33.2 Schematic representation and photograph of a 2-D IEF/SDS–CGE microfluidic chip containing 10 spatially-multiplexed channels for SDS–CGE separation. (From Li, Y. et al., Anal. Chem., 2004, 76, 742–748.)

Multidimensional Microfluidic Systems for Protein and Peptide Separations

(a)

1005

(b)

FIGURE 33.3 (See color insert following page 810.) (a) Microfluidic PMMA chip for 2-D IEF/SDS–CGE protein separations and (b) pseudo-gel image resulting from a yeast cell lysate separation.

separations in the second dimension, there is no accompanying increase in the analysis time. Further improvements in separation resolution were later achieved through the use of individual reservoirs for the injection of PEO gel and SDS solutions into the chip, and through electrical isolation of each second dimension channel. A revised chip design is shown in Figure 33.3a, with an example separation of intact proteins from yeast cell lysate depicted in Figure 33.3b. Ivory and coworkers19 demonstrated a multistage separation platform based solely on multiple IEF separations, with an initial separation performed in a straight channel using broad-range ampholytes, followed by a second separation using a set of orthogonal microchannels that terminated in reservoirs containing ampholytes covering more narrow pI ranges for higher focusing resolution. By providing multiple sets of the orthogonal channels, progressively narrower pI ranges were established within the main separation channel. Alternately, multiple parallel narrow-range IEF separations could be performed using the apparatus. A three-stage transient IEF separation was demonstrated in this system, and additional stages were proposed. Multidimensional microfluidic separation platforms with even higher dimensionality have been proposed, such as a concept for a four-dimensional (4-D) platform combining IEF, isotachophoresis (ITP), SDS–PAGE, and reversed-phase liquid chromatography (RPLC)20 (Figure 33.4). It is not yet clear whether significant advantages can be realized by such high-dimensional fractionation, since analyte loss, dispersion, and dilution may ultimately prevent effective downstream detection (e.g., by MS), in particular given the relatively low loading capacities typical of many microfluidic systems. Regardless, this 4-D concept suggests an intriguing direction for new platforms and multidimensional separation modalities that can be enabled by the unique advantages offered by integrated microfluidics technology.

33.2.3 INTERFACING MULTIDIMENSIONAL SEPARATIONS WITH MS Mass spectrometry is an essential tool for the characterization of biomolecules, revealing the chargeto-mass ratio of analyte molecules following ionization. There is a strong need for effective MS in the analysis of proteins and peptides, where MS coupling is required to provide sufficient mass resolution and sequence information for modern proteomic analyses. Unlike deoxyribonucleic acid (DNA), no techniques exists for the direct amplification of proteins, and thus detection sensitivity is paramount for the identification of low abundance species from limited samples. The matter is further complicated by the enormous dynamic range of protein abundance in complex samples. For

1006

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

IEF

ITP SDS-PAGE

rp-HPLC

FIGURE 33.4 Conceptual layout of a spatially-multiplexed 4-D microfluidic system combining IEF, ITP, SDS–PAGE, and RPLC. (From Ivory, C. F., Electrophoresis, 2007, 28, 15–25.)

multidimensional microfluidic technologies to substantially impact the field of proteomics, effective methods for MS interfacing are a necessity. The integration of microfluidic systems and MS received substantial attention following demonstration of the first microfluidic electrospray ionization (ESI)–MS interfaces in 1997 by the groups of Karger and coworkers,21 Ramsey and Ramsey,22 and Aebersold and coworkers.23 Extensive studies on ESI–MS interfacing for coupling microfluidics to MS have been reported, with online ESI–MS offering a simple approach for directly interfacing microfluidic analyses with MS, while providing good sensitivity and mass accuracy, and reproducible signals for effective analyte quantification. A number of reviews have appeared in recent years, which illuminate this topic. Oleschuk and Harrison24 offer an early summary of microfluidic ESI–MS interfacing, and Sung et al.25 provide a more recent review addressing this area. A further review by Limbach and Meng26 in 2002 updates the developments in microfluidics for ESI–MS, and reviews by Figeys and Pinto,27 Lion et al.,28 and Marko-Varga et al.29 focus on microfabricated systems in proteomics, including summaries of relevant microfluidic-to-MS interfaces. A more detailed discussion of microfluidic ESI–MS interfacing is provided in Chapter 53 by Lazar. While ESI–MS is well suited as an interface between serial time-multiplexed microfluidic systems and MS, as reviewed by Foret and Kusý,30 interfaces based on matrix-assisted laser desorption/ionization (MALDI) offer several important advantages for spatially-multiplexed (i.e., parallel analysis) microfluidics. Unlike online ESI–MS, MALDI–MS is an offline soft ionization method, which is typically applied to the analysis of solid phase analyte co-crystallized with an energyabsorbing matrix material on the surface of a supporting target plate. In comparison with ESI–MS, MALDI–MS tends to be more tolerant of salts and other sample contaminants, offers excellent detection limits, and produces mass spectra which are relatively simple to interpret due to the absence of multiple charge states. Although generally limited to the analysis of higher molecular weight species due to interference from matrix components below ∼500 Da, MALDI–MS can be applied to a wide range of sample mass, with good sensitivity above 300 kDa, compared with less than 100 kDa for ESI–MS. From the point of view of multidimensional microfluidic systems, the off-line nature of MALDI– MS analysis represents a key advantage. First, MALDI–MS allows on-chip sample processing steps to be decoupled from back-end MS analysis. This is necessary when, for example, the time scales for biomolecular separations and online MS data acquisition are incompatible. More importantly for the present discussion, off-line MALDI–MS analysis provides a method for coupling simultaneous parallel on-chip analyses with MS detection, where ESI–MS from multiple microchannels is simply

Multidimensional Microfluidic Systems for Protein and Peptide Separations

1007

not practical or even feasible. It should be noted that although MALDI–MS is a serial process once the sample plate has been prepared, its high duty cycle enables high throughput for large numbers of deposited samples. For reviews of work in this area, Foret and Preisler31 presented an early summary of microfabricated systems for MALDI–MS interfacing, and DeVoe and Lee32 recently reviewed the current state of microfluidic interfacing with MALDI–MS. Additional discussion of microfluidic MALDI–MS interfacing may be found in Chapter 49 by Laurell. For microfluidic systems that employ parallel chromatography channels, MALDI–MS interfacing is particularly straightforward. The concept of interfacing multiplexed capillary chromatography separations with MALDI–MS for high-throughput proteome analysis was first described by Aebersold and coworkers33 using a system comprising four parallel capillary chromatography columns coupled with MALDI–MS/MS. In their system, a network of capillaries connected together with microjunctions was used to interface a single gradient pump to four parallel separation columns, with eluent from each column deposited by direct mechanical spotting onto a MALDI target at a flow rate of 1 µL/min using a commercial fraction collector. Flow rate variations of ±5–6% were observed between the columns, presumably due to variations in hydrodynamic flow resistance resulting from inhomogeneity of the packed silica beads. In a recent demonstration by Knapp and coworkers,34 a multichannel cyclic olefin copolymer (COC) chip was developed for parallel nanoflow RPLC analysis. Chip effluent was deposited onto a MALDI target through an electrically mediated deposition technique,35 which was shown to be robust for the case of a two-channel chromatography chip. While electrospray36 and pulsed electric field35 deposition has been demonstrated from multichannel chips, there are several disadvantages associated with the use of electric fields to assist sample deposition. In addition to increased system complexity, the reliability and repeatability of droplet deposition degrade as the number of parallel channels is increased. In contrast, the direct mechanical spotting strategy described by Aebersold and coworkers33 in a capillary format is well suited for multichannel microfluidic chips employing a chromatographic separation in the second dimension. For example, the eight-channel COC chip shown in Figure 33.5 contains 100-µm wide and 60-µm deep channels, with a single fluid port connected to all eight channels through a symmetric on-chip splitter. The end of the chip was cut with a high-speed semiconductor dicing saw to expose the channel exits. Using a syringe pump, a panel of nine model peptides were deposited onto 64 target spots (8 channels × 8 spotting events) on a custom MALDI target, with a target deposition volume of 150 nL per spot corresponding to sample amounts ranging from 25 to 75 fmol for each peptide. Using a Kratos Amixa MALDI-time-of-flight (TOF) instrument, the resulting spectra showed good uniformity, with an average relative standard deviation (RSD) of 44% across all peptides.

(a)

(b)

FIGURE 33.5 (a) Photo of an eight-channel COC chip aligned to a MALDI target mounted on a three-axis robotic positioner and (b) close up of chip exit showing droplet formation.

1008

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

This compares well with an RSD of 50% in MALDI signal intensities reported for multiple peptide spots deposited from a single capillary using an optimized matrix seed layer technique.37 Further improvements in deposition volume uniformity were realized for an eight-channel chip containing a methacrylate monolith developed for reversed-phase chromatography in the second separation dimension, with negligible variations at flow rates down to 200 nL/min per channel.

33.3 METHODS DEVELOPMENT GUIDELINES 33.3.1 FABRICATION OF SPATIALLY MULTIPLEXED MICROFLUIDIC CHIPS There are a number of practical issues that must be addressed to realize functional multidimensional microfluidic separation platforms. For devices based on spatially-multiplexed microfluidics, challenges exist even at the fabrication level, since producing microfluidic substrates with large numbers of parallel channels can be difficult. For example, when rapid prototyping methods based on hot embossing of thermoplastics such as polymethyl methacrylate (PMMA) or COC are used for substrate patterning, high frictional forces imposed during demolding of the plastic from dense microchannel arrays can lead to distortion of channel geometry, substrate warping, template damage, and in extreme cases prevent demolding from occurring entirely. It can also be difficult to achieve sufficient adhesion strength when bonding chips containing closely-spaced microchannels due to the reduced bond surface area. While these problems can be alleviated by employing low aspect ratio and tapered channel geometries, limiting the overall chip real estate occupied by the channels, and by keeping “reasonable” spacing between adjacent channels, such solutions can often be in conflict with desired performance goals for the system. The appropriate channel spacing must be evaluated on the basis of the particular fabrication methods, materials, and channel dimensions being employed, and weighed against potential performance losses resulting from a compromised design.

33.3.2 ELECTRICAL AND HYDRODYNAMIC CROSSTALK Another fundamental challenge results from electrical and hydrodynamic crosstalk between multiple interconnected channels that are inherent in spatially-multiplexed systems. To illustrate this point, consider the schematic representation of the 2-D IEF/SDS–CGE protein separation chip previously depicted in Figure 33.2a. This chip contains 11 parallel channels above the horizontal IEF channel, and 10 parallel channels below. As shown in the figure, each of these channels terminate in one of two common reservoirs, which connect the distal ends of the upper and lower channel arrays. In practice, this configuration results in a resistive network that shunts electrical current through the reservoirs during IEF separation, thereby producing an uneven electric field distribution within the IEF channel. More significantly, both analytes and ampholytes (for the creation of pH gradient in the IEF channel) can be unintentionally injected into the upper and lower channel networks by electrophoresis, resulting in potentially severe sample loss, and preventing establishment of the desired pH gradient for the present example of IEF as the first separation dimension. Issues with crosstalk can also manifest during hydrodynamic sample injection, since the resistive network introduced by the interconnected SDS–CGE channels applies equally to hydrodynamic and electrical current shunts. One approach to solve this problem would be to integrate flow control elements such as elastomeric valves into the system, but the additional complexity and inherent fabrication constraints make this solution less than ideal for real-world applications. Another obvious approach for eliminating crosstalk would be to provide individual reservoirs for all channels within the second dimension array, thereby isolating each channel both electrically and fluidically. As an example, this approach was chosen for the IEF/SDS–CGE device shown in Figure 33.3. However, there are practical limitations to this solution. As the number of reservoirs increases, so does the challenge of achieving efficient interfacing between off-chip plumbing and on-chip channels. For applications requiring

Multidimensional Microfluidic Systems for Protein and Peptide Separations

(a)

1009

(b)

FIGURE 33.6 (a) Manifold interface for providing fluidic interconnects between on-chip microchannels and off-chip pumps, valves, and electrodes and (b) manifold with integrated valves positioned above an IEF/SDS– CGE chip during testing.

pressure injection of sample, buffer, or other solutions into the device, on-chip connections are generally required at each reservoir (with the potential exception of waste reservoirs) to provide interfacing to pumps, valves, and/or electrodes, leading to a complex overall interface. For applications in which simple open reservoirs are suitable, for example, when electrokinetic sample injection is favored over pressure injection, the need for large numbers of interconnects to off-chip components may be alleviated, but additional challenges are introduced by bulk flow induced by hydrodynamic pressure gradients resulting from uneven fluid heights among large numbers of reservoirs in the chip. Furthermore, each added reservoir represents an additional point for bubble formation and entrapment, reducing the robustness of the microfluidic system. There are several solutions that may be used to address this interface challenge. For the case of the relatively simple IEF/SDS–CGE chip previously described in Figure 33.3a, 38 individual reservoirs were ultimately required for effective operation of the system, with a manifold interface used to simplify the fluidic and electrical interface. The manifold consisted of a machined PMMA block as shown in Figure 33.6, with O-ring compression seals and integrated valves, which provided leakfree hydrodynamic pressure control for chip preparation and sample introduction, and electrodes for electrokinetic control. The use of manifolds for microfluidic interfacing can substantially reduce the time and effort required for chip preparation and testing by eliminating the need for discrete fluidic or electrical ports to be added to each on-chip reservoir. Another solution is to reduce the required number of reservoirs by connecting limited numbers of channels to common fluid ports for off-chip access. Consider the example of a chip design comprising 32 second dimension channels shown in Figure 33.7a. Each set of eight adjacent channels within the upper and lower arrays are connected to a single reservoir, substantially reducing the total number of required ports. Although channels within each group remain subject to crosstalk during the first dimension separation, the overall current uniformity is greatly improved by segregating each group from the others as revealed in Figure 33.7b.

33.3.3 SAMPLE DISPERSION AND LOSS AT CHANNEL INTERSECTIONS When spatial or time multiplexing is employed as a multidimensional separation strategy with a first dimension based on an electrokinetic separation mode, an inherent difficulty arises from the need for the first dimension separation channel to be intersected by one of more second dimension channels. Regardless of whether the first dimension separation is operated in a transient (e.g., CZE) or steady-state (e.g., IEF) mode, electric field lines extending into the intersecting channels result in dispersion of sample out of the first dimension channel, and ultimately to sample loss as diffusion

1010

(b) 1.0 Channel current (normalized)

(a)

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

1 reservoir 0.9

0.8 4 reservoirs 0.7 2 reservoirs 0.6

0

0.2

0.4

0.6 0.8 1.0 1.2 Channel position (cm)

1.4

1.6

FIGURE 33.7 (a) Schematic representation of 2-D separation chip containing 32 second dimension microchannels, with groups of 8 second microchannels terminating at one of four common reservoirs and (b) model results showing relative current uniformity within the first dimension channel when using 1, 2, and 4 reservoirs.

allows sample to travel beyond the influence of the radiating electric field lines. This is of particular concern when analyte passing the intersections is in a concentrated band, since a substantial portion of the analyte may be lost into the interconnecting channel, while the remaining portion suffers from broadening after passing the intersection. Li et al.18 addressed this issue by integrating multiple separation media, with relatively low resistivity ampholyte and buffer in the first separation dimension for IEF and higher resistivity PEO gel used for the second dimension for CGE. As a result, the electric field remained largely constrained within the first dimension microchannel during IEF, and minimal sample dispersion was observed. In addition to constraining the electric field, low diffusion in the gel allowed dispersed analyte to be mobilized back into the first dimension channel without suffering a high degree of loss. As discussed previously, another benefit of this approach is that the multiple separation media further enabled effective sample stacking as focused analyte was electrokinetically transferred to the second separation dimension. Cui et al.38 proposed a solution based on the integration of addressable electrodes near each channel intersection to control the local current streamlines and thereby limit sample dispersion and loss. This is a more general approach that may be applied to a range of multidimensional systems. A potential disadvantage is that additional current is injected by each electrode, leading to current and electric field gradients within the first dimension separation channel, which can affect separation performance.

33.3.4 INHIBITING BULK FLOW IN MULTIDIMENSIONAL MICROCHANNEL NETWORKS An important practical issue that affects many microfluidic systems, and multidimensional microfluidics in particular, is unwanted hydrodynamic flow induced by on-chip pressure gradients. Owing to the complexity of the microchannel networks and relatively large numbers of reservoirs or other off-chip interconnects required to operate multidimensional separation chips, there is a large potential for pressure gradients to be induced by sources such as gravity (due to uneven fluid heights in the reservoirs), variations in capillary forces at the reservoirs, or trapped air bubbles within the reservoirs or microchannel networks. In the authors’ experience, fully eliminating these pressure

Multidimensional Microfluidic Systems for Protein and Peptide Separations

1011

gradient sources in complex microfluidic devices is rarely feasible, even when employing automated manifold interfaces, taking great care in chip preparation, and allowing pressure gradients to equilibrate over long periods of time. A more effective approach is to minimize hydrodynamic flow through the use of media, which inhibit bulk fluid motion, either by increasing the effective viscosity of fluids or by generating a physical barrier to flow. For example, Cui et al.19 added a 2.5% solution of methylcellulose to all reservoirs in a multidimensional IEF chip to limit the intrusion of reservoir solutions into the separation channels, while the replaceable PEO gel used as a second dimension separation medium by Li et al.18 served a similar purpose. In general, multidimensional systems employing gel-based separations are likely to suffer from less bulk flow than those using free solutions, as are systems employing chromatographic media that require high-pressure gradients to generate bulk flow.

33.4 CONCLUDING REMARKS The development of novel platforms for multidimensional separations of proteins and peptides remains an area of great interest throughout the analytical community. Microfluidic technologies offer real advantages over capillary-based systems in this area, particularly for the analysis of limited samples. Despite the promise and numerous examples of multidimensional microfluidic platforms for proteomic analysis, a number of challenges remain before these systems can become competitive with modern capillary systems such as 2-D liquid chromatography system for the analysis of complex substrates. The authors believe that as multidimensional microfluidic systems continue to evolve, the benefits of the technology will become increasingly evident. Specific areas that require attention include improved methods for fluidic interfacing and on-chip valving, which do not substantially impact system complexity or cost. Another important area of research currently being addressed by a number of groups is the development microfluidic systems with novel coupled separation dimensions that may offer improvements in separation efficiency. Interfacing spatially-multiplexed systems with MS also remains a significant challenge, and although a variety of solutions have been proposed and explored more efforts are needed to achieve parity with serial capillary ESI–MS. New development efforts in the area of time-multiplexed multidimensional separations may also provide real-world benefits for proteomic analysis. One of the difficulties with time multiplexing relates to sample loss resulting from the inability to fully sample all separated analyte bands within the first dimension microchannel. When a transient separation is employed in the first dimension, sample loss is inherent in this configuration due to the inability to collect fractions while the second dimension separation is proceeding. The difficulties with sample loss imposed by the serial nature of the second separation dimension could potentially be alleviated through the use of on-chip trap columns to store fractions before injection into the second dimension channel, providing parity between the desired rate of fraction acquisition and the duty cycle of the second dimension separation. Alternately, a combined time- and spatial-multiplexing scheme could be envisioned, wherein multiple second dimension separation channels sequentially sample the first dimension. Clearly, there are research issues that remain to be addressed before multidimensional microfluidics is commercially competitive against more traditional multidimensional capillary systems. However, with the demand for higher analytical throughput and smaller sample quantities in proteomics continuing to increase, there is an open opportunity for novel multidimensional platforms based on microfluidics to meet this demand. Beyond the concepts described in this chapter such as rapid and parallel separations, zero dead-volume flow control, and effective multichannel MS interfacing, further functionalities are also of great interest, for example, the combination of cell lysis, protein preconcentration, and digestion before multidimensional separation. Such technologies, already demonstrated for related lab-on-a-chip applications, offer an intriguing glimpse of what could ultimately be possible in a fully integrated system for protein or peptide analysis.

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

1012

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38.

Manz, A., Graber, N., Widmer, H. M., Sens. Actuat. 1990, B1, 244–248. Giddings, J. C., J. High Res. Chromatogr. 1987, 10, 319–323. Giddings, J. C., United Separation Science, Wiley, New York, 1991. Corthals, G. L., Wasinger, V. C., Hochstrasser, D. F., Sanchez, J.-C., Electrophoresis 2000, 21, 1104–1115. Jacobs, J. M., Adkins, J. N., Qian, W.-J., Liu, T., Shen, Y., Camp, D. G. I., Smith, R. D., J. Prot. Res. 2005, 4, 1073–1085. Rabilloud, T., Anal. Chem. 2000, 72, 48A–55A. Gygi, S. P., Rist, B., Griffin, T. J., Eng, J., Aebersold, R., J. Prot. Res. 2002, 1, 47–54. Peng, J., Elias, J. E., Thoreen, C. C., Licklider, L. J., Gygi, S. P., J. Prot. Res. 2003, 2, 43–50. Rocklin, R. D., Ramsey, R. S., Ramsey, J. M., Anal. Chem. 2000, 72, 5244–5249. Ramsey, J. D., Jacobson, S. C., Culbertson, C. T., Ramsey, J. M., Anal. Chem. 2003, 75, 3758–3764. Gottschlich, N., Jacobson, S. C., Culbertson, C. T., Ramsey, J. M., Anal. Chem. 2001, 73, 2669–2674. Herr, A. E., Molho, J. I., Drouvalakis, K. A., Mikkelsen, J. C., Utz, P. J., Santiago, J. G., Kenny, T. W., Anal. Chem. 2003, 75, 1180–1187. Wang, Y.-C., Choi, M. H., Han, J., Anal. Chem. 2004, 76, 4426–4431. Shadpour, H., Soper, S. A., Anal. Chem. 2006, 78, 3519–3527. Evans, C. R., Jorgenson, J. W., Anal. Bioanal. Chem. 2004, 378, 1952–1961. Becker, H., Lowack, K., Manz, A., J. Micromech. Microeng. 1998, 8, 24–28. Chen, X., Wu, H., Mao, C., Whitesides, G. M., Anal. Chem. 2002, 74, 1772–1778. Li, Y., Buch, J. S., Rosenberger, F., DeVoe, D. L., Lee, C. S., Anal. Chem. 2004, 76, 742–748. Cui, H., Horiuchi, K., Dutta, P., Ivory, C. F., Anal. Chem. 2005, 77, 7878–7886. Ivory, C. F., Electrophoresis 2007, 28, 15–25. Xue, Q., Foret, F., Dunayevskiy, Y. M., Zavracky, P. M., McGruer, N. E., Karger, B. L., Anal. Chem. 1997, 69, 426–430. Ramsey, R. S., Ramsey, J. M., Anal. Chem. 1997, 69, 1174–1178. Figeys, D., Ning, Y., Aebersold, R., Anal. Chem. 1997, 69, 3153–3160. Oleschuk, R. D., Harrison, D. J., Trends Anal. Chem. 2000, 19, 379–388. Sung, W.-C., Makamba, H., Chen, S.-H., Electrophoresis 2005, 26, 1783–1791. Limbach, P. A., Meng, Z., Analyst 2002, 127, 693–700. Figeys, D., Pinto, D., Electrophoresis 2001, 22, 208–216. Lion, N., Rohner, T. C., Dayon, L., Arnaud, I. L., Damoc, E., Youhnovski, N., Wu, Z. Y., et al., Electrophoresis 2003, 24, 3533–3562. Marko-Varga, G., Nilsson, J., Laurell, T., Electrophoresis 2004, 25, 3479–3491. Foret, F., Kusý, P., Electrophoresis 2006, 27, 4877–4887. Foret, F., Preisler, J., Proteomics 2002, 2, 360–372. DeVoe, D. L., Lee, C. S., Electrophoresis 2006, 27, 3559–3568. Lee, H., Criffin, T. J., Gygi, S. P., Rist, B., Aebersold, R., Anal. Chem. 2002, 74, 4353–4360. Ro, K. W., Liu, J., Knapp, D. R., J. Chromatogr. A 2006, 1111, 40–47. Ericson, C., Phung, Q. T., Horn, D. M., Peters, E. C., Fitchett, J. R., Ficarro, S. B., Salomon, A. R., Brill, L. M., Brock, A., Anal. Chem. 2003, 75, 2309–2315. Wang, Y., Zhou, Y., Balgley, B., Cooper, J., Lee, C. S., DeVoe, D. L., Electrophoresis 2005, 26, 3631–3640. Onnerfjord, P., Ekstrom, S., Bergquist, J., Nilsson, J., Laurell, T., Marko-Varga, G., Rapid Comm. Mass Spectrom. 1999, 13, 315–322. Cui, H., Huang, Z., Dutta, P., Ivory, C. F., Anal. Chem. 2007, 79, 1456–1465.

34 Microchip Immunoassays Kiichi Sato and Takehiko Kitamori CONTENTS 34.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.2 Basics of Immunoassay Microchips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.3 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.4 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1013 1014 1016 1018 1019

34.1 INTRODUCTION It is predicted that an enormous number of analyses will be required in the future in our daily lives, especially in the clinical and biochemical fields. Rapid progress in molecular biology and biochemistry has brought about a deluge of useful and important information. From these achievements, a significant number of genetic mutations have been elucidated, and a number of biomarkers concordant with disease have been established. These are not only useful for revealing the presence and, potentially, staging the disease but also for the a priori prediction of patient predisposition to diseases. By simple calculation, the number of required diagnostic analyses is thought to be the product of the population size and the number of indices. To perform all of these analyses, it is clear that significant investment of funds, time, and resources will be required, which is very difficult to realize with conventional analytical methods. One form of evolving technology, microfluidic devices or microchips can be a powerful solution to this problem. Microchip devices for chemical and biochemical analyses have been greatly advanced owing to the progress of microfabrication techniques. Microchemical systems using these devices have attracted much attention, not only from scientists and engineers but also from the clinical, forensic, and biomedical research communities. Thus far, most studies describing microchip-based analytical systems have focused on deoxyribonucleic acid (DNA) analysis by microchip electrophoresis with laser-induced fluorescence detection. These microchip-based electrophoretic systems have great advantages in some applications, especially in the clinical diagnostics and molecular biological fields. Other analytical methods must also be performed for select applications, which involve several distinct biochemical processes. To realize process functionality in these complicated systems, it is necessary to utilize the chemical properties and potential of molecules involved in an effective way. Immunoassay, founded by Rosalyn Yalow [1] in the late 1950s in the form of the “radioimmunoassay,” is still today one of the most powerful and important analytical methods used in clinical diagnoses and biochemical studies, primarily because of its extremely high selectivity and sensitivity. The antibody-based method used today has two generic classifications: homogeneous and heterogeneous. In heterogeneous immunoassays, the antibody is immobilized on a solid surface, while homogeneous assays take place in a solution phase. Integration of both types of immunoassay into microchips has been accomplished with reasonable success by numerous research groups. Most of the homogeneous immunoassays carried out in the microchip format have been based on a microchip electrophoresis [2,3], where the benefit of the chip-based integration was primarily the 1013

1014

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

reduction in separation time. In heterogeneous microchip immunoassay, however, a number of other advantages exist—and this represents the focus of this chapter. Enzyme-linked immunosorbent assay (ELISA) or other immunosorbent assay systems, in which antigen or antibodies are fixed on a solid surface, are applicable to many analytes with high sensitivity. These are used practically in many fields including clinical diagnoses and life science researches. The conventional heterogeneous immunoassay, however, requires a relatively long assay time and involves troublesome liquid-handling procedures and large quantities of expensive antibody reagents. Moreover, realization of point-of-care (POC) testing is difficult with the conventional immunoassay, since rather large devices are necessary for automated practical diagnosis systems. To overcome these drawbacks, a microchip-based system is effective. The integration of analytical systems into a microchip should bring about enhanced reaction efficiency, a reduced assay time, simplified procedures, and a lowered consumption of samples, reagents, and energy. The main reason of the long assay time in the conventional heterogeneous assay is that the reaction efficiency is very poor. This is because the reactions occur only on the solid surface. Moreover, it takes long time to complete the reaction due to the long molecular diffusion time. In a microchip, since it is easy to reduce the diffusion distance and increase the surface area-to-volume ratio, the reaction time can be reduced to several minutes rather than hours or days. A major source of the loss of reproducibility in the conventional assay is human error, because the assay procedure contains many troublesome manual operations. In the microchip format, all procedures can be controlled by pumps and valves that are regulated by a computer. Automation is very favorable without the need for any large-scale robotic equipment. Antibodies, enzymes, and other reagents used for immunoassay are expensive, hence, when used in the conventional method causes a large problem. Moreover, in many cases, samples are very precious or only a very small amount of samples can be obtained. Therefore, any reduction of the consumption in the microchip systems is welcomed.

34.2 BASICS OF IMMUNOASSAY MICROCHIPS Microfluidic devices for heterogeneous immunoassays consist of microchannels for transporting solutions, reaction solid phase, and detection area. There are several important factors to develop a high performance system. Microchip substrate is open to a variety of materials. Silicon wafer is a good material to build up microstructures if fabrication facilities are available. Glass has good chemical and optical properties, and some polymers are cost effective for mass production. The surface treatment of the microchannel is critically important for all materials. This is because of the nonspecific adsorption of the analytes and antibodies to the channel wall that will result in considerable analytical error. It is very important to modify the surface with some blocking reagents or other materials to prevent protein adsorption before experiments. Therefore, we must choose a material of which, surface chemistry is well understood. Reaction solid phase is the core of the immunoassay microchip where the primary antibody (capture antibody) is immobilized on the surface. There are some formats for the immobilization (Figure 34.1). The simplest method is a fixation of the antibody to the channel wall [4]. Several strategies have been used to adsorb or fix antibodies on the surface (e.g., direct adsorption), covalent bonding to react with function groups on the surface, and microcontact printing. While direct adsorption is common on hydrophobic polymer surfaces, covalent bonding by silanization reagents is used for silicon and glass surface. Since the amount and activity of the immobilized antibody is strongly dependent on the immobilization method, study of the fixation method is important [5]. Another type of solid support for capture antibody is microbeads. Beads coated with antibodies are packed in the microchannel with a dam or cage structure. Bead format is considerably superior to

Microchip Immunoassays

1015

(a)

(b)

(c)

(d)

FIGURE 34.1 Formats of the reaction solid phase for microchip immunoassay. (a) channel wall, (b) packed microbeads, (c) monolithic structure, and (d) magnetic beads.

(a)

(b)

100 µm

1.5 mm

6.5 mm

Antigen

Antibody

45 µm

FIGURE 34.2

Schematic illustrations of the antigen–antibody reaction. (a) Microtiter plate and (b) microchip.

the wall surface immobilization methods. Microbeads bring extremely large surface area-to-volume ratio and decrease the size of the liquid phase (i.e., decreasing the diffusion distance). These effects are useful to reduce the assay time dramatically. An example is shown in Figure 34.2. The surface area-to-volume ratio of a 50-µL solution in a microtiter plate well (0.65 mm in diameter) was estimated to be 13 cm–1 , whereas that of the microchannel (11 beads in 100 µm × 100 µm × 200 µm channel space) was 480 cm–1 . Therefore, the surface area-to-volume ratio the microchannel was 37 times larger than that of the microtiter plate, and the reaction rate may be increased because of this larger reaction field. In the case of the conventional microtiter plate assay, a 1.5-mm movement would be necessary for the most distant located antibody molecule to react with an antigen fixed on the surface of the well, since the liquid depth was 1.5 mm. On the other hand, the liquid phase of the microchannel filled with polystyrene beads was much smaller. The maximum distance from an antibody molecule to the reaction-solid surface will not exceed 20 µm. Because the diffusion time is proportional to the squares of the diffusion distance, the diffusion time of the antibody molecule to the antigen in the microchip would be more than 5600 times shorter than in the conventional method [6].

1016

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

The monolithic porous structure has also the same properties and suitable for the reaction solid phase. However, it is difficult to remove the monolith from the chip after an assay so the chip is then disposed. Hence, this happens even though the change of the beads is very easy. Therefore, the monolithic format is not favorable unless the cost of the chip is very low. Magnetic fine particles are another choice for the solid support. These particles can be captured in the channel easily with a magnet without any microstructures inside. In many cases, however, the particles are gathered very tightly and reagents are unable to go through them. That limits the number of particles contributed to the reaction. There are several kinds of detection methods for microchip-based heterogeneous immunoassays. The most popular technique is a fluorescent detection. Fluorescent molecules are used as a labeling material and detection is done at the reaction solid phase. A combination of a laser-induced fluorescence microscope, a highly sensitive detector, and a good labeling material may bring very sensitive determination, while the beads may scatter light. Nonfluorescent material, colloidal gold nanoparticle, is also useful for detection with a laser-induced thermal lens microscope (TLM). Instead of the optical labels, enzymes are also used as a labeling material [7]. ELISA is one of the most popular techniques in the conventional immunoassay and suitable in the microchip format. By using a fluorogenic reaction, the fluorescent signal can be taken at the downstream microchannel without interruption by microbeads. In the case of chromogenic reaction, a TLM is useful [8].

34.3 APPLICATIONS An example of the application of the microchip-based immunosorbent assay is shown below. The human carcinoembryonic antigen (CEA), one of the most widely used tumor markers for serodiagnosis of colon cancer, was assayed with a microchip-based system [9]. An ultratrace amount of CEA dissolved in serum samples was successfully determined within a short time. Polystyrene beads pre-coated with anti-CEA antibody were introduced into a microchannel. Then a serum sample containing CEA, the first antibody, and the second antibody conjugated with colloidal gold, First reaction

Mouse anti-CEA

Washing

CEA

Second reaction

Rabbit anti-CEA

Detection Washing

Third reaction

Washing

Anti-lgG Colloidal gold

FIGURE 34.3 Schematic illustrations of the microchip-based immunosorbent assay using microbeads.

Microchip Immunoassays

1017

(a)

(b)

100

Microchip assay (ng/mL)

Signal intensity (Arbitrary unit)

8

6

4

2

0

10 C

0.01

0.1

1

10

CEA in serum (ng/mL)

100

0

2

4

6

8

Conventional ELISA (ng/mL)

FIGURE 34.4 Determination of a tumor marker, human CEA. (a) Calibration curve for CEA in human sera. (b) Correlation between the conventional ELISA and the microchip-based immunoassay.

were reacted successively (Figure 34.3). The resulting antigen–antibody complex, fixed on the bead surface, was detected using a TLM. A highly selective and sensitive determination of an ultratrace amount of CEA in human sera was made possible by a sandwich immunoassay system that requires three antibodies for an assay. A detection limit in the orders of magnitude lower than the conventional ELISA was achieved (Figure 34.4). Moreover and pertinent to the clinical application of this technology, the analysis of serum samples from 13 patients showed a high correlation (r = .917) with the conventional ELISA. This integration reduced the time necessary for the antigen–antibody reaction to ∼1%, thus shortening the overall analysis time from 45 h to 35 min. This microchip-based diagnostic system might be the first micrototal analysis system (µTAS) to show a practical usefulness for clinical diagnoses with short analysis times, high sensitivity, and easy procedures. In numerous microchip systems, researchers have shown that multiplexed analysis can be achieved by increasing the number of channels available for parallel processing of numerous samples simultaneously for protein crystallization [10] and genetic [11] analysis. This type of higher integration has been achieved for multiplexed immunoassay analysis capable of processing several samples simultaneously [12]. In this integrated system, the chip had branching multichannels connected to four reaction and detection regions; thus, the system could process four samples at a time with only a single fluid pumping unit (Figure 34.5). To show utility, interferon gamma was assayed by a three-step sandwich immunoassay with the system coupled to a TLM as a detector. The biases of the signal intensities obtained from each channel were within 10%, with percent coefficient of variance (CVs) were almost the same level as the single straight channel assay. The total assay time for all four samples was 50 min when compared with 35 min for one sample in the single channel assay; hence, a higher throughput was realized with the branching structure chip. The simultaneous assay of many samples may also be achieved by simply arraying many channels in parallel on a chip. This approach, however, requires many pumps and capillary connections, and high integration seems to be difficult. On the other hand, a microchip with branching microchannels seems to be suitable for carrying out the simultaneous assay, since the numbers of pumps and capillary connections required for the system can be minimized.

1018

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Sample injection

Suction

Reagent injection

Injection Drain

FIGURE 34.5 Immunoassay chip with a branching microchannels for simultaneous assay.

34.4 FUTURE DIRECTIONS It is clear that much research has been done in this area, with the general sense that the microfluidic immunoassay can be superior to conventional immunoassay systems. It is interesting that, despite the large potential market size, microchip-based immunoassay systems have not yet been commercialized. One of the main reasons is the difficulty associated with creating a truly high-throughput assay without interfacing with automated fluidic control systems. While one microchip immunoassay can be completed in 30 min with manual operation, 96 assays can be completed within several hours in the conventional format. Therefore, it is clear that parallelization and automation will be required to realize the full potential. However, this will require the development and perfection of compact, cheap, precise, and automatic fluidic controller systems (including pumps, valves, and mixers). Precise control of submicroliter solutions without dead volume is a technical challenge and it is difficult to realize it only with inexpensive devices—however, such developments are underway in numerous labs. In addition, assay cost presents a problem, largely due to the cost of a 96-well microtiter plate at less than $2 dollars, while microchip costs could approach $100 dollars. Clearly, developing a cost-effective mass production approach will be a necessity. A candidate design for the automated and compact system might be a compact disk (CD) format. It is very easy to array more than 96 microchannels radially on a CD and all solutions/reagents can be controlled by centrifugal force without any pumping. The structured portable CD player or CD drive in a laptop PC might be a good analyzer as it can control the rotation speed precisely and has optical readout and could accommodate fluorescent detection. Ultimately, development of a functional, easy-to-use, and cost-effective system is a key to the acceptance and dissemination of such technology.

Microchip Immunoassays

1019

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

Yalow, R. S., Berson, S. A. J. Clin. Invest. 1960, 39, 1157–1175. Koutny, L. B., Schmalzing, D., Taylor, T. A., Fuchs, M. Anal. Chem. 1996, 68, 18–22. Chiem, N. H., Harrison, D. J. Clin. Chem. 1998, 44, 591–598. Bernard, A., Michel, B., Delamarche, E. Anal. Chem. 2001, 73, 8–12. Yakovleva, J., Davidsson, R., Lobanova, A., Bengtsson, M., Eremin, S., Laurell, T., Emneus, J. Anal. Chem. 2002, 74, 2994–3004. Sato, K., Tokeshi, M., Odake, T., Kimura, H., Ooi, T., Nakao, M., Kitamori, T. Anal. Chem. 2000, 72, 1144–1147. Eteshola, E., Leckband, D. Sens. Actuators B-Chem. 2001, 72, 129–133. Sato, K., Yamanaka, M., Hagino, T., Tokeshi, M., Kimura, H., Kitamori, T. Lab Chip 2004, 4, 570–575. Sato, K., Tokeshi, M., Kimura, H., Kitamori, T. Anal. Chem. 2001, 73, 1213–1218. Hansen, C.L., Skordalakes, E., Berger, J.M., Quake, S.R. Proc. Natl Acad. Sci. USA. 2002, 99, 16531–16536. Yeung, S.H., Greenspoon, S.A., McGuckian, A., Crouse, C.A., Emrich, C.A., Ban, J., Mathies, R.A. J. Forensic Sci. 2006, 51, 740–747. Sato, K., Yamanaka, M., Takahashi, H., Tokeshi, M., Kimura, H., Kitamori, T. Electrophoresis 2002, 23, 734–739.

35 Solvent Extraction on Chips Manabu Tokeshi and Takehiko Kitamori CONTENTS 35.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.2 Cocurrent Solvent Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.2.1 Simple Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.2.2 Ion Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.2.3 Liquid Membrane. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.2.4 Continuous Flow Chemical Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.2.5 Sample Preparation for Gas Chromatography and Mass Spectroscopy . . . . . . . . . . . 35.3 Counter-Current Solvent Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.4 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1021 1021 1021 1025 1027 1028 1030 1032 1033 1033 1033

35.1 INTRODUCTION Solvent extraction is a method capable of separating compounds based on their solution preferences for two different immiscible liquids. In other words, it is the extraction of a substance from one liquid phase into another liquid phase. Solvent extraction is a fundamental technique in chemical/biological laboratories and chemical industries, where it is carried out in the macroscale using separatory funnels. The advantage of doing a solvent extraction in the microspace on a chip is the scale merit of using small dimensions, that is, a large specific interface area (the interface-to-volume ratio) and a short diffusion distance, which together results in a short diffusion time. High speed and high performance solvent extraction systems are possible on chips, without the need for mechanical stirring, mixing, or shaking, all of which are necessary for the conventional method. This chapter focuses on the analytical aspects of solvent extraction on microchips. Although there are several reports in the literature on the application of chemical reactions,1−7 the topics dealt with here are limited only to application using parallel (side-by-side) flow schemes inside microchannels. Other examples, such as using a porous membrane,8 metal mesh,9 or droplets10 lie outside the scope of this chapter.

35.2 COCURRENT SOLVENT EXTRACTION 35.2.1 SIMPLE APPROACHES In order to perform an on-chip solvent extraction, a chip with at least two inlets for two immiscible liquids is required. However, the difficulty associated with its use, and the resultant extraction efficiency depends on the shape and dimensions of the channel, as well as other variables. Glass is the substrate usually used when performing solvent extraction using chips due to its chemical inertness toward organic solvents. If a polymer is to be used as the substrate, chemical modification of the channel is typically required. 1021

1022

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

The first solvent extraction carried out using a two-phase parallel flow in a microchannel was demonstrated by Tokeshi et al.11 using the typical extraction set up shown in Figure 35.1. An aqueous solution of Fe-bathophenanthrolinedisulfonic acid complex and a chloroform solution of capriquat were introduced into the microchannel, using syringe pumps at a constant flow rate to establish a parallel two-phase flow in the microchannel (Figure 35.2). In the case of ordinary solvent extraction using a separatory funnel, the two solutions in the separatory funnel are separated vertically by the difference in their specific gravities. Generally, in a microchannel, the liquid–liquid interface does not have this vertical arrangement, instead the flows are positioned parallel to the sidewalls

Capillary tube

Teflom screw

O-ring Waste

13 mm 10 mm 13 mm 8 mm 8 mm

46 mm φ 0.5

0.77 mm 66 mm

FIGURE 35.1 Schematic illustration of experimental setup. (From Tokeshi, M. et al., Anal. Chem., 72, 1711, 2000.)

Solvent Extraction on Chips

1023

Aqueous phase Fe-complex Organic phase

Microsyringe pump Drain

Capillary tube Capillary tube

FIGURE 35.2 1711, 2000.)

Schematic illustration of microextraction system. (From Tokeshi, M. et al., Anal. Chem., 72,

of the microchannel. This is due to the fact that the interfacial tension and friction force are much stronger than the force of gravity in the microspace provided by the microchannel.12 However, this also depends on the microchannel geometry—for example, the properties of the wall surface (hydrophobicity, hydrophilicity, wettability), the properties of liquids (viscosity, interfacial tension), and so forth.13 The time taken for the extraction in the 250 µm (w) × 100 µm (d) microchannel of Figure 35.2 was determined to be 45 s—this roughly coincides with the molecular diffusion time. The extraction time was also at least one order of magnitude shorter when compared with the conventional extraction time using a separatory funnel and mechanical mixing (shaking). In a similar manner, the Ni-dimethylglyoxime complex,14 Co-2-nitroso-5-dimethylaminophenol (DMAP) complex,15 and Al-2,2′ -dihydroxyazobenzen (DHAB) complex16 were all analyzed by performing an extraction at a two-phase interface, water/chloroform, water/toluene, and water/1-butanol, respectively. The determination of radioactive nuclides, such as uranium, was carried out by the extraction of U(IV) from an HNO3 aqueous solution to 30% (70% n-dodecane) and 100% tri-n-butyl phosphate phase (TBP).17 In the conventional method, 100% TBP was not used as an extraction solvent because it is impossible to get a separation between these two phases due to the fact that the specific gravity of the two liquids is almost identical. Alternatively, with the chip, the extraction of U(IV) can be achieved by using a two-phase aqueous/100% TBP flow because specific gravity is not a dominant variable. Moreover, using a three-phase flow, it is possible to obtain a rapid solvent extraction as illustrated in the following example, in which the extraction solvent m-xylene, was sandwiched by two aqueous solutions of Co-2-nitroso-DMAP complex to form a microchannel-based three-phase flow (Figure 35.3).12 The concentration of Co-2-nitroso-DMAP complex in the m-xylene phase was determined by using thermal lens microscopy (TLM)18 and corresponded well with the time taken for the extraction process. The dependence of the TLM signal on the distance x from the junction is shown in the graph of Figure 35.3. The extraction process reached equilibrium after only a distance of 1.5 cm, indicating that the extraction equilibrium was attained within 3 s of contact. The extraction efficiency can, however, be improved by using specially designed microchannels, for example, an asymmetrical zigzag-side-walled microchannel19 or a microchannel, which has intermittent partition walls in the center of the confluent part.20 Uniquely, Minagawa et al.21 demonstrated the integration of two chemical processes on a chip: a chelating reaction and solvent extraction. The layout of the microchannels fabricated on the glass plates is shown in Figure 35.4 and a schematic illustration of the molecular behavior demonstrated

1024

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a)

CoDMAP aq.

m-xylene

CoDMAP aq. x

0

TLM signal (arbitrary units)

(b)

150

100

50

0

0

5 10 15 20 25 30 35 40 x (mm)

FIGURE 35.3 (a) Schematic illustration of multilayer extraction system. (b) Position dependence of thermal lens signal intensity on distance x. (From Hibara, A. et al., Anal. Sci., 17, 89, 2001.)

16

8


= 0.5

30º

32 mm

8 8 8

15

15

5

15

8

66 mm

FIGURE 35.4 Layout and dimensions of microchip. Microchannel was 100-µm deep and 250-µm wide. (From Minagawa, T. et al., Lab Chip, 1, 72, 2001.)

in the microchannels is given in Figure 35.5. First, the m-xylene and the 2-nitroso-1-naphthol (NN) aqueous solutions were introduced into the two inlets of the first Y-shaped microchannel by the first syringe pump at a constant flow rate. These two liquids met at the intersection point, and a parallel two-phase flow, that is, an organic/aqueous interface, was formed in the microchannel. In this region, the NN in the aqueous phase could not be extracted into the m-xylene, whether or not the liquids were flowing. Next, the sample solution of cobalt ion [Co(II)] was introduced into the third inlet of the second Y-shaped microchannel using a second syringe pump at the same constant flow rate. Therefore, a three-phase parallel flow was formed upstream of the intersection point of the second Y-shaped structure. There was almost no interdiffusion in the three fluids under the flow condition, although the lack of interdiffusion depended on the flow rate. However, under these experimental conditions, interdiffusion can never occur. Thus, the chelating

Solvent Extraction on Chips

1025

m-xylene

NN/NaOH Co complex

Co2+ solution(aq.)

Co2+

NN

FIGURE 35.5 Schematic illustration of the chelating reaction and solvent extraction in microchannel. (From Minagawa, T. et al., Lab Chip, 1, 72, 2001.)

reaction of Co(II) with the NN and the extraction of Co chelates did not occur under flow. When the flow was stopped, the NN solution and the Co(II) sample solution promptly mixed, and Co(II) reacted with NN. Then, the reaction product, Co chelate in the aqueous phase was extracted into the organic phase. By using on-chip detection with a thermal lens microscope, troublesome operations such as the phase separation necessary for the conventional system can be avoided.

35.2.2 ION SENSING The liquid microspace provides a short diffusion distance, and large specific interfacial area, of the liquid–liquid interface. As a result several novel and attractive analytical features arise including extremely fast ion sensing, and a need for only an ultra small volume of reagent solution. In contrast to the slow response time of a standard ion-selective optode, where the response time is basically governed by the slow diffusion of ionic species in the viscous polymer membrane, that of the on-chip ion-sensing system is clearly faster due to the short diffusion distance and low viscosity of organic solution. The ion-pair extraction scheme is an established methodology employed for ion-selective optodes, which provide highly selective optical ion determination for various kinds of ions, using a single lipophilic pH indicator dye and a highly selective neutral ionophore. However, exploitation of an ion-pair extraction reaction and chip technology provide attractive advantages that would not be achieved by conventional ion sensors. The advantages of on-chip ion-sensing systems are as follows: 1. Reaction time (response time) can be dramatically reduced by the fast molecular transport achievable in a microspace. 2. Since the required reagent solution volume is extremely small (∼100 nL), a fresh organic phase can be used in every measurement. Subsequently, response degradation caused by the leaching of ion-sensing components, a typical problem of ion-selective optodes, would not need to be taken into consideration. This merit is directly reflected in the excellent reproducibility of the response during continuous measurements and the effective reduction of the amount of expensive reagents consumed in one measurement. 3. Ion determination can essentially be carried out by detection of the protonation/ deprotonation process of a single lipophilic anionic dye in the organic phase. Therefore, no special color-responsive chelating reagents are required for the measurement of another kind of analyte ion. One only need to alter the neutral ionophore for different ion-selective ionophores. From the viewpoint of optical instrumentation, this merit is

1026

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

quite important; there is no need to change the excitation source to match the excitation wavelength of the different chelating reagents. 4. Highly selective ionophores or carriers, for various kinds of ions developed for application to ion-selective electrodes, are commercially available and can be used without any chemical modification. Figure 35.6 shows a schematic illustration of the experimental setup and ion-pair extraction scheme.22 The organic solution containing a neutral ionophore, a lipophilic pH indicator dye, and an aqueous solution containing sample ion (K+ or Na+ ) were independently introduced into the microchannel to form an organic/aqueous interface. Then determination of the ion was done using a thermal lens microscope positioned downstream, where the ion-pair extraction occurred in the organic phase under continuous-flow conditions. The response time and minimal required volume of reagent for the on-chip ion-sensing system were about 8 s and 125 nL, respectively. In other work, a sequential multi-ion-sensing system using a single chip was successfully identified by expanding the concept of on-chip ion sensing.23 Figure 35.7 shows the basic concept for this chip-based multi-ion sensor. Different organic phases containing the same lipophilic pH indicator dyes, but different ionophores were introduced sequentially into the microchannel by the on–off switching of syringe pumps. In this case, having each organic phase contain different ionophores avoids contamination. The aqueous sample solution containing the different ions is introduced from the other inlet, to form a parallel two-phase flow with the intermittently pumped Organic phase containing lipophilic dye and ionophore

Thermal lens microscope detection

Syringe pump mm

Waste

30

10 mm

Water phase containing sample ion

60 mm Channel width 250 µm Channel depth 10 µm

TLM Protonated detection lipophilic dye Ionophore

Ion pair

Org.

250 µm

Aq.

Ion-pair extraction Ion

Proton

FIGURE 35.6 Schematic illustration of experimental setup and ion-pair extraction in microchannel. (From Hisamoto, H. et al., Anal. Chem., 73, 1382, 2001.)

Solvent Extraction on Chips

1027

C + io

noph

ore so lvent

B + io

nopho

re so lve

(a)-(c) Organic phases containing lipophilic dye and different ionophores (a) (b) (c) (d)

nt A + io

nopho

re

(d) Organic phase without ionophore Aqueous phase containing multiple ions

TLM Defection Waste

Recorder

FIGURE 35.7 Concept of sequential ion-sensing system using single microchip. (From Hisamoto, H. et al., Anal. Chem., 73, 5551, 2001.)

organic phases. The selective ion-pair extraction reaction proceeds during flow; thus, different ions can be selectively extracted into different organic phases, depending on the selectivity of the neutral ionophores contained in the respective organic phases. Downstream in the flow, the ion-pair extraction reaction becomes equilibrated; thus, downstream detection of the color change of the organic phase, allowed for sequential and selective multi-ion sensing in the single aqueous sample solution containing multiple ions. In this case, valinomycin and 2,6,13,16,19-pentaoxapentacyclo[18.4.4.4.7,12 0.1,20 07,12 ]dotriacontane (DD16C5)—known to exhibit high selectivity when used in conventional ion sensors—were used as highly selective potassium and sodium ionophores, respectively. Three types of aqueous sample solutions were analyzed with the system: a buffer solution containing 10−2 M K+ , 10−2 M Na+ , or both ions. When the aqueous phases containing a single type of ion were used, selective extractions occurred in each case; that is, potassium ions were extracted only for an organic phase segment containing valinomycin, and sodium ions were extracted only for that containing DD16C5. When the aqueous phase containing both ions was examined, both ions were independently extracted into different organic phases, depending on the nature of the ionophores in the respective organic phase (Figure 35.8). The minimum volume of single organic phase needed to obtain an equilibrium response without dilution by cross dispersion of two organic phases was ∼500 nL in the system, indicating that the required amounts of expensive reagents in one measurement could be reduced to a few nanograms.

35.2.3 LIQUID MEMBRANE A liquid membrane, composed of an organic phase with two aqueous phases, has a wide variety of industrial and analytical applications, including separation, concentration, and removal of

1028

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a) KCL aq.

TLM signal intensity (au)

K+

K+

K+

(b) NaCL aq. Na+

Na+

Na+

(c) KCI aq. & NaCI aq. K+

Na+

2

K+

K+

2

2

Na+

Na+

2 (min)

Org. Aq. Time

FIGURE 35.8 Response profiles for different aqueous solutions obtained by intermittent pumping of organic phase: (a) Aqueous solution containing 10−2 M KCl, (b) aqueous solution containing 10−2 M NaCl, and (c) aqueous solution containing both 10−2 M KCl and 10−2 M NaCl. (From Hisamoto, H. et al., Anal. Chem., 73, 5551, 2001.)

analytes from wastewater, environmental, and biomedical samples. Surmeian et al.24 demonstrated molecular transport on a chip using a stable three-layer flow membrane system, water/organic and solvent/water. Under continuous-flow conditions, the analyte (methyl red) was rapidly extracted across the microchannel from the donor to the acceptor phase through the organic phase (cyclohexene). The thickness of the organic phase, sandwiched by the two aqueous phases, was ∼64 µm, and it was considered a thin liquid organic membrane. Permeability studies identified the effects of molecular diffusion, layer thickness, and organic solvent-water partition coefficient on the molecular transport. In the chip, complete equilibration was achieved in several seconds, in contrast to a conventional apparatus, where it would require tens of minutes for the comparable extraction. Maruyama et al.25 showed another selective separation by using a three-layer flow membrane, water/n-heptane/water, to separate yttrium ions within a few seconds. In studies such as these, the formation of a stable liquid/liquid interface inside the microchannel under flow conditions is very important. The use of the microchannel structures and surface modification are particularly effective for stabilization of the liquid/liquid interface inside the microchannel.7,20,24−28

35.2.4 CONTINUOUS FLOW CHEMICAL PROCESSING The integration of chemical processes on a chip by using continuous-flow chemical processing (CFCP) in combination with microunit operations, such as solvent extraction, phase separation, and

Solvent Extraction on Chips

1029

Reaction and extraction area Sample Co2+ NN

NN/ NaOH

Decomposition and removal area

Co chelate

m-xylene

TLM detection

HCL Metal ions

Reaction & extraction area Sample NN/NaOH m-xylene

Waste HCI Waste

Metal chelates/ m-xylene

NaOH Decomposition & removal area

NaOH

Metal chelate

NN

FIGURE 35.9 Schematic illustration of Co(II) determination by using CFCP.

so forth, under continuous flow conditions is a strategy used for the construction of real micrototal analysis systems.27−30 The integration of a Co(II) wet analysis provides a good example of CFCP,27 and Figure 35.9 schematically illustrates this analysis. The microchip consists of two different areas: the former is the reaction and extraction area, and the latter is the washing, that is, decomposition and removal, area. In the former area, a sample solution containing Co(II) ions, a NN solution and m-xylene, are introduced at a constant flow rate through three inlets using syringe pumps. The three liquids meet at the intersection point, and a parallel two-phase flow, consisting of an organic/aqueous interface, forms in the microchannel. The chelating reaction of Co(II) and NN and extraction of the resulting Co(II) chelate proceed as the reacting mixture flows along the microchannel. Since the NN reacts with the coexisting metal ions [such as Cu(II), Ni(II), and Fe(II)], these coexisting metal chelates are also extracted into m-xylene. Therefore, a postextraction washing process is necessary for the decomposition and removal of the undesired coexisting metal chelates. The coexisting metal chelates decompose when they make contact with hydrochloric acid, and the metal ions are moved into an HCl solution (back extraction). The decomposed chelating reagent, NN, is dissolved in a sodium hydroxide solution. In contrast to the coexisting metal chelates, the Co chelate is stable in HCl and NaOH solutions, and remains. In the latter (washing) area, the m-xylene phase containing Co chelates and the coexisting metal chelates from the former (reaction and extraction) area is interposed between the other two inlets at a constant flow rate. Then, the three-phase flow, HCl/m-xylene/NaOH, forms in the microchannel. The decomposition and removal of the coexisting metal chelates proceed along the microchannel in the same manner as described above. Finally, the target chelates in m-xylene are detected downstream. Cobalt in an admixture sample was successfully determined with a rapid analysis time of less than 1 min. The advantages of this approach, compared with a conventional method, are the simplified nature of the procedure and avoiding troublesome operations. In the conventional method, the acid and alkali solutions cannot be used simultaneously, and alternative washing procedures must be repeated several times. The same effect can be obtained by using three-phase flow in the microchannel. In a subsequent paper, Kikutani et al.31 expanded the concept of CFCP to integrate four parallel analyses of Co(II) and Fe(II) on a chip in a three-dimensional microchannel network.

1030

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

b

a

c d

e (a) Cells (b) NO3–

(c)

LPS

NO

O2

Nitrate reductase

NO2– + H2NO2S

NO2– + NO3– NO2– NH2

H+

H2NO2S

N N+

NH (d) H2NO2S

N N+

NH2 H NO S 2 2

N N

NH NH2

(e) Detection with a thermal lens microscope

FIGURE 35.10 Chemical processes carried out in microchip for bioassay of macrophage-stimulating agent. (From Tokeshi, M. et al., Anal. Chem., 74, 1565, 2002.)

Goto et al.32 demonstrated the integration of a bioassay system that illustrated all processes required for a bioassay, that is, cell culture, chemical stimulation of cells, chemical and enzymatic reactions, and detection, on a chip using CFCP (Figure 35.10). By using the temperature control device, spatial temperature control of the system was possible, with areas on the chip maintained at different temperatures. Nitric oxide released from macrophage-like cells stimulated by lipopolysaccharide was successfully monitored by this system. The total assay time was reduced from 24 to 4 h, and the detection limit for nitric oxide was improved from 1 × 10−6 m to 7 × 10−8 M compared with the conventional batch methods. Moreover, the system could monitor a time course of the release, which is difficult to measure by conventional methods. Smirnova et al.33 demonstrated the determination of the insecticide, carbaryl, using a two-chip system. The first chip (for the hydrolysis of carbaryl) had a simple Y-shaped channel while the second chip (for the diazo coupling reaction between hydrolyzed products and 2,4,6-trimethylaniline)—the extraction required special channel shapes with a partial surface—modification obtained by using capillary-restricted modification (CARM) (Figure 35.11).34 Determination of carbaryl pesticide in water with sufficient sensitivity was carried out with an analysis time of 8 min. In a similar manner, Honda et al.35 developed a combination of a tube-type enzyme-immobilized microreactor and a microextractor with partial surface modification to produce optically pure amino acids.

35.2.5 SAMPLE PREPARATION FOR GAS CHROMATOGRAPHY AND MASS SPECTROSCOPY As the successful applications of a commercially available liquid chromatography (LC) chip for mass spectroscopy (MS) have been made apparent, the development of sample preparation chips for other

Solvent Extraction on Chips

1031

 = 532 mm

Toluene

Chip 2 Coupling reaction, extraction & detection

TMA NaOH Carbaryl

Toluene

Extraction

Chip 1 Carbaryl hydrolysis ODS

Buffer

Interface

FIGURE 35.11 Concept of integration of carbaryl determination onto microchips. Target carbaryl was hydrolyzed in chip 1 to the product 1-naphthol. In chip 2,1-naphthol was coupled with diazonium ion and was extracted into the organic phase. (From Smirnova, A. et al., Anal. Chim. Acta, 558, 69, 2006.)

1/ether, 100 mM

Synthesis microchip Capillary tubing Extraction microchip

TF AA/ether, 100 mM

Mass spectrometer Capillary tubing Syringe pumps

Phosphate buffer pH 6.8

FIGURE 35.12 Schematic illustration of synthesis microchip-extraction microchip-ESI-MS. (From Takahashi, Y. et al., J. Mass Spectrom. Soc. Jpn., 54, 19, 2006.)

conventional analytical systems, such as gas chromatography (GC), are also strongly desirable. As described above, microfluidic chips have great potential to integrate chemical processing such as sample preparation (see Chapter 44 by Biemvenue). Recently, several groups have developed the sample preparation chips for GC and GC–MS, and MS analyses. Miyaguchi et al.36 developed a solvent extraction chip for the off-chip GC analysis of amphetamine-type stimulants (a class of illegal drugs) in urine. Microchannels, partially fluoroalkylated by the CARM method,34 were employed for stabilizing the 1-chlorobutane/alkalize urine interface and obtaining a GC-amenable fraction. As a practical demonstration, methoxyphenamine hydrochloride was administered to three healthy volunteers, and the concentration of methoxyphenamine in their urine determined. This showed the potential of chip-based sample preparation to contribute to the rapid automated analysis in forensic toxicology. Similarly, Xiao et al.37 carried out solvent extraction of ephedrine using a polydimethylsiloxane (PDMS)/glass chip that had been obtained by surface modification of half of the glass wall, with octadecyltrichlorosilane. On-chip extraction of hydrocarbons from a North Sea oil for GC-MS analysis was also reported.38 Figure 35.12 shows a schematic diagram of an online high-throughput detection system for a reaction product synthesized by a microreactor.39 The system has a synthesis chip that a microreactor for the synthesis of 2,2,2-trifluore-N-phenetyl acetamide (TPA), an extraction chip for the purification of TPA from the reaction mixture, and an electrospray ionization mass spectrometer (ESI-MS) for

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

1032

detection. In this system, the extraction chip is placed between the microreactor and the ESI-MS to reduce the ionization hindrances for the reaction products. This placement makes the detection of the reaction product possible under any reaction conditions. The system as described here may be useful for the optimization of reaction conditions and the screening of valuable chemicals such as precursors of pharmaceuticals.

35.3 COUNTER-CURRENT SOLVENT EXTRACTION From the viewpoint of recovery efficiency, cocurrent solvent extraction as described above, can reach a theoretical plate number of only unity. In contrast to cocurrent solvent extraction, a higher theoretical plate number is expected using counter-current solvent extraction. However, counter-current flow in a microchannel cannot be easily formed because of interfacial tension and viscous force of both (aqueous and organic) phases. In an ordinary microchannel, counter-current flow cannot occur because the two phases collide, and high shear stress at the liquid/liquid interface causes breakup. To form counter-current flow, the aqueous solution must flow along one side of the channel, and the organic solution flow along the other, without breakup. Aota et al.40 successfully developed a counter-current flow system, which was obtained by selectively modifying the lower half of a microchannel with a hydrophobic group (Figure 35.13). Using this system, they realized that a theoretical plate number of 4.6 was needed for the extraction of a Co(II)-2-nitroso-dimethylaminopyridine (DMAP) complex in an aqueous-toluene counter-current flow. Counter-current flow is expected to be applicable to enrichment processes for various environmental analyses and biomolecular separations.

70 mm 30 m

m

(a)

Hydrophilic surface

Liquid/liquid interface

Microchannel Top plate (hydrophilic) Bottom plate (hydrophobic)

aq.

org.

Hydrophobic surface 20 mm

(b)

Butylacetate

Water

Water

300 µm

Water

Butylacetate Butylacetate 20 mm

FIGURE 35.13 (a) Photograph of microchip and (b) schematic illustration of counter-current extraction system. (From Aota, A. et al., Angew. Chem. Int. Ed., 46, 878, 2007.)

Solvent Extraction on Chips

1033

35.4 FUTURE DIRECTIONS The development and application of chip-based solvent extraction is gradually expanding as shown in this chapter. With methods for surface modification, the liquid–liquid interface in a microchannel can be stabilized, and so special techniques for solvent extraction are not required. Therefore, since solvent extraction is one of the basic chemical processes, it is expected that it will be used more often in the future. In the next 5 years, from a practical viewpoint, the combination of solvent extraction chips with conventional analytical apparatuses, such as GC, LC, MS, and so forth, and the integration of solvent extraction and other chemical processes will become more and more important. From the viewpoint of the basic science of two-phase and multi-phase flows, the flow itself near the liquid– liquid interface is very interesting. In fact, chaotic vortices and transient vortex-like flows have been observed around the liquid–liquid interface in the counter-current flow.41

ACKNOWLEDGMENTS We would like to thank our coworkers whose work has been referenced in this chapter. Kanagawa Academy of Science and Technology and The Ministry of Education, Culture, Sports, Science and Technology of Japan are acknowledged for financial support.

REFERENCES 1. Hisamoto, H., Saito, T., Tokeshi, M., Hibara, A. and Kitamori, T., Fast and high conversion phase-transfer synthesis exploiting the liquid–liquid interface formed in a microchannel chip, Chem. Commun., 2662, 2001. 2. Ueno, M., Hisamoto, H., Kitamori, T. and Kobayashi, S., Phase-transfer alkylation reactions using microreactors, Chem. Commun., 936, 2003. 3. Shaw, J., Nudd, R., Naik, B., Turner, C., Rudge, D., Benson, M. and Garman, A., Liquid/liquid extraction systems using micro-contactor arrays, Proc. µTAS 2000 Symposium, Enschede, The Netherlands, 371, 2000. 4. Ehrfeld, W., Hessel, V. and Löwe, H., Microreactors, Wiley-VCH, Weinheim, 2000, Chapter 5. 5. Kikutani, Y., Horiuchi, T., Uchiyama, K., Hisamoto, H., Tokeshi, M. and Kitamori, T., Glass microchip with three-dimensional microchannel network for 2 × 2 parallel synthesis, Lab Chip, 2, 188, 2002. 6. Kikutani, Y., Hibara, A., Uchiyama, K., Hisamoto, H., Tokeshi, M. and Kitamori, T., Pile-up glass microreactors, Lab Chip, 2, 193, 2002. 7. Maruyama, T., Uchida, J., Ohkawa, T., Futami, T., Katayama, K., Nishizawa, K., Sotowa, K, Kobota, F., Kamiya, N. and Goto, M., Enzymatic degradation of p-chlorophenol in a two-phase flow microchannel system, Lab Chip, 3, 308, 2003. 8. Cai, Z.-X., Fang, Q., Chen, H.-W. and Fang, Z.-L., A microfluidic chip based liquid–liquid extraction system with microporous membrane, Anal. Chim. Acta, 556, 151, 2006. 9. Wenn, D. A., Shaw, J. E. A. and Machenzie, B., A mesh microcontactor for 2-phase reactions, Lab Chip, 3, 180, 2003. 10. Chen, H., Fang, Q., Yin, X.-F. and Fang, Z.-L., Microfluidic chip-based liquid-liquid extraction and preconcentration using a subnanoliter-droplet trapping technique, Lab Chip, 5, 719, 2005. 11. Tokeshi, M., Minagawa, T. and Kitamori, T., Integration of a microextraction system on a glass chip: Ion-pair solvent extraction on Fe(II) with 4,7-diphenyl-1,10-phenanthrolinedisulfonic acid and tri-n-octylmethylammonium chloride, Anal. Chem., 72, 1711, 2000. 12. Hibara, A., Tokeshi, M., Uchiyama, K., Hisamoto, H. and Kitamori, T., Integrated multilayer flow system on a microchip, Anal. Sci., 17, 89, 2001. 13. Kuban, P., Berg, J. and Dasgupta, P. K., Vertically stratified flows in microchannels. Computational simulations and applications to solvent extraction and ion exchange, Anal. Chem., 75, 3549, 2003. 14. Sato, K., Tokeshi, M., Sawada, T. and Kitamori, T., Molecular transport between two phases in a microchannel, Anal. Sci., 16, 455 2000.

1034

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

15. Tokeshi, M., Minagawa, T. and Kitamori, T., Integration of a microextraction system: solvent extraction of Co-2-nitroso-5-dimethylaminophenol complex on a microchip, J. Chromatogr. A, 894, 19, 2000. 16. Kim, H.-B., Ueno, K., Chiba, M., Kogi, O. and Kitamura, N., Spatially-resolved fluorescence spectroscopic study on liquid/liquid extraction processes in polymer microchannels, Anal. Sci., 16, 871, 2000. 17. Hotokezaka, H., Tokeshi, M., Harada, M., Kitamori, T. and Ikeda, Y., Development of the innovative nuclide separation system for high-level radioactive waste using microchip-extraction behavior of metal ions from aqueous to organic phase in microchannel, Prog. Nucl. Energy, 47, 439, 2005. 18. Kitamori, T., Tokeshi, M., Hibara, A. and Sato, K., Thermal lens microscope and microchip chemistry, Anal. Chem., 76, 52A, 2004. 19. Ueno, K., Kim, H.-B. and Kitamura, N., Channel shape effects on the solution-flow characteristics and liquid/liquid extraction efficiency in polymer microchannel chips, Anal. Sci., 19, 391, 2003. 20. Maruyama, T., Kaji, T., Ohkawa, T., Sotowa, K., Matsushita, H., Kubota, F., Kamiya, N., Kusakabe, K. and Goto, M., Intermittent partition walls promote solvent extraction of metal ions in a microfluidic device, Analyst, 129, 1008, 2004. 21. Minagawa, T., Tokeshi, M. and Kitamori, T., Integration of a wet analysis system on a glass chip: determination of Co(II) as 2-nitroso-1-naphthol chelates by solvent extraction and thermal lens microscopy, Lab Chip, 1, 72, 2001. 22. Hisamoto, H., Horiuchi, T., Tokeshi, M., Hibara, A. and Kitamori, T., On-chip integration of neutral ionophore-based ion pair extraction reaction, Anal. Chem., 73, 1382, 2001. 23. Hisamoto, H., Horiuchi, T., Uchiyama, K., Tokeshi, M., Hibara, A. and Kitamori, T., On-chip integration of sequential ion-sensing system based on intermittent reagent pumping and formation of two-layer flow, Anal. Chem., 73, 5551, 2001. 24. Surmeian, M., Slyadnev, M. N., Hisamoto, H., Hibara, A. and Kitamori, T., Three-layer flow membrane system on a microchip for investigation of molecular transport, Anal. Chem., 74, 2014, 2002. 25. Maruyama, T., Matsushita, H., Uchida, J., Kubota, F., Kamiya, N. and Goto, M., Liquid membrane operations in a microfluidic device for selective separation of metal ions, Anal. Chem., 76, 4495, 2004. 26. Hibara, A., Nonaka, M., Hisamoto, H., Uchiyama, K., Kikutani, Y., Tokeshi, M. and Kitamori, T., Stabilization of liquid interface and control of two-phase confluence and separation in glass microchips by utilizing octadecylsilane modification of microchannels, Anal. Chem., 74, 1724, 2002. 27. Tokeshi, M., Minagawa, T., Uchiyama, K., Hibara, A., Sato, K., Hisamoto, H. and Kitamori, T., Continuous-flow chemical processing on a microchip by combining microunit operations and a multiphase flow network, Anal. Chem., 74, 1565, 2002. 28. Kikutani, Y., Ueno, M., Hisamoto, H., Tokeshi, M. and Kitamori, T., Continuous-flow chemical processing in three-dimensional microchannel network for on-chip integration of multiple reaction in a combinatorial mode, QSAR Comb. Sci., 24, 742, 2005. 29. Sato, K., Hibara, A., Tokeshi, M., Hisamoto, H. and Kitamori, T., Integration of chemical and biochemical analysis systems into a glass microchip, Anal. Sci., 19, 15, 2003. 30. Tokeshi, M., Kikutani, Y., Hibara, A., Sato, K., Hisamoto, H. and Kitamori, T., Chemical process on microchips for analysis, synthesis and bioassay, Electrophoresis, 23, 3583, 2003. 31. Kikutani, Y., Hisamoto, H., Tokeshi, M. and Kitamori, T., Micro wet analysis system using multi-phase laminar flows in three-dimensional microchannel network, Lab Chip, 4, 328, 2004. 32. Goto, M., Sato, K., Murakami, A., Tokeshi, M. and Kitamori, T., Development of a microchip-based bioassay system using cultured cells, Anal. Chem., 77, 2125, 2005. 33. Smirnova, A., Mawatari, K., Hibara, A., Proskurnin, M. A. and Kitamori, T., Micro-multiphase laminar flow for the extraction and detection of carbaryl derivative, Anal. Chim. Acta, 558, 69, 2006. 34. Hibara, A., Iwayama, S., Matsuoka, S., Ueno, M., Kikutani, Y., Tokeshi, M. and Kitamori, T., Surface modification method of microchannels for gas-liquid two phase flow in microchips, Anal Chem, 77, 943, 2005. 35. Honda, T., Miyazaki, M., Yamaguchi, Y., Nakamura, H. and Maeda, H., Integrated microreaction system for optical resolution of racemic amino acids, Lab Chip, 7, 366, 2007. 36. Miyaguchi, H., Tokeshi, M., Kikutani, Y., Hibara, A., Inoue, H. and Kitamori, T., Microchip-based liquid-liquid extraction for gas-chromatography analysis of amphetamine-type stimulants in urine, J. Chromatogr. A, 1129, 105, 2006.

Solvent Extraction on Chips

1035

37. Xiao, H., Liang, D., Liu, G., Guo, M., Xing, W. and Cheng, J., Initial study of two-phase laminar flow extraction chip for sample preparation for gas chromatography, Lab Chip, 6, 1067, 2006. 38. Bowden, S. A., Monaghan, P. B., Wilson, R., Parnell, J. and Cooper, J. M., The liquid–liquid diffusive extraction of hydrocarbons from a North Sea oil using a microfluidic format, Lab Chip, 6, 740, 2006. 39. Takahashi, Y., Sakai, R., Sakamoto, K., Yoshida, Y., Kitaoka, M. and Kitamori, T., On-line highthroughput ESIMS detection of a reaction product using synthesis and extraction microchips, J. Mass Spectrom. Soc. Jpn., 54, 19, 2006. 40. Aota, A., Nonaka, M., Hibara, A. and Kitamori, T., Countercurrent laminar microflow for highly efficient solvent extraction, Angew. Chem. Int. Ed., 46, 878, 2007. 41. Aota, A., Hibara, A. and Kitamori, T., Dependence of the number of theoretical plates of micro counter-current extraction of flow rate, Proc. µTAS 2005 Symposium, Boston, USA, 118, 2005.

Microdevices 36 Electrophoretic for Clinical Diagnostics Jerome P. Ferrance CONTENTS 36.1 36.2 36.3 36.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practical Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.1 DNA Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.1.1 Genotyping and Mutation Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.1.2 Other Nucleic Acid Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.2 Protein Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.2.1 Protein Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.2.2 Microchip Immunoassays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.3 Additional Clinically Relevant Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.4 Sample Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.4.1 Enzymatic Reactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.4.2 Purification and/or Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.5 Chip-Based Integration of Sample Preparation Steps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.5.1 DNA Amplification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4.5.2 Onboard Sample Preparation for Other Analytes. . . . . . . . . . . . . . . . . . . . . . . . 36.5 Method Development Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.6 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1037 1038 1038 1039 1039 1039 1041 1042 1043 1044 1046 1047 1049 1050 1051 1052 1054 1054 1056 1057

36.1 INTRODUCTION Microfluidic devices are changing the way analytical procedures are performed. The speed that they bring to an analysis, the small sample size and limited reagent use, integration of multiple processing steps into a single unit, as well as the possibility for on-site analysis, all point to changes in the way analytical methods are to be used in the relatively near future. Nowhere is this going to be more important than in the field of clinical analysis. Microfluidic devices in the doctor’s office, at the patient’s bedside, or at any point-of-care will be as common and easy to use as the blood glucose and home pregnancy testing devices currently available. In the operating room, where rapid analysis while a patient is still in surgery can provide information to the surgeon regarding the extent of damage, to the emergency room, where rapid feedback could identify agents that might cause an epidemic, or are the result of a terrorist attack, the clinical analysis methods provided by microchips will be invaluable. Developments in microfluidics methods directed toward the clinical laboratory have been progressing over the past few years, and fully integrated devices are now being reported. This chapter will review what has been achieved, and look to what is still needed to move this technology into the mainstream of clinical diagnostics. 1037

1038

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

36.2 BACKGROUND Not long after the early development of microchips for electrophoretic separations of DNA, the potential of these devices for clinical applications was realized.1−3 While microchip electrophoresis remains a major area in clinical analysis to which microchip processing is applied, implementation of additional processing and analysis steps on microchips over the past few years is also providing new devices designed to replace existing methods utilized in the clinical laboratory.4−6 The goal of many of these efforts is to eliminate the need for sample transportation to and processing in a central laboratory, by bringing the microfluidic technology directly to the point of care.7 A report sponsored by the National Cancer Institute discusses the development of point-of-care systems for cancer diagnosis and prognosis, in which integrated microfluidic systems play an essential role.8 The purpose of this chapter is to explore clinical analyses that have been transferred to an electrophoretic microchip, along with some of the microchip preparatory methods that are now being integrated with an electrophoretic analysis step to provide lab-on-a-chip type devices. There are a vast number of clinical analyses that are now being transferred to microdevices, but the scope of this chapter is limited to only those that have an electrophoretic component in accord with the nature of this volume. A number of clinical analysis methods simply take advantage of the microfluidic architecture in microchips to decrease sample size and reaction times without requiring a separation step. The most common example of these is microarrays, recently reviewed by Situma et al.,9 that utilize oligonucleotide to detect the presence of specific DNA sequences.10−14 Similar types of immunoassays for clinical detection of specific proteins have also been developed on microfluidic devices by immobilizing antibodies within microchannels;15,16 while these are not included in this chapter, a number of reviews17,18 as well as an earlier Chapter 34 in this book by Sato and Kitamori cover these types of microchip immunoassays with applications to clinical analyses in more detail. Protein immunoassays on microchips that incorporate a separation step are included in this chapter along with analyses for specific ions,19 glucose,20 pH,21 enzymatic activity,21,22 and other types of bioaffinity interactions.23 The use of mass spectroscopy in clinical analysis is increasing, but is not yet widely employed, thus electrophoretic microchips designed for interfacing with MS detection are also not included here; these devices are covered in Chapter 49 by Laurell for coupling with matrix-assisted laser desorption/ionization-mass spectrometry (MALDI-MS) detection, and Chapter 53 by Lazar for electrospray-MS. A number of microchip applications also relate to the trapping, growth, and analysis of cells on microdevices, but these are not methods utilized in clinical analysis and thus are not included. The reader is directed to a recent review of these techniques by El-Ali et al.24 as well as in Chapter 32 by Keenan and Beebe. Even with these limitations, the number of clinical methods that have been transferred to microchips is impressive and a comprehensive review is not possible. A number of reviews on microchip separations for clinical analyses have been written,25−29 the most recent of which reported on microchip separations utilized for detection of cancer, cardiovascular disease, renal disease, neurological disease, immune disorders, diabetes, hereditary disorders, thyroid functioning, and infectious agents in body fluids.29

36.3 THEORETICAL ASPECTS Microchip electrophoretic separations themselves and the theory behind them have been covered in Chapters 33 by DeVoe and Lee and 55 by Mahmoudian et al. This is also true of sample processing methods on microdevices, which have been covered in Chapter 43 by Bienvenue and Landers, Chapter 50 by van Midwoud and Verpoorte. The theory and engineering behind integration of multiple processes on a single device, in which both sample pretreatment and analysis steps are performed without removal of the sample from the device, are also explored in Chapter 43 by Bienvenue and Landers, as well as Chapter 40 by Easley and Landers describing valves to control flow within and between processes. This chapter seeks to show how the methods and theory discussed

Electrophoretic Microdevices for Clinical Diagnostics

1039

in these other chapters are being applied to real problems in clinical analysis. In addition, some of the issues involved in developing microdevices for utilization in the clinical laboratory not covered elsewhere will be described.

36.4 PRACTICAL APPLICATIONS 36.4.1 DNA ANALYSES The ability to use laser-induced fluorescence (LIF) for detection allowed DNA separations on microchips to advance more rapidly than separations of other types of clinically relevant compounds. Chapter 6 by Szántai and Guttman reviewed the use of capillary electrophoresis of DNA for clinical diagnostic purposes, and any of those methods can easily be transferred to the microchip platform. Most often, clinical microchip DNA diagnostic devices utilize separations of specific fragments of DNA amplified using the polymerase chain reaction (PCR). Nucleic acid targets of clinical interest for amplification include mutations associated with particular diseases, the presence of exogenous DNA or RNA from pathogens, or over expressed DNA indicating the presence of a single clonal variation. In the coming genetic medicine revolution, where treatment options are also based on particular DNA sequences within a patient’s genome, the use of microchip DNA separations will expand even further in the clinical laboratory.

36.4.1.1 Genotyping and Mutation Detection One of the most common methods for detection of mutations in PCR-amplified DNA fragments is single-stranded conformational polymorphism (SSCP) analysis. Tian et al.30 reported microchip separations for the detection of common mutations in BRCA1 and BRCA2, two breast cancer susceptibility genes specific for the Ashkenazi Jewish population. Separations were performed in 140 s, fourfold faster than the CE-based assay, with profiles from the wild-type and mutant alleles easily distinguishable. As was important in the development of all of these clinical microchip methods, no loss of diagnostic ability was seen in moving to the microfluidic analysis method. This method was further pursued by Kang et al.31 for rapid detection of a point mutation in the obesity gene. In addition to the SSCP technique, heteroduplex analysis (HDA) is a second method utilized for mutation analysis that has been transferred to microchips. Hestekin et al.32 used a combination of SSCP and HDA separations on a microchip to evaluate mutations in specific exons in the p53 gene, a gene for a transcription factor important in regulating the cell cycle and preventing tumor growth. Figure 36.1 shows microchip separations of fragments generated from samples of a wild type and an exon 8 mutant of the p53 gene. Tian et al.33 also utilized HDA on a microchip, combining it with allele-specific PCR to identify additional mutations in the BRCA genes as well as the PTEN tumor suppressor gene. The clinical potential for all of these electrophoretic methods based on migration differences due to the presence of mutations was recently reviewed by Hestekin and Barron.34 The mutations for which the above methods are most applicable are simple deletion and insertion mutations affecting only a few bases. Substitution mutations are difficult to monitor in this way,30 but represent a significant class of DNA mutations called single nucleotide polymorphisms (SNP), which will play a significant role in the area of personalized medicine. Woolley et al.35 performed SNP detection in HLA-H, a marker gene for hemochromatosis, on a 12-channel microchip using selective restriction digestion of a PCR-amplified product followed by traditional DNA sizing electrophoresis. Shi et al.36 utilized this same method on a 96-channel radial array microchip for SNP detection in the gene for methylenetetrahydrofolate reductase, the protein product of which regulates folate and methionine metabolism. This same 96-channel microchip design was used by Medintz et al.37 to evaluate a polymorphism in the HFE gene resulting in hereditary hemochromatosis using

1040

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a)

(b)

p56 Exon 8 Wild-type Temperature = ambient Rev ssDNA

1.8 dsDNA 1.5

0.5 0.4

0.9

RFU

RFU

1.2

0.6

Fwd ssDNA

0.3 380

400

420

440

Migration time (s)

0.2 0.1

0 360

0.3

460

480

0 405

p53 Exon 8 Wild-type + Mutant Codon 282, CGG to TGG Rev Temperature = ambient wt + mut HomessDNA duplexes Heteroduplexes Fwd mut Fwd ssDNA wt ssDNA

425

445

465

485

505

525

Migration time (s)

FIGURE 36.1 Electropherograms showing analysis of p53 exon 8 amplicons for mutation detection by tandem SSCP/HA. dsDNA peaks are identified by the overlap of the FAM (forward strand) and JOE (reverse strand) dyes. ssDNA peaks are identified by the predominance of one fluorescent dye. Peaks due to the mutation are indicated by the arrows. Separation conditions: ambient temperature, 0.1% PHEA dynamically coated channel, 350–450 V/cm applied electric field strengths. (Reprinted from Hestekin, C. N., et al., Electrophoresis, 27, 3823, 2006. With permission.)

allele-specific primers to generate different sized products. Figure 36.2 shows the separation and detection of both homozygote and heterozygote mutations in the S62C variants of the HFE gene on a microchip. The authors extended this work to simultaneously detect three SNP mutations within this same gene based on the sizes of the DNA fragments produced.38 Simple size-based electrophoretic separations in microchips can also be employed for insertion and deletion mutations, which modify the size of the PCR-amplified DNA fragments. Sung et al.39 utilized this method on a plastic microchip to evaluate fragile X alleles that have multiple repeat units of a trimeric CGG repeat. A tetranucleotide repeat in microsatellite alleles, related to hypercholesterolemia in families, was investigated by Cantafora et al.40 using the Agilent 2100 Bioanalyzer, a commercially available microchip electrophoresis instrument capable of analyzing up to 12 samples in 30 min. This same instrumentation was used in the evaluation of deletion and duplication mutations associated with Duchenne muscular dystropy in two papers by Ferrance et al.,41,42 which looked at a total of 13 loci in the dystrophin gene using two PCR amplifications and electrophoretic separations. A total of 50 samples were evaluated in the initial work, with all 35 samples with mutations being identified.41 Figure 36.3 shows the traditional Southern blot analysis versus the microchip separation for the detection of a particular deletion. The subsequent work went on to establish normal ranges for each exon product using 40 control samples, which were then used to correctly evaluate patient samples for the presence of mutations.42 The Bioanalyzer was also used in two studies by Sohni et al.43,44 that looked at both variable number of tandem repeat and deletion mutations in genes of clinical interest. The first study investigated polymorphisms in the cytokine genes IL-1RN and CCR5, both of which are important in immune system functioning.43 The second study investigated nucleotide repeats in the promoter region of the inducible NO2 synthase gene and the endothelial NO2 gene, both associated with complications in diabetes, and a deletion/insertion mutation in the angiotensin-converting enzyme gene associated with NO2 activity.44 In addition to the Agilent instrument, Hitachi also developed a commercial microchip electrophoresis instrument (SV 1210) both of which were utilized by Zhang et al.45 to evaluate gene fragments generated from 3 µL of blood in less than 20 min using a DNA extraction disk and a capillary PCR instrument.

Electrophoretic Microdevices for Clinical Diagnostics

1041

193A/A Homozygote sample

193A-> T Heterozygote sample (d) A-allele specific reaction

(a) A-allele specific reaction 211

400

136

268

211

(e) T-allele specific reaction Relative fluorescence intensity

136 400

211

211

(c) A- and T-allele reacitons 136

Relative fluorescence intensity

(b) T-allele specific reaciton

281

136 201 281

Red channel

Blue channel 211

Blue channel 201

268 281 136

400

60

70

80

90

Time (s)

400

(f) A- and T-allele specific reactions

Red channel

136

400

268

136

136

60

211 268 281

70

400

80

90

Time (s)

FIGURE 36.2 Detection of a heterozygous mutation (right panel) in the HFE gene using allele-specific PCR with two-color (red/blue) detection. (a and d) Separation of the A-allele specific amplicon (211 bp) from the 136 bp and 400 bp internal standards (red channel); (b and e) separation of T-allele specific amplicon (201 bp) from the same samples (blue channel); (c and f) separation of both amplicons showing only the A-allele fragment in the homozygous sample but both fragments in the heterozygous mutant sample. Numbers above the peaks indicate fragment length (bp). (Reprinted from Medintz, I., et al., Electrophoresis, 21, 2352, 2000. With permission.)

36.4.1.2 Other Nucleic Acid Analyses Mutation detection and genotyping represent a significant portion of clinical DNA analyses, but other DNA separations of clinical interest have also been transferred to microchips. In lymphoproliferative disorders, a single clonal variation is often overexpressed in T- or Bcells resulting in a loss of heterogeneity in the immune system. Munro et al.46 translated slab gel separations of PCR-amplified fragments, traditionally used to detect these disorders, to a microchip, effectively detecting clonal populations in the variable region of the T-cell receptor-γ gene and the immunoglobulin heavy chain gene. Again, no loss of diagnostic information was seen in the transition to the microchip separation system, in fact, a patient diagnosed as not having a lymphoma using the gel-based method was re-evaluated based on the microchip results and found to be in an early stage of the disease. Quantification of mRNA species has also been proposed as a clinical method, and reverse transcriptase-PCR (RT-PCR) products from renal biopsies have also been separated on microchips as a diagnostic method.47

1042

(a)

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

1 2

3

(b) - Exon 47 - Exon 52

Fluorescence

200

Exon 6

- Exon 51

Exon 47

- Exon 48 - Exon 50

Microchip separation Control Patient 22

100

0 - Exon 49

50

60

70

80

90

100

Seconds - Exon 48

FIGURE 36.3 (a) Southern blot results obtained using DNA digested with HindIII and probed with cDMD 8. Lane 1: normal male control; lane 2: normal female control; lane 3: female patient with a deletion in exon 47. (b) Microchip electropherograms of a multiplex amplification for a control and a female patient with a deletion in exon 47.

The presence of exogenous DNA or RNA in a patient is also of clinical interest, as this can be used to indicate the presence of a bacterial or viral pathogen. Höfgartner et al.48 used a singlechannel glass microchip for the detection of herpes simplex virus (HSV) in cerebrospinal fluid (CSF) through separation of PCR-amplified products. The microchip-based electrophoresis method allowed for rapid diagnosis of herpes simplex encephalitis (HSE), in less than 100 s, compared to 18 h for identical results obtained by hybridization. Chen et al.49 were able to detect viral RNA from the hepatitis C virus through RT-PCR amplification with product identification performed using an electrophoretic separation on a plastic microdevice. This group went on to further quantify the amount of viral DNA present using a competitive PCR procedure, with product analysis again on the plastic microchips.50 The Agilent Bioanalyzer also provides an efficient method for the separation of PCR products generated from exogenous DNA in CSF to show the presence of viral infections. Separations showing the presence of infections from a number of viral pathogens are shown in Figure 36.4. The Agilent instrument was also used to evaluate repetitive sequence PCR products from Mycobacterium species by Goldberg and coworkers.51 This kind of genotypic analysis can be used to identify particular strains, track tuberculosis transmission in outbreaks, and identify cross contamination in the laboratory.

36.4.2 PROTEIN ANALYSIS Electrophoretic analysis of proteins was one of the first methods transitioned to capillaries in clinical laboratories—this is highlighted in Chapter 2 by Hempe. Unfortunately, protein analysis has not made the same transition to microchips. This is due to the difficulties in implementing a ultraviolet (UV) absorbance method on microchips, the method normally utilized for proteins detection on CE. A number of fluorescent methods have been employed both for standard separations as well as for immunoassay analyses that utilize separations. Two-dimensional separations have also been performed on microchips,52 as reviewed in Chapter 33 by Lee, and these might have clinical significance at some point in the future.

Electrophoretic Microdevices for Clinical Diagnostics

1043

CMV samples HSV samples

250 200

200 Fluorescence

Fluorescence

250

150 100 50 0 30

150 100 50

40

50

60

70

80

90

100

110

0

30

40

Time (s) Blank

Control

50

60

70

80

90

100 110

Time (s) Patient 2

Patient 6

Blank

Control

Patient 9

Patient 20

FIGURE 36.4 Electropherograms of PCR-amplified cerebral spinal fluid samples from patients with suspected infectious agents. (a) Samples from patients possibly infected with HSV; clinical diagnosis by traditional means: patient 2—infected HSV type 1, patient 6—negative. (b) Samples from patients possibly infected with cytomegalovirus; clinical diagnosis by traditional means: patient 9—infected, patient 20—negative.

36.4.2.1 Protein Separations Human serum protein quantification is an important method for determining the relative abundances of albumin, IgG, and other important serum proteins. Colyer et al.53 investigated serum proteins through microchip separations of fluorescently labeled proteins, showing that quantification was possible even after labeling. Real samples, however, could not be evaluated because of poor sensitivity. Callewaert et al.54 also utilized a fluorescent labeling scheme to tag N-glycan components in serum, but was able to separate these real clinical samples on a microchip; a typical separation is shown in Figure 36.5, in which the difference between serum from a patient with cirrhosis and one with chronic hepatitis is illustrated. Separation of lipoproteins in serum on a microchip was reported by Weiller et al.55 who utilized NBD as a fluorescent label to allow detection of both highand low-density lipoproteins, important risk factors for cardiac disease. This group further refined the work to allow analysis of the low-density lipoprotein (LDL) components to show the presence of different forms of LDL in serum samples.56 A number of other covalent and dynamic fluorescent labeling schemes have now been proposed for the detection of proteins on microchips, all of which could be applied for clinical protein analysis.57−60 Bio-Rad has adopted an on-column dynamic labeling scheme in their commercial microchip electrophoresis instrument, which Chan and Herold61 utilized to evaluate microalbuminuria, a marker for cardiovascular disease and nephropathy in diabetic patients. Through the use of an SDS-based separation, these experiments can provide quantitative and qualitative information about the proteins. However, the use of covalent and dynamic labeling of native proteins must be evaluated for each system studied as fluorescent tags are not incorporated homogeneously across proteins, and are known to be influenced by the hydrophobicity and secondary structure of the proteins. This heterogeneity affects both the electrophoretic mobility and fluorescence response of each protein differently, and thus must be accounted for in the results obtained from any clinical protein separation. Giordano et al.60 developed a partially denaturing technique for fluorescent labeling, to minimize preferential binding, in which SDS was added to the sample, but no heat denaturation step was employed. Figure 36.6 shows the microchip protein profiling of human sera, using this method for the detection of gammopathies. The separations show clear spikes in the protein profiles of severely affected patient sera, but detection of more subtle changes was not evaluated. UV detection on microchips has not been completely bypassed, but it does require quartz microdevices, as the normal glass and plastic microchips absorb too much UV radiation in the wavelengths of interest. Zhuang et al.62 have developed quartz microchips for protein detection, and

1044

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

3

Chronic hepatitis

0.6 0.5 6 0.4

8

7

9

1 RFU

0.3

Cirrhosis

0.4

0.3

0.2

8

9

10

11

12

13

Time (min)

FIGURE 36.5 Profiling of two serum samples: one from a noncirrhotic chronic hepatitis patient and one from a patient with cirrhosis. Log (peak 7/peak 8) is diagnostic for liver cirrhosis, and these two peaks are well resolved in the microfluidics profiles. (Reprinted from Callewaert, N., et al., Electrophoresis, 25, 3128, 2004. With permission.)

have utilized them for electrophoretic analysis of urinary proteins of clinical interest.63 This group has also now applied quartz microchips to the analysis of serum lipoproteins64 without the need for fluorescent tagging as discussed above.55 UV detection is also covered in a review by Wang et al.65 who detail the use of electrophoretic microchips for the clinical analysis of hemoglobin, with various detection methods. 36.4.2.2 Microchip Immunoassays A number of microchip immunoassays have been developed that utilize an on-chip reaction followed by a separation component to separate free antigen from that bound to the antibody—if a labeled antigen is used—or to separate free antibody from the antigen/antibody complex—if labeled antibodies are used. Schmalzing et al.66 utilized the latter method to demonstrate an immunoassay for T4 in serum for clinical evaluation of thyroid function on fused-silica microchips. The immunoassay used a competitive format with labeled antigens competing against antigens in the serum for antibody-binding sites; the amount of free labeled antigen was used to quantify the amount of T4 in the samples. The same group also reported a microchip electrophoretic immunoassay for cortisol in serum using a similar method,67 which could quantify cortisol in serum within the clinical range without extraction from the sample or other preparation steps. Qiu and Harrison68 were able to incorporate a calibration step into their immunoassay microchip that will be important for eventual implementation in clinical analyses. Using a microchip with multiple inlets, they could control the amount of antibody mixed with labeled antigen, in this case bovine serum albumin (BSA). While this antigen is not clinically relevant, the ability to calibrate the microchip response, as shown in Figure 36.7, will be important for protein quantification. In addition to fluorescent-based assays, electrochemical detection has also been utilized for microchip immunoassays. Wang et al.69 initially utilized an alkaline phosphatase-labeled antibody,

Electrophoretic Microdevices for Clinical Diagnostics

(a)

1045

2.4 Albumin

2.2

RFU

2 1.8 1.6

β α

γ

1.4 1.2 1 200

RFU

(b)

(c)

2.4 2.3 2.1 1.9 1.7 1.8 1.4 1.1 0.8 0.7 0.4 200

250

300

350

400

350

400

β Spike

250

300

3.5 3 γ Spike 2.5 2 1.5 1 0.5 200

250

300

350

400

Seconds

FIGURE 36.6 Comparison of CZE analysis of partially denatured human sera using run buffer containing 0.04% NanoOrange dye. Sera were diluted 1:500 in 0.5% SDS, with 1% dye included in the sample buffer. Separation conditions: 100 mM borate, 3 mM diaminobutane, pH 8.5. (Reprinted from Giordano, B. C., et al., Anal. Chem., 76, 4705, 2004. With permission.)

which was mixed with the antigen (mouse IgG), then separated on-chip to resolve the free antibody from the complex. Using a post column reaction of the alkaline phosphatase to produce 4-aminophenol, they were able to monitor the products using electrochemical detection. This same group also used ferrocene tagged antibodies, simply measuring the reduction of the ferrocene in the free antibody and the complex using amperometric detection.70 Measurement of both IgG and T3 antigens were shown in this format. More recently, Herr et al.71 showed fluorescent microchip immunoassays for both tetanus antibody and tetanus toxin, utilizing fluorescently labeled tetanus toxin C-fragment and a fluorescent marker as an internal standard for quantification. The microchip separation provided a direct immunoassay for the presence of tetanus antibody, and a competitive

1046

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

Buffer

Ab Buffer Reactor

BSA*

Waste

Sep. channel Waste

500 AFU

Detector

Impurities Ab (µg/ml)

46.1

27.7

19.3 Complex BSA* 0

20

40

7.0 60

80

Time (s)

FIGURE 36.7 Traces from an on-chip calibration series. Intensity changes as a function of the amount of antiBSA delivered to the reaction, the complex peak increases in intensity while the labeled BSA peak decreases. At the highest antibody concentration impurities can be seen in the BSA peak region. (Reprinted from Qiu, C. X., and Harrison, D. J., Electrophoresis, 22, 3949, 2001. With permission.)

assay for the presence of tetanus toxin could be used to generate a dose-response curve as shown in Figure 36.8; the microchip design, however, did not include an integrated incubation step.

36.4.3 ADDITIONAL CLINICALLY RELEVANT ANALYSES In addition to DNA and proteins, a number of other important clinical markers have been evaluated using electrophoretic separations on microchips. Fanguy and Henry72 quantified uric acid in urine using a microchip electrophoretic separation with electrochemical detection. This work was extended to include three markers of renal function normally measured in clinical assays: uric acid, creatinine, and creatine.73 Detection of glucose has also been performed directly from human plasma on a microchip, with no need for sample pretreatment. Du et al.74 utilized a poly(dimethylsiloxane) (PDMS) microchip containing a copper electrode for electrocatalytic oxidation of the glucose as the detection method. Additional investigations involving the analysis of carbohydrates by microchip electrophoresis, a few of which are of interest for clinical analyses, are covered in a review by Suzuki and Honda.75 This review covers not only electrochemical detection methods but also reports on the use of fluorescent and UV absorbing tags; the authors suggest, however, that the need for a derivatization step eliminates the rapid separation advantage provided by the microchip analysis. A number of small molecules of interest in clinical analyses have also been analyzed using microchip-based separations. Zhao et al.76 utilized a microchip separation for detection on the antibiotic lincomycin directly from urine samples. This method, also employing electrochemical detection, allowed lincomycin analysis in under 40 s. Crevillen et al.77 utilized a similar method for

Electrophoretic Microdevices for Clinical Diagnostics

Normalized detector response

7

Normalized peak height

6 5 4 3

1047

4

2

(a)

1 (b) (c) 0

2

3

40 80 Separation time (s)

120

1 0 0

0.025

0.05

0.075

[TTC]2 µM

FIGURE 36.8 Dose–response curve for the tetanus toxin C-fragment using a microchip competitive immunoassay. Symbols show the normalized response (peak height of complex/peak height of free dye) with error bars indicating the standard error in the measurements (n = 3–5). A correlation coefficient of .95 was obtained for the linear fit through the points indicated. The inset shows electropherograms with TTC concentrations of (a) 2.0, (b) 7.8, and (c) 15.6 nM, with a labeled TTC* concentrations of 13.0 nM and 6.0 nM anti-TTC. Peaks 1 and 2 are due to free dye with peak 2 used for normalization; peak 3 is free TTC*, and peak 4 is the complex. Separation conditions: 300 V/cm, total length 6.1 cm, and length to detector 6 mm. (Reprinted from Herr, A. E., et al., Anal. Chem., 77, 585, 2005. With permission.)

separation and detection of hydroquinone and hydroquinone glycoside, important in clinical analyses because of their presence in pharmaceuticals and natural remedies. They were able to separate and detect these components using different injection methods, and also applied the microchip method to the analysis of urine samples. Figure 36.9 shows application of this method to a variety of sample types. Electrochemical detection was also utilized on microchips for the separation of six organic acids (fumaric, citric, succinic, pyruvic, acetic, and lactic acids) and three cations (K+ , Na+ , and Li+ ).78 This microchip used a four-electrode system with no direct contact between the electrodes and the separation channel, but was not tested with real samples. Munro et al.79 showed separation and detection of amino acids on microchips using an indirect fluorescence detection method. Figure 36.10 shows application of this method to urine samples with no pretreatment other than dilution in the appropriate separation buffer. Abnormal amounts of amino acids can easily be detected in the two patient samples compared to the healthy control sample. An absorbance detection based approach was utilized for the clinical analysis of calcium ion in serum, which is important in the regulation of a number of physiological processes.80 Beads with an immobilized calcium reactive dye were placed into the detection region, and the samples mobilized past the beads using electrophoretic flow. While a true separation was not intended, the interference of magnesium ions was significantly less than seen in pressure flow devices,19 since Mg2+ moves through the channel faster than Ca2+ allowing less interaction time with the beads.

36.4.4 SAMPLE PREPARATION There are a large number of biological sample processing steps that have now been translated to microchip methods and been shown to provide the same results, while utilizing smaller amounts of sample and less reagents, thus reducing both cost and waste generated. At the same time, the miniaturization of these processes has often produced a significant reduction in the time required,

1048

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(a)

(b)

2 10 nA

20 nA

1

1

2 b

b 1

2

a a

0

50

100

150

0

200

50

100

Time (s) (c)

150

200

Time (s) (d)

1

1 2 20 nA 2

20 nA

b

b

1 2

0

50

100 150 Time (s)

a 0

200

50

100

150

200

Time (s)

FIGURE 36.9 Electropherograms of (a) the nutraceutical HMP, (b) urine, (c) a cosmetic formulation, and (d) a wine sample. Peak labels correspond to (1) hydroquinone and (2) arbutin. The (a) traces are unspiked sample (b and d) or spiked with 150 ppm of (a) hydroquinone and (c) arbutin. The (b) traces are spiked with 150 ppm of hydroquinone and 125 ppm of arbutin. (Reprinted from Crevillen, A. G., et al., Anal. Chim. Acta, 562, 137, 2006. With permission.)

(a)

(b)

(c)

S T A.Q H

3 RFU

3 RFU

50

I.L

I.L

T

G S A. M. Q

G S N.T A.Q H.V

G

Microchip

70

90

110

50

5 RFU

70

90

110

50

70

90

110

Time (s)

FIGURE 36.10 Microchip electrophoretic separations of (a) normal and (b and c) abnormal urine samples. Urine samples were diluted 1:10 in water with 0.5 mM fluorescein, 1 mM sodium carbonate and 0.2 mM CTAOH. Separation conditions: 1.0 mM sodium carbonate, 0.5 mM fluorescein, and 0.2 mM CTAOH, pH 10.3; leff 5.5 cm, 15-s injection at 417 V/cm (reversed polarity), 183 V/cm separation voltage. Abbreviations: G, glycine; S, serine; N, asparagine; T, threonine; A, alanine; M, methionine; Q, glutamine; H, histidine; V, valine; L, leucine; I, isoleucine. (Reprinted from Munro, N. J., et al., Anal. Chem., 72, 2765, 2000. With permission.)

Electrophoretic Microdevices for Clinical Diagnostics

1049

by reducing diffusional lengths, thermocycling times, and column volumes. This chapter will not address all of these processes, but again will be limited to those processes directly related to clinical analysis and those that would be utilized in conjunction with a subsequent electrophoretic separation for analysis. Chapter 43 by Bienvenue and Landers, Chapter 50 by van Midwoud and Verpoorte in this volume address additional methods and provide more in-depth coverage of the microchip sample preparation literature. In addition, early work on sample preparation methods utilized for clinical molecular diagnostics that have been translated to microfluidic formats is covered in a review by Huang et al.81 These methods include cell sorting/selection, which can be used for clinical diagnostics but are not covered here, as well as nucleic acid purifications, and nucleic acid amplifications methods. 36.4.4.1 Enzymatic Reactions Enzymatic amplification of specific fragments of a nucleic acid template, particularly through the use of the PCR, is one of the most widely translated methods to microchips associated with a subsequent electrophoretic analysis. A number of microdevices and thermocycling methods have been described and are summarized in a recent review by Zhang et al.82 covering the basic designs and practical applications of microchip PCR. Not all of the microchip PCR developments showed direct clinical utility, and thus are not included here, but most of the methods could theoretically be utilized for clinical analysis. In addition to the traditional PCR-based DNA amplification, amplifications from RNA using RT to first generate a cDNA are also of clinical interest, as well as other enzymatic methods, which are not as widely utilized. Most clinically relevant microchip PCR amplifications have been utilized for the detection of infectious agents. Mikhailovich et al.83 showed allele-specific amplifications of mycobacterium tuberculosis to evaluate rifampicin resistance as a method for monitoring the use of this drug for treatment. They compared this method to a hybridization method, providing similar results but greatly decreasing the analysis time. Yang et al.84 utilized a plastic microchip to test the sensitivity and specificity of detection of a K12-specific fragment from Escherichia coli. They were able to detect as low as 10 E. coli cells with this method in the presence of 2% blood, and could do multiplex PCR to show specificity. For the detection of hepatitis B, a common infectious disease and blood borne pathogen, Cho et al.85 conducted a large-scale study of a microchip PCR method for the virus in real clinical samples. The authors utilized a real-time detection method, employing a miniature fluorescence detection system using light emitting diodes and photodiode detectors, to monitor amplification of sequences specific for the hepatitis B virus, thus a separation was not needed. The study, however, evaluated specificity, sensitivity, cross reactivity, reproducibility, and limits of detection in a large-scale clinical evaluation of the method (n = 563) to show the applicability of microchip methods for routine clinical analysis. Real-time detection was also used for an isothermal amplification procedure termed nucleic acid sequence-based amplification, for the detection of artificial human papilloma virus sequences in SiHa cells.86 This method was not applied to real samples, but demonstrates the clinical applicability of additional microchip enzymatic amplification methods. Not all clinical microchip PCR amplifications looked for infectious agents, however. Cheng et al.87 showed a microchip PCR method for multiplex amplifications of dystrophin gene fragments for the detection of mutations associated with muscular dystrophy. Other types of enzymatic reactions have also been implemented on microchips. Using PCRamplified DNA as the template, Lou et al.88 utilized ligase chain reaction to identify a mutation in the NOD2/CARD15 gene, a marker for inflammatory bowel disease, using DNA from blood of healthy volunteers. The reaction used the ligase enzyme to join short oligonucleotides sequences when the corresponding complementary wild-type or mutant DNA template was present. Digestion of DNA on a microchip using restriction enzymes has also been reported, but at this time, this has not yet been utilized for analyses relevant to clinical work.89 Likewise, digestions of proteins have been shown on microchips, but have not been applied for clinical analyses.90,91 A number of

1050

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

these enzymatic reactions and others have been transferred to microchips, but are covered within Section 36.4.5 below since they have been directly coupled to the subsequent electrophoretic separation on the same microchip. 36.4.4.2 Purification and/or Concentration A number of diverse methods fall under this category with respect to clinical analysis. One important process in clinical analyses is the isolation of plasma from whole blood, which has been performed on a microchip using a transverse flow microfilter design.92 Utilizing capillary action for the separation, both the plasma and the cell components of the blood were available on-chip for additional processing. Processing of blood to separate the leukocytes from other components has also been shown on a microchip for clinical applications requiring removal of erythrocytes.93 This microchip was designed to mimic intrinsic blood flow processes, such as leukocyte margination, to separate the two cell types and allow collection of the white cells. Another interesting technique applied to serum and urine samples is laminar flow diffusion, which takes advantage of the laminar flow and short diffusion lengths provided by microchips.94 By flowing the sample in parallel with a water stream, low molecular weight components are extracted into the water phase, which can then be collected in a separate stream from the original sample. While used only for spectroscopic detection in the original work, this could be implemented for the removal of small molecular weight components for analysis, such as urea and creatine from urine. Additional details on this process can be found in Chapter 35 on liquid/liquid extraction by Tokeshi and Kitamori. One of the most widely investigated purification methods on microchips is DNA extraction, which has been shown using a number of different phases and microchip designs, but only those methods that have been shown to deal with real samples such as whole blood are of clinical interest. Breadmore et al.95 capture DNA directly from blood utilizing silica beads immobilized in place with a silica sol–gel matrix, showing that it was possible to perform PCR amplification on the eluted genomic DNA. Chung et al.96 used beads with organic surface groups for capture of DNA on a poly(methyl methacrylate) (PMMA) microchip. The beads were immobilized to the surface of the PMMA through the organic surface groups, and the DNA containing solution passed back and forth over the immobilized beads. DNA from E. coli cells could be extracted and amplified from serum and whole blood, with the immobilized beads providing significantly better extraction efficiency than free beads as the number of E. coli in the blood sample decreased. Rather than packing beads into a microchannel, Wu et al.97 utilized the liquid precursor of the silica sol-gel to form a monolithic matrix in a microchannel for DNA extraction. This solid phase provided extraction efficiency with good reproducibility using standards, but extraction efficiency fell off quickly with whole blood when the devices were used repeatedly. This is not necessarily a problem, in that these devices for clinical applications would be designed for only a single use to prevent cross contamination of samples. Figure 36.11 shows the protein and DNA profiles generated on the sol-gel microchip during purification of DNA directly from lysed blood. While the sol-gel matrix was easier to generate in the channel, the liquid precursor filled all of the channels. Wen et al.98 solved this problem through the use of a photopolymerizable matrix, which they coated with a sol-gel precursor to increase the DNA-binding capacity. For whole blood application, they found that the proteins competed for binding sites, and significantly decreased the capacity of the monolith for DNA. A dual phase microchip was developed using an initial C18 phase to capture the proteins from the lysed blood as the DNA passed through for capture on the monolith. Volumes as large as 10 µL of blood could be processed on this device with good extraction efficiencies. In addition to proteins competing for binding sites in the DNA extraction process, removal of proteins from the solid phase is normally performed using an alcohol-based wash step to prevent protein coelution with the DNA and possible interference with the subsequent processing steps. As reported by Legendre et al.99 this wash step causes its own problems in that isopropanol inhibits PCR, which is often the next process on an integrated device. Wen et al.98 were able to eliminate this

Electrophoretic Microdevices for Clinical Diagnostics

Load

1051

Wash

Elution

Protein (µg)

55

Protein

2.6

Genomic DNA

2.2

45

1.8

35

1.4

25

1

15

0.6

5

0.2

–5 0

20

40

60

80

100

120

140

160

180

DNA (ng)

3 65

–0.2 200

Volume (µL)

FIGURE 36.11 Protein and DNA extraction profiles from a lysed human blood sample loaded onto a sol-gel filled microchip. Protein (closed squares) is removed in the wash step and DNA (open circles) is released during the elution step.

step using the dual phase method, but other approaches have also been investigated. Witek et al.100 utilized a photoactivated polycarbonate (PPC) microfluidic device for DNA extraction from E. coli, which incorporated an ethanol wash step, but the device could be dried before elution of the DNA to eliminate inhibition from the wash buffer. Larger elution volumes were required, however, and this work did not address DNA purification from human whole blood. An amine-coated microchip that had lower protein binding was used in work by Nakagawa et al.101 to extract DNA from whole blood. No wash step was required, avoiding the use of PCR-inhibiting reagents, but the amine phase showed only moderate efficiency. A competing approach was developed by Cao et al.102 using the protonated amine groups on chitosan at a low pH to capture the negatively charged DNA. The DNA was easily released at a high pH at which the amine groups became neutral. Very little protein was found to bind to the chitosan phase, thus a wash step was also not required in this method. Using the chitosan technology, a microchip capture device could be constructed simply by immobilizing the chitosan directly on the surface of a bifurcated channel structure within a microchip. While DNA represents a major focus of purification on microchips, capture of RNA has also been explored. Kokoris et al.103 utilized a silica membrane immobilized in a microchip for RNA purification. The “lab card” device provided RNA suitable for reverse transcriptase amplification with a yield and quality better than that obtained using traditional RNA extraction methods. A number of protein preconcentration or extraction microchip devices have also been developed, using either a general hydrophobic capture phase to capture all proteins, or an antibody or ligand-based phase for specific capture of a protein of interest. Most of these are designed for use with MS analysis of the collected products, however, and are covered in Chapter 53 by Lazar. One exception is the use of a phosphocholine ligand capture phase for selective capture of C-reactive protein (CRP), of interest as a cardiac and inflammation marker.104 This work required subsequent electrophoretic separation to quantify the CRP because of capture of additional phosphocholine-binding proteins in the extraction; quantitative results generated from the microchip extraction process agreed well with current CRP analysis method results.

36.4.5 CHIP-BASED INTEGRATION OF SAMPLE PREPARATION STEPS The ultimate motivation for transfer of electrophoretic separations from a capillary to a microchip was the ability to perform these additional processing steps in the microchip format, which was not as easily achieved in a capillary system. This provides the microchip the advantage of sequential integration of a number of processing steps directly with the electrophoretic analysis, and many

1052

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

of the integrated devices that resulted are of interest in clinical analyses.105 As with microchip electrophoresis itself, much of this work has focused on DNA for genetic analysis106 (see Chapter 43 by Bienvenue and Landers), but a number of integrated protein microchips have also been reported, along with multiprocess assays for small molecules. 36.4.5.1 DNA Amplification

1200

400

600

194

400

118

200

70

271 281 310

60

234

50

Salmonclla PCR product

0 Primer-dimer

Fluorescence (arbitrary units)

800

Salmonclla PCR product

Primer-dimer

A significant number of groups have reported on integration of the PCR process directly with the subsequent electrophoretic separation on the same microchip.87,107−115 As with the microchip PCR for clinical applications, most of the integrated microchip work focused on detection of infectious agents, starting in 1996 with Woolley et al.107 who showed amplification and electrophoresis of a fragment of salmonella DNA. Figure 36.12 shows the microchip separation of the amplified salmonella fragment along with a co-injection of the amplified product with a size standard for peak identification. Waters et al.108,109 incorporated cell lysis, and amplified and separated fragments from multiple regions of the E. coli genome in their integrated device. Bacterial detection of both E. coli 0157 and salmonella was performed by Koh et al.113 in a plastic integrated device fabricated in cyclic olefin copolymer (COC). Limits of detection as low as six copies were possible. Lagally et al.114 also evaluated an integrated microchip for pathogen detection, amplifying genes from both E. coli and Staphylococcus aureus. Limits of detection in this system were as low as two to three bacterial cells. Easley et al.115 developed an integrated microchip that utilized valves for pressure-based injections from the PCR chamber into the separation channel for the detection of salmonella DNA. This allowed dilution

0 40

50

60

70

80

Time (s)

FIGURE 36.12 Integrated PCR-electrophoresis microdevice assay for the detection of salmonella DNA. (a) Microchip separation of the PCR product was performed immediately following a 39-min amplification in the integrated microdevice. Primer-dimer (light gray peak) appears at 51 s and the PCR product (dark gray peak) appears at 61 s. (b) Sizing of the Salmonella product (1:100 dilution) in a separate microchip using a X174 HaeIII digest DNA standard (1 ng/µL). Total analysis time for the salmonella sample using the integrated microdevice was less than 45 min. (Reprinted from Woolley, A. T., et al., Anal. Chem., 68, 4081, 1996. With permission.)

Electrophoretic Microdevices for Clinical Diagnostics

1053

of the PCR product with water to provide a stacking effect as it was co-injected with a DNA size standard for accurate sizing. For the detection of a severe acute respiratory syndrome coronavirus, Zhou et al.116 used an integrated microchip to perform RT-PCR along with electrophoretic sizing. Nasopharyngeal swab samples were utilized in the analysis with the microchip method showing infection in 17 of 18 patient samples; this was better than the conventional RT-PCR method that identified only 12 infected patients. Pal et al.117 incorporated an additional reaction step in their integrated microchip for the identification of hemagglutinin A subtypes of influenza virus. Identification was based on restriction fragments, thus an on-chip enzymatic digestion was performed after the PCR. Separation of the digested products allowed successful discrimination of two influenza strains. In more recent work, Kaigala et al.118 used a PCR electrophoresis microchip to assess the risk of nephropathy in renal transplant patients. The analysis could distinguish between low, medium, and high BK viral loads, indicating patients at risk for complications; detection of as low as two viral copies was possible. Easley et al.119 went further, incorporating a DNA extraction step with the PCR and separation steps to show a fully-integrated analysis following off-chip cell lysis. They reported on analysis of mouse blood for the detection of Bacillus anthracis infections showing detection in asymptomatic mice from less than 1 µL of blood samples. A similar sized sample of nasal aspirate was used to show the presence of Bordetella pertussis in a patient confirmed to have whooping cough. Figure 36.13a shows a picture of the integrated microchip utilized in this work, Figure 36.13b shows the IR-mediated thermocycling profile on the device, and Figure 36.13c shows consecutive injections and separations of the amplified product for the B. anthracis analysis. For detection of mutations, Cheng et al.87 used an integrated PCR-microchip electrophoresis device for a multiplex amplification to identify deletions in the dystrophin gene, with successful amplification and resolution of all of the products. Taylor et al.120 detected mutations in mitochondrial DNA, which are indicated in a wide range of human disease states, utilizing a different set of on-chip processes. DNA was digested using a restriction enzyme, then denatured and reannealed to perform a HDA in the separation step for the identification of mutations. (a)

(b)

SPE-PCR-ME anthrax 1 thermocycling profile

Temp (c)

90 80 70 60 50 0

(c)

2

4

6 Time (min)

8

10

12

SPE-PCR-ME anthrax 1 7

Run 01_1

RFU

6 Run 01_2

5 4 3 2 60

120

180

240

Time (s)

FIGURE 36.13 (a) Integrated microchip used for the detection of B. anthracis infection in murine blood. The right channels are the solid phase extraction region; the dark filled chambers are the PCR area; the left channel is the separation channel, with the cross-t injection region covered by the pneumatic actuation lines used for an on-chip pumped injection. (b) Thermocycling profile for the microchip using IR-mediated PCR. (c) Electrophoretic analysis of the amplified product showing sequential pressure injections and separations on the integrated device to confirm the presence of the amplified fragment.

1054

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

36.4.5.2 Onboard Sample Preparation for Other Analytes Integrated microdevices have also been developed for protein analysis, most often using MS detection, as in the discovery of new biomarkers useful for clinical analysis.121 Gottschlich et al.122 showed early integration on a microchip, combining a protein digestion step directly with an electrophoretic separation of the products, but often only a preconcentration step is integrated with the on-chip separation. Membranes have been formed in a cross channel for concentrating proteins during the injection step using both laser patterning123 and photopolymerization.124 While these methods might be applicable to clinical analysis, no clinical sample was evaluated. Both concentration and selective protein removal was shown on a microchip that integrated two electrophoretic channels with builtin electrochemical detection.125 The first channel was used for sample preconcentration utilizing an isotachophoresis method, followed by a normal CZE separation; a myoglobin sample could be concentrated by greater than 60-fold in this system. Selective removal of proteins from a peptide sample was also shown in conjunction with desalting of the sample for subsequent analysis of the peptides. In addition to proteins, small molecules have also been analyzed on integrated microchips. Analyses of three markers of renal function, uric acid, creatinine, and creatine, along with p-aminohippuric acid were investigated using integrated microchips by Wang and Chatrathi.126 This work utilized fluid control to mix in creatinase and creatininase enzymes with the samples for on-chip enzymatic reactions before the electrophoresis. The unreacted acid species were separated from the neutral hydrogen peroxide, generated in the two enzymatic reactions, with detection using an amperometric method. To individually quantify the creatine and creatinine concentrations in a sample, the creatininase was simply eliminated from the reaction. A similar method was employed by Wang et al.127 for the measurement of uric acid integrated with measurements of ascorbic acid, glucose, and acetaminophen on a single device. This work utilized the integrated microchip to mix the sample with glucose oxidase for generating hydrogen peroxide from the glucose present. Again, electrophoretic separation from the acidic species was possible with amperometric detection used for quantification. Acetaminophen, another neutral species, which migrated with the hydrogen peroxide, was detected by analyzing the sample with and without glucose oxidase to determine the difference in detector response when the neutral components passed the detector.

36.5 METHOD DEVELOPMENT GUIDELINES Implementation of microchip methods for clinical analysis has thus far been almost exclusively confined to the research laboratory. This is changing with the availability and ease of use of the commercially available microchip electrophoresis instrumentation. These instruments provide both quantification and size determination of DNA or protein samples, thus could be immediately implemented into any clinical laboratory. Microdevices for sample preparation steps, at least for DNA processing, are currently under commercial development, but have not yet been introduced. Sample extractions and concentrations on microdevices are easily implemented in any laboratory, simply by placing the appropriate solid phase matrix into a microchip rather than using a suspension or large column, or centrifugal tube format. The only issue is fabrication or purchase of the microchips that have the appropriate sized channel and a constriction or frit for retention of the solid phase within the channel. Interfaces for the connection of syringe pumps to flow material through the microchip column have been described in the literature,95 and microchip reservoirs/connectors are also commercially available from IDEX® Corporation. In-house development of a microchip PCR amplification system is a bit more challenging than the DNA extraction system, but a noncontact IR heating design has been described in the literature.128 As an easier approach, simply putting the microfluidic device into a conventional thermocycler also works, and Legendre et al.99 showed that it was possible to integrate microchip DNA extraction with the PCR amplification using this technique. Integration of capture and separations, or enzymatic

Electrophoretic Microdevices for Clinical Diagnostics

1055

or labeling reactions with the electrophoretic separation, are also possible in many laboratories, again obtaining or fabricating the specifically designed microchips will be a significant part of the processes. Development of fully-integrated devices, however, is at this point beyond the realm of most laboratories that do not have experience with microfluidic design and utilization. In addition to simply designing and fabricating the devices and instrumentation to use microchip methods, a number of issues also have to be dealt with that are more specific for clinical application of these devices. The use of microfluidic devices with biological samples provides some important differences relative to the use of these devices for standards or “real samples” that have been cleaned up or spiked to greatly increase the concentration of the compound of interest. The proteins, lipids, and DNA mixture found in any biological sample have numerous charged species, as well as both hydrophobic and hydrophilic compounds. Thus, surfaces in microchips formed from almost any substrate will attract or interact with some component in the sample. In addition, there are issues with the sheer mass of material, such as proteins that can compete nonspecifically with an assay, the particulate matter left by cell lysis that can clog the microstructures, and the viscosity of biological samples that make it hard to pump material through the microdevice. The development of microchips for clinical applications then must look beyond the initial testing of the microchip assays using standards, and show that the assay, method, and microchip are rugged and reliable enough for use with real samples at the required concentrations, which is not always the case.2 Often, the primary issue when implementing on-chip clinical analyses is biofouling, where proteins or lipids bind to the microchip surface interfering with the assay being performed on the device; this is particularly an issue in microchips because of the large surface-to-volume ratio presented by these devices. One of the biggest clinical applications in which this is important is PCR amplification of DNA, in which binding of Taq polymerase and Mg2+ to the microchip surface can inhibit the reaction.129,130 A review by Kricka and Wilding131 detailed the various methods that had been utilized for passivating microchip surfaces to prevent PCR inhibition, and this group has shown increased efficiency for PCR through coating selection.132 Preventing reaction components from binding to the surface can also be accomplished by preventing the components in the aqueous solutions from reaching the surfaces by surrounding the aqueous phase with an organic phase such as a fluorocarbon oil, as reported by Roach et al.133 Of more interest is some recent work that looks at the modification of the microchip surfaces to resist nonspecific adsorption in polymeric devices, since these represent the future of inexpensive, disposable microchips that will be needed for clinical analyses. Bi et al.134 utilized an acrylate copolymer to modify the surface of PMMA devices to exhibit poly(ethylene glycol) (PEG) on the surface. PEG is known to prevent protein adsorption, and these microdevices were utilized in electrophoretic separations of proteins, with good theoretical efficiencies. Application of these devices directly to serum or plasma should be possible with no biofouling of the separation channel surface. This same group was also able to generate a phospholipid coating on the surface of PMMA microchip channels as a biomimetic surface to prevent protein binding.135 This coating was also evaluated for the electrophoretic separation of proteins, showing stable electrophoretic mobility, with little nonspecific adhesion of proteins or platelets when serum or plasma was utilized in the device. For PDMS devices, Wu et al.136 grafted hydrophilic polymers onto the surface to create a surface with a strongly suppressed electroosmotic flow (EOF), and significantly decreased adsorption of BSA and lysozyme than the native PDMS surface. These microchips were also utilized for electrophoretic separations of protein, both native and denatured, as well as peptides. In addition to adsorption, issues to be addressed for the routine use of microchips for clinical analysis include the development of lower cost, compact instrumentation for detection on microchips. Shrinivasan et al.137 detailed the development of a miniaturized LIF detector applicable for fluorescent detection of DNA on microchips, which could also be applied to proteins with the necessary fluorescent tags. By replacing the argon-ion laser with a diode laser, the cost and power requirements were significantly decreased, with only a small loss in sensitivity. Integration of fluid flow and mixing

1056

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

components into the microchips must also be addressed. Kim et al.138 describe a microchip for blood typing that incorporates flow splitting designs, micromixers, and microfilters directly fabricated into an injection molded COC device.139 This group also developed a COC microchip for partial oxygen, glucose, and lactate levels in blood that incorporates an on-chip power source for fluid manipulation. Vestad et al.140 addressed the need to reduce the associated hardware for flow control through microchips through the use of capillary pumping, allowing surface tension to pull fluid through their devices. Flow path can be controlled through the use of elastomeric valves, similar to those utilized in the integrated devices. Again, the goal is to eliminate the need for external power sources in the instrumentation for fluid flow control, to simplify the instrumentation and make it cost effective and compact enough for use in point-of-care applications.

36.6 CONCLUDING REMARKS There are a number of clinical areas in which microchip devices will be implemented over the next decade. For DNA, the switch will actually be away from the devices described here to a large extent as real time-PCR analysis becomes the predominant method for detecting and quantifying specific DNA sequences. Real-time PCR takes advantage of the amplification method while eliminating the separation step to reduce the complexity of the analysis and the microchip design. These devices will be implemented both for mutation detection and infectious agent analysis. For proteins, the separation will remain an integral part of the analysis, thus microchip separations will expand to form a larger part of the clinical diagnostic arena. Most of these devices will utilize MS detection, however, unless new breakthroughs in protein labeling or UV detection on microchips are seen. Electrophoretic microchips for small molecule detection will play some part in clinical analysis, but individualized instrumentation specific for a particular analyte or group of analytes may be the final format. For this to become reality, reductions in the costs of both the microdevices and the instrumentation will be needed to move these methods into routine clinical laboratory use. On the instrumentation side, replacing research designs with commercial grade instruments, with simple user interfaces and integrated components for sample and reagent loading, alignment, and thermal processing must be completed. Initial commercial microchip electrophoresis instruments show that this is possible, but more demand from the clinical community may be needed to show the potential for additional microchip instrumentation to commercial entities. The microchips themselves must also be cost effective for the analyses being performed, with disposable devices required to prevent contamination or false positive and false negative results. The cost of a PCR microchip, even a plastic and disposable one, will never be as inexpensive as a PCR tube, thus additional advantages must be provided by the microchip methods to justify their implementation. These can include reduced reagent consumption in the microchip, or incorporation of additional processing steps such as the DNA extraction, but some added benefit must be present in the microchips if they are to be accepted and used by the clinical community. In terms of integrating additional steps onto a microdevice, issues related to fluid flow control, either through on-chip or off-chip valving or flow metering, still need to be more fully addressed. Handling of the nanoliter volumes utilized in microchips presents additional challenges for instrumentation development, particularly with reagent addition and mixing with components already on a microdevice. Further work is also needed in the area of on-chip storage of reagents, possibly already lyophilized in the necessary compartments, or sealed off for storage but easily available on-chip when needed. Development of integrated microchips will continue, however, and these will begin to focus more on the actual end applications now that the concepts have been shown using standards and simulants. More specific clinical analyses will be translated to microchips over the next few years, and the complexity of the designs and instrumentation will determine how quickly they will be transitioned into the clinical laboratory.

Electrophoretic Microdevices for Clinical Diagnostics

1057

REFERENCES 1. Burtis, C. A., Converging technologies and their impact on the clinical laboratory, Clin. Chem., 42, 1735, 1996. 2. Colyer, C. L., Tang, T., Chiem, N., and Harrison, D. J. V., Clinical potential of microchip capillary electrophoresis systems, Electrophoresis, 18, 1733, 1997. 3. Kricka, L. J., Miniaturization of analytical systems, Clin. Chem., 44, 2008, 1998. 4. Tudos, A. J., Besselink, G. A. J., and Schasfoort, R. B. M., Trends in miniaturized total analysis systems for point-of-care testing in clinical chemistry, Lab Chip, 1, 83, 2001. 5. Schulte, T. H., Bardell, R. L., and Weigl, B. H., Microfluidic technologies in clinical diagnostics, Clin. Chim. Acta, 321, 1, 2002. 6. Verpoorte, E., Microfluidic chips for clinical and forensic analysis, Electrophoresis, 23, 677, 2002. 7. Pugia, M. J., Blankenstein, G., Peters, R. P., Profitt, J. A., Kadel, K., Willms, T., Sommer, R., Kuo, H. H., and Schulman, L. S., Microfluidic tool box as technology platform for hand-held diagnostics, Clin. Chem., 51, 1923, 2005. 8. Soper, S. A., Brown, K., Ellington, A., Frazier, B., Garcia-Manero, G., Gau, V., Gutman, S. I., et al., Point-of-care biosensor systems for cancer diagnostics/prognostics, Biosens. Bioelec., 21, 1932, 2006. 9. Situma, C., Hashimoto, M., and Soper, S. A., Merging microfluidics with microarray-based bioassays, Biomol. Eng., 23, 213, 2006. 10. Aytur, T., Foley, J., Anwar, M., Boser, B., Harris, E., and Beatty, P. R., A novel magnetic bead bioassay platform using a microchip-based sensor for infectious disease diagnosis, J. Immunol. Meth., 314, 21, 2006. 11. Soper, S. A., Hashimoto, M., Situma, C., Murphy, M. C., McCarley, R. L., Cheng, Y. W., and Barany, F., Fabrication of DNA microarrays onto polymer substrates using UV modification protocols with integration into microfluidic platforms for the sensing of low-abundant DNA point mutations, Methods, 37, 103, 2005. 12. Peytavi, R., Raymond, F. R., Gagne, D., Picard, F. J., Jia, G., Zoval, J., Madou, M., Boissinot, K., Boissinot, M., Bissonnette, L., Ouellette, M., and Bergeron, M. G., Microfluidic device for rapid (5 ng) can result in off-scale peaks, “pull-up” (color bleeding usually due to off-scale peaks), increased levels of stutter (small peaks typically one repeat unit shorter than the STR allele, believed to be due to slippage of the primer or template during replication) or other artifacts in the subsequent electrophoretic separation. Until recently, DNA quantification in forensic analyses has most commonly involved the use of slot-blot methods. In the past few years, the forensics community has shifted focus to the use of quantitative PCR as the preferred method for quantification. Both methods allow for human-specific DNA quantification, although qPCR has significantly lower limits of detection and gives an indication of the PCR-amplifiability of the DNA as well. Recently, multiplex qPCR assays have been developed to simultaneously determine the total amount of genomic and mitochondrial DNA,23−25 total genomic and male DNA,26,27 or assess the extent of DNA degradation for specific forensic applications.28,29 Because of the utility of qPCR to the forensic community, we focus on the development of qPCR on microdevices here. Although real-time PCR has been demonstrated on microdevices, no qPCR has been reported to date. That is, fluorescence detection during the accumulation of PCR products throughout cycling has been demonstrated. However, no standard curve was simultaneously amplified to allow for DNA quantification based on the signal generated. Real-time PCR has been demonstrated by Lin et al.30 in 25 µL reactions using fluorogenic probes and the SYBR Green I intercalating dye, respectively. In both cases, multiplex PCRs that include generation of a standard curve for quantification have not been described. In 2003, Quake and coworkers31 reported the development of picoliter volume PCR chambers on a microfluidic device for parallel real-time PCR assays. Although not demonstrated, the authors indicate that the eight flow channels can be used for the generation of a standard curve and, thus, the chip can be used for quantitative PCR. While DNA quantification remains the least-developed sample processing step on microchip for forensic DNA analysis, it is anticipated that extension of the reported methods to qPCR will be forthcoming. Any such qPCR method should be directly applicable to forensic DNA quantification by use of the appropriate primers and probes described in the literature for the macroscale counterpart. Following successful demonstration of qPCR on-chip, integration of this

1070

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

method with DNA extraction and DNA amplification would be the next microfluidic challenge; however, recent advances in this arena suggest this is possible.32−34 In addition, development of valving or other means of metering the appropriate volume of purified DNA into the amplification reaction must be carried out. This then results in the need for a storage mechanism on chip or off chip for the unused extracted DNA, which is derived from the original evidence sample. This material must be saved for subsequent reprocessing, permanent storage, or use by opposing legal council.

37.5 DNA AMPLIFICATION DNA amplification on microdevices has been demonstrated routinely, and Roper et al.35 have not only provided an extensive review of the literature on this topic (also see Chapter 43 by Bienvenue) but also provided a scoring method for device comparison. However, it is noteworthy that few examples of the complex PCR associated with forensic DNA typing have been demonstrated on chip. Forensic DNA typing most commonly involves the amplification of STRs via PCR using multiplex PCRs involving up to 16 primer pairs for coamplification of STRs from multiple loci. Primers are tagged with fluorescent tags (typically one of three or four fluorophores) for discrimination of loci. Commercially available kits have been developed, are the industry standard, and are used by every forensic DNA laboratory in the United States. Of primary interest to the forensics community in the development of a microfluidic device for PCR is the limit of detection for the amplification fragments (low-copy number amplifications), unbiased amplification, limiting the DNA template required for effective amplification, total reaction time, the ability to perform multiplex amplifications (as in the commercially available kits), and multiple chamber amplification (to amplify >1 sample at a time). Of course, any microdevice method must show that the PCR amplification product is comparable to that obtained on the macroscale, permitting any future results to be related to those previously obtained and are contained in the national DNA database. Microchip DNA amplification, to date, has already demonstrated the potential for low limits of detection, with amplification demonstrated from single-copy sources.36 The useful limit of detection (so as to prevent stochastic effects in forensic analyses) with microchip PCR is yet to be determined. However, the reduced reaction volume inherent to the microchip supports the expectation of lower limits of detection. Fast thermocycling has been demonstrated by a number of microchip PCR designs and methods with amplification rates approaching the biological limit of the processivity of Taq polymerase, that is, with 20–25 cycles in under 5 min.1,2 This fast thermocycling has not yet been applied to STR amplifications, although preliminary data have shown a reduction in total amplification time by almost half.37 Therefore, it is not unreasonable to anticipate that the approximately 3-h conventional thermocycling protocol will be reduced to 45 min or less. There are a few demonstrations of STR amplifications in microdevices. Liu et al.4 have demonstrated the amplification of a mini Y-STR multiplex, Bienvenue et al.5 and Legendre et al.37 have demonstrated amplification using the AmpFlSTR COfiler and Profiler Plus kits, and Schmidt et al.40 demonstrated microchip amplification using the PowerPlex 16 STR kit. Recently, amplification of mitochondrial DNA has been reported on a glass microchip.41 Following amplification, a subsequent mitochondrial sequencing reaction was performed. These data indicated amplification with template amounts as low as 1 pg, where product was not detected by the conventional method. Large masses of mitochondrial DNA (500 pg–1 ng) resulted in some nonspecific product formation when amplified in the low volume reaction. Table 37.1 summarizes the characteristics of these reported microchip forensic DNA amplifications. In addition to multiproduct amplification in each chamber (corresponding to the multiple loci), simultaneous amplification in multiple chambers will likely be needed to compete with the conventional thermocyclers. This has been shown by Waters et al.42 (four chambers), Legendre et al.37 (three reactions performed—although four chambers are fabricated in the device) (Figure 37.2), Lutz-Bonengel et al.41 (60 reaction spots), and Schmidt et al.40 (23 reactions

200 nL 1 µL 1.2 µL 1 µL

Legendre et al.37 Schmidt et al.40 Bienvenue et al.5 Lutz-Bonengel et al.41 100 min (28 cycles) 210 min (32 cycles) 180 mina (28 cycles) 180 minb (32 cycles)

64 min (35 cycles)

Reaction Time

Block (conventional) Block (conventional) Block (conventional) Block (conventional)

Resistive

Heating Method Mini Y-STR multiplex incl. Amelogenin COfiler, Profiler Plus PowerPlex 16 COfiler, Profiler Plus Mitochondrial

Amplification Demonstrated

Integrated? Yes, to microchip electrophoresis No No Yes, to DNA purification No

b This reaction time is reported for the PCR amplification only, not the sequencing step, which requires approximately an additional 220 min.

a This reaction time is reported for the PCR amplification and the DNA extraction in total.

160 nL

Reaction Volume

Liu et al.31

Author (Citation)

TABLE 37.1 Forensic DNA Amplifications on Microdevices

Yes, 4 Yes, 60 Yes, 4 Yes, 60

No

Multichamber?

Advances in Microfluidics: Development of a Forensic Integrated DNA Microchip (IDChip) 1071

1072

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

(b) COfiler®

(a)

(c)

(d)

Amel

TH01

D3S1358

TPOX D16S539 D7S820

100

150

200

250

300

CSF1PO 350

Fragment size (bp)

FIGURE 37.2 Multiple simultaneous STR amplifications on a single glass microdevice. (a) Photograph of the microdevice for multiple chamber PCR microdevice. The chip contains five chambers: one for temperature control (center) and four surrounding PCR chambers for parallel amplification. Amplification of the COfiler multiplex STR loci was performed simultaneously in three of the four reaction chambers on this microdevice using less than 1 ng template DNA. (Adapted and reproduced from Legendre, L. et al. Multiplex microchip PCR for STR Analysis. Poster presented at The 15th International Symposium on Human Identification, 2004. With permission.) The resulting STR profiles (following conventional DNA separation) are shown in panels B–D, indicating amplification in each chamber with no apparent PCR artifacts that would complicate analysis of the DNA profile. (Figures 37.2B through 37.2D are courtesy of L. Legendre.)

performed—although the chip contains 60 reaction spots) (Figure 37.3). Multiple PCR chambers per device will enable high-throughput processing but also allow analysts to simultaneously process the necessary controls (reagent blanks, positives, negatives). The application of microfabricated devices to the amplification of STR fragments in forensic DNA typing provides numerous advantages over the conventional methodology—specifically, lower limits of detection,36 less reagent and sample consumption, enhanced processing speed,1 and so forth. However, it also brings along issues that are not currently encountered with conventional methodology. For example, the STR amplifications on microchip discussed above involve total reaction volumes that range from 160 nL to 1.2 µL. Thus, the DNA added to the reaction must be contained in a fraction of a microliter, rather than the common 10–200 µL obtained from a Qiagen purification and then used for a conventional 25 µL amplification reaction. Consequently, the DNA must be concentrated to a greater extent than needed for conventional amplification. As a result, for microchip PCR to be utilized in forensic laboratories it will most likely require direct integration with microchip DNA extraction, where the DNA is routinely eluted in microliter volumes. This need to amplify the entire mass of DNA and, thus, concentrate the purified DNA to a high degree, is most critical in cases where low starting copies of DNA template are available. Bienvenue et al.5 have demonstrated DNA purification and PCR amplification on the same microdevice, carrying out microliter-scale DNA extraction as reported previously in a glass

Advances in Microfluidics: Development of a Forensic Integrated DNA Microchip (IDChip)

1073

(a)

(b)

FIGURE 37.3 (a) Photograph of a glass microchip containing 60 reaction spots, each with a hydrophilic surface surrounded by a hydrophobic ring. Each spot is loaded with 0.5 µL PCR mix and 0.5 µL of DNA template and covered with 5 µL oil to decrease evaporation and contamination. Thus, the total reaction volume on this device is 1 µL. The chip was thermocycled on conventional instrumentation using an in situ adapter. (Photo courtesy of Schmidt, U.) (b) A complete STR profile generated by PowerPlex 16 amplification of 32 pg genomic DNA on the chip shown in (a). Allele names are identified in the gray boxes above each set of peaks. Repeat numbers for the corresponding alleles are indicated below each peak. (Adapted and reproduced from Schmidt, U. et al., Int J Legal Med 2006, 120, 42–48. Copyright 2005. With permission from Springer-Verlag.)

1074

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques

microdevice.17,22,43 However, in a unique approach, the DNA was eluted from the solid phase using PCR master mix and retained in a chamber. Subsequent placement of the device into a conventional (block) thermocycler allowed for amplification of the STR fragments in the Cofiler and Profiler Plus kits. These results are significant as they represent the first work demonstrating the potential for interfacing microfluidic DNA extraction technology with conventional PCR instrumentation. Of course, for application to forensic casework, the integration of a DNA quantification step will also be necessary.

37.6 SEPARATION OF PCR PRODUCTS As described in detail in Chapter 25 by McCord, forensic DNA profiling is routinely accomplished by capillary electrophoresis (CE) in forensic laboratories. Recently, microchip electrophoresis is being evaluated for use in casework by the Virginia Department of Forensic Science and the Forensic Science Service (U.K.). While DNA separations on microdevices are becoming more commonplace, separations for forensic DNA analysis have been less-frequently demonstrated. DNA separations for forensic analyses, like DNA sequencing, require single base pair resolution and multicolor fluorescence detection (see Chapters 6, 27, and 44). In addition, the STR separation must be rapid (less than 30 min) to compete with the current capillary-based method. For nondisposable, single-process devices, the sieving matrix must be of relatively low viscosity to enable loading and replacement of the matrix after each separation (as in the capillary) to prevent carryover. Table 37.2 provides some of the pertinent details associated with the STR separations carried out on microdevices to date. All of these STR separations have utilized long-read polyacrylamide (LPA) or polydimethylacrylamide (PDMA) [the main component of the commercially available polyolefin plastomer polymers, known as “POP” polymers]; POP-4 (where the number dictates the %PDMA) is the polymer used for capillary STR separations on ABI instruments. Acceptable polymers for these separations must be capable of single base pair resolution at low temperatures (preferably, room temperature) with short effective separation length (Leff – distance from injection point to detection point) and of relatively low viscosity to enable replacement and ease of filling the microchannel. Overall, it appears that LPA is capable of providing the requisite resolution with microchannels having an Leff of 11.5 cm,44 although shorter distances are unlikely to be adequate with the current methodology.45 In general, the separations require 25 or more minutes,46−48 which is comparable to that obtained on the conventional capillary instrumentation. One embodiment46,48 , however, utilizes a 96-microchannel device (Figure 37.4) and can accomplish 96 simultaneous separations, resulting in greater throughput capabilities than the commonly encountered 4- and 16-capillary conventional CE instruments. An example of data obtained on this microdevice is shown in Figure 37.5. POP-4 has also been used for the separation of STR fragments on microdevices, where the separation of an Identifiler allelic ladder was accomplished with an Leff of only 8 cm and in under 12 min.49 This separation is distinct in that the optical requirements for the detection system are relatively simple—containing only a laser and a single photomultiplier tube (PMT), because of the electronic filtering capabilities of an inline acousto-optic tunable filter (AOTF) to select the desired emission wavelengths sequentially (see Chapter 45 by Karlinsey). The value of the AOTF stems from the ease with which four-color detection can be expanded to five or more colors, which may be optimal for portable systems. One of the benefits of microchip electrophoresis reported by Goedecke et al.47 (Figure 37.6) was the improved data quality and stability when compared to the commercially available capillary array instruments. This advantage arises from the enhanced thermal and mechanical stability of the separation channel, as well as the improved heat-sinking inherent to glass microchips. As mentioned earlier, devices similar to those described in Schmalzing et al.50 and Yeung et al.48 are currently undergoing evaluation in the United States and the United Kingdom. It is anticipated

Medintz et al.53

Glass (borofloat)

5.5

Glass (fused silica)

Desalted (Qiagen), resuspended in 0.5 × TE, diluted in formamide, heat denatured, snap cooled

11.5

Glass (fused silica)

Diluted in buffer, heat denatured, snap cooled Diluted in H2 O, heat denatured, snap cooled

Schmalzing et al.52

Schmalzing et al.50

2.6

Microchip Substrate

Sample Treatment

Separation Channel (Effective Length, cm)

Author (Citation)

TABLE 37.2 STR Separations on Microdevices

4% LPA in one time TBE/3.5 M urea/30% formamide, 50◦ C Long-read LPA (Amersham), 40◦ C

4% LPA in one time TBE/3.5 M urea/30% formamide

Sieving Matrix

Modified Hjerten

Modified Hjerten

Modified Hjerten

Channel Coating

CTTv

(ET dyes)

a

CTTv

STR Kit Utilized

200 V/cm

200 V/cm

Separation Field Strength