Producing Biomolecular Substances with Fermenters, Bioreactors, and Biomolecular Synthesizers

PH2270_C000.fm Page ii Tuesday, May 23, 2006 10:00 AM

William L. Hochfeld Ph.D., C.Ch.E., B.C.L.D. (A.B.B.), F.A.I.C., F.A.C.B.

PH2270_Discl.fm Page 1 Friday, September 30, 2005 2:37 PM

Published in 2006 by CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2006 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group No claim to original U.S. Government works Printed in the United States of America on acid-free paper 10 9 8 7 6 5 4 3 2 1 International Standard Book Number-10: 0-8493-2270-7 (Hardcover) International Standard Book Number-13: 978-0-8493-2270-9 (Hardcover) Library of Congress Card Number 2005050643 This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials or for the consequences of their use. No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC) 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data Hochfeld, William L. Producing biomolecular substances with fermenters, bioreactors, and biomolecular synthesizers / William L. Hochfeld. p. cm. Includes bibliographical references and index. ISBN 0-8493-2270-7 1. Biomolecules--Synthesis. 2. Bioreactors. I. Title. TP248.25.B55H67 2005 660.6'3--dc22

2005050643

Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com Taylor & Francis Group is the Academic Division of Informa plc.

and the CRC Press Web site at http://www.crcpress.com

PH2270_C000.fm Page v Tuesday, May 23, 2006 10:00 AM

Preface Producing Biomolecular Substances with Fermenters, Bioreactors, and Biomolecular Synthesizers provides a discussion of the substance, character, makeup, and quality of the materials used in the production and downstream processing of biomolecular substances: raw materials, reagents, intermediates, and consumables. In addition to a Technology Overview and Introduction to the Bioprocess, it contains information about: • • • • • •

Recombinant DNA Materials DNA Synthesis Reagents Enzymes DNA Amplification Reagents Cell Growth Media Peptides and Synthesis Reagents

This book is designed for biochemical and biopharmaceutical engineers, technicians, scientists new to biotechnology, production, QA personnel, and managers. Fully equipped with a biotechnology glossary, numerous references, and an exceptionally extensive index, it emphasizes practical techniques, methods, and applications.

PH2270_C000.fm Page vi Tuesday, May 23, 2006 10:00 AM

PH2270_C000.fm Page vii Tuesday, May 23, 2006 10:00 AM

Acknowledgments I would like to dedicate this book to the late Dr. C. Thomas Dean of blessed memory, Dean Emeritus at California State University, Long Beach — my doctoral advisor, mentor, and long-time friend — for his support and inspiration over the years. Many thanks to Dr. Austin Reed, who after an arduous journey “across the pond,” read my book and subsequently wrote a gracious review. Many thanks to the staff of the Howard Street Veterans’ Center in Evanston, Illinois, for valuable support and to Ron Frym for encouragement when it was really, really needed. Thanks a bunch for the helpful advice and support of Marsha Hecht, Helena Redshaw, Jessica Vakili, Steve Zollo, and Harvey Kane of Taylor & Francis. My appreciation also goes to Northeastern Illinois University for the use of its research and library facilities. And finally, appreciation to Microsoft Word™ and my trusty Dell© IBM® clone for secretarial support. William L. Hochfeld, Ph.D., C.Ch.E., B.C.L.D. (A.B.B.), F.A.I.C., F.A.C.B.

PH2270_C000.fm Page viii Tuesday, May 23, 2006 10:00 AM

PH2270_C000.fm Page ix Tuesday, May 23, 2006 11:11 AM

Author Arthur’s Photography, Des Plaines, IL

Dr. William L. Hochfeld is an internationally recognized biochemist, immunologist, molecular biologist, chemical engineer, prolific writer, and decorated Vietnam veteran, attaining the rank of Lt. Colonel, in the Medical Service Corps, of the U.S. Army Reserve. An acknowledged authority on communicable disease, biopharmaceutical research, development, and regulatory affairs, during part of his reserve service Dr. Hochfeld was attached to the United States Army Medical Research Institute for Infectious Diseases (USAMRIID), and also served as CEO of Biotechnical Consulting Ltd/Biotech Laboratories, Ltd., while employing creative approaches to solving complex biological problems and producing seminal contributions to epidemiology, clinical immunology, bioprocess engineering, and vaccine development. Dr. Hochfeld holds national board certifications in biochemistry, chemical engineering, and as a bioanalyst clinical laboratory director. He served as a consultant to the European research team that developed a yellow fever vaccine in transgenic eggs, which enabled the sponsoring pharmaceutical company to offer costeffective, yet comprehensive, yellow fever vaccination programs to third-world countries. This initiative earned the team a nomination for the Nobel Prize in Medicine. Dr. Hochfeld’s observation that Gulf War Syndrome was caused by the administration of mycoplasma-contaminated anthrax vaccine led him to be one of the first to propose punitive action by the Defense Department against the vaccine manufacturer and its corporate officers. Dr. Hochfeld was selected from a list of preeminent U.S. scientists and engineers to lead a validation engineering team to coordinate validations of three vaccine manufacturing plants scheduled for construction in early 2006 and an M5B1 “Bird-Flu” vaccine plant scheduled for construction later that year.

PH2270_C000.fm Page x Tuesday, May 23, 2006 10:00 AM

PH2270_bookTOC.fm Page xi Tuesday, May 23, 2006 10:01 AM

Contents Chapter 1

Introduction ..........................................................................................1

Biotechnology ............................................................................................................1 Bioproduction ............................................................................................................1 The Bioprocess ..........................................................................................................5 Cell Lines...................................................................................................................5 Cell-Culture Process ..................................................................................................6 Separation, Recovery, Purification ............................................................................7 Virus and Foreign-DNA Removal.............................................................................7 Quality Assurance......................................................................................................8 Chapter 2

Technology Overview ..........................................................................9

Biomolecular Foundations.........................................................................................9 Cellular Variation .......................................................................................................9 Key Molecular Interactions .....................................................................................10 Base-Pair Complementarity.....................................................................................10 Genetic Coding ........................................................................................................11 Interrupting Gene Segments ....................................................................................12 Sequence Determinations ........................................................................................13 Cloning.....................................................................................................................13 Applications ...................................................................................................15 Screening and Selection.................................................................................15 DNA Synthesis ........................................................................................................16 Cell Lines.................................................................................................................19 Expression Systems .................................................................................................19 Vector Construction........................................................................................20 Prokaryotic Cells............................................................................................20 Eukaryotic Cells.............................................................................................21 Expression Levels ..........................................................................................22 Intra- vs. Extracellular Expression ................................................................23 Glycosylation .................................................................................................23 General Considerations ..................................................................................24 Further Posttranslational Modifications.........................................................24 Economic Concerns .......................................................................................24 Regulatory Concerns......................................................................................25 Chapter 3

Introduction to the Bioprocess...........................................................27

Overview..................................................................................................................27

PH2270_bookTOC.fm Page xii Tuesday, May 23, 2006 10:01 AM

Optimizing Product Yield ........................................................................................27 Kinetic Models ........................................................................................................28 cGMP Standards ......................................................................................................28 Contract Manufacturing...........................................................................................28 Complex Mixtures ...................................................................................................29 Biomolecular Products ............................................................................................29 Joint Manufacturing.................................................................................................29 Downscaling to Scale Up ........................................................................................30 Large-Scale Bioprocesses ........................................................................................31 Scaling-Up In-House or in a Contract Facility .......................................................31 Lab-to-Pilot-Plant-to-Production .............................................................................31 Increasing Bioprocess Scale ....................................................................................32 Differing Features at Larger Scale ..........................................................................32 Simulating Environment at Scale ............................................................................32 Detailed Specifications ............................................................................................33 Impact on Downstream Bioprocessing ...................................................................33 FDA Biologic Review Process ................................................................................34 Regulating Biological Products...............................................................................35 FDA History and Agency Structure ..............................................................35 The Biological License Application (BLA) ..................................................36 Responsibility for Therapeutic Biologics ......................................................37 Chapter 4

Recombinant DNA Materials and Methods ......................................39

Cloning.....................................................................................................................39 Applications .............................................................................................................41 Screening and Selection ..........................................................................................43 DNA Synthesis ........................................................................................................43 Cell Lines.................................................................................................................47 Expression Systems .................................................................................................48 Vector Construction .................................................................................................48 Prokaryotic and Eukaryotic Growth Characteristics...............................................49 Bacteria...........................................................................................................49 Yeasts..............................................................................................................49 Eukaryotic Systems (Insect and Mammalian Cells) .....................................50 Expression Levels ....................................................................................................51 Intra- vs. Extracellular Expression..........................................................................52 Intracellular Expression ...........................................................................................52 Extracellular Expression..........................................................................................53 Glycosylation ...........................................................................................................54 Mammalian Cell Glycosylation...............................................................................54 Insect Cell Glycosylation ........................................................................................55 General Considerations............................................................................................55 Further Posttranslational Modifications ..................................................................55 Economic Concerns .................................................................................................56 Regulatory Concerns ...............................................................................................58

PH2270_bookTOC.fm Page xiii Tuesday, May 23, 2006 10:01 AM

Manufacturers’ Directory.........................................................................................59 Cloning Vectors, Bacterial Expression Systems............................................59 Cloning Vectors, Yeast Expression Systems..................................................60 Cloning Vectors, Mammalian Expression Systems.......................................60 Cloning Vectors, Insect Expression Systems.................................................60 Prokaryotic Cell Lines ...................................................................................60 Eukaryotic Cell Lines ....................................................................................61 Chapter 5

Enzymes .............................................................................................63

Overview..................................................................................................................63 Enzyme Protein Engineering...................................................................................67 Transformation in Nonaqueous Systems.................................................................68 Restriction Enzymes ................................................................................................69 Restriction-Enzyme Classes ....................................................................................69 Class I Restriction Enzymes ..........................................................................70 Class II Restriction Enzymes.........................................................................70 Class III Restriction Enzymes .......................................................................70 Restriction-Enzyme Specificity ...............................................................................70 Star Activity .............................................................................................................70 Enzyme Families and Compatible Ends .................................................................71 Isoschizomers...........................................................................................................71 Reverse Transcriptases.............................................................................................71 DNA and RNA Nucleases .......................................................................................71 Nucleases........................................................................................................71 DNase I.................................................................................................71 RNase-free DNase I .............................................................................72 Exonuclease III.....................................................................................72 Endonuclease from Neurospora crassa ...............................................72 Nuclease Bal 31....................................................................................72 Mung Bean Nuclease ...........................................................................72 Nuclease P1 ..........................................................................................72 Nuclease S7 ..........................................................................................72 Nuclease S1 ..........................................................................................73 Uracil-DNA Glycosylase......................................................................73 Ribonucleases.................................................................................................73 DNase-free RNase ................................................................................73 RNase A................................................................................................73 RNase CL3 ...........................................................................................74 RNase H................................................................................................74 RNase T ................................................................................................74 RNase U2..............................................................................................74 Proteases.........................................................................................................74 Pronase..................................................................................................74 Proteinase K..........................................................................................74 Recognition Sequence of Factor Xa ....................................................75

PH2270_bookTOC.fm Page xiv Tuesday, May 23, 2006 10:01 AM

Sequencing-Grade Proteases ................................................................75 Modifying Enzymes .............................................................................76 DNA Phosphorylation and Dephosphorylation .............................................76 Alkaline Phosphatase............................................................................76 T4 Polynucleotide Kinase ....................................................................76 DNA Ligases ..................................................................................................76 E. coli DNA Ligase ..............................................................................76 T4 DNA Ligase ....................................................................................77 DNA and RNA Polymerases .........................................................................77 Taq DNA Polymerase...........................................................................77 E. coli DNA Polymerase I ...................................................................77 Labeling-Grade Klenow Enzyme.........................................................77 Sequencing-Grade Klenow Enzyme ....................................................77 T4 DNA Polymerase ............................................................................78 Exonuclease III from E. coli................................................................78 T4 Gene 32 Enzyme.............................................................................78 Terminal Transferase ............................................................................78 E. coli RNA Polymerase ......................................................................78 RNA Polymerases (SP6, T3, T7) .........................................................78 Polynucleotide Phosphorylase..............................................................79 RNase Inhibitor ....................................................................................79 Phosphodiesterases.........................................................................................79 Phosphodiesterase from Calf Spleen ...................................................79 Phosphodiesterase from Snake Venom ................................................80 Miscellaneous Modifying Enzymes...............................................................80 Methylase Hpa II..................................................................................80 Protoplast-Forming Enzyme.................................................................80 Manufacturers’ Directory.........................................................................................80 Endonucleases, Restriction, DNA..................................................................80 Reverse Transcriptases ...................................................................................81 Nucleases........................................................................................................81 Ribonucleases.................................................................................................82 Proteases.........................................................................................................82 Modifying Enzymes, DNA ............................................................................83 Chapter 6

DNA-Amplification Reagents ............................................................85

Overview..................................................................................................................85 Amplification by DNA Synthesis (PCR) ................................................................85 Amplification by RNA Transcription (TAS and 3SR)............................................87 Amplification by Ligation (LAR, LCR, LAS)........................................................90 Amplification by RNA Replication (Qβ Replicase) ...............................................92 Sequence-Based Amplification................................................................................94 Manufacturers’ Directory.........................................................................................95 DNA-Amplification Reagents ........................................................................95

PH2270_bookTOC.fm Page xv Tuesday, May 23, 2006 10:01 AM

Chapter 7

Cell-Culture Media ............................................................................97

Overview..................................................................................................................97 Media .......................................................................................................................97 Prokaryotic Cell-Culture Medium .................................................................97 Eukaryotic Cell-Culture Medium ..................................................................97 Defined Medium ............................................................................................98 Serum .......................................................................................................................98 Bovine Serum...............................................................................................100 Other Sera ....................................................................................................101 Serum Contamination ..................................................................................101 Serum-Free (Defined) Medium .............................................................................102 Cell-Culture Gels ...................................................................................................103 Media Supplements ...............................................................................................104 Amino Acids ................................................................................................104 Antibiotics ....................................................................................................104 Amphotericin-B............................................................................................105 Ampicillin.....................................................................................................105 Tylosin Solution ...........................................................................................105 Attachment Factors................................................................................................105 Buffers....................................................................................................................105 Sodium Bicarbonate.....................................................................................106 Hanks’ Balanced Salt Solution ....................................................................107 Earle’s Balanced Salt Solution ....................................................................107 L-15 Medium ...............................................................................................107 HEPES Buffer ..............................................................................................107 Growth Factors ......................................................................................................109 Epidermal Growth Factor (EGF) .................................................................109 Fibroblast Growth Factors (FGF) ................................................................110 Acidic FGF .........................................................................................110 Basic FGF...........................................................................................110 Other Growth Factors ..................................................................................110 Lectins....................................................................................................................110 Toxins.....................................................................................................................111 Transport Factors ...................................................................................................111 Vitamins .................................................................................................................111 Water ......................................................................................................................112 Manufacturers’ Directory.......................................................................................112 Media, Cell Culture, Basic, Liquid .............................................................112 Media, Cell Culture, Basic, Powdered ........................................................112 Media, Cell Culture, Conditioned ...............................................................113 Media, Culture, Recombinant-Bacterial ......................................................113 Media, Sera ..................................................................................................114 Media, Amino Acids ....................................................................................114 Media, Antibiotics........................................................................................115

PH2270_bookTOC.fm Page xvi Tuesday, May 23, 2006 10:01 AM

Media, Cell-Attachment Factors..................................................................115 Media, Matrices & Microparticles...............................................................116 Media, Balanced Salt Solution ....................................................................116 Media, HEPES Buffer..................................................................................117 Cell-Culture Media, Growth Factors ...........................................................117 Cell-Culture Media, Lectins ........................................................................118 Cell-Culture Media, Toxins .........................................................................118 Cell-Culture Media, Transport Factors........................................................119 Cell-Culture Media, Vitamin Supplements..................................................119 Cell-Culture Media, Water for Injection .....................................................119 Chapter 8

Fermenters ........................................................................................121

Introduction to Fermentation.................................................................................121 Historical Perspective ............................................................................................121 1700s: Dawn of the Scientific Approach to Fermentation..........................121 Early 1800s: Pasteur ....................................................................................121 Mid-1800s–1858: Virchow, Schwann, and Flemming ................................122 1916–1918 WWI: Impetus for the Development of Industrial Fermentation ..............................................................................................122 The Roaring Twenties: Fermentation Spreads; Most Solvents Are Fermentation Products ...............................................................................122 The 1930s and 1940s: Synthetic Chemical Processes Replace Fermentation ..............................................................................................123 1941–1949 WWII: Fermentation Used for Penicillin and Pharmaceuticals .........................................................................................123 The 1950s: Fermentation as a Discrete Science .........................................123 The 1960s: Microbial Metal Mining Revived; Designer Microbes Assemble Proteins .....................................................................................123 The 1970s: Fermentation for Biopharmaceutical Manufacturing...............124 The Mid-1970s: Creating Chimeras by Inserting Foreign DNA Fragments into Plasmids ...........................................................................124 Mid-1980s: Fermentation for Producing Bioweapons ................................124 The Fermentation Process .....................................................................................127 Expression Systems......................................................................................127 Expression Modes ........................................................................................129 Enzyme Fermentation in the Organic Phase.........................................................133 Cell Growth and Production..................................................................................134 Kinetic Growth.............................................................................................134 Factors Affecting Specific Growth Rate ...............................................................143 Kinetic Models ......................................................................................................145 Optimal Conditions................................................................................................151 Maximizing Continuous Culture Productivity ............................................152 Continuous Culture Dilution Rate ...............................................................152 Recombinant Culture Kinetics ..............................................................................152 Plasmid-Carrying Cell Growth-Rate Kinetics .............................................152

PH2270_bookTOC.fm Page xvii Tuesday, May 23, 2006 10:01 AM

Fermenter Operating Modes ........................................................................156 Structure and Configuration.........................................................................157 Size and Scale ..............................................................................................158 Containment Vessel ......................................................................................160 Mixing and Aeration ....................................................................................160 Increased Size and Complexity ...................................................................161 Process Monitoring and Control..................................................................162 Process Monitoring: Sensor Location and Function ...................................166 Process Control ............................................................................................166 Feedback Controllers ...................................................................................166 Automatic Control Systems .........................................................................167 Integral Controllers ......................................................................................167 Derivative Controllers ..................................................................................168 Feed-Forward Controllers ............................................................................168 Adaptive Bioreactor Controls ......................................................................168 Complex Bioreactor Controls ......................................................................168 Cleaning and Sterilization............................................................................168 Sterilization ..................................................................................................168 Means of Sterilization..................................................................................169 The Del Factor (∇) ......................................................................................170 Continuous Sterilization...............................................................................171 Steam Sterilization .......................................................................................173 Steam System Components .........................................................................173 Plant or Utility Steam ..................................................................................174 Clean Steam or Pure Steam .........................................................................174 Steam System Design ..................................................................................175 Supplementary Fermenter Equipment .........................................................176 Other Fermenter Considerations..................................................................176 Fermenter Design ..................................................................................................176 Cell Kinetics.................................................................................................176 Bubble-Column Fermenter ..........................................................................179 STF Design Additions..................................................................................180 Basic Physical Elements ..............................................................................181 Complex Physical Elements ........................................................................184 Kinetic Modeling of the Fermentation Process ....................................................184 Agitation Dimensional Analysis ..................................................................184 Kinetic Fermentation Modeling...................................................................188 STF Batch Scale Up ..............................................................................................192 Scale-of-Agitation Method ..........................................................................192 Geometric Method for STF Scale Up .........................................................192 Dimensionless Numbers Method for STF Scale Up...................................193 Froude Number for STF Scale-Up Applications.........................................194 Power Number Method for STF Geometric Scale Up .........................................194 Scale-of-Agitation Method for Geometric Scale Up ..................................194 Continuous-Culture Scale Up................................................................................196 Theoretical Basis..........................................................................................196

PH2270_bookTOC.fm Page xviii Tuesday, May 23, 2006 10:01 AM

Laboratory to Pilot Plant to Production ......................................................196 Key Biological Aspect .................................................................................197 Maintaining the Steady State.......................................................................197 Contamination and Mutant Selection ..........................................................198 Assessing Productivity Loss ........................................................................198 Other Problems ............................................................................................198 Fermenter Manufacturers’ Directory .....................................................................199 Chapter 9

Bioreactors .......................................................................................201

Introduction............................................................................................................201 Historical Perspective...................................................................................202 Fermenters vs. Bioreactors ..........................................................................202 Bioreactor Operating Modes........................................................................202 Structure and Configuration.........................................................................202 Size and Scale ..............................................................................................203 Mixing and Agitation ...................................................................................204 Containment Vessels ....................................................................................205 Increased Size vs. Complexity.....................................................................205 Aseptic Conditions ................................................................................................206 Monoclonal Antibody Production .........................................................................206 Introduction ..................................................................................................206 Antigen-Antibody Binding ..........................................................................207 Equilibrium Dialysis ....................................................................................208 Determining Antibody Affinity ....................................................................208 Immunoassay................................................................................................209 Antibody-Based Immunoassays...................................................................210 Optimizing and Validating Immunoassays ..................................................210 Polyclonal Antibodies ..................................................................................210 Affinity-Purified Antibodies.........................................................................210 Monoclonal Antibodies ................................................................................211 Hybridomas ..................................................................................................211 Creating Monoclonal Antibodies .................................................................212 mAB Production in Ascites Fluid .........................................................................213 Overview ......................................................................................................213 Host Animals................................................................................................214 Antibiotics ....................................................................................................214 Injecting Hosts .............................................................................................215 Sacrificing Hosts ..........................................................................................216 mAB Production by Cell Culture ................................................................216 Eukaryotic Growth Medium ........................................................................217 Serum-Free Medium (SFM) ........................................................................218 Protein-Free Medium (PFM) .......................................................................218 Suspension vs. Anchorage-Dependent Cells ...............................................219 Porous Microcarriers....................................................................................219

PH2270_bookTOC.fm Page xix Tuesday, May 23, 2006 10:01 AM

Adventitious Agents .....................................................................................221 Bioreactor vs. Ascites-Fluid Production......................................................223 Bioreactor Engineering and Design ......................................................................224 Design and Engineering Considerations......................................................224 Basic Physical Elements ....................................................................224 Complex Physical Elements...............................................................227 Heat Generation..................................................................................230 Process Control System Design...................................................................232 Bioreactor Dynamics ..........................................................................232 Process Measurement and Control ..............................................................233 Process Monitoring: Sensor Location and Function .........................235 Process Control...................................................................................236 Feedback Controllers..........................................................................236 Automatic Control Systems ...............................................................236 Integral Controllers.............................................................................236 Derivative Controllers.........................................................................237 Feed-Forward Controllers...................................................................237 Adaptive Bioreactor Controls.............................................................237 Complex Bioreactor Controls.............................................................237 Bioreactors II: Kinetic Modeling and Bioreactor Dynamics and Bioreactor Design Program.................................................................................238 Modeling Cell Growth and Product Formation ..........................................238 Overview.............................................................................................238 Growth Reaction.................................................................................238 Structured Models ..............................................................................238 Cell Yield and Stoichiometric Coefficients........................................242 Cell Yield ............................................................................................242 Stoichiometric Coefficient Measurements .........................................243 Cell Growth Thermodynamics.....................................................................243 Heat Release Due to Growth .............................................................243 Heat Release from Extracellular Products.........................................245 Modeling Bioreactor Sterilization......................................................245 Values of ∇Heat and ∇Cool ....................................................................245 Heating................................................................................................246 Del Factor (∇) ....................................................................................247 Growth Kinetics and Product Formation.....................................................248 Growth Kinetics..................................................................................248 The Specific Growth Rate Value (μ)..................................................250 Metabolic Quotient and Rate Expression ..........................................252 Factors Affecting Growth Rate ..........................................................253 Product Formation Kinetics ...............................................................253 Oxygen Transfer in Bioreactors...................................................................257 Overview.............................................................................................257 Metabolic Oxygen Demand ...............................................................257 Volumetric Oxygen Mass-Transfer Coefficient .................................257

PH2270_bookTOC.fm Page xx Tuesday, May 23, 2006 10:01 AM

Bioreactor Oxygen Balance ...............................................................260 Factors Affecting Mass-Transfer Coefficient .....................................261 Modeling Eukaryotic Cell Growth and Production ....................................262 Kinetic Growth ...................................................................................262 Factors Affecting Specific Growth Rate ............................................270 Kinetic Models ...................................................................................271 Optimal Conditions ............................................................................275 Maximizing Productivity....................................................................276 Measurement of Stoichiometric Coefficients.....................................277 Cell Composition................................................................................277 Growth Reaction.................................................................................277 Cell Yield and Stoichiometric Coefficients........................................278 Cell Yield ............................................................................................279 Stoichiometric Coefficient Measurements .........................................282 Cell-Growth Thermodynamics.....................................................................283 Heat Release Due to Growth .............................................................283 Extracellular Product Heat Release ...................................................284 Growth Kinetics and Product Formation.....................................................285 Growth Kinetics..................................................................................285 On What Does the Specific Growth Rate (μ) Depend? ....................287 Metabolic Quotient and Rate Expression ..........................................288 Factors Affecting Growth Rate ..........................................................290 Product Formation Kinetics ...............................................................290 Oxygen Transfer in Bioreactors...................................................................292 Overview.............................................................................................292 Metabolic Oxygen Demand ...............................................................292 Volumetric Oxygen Mass-Transfer Coefficient .................................292 Bioreactor Oxygen Balance ...............................................................294 Factors Affecting Mass-Transfer Coefficient .....................................295 Cell Growth and Production ........................................................................297 Kinetic Growth ...................................................................................297 Batch Culture......................................................................................301 Bioreactor Design...............................................................................302 Factors Affecting Specific Growth Rate ............................................304 Kinetic Models ...................................................................................305 Optimal Conditions ............................................................................310 Maximizing Productivity....................................................................310 Recombinant Culture Kinetics ...........................................................311 Cell-Growth Modeling in a Batch Bioreactor ...................................314 Continuous Bioreactor Dynamics ......................................................315 Multiplicity and Steady-State Stability in a Continuous-Culture Bioreactor .........................................................................................317 Calculating Multiple Steady States....................................................317 Steady State Stability .........................................................................317 Proportional Control of STBR with Monod......................................317

PH2270_bookTOC.fm Page xxi Tuesday, May 23, 2006 10:01 AM

Linear Stability Analysis of an Open-Loop STBR ...........................318 Phase Plane Analysis for Open-Loop CSTBR ..................................319 Stability Analysis of Closed-Loop Bioreactor...................................319 Controlling Continuous-Culture Bioreactors with Substrate Inhibition ..........................................................................................320 Fed Batch Reactor Dynamics.............................................................320 Bifurcation Analysis ...........................................................................321 Bioreactor Design Program .........................................................................321 Bioreactor Manufacturers’ Directory ....................................................................324 Chapter 10 Biomolecular Synthesizers...............................................................325 DNA Synthesizers..................................................................................................325 DNA Synthesis.............................................................................................325 Coupling Chemistries...................................................................................326 Advantage of Multiple Chemistries.............................................................327 DNA Synthesizers ........................................................................................328 Automated Units ..........................................................................................328 Throughput Capability .................................................................................328 Scale Up .......................................................................................................329 Manufacturers’ Directory.......................................................................................329 DNA Synthesizers ........................................................................................329 DNA Synthesizer Accessories and Reagents ..............................................329 Peptide Technology Overview...............................................................................330 Peptides by Solid-Phase Synthesis........................................................................332 Peptide Synthesis .........................................................................................332 Solid-Phase Peptide Synthesis .....................................................................334 Solid-Phase Peptide Synthesizers ................................................................335 Solution Synthesis........................................................................................336 Sequential Peptide Synthesis .......................................................................338 Fragment Condensation Synthesis...............................................................338 Recombinant Peptide Synthesis...................................................................338 Screening for Peptides with Monoclonal Antibodies..................................338 Phage-Generated Peptide Libraries .............................................................339 Split Peptide Synthesis on a Bead Library..................................................340 Cleavable Linkers.........................................................................................340 Enzymatic Synthesis ....................................................................................341 Cell-Free Translation Systems .....................................................................341 Peptide Synthesis Equipment.......................................................................342 Process-Scale Peptide Synthesizer...............................................................343 Scaling-Up Peptide Synthesis......................................................................343 Solution Peptide Synthesis Scale-Up ..........................................................344 Solid-Phase Peptide Synthesis Scale-Up.....................................................345 Recombinant Peptide Synthesis Scale-Up...................................................345 Enzymatic Peptide Synthesis Scale-Up.......................................................346

PH2270_bookTOC.fm Page xxii Tuesday, May 23, 2006 10:01 AM

Manufacturers’ Directory.......................................................................................346 Peptide Synthesizers ....................................................................................346 Peptide Synthesizer Reagents ......................................................................346 Biotechnology Glossary .......................................................................................347 Bibliography and Recommended Reading .......................................................357 Bibliography ..........................................................................................................357 Recommended Reading Topics .............................................................................368 Airlift Bioreacters ........................................................................................368 Biokinetics, Control, etc. .............................................................................368 Biomass and Secondary Metabolytes ..........................................................369 Bioplastics ....................................................................................................369 Bioprocess Engineering ...............................................................................369 Engineering for Monoseptic Operations......................................................370 General .........................................................................................................370 Heat Transfer................................................................................................370 Mass Transfer and Hydrodynamics .............................................................370 Other Bioreactors .........................................................................................371 Photo Bioreactors.........................................................................................371 Proteins and Enzymes..................................................................................371 Shear Effects ................................................................................................372 Index ......................................................................................................................373

PH2270_book.fm Page 1 Tuesday, May 16, 2006 9:37 AM

Introduction BIOTECHNOLOGY

Biotechnology is the use of living organisms, their components, secretions, or metabolites for the development of useful products. Through biotechnology, complex cellular chemistry can be technically enhanced and then employed to provide new products and processes. The focus of biotechnology is generally on the cell, existing in culture as genetically altered prokaryotic (bacteria) or eukaryotic cells (mammalian and insect cells). Historically, biotechnology has been concerned with the development of hybrid microorganisms used for making beer, wine, bread, cheese, and for developing hybrid crops and livestock. Modern techniques have led to the development of genetically altered cells that effectively produce either new products or large quantities of scarce products including antibiotics, antigens, enzymes, and other such biosubstances. Such techniques typically improve industrial processes where a biological catalyst replaces a chemical catalyst. Table 1.1, The Biochemistry, Biotechnology, Fermentation, and Enzyme Timeline, shows these developments over the years. It is expected that the biotechnology industry will grow significantly during the balance of this century. The U.S. Office of Technology Assessment predicted that biotechnology-based processes will replace a large percentage of those utilized in standard product manufacture by the year 2025. Thus, biotechnology can strengthen and diversify the national economy by innovating existing biotech-based businesses and by providing the necessary technology to establish new industrial enterprises. Industrial biotechnology applications can be found in many sectors including pharmaceuticals, commodities, specialty chemicals, food additives, bioelectronics, energy production, and the environment. While the investment community and the media have focused primarily on the beginnings of the biomolecular production process, the cloning and expression of desired genes, biochemical synthesis, and the growth of genetically altered cell cultures to produce intrinsic or extrinsic cell products, technical challenges at the downstream stages of the production process have gone relatively unnoticed.

BIOPRODUCTION The cost of producing large quantities of products of acceptable purity with the early bioprocessing techniques remained so high that only exceedingly high value-added products, such as pharmaceuticals, could generate sufficient revenue to justify the costs of the elaborate manufacturing steps that were required. Over the years, improvements in production and purification have greatly cut costs and generated a flood of new products that otherwise would not have been marketed because of prohibitive manu1

PH2270_book.fm Page 2 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

TABLE 1.1 Biochemistry, Biotechnology, Fermentation, and Enzyme Timeline ca. 10,000 BC Fermentation discovered/used. ca. 2500 BC “Ancient Greek” view of chemistry: • Air, Earth, Fire, Water, Dry, Wet, Cold, Hot The emergence of modern biochemistry resulted, in large measure, through the gradual replacement of the vital force theory by the enzyme theory. ca. 350 BC Aristotle: • “The vital force.” 1000–1600 AD Alchemists identify alcohol as the product of fermentation. • Addition to “Greek” chemistry theory: the quinta essential. 1200 AD Distillation: • Aqua ardens 1734–1794 AD Lavoisier: • C6H12O6 2C2H5OH + 2 CO2 (conservation of mass) 1800 AD The role of yeast in fermentation: • Fermentation yeast 1833 AD Payen and Persoz: • Diastase converts starch to sugar; the chemical transformation is not associated with living cells. 1835 AD Schwann: • Pepsin from gastric mucosa 1839 AD Berzelius: • The concept of catalysis: “The catalytic force is reflected in the capacity that some substances have, by their mere presence and not by their own reactivity, to promote changes in otherwise stable and unreactive molecules ... in living plants and animals, thousands of catalytic processes occur within the tissues and fluids, generating a multitude of substances of differing chemical compositions.” Liebig: • The chemical theory of fermentation: “... vibrational impact of decomposing yeast induces sugar decomposition ...” • Liebig’s reply to Berzelius: “To call a phenomenon catalytic is not to explain it; it is nothing but the replacement of a common word by a Greek word.” 1840 AD Schwann, Cagniard-Latour: • “… the driving force for chemical transformation comes from the living organism.” 1842 AD Pasteur: • “... fermentation is essentially a phenomenon correlative with a vital act, beginning and ending with the growth of a living cell ...” 1860 AD Traube: Berzelius: • “… soluble ferments exist inside cells, where they carry out specific chemical transformations ...” Berthelot: • Invertase was the first “ferment,” (i.e., enzyme) extracted from living yeast cells. • This essentially suggests that: “The vital force of all chemical transformations in living systems is promoted by ferments in cells.” 1860–1917 AD Büchner: • Zymase fermentation in a test tube 1878 AD Kühne: • Name “Enzym” • 1887 AD Bormuschel Dubois:

PH2270_book.fm Page 3 Tuesday, May 16, 2006 9:37 AM

Introduction

TABLE 1.1 (Continued) Biochemistry, Biotechnology, Fermentation, and Enzyme Timeline • “Extract with cold water • Extract with hot water • Emission (während min.) • Combine: hot water and spent cold water extracts: no emission” • Emission: Luciferin (from Lucifer: the carrier of light) (heat stable, but will be consumed) • Luciferase: (“-ase” → enzyme, which is a catalyst labile to heat, but remains unchanged) 1895 AD–1900 AD Fischer: • Concept of enzyme action: “What is an enzyme?” • “Lock-and-key” enzyme mechanism • Willstätter-Sumner controversy: Colloids and Chaos Proteins as Enzymes: “… each vital cellular reaction will be discovered to be carried out by a particular specific ferment.” • The beginnings of biochemistry • Postulated metabolic pathways • Theory: “Enzymes are capable of generating all the compounds found in living tissue.” 1916–1918 AD (WWI): Impetus for Fermentation Growth 1927 AD Sumner: • Crystallization of Urease from jack beans, acetone precipitation 1930 AD Northrop and Kunitz • Introduction of enzymological methods: • Electrophoresis • Centrifugation • Denaturation and renaturation • Henri Brown: • Enzyme – substrate complex 1930s and 1940s AD Replacement of Fermentation by Synthetic Chemical Processes 1935 AD Hugo Theorell: • (Holo) enzyme = Apoenzyme + Cofactor • Theorell reported on the discovery of a reversibly bound cofactor associated • With what is now known as the “Old Yellow Enzyme” (excerpted from his original 1935 report): “The significance of new investigations on the yellow enzyme may be summarized as: 1. The reversible splitting of the yellow enzyme to a poenzyme + coenzyme in the simple molecular relation 1:1 proved that we had here to do with a pure enzyme; the experiments would have been incomprehensible if the enzyme itself had been only an impurity. 2. The enzyme was thus demonstrably a protein. In the sequellae, enzymes have been isolated and proved to be proteins. 3. The first coenzyme, FMN, was isolated and found to be a vitamin phosphoric acid ester. This has since proved to be something occurring widely in nature; the vitamins nicotinic acid amide, thiamine and pyridoxine form in an analogous way nucleotide-like coenzymes, which like the nucleic acids themselves combine reversibly with proteins.” • (Holo) enzyme = Apoenzyme + Cofactor • {X}Old Yellow Enzyme Ü Protein FMN • Holoenzyme, apoenzyme, cofactor • {X}: acid ammonium sulfate [precipitate] 1941–1944 AD (WWII): Fermentation for penicillin and pharmaceuticals 1948 AD Pauling: • Transition state theory (stabilization of TS)

PH2270_book.fm Page 4 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

TABLE 1.1 (Continued) Biochemistry, Biotechnology, Fermentation, and Enzyme Timeline • Northrop: “… on controversies.“ • “... The history of biochemistry is a chronicle of controversy. Such controversies exhibit a common pattern … There is a complicated hypothesis, which usually entails an element of mystery and several unnecessary assumptions … This is opposed by a more simple explanation, which contains no unnecessary assumptions. The complicated explanation is always popular at first, but the simpler one, as a rule, is eventually found to be correct. This process frequently requires two to twenty years. The reason for this long time lag was explained by Max Planck, who reminded us that scientists never change their minds, but do eventually die.” 1950s AD Fermentation as a Discrete Science 1960s AD Microbial Mining of Metals; Designer Microbes for Producing Bioactive Proteins 1970s AD Fermentation for Biopharmaceutical Manufacturing 1973 AD Inserting Foreign DNA Fragments to Create Chimeras 1974 AD Biological Weapons Programs to Make Weapons of Mass Destruction (WMDs)

facturing costs — or would have been confined to relatively small market niches in which volume, characterization, pricing, and regulatory requirements would not represent significant constraints. Like fingerprints, no two fermenter or bioreactor processes are quite alike, and the study of the growth processes of mammalian, bacterial, and yeast cells has led to techniques that optimize product yields. Such techniques include: • • • •

The design of fermenters and bioreactors and how their configurations at different volumes affect culture systems. How variables such as pH, cell mass, oxygen and carbon dioxide levels, and agitation parameters affect productivity. Understanding cellular metabolism, mass balance, and kinetics, and using this knowledge to develop models to help optimize cell growth. Simulating genetically engineered cell-population dynamics to gain knowledge of factors that contribute to increased cell productivity.

Such diverse approaches notwithstanding, a detailed knowledge of cell biology and metabolism is key to optimizing cell growth. However, modeling cell synthesis is complicated by the interaction of sophisticated control systems, thousands of individual biochemical reactions, and the cellular environment. Nevertheless, highly detailed kinetic models incorporating many of these variables and an ability to predict cellular physiology represent some of the essential work necessary for understanding factors essential for optimizing cell growth. Since such models are typically too complex to be employed for actual process control, methods are needed to simplify the mathematics of these models by developing new algorithms and utilizing artificial intelligence to maintain optimal product yield. Methods for monitoring parameters detrimental to cell growth and productivity are also necessary in such a process control scheme, suggesting the use of biosensors to measure constituents of interest, a feedback mechanism, and a process control system to maintain optimal environment. Whether developed empirically or with computer models, the resultant control

PH2270_book.fm Page 5 Tuesday, May 16, 2006 9:37 AM

Introduction

Lab & Pilot-Scale

Production Scale

Cell-Line R&D Testing Cell Culture Cell-Culture Process Testing

Purification Purification Process Testing

QC Sterile Bulk Product QC Final QC Batch Records

Materials

Finished Product Release

Documentation

FIGURE 1.1 Typical bioprocess scale-up flow sheet.

system would, notwithstanding, still be used to maintain productivity. For highly complex systems, programmable logic controllers would provide command function at the fermenter/bioreactor level, while the computer algorithm would provide system control — thus providing two levels of command redundancy. Even though the development of large-scale cell culture processes, while critical to this new generation of products, is only one aspect of the complete bioprocess, cell-derived biomolecules licensed for human use include: natural interferons, monoclonal antibodies, human enzymes, and hormones produced by recombinant technology.

THE BIOPROCESS Figure 1.1 outlines the steps involved in the development and manufacture of a cellderived therapeutic product. All technologies should be integrated and performed to FDA current Good Manufacturing Practice (cGMP) standards, as followed by the biopharmaceutical industry. In particular, emphasis should be placed on using a reliable bioprocess to ensure a product of consistent, reproducible biological activity. This is accomplished by validation of both the bioprocess and the plant. Validation means the documented experimental proof that the particular manufacturing process can consistently produce a product of the required quality, purity, and character, and that the manufacturing plant is functioning acceptably during each production operation.

CELL LINES In the United States, most new biomolecular products are either monoclonal antibodies grown from hybridoma cell lines or recombinant products expressed in E. coli

PH2270_book.fm Page 6 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

TABLE 1.2 E. coli Host Strain for Expression Author: Date: Notes: Supersedes: Host genus and species: Designation: Other: Genotype: Derivation: Biosafety level: Reference: Pathogenic characteristics: Disposal: General:

John Doe November 4, 1989 Description of E. coli host strain (W3110/DE3) including organism, description, biosafety level, and genetic markers None E. coli K-12 W3110 Lambda-DE3 lysogen F-mcrA mcrB IN(rrnD-rrnE)1 Lambda-DE3 Lambda-DE3 lysogen of ATCC Strain W3110 BL-1 Federal Register, Vol. 51, No. 88, May 7, 1986 None known Autoclave Exhibits normal E. coli growth characteristics; no special regulatory issues pertain to the use of this strain.

(see Table 1.2). Therefore, hybridoma and recombinant cell-line expression systems are important to bioproduction, since cell-line productivity varies enormously with selection of a host-cell type, which can determine factors such as growth characteristics, expressed protein production rate, or whether the cells will only grow attached to a solid surface (anchor dependent), or grow in free suspension (anchor independent).

CELL-CULTURE PROCESS Many cell-culture products are produced by way of deep-tank suspension cultures in either batch or continuous operation (see Table 1.3). Tanks of up to 10,000 liters are frequently used in industrial production. Bioprocesses in hom*ogeneous, wellmixed fermenters (or bioreactors) are typically monitored through probes in the vessels, or by removing samples for the offline analysis of residual nutrients, toxic metabolites, cell debris, and the desired biomolecular product. One of the most important aspects of working with a suspension culture is the ability to develop a unit process and scale it up as required. Particular attention is given to validating the process and its associated systems. Process monitoring and control is typically automated, and results are documented by an online computer system. For example, in a hypothetical 2,000-liter product-run, a culture is grown from a working cell bank to perhaps 2 liters (at maximum production density in the laboratory), and then further scaled up in pilot and finally in production fermenters (or bioreactors) in stages from 20 to 200 to 2,000 liters. Transfers generally take place aseptically through steam-sterilized hard-piped lines.

PH2270_book.fm Page 7 Tuesday, May 16, 2006 9:37 AM

Introduction

TABLE 1.3 Example of a Pilot Fermentation Procedure 1.

2. 3.

4. 5.

A 300-liter fermenter will be batched with approximately 225 liters of presterilization medium, and sterilized at >121ºC for 30 minutes. After cool-down, poststerilization media is added to the fermenter from separate sterile stock solutions. A 10-liter fermenter seed culture is used to inoculate the 300-liter fermenter. The 300-liter fermenter will operate under the following parameters: DO2: >10% Temperature: 37ºC Agitation: 150 RPM Pressure: 5.0 PSIG pH: 7.0* The culture is monitored for glucose concentration and feeds are adjusted during the run to maintain 1–2 g/l glucose until it reaches an O.D. of 25 @ 600 nm.** When an O.D. of 25 @ 600 nm is reached the culture will be induced with 1 ml/l of 1.0 M IPTG.

* Controlled with 10% NH4OH; feed, adjusted to maintain growth. ** Continue growth for 3 hours after culture induction.

SEPARATION, RECOVERY, PURIFICATION After lysing (mechanically breaking up) the cells and removing cell debris (for intracellular production), or after removing cells and cell debris (for cells expressing in the medium) either by filtration and/or centrifugation, the crude product is then typically concentrated by ultrafiltration. Subsequently, purification is carried out using a series of chromatography steps, the precise details depending, of course, on the particular product being isolated. Typically, open-column chromatography is used, with specific biological matrices targeted for affinity, ion-exchange, or gelfiltration separation. Similarly, protein contaminants, cell DNA, cell RNA, and other materials used in or produced by the bioprocess, are removed from the culture medium. Additionally, specific virus inactivation steps (e.g., high- or low-pH, heat, or organic solvent treatment) may be introduced to provide additional reassurance that a theoretical virus, undetectable by present methods, will not be present in the final product.

VIRUS AND FOREIGN-DNA REMOVAL Bioprocess efficiency also depends on the ability to quantify the adequacy of each step to eliminate viruses from the end product. This is generally demonstrated in small-scale lab experiments where the crude product is spiked with a high titer of a known virus and passed through the process purification steps, subsequently determining the virus titer of the eluate. Reduction in virus titer through a particular processing step may be as great as 105 or greater. The clearance factor is a semiquantitative measure of assurance that an end product will not contain a virus. And, a similar approach is used to demonstrate the elimination of foreign DNA from

PH2270_book.fm Page 8 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

eukaryotic cell-derived products. After purification, the product is sterile-filtered and stored in bulk, prior to final formulation and packaging.

QUALITY ASSURANCE Both intermediate and end products are tested to demonstrate that they meet required specifications. Creation of appropriate test methods (major components of the process validation) requires tremendous developmental effort. For example, ultra-sensitive hybridization methods, used for detecting foreign-cell DNA at 1 pg or less in a product’s dose, have been developed for many eukaryotic cell-derived products. Specific cell proteins are detectable at very low concentration using monoclonal antibodies against these proteins in complex protein mixtures. Immunoassays can also be developed for many of the reagents used in bioprocessing. Quality control also plays a major role in bioproduction through quality-assured raw materials. For example, the injection-grade water used in formulating cell-culture medium is regularly tested. Biologics that are used in the process (such as serum, albumin, insulin, etc.) are purchased from suppliers who have their own quality assurance programs, and their facilities are even inspected by some purchasers. The absence of viruses, mycoplasma, and endotoxins are key concerns in controlling these materials. There is also a major R&D effort to minimize or eliminate many biologicals from cellculture media. Over the last few years, DNA, hybridoma, large-scale cell culture, protein purification, immunoassay development, and many other technologies have been integrated to provide effective manufacturing processes that can be implemented by the biopharmaceutical industry. The final product is only released after all quality tests have been performed and the batch records reviewed carefully to ensure compliance. The end result is a culmination of the total quality assurance program that began during the development of the product, continued with the bioprocess design and monitoring, and concluded with end-product inspection to satisfy the regulatory authorities of each country to which the product is shipped. Over the last few years, rDNA, hybridomas, large-scale cell culture, protein purification, immunoassay development, and many other technologies have been integrated to provide effective bioprocesses that can be implemented by the biopharm industry. Manufacturing operations such as cell culture, fermentation, recovery, purification, packaging, material movement, and all other aspects of bioprocessing facility operations are conscientiously performed procedures that contribute to the production of an exact product of specified biological activity. The production facilities, as well as the component manufacturing, testing, labeling, and packaging procedures, are specified for both the establishment and in the product license applications, and must be approved before the product can be legally marketed. After license is granted, the manufacturer is still held responsible for assuring that the product is manufactured in accordance with the specified procedures. Compliance with regulations and guidelines, while demanding, actually helps the manufacturer increase productivity and assure quality because they manufacture the product correctly.

PH2270_book.fm Page 9 Tuesday, May 16, 2006 9:37 AM

Technology Overview BIOMOLECULAR FOUNDATIONS

When Watson and Crick proposed their three-dimensional double-helix DNA model in 1953, implications for the coding and replication of genetic information could not be experimentally verified. The concept of the gene and gene-expressed products gradually attained a concrete form. By 1978, it had become clear how to combine multicellular organism genes or genetic fragments with those of viruses, fungi, and bacteria to yield new metabolic instructions for producing novel products; scientists were then ready to apply those methods to exploit new technologies. The success of biomolecular procedures is based upon several related, but independent, developments: 1. The ability to clone genetic information (i.e., to isolate a selected segment and accurately reproduce it in large amounts). 2. The ability to determine the nucleic acid sequence of the selected gene segment(s) (i.e., to read the complete molecular structure of a gene). 3. The ability to practice genetic engineering (i.e., to alter and control gene expression and change the structure of gene-induced products by chemically modifying specific sites in the molecular structure of the gene). 4. Increased possibility for applying such procedures was implied by two additional technologies: (a) the polymerase chain reaction (PCR), by which large amounts of specific nucleic acid sequences can be produced without prior purification, cloning, or complete knowledge of their sequence, and (b) the ability to create transgenic animals by transferring synthetic genes into embryonic cells to make genetically modified animals to specification.

CELLULAR VARIATION All mammalian cells, except for the erythrocytes, have a nucleus separating the units of genetic information from the cytoplasm. The cytoplasm and its component organelles allow cells to generate the energy necessary to synthesize structural and enzymatic molecules that furnish the functional properties whereby they operate the organism. Each and every somatic cell, except for antibody-producing lymphocytes, carries precisely the same basic set of genetic information. An individual’s complete set of genetic instructions (genotype), is composed of genes (basic metabolic instructional units), having particular locations on specific chromosomes. Each type of cell 9

PH2270_book.fm Page 10 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

DNA

hnRNA

mRNA

Protein

mRNA Screen cDNA

Insert

Replicate DNA Clone Selected Plasmid

FIGURE 2.1 Determining cell phenotype and cloning mRNAs.

expresses a subset of genes that encode for special structural and enzymatic proteins that provide the cell with its functional characteristics (phenotype). Cell organelle systems, essential for selective transcription, translation, separation, and packaging of proteins, by which a given type of cell attains its specific phenotype, were well known to typical cytologists long before the molecular mechanisms underlying these events was understandable (see Figure 2.1).

KEY MOLECULAR INTERACTIONS The Watson-Crick DNA model provided a coherent integration of the X-ray crystallographic structural data of partially purified DNA with the previously known quantitative chemical data, and demonstrated the equal occurrence frequency of the two purine-pyrimidine pairs within the DNA molecule [adenine A with thymidine T (A-T) and guanine G with cytosine C (G-T)], suggesting the basic mechanism of DNA replication.

BASE-PAIR COMPLEMENTARITY In the DNA double helix, two right-handed, helical, polynucleotide chains coil around the same central axis, making a complete helical turn every ten nucleotides. In the interior, the purine and pyrimidine bases are paired through hydrogen bonding of their complementary structures (A-T and G-C), placing the phosphate groups at

PH2270_book.fm Page 11 Tuesday, May 16, 2006 9:37 AM

Technology Overview

T A

A T

C G

G C

T A

C G

DNA Sense Strand -A-T-G-C-A-G-T-A-C-G-T-CAnti-Sense Strand Transcription RNA Transcript -A-U-G-C-A-GRNA Triplets Translation Amino Met Gin Acids

NH2 - G - M - A - M - Q - ...... COOH

Proteins

FIGURE 2.2 Arrangement and relationship of the DNA bases.

the periphery of the helix. Gene expression is enabled by the precise molecular complementarity between the primary nucleotide base sequences in the one strand of the helix and the analogous sequence of the second strand of the helix. The strand that encodes genetic information is termed the sense strand and its complement the antisense strand; wherever a particular base exists in the sense strand, there will exist a complementary base in the antisense strand. This base-pair complementarity (see Figure 2.2) allows for genetic information duplication in dividing cells through enzymes known as DNA polymerases that open the helix and again convert each single strand into double strands according to the single strand’s template. This double-strand complementarity also provides a repair mechanism for the DNA, should it be damaged, since the single strand surviving the damage acts as a template for the repair. In a similar manner, the information-bearing sense strand is copied into a complementary single-stranded RNA during the process of transcription. RNA, therefore, is a single-stranded complementary copy of the DNA antisense strand, its sequence resembling that of the sense strand. RNA differs chemically from DNA by the substitution of uridine for thymidine, and ribose phosphates for deoxyribose phosphates. Transcription is performed by RNA polymerase enzymes. Base-pair affinity for complementary base-pair sequences in DNA or RNA along the sequences of a single strand is so precise that small segments can be used as probes for the detection of hom*ologous sequences between large DNA and RNA domains. The ability of a single-stranded nucleic acid to hybridize (bind) with its complementary sequence is an essential component of many molecular biology techniques.

GENETIC CODING Meticulous coding is required to translate genetic information from RNA sequences into linear protein amino acid sequences so that protein amino acids can be specified by various combinations of the four nucleotides. Sets of three RNA bases, or triplets, provide code templates that specify the order in which amino acids are incorporated

PH2270_book.fm Page 12 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

into the protein (see Figure 2.2), and other triplet sequences mark the point at which synthesis begins, ends, or is modified. Often, mRNAs encode more protein sequence than represented in the final form of the processed gene product. The signal peptide is a 19 to 30 amino acid sequence at the N-terminus of the encoded gene product in which the amino acids are highly hydrophobic. The signal peptide is a constant feature of proteins intended for secretion, such as neuropeptides, and its apparent function is to guide the newly formed protein chain through the endoplasmic reticulum for later envelopment by the Golgi apparatus.

INTERRUPTING GENE SEGMENTS The genetic organization of the eukaryotic cells is basically different from that of the prokaryotic cells that typically lack a separate genetic compartment (nuclear membrane). Instead, in the higher organisms, and in some viruses, the cells have their gene segments split into expressed coding regions (exons) separated by intervening regions (introns) that are not found in the mRNA. The interrupted DNA coding produces two outcomes: (1) the primary gene transcript — or heterogeneous nuclear RNA (hnRNA) containing extra RNA sequences — and introns must be removed by opening the hnRNA, removing them, and re-splicing the cut ends before the mRNA can successfully direct protein synthesis by ribosomes (see Figure 2.3), or (2) the composition of the transcribed mRNA in some cells must also be revised by splicing out certain exon segments. This process provides a means by which a gene containing several exons can trigger the production of multiple protein gene products where certain protein domains are shared and others are unique. Gene products that share similar nucleotide and protein sequences are known as a structural family. DNA

hnRNA

Proteins

Carbohydrate Groups

FIGURE 2.3 Relationship of DNA, RNA, and mRNA sequences and proteins to each other.

PH2270_book.fm Page 13 Tuesday, May 16, 2006 9:37 AM

Technology Overview

SEQUENCE DETERMINATIONS A final development that facilitated current breakthroughs in molecular biology was a group of techniques for discerning DNA sequences up to several thousand base pairs in length. From such DNA sequences it is possible to derive the nucleotide sequences of RNA and, thereby, the amino-acid sequences of expressed proteins. Currently, DNA and gene-product structures are analyzed by computer and their sequences compared with libraries of previously characterized proteins and/or nucleic acids. In addition to nucleic-acid-sequence structural comparison, other indications of the encoded molecule’s function may be inferred from hydrophobic, hydrophilic, or other structural sites for posttranslational modification.

CLONING When mRNAs are converted into DNA by the enzyme reverse transcriptase (see Chapter 5) the copied double-stranded form (cDNA) can be incorporated or inserted into specific sites within an infectious vector, usually a virus or bacteriophage, or into an extra chromosomal genetic element (plasmid) found among various strains of bacteria. The insertion sites are selected by identifying DNA sequences that can be cut by the actions of restriction endonucleases (see Chapter 5) — enzymes from purified bacterial sources that cleave DNA sequences at specific palindromicallyrepeated sequence sites. By using plasmids of known DNA sequence, tailored to include restriction cleavage sites that allow for DNA insertion, the same enzyme can later be used to cleave out the insert. The restriction sites for insertion are typically chosen within plasmid genes that code for some discernible functional property (such as antibiotic resistance). Thus, when insertion has been successful, interruption of coding and expression leads to the loss of functional property and permits the identification of plasmids with effective inserts. In general, each plasmid can only incorporate one cDNA insert, and with a great excess of host bacteria, each insert-bearing plasmid will infect only a single host, usually E. coli. By growing these genetically altered bacteria in such a way that each individual bacterium gives rise to a colony of identical bacteria carrying replicates of the plasmid and insert, the DNA is successfully cloned. The cDNA can then be recovered from the plasmid through another exposure to the restriction enzyme selected for the original opening of the plasmid insertion site. Thus, in a relatively few steps the process can begin with a collection of mRNAs — from common to very rare — purify them individually, and develop virtually unlimited pure copies of the DNA insert (see Figure 2.1 and Table 2.1). The PCR (see Chapter 7) processes large amounts of specific rare nucleic acid sequences through amplification in vitro without the necessity of first purifying the desired sequences through cloning (see Chapter 5). The changes in biotechnology research brought about by this new technology were immediate and dramatic. Using previously selected restriction enzymes, human DNA was sliced into small sections, one of which contained the target gene sequence. When the DNA was heated to 95ºC, the double-stranded DNA dissociated (“melted”), and the strands separated. Two small DNA sequences synthesized oligonucleotides to be complementary to

PH2270_book.fm Page 14 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

TABLE 2.1 Construction of Host Strain for Production Author: Date: Notes: Supersedes: Bacterial strains: Phage strains:

Protocol: Phage strains for selecting lysogens:

Earl Jones November 4, 1989 Protocol for construction of DE3 lysogen; origin of strains and phages used None W3110 and LE392.23 Lambda-DE3: imm21 cI+ carries T7 RNA polymerase under lac control Lambda-B10: imm21 cILambda-B482: imm21 cI-(att) Delta-1 h80 T4107: T-phage from which entire the T7 RNA polymerase gene is deleted Procedure for making DE3 lysogens. DE3 imm21 cI+ carries inducible gene for T7 RNA polymerase B10 imm21 cIB482 imm21 cI-(att) Delta-1 h80 Use lysates grown on a modifying host when making lysogens of Eco*k+ strains. Tryptone broth or equivalent is suitable for cultures and in agar plates. Mix 108 each of DE3, B10, and B482 in 2.5 ml top agar, add 1–10 ml of a fresh culture of the cells to be lysogenized, and pour onto a 20 ml agar plate. Most colonies that grow after incubation overnight at 37°C should be DE3 lysogens. Grow a small culture from one of the colonies and purify a single colony from this culture. Test the purified strain for immunity and for inducible T7 RNA polymerase activity. T7 grows on many female, but few male, strains of E. coli. A simple test for T7 RNA polymerase activity in cells that plate T7 is to test for ability to plate 4107, a T7 mutant from which the entire T7 RNA polymerase gene has been deleted. 4107 is totally unable to form a plaque on cells that lack T7 RNA polymerase, but it forms normal plaques on a DE3 lysogen in the presence of an inducer (0.025 ml of 0.1M IPTG added to the top agar). When plated on a DE3 lysogen in the absence of the inducer, 4107 typically forms small plaques that take a long time to develop.

opposite strands of the target gene. These oligonucleotide primers hybridized to their complementary sequences on the single-stranded DNA (see Figure 2.2). In the presence of large amounts of a purified DNA polymerase and deoxynucleotides, the primers are extended to the end of the single strand. Then, the reaction cycles of denaturing, annealing, and extending, each lasting only a few seconds, can be rapidly repeated by separating the dual-helical strands of the newly synthesized material through heat denaturation, cooling the reaction mixture, and adding fresh DNA polymerase and nucleotides. As long as the polymerase and the nucleotide substrates are in excess, the extended sequences of the first reaction serve as templates for opposite strand synthesis in subsequent cycles, providing a geometric rate of amplification. With DNA polymerase, which is isolated from a specific bacterial strain that grows in the extreme heat of geysers, it is possible to develop a method permitting large quantities of polymerase to survive the heat denaturation step. With this polymerase and large beginning amounts of d-nucle-

PH2270_book.fm Page 15 Tuesday, May 16, 2006 9:37 AM

Technology Overview

otides, a series of rapid cycles produce the exponential in vitro amplification of the desired DNA sequence. If selected sequences are known to be generally constant in the genome of a species, the same primers can be used to select, amplify, and subsequently analyze the same intervening gene segment from many individuals. If the amplified segment is simply analyzed for its length, it is possible to determine whether a given individual has a major genetic mutation (e.g., a deletion or an insertion). Additional applications of the technology have increased its advantages. PCR can also be applied to mRNAs and can even provide a quantitative analysis by first using reverse transcriptase to make double-stranded copy DNA (cDNA). By modifying the end of the probes to be used, it is also possible to incorporate special synthetic sequences that make it easier to clone or sequence amplified segments, or to re-insert modified versions to determine the functional importance of a specific sequence. Even if the sequences do not match precisely, it is also possible to amplify hom*ologous sequences from different but related genes by using special nucleotides that will form complementary base pairs. The above examples indicate the simplicity of cloning DNA segments taken directly from genomic digests or from mRNA copies, sequencing the cloned segments, and discerning the structure of the product. Nevertheless, how does one determine which clone is carrying the insert that encodes a specific gene product? And what if this sequence is not determined so that it can be located and cloned? Scientists have found various techniques to gain these objectives, although some of the screening methods are extremely tedious.

APPLICATIONS The cell phenotype depends upon the structural, metabolic, and regulatory proteins by which recognizable physical and functional properties are determined. Complex, multifunctional cells rely upon myriads of special-purpose proteins, many of which exist in limited amounts. Purifying these scarce proteins by premolecular cloning methods, especially in the absence of a functional assay, was an enormous task requiring patience, resources, and a very large supply of cellular material. The structure of a specific protein directly relates to the mRNA or gene segment that encodes the protein, and depends upon the nature of the cDNA sought after, and whether or not you can locate its insertion or determine its genetic fusion product translation in a biological infrastructure capable of processing its translation product into a form that replicates natural structure and function.

SCREENING

AND

SELECTION

A practical beginning is selecting cells that are known or are presumed to express a target molecule, then enriching mRNA sources that favor target molecule detection. Hormone-producing cell lines and tumors or cells bearing large numbers of desired receptors or channels (e.g., striated muscle) are excellent starting materials. Once a cell source is selected, the desired mRNAs are further enriched by sucrose gradient centrifugation or electrophoresis, provided that some of the characteristics of the target mRNA are known. In colony hybridization, a common strategy for detecting

PH2270_book.fm Page 16 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

desired cloned bacteria colonies is growing the bacteria on a special culture plate from which the colonies can be transferred as a group by replica plating (e.g., lightly pressing them to another supporting surface), and thereby sample and preserve the spatial identity of all the colonies on the plate. Bacteria on the replicate supports are then screened with nucleic acid probes to identify the target colonies. If the plasmid-carrying inserts are tailored to allow for protein expression encoded by the transferred genetic material, it is also possible to identify desired clones by immunoassay. When a target colony has been identified, the original colony is recovered from the primary culture plate and subsequently cultured in large quantities, thus providing material for DNA sequence analysis.

DNA SYNTHESIS An mRNA sequence can be predicted if a partial protein or peptide sequence is known by back-translating the genetic amino acid code and including enough alternatives to overcome ambiguous cases where a specific amino acid might be encoded by several variant triplets; then the cDNA can be designed and synthesized from the predicted RNA structure. This approach has been used to create theoretical cDNAs for hormones, with sequences determined one amino acid at a time from highly purified tissue extracts. Generally performed with a structurally known, biologically active, complex protein, or peptide, the procedures determine either the product’s structure, or the complete genomic configuration necessary to produce its regulatory and expression mechanisms. A target peptide’s precursor, however, is generally found to encode more than one active product, and because of the redundancy of triplet RNA codons for some amino acids, it is generally difficult to acquire a functional full-length mRNA by predictive synthesis. An alternative approach is to synthesize a DNA probe, which is a shorter, complementary, single-stranded DNA, used as a probe to screen clone libraries which are prepared either from mRNA extracts or from whole genomic digests. In the former case the starting material would be tissue, while in the latter case any somatic cells could be used to prepare the library. With the availability of automated genetic sequencers, it is now possible to synthesize probes overnight, use them to screen a genomic or cDNA library, and subsequently determine the complete coding sequence for a partially purified protein within a few weeks. Candidate colonies can also be cross-screened by a second probe based upon another separate domain of the full protein. Clones positive for both probes, therefore, should contain the gene sequence that encodes the two sequences against which the probes were made, as well as the sequence between them. This strategy has been used with many neuropeptide mRNAs. Utilizing the Southern blot (named for Dr. E.M. Southern, who devised the method), after DNA is cleaved by restriction enzymes and separated by gel electrophoresis, it is possible to transfer or blot the resulting fragments (separated mainly on the basis of their lengths) from the gel to a nitrocellulose support and then to analyze them for their ability to hybridize with cDNA or RNA probes. A similar method uses RNA as the starting material, where the separated RNAs are blotted for probing with single-stranded cDNA probes. Because the starting material

PH2270_book.fm Page 17 Tuesday, May 16, 2006 9:37 AM

Technology Overview

is the opposite of that of the Southern method, the RNA blot is referred to as a Northern blot. Recently, immunological methods have been employed to examine protein extracts separated by Polyacrylamide Gel Electrophoresis (PAGE) and then blotted by electrical transfer for subsequent identification by labeled antibodies to specific protein antigens. This method is called a Western blot. If RNA or protein specimens are dried on nitrocellulose for analysis without first separating them for size, the resulting blots are called slot blots or dot blots. These blots are useful when quickly screening a large number of clone extracts for inserts or expressed products. Increasingly, a large number of molecules are being fully characterized, thereby enabling scientists to create biodrug molecules that precisely fit specific receptors or enzymes, and thus demonstrating that any biological event mediated by proteins is receptive to biomolecular analysis. The range of these biomolecules includes enzymes needed for hormone synthesis, storage, release, and catabolism, receptors and related macromolecules needed for response, and those needed for response mediation of broad-scale physiological events. Methods that provide innovative, effective, and accurate means of identifying, isolating, and characterizing the amino acid sequences of various intracellular proteins and metabolites do not address the important mechanisms investigated by recombinant methods. Once cDNA has been proven to represent an mRNA for a specific molecule, the determined sequence of the protein suggests its functional properties. To illustrate, the acetylcholine receptor molecule and the myelin proteolipid protein exhibit several stretches of twenty to twenty-four hydrophobic amino acids — strongly suggestive of membrane-crossing domains, which therefore implies the presence of plasma membrane activation proteins. When the protein structure is finally determined, the entire molecule or selected fragments of it can be synthesized and used to produce antisera in order to develop immunoassays for the protein. The antisera can also be used for immunocytochemical analysis of the system to determine which cells and cell loci exhibit the identified target protein. Synthetic fragments can be used to determine whether the protein’s domains are substrates for posttranslational modification, are processed by further proteolytic cleavage, or are structurally modified by glycosylation, phosphorylation, sulfation, or acylation. Subcellular loci may suggest organelle specialization and cell surface marker associations. Regions around genomic exons can be probed with cDNA to discover molecular mechanisms for expression control. Once the position of genomic units has been located, cDNA probes can be used to determine the degree to which mRNA or underlying gene exons have been conserved across eukaryotic species lines and to locate their position on respective chromosomes. Although the chromosomal loci of many proteins have been determined, both human and mammalian genomes are on the order of 3 × 109 base pairs in length, of which fewer than 1,000 genes have been mapped (most on the relatively small X-chromosome), with enormous genomic expanses having no identified markers of any kind. In view of the length and complexity of gene expression, linking specific DNA-polymorphic patterns is important. Genetic linkages are usually determined by Southern blot analysis of the genomic DNA of family members, which is treated with different restriction enzymes to produce restriction fragment length polymorphisms (RFLPs). Since there can be

PH2270_book.fm Page 18 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

considerable individual variation in nucleotide sequences without disturbing an encoded protein function, the degree to which the restriction-endonuclease-digestedfragment patterns differ reflects individual differences. The ability to connect DNA fragments with the inheritance of genetic disorders and to specific markers within these digested fragments helps determine the approximate loci of mutations on a particular chromosome (e.g., determining the locus for Huntington’s chorea on human chromosome 4). Customizing purified mRNAs can be expanded far beyond simply inserting them into a plasmid. Customized designer genes have been included in and expressed by many prokaryotic or eukaryotic cellular hosts. Directly injecting mRNAs into frog eggs and allowing them to be translated and incorporated into the cellular infrastructure is useful for detecting particular mRNAs for functional proteins. A cell with a surface receptor of a known amino acid sequence has been used as a template for rational drug design by combining information from X-ray crystallography of the gene product, amino acid analysis, and functional reactions in living cells and tissue. Similar methods have already been employed to derive inhibitors for membrane lipolytic enzymes. In addition, once the mRNA has been identified, synthetic mRNAs lacking specific nucleotide segments can be evaluated to determine the function of these modified structures by indicating the active receptor sites, the membranereceptor interactions, and the ion channel or enzyme-interactive sites they regulate. Mutations can also be produced for the same purpose. Applications for gene expression go considerably beyond frog eggs. Cell lines derived from spontaneously, chemically, or virally induced tumors, or from fused hybrids of original tumor cells (hybridomas), have been effectively used for years. Depending on procedural details, ideal cell lines characteristically differ (i.e., activating a receptor cell line that transduces receptor function, or in the case of cell lines with no receptors but abundant second messenger systems, the genetic sequence encoding a receptor can be screened for its natural ligand and/or its transduction mechanism). Even more promising is the microinjection of a segment of cloned DNA into the pronucleus of a single-cell zygote to create a transgenic animal. After injecting a large number of fertilized ova, the eggs are transferred to the uterine cavity of a surrogate that has been mated with a sterile male. When the progeny are born, they are screened by checking epidermal fibroblast cultures for incorporation of the injected DNA and also for genetic sequence product expression. After puberty, the progeny’s sperm can also be evaluated for integration of a foreign DNA sample by their ability to transmit the integrated gene to the offspring. If integration of the foreign DNA is successful and if the resultant structure at the integration site is not mutagenic, the progenitor animals give rise to lines of offspring carrying the transgene (e.g., it has been possible to produce giant mice, pigs, and goats by preparing genetic constructs that induce overproduction of growth hormone). Newer applications include implementation of a retrovirus to transfect a genetic segment into cells beyond the single-cell embryo level — especially useful in systems where partial development can reveal effects of the added gene. In another application, where transfected blastocystic embryonic stem cells are returned to blastocyst embryos, they spread throughout the developing organism, and sometimes the foreign gene will appear in germ-cell lines. Changes can be made by

PH2270_book.fm Page 19 Tuesday, May 16, 2006 9:37 AM

Technology Overview

chemical modification of amino acid residues (e.g., acetylation) or by covalent attachment of ligands, although the most accurate and predictable modifications are achieved by site-directed mutagenesis of the gene encoding the protein, followed by its expression in a suitable host organism. Such advances in DNA technology have enabled the cloning of genes isolated directly from bacterial sources or via mRNA from higher organisms into expression vectors to produce both native and mutated proteins.

CELL LINES Cell-line expression technology is crucial to the success of cell-culture production. Cell-line productivity varies enormously with the host cell type, and it can be determined whether the cells will only successfully grow when attached to a solid surface (anchor-dependent) or in free suspension in a liquid (anchor-independent). The vast majority of cell-products are either monoclonal antibodies from hybridoma cell lines, or recombinant products, typically expressed in E. coli. Both cell types can be readily grown in suspension culture by careful attention to media and fermenter/bioreactor design. To insure a reproducibile process, the cell line must be stable over the number of cell divisions likely to be encountered during the process. Thus, it is important for the cell line to be clonal, and for a seed-lot system consisting of a master and working cell bank to be set up and stored in liquid nitrogen. Representative vials from the cell banks can then be tested for viruses and other potential contaminants, in addition to testing for cell-line stability. Such work should be carried out and documented to cGMP standards.

EXPRESSION SYSTEMS Recent advances have facilitated the availability of various methods for cell expression and recombinant protein production. Since many genes can be expressed in multiple systems, it should be determined which has the greatest applicability for producing a particular recombinant product. As expression systems, fermentation techniques, and downstream processing technologies develop, selection must accommodate this progress. Two key factors affect the choice of expression systems: (1) the required quantity of the designated protein, and (2) the protein’s indigenous structural complexity. Additional key factors also must be determined, including the product’s estimated market size, whether or not specific modifications are necessary for retaining biological activity, and the chemical stability of the product. The optimal expression system, then, would be one that yields maximum quantity of properly folded bioactive material. Nevertheless, there are situations in which the system expressing the highest level of a particular protein is not necessarily the system that produces the most bioactive or properly-folded product, and therefore optimal performance of a particular expression system must sometimes be weighed against maximum product quantity. Since there is no optimal system for expressing and commercially producing all recombinant proteins, each recombinant system presents its own unique challenge.

PH2270_book.fm Page 20 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

VECTOR CONSTRUCTION Specialized vectors enable efficient introduction of heterologous genetic material into host cells, and many have been developed for specific expression systems, the heterologous gene being integrated into the host genome in single or multiple copies, or located in extra chromosomal DNA fragments (plasmids) that independently replicate after incorporation into a given cell to produce many genetic copies. A recombinant plasmid incorporates an expression cassette consisting of a promoter component to regulate transcription, heterologous recombinant genetic expression material, and a transcriptional terminator. The expression cassette also contains components for optimizing transcription, translation, and secretion in the host cell — all required to express recombinant protein and transport it to its extracellular habitat. Initiation of heterologous expression, as well as endogenous expression (operational intensity), can be controlled by incorporating an appropriate promoter in the DNA construct consisting of a small portion of the DNA sequence that partially determines the endogenous expression level in the host cell (e.g., during the lytic phase of baculovirus infection of an insect cell, the polyhedron gene product represents about 50 percent of the cell’s total protein, so insect cells can be driven to express large quantities of heterologous recombinant protein following incorporation of the gene promoter into the expression cassette). Additional factors, such as the stability of its mRNA or the terminator signal employed, may also affect expression. Within the protein-coding sequence of many eukaryotic genes, there are noncoding sequences (introns) that are spliced out during genetic expression. While heterologous gene expression in bacteria and yeasts is limited to clones lacking introns, bacterial systems are unable to remove introns from mRNA sequences; and although certain yeast genes do contain introns, the yeast’s splicing mechanisms are not efficient in removing eukaryotic introns. Intron-containing genes, however, can be expressed in prokaryotic systems by using either cDNAs (e.g., DNAs that have been synthesized from a gene’s mRNA) or chemically synthesized genes that lack introns. Many heterologous proteins have been expressed in insect cells; in many instances, these proteins were demonstrated as similar to the natural moieties both in antigenicity and function. Insect cell expression levels typically range from one mg per liter to 75 grams per liter, as cell lines from other tissues and host species are cultivated, further developments in protein synthesis can be anticipated. Translational efficiency must also be high enough to attain production levels of heterologous protein expression. In translation initiation, codon implementation also affects translation elongation and termination. Highly expressed genes encoding hom*ologous host proteins demonstrate strong codon partiality that correlate with tRNA levels, although absolute correlation between codon usage and heterologous gene expression has not been demonstrated.

PROKARYOTIC CELLS The gram-negative bacterium E. coli is the best characterized, most easily grown, and most frequently used microorganism for the industrial production of recombinant

PH2270_book.fm Page 21 Tuesday, May 16, 2006 9:37 AM

Technology Overview

protein. Its superior growth characteristics are well documented. An E. coli colony can double in about twenty to thirty minutes when grown in enriched medium — one bacterium is capable of generating about 150 grams dry-cell-weight/liter in a defined medium. Yeasts such as S. cerevisiae also offer advantages for large-scale cell culture. Unlike bacteria, yeasts lack detectable endotoxins and are generally recognized as safe for the production of food and healthcare products. Yeasts provide a well-established fermentation technology and extensively characterized genetics. Yeast culture is simple, cost-effective, and rapid, with populations doubling in about 90 minutes in an enriched medium containing glucose as a carbon source. S. cerevisiae also grows on minimal medium and utilizes a variety of nonglucose carbon sources. In addition to S. cerevisiae, Pichia pastoris is used, since its gene expression is tightly regulated by methanol, providing a simple and cost-effective method for industrial fermentation. Although the produced heterologous protein may be potentially toxic to the host cell, selection against heterologous gene-containing cells can be reduced initially by growing Pichia pastoris to high densities on glucose or glycerol, which repress heterologous gene expression when regulated by a methanolinduced promoter. Subsequently, a switch to methanol as a carbon source will then initiate expression. Although inducible promoters are often used in E. coli and S. cerevisiae expression, P. pastoris strains are typically selected for efficient growth on methanol in defined minimal medium at high cell density. From the bioprocess standpoint, these yeast strains may be ideal hosts for heterologous expression, with concentrations of 130 grams dry-cell-weight/liter being reported for continuous culture fermentation.

EUKARYOTIC CELLS Scaling up recombinant protein synthesis in cultured insect or mammalian cells is usually a complex and rather costly process, since cultured insect and mammalian cells grow more slowly than microbial cells. Choosing insect or mammalian expression systems add significantly to total product costs since the complex medium required is about fifteen times more expensive than bacterial media. Media supplementation with fetal bovine serum (FBS) and/or growth promoters is expensive — FBS supplies are limited and its composition generally varies, requiring adjustments in both fermentation and purification. Mammalian cells are typically much more sensitive to the shear forces associated with traditional fermenter mixing methods using impellers and agitators. Insect cells have a greater oxygen demand than mammalian cells, are also sensitive to shear, and are subject to damage by sparging. Mammalian cells are delicate and can tolerate only narrow ranges of temperature, pH, oxygen level, and waste metabolites; large-scale mammalian bioreactors are complicated and must be designed to meet stringent sterility requirements. Scale up for insect cell systems is still in the middle stages of development, and in time, more technical obstacles will be overcome (e.g., large-scale commercial immobilized cell systems developed to protect delicate cells, reduce media volume, concentrate the product, and perform simple cell separations). In addition, airlift fermenters, hollowfiber bioreactors, and low-cost serum-free culture media are being developed specifically for insect cell culture.

PH2270_book.fm Page 22 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

EXPRESSION LEVELS Since commercial products generally must be produced in large scale to be costeffective, the protein expression level in a given host system is a major factor in choosing a heterologous expression system. Historically, bacterial systems have been the most extensively used, and such utilization has lead to a large body of literature on the subject. To protect heterologous proteins from cellular proteases, bacterial cells often accumulate them as insoluble complexes, concentrating the protein complexes within the cell as inclusion bodies. The high expression levels achieved with E. coli may be due to this protection, with expression levels as high as 30 percent of total cell protein having been obtained. Although produced in large quantities, the recombinant protein may not be properly folded, reducing the amount of bioactive product that can be recovered. Nevertheless, a large number of mammalian gene products have been synthesized in E. coli, and high expression is still considered a significant advantage. Despite the fact that secreted proteins have been produced recombinantly in bacteria, protein levels are generally much lower than those produced intracellularly, and the product is usually trapped in the periplasmic space between the plasma membrane and cell wall. Bacterial secretion systems have been developed using fusion proteins (e.g., a synthetic fragment of staphylococcal protein-A, normally secreted by gram-positive bacteria, is linked with heterologous genes, and directs 80 percent of the fusion product to the culture medium). In a defined medium with optimized conditions, the yield of a staphylococcal protein-A fragment bound to a recombinant cytokine has been reported at as much as one gram per liter. The fusion protein is then recovered using IgG affinity chromatography. The product is then separated from the protein-A fragment by chemical cleavage. Yeasts generally synthesize heterologous proteins at lower levels, but when a primary consideration is the production of a product with appropriate biological activity, yeasts have demonstrated the ability to produce biologically active moieties. Yeasts have been utilized more by the Europeans than bacteria for recombinant protein production. Super oxide dismutase (SOD) is produced in S. cerevisiae at levels of 25–30 percent total cellular protein, and, when grown in high cell density, P. pastoris synthesizes human tumor necrosis factor (hTNF) at levels reaching nearing 30–35 percent soluble protein. Many heterologous proteins have been expressed in insect cells. Also, further development in protein synthesis can be anticipated as cell lines from other tissues and host species are cultivated. Mammalian cells express relatively low protein levels on the order of tens of milligrams per liter per day under optimal conditions. Some of the highest levels of mammalian system heterologous protein expression have been produced by amplifiable vectors from Chinese hamster ovary (CHO) cells. These cells typically reach production levels per day of 3-6 × 1014 molecules of heterologous protein per 106 cells. High expression levels have also been achieved with bovine papillomavirus (BPV) vectors in mouse C127 cells, with protein production per day ranging from 1012 to 1013 molecules of heterologous protein per 106 cells.

PH2270_book.fm Page 23 Tuesday, May 16, 2006 9:37 AM

Technology Overview

INTRA-

VS.

EXTRACELLULAR EXPRESSION

There are two basic modes of expression: (1) intracellular, in which the heterologous protein accumulates within the host-cell cytoplasm, either as a soluble protein or an insoluble aggregate; and (2) extracellular, in which the heterologous protein genetic sequence has been manipulated to secrete the protein into the culture medium. Selection of either mode affects expression levels, recovery, and purification. The chosen expression mode also governs the biomolecular nature of the end product, since expression can also influence protein folding, disulfide bond formation, and posttranslational modifications.

GLYCOSYLATION The quality and extent of a protein’s glycosylation depends on the host; if glycosylation plays a significant role in the activity of a potential therapeutic, the choice of an expression system is restricted. For example, with the glycoprotein hormone erythropoietin (EPO), principal regulator of red blood cell formation, oligosaccharides are essential for in vivo function. The serum half-life of nonglycosylated EPO is measured in minutes, although when the protein is glycosylated in a CHO-cell, its half-life is about 2 hours. In another example, deglycosylation of a subunit of the protein hCG has various effects: (1) it cannot activate adenylate cyclase, although the product has an increased affinity for its receptor, resulting in a significant loss in bioactivity; (2) glycosylation may also affect protein folding, solubility, and, thereby the structural and biological characteristics of the therapeutic, and (3) glycosylation can influence the product’s antigenicity since the carbohydrate may (a) be a part of an antigenic determinant, (b) mask a potential antigenic site, or (c) affect protein conformation by influencing other potential antigenic determinants; and (4) glycosylation can affect the stability of the protein, either by stabilizing the structural conformation or by protecting it from proteases, thereby influencing recovery, purification, and in vivo half-life. In biosynthesizing mammalian cell N-linked protein glycosylation, the structures of typical mammalian N- and O-linked oligosaccharides consist of short sequences of mannose residues attached to serine or threonine, and these O-linked oligosaccharides differ from those in mammalian systems. Insufficient information is known about the structure and synthesis of insect cell glycoproteins or of their glycosylation, although it appears that insect cells can add only the high mannose N-linked oligosaccharides. Studies in mosquito cell lines indicate that the N-linked oligosaccharides produced in these species are deficient in sialic acid, galactose, and fructose. In addition, enzymes responsible for the progression of high-mannose oligosaccharide structures into complex mammalian N-glycans have only been detected at low levels in these species. The oligosaccharides of glycoproteins produced in Spodoptera frugiperda cells have only been examined indirectly, and evidence has yet to be demonstrated for the synthesis and addition of complex oligosaccharide moieties.

PH2270_book.fm Page 24 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

GENERAL CONSIDERATIONS Selecting correct cell lines can be critical for recombinant proteins that require proper glycosylation for therapeutic function. Bacterial expression systems cannot be used since prokaryotic cells cannot glycosylate. If the recombinant therapeutic requires only high-mannose oligosaccharides, then yeast, insect, or mammalian cells can do the job. However, if the therapeutic protein contains hybrid or complex oligosaccharides required for biological activity, then yeast and insect cells can be eliminated and mammalian cells should be employed. Glycosylation in mammalian cells is species-, tissue-, and cell-type-specific. For example, rat and human acid glycoproteins differ in the degree of oligosaccharide branching and glycosylation and, likewise, the amino acid sequences for rat brain and thymus tissue glycoproteins are identical — although their oligosaccharide structures are dissimilar.

FURTHER POSTTRANSLATIONAL MODIFICATIONS Further posttranslational modifications such as acetylation, phosphorylation, acylation, and carboxylation can affect the ultimate choice of an expression system, and each of these modifications should be evaluated to determine if they are required for the biological activity or stability of the product. Human super oxide dismutase (hSOD) typifies the way which a posttranslational modification can determine the choice of expression system. Human SOD, an N-acetylated enzyme that prevents oxidation damage by scavenging super oxide radicals, has been expressed at high levels in both bacteria and yeasts.

ECONOMIC CONCERNS Cost considerations are essential when one evaluates expression systems. Examples of detailed cost analyses are not readily available because most bioprocesses employed in the biopharmaceutical industry are proprietary. Product manufacturing costs, capital investment requirements, and investment returns can be developed using basic engineering principles and company experience. A thorough cost analysis should be completed early in the product development cycle, covering all available options for its expression, synthesis, and large-scale production. This analysis is critical in situations when an expression system has not been demonstrated at production levels or when the process or recovery strategies differ significantly from established procedures. Initial values must be calculated to determine the necessary capital investment and associated operating costs (including market value in dollars per unit or dose, potential market size, and share in units necessary to show the anticipated production volume), manufacturing cost, which is a function of production efficiency (protein produced per unit volume), and production cost (dollars per unit volume). Production cost is further broken down into manufacturing cost and required capital investment, which is a function of the scale of manufacturing and the equipment cost.

PH2270_book.fm Page 25 Tuesday, May 16, 2006 9:37 AM

Technology Overview

PH2270_book.fm Page 26 Tuesday, May 16, 2006 9:37 AM

PH2270_book.fm Page 27 Tuesday, May 16, 2006 9:37 AM

Introduction to the Bioprocess OVERVIEW

The distinction between the two classes of pharmaceutical substances lies in the fact that biologicals are complex mixtures of proteins and other substances, and that their potency is typically based upon their specific capacity to effect a given result. On the other hand, drugs are considered discrete chemical entities with their purity characterized and quantified.

OPTIMIZING PRODUCT YIELD The study of prokaryotic and eukaryotic cell growth fostered the development of techniques currently used to optimize product yield; such studies included: 1. Fermenter and bioreactor design, and how their distinct configurations affect the different cell-culture systems. 2. How variables such as pH, cell mass, dissolved oxygen (DO2), carbon monoxide levels, and agitation parameters affect fermenter and bioreactor productivity. 3. Understanding cellular metabolism, mass balance, and kinetics, to develop models in order to optimize fermenter and/or bioreactor processes. 4. Studying factors that contribute to productivity loss by simulating genetically engineered systems’ population dynamics. These diverse approaches notwithstanding, a detailed knowledge of cell biology and metabolism is key to optimizing biocellular fermenter and/or bioreactor processes. Modeling biosynthetic cellular processes, however, is complicated by the interaction of process control systems, thousands of individual biochemical reactions, and a widely variable extracellular environment. Nevertheless, highly detailed models incorporating many of these variables and their utility for predicting various aspects of cellular physiology have been reported.

PH2270_book.fm Page 28 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

KINETIC MODELS Highly detailed kinetic models have aided scientists investigating cell-culture metabolism in understanding the factors necessary for fermenter and bioreactor process optimization. However, in many cases, such models exceed computer memory and require too much processing time to be employed for routine process control. In process control systems, methods for monitoring parameters adverse to productivity are vital, especially the biosensors that affect the real-time measurement of key constituents. Designing feedback mechanisms and process-control algorithms to maintain optimal culture environment is also imperative. An alternative approach for decidedly complex systems might employ programmable logic controllers to provide local control at the fermenter or bioreactor level, while the process-control algorithm and a microcomputer can contribute systemwide supervisory control, thereby giving the system two levels of redundant control.

cGMP STANDARDS The development of large-scale cell-culture processing, while important to new generations of biomolecular products, is only one aspect of the bioprocess. In manufacturing cell-derived biopharmaceuticals, all process technology must be integrated and implemented to cGMP standards, as typically practiced in the pharmaceutical industry. In particular, emphasis must be placed on the use of reliable and repeatable processes to assure product consistency and reproducibility. This is accomplished by validation of both the process and the plant facilities, in other words, documented experimental proof that the process consistently yields product of the required quality, and that the manufacturing plant is functioning correctly during each production run.

CONTRACT MANUFACTURING Contract manufacturing represents a fast, efficient, and cost-effective way of moving biotechnology-derived drugs and biologicals from R&D into commercial production. It also offers expertise that may not be available to both small and large companies. Presently jurisdiction for the review of biotherapeutics is determined by indications for which similar drugs were reviewed in the past, and by applying the existing regulations. Substances classified as biologicals are regulated as licensed products by the FDA’s Center for Biologics Evaluation and Research (CBER) under the U.S. Public Health Service Act; this includes bacterial and viral vaccines, human blood and its derivatives, cytokines, and certain recombinant DNA and monoclonal antibody products. Substances classified as drug-regulated biologicals include antibiotics, hormones, and enzymes synthesized by fermentation, chemical processes, and/or genetic engineering techniques that are regulated as drugs — or more correctly, biodrugs — by the FDA’s Center for Drug Evaluation and Research (CDER).

PH2270_book.fm Page 29 Tuesday, May 16, 2006 9:37 AM

Introduction to the Bioprocess

COMPLEX MIXTURES The distinction between the two classes of pharmaceutical substances lies in the fact that biologicals are considered as complex mixtures of proteins and other substances and that their potency is usually based on their specific ability or capacity, when appropriately administered, to effect a given result. On the other hand, drugs are considered as discrete chemical entities whose purity can be characterized and quantified; another difference, which centers on the concept of regulating biologicals as processes and drugs as products, concerns manufacturing regulations, and also lies at the heart of the contrasting regulatory approaches of CBER and CDER. Manufacturing operations such as fermentation and cell culture, recovery, purification, packaging, material flow, and aspects of facility operation are meticulously performed actions that contribute to creation of the appropriate product with desired biological activity. The facilities, manufacturing, testing, labeling, and packaging procedures are delineated in both the establishment and product license applications that must be approved by CBER before the product can legally be marketed. When such licenses are granted, the manufacturer is held responsible for assuring the product is manufactured in accord with the licensing procedures.

BIOMOLECULAR PRODUCTS For biomolecular products, the Division of Product Certification at CBER stated the FDA considers the regulations listed in the Code of Federal Regulations Title 21, parts 600–680 (21 CFR 600–680), and the Good Manufacturing Practice regulations (21 CFR 200–211), to be applicable for the manufacture of these new biomolecular substances. They also note that FDA’s Guidelines for Sterile Drug Products Produced by Aseptic Processing, the various Points to Consider documents (regarding characterizing cell lines, manufacturing, and testing recombinant DNA products and therapeutic monoclonal antibodies), and the National Institutes of Health’s (NIH) Guidelines for Research Involving Recombinant DNA Molecules provide useful information for describing a manufacturing bioprocess. Such information must be included in the product license application, where applicable. Compliance with these regulations and guidelines, while onerous, can help manufacturers increase productivity and assure quality given that the biomolecular product is manufactured correctly the first time.

JOINT MANUFACTURING Another important issue affecting contract manufacture is licensing strategy for joint manufacturing. CBER has received requests for licensing multiple manufacturers under shared manufacturing relationships for biomolecular products. While these arrangements are complex in themselves, they can be even more complicated when they include contract manufacturing relationships within them; consequently, such requests are reviewed on a case-by-case basis. According to CBER, sole, shared, and divided manufacturing relationships are considered practical for licens-

PH2270_book.fm Page 30 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

ing. A sole manufacturer has a product and establishment license for all of its production phases. Also, there are no restrictions on the number of sites that can be used to manufacture the product if all sites in the establishment license application are approved by CBER and the FDA inspects and finds all these locations to be in compliance with the cGMP regulations. In a divided manufacturing relationship, individual companies are allowed to perform certain steps of the manufacturing process. Each company, however, has a product license for the entire process and has the capacity to manufacture the product in its entirety. Other requirements include an agreement between the manufacturers defining their responsibilities, that all steps be performed in licensed facilities, and that the names, addresses, and license numbers of all manufacturers appear on the label. In a shared manufacturing relationship, a license for the entire process is not needed. Rather, the final license allows several manufacturers to perform different steps of the manufacturing process. These steps are specifically defined for each manufacturer and must be performed in facilities with establishment licenses and product licenses that describe their part of the manufacturing procedure. Another requirement for these licensing arrangements is that the final licensee controls what goes on at the other facilities and that they test against the specifications of the product when they get it. Unless control is established, there can be no license. Regardless of the number of licensed manufacturers, the final use of the product as marketed is the responsibility of the final licensee. That the FDA permits these manufacturing arrangements is considered an innovation in regulatory policy. However, since FDA policy regarding such arrangements is still in a state of flux, both FDA officials and regulatory professionals emphasize the importance of early and frequent communications between the manufacturer and its regional FDA offices.

DOWNSCALING TO SCALE UP In scaling-up a bioprocess [e.g., multiple progressions (scalings) of fermenter and/or bioreactor culture capacities and downstream purification volumes], there are technical procedures that are not as efficient or as cost-effective in full-scale production as they are in either lab- or pilot-scale systems. Anticipating such problems is largely a matter of specialized expertise. Therefore, a major stumbling block for many startup and medium-sized biopharmaceutical companies is a lack of independent cGMPcertifiable contract production facilities staffed by experienced project managers familiar with the various bioprocess scale-up idiosyncrasies. Years of scaling up fermentation and cell-culture production established that many production systems could have been scaled up more quickly and more cost-effectively if both the laboratory and pilot-plant experimental runs were designed better. By reducing bioprocesses to a smaller scale, various downscaling techniques can identify and solve potential large-scale production problems before they occur (i.e., discarding a small batch because of an error is less costly in time and money than discarding a large batch). In addition, superior experimental data can be generated for scale-up process design, allowing a more accurate estimation of future bioprocess installation and operating costs. The resulting optimized production protocols can typically provide greater operation efficiency for scaled-up production.

PH2270_book.fm Page 31 Tuesday, May 16, 2006 9:37 AM

Introduction to the Bioprocess

LARGE-SCALE BIOPROCESSES Before a decision is made to build a production facility, it is necessary to know whether or not the bioprocess will work in large scale and whether it will function predictably. When production- or even large pilot-scale fermenters are not available for such process evaluation, downscaling techniques are of inestimable value in costeffectively resolving these problems and further increasing confidence in bioprocess reliability. Downscaled systems can predict the requirements of the scale-up design and the tough production environments encountered. By applying downscaling, protocols developed in both the laboratory and pilot plant can be scaled up more efficiently, bioproduction can function with fewer problems, and the end product can be brought to market in less time and at a lower cost.

SCALING-UP IN-HOUSE OR IN A CONTRACT FACILITY Although engineering firms are periodically hired to carry out bioprocess scale-up, design, and construction, downscaling techniques should begin in either the manufacturer’s laboratory and pilot facility or in an independent contract bioprocess facility experienced in scale-up technique. Unfortunately, with the wave of recent acquisitions, there seem to be few, if any, independent commercial contract facilities available to carry out these projects, with the fermenter or bioreactor culture capacities and downstream purification volumes invariably required to manufacture a biopharmaceutical product in production quantities.

LAB-TO-PILOT-PLANT-TO-PRODUCTION The movement of a process from the laboratory to a pilot-plant to production facility is called scale up. If a scale up is to be successful, projections for yield, cost, and efficiency must be accurate. Scale-up processes must also overcome technical difficulties and limiting factors, as must be identified in bioprocesses. Common scale-up parameters in continuous-culture systems are generally based on tank geometry, oxygen supply, chemostat volume, substrate concentration, critical dilution rate, cooling capacity, mixing effect, loss of cell population (washout), and agitator power input. As the process scale increases, certain parameters change as the result of either the increased fermenter size or the changed nature of its operation. These features must be considered when scaling either up or down. Chemical features include pH control agents, growth medium, and water quality. Physical features include tank configuration, fermenter volume, aeration, agitation, pressure, sterilization, and temperature control. Finally, there are biological features including the probability of mutation, contamination, and selection pressure. Lab-scale processes are typically used to screen microorganism or cell strains, to optimize media and culture conditions, and to develop separation and purification methods. Pilot-scale production is generally used to verify optimal large-scale process conditions and to produce sufficient data for the appraisal of large-scale production efficiency. Finally, to commercialize a product, the process must be transferred to a large-scale production facility.

PH2270_book.fm Page 32 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

INCREASING BIOPROCESS SCALE As the process scale increases, certain parameters change as a result of either the increased fermenter size or the changed nature of its operation. Such features must be considered when scaling up or down. Chemical features include pH control agents, growth medium, and water quality. Physical features include tank configuration, aeration, agitation, pressure, medium sterilization, and temperature control. Finally, biological features include the probability of mutation and contamination as well as selection pressure. A key biological factor influenced by up-scaling is the total number of generations necessary for the cultivated strain to produce a given amount of product. For example, production volumes 1,000 times larger than laboratory or pilot scale require 10 additional generations to reach end stage. For genetically engineered strains, this larger generation must maintain stable plasmids harboring the production gene. Thus, a bioprocess may have many differing factors when implemented at different scales. A recombinant strain and its bioprocess developed under laboratory- and pilot-scale conditions may be neither optimal nor practical at production scale. Prudent planning requires bioprocess analysis, verification in a pilot plant, and then another scale up for production. Certain biological features that seem unimportant at smaller scale may have significant influences on the function, design, and operation of production processes.

DIFFERING FEATURES AT LARGER SCALE A bioprocess, therefore, may have many features that differ when implemented at different scale. A recombinant strain and its bioprocess developed under laboratory and pilot-scale conditions may be neither optimal nor practical at production scale. Prudent planning requires that the bioprocess be analyzed, verified in a pilot plant, and then scaled up again for production. Unfortunately, additional effort and time are required for such process revisions, and this may result in increased development costs and delayed marketing schedules. Downscaling techniques, on the other hand, simulate production conditions in small-scale runs so that technical problems in production-scale bioprocesses can be pinpointed and solved before their implementation. For those companies or researchers who do not have access to large- or pilotscale fermentation facilities to assess their process protocols, downscaling techniques can be quite cost-effective as well as practical. Certain biological features that seem unimportant at smaller scales may have significant effects on the function, design, and operation of production processes. These features should be identified and studied at the downscaling stage and during a strain’s screening and improvement program; and after an optimal strain has been selected, the production process itself must be optimized. Downscaling techniques can help to identify and implement the best conditions for a specific production-scale bioprocess.

SIMULATING ENVIRONMENT AT SCALE Optimal laboratory conditions may not correspond to those in a production environment, and the key to developing an optimal production protocol is simulating the

PH2270_book.fm Page 33 Tuesday, May 16, 2006 9:37 AM

Introduction to the Bioprocess

particular scale’s environment as much as possible during bioprocess development. For example, if an entire fermentation process in production scale is known to require 22 generations, then subcultures should be used to simulate this generation number in the laboratory fermenter.

DETAILED SPECIFICATIONS Scaling-up a fermentation or downstream bioprocess requires detailed specifications, usually available from process development and optimization studies. Basic data, such as DO2, pH profile, and the amount of acid or base added, should be gathered with great care since a scale up is only as good as the input data provided (“Garbage-in, Garbage-out!”). Other than basic data, much information required for successful scale up design is generally overlooked during process development and optimization stages. Some of this data is obtained by monitoring obscure parameters while other data can only be generated through experiments designed specifically to produce such scale up data. For a more accurate scale up design, lab- or pilot-scale fermenters with standard geometry should be used since many scale up calculations are based on empirical equations that assume similar tank geometries. Also of importance is the size of the pilot-scale fermenters used to collect scale-up data. In general, 10 liter laboratory fermenters are not good choices for collecting data, since their design usually differs significantly from that of large-scale production fermenters in which the liquid-surface area to tank-volume ratio is much smaller and the surface vortex is insignificant. To scale up from a small fermenter, with its more prominent surface aeration, almost requires the exclusion of surface aeration effects from total oxygen transfer.

IMPACT ON DOWNSTREAM BIOPROCESSING The objective of a production-scale downstream purification process for a recombinant therapeutic protein is to secure a fully bioactive and unaltered product, free of impurities and pyrogens that contains from about 10 to 100 picograms of foreign nucleic acid per dose. Even in the early stages of developing the expression system, purification designers must consider the impact of fermentation or cell culture on downstream processing (i.e., when selecting the host promoter and vector). Production host, location, and the physical form of the biomolecular end product dictate the selection and sequence of biomolecular isolation steps in the total purification scheme. Prechromatographic steps include separating cells from the medium, biomolecular release (sometimes with inclusion bodies), as well as protein unfolding and refolding. Only chromatography offers the high resolution necessary to separate an ultra-pure and active end product. Thus, prepurification steps are aimed at volume reduction, clarification (removing nucleic acids and cell debris), and chemical adjustment of the crude extract. These can then be viewed as preparation for later chromatographic separation. Expression levels greater than 20 percent of total soluble cellular protein can be attained from bacterial hosts, three to five percent from yeasts, and milligram-per-liter concentrations from eukaryotes — which obviates the need for multi-thousand-fold expressed protein purification. The challenge

PH2270_book.fm Page 34 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

here is to eliminate impurities and increase purification levels from 95% to nearhom*ogeneity. This requires the precise separation of the target biomolecular protein from very similar expressed molecules that may have resulted from errors in coand posttranslational modifications. During fermentation and downstream processing, improper refolding, dimerization, deamidation, proteolytic nicking, and oxidation also contribute to heterogeneity.

FDA BIOLOGIC REVIEW PROCESS While most discussions of the FDA focus on the agency’s authority to decide which new treatments reach the U.S. marketplace, the agency also plays the role of regulatory gatekeeper at a critical point in the biological development process. In reviewing investigational new drug applications (INDs), the FDA determines which experimental therapies, vaccines, and other biological products advance from preclinical through clinical development to the licensing phase. When a sponsor submits an IND, the FDA assumes a key role in developing a new biological product. Most sponsor activities beyond preclinical development are subject to some form of FDA oversight, since these activities involve human subjects, their health, and their welfare. This dose of reality is demonstrated by the following statement a sponsor must sign in filing an IND (Form FDA 1571): “I agree not to begin clinical investigations until 30 days after FDA’s receipt of the IND unless I receive earlier notification by the FDA that the studies may begin. I also agree not to begin or continue clinical investigations covered by the IND if those studies are placed on clinical hold. I agree that an Institutional Review Board (IRB) that complies with [federal regulations] will be responsible for the initial and continuing review and approval of each of the studies in the proposed clinical investigation. I agree to conduct the investigation in accordance with all other applicable regulatory requirements.” While significant essentials of the FDA’s biological product review process have been reformed under the Prescription Drug User Fee Acts I, II, and III (PDUFA I, II, and III) and the FDA Modernization Act of 1997 (FDAMA), the agency’s longstanding IND review process has largely survived the wave of regulatory reform implemented over the past decade. In contrast to biological product development and approval, which has weathered everything from review timelines to redefined, tightened, or tweaked sponsor communication processes, and which face several additional changes under the newly enacted PDUFA III, it is arguable that only peripheral aspects of the IND review process were affected by these legislatively driven reform efforts. However, the biological licensing and the biological IND review processes have been equally affected by a restructuring initiative that the FDA implemented in mid-2003. The drug development process for biological products comprises several key phases, including (1) preclinical development, (2) clinical development, (3) licensing (biologics license application), and (4) approval. This chapter outlines the biological IND review process, through which an experimental product must pass to advance from the preclinical phase to the clinical development phase. Before outlining this process, however, it is appropriate to first profile CBER and CDER, the two regulatory and scientific units responsible for processing and evaluating biological INDs and biological license applications (BLAs) for these

PH2270_book.fm Page 35 Tuesday, May 16, 2006 9:37 AM

Introduction to the Bioprocess

products. Editor’s note: Although the term “biological IND” is used throughout this chapter, CBER sometimes also reviews INDs for drug and investigational device exemptions (IDE) for medical devices. In most cases, CBER’s review processes for, and its approaches to, these applications are similar — if not identical — to those for biological INDs.

REGULATING BIOLOGICAL PRODUCTS FDA HISTORY

AND

AGENCY STRUCTURE

Although the BLA served as the basis for biological product licensing only since 1996, it has been at the center of regulatory reform throughout it relatively brief existence. To provide a better understanding of the regulatory basis for the BLA, it is appropriate to briefly recount the history of U.S. biological product licensing. The FDA’s current authority over biological products derives from Section 351 of the Public Health Service Act and from selected sections in the Food, Drug and Cosmetic Act. Because of this relationship, biological products are considered to be drugs. Specifically with respect to biological products, the FDA is responsible for ensuring: •

• •

The safety of the nation’s blood supply, derivative products, test kits used to screen blood donors, and devices used for the manufacture blood products; The approval and availability of safe and effective vaccines and allergenic products; The approval, safety, and effectiveness of biological therapeutic products, including monoclonal antibodies, cytokines, cellular and gene therapies, and xeno-transplantation.

From late 1987 through June, 2003, all biological products were regulated by the FDA’s CBER. CBER was created in 1987 as part of an FDA reorganization plan to separate the reviews of drugs and biologics, which were previously managed by a single scientific and administrative organization. The reorganization split the FDA’s Center for Drugs and Biologics into two distinct centers: CBER and the CDER. At the time, the FDA believed this new structure provided a more definitive authority for drug and biologic review. In September 2002, the FDA again announced their intention to restructure the scientific and administrative units regulating biological products in order to eliminate inefficiencies and provide a more consistent regulation and product review, and decided to shift the review responsibilities and the staff involved in regulating therapeutic biological products from CBER to CDER. Agency officials asserted that this controversial decision, met with several high-level resignations within CBER, was predicated on simplifying procedures and thereby effecting cost savings by merging practices and resources to regulate biological therapeutics. By supplementing CBER’s product and technology-related expertise with CDER’s disease-specific medical review expertise, FDA officials argued this would make the decision process even more patient-centric and scientifically based.

PH2270_book.fm Page 36 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

Following this announcement, a CDER-CBER consolidation task force initiated a three-phase process that culminated in the reassignment of an estimated 208 CBER staff, the responsibility for more than 54 approved therapeutic biological products, and 1,500 active INDs to CDER — first through a temporary June 2003 staff directive, and subsequently by an October 2003 transfer of staff and funding. Once implemented, the reorganization shifted the responsibility for the following therapeutic biological products to CDER: • • • •

Monoclonal antibodies for in vivo use Cytokines, growth factors, and immunomodulators Enzymes Thrombolytics

THE BIOLOGICAL LICENSE APPLICATION (BLA) U.S. biological product regulatory structure never experienced such a fundamental change as it has in the past decade. Only a few years after completing the transition from the FDA’s long-standing dual-license approval format to a more streamlined single-license system, the FDA undertook yet another series of reforms that continued to reshape the future of biological product regulation in the United States. As before, the biological licensing process is at the center of a current series of regulatory reforms. As of this writing, the process has entered its latest evolutionary phase, driven by two important developments: in late 2001, CBER and CDER implemented a temporary and optional BLA format, based upon the International Conference on Harmonization, i.e., a common technical format that companies could employ in lieu of the standard BLA or new drug application (NDA) formats. In July, 2003, this transition phase expired, and the use of the international harmonization format was recommended by the FDA. Although the traditional formats continued to be those typically employed by industry, by late 2003, international harmonization conference-formatted submissions were more frequent and CDER was feverishly laying the groundwork for its next evolutionary stage: a shift to electronic format submission (E-format). In an August 2003 draft guidance Providing Regulatory Submissions in Electronic Format — Human Pharmaceutical Product Applications and Related Submissions, the FDA explained how to submit BLAs, NDAs, INDs, and other applications in the new electronic format. From 2001 through 2003 nearly half of all BLAs submitted were fully electronic (i.e., no accompanying paper other than signatory or other legal documents), and CBER was already well on its way to completely transitioning to e-format submission and electronic review. In July, 2003, after receiving numerous electronic submissions, CDER prepared a web posting announcing that they wished to work closely with any companies that planned submissions in the new e-format. Although CBER and some applicant companies had discussed submissions based on the electronic format in late 2003, CDER developed the specifications and appeared to encourage use in this early pilot phase, since they hoped that pilot applications would focus on e-format for filing INDs, where this could show maximum benefit, since IND submissions typically extend over longer time periods.

PH2270_book.fm Page 37 Tuesday, May 16, 2006 9:37 AM

Introduction to the Bioprocess

RESPONSIBILITY

FOR

THERAPEUTIC BIOLOGICS

After nearly two decades, the FDA implemented a consolidation where the regulation, review, and approval responsibility for therapeutic biological products was transferred from CBER to CDER. Designed to eliminate inefficiency and provide consistency in both regulation and review, the move supplemented CBER’s product and technology-related expertise with CDER’s disease-specific expertise. In October 2003, after assigning to CDER more than 200 CBER staffers from the Office of Therapeutics Research and Review, the reorganization was formalized through this transfer of staff and funding. This transfer, however, did not affect the legal or regulatory status of therapeutic biological products, since they continue to be regulated as licensed biological products and were reviewed under the same BLAs relevant to all other biological products. For reasons of administrative simplicity and continuity, four divisions of CBER’s former Office of Therapeutics Research and Review (OTRR) were transferred intact to CDER, and have been operating within two new biologic-specific CDER offices. CDER officials characterized this initial organization as interim and added that former OTRR divisions and staff would ultimately be absorbed into the CDER structure as part of a comprehensive restructuring of the drug center in mid-2005. Company submission of a BLA represents a culmination of years of clinical investigation and product development. After a biologic’s safety and efficacy have been demonstrated by means of the IND, the company’s sponsor must compile a BLA, which is reviewed as the documented basis for the product’s approval, and subsequent commercial distribution in the United States. By providing both clinical and nonclinical data, information on physico-chemical characterization, biological activity, product manufacturing, and manufacturing facility information, the BLA enables FDA reviewers to make determinations as to whether: • • • •

The biologic is safe and effective in its indicated use; The biologic’s benefits outweigh its risks; The proposed labeling is appropriate; Manufacturing methods and quality control are adequate to preserve the identity, strength, quality, potency, and purity of the biomolecule.

The manufacturing process is capable of producing a product that is consistent with the specifications in the application, cGMP, and other relevant regulations.

PH2270_book.fm Page 38 Tuesday, May 16, 2006 9:37 AM

PH2270_book.fm Page 39 Tuesday, May 16, 2006 9:37 AM

Recombinant DNA Materials and Methods CLONING

When mRNAs are converted into DNA by the enzyme reverse transcriptase (see Chapter 5), the cDNA can be incorporated or inserted into specific sites within an infectious vector, usually a virus or bacteriophage, or even into an extrachromosomal genetic element (plasmid) found among various strains of bacteria. The insertion sites are selected by identifying DNA sequences that can be cut by the actions of restriction endonucleases (see Chapter 5), enzymes from purified bacterial sources that cleave DNA sequences at specific palindromically repeated sequence sites. By using plasmids of known DNA sequence, tailored to include restriction cleavage sites that allow for DNA insertion, the same enzyme can later be used to, again, cleave out the insert. Restriction sites for insertion are typically chosen within plasmid genes that code for some discernible functional property (such as antibiotic resistance). Thus, when insertion has been successful, interruption of coding and expression leads to the loss of functional property and permits the identification of plasmids with effective inserts. In general, each plasmid can only incorporate one cDNA insert, and with a great excess of host bacteria, each insert-bearing plasmid will infect only a single host, usually E. coli. By growing these genetically altered bacteria in such a way that each individual bacterium gives rise to a colony of identical bacteria carrying replicates of the plasmid and insert, the DNA is successfully cloned. The cDNA can then be recovered from the plasmid through another exposure to the restriction enzyme selected for the original opening of the plasmid insertion site. Thus, in relatively few steps one can begin with a collection of mRNAs, from common to very rare, purify them individually, and develop virtually unlimited pure copies of the DNA insert (see Figure 4.1 and Table 4.1). The PCR processes large amounts of specific rare nucleic acid sequences through amplification in vitro without the necessity of first purifying the desired sequences through cloning (see Chapter 5). The changes in biotechnology research brought about by this new technology were immediate and dramatic. Using previously selected restriction enzymes, human DNA was sliced into small sections, one of which contained the target gene sequence. When the DNA was heated to 95°C, the double-stranded DNA dissociated and the strands separated. Two small DNA sequences synthesized oligonucleotides to be complementary to opposite strands of the target gene. These oligonucleotide primers hybridized to their complementary sequences on the single-stranded DNA (see Figure 4.2). In the presence of large 39

PH2270_book.fm Page 40 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

Tissues, Cells, Fractions, or Extracts

Protein mRNA

mRNA

Tissue cDNA (mRNA) Library

Southern Blots

Northern Blots

Known mRNAs Genomic Library

Synthetic cDNA Probe

Antibodies Western Blots & RIAs Expression Systems

Microscopic Localization

In Situ Hybridization Protein Sequence Data Bank

Identified Proteins (Natural or Synthetic)

FIGURE 4.1 Cloning.

amounts of a purified DNA polymerase and large amounts of all the deoxynucleotides, the primers were extended to the end of the single strand. Thereafter, the cycles of denaturing, annealing, and extending reactions, each lasting only a few seconds, can be rapidly repeated by separating the dual-helical strands of the newly synthesized material through heat denaturation, cooling the reaction mixture, and adding fresh DNA polymerase and nucleotides. As long as the polymerase and the nucleotide substrates are in excess, the extended sequences of the first reaction serve as templates for opposite strand synthesis in subsequent cycles, providing a geometric rate of amplification. With DNA polymerase isolated from a specific bacterial strain that grows in the extreme heat of geysers, it is then possible to develop a method permitting large quantities of polymerase to survive the heat denaturation step. With this polymerase and large beginning amounts of dnucleotides, a series of rapid cycles could produce the exponential in vitro amplification of the desired DNA sequence. If selected sequences are known to be generally constant in the genome of a species, the same primers can be used to select, amplify, and subsequently analyze the same intervening gene segment from many individuals. If the amplified segment is simply analyzed for its length, it is possible to determine whether a given individual has a major genetic mutation (e.g., a deletion or an insertion). Additional applications of the technology have increased its advantages. PCR can also be applied to mRNAs and can even provide a quantitative analysis by first using reverse transcriptase to make double-stranded cDNA. By modifying the ends of the probes to be used, it is also possible to incorporate special synthetic sequences that make it easier to clone or sequence amplified segments, or to reinsert modified versions to determine the functional importance of a specific sequence. Even if the sequences do not match precisely, it is also possible to amplify hom*ologous sequences from different, but related, genes by

PH2270_book.fm Page 41 Tuesday, May 16, 2006 9:37 AM

Recombinant DNA Materials and Methods

TABLE 4.1 Construction of Host Strain for Production Author: Date: Notes: Supersedes: Bacterial strains: Phage strains:

Protocol: Phage strains for selecting lysogens:

Earl Jones November 4, 1989 Protocol for construction of DE3 lysogen; origin of strains and phages used None W3110 and LE392.23 Lambda-DE3: imm21 cI+ carries T7 RNA polymerase under lac control Lambda-B10: imm21 cILambda-B482: imm21 cI-(att) Delta-1 h80 T4107: T-phage from which the entire T7 RNA polymerase gene is deleted Procedure for making DE3 lysogens DE3 imm21 cI+ carries inducible gene for T7 RNA polymerase B10 imm21 cIB482 imm21 cI-(att) Delta-1 h80 Use lysates grown on a modifying host when making lysogens of Eco*k+ strains. Tryptone broth or equivalent is suitable for cultures and in agar plates. Mix 108 each of DE3, B10, and B482 in 2.5 ml top agar, add 1–10 ml of a fresh culture of the cells to be lysogenized, and pour onto a 20 ml agar plate. Most colonies that grow after incubation overnight at 37°C should be DE3 lysogens. Grow a small culture from one of the colonies and purify a single colony from this culture. Test the purified strain for immunity and for inducible T7 RNA polymerase activity. T7 grows on many female, but few male, strains of E. coli. A simple test for T7 RNA polymerase activity in cells that plate T7 is to test for ability to plate 4107, a T7 mutant from which the entire T7 RNA polymerase gene has been deleted. 4107 is totally unable to form a plaque on cells that lack T7 RNA polymerase, but it forms normal plaques on a DE3 lysogen in the presence of an inducer (0.025 ml of 0.1M IPTG added to the top agar). When plated on a DE3 lysogen in the absence of the inducer, 4107 typically forms small plaques that take a long time to develop.

using special nucleotides that will form complementary base pairs. The above examples indicate the simplicity of cloning DNA segments taken directly from genomic digests or from mRNA copies, sequencing the cloned segments, and discerning the structure of the product. Nevertheless, how does one determine which clone is carrying the insert that encodes for a specific gene product? And what if one has not determined this sequence so that it can be located and cloned? Scientists have found various techniques to gain these objectives, although some of the screening methods are extremely tedious (see Figure 4.3).

APPLICATIONS The phenotype of a cell depends upon the structural, metabolic, and regulatory proteins by which recognizable physical and functional properties are determined. Complex multifunctional cells rely upon myriads of special-purpose proteins, many

PH2270_book.fm Page 42 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

T A

G A

TA A T TA AT

A T G

T A C C

C G

GC TA CG

T C

A A

T T

G C

A T

AT GC CG

G T

T T

Complementary Strands of Nucleic Acid

Hybrid

FIGURE 4.2 Hybridization of complementary DNA strands.

Double-Stranded Sample DNA Denaturation

Single-Stranded DNA Bind to Solid Support

Labeled, Single-Stranded DNA Probe

Membrane

Hybridization

Wash Off Excess Probe Detect Bound Probe Positive Signal

FIGURE 4.3 Schematic of a solid-phase DNA probe assay.

PH2270_book.fm Page 43 Tuesday, May 16, 2006 9:37 AM

Recombinant DNA Materials and Methods

of which may exist in rather limited amounts. Purifying such scarce proteins by premolecular cloning methods, especially in the absence of a functional assay, was an enormous task requiring patience, resources, and a very large supply of cellular material. The structure of a specific protein directly relates to the mRNA or gene segment that encodes the protein. It depends on the nature of the desired cDNA, whether or not you can identify its insertion or determine the genetic fusion product translation in a biological infrastructure capable of processing the translation product into a form that replicates its natural structure and function.

SCREENING AND SELECTION Selecting cells that express or are presumed to express a target molecule, and then enriching sources of mRNA to favor target molecule detection is a practical beginning. Hormone-producing cell lines and tumors or cells bearing large numbers of desired receptors or channels (e.g., striated muscle) are examples of excellent starting material. Once the cell source is selected, the desired mRNAs are further enriched by sucrose gradient centrifugation or electrophoresis, provided some of the characteristics of the mRNA being sought are known. In colony hybridization, a common strategy for detecting desired colonies of cloned bacteria, the bacteria are grown on a special culture plate from which the colonies can be transferred as a group by replica plating (e.g., lightly pressing them to another supporting surface), thereby sampling and preserving the spatial identity of all colonies on the plate. Bacteria on the replicate supports are screened with nucleic acid probes to identify target colonies. If the plasmid-carrying inserts are tailored to allow for expression of the protein encoded by the transferred genetic material, it is also possible to identify desired clones by immunoassay. When a target colony has been identified, the original colony is recovered from the primary culture plate and subsequently cultured in large quantities to provide material for DNA sequence analysis.

DNA SYNTHESIS It is possible to predict an mRNA sequence, if partial protein or peptide sequences are known, by back-translating the genetic amino acid code and including enough alternatives to overcome ambiguous cases where a specific amino acid might be encoded by several variant triplets, then, design and synthesize the cDNA from the predicted RNA structure. This approach has been used to create theoretical cDNAs for hormones with sequences that have been determined one amino acid at a time from highly purified tissue extracts. Generally performed with a structurally known biologically active complex protein or peptide, the procedures determine either the product’s structure or the complete genomic configuration necessary to produce its regulatory and expression mechanisms. A target peptide’s precursor, however, is generally found to encode more than one active product, and because of the redundancy of triplet RNA codons for some amino acids, it is generally difficult to acquire a functional full-length mRNA by predictive synthesis. An alternative approach is to synthesize a DNA probe which is a shorter, complementary, single-

PH2270_book.fm Page 44 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

stranded DNA used as a probe to screen clone libraries, which are prepared either from mRNA extracts or from whole genomic digests (see Figure 4.3). In the former case, the starting material would be tissue, while in the latter case any somatic cells could be used to prepare the library. With the availability of automated genetic sequencers, it is now possible to synthesize probes overnight, use them to screen a genomic or cDNA library, and subsequently determine within a few weeks the complete coding sequence for a partially purified protein. Candidate colonies can also be cross-screened by a second probe based upon another separate domain of the full protein. Clones positive for both probes, therefore, should contain the gene sequence that encodes the two sequences against which the probes were made, as well as the sequence between them. This strategy has been used with many neuropeptide mRNAs. It is also possible to examine many cellular proteins that, as yet, have not been identified by conventional methods. Given the length of the mammalian genome and the relatively small group of specific identified intracellular biomolecules, it can be surmised that there are countless more unidentified biomolecules and few clues as to their identity. By exploiting the fact that a high proportion of expressed tissue proteins are quite similar, the method can become highly sensitive, revealing unique differences in cell-specific gene expression. The technique can also be expanded to track large groups of tissue-specific genetic sequences (e.g., such as employing molecular cloning methods to find out what proteins are ordinarily manufactured by specific tissues but not found in other major tissue groups) and to determine the degree to which cells differ in specific phenotypic proteins. Determining their nucleotide sequences and then deriving the amino acid sequence from the encoded protein can further examine individual tissue-specific clones. Developing antisera against synthetic peptides that mimic selected regions of the derived protein structure can further identify proteins unique to known sequences. Another technique involves injecting mRNAs from enriched or prepared sources into frog eggs (oocytes), where the complex eukaryotic genes can be expressed more efficiently. Because the frog eggs are relatively large, their expression of a novel functional protein can be assessed either physiologically or biochemically to identify mRNA species for isolation and cloning. By injecting groups of mRNAs and evaluating the eggs for the desired response, through trial and error, it is possible to identify the mRNA for a specific functional protein (i.e., for a cell surface receptor). Utilizing the Southern blot (named for Dr. E.M. Southern, who devised the method) after the DNA is cleaved by restriction enzymes and separated by gel electrophoresis, it is possible to transfer or “blot” the resulting fragments (separated primarily on the basis of their lengths) from the gel to a nitrocellulose support and then analyze them for their ability to hybridize with cDNA or RNA probes. A similar method uses RNA as the starting material, where the separated RNAs are blotted for probing with single-stranded cDNA probes. Since the starting material is the opposite of that of the Southern method, the RNA blot has been referred to as the “Northern” blot. Recently, immunological methods were used to examine protein extracts separated by PAGE and then blotted by electrical transfer for subsequent identification by labeled antibodies to specific protein antigens. This method is called a Western blot. If RNA or protein specimens

PH2270_book.fm Page 45 Tuesday, May 16, 2006 9:37 AM

Recombinant DNA Materials and Methods

are dried on nitrocellulose for analysis without first separating them for size, the resulting blots are called “slot” or “dot” blots, which are useful when quickly screening large numbers of clone extracts for inserts or expressed products. Increasingly, a large number of molecules are being fully characterized, enabling scientists to create biodrug molecules that precisely either fit specific receptors or enzymes, thus demonstrating that any biological event mediated by proteins is receptive to biomolecular analysis. The range of such biomolecules includes enzymes needed for hormone synthesis, storage, release, and catabolism; receptors and related macromolecules needed for response; and those needed for response mediation of a broad-scale of physiological events. Methods that provide innovative, effective, and accurate means of identifying, isolating, and characterizing the amino acid sequences of various intracellular proteins and metabolytes do not address the important mechanisms investigated by recombinant methods. Once cDNA has been proven to represent an mRNA for a specific molecule, the determined sequence of the protein suggests its functional properties. To illustrate, the acetylcholine receptor molecule and the myelin proteolipid protein exhibit several stretches of twenty to twenty-four hydrophobic amino acids that are strongly suggestive of membranecrossing domains and, therefore, imply the presence of plasma membrane activation proteins. When the protein structure is finally determined, either the entire molecule or selected fragments of it can be synthesized and used to produce antisera to develop immunoassays for the protein. The antisera can also be used for immuno-cytochemical analysis of the system to determine which cells and cell loci exhibit the identified target protein. Synthetic fragments can be used to determine whether the protein’s domains are substrates for posttranslational modification, are processed by further proteolytic cleavage, or are structurally modified by glycosylation, phosphorylation, sulfation, or acylation. Subcellular loci may suggest organelle specialization and cell-surface marker associations. Regions around genomic exons can be probed with cDNA to discover molecular mechanisms for expression control. Once the position of genomic units has been located, cDNA probes can be used to determine the degree to which mRNA or underlying gene exons have been conserved across eukaryotic species lines and to locate their position on the respective chromosomes. Although the chromosomal loci of many proteins have been determined, both human and mammalian genomes are on the order of 3 × 109 base-pairs long, of which fewer than 1,000 genes have been mapped (most on the relatively small X-chromosome), with enormous genomic expanses having no identified markers of any kind. In view of the length and complexity of gene expression, linking specific DNA polymorphic patterns is important. Genetic linkages are usually determined by Southern blot analysis of the genomic DNA of family members, which is treated with different restriction enzymes to produce restriction-endonuclease-digested-fragment patterns (RFLPs). Since there can be considerable individual variation in nucleotide sequences without disturbing an encoded protein function, the degree to which the RFLPs differ reflects individual differences. The ability to connect DNA fragments with the inheritance of genetic disorders and to specific markers within these digested fragments helps determine the approximate loci of mutations on a particular chromosome (e.g., determining the locus for Huntington’s chorea on human chromosome 4).

PH2270_book.fm Page 46 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

The customization of purified mRNAs can be expanded far beyond simply inserting them into a plasmid. Customized designer genes incorporate the basis for being included in and expressed by either prokaryotic or eukaryotic novel cellular hosts. Directly injecting mRNAs into frog oocytes and allowing them to be translated and incorporated into the cellular infrastructure is useful for detecting particular mRNAs for functional proteins. A cell with a surface receptor of known amino acid sequence could be used as a template for rational drug design by combining information from X-ray crystallography of the gene product, amino acid analysis, and functional reactions in living cells or tissue. Such methods have already been employed to derive inhibitors for membrane lipolytic enzymes. In addition, once the mRNA has been identified, synthetic mRNAs lacking specific nucleotide segments can be evaluated to determine the function of these modified structures by indicating the active receptor sites, the membrane-receptor interactions, and the ion channel or enzyme-interactive sites they regulate. Mutations can also be produced for the same purpose. The application of genetic expression goes considerably beyond frog eggs. Cell lines derived from spontaneously, chemically, or virally induced tumors, or from fused hybrids of original tumor cells (hybridomas), have been effectively used for years. Depending upon procedural details, ideal cell lines characteristically differ (i.e., activating a receptor cell line that transduces receptor function; or in the case of cell lines with no receptors but abundant second messenger systems, the genetic sequence encoding a receptor can be screened for its natural ligand and/or its transduction mechanism). Even more promising in its biopharmaceutical implications is the microinjection of a segment of cloned DNA into the pronucleus of a single-cell zygote to create a transgenic animal. After injecting a large number of fertilized ova, the eggs are transferred to the uterine cavity of a surrogate that has been mated with a sterile male. When these progeny are born, they are screened by checking epidermal fibroblast cultures for incorporation of the injected DNA and also for genetic sequence product expression. After puberty the progeny’s sperm can also be evaluated for integration of a foreign DNA sample by their ability to transmit the integrated gene to the offspring. If integration of the foreign DNA is successful, and if the resultant structure at the integration site is not mutagenic, the progenitor animals give rise to lines of offspring carrying the transgene (e.g., it has been possible to produce giant mice, pigs, and goats by preparing genetic constructs that induce overproduction of growth hormone. Outstanding among many novel genetic engineering applications is the preparation of a gene construct in which the regulatory domains of a known gene (e.g., regulatory domains for expressing a neuropeptide) are coupled with the expression of a novel reporter gene (e.g., an enzyme that is normally absent). Under such conditions, one can evaluate whole-animal or whole-cell behavior, so that if mature cell lines are transfected and activate the reporter gene, this activation provides deductive evidence of activation control of the natural gene. Newer applications include implementation of a retrovirus to transfect a genetic segment into cells beyond the single-cell embryo, which is especially useful in systems where partial development can reveal effects of the added gene. In other applications, where transfected blastocystic embryonic stem cells are returned to blastocyst embryos, they spread throughout the developing

PH2270_book.fm Page 47 Tuesday, May 16, 2006 9:37 AM

Recombinant DNA Materials and Methods

organism and sometimes the foreign gene will appear in the germ-cell lines. Changes can be made by chemical modification of amino acid residues (e.g., acetylation) or by covalent attachment of ligands, although the most accurate and predictable modifications are achieved by site-directed mutagenesis of the gene encoding the protein, followed by expression in a suitable host organism. Such DNA technology advances enable the cloning of genes isolated directly from bacterial sources or through mRNA from higher organisms transferred into expression vectors, producing both native and mutated proteins.

CELL LINES The expression technology of the cell line is crucial to the success of bioproduction by cell culture or fermentation. Cell-line productivity can vary enormously and the choice of a host-cell type can determine whether the cells will successfully grow when attached to a solid surface (anchor-dependent), or grow in free suspension in a liquid (anchor-independent). The vast majority of biomolecular cell products are either monoclonal antibodies from hybridoma lines, or recombinant products expressed in E. coli. (see Table 4.2). Both cell types can be readily grown in suspension culture by careful attention to media formulation and fermenter/bioreactor design. To insure a reproducible process, the cell line must be stable over the number of cell divisions likely to be encountered during production. Thus, it is important for the cell line to be clonal, and for a seed-lot system consisting of a master and working culture or cell bank to be set up and stored in liquid nitrogen. Representative

TABLE 4.2 E. coli Host Strain for Expression Author: Date: Notes: Supersedes: Host organism genus and species: Designation: Other: Genotype: Derivation: Biosafety level: Reference: Toxic or pathogenic characteristics: Disposal: General:

John Doe November 4, 1989 Description of E. coli host strain (W3110/DE3) including organism, description, biosafety level, and genetic markers None E. coli K-12 W3110 Lambda-DE3 lysogen F- mcrA mcrB IN(rrnD-rrnE)1 Lambda-DE3 Lambda-DE3 lysogen of ATCC Strain W3110 BL1 Federal Register, Vol. 51, No. 88, May 7, 1986 None known Autoclave Organism exhibits normal E. coli growth characteristics. There are no special regulatory issues that pertain to the use of this strain.

PH2270_book.fm Page 48 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

vials from cultures or cell banks can then be tested for viruses and other potential contaminants in addition to testing for stability. Such work should be carried out and documented to cGMP standards.

EXPRESSION SYSTEMS Recent advances have facilitated the availability of various methods for cell expression and recombinant protein production. Since many genes can be expressed in multiple systems, it should be determined which system has the greatest applicability for producing a particular recombinant product. Thus, as expression systems, fermentation techniques, and downstream processing technologies develop, the selection process must change to accommodate this progress. Two key factors in particular affect the choice of an expression system: (1) the required quantity of the designated protein, and (2) the protein’s indigenous structural complexity. Additional key factors must also be determined, including the product’s estimated market size, whether or not specific modifications are necessary for retaining biological activity, and the chemical stability of the product. The optimal expression system, then, would be the one that yields the maximum quantity of properly folded bioactive protein. Nevertheless, there are situations in which the system expressing the highest level of a particular protein is not the system that produces the most bioactive or properly folded protein, and therefore, the optimal performance of a particular expression system must sometimes be weighed against maximum production quantity. Since there is no optimal system for expressing and commercially producing all recombinant proteins, each recombinant protein presents its own unique challenge.

VECTOR CONSTRUCTION A detailed discussion of recombinant DNA methods and expression vectors for each heterologous expression system is beyond the scope of this work; the subjects have been reviewed in detail elsewhere (see References). Specialized vectors enabling efficient introduction of heterologous genetic material into a host cell have been developed for each expression system, the heterologous gene being integrated into the host genome, either in single or in multiple copies, or located in an extrachromosomal DNA fragment (plasmid) that independently replicates after incorporation into a given cell to produce many genetic copies. This recombinant plasmid incorporates an expression cassette consisting of a promoter component to regulate transcription, the heterologous recombinant genetic expression material, and a transcriptional terminator. The expression cassette also contains components for optimizing transcription, translation, and secretion in the host cell — all required to express recombinant protein and to transport it to its extracellular habitat. Initiation of heterologous expression gene activity, as well as its endogenous expression level (operational intensity), can be controlled by incorporating an appropriate promoter in the heterologous DNA construct, consisting of a small portion of the DNA sequence that partially determines the endogenous expression level in the host cell (e.g., during the lytic phase of a baculovirus infection of an insect cell, the polyhedron gene product represents around 50% of the cell’s total protein, so an insect cell can

PH2270_book.fm Page 49 Tuesday, May 16, 2006 9:37 AM

Recombinant DNA Materials and Methods

be driven to express large quantities of heterologous recombinant protein following incorporation of the genetic promoter into its expression cassette). Additional factors may also affect expression, such as the stability of its mRNA or the terminator signal employed. Within the protein-coding sequence of many eukaryotic genes, there are noncoding sequences (introns) that are spliced out during genetic expression. While heterologous gene expression in bacteria and yeasts is limited to clones lacking introns, bacterial systems are unable to remove introns from mRNA sequences. And, although certain yeast genes do contain introns, the yeast’s splicing mechanisms are not particularly efficient at removing the higher eukaryotic introns. Intron-containing genes, however, can be expressed in prokaryotic systems by using either cDNAs (DNAs that have been synthesized from a gene’s mRNA) or chemically synthesized genes that lack introns. Little is known about mRNA splicing in insect cells. Universal expression of intron-containing genes in insect cells remains to be demonstrated, although splicing signals of a genomic fragment of SV40 small-t antigen have been correctly and preferentially implemented when expressed in insect cells. Recently both light and heavy chain cDNA of murine monoclonal antibody have been expressed separately, together, and as a dual construct in insect cells infected with recombinant baculoviruses with high levels expressed under the control of the polyhedron promoter. Mammalian cells, of course, can correctly process and express intron-containing genes. A multitude of heterologous proteins have been expressed in insect cells. In many instances, these proteins were shown to be similar to natural moieties, both in antigenicity and function. Insect-cell expression levels have typically ranged from one mg per liter to 75 grams per liter. As cell lines from other tissues and hosts are cultivated, further developments in protein synthesis may be anticipated. Translational efficiency must also be high enough to attain heterologous protein expression of a production level. In translation initiation, implementation of the codon affects translation elongation and termination. Highly expressed genes encoding hom*ologous host proteins demonstrate strong codon partiality correlating with tRNA levels, although absolute correlation between codon usage and heterologous gene expression has not been demonstrated.

PROKARYOTIC AND EUKARYOTIC GROWTH CHARACTERISTICS BACTERIA E. coli, a gram-negative bacterium, has been best characterized, most easily grown, and most frequently used for industrial recombinant protein production. Its superior growth characteristics are well documented. An E. coli colony can double in about twenty to thirty minutes when grown in enriched medium, and one bacterium is capable of generating about 150 grams dry-cell-weight/liter in defined medium.

YEASTS Yeasts such as S. cerevosiae also offer advantages in large-scale culture. Yeasts, unlike bacteria, lack detectable endotoxin levels and are generally recognized as

PH2270_book.fm Page 50 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

safe for producing food and healthcare products. Yeasts provide a well-established fermentation technology and extensively characterized genetics. Yeast culturing is simple, cost-effective, and rapid, with populations doubling in about 90 minutes when grown in an enriched medium containing glucose as a carbon source. S. cerevosiae also grows on minimal medium and utilizes a variety of nonglucose carbon sources. In addition to S. cerevosiae, Pichia pastoris can be used since its gene expression is tightly regulated by methanol, providing a simple and costeffective method for industrial fermentation. Although heterologous proteins may be potentially toxic to a host cell, selection against heterologous gene-containing cells can be reduced by initially growing Pichia pastoris to high densities on carbon sources such as glucose or glycerol which when regulated by a methanol-induced promoter, repress heterologous gene expression. Subsequently, a switch to methanol as a carbon source will initiate expression. Although inducible promoters are also often used in E. coli and S. cerevosiae expression, P. pastoris strains were selected for efficient growth on methanol in defined medium at high cell density. From a bioprocess standpoint, these yeast strains might be ideal hosts for heterologous gene expression, since concentrations of 130 grams dry-cell-weight/liter have been reported in continuous culture fermentation.

EUKARYOTIC SYSTEMS (INSECT

AND

MAMMALIAN CELLS)

Scaling-up recombinant protein synthesis in cultured insect or mammalian cells is usually a complex and costly process: 1. Cultured insect and mammalian cells grow more slowly than microbial cells (when compared to the 20–90 minute doubling time of the commonly used microbial systems, the average doubling time of insect and mammalian cells is 18–24 hours). 2. Choosing an insect or mammalian expression system significantly adds to total product costs in view of the fact that the complex media necessary is about fifteen times more expensive than common bacterial media. 3. Fetal bovine serum (FBS) medium supplementation along with growth promoters is expensive in view of the fact that FBS supplies are limited and their composition generally varies, requiring adjustments to both growth and purification. 4. Mammalian cells are much more sensitive to the shear forces associated with traditional fermenter mixing methods employing marine impellers or agitators; insect cells have a greater oxygen demand than mammalian cells, and they are also sensitive to shear and subject to damage by sparging. 5. Mammalian cells are delicate and tolerate only narrow ranges of temperature, pH, DO2 level, and metabolite waste levels in the medium; largescale mammalian bioreactors are thus quite complicated and are also designed to meet stringent sterility requirements. 6. Scaling-up insect-cell-system growth is still in the early-to-middle stage of development, and after more technical obstacles are overcome, largescale commercial immobilized cell systems will subsequently be devel-

PH2270_book.fm Page 51 Tuesday, May 16, 2006 9:37 AM

Recombinant DNA Materials and Methods

oped to protect delicate cells, reduce medium volume, concentrate end product, and perform cell separation. Thus, airlift, hollow-fiber, and other bioreactors, many not conceived of as yet, as well as low-cost serum-free media, will be developed specifically for insect-cell culture.

EXPRESSION LEVELS Since commercial products generally must be produced in large-scale to be costeffective in the marketplace, the expression level of a particular protein in a given host system is generally a major factor in choosing the heterologous expression system. Historically, bacterial expression systems have been the most extensively used, and such utilization over the years has lead to the accumulation of a large body of literature on the subject. To protect heterologous proteins from cellular proteases, bacterial cells often accumulate them as insoluble complexes, concentrating protein complexes within the cell as inclusion bodies. The high expression levels that have been achieved with E. coli may be due to this protection, with expression levels as high as 30% of total cell protein having been obtained. Although produced in large quantity, the recombinant protein may not be properly folded, thereby reducing the amount of bioactive product that can be recovered from the crude material. Nevertheless, a large number of mammalian gene products have been synthesized in E. coli, and its high expression level is still considered one of its most significant advantages. Despite the fact that secreted proteins have been produced recombinantly in bacteria, protein levels are generally much lower than those produced by intracellular expression and the product usually is trapped in the periplasmic space located between the cell’s plasma membrane and cell wall. Bacterial secretion systems have also been developed using fusion proteins. To illustrate, a synthetic fragment of staphylococcal protein A, normally secreted by gram-positive bacteria, has been linked with heterologous genes to direct about 80% of the fusion product into the culture medium. In defined medium with optimized fermentation, the fusion protein yield, consisting of a staphylococcal protein-A fragment bound to a recombinant cytokine, has been reported at as much as one gram per liter. The fusion protein is then recovered using IgG affinity chromatography. The product is then separated from the protein-A fragment by chemical cleavage. In comparison with bacteria, yeasts generally synthesize heterologous proteins at lower levels. Nevertheless, production of a product with appropriate biological activity is a primary consideration and yeasts have demonstrated their ability to produce biologically active proteins. Yeasts have been utilized by the Europeans to a greater extent than bacteria for recombinant protein production. SOD is produced in S. cerevosiae at levels equal to 25–30% total cellular protein, and, when grown in a high cell-density fermenter, P. pastoris synthesized hTNF at levels reaching 30–35% soluble protein. Numerous heterologous proteins have been expressed in insect cells. As cell lines from other tissues and host species are cultivated, further developments in protein synthesis can be anticipated. Unlike other systems, mammalian cells express relatively low protein levels — on the order of tens of milligrams per liter per day under optimal production conditions. Some of the

PH2270_book.fm Page 52 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

highest levels of heterologous protein expression in mammalian systems have been produced by using amplifiable vectors from CHO cells. These cells typically reach production levels of 36 × 1014 molecules of heterologous protein per 106 cells per day. High expression levels have also been achieved using BPV vectors in mouse C127 cells. Protein production has ranged with this system from 1012–1013 molecules of heterologous protein per 106 cells per day. In each expression system, overproduction of a heterologous protein frequently leads to complications such as protein aggregation, host-mediated protein degradation, host cell toxicity, inaccurate or incomplete synthesis, incorrect modifications, and improper protein folding. To prevent degradation, host cells with reduced proteolytic activity can be used, and/or a system can be selected that allows extracellular secretion of the heterologous protein in order to avoid cellular proteases and eliminate host cell toxicity. Also, by over-expressing endogenous proteins required for correct protein processing, production of correctly processed protein may be enhanced. Although some guidelines have been established for each expression system, each recombinant protein presents its own unique problems and expression levels should be optimized on a case-by-case basis.

INTRA- VS. EXTRACELLULAR EXPRESSION There are two basic modes of expression: (1) intracellular, in which the heterologous protein accumulates within the host cell cytoplasm either as a soluble protein or an insoluble aggregate; and (2) extracellular, in which the heterologous protein genetic sequence has been manipulated so that the protein is secreted into the culture medium. Selecting either mode affects expression levels, recovery, and purification of the protein. The biomolecular nature of the final product is also governed by the expression mode since protein folding, disulfide bond formation, and posttranslational modifications can also be influenced.

INTRACELLULAR EXPRESSION High synthesis levels have been produced with intracellular expression systems, particularly in E. coli. High heterologous protein expression in E. coli cytoplasm often results in compartmentalization of the protein into inclusion bodies. Isolating these inclusion bodies from bacterial cells by centrifugation represents a significant first step in recombinant protein recovery. Inclusion bodies are not generally found in eukaryotic expression systems, although bodies of aggregated protein have been observed in insect cells when large quantities of viral nucleoproteins are expressed. In all expression systems, however, there are disadvantages associated with intracellular expression. Protein synthesis begins with the addition of a methionine at the NH2-terminus of the protein; however, because many target proteins are processed from larger precursors, they have a specific NH2 terminus other than methionine. Since special cellular mechanisms are required to remove the methionine from rudimentary intracellular proteins, there is a possibility that a heterologous protein expressed intracellularly may have an incorrect or heterogeneous

PH2270_book.fm Page 53 Tuesday, May 16, 2006 9:37 AM

Recombinant DNA Materials and Methods

protein NH2 terminus. Although enzymes (methionine aminopeptidases) used to remove NH2-terminus methionines have the same general specificity in both prokaryotic and eukaryotic cells, protein over-expression may result in incomplete processing of the NH2 terminus methionine, yielding a mixture of improperly processed protein. If the native protein has a specific NH2-terminus other than methionine, it is critical to determine whether or not the variant containing the NH2-terminus methionine retains the desired biological properties and does not exhibit increased or altered immunogeneity. Alternative approaches have been employed to avoid the problem of NH2terminal heterogeneity. In one approach the genetic sequence of the native protein is combined with that of a polypeptide. The target recombinant protein can then be separated from the complex by specific chemical cleavage or endopeptidase treatment. Similarly, enzymatic treatment can be employed to process polypeptide hormones, frequently endogenously expressed as long precursor proteins that encounter posttranslational processing. For example, an hGH precursor produced in E. coli at levels corresponding to 20% total cell protein can be converted by an exopeptidase, dipeptidyl aminopeptidase I (DAP I), to authentic hGH. Intracellular expression systems may not yield soluble, properly folded bioactive products that are easily isolated since the reducing conditions of the host’s cytoplasm may be greatly different from those encountered by the protein in its native environment, causing the protein to improperly fold or aggregate. Should this occur, the isolated protein then requires additional processing to solubilize and refold it. Clearly, adding these steps has a tremendous effect on purification strategy, scale up, and end product cost.

EXTRACELLULAR EXPRESSION The alternative mode, extracellular expression, offers multiple advantages. First, an NH2 terminus corresponding to that of the native protein can be obtained because the heterologous protein’s NH2- terminal signal peptide is removed from the protein during the secretion process — although incorrect or incomplete processing of the signal peptide has been known to occur. Second, if the heterologous protein is normally a secreted cellular protein, extracellular expression in the heterologous host can yield a properly folded product with correct disulfide bonds. For example, peptide mapping by tryptic digest demonstrated that a recombinant interferon secreted by a yeast (S. cerevisiae) has the identical disulfide bond structure as that of the natural human protein. And, third, the secretion pathway of the eukaryotic system provides a means to obtain a correctly glycosylated product and this secretion may protect it from degradation by intracellular proteases. Secretion of a recombinant protein into the culture medium may also facilitate production because it is removed from contaminating intracellular proteins, thus simplifying purification. Additionally, certain hosts employing secretion can be immobilized and used in high cell-density production systems, and although optimal host cell growth conditions for protein expression must be maintained, it is conceivable that medium conditions could also be adjusted to optimize retention of the protein’s biological activity. There

PH2270_book.fm Page 54 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

are also disadvantages associated with extracellular expression (e.g., extracellular production levels may be less than those obtained in a comparable intracellular expression system and with the E. coli host, and the heterologous protein is often trapped in the periplasmic space). When this happens, the protein must be released and these methods must also be suitable for scale up.

GLYCOSYLATION The quality and extent of a protein’s glycosylation depends upon the host system; thus, glycosylation playing a significant role in the biological activity of a potential therapeutic restricts the choice of expression system. For example, with the glycoprotein hormone EPO, the principal regulator of RBC formation, oligosaccharides are essential for this in vivo function. The serum half-life of unglycosylated EPO is measured in minutes, although when the protein has been glycosylated in the CHO cell host, its half-life is about 2 hours. In another example, the deglycosylation of a subunit of hCG has various effects: 1. Deglycosylation of a subunit of hCG cannot activate adenylate cyclase, although the product has an increased affinity for its receptor, resulting in a significant loss in bioactivity. 2. Glycosylation, however, may affect protein folding, solubility, and in so doing, affect the structural and biological characteristics of the therapeutic. 3. Glycosylation may influence the product’s antigenicity since the carbohydrate may (a) be a part of an antigenic determinant, (b) mask a potential antigenic site, or (c) affect protein conformation, thus influencing other potential antigenic determinants. 4. Glycosylation can affect protein stability, either by stabilizing the protein’s conformation, or by protecting it from the proteases, thereby influencing protein recovery, purification, and in vivo half-life. 5. Oligosaccharides are linked to their base protein through either N- or O-linkages. In N-linked oligosaccharides, an N-glycosidic bond is formed between the oligosaccharide and the amide nitrogen of an aspargine amino acid. The attachment site of an N-linked oligosaccharides is within an Asn-X-Thr/Ser sequence (where X is any amino acid, Thr is threonine, and Ser is serine); however, not all Asn-X-Thr/Ser sequences are glycosylated. In O-linked oligosaccharides, an O-glycosidic bond is formed between the oligosaccharide and the hydroxyl group of either serine or threonine.

MAMMALIAN CELL GLYCOSYLATION In the biosynthesis of mammalian cell N-linked glycosylation, the structures of typical mammalian N- and O-linked oligosaccharides consist of short sequences of mannose residues that are attached to serine or threonine, and these O-linked oligosaccharides differ from those in mammalian systems.

PH2270_book.fm Page 55 Tuesday, May 16, 2006 9:37 AM

Recombinant DNA Materials and Methods

INSECT CELL GLYCOSYLATION There is insufficient documented information about the structure and synthesis of insect cell glycoproteins or of the glycosylation of recombinant proteins produced in baculovirus-infected insect cells, although it appears insect cells add only highmannose N-linked oligosaccharides. Studies in mosquito cell lines indicate that their N-linked oligosaccharides are deficient in sialic acid, galactose, and fructose. Additionally, enzymes accountable for the progression of high-mannose oligosaccharides into complex mammalian N-glycan structures are only detected at low levels in these species. The oligosaccharides of glycoproteins produced in Spodoptera frugiperda cells have only been examined indirectly, and evidence for their synthesis and addition of complex oligosaccharide moieties has yet to be demonstrated.

GENERAL CONSIDERATIONS Selecting the correct cell line can be very important for those recombinant proteins that require proper glycosylation for therapeutic function. Bacterial expression systems cannot be used since prokaryotic cells cannot glycosylate. If the recombinant therapeutic requires only high-mannose oligosaccharides, then yeast, insect, or mammalian expression systems may do the job. If the therapeutic protein contains hybrid or complex oligosaccharides that are required for its biological activity, then both yeast and insect cells should be eliminated and a mammalian cells expression system should be employed. Glycosylation in mammalian cells is specific to species, tissue, and cell type. Rat and human acid glycoproteins differ in the degree of oligosaccharide branching and glycosylation and, likewise, the amino acid sequences for rat brain and for thymus tissue glycoproteins are identical although their oligosaccharide structures are dissimilar. In addition to the differences in glycosylation associated with each expression system, the biopharmaceutical industry is also faced with the problem of microheterogeneity. At each glycosylation site within a protein, the expression system could generate subtle differences in the oligosaccharide structure or composition. Since the number of possible expressed variants increases significantly with each additional glycosylation site, the degree of heterogeneity can be significant for proteins with a large number of sites. A given glycoprotein consisting of different glycoforms will have a composite activity, reflecting the weighted average of the activity and incidence of each glycoform present. Defining the structure of native proteins is difficult if the protein is microheterogeneic in its endogenous state. It is also difficult to determine the applicability of a recombinant product if the native protein is microheterogeneous.

FURTHER POSTTRANSLATIONAL MODIFICATIONS Further posttranslational modifications such as acetylation, phosphorylation, acylation, and carboxylation can affect the ultimate choice of a host expression system, and each of these modifications should be evaluated to determine if it is required for the biological activity or stability of the product. hSOD typifies the way which a posttranslational modification can determine the choice of expression system.

PH2270_book.fm Page 56 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

hSOD, an N-acetylated enzyme that prevents oxidation damage by scavenging superoxide radicals, has been expressed at high levels in both bacteria and yeasts. E. coli normally does not acetylate the N termini of its proteins, and hSOD is expressed in the unacetylated form with the N terminus methionine removed. In contrast, hSOD expressed in S. cerevisiae is N-acetylated and is identical to that of the human erythrocyte. Thus, for pharmaceutical purposes, it would be preferable to use yeast products that are identical to human SOD, thereby eliminating the possibility of an adverse immunoresponse. The specificity of the yeast and human acetylation enzyme is generally the same, and yeast is sometimes useful as an alternative host for the production of N-acetylated heterologous proteins. Phosphorylation is another example of host-cell-specific posttranslational modifications. E. coli expression of the human c-myc protein, the cellular hom*ologue of the myc oncogene, yields a 60–64-kDa protein. There are no apparent modifications in the E. coli — expressed proteins because they comigrated with the protein derived from an in vitro-coupled transcription-translation system. The myc proteins expressed in S. cerevisiae, however, are modified by phosphorylation. Yeast proteins are not as extensively modified as c-myc phosphoproteins synthesized in the baculovirus-insect cell system, however, and the introduced phosphate-free yeast gene expression limits this modification. Still, the relationship remains cloudy between the different c-myc species produced in human cells and those synthesized in heterologous expression systems. Typically, if large quantities of the product are required, and the target product is unmodified and has few disulfide bonds, prokaryotic bacterial or yeast expression systems should be considered first. These expression systems offer significant advantages over eukaryotic insect or mammalian cell systems, because they require less time to develop transformed cell lines that express the product, levels of crude product are typically higher, the cells have simple culture requirements, and large-scale bioprocess strategies are well established. Although prokaryotic microbial systems can produce large quantities of crude material, they might not be able to produce properly folded end products. If the protein requires a particular modification that can only be carried out in the insect or mammalian cells, or if the prokaryotic protein product is inactive, then these prokaryotic systems cannot be used. If the in-house expertise is available, simultaneous investigation of various expression systems can speed the product’s initial development.

ECONOMIC CONCERNS Cost consideration is crucial when evaluating an expression system. Examples of detailed cost analyses are not readily available for this book because most of the bioprocesses used in the biopharm industry are proprietary. Capital investment, product manufacturing costs, and investment return are typically determined by employing basic engineering principles and company experience. Cost analysis is especially important in situations when an expression system has not been proven at production level or when bioprocess or recovery strategies differ significantly from established procedures. A thorough cost analysis should be completed early in the project, covering all available options for protein expression, synthesis, and large-scale production. Initial values that must be calculated to determine the capital

PH2270_book.fm Page 57 Tuesday, May 16, 2006 9:37 AM

Recombinant DNA Materials and Methods

investment and associated operating costs include market value in dollars per unit or dose, potential market size, the required market share in units necessary to yield the anticipated production volume, and the manufacturing cost as a function of both production efficiency (protein produced per unit volume) and production cost (dollars per unit volume). The production cost can be further broken down into production cost and required capital investment, a function of both plant scale and equipment cost. Choosing the expression system can significantly affect the total manufacturing cost, although downstream processing is the major expense. For example, production cost is significantly affected by operating expenses including medium cost and the chemicals necessary for protein over-expression; the selection of a eukaryotic expression system requires a more complex and expensive medium than a typical prokaryotic system, which amplifies production costs. Often, insect and mammalian cells require additional media supplements such as FBS, although serum-free media have been developed for the production of mammalian and insect-cell-derived proteins. Because labor costs represent a significant proportion of overhead, bioprocess length, downtime between runs, and potential production time lost due to containment vessel contamination are all important in comparing production systems,. Expression systems with higher production efficiencies are preferred since a smaller number of batches are required to produce given volumes. Expression systems also determine the design and complexity of the required fermenter or bioreactor and its associated equipment. Parameters such as cell concentration (expressed as dry-cell-weight/liter) for a prokaryotic (microbial) growth process, total intracellular protein concentration, expression levels of target protein (expressed as percent total cell protein), and time should be considered when estimating the efficiency of intracellular product synthesis. In contrast, extracellular production efficiency is appraised by expression levels of target protein per unit volume, total extracellular protein concentration, and growth time (e.g., production levels in mammalian cell expression systems are stated as the amount of product per cell per day or the amount of product per volume per day). The amount of purified bioactive product that can be produced from the crude material must also be estimated when comparing production costs. For example, a mammalian cell production system that secretes small amounts of a properly folded bioactive product into the culture medium might actually outperform an intricate, expensive fermenter system. If the downstream processes necessary to obtain pure bioactive product are inefficient or costly, an alternate production system should be designed. Two important values that must be obtained when conducting a cost analysis of downstream product recovery and purification are: (1) the level of purity required in the final product and (2) the necessary degree of purity in the starting material. The purification process is affected by a number of variables, including the starting material’s accumulation (i.e., is the protein deposited in intracellular inclusion bodies or is it secreted into the culture medium?), cost, complexity, number of required purification steps, protein refolding efficiency, and the length of time for adequate purification. A simplified strategy for evaluating a potential production system involves an initial step of defining critical product requirements, including the protein’s properties and its potential market size. On the other hand, if the native

PH2270_book.fm Page 58 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

material is in limited supply, information regarding its properties may be unavailable at initiation of the project (i.e., it may not be known whether the protein’s oligosaccharides play a role in in vivo bioactivity or if the unfolded purified protein can be refolded). Likewise, the optimal dose and treatment duration (i.e., the market volume) may not be accurately determined until the clinical trials are completed.

REGULATORY CONCERNS Selecting an expression system for the production of a recombinant protein may depend as much upon the government as it does on technical concerns. The regulatory requirements and guidelines generated by national and international authorities for the control and licensing of biotechnology products include the information necessary for the assessment of product quality and safety. Initially, the regulatory focus was on the production and use of first-generation recombinant products that were identical to native counterparts. Evaluation by the FDA of human insulin, the first biotechnology product to reach the market, established a basis for this review process for all future materials. Moreover, in doing this initial review, the FDA was forced to adapt the regulations governing traditional pharmaceutical and chemical products to genetically engineered proteins. Currently, regulatory agencies face more complex issues such as evaluating recombinant proteins that differ from native products by their amino acid sequence, oligosaccharide composition, or posttranslational modification. Such second-generation products result from research efforts to improve efficacy or to diminish side effects. First-generation recombinant products have also been redesigned to improve stability and solubility, to lengthen in vivo half-life, and to increase specificity. Regulatory agencies are now faced with the task of establishing guidelines that address safety issues pertaining to these modified protein products. Microheterogeneity is another issue. As previously discussed, glycoprotein expression yields proteins with oligosaccharide structures differing significantly from those of native molecules and proteins that comprise many glycoforms, each with fine differences in oligosaccharide structure. More must be known about the influence of oligosaccharide heterogeneity on bioactivity, pharmaco*kinetics, or immunogenicity, and it is these issues that concern regulatory agencies. Nevertheless, defining glycosylation patterns and determining the degree of natural glycoprotein microheterogeneity must be accomplished before the above issues can be addressed. For example, if the regulatory agencies were to require detailed analyses of final product oligosaccharide structure or insist on procedures that minimize microheterogeneity, such as complex purification schemes or manipulations of the expression system, manufacturing costs could become prohibitive. These issues are receiving considerable attention and are even becoming integral to patent applications and proactive patent defenses. Advances that led to increased protein production in cultured mammalian cells also opened up new issues regarding recombinant product safety and regulation. Although transformed cell lines are essential for the high-volume extended production of human protein therapeutics, it took considerable reformulation of earlier regulatory positions before the FDA and the World Health Organization (WHO) accepted the use of genetically altered

PH2270_book.fm Page 59 Tuesday, May 16, 2006 9:37 AM

Recombinant DNA Materials and Methods

cells. The use of transformed cell lines raises major safety concerns as to whether the final product will contain viral DNA or cell-derived DNA fragments with oncogenic potential, or whether there is contamination from virus particles, cellular proteins, or pyrogens. To ensure safety, recombinant therapeutic products derived from transformed cell lines must undergo stringent purification procedures and extensive characterization — particularly important for products used in repeated or concentrated doses. From a regulatory standpoint, insect cell expression offers advantages over mammalian. Since the baculovirus is not pathogenic to vertebrates, it cannot replicate, express, or integrate its genes in vertebrate cells; furthermore, transformed cells or transforming elements are not required in this system. Nevertheless, regulatory agencies always will require data on the safety of any system used in the production of recombinant therapeutic proteins, and these requirements should be a primary consideration in choosing an expression system. In a schematic overview of molecular cloning methods (Figure 4.1) applied to resolving the structures of various proteins, mRNA from tissue fractions is the basic material for a tissue cDNA library that can be selected for clones by the indicated strategic protocols. Alternatively (lower left), DNA can be prepared as a total genomic library or (lower right) selected probes can be tailored to proteins; other immunological methods can be employed in evaluating the expressed product’s location and functional properties.

MANUFACTURERS’ DIRECTORY CLONING VECTORS, BACTERIAL EXPRESSION SYSTEMS Allelix Biopharmaceuticals, Inc. (416) 677-0831 American Type Culture Collection (800) 638-6597 Bio/Can Scientific (416) 828-2455 Bio-Rad Laboratories (510) 741-1000 Boehringer Mannheim Corp. (317) 845-2000 Carolina Biological Supply Co. (800) 334-5551 Clontech Labs, Inc. (800) 662-2566 Fisher Scientific (800) 562-1729 5 Prime 3 Prime, Inc. (800) 533-5703 Gibco BRL Life Technologies (800) 955-6288 Invitrogen (760) 603-7200 New England Biolabs, Inc. (800) 632-5227 Pharmacia Biotech, Inc. (800) 526-3593 Promega Corp. (800) 356-9526 Qiagen, Inc. (800) 426-8157 Sigma Chemical Co. (800) 325-3010 Stratagene (800) 424-5444 United States Biochemical Corp. (800) 321-9322 Wako Chemicals USA (800) 992-WAKO

PH2270_book.fm Page 60 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

CLONING VECTORS, YEAST EXPRESSION SYSTEMS American Type Culture Collection (800) 638-6597 Bio/Can Scientific (416) 828-2455 BIO 101, Inc. (800) 424-6101 Clontech Labs, Inc. (800) 662-2566 Sigma Chemical Co. (800) 325-3010 Stratagene (800) 424-5444 Wako Chemicals USA (800) 992-WAKO

CLONING VECTORS, MAMMALIAN EXPRESSION SYSTEMS Bio/Can Scientific (416) 828-2455 Bio-Rad Laboratories (510) 741-1000 Clontech Labs, Inc. (800) 662-2566 Fisher Scientific (800) 562-1729 5 Prime 3 Prime, Inc. (800) 533-5703 Gibco BRL Life Technologies (800) 955-6288 Invitrogen (760) 603-7200 National Biosciences, Inc. (800) 747-4362 Pharmacia Biotech, Inc. (800) 526-3593 Promega Corp. (800) 356-9526 Stratagene (800) 424-5444 Wako Chemicals USA (800) 992-WAKO

CLONING VECTORS, INSECT EXPRESSION SYSTEMS American Type Culture Collection (800) 638-6597 Pharmacia Biotech, Inc. (800) 526-3593 Promega Corp. (800) 356-9526 Sigma Chemical Co. (800) 325-3010 Stratagene (800) 424-5444 United States Biochemical Corp. (800) 321-9322 Wako Chemicals USA (800) 992-WAKO

PROKARYOTIC CELL LINES American Type Culture Collection (800) 638-6597 Becton Dickinson Microbiology Sys. (800) 638-8663 Biotrol Inc. (612) 448-2515 Carolina Biological Supply Co. (800) 334-5551 Clontech Labs, Inc. (800) 662-2566 Difco Laboratories, Inc. (800) 521-0851 Lallemand, Inc. (514) 522-2133 Medical Chemical Corp. (310) 829-4304

PH2270_book.fm Page 61 Tuesday, May 16, 2006 9:37 AM

Recombinant DNA Materials and Methods

PML Microbiologicals (800) 547-0659 Promega Corp. (800) 356-9526 REMEL (800) 255-6730

EUKARYOTIC CELL LINES American Type Culture Collection (800) 638-6597 Buckshire Corp. (215) 257-0116 Cell Associates, Inc. (707) 785-3181 Clonetics Corp. (800) 852-5663 Harlan Bioproducts for Science, Inc. (317) 894-7536 ICN Biomedicals, Inc. (800) 854-0530 Thomas Scientific (800) 345-2100 Unisyn Technologies, Inc. (619) 673-1131 Whittaker Bioproducts (800) 654-4452

PH2270_book.fm Page 62 Tuesday, May 16, 2006 9:37 AM

PH2270_book.fm Page 63 Tuesday, May 16, 2006 9:37 AM

Enzymes OVERVIEW

Man has used the products of enzyme technology for domestic and commercial purposes for thousands of years. Beer, wine, and bread are probably the earliest products of fermentation. It is not surprising, then, that yeast enzymes, upon which such fermentations depend, were the first to be chemically identified in the early nineteenth century. In cheese making, a technology of almost equal antiquity, rennet was the first relatively pure enzyme used industrially. The appalling stench of hides being soaked in urine to soften them and remove fats has not been characteristic of the tanning industry for nearly a century; trypsin and lipase are now used. Well over 2,000 enzymes have been identified; however, only about 10% have been isolated in pure crystalline form. Consequently, most enzymes used in biotechnology are sold on the basis of units of activity rather than by weight, and many product descriptions include a listing of other enzymes known to be present. Enzymes are sometimes named by their substrate plus -ase, followed by source (e.g., lipase, human pancreas). Some, such as pepsin or trypsin, are not. To avoid confusion, a classification system was adopted by international agreement that organized enzymes into six major classes (see Table 5.1) according to the type of reaction catalyzed. This classification is used in some catalogs (although many do not) in addition to the enzymes’ common names and sources. Well over 200 enzymes are used in industry, with fewer than 20 having attained major commercial importance at this time. Most of these are enzymes used in the food processing, pharmaceutical, and detergent industries. Enzymes of vital importance to biotechnology are those used in DNA and RNA sequencing and synthesizing, and in polysaccharide and polypeptide sequencing and synthesizing. Also important are enzymes such as trypsin, dispase, collagenase, elastase, and hyaluronidase (collectively called tissue dissociating and cell harvesting enzymes) that are generally used in vitro to remove cells from their extracellular matrix. Some examples are listed in Table 5.2. Enzymes are highly specialized proteins with extraordinary catalytic power and activity, and like other catalytic substances, they facilitate or accelerate a chemical reaction but are not themselves changed or chemically involved in that reaction. Also known as biological catalysts, enzymes are highly substrate-specific, accelerating chemical reactions without the formation of by-products and functioning in dilute aqueous solutions under very mild but very precise conditions of temperature and pH. As functional units of cellular metabolism in precisely organized sequences, enzymes act by degrading nutrient molecules, conserving and transforming energy, synthesizing macromolecules from simpler precursors, and regulating and sustaining

PH2270_book.fm Page 64 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

TABLE 5.1 Enzyme Categories Class

Type of Reaction Catalyzed

Oxidoreductases Transferases Hydrolases Lyases Isomerases Ligases

Transfer of electrons Group-transfer reactions Hydrolysis reactions Addition or removal of groups from double bonds Transfer of groups within molecules to yield isomeric forms Formation of CC, CS, CO, and CN bonds by condensation reactions coupled to ATP cleavages

TABLE 5.2A Enzyme Product Summary: Restriction Enzymes Restriction Enzymes Aat II Acc I Acy I (Aha II) Afl III Alu I Apa I Apy I Asn I Asp I (Tth III) Asp HI (Hgi Al) Asp 700 (Xmn I) Asp 718 (Kpn I) Ava I Ava II Avi II (Aos I) Bam HI Ban I Ban II Bbr PI (Pma CI) Bcl I Bfr I (Afl II) Bgl I Bgl II Bmy I Bss HII

Bst EII Bst XI Cel II (Esp I) Cfo I (Hha I) Cfr 10 I Cla I Dde I Dpn I Dra I Dra II Dra III Dsa I Dsa V (Scr FI) Eae I Ecl XI (Xma III) Eco RI Eco RV Eco 47 III Fok I Hae II Hae III Hind II Hind III Hinf I Hpa I

Hpa II Kpn I Ksp I (Sac II) Ksp 632 I Mae I Mae II Mae III Mam I Mlu I Mro I (Acc III) Msp I Mva I Mvn I (Fnu DII) Nae I Nar I Nci I Nco I Nde I Nde II (Mbo I) Nhe I Not I Nru I Nsi I Nsp I Pst I

Pvu I Pvu II Rsa I Sac I (Sst I) Sal I Sau I (Mst II) Sau 3A Sau 96 I Sca I Scr FI(Dsa V) Sfi I Sfu I (Asu II) Sgr AI Sma I Sna BI Sno I Spe I Sph I Ssp I Stu I Sty I Taq I Xba I Xho I Xho II

PH2270_book.fm Page 65 Tuesday, May 16, 2006 9:37 AM

Enzymes

TABLE 5.2B Enzyme Product Summary: Modifying Enzymes Modifying Enzymes DNA Phosphorylation/Dephosphorylation Alkaline phosphatase Polynucleotide kinase Polynucleotide kinase, 3′-phosphatase-free DNA Ligases E. coli DNA ligase T4 DNA ligase DNA and RNA Polymerases Taq DNA polymerase T4 DNA polymerase E. coli DNA polymerase I, nick translation grade E. coli DNA polymerase I, endonuclease-free Klenow enzyme (DNA polymerase I, large T4 Gene 32 protein fragment) Terminal transferase Reverse transcriptase SP 6 RNA, polymerase E. coli RNA polymerase T7 RNA polymerase T3 RNA polymerase RNase inhibitor Polynucleotide phosphorylase Phosphodiesterases Phosphodiesterase from calf spleen Phosphodiesterase from Crotalus durissus

DNase I, RNase free Exonuclease III Nuclease Bal 31 Nuclease P1 Nuclease S1

RNase, DNase-free RNase H

Pronase Restriction protease

Nucleases DNase I Neurospora crassa Endonuclease Nuclease, mung bean Nuclease S7 (Microccocal nuclease) Uracil-DNA glycosylase Ribonucleases RNase A RNase T Proteases Proteinase K Sequencing grade proteases

RNase CL3 RNase U2

PH2270_book.fm Page 66 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

TABLE 5.2C Enzyme Product Summary: Tissue Dissociation and Cell Isolation Enzymes Tissue Dissociation and Cell Isolation Enzymes Chymotrypsin Papain Subtilisin Chondroitinase DNase I

Clostripain Pronase Protease Hyaluronidase

Elastase Proteinase K Trypsin Neuraminidase

metabolic activity. Restriction enzymes, found in certain bacterial cells are used to protect those cells from attack by other bacteria by slicing up foreign cellular DNA at sequences that are either absent in the original cell or that are protected by modifying enzymes. Modifying enzymes modify the structure of DNA to permit methylation, dephosphorylation, and so on for genetic mapping, marking, and sequencing. Proteolytic enzymes break down proteins into component peptides and are used for protein characterization studies and as tissue dissociation and cell harvesting tools. All enzymes are proteins, with molecular weights ranging from 12,000 to over 1 million. Some enzymes consist only of amino acid residues, while others require additional components (cofactors) to function, which may either be loosely and transiently bound or tightly and permanently bound. Cofactors may be simple inorganic ions (e.g., Fe+2, Mn+2, Zn+2) or complex organic molecules called coenzymes (e.g., thiamine pyrophosphate), and some enzymes require both types of cofactor. Enzyme catalytic activity depends on the maintenance of structural integrity (i.e., the enzyme must not be denatured by chemical activity, heat, cold, pH, etc.), and in order to compete with chemical catalysts in industrial processes, some commercial usages may require recovery. Although functioning at pH and temperature conditions that are an advantage over some chemical catalysts, enzymes are water soluble and can be lost or degraded during processing. Various methods are employed to maintain the catalytic activity of enzymes at commercially feasible levels, including: 1. Chemical methods: direct covalent attachment of the enzyme to a matrix (usually a polymer) 2. Physical methods: electrostatic attraction to secure the enzyme to an inert support (adsorption) 3. Cavitation: (introducing the enzyme into the cavities of porous material) which chemically prevents removal of the enzyme but allows penetration of the substrate 4. Microencapsulation: of the enzyme within membranes that are physically impermeable to the enzyme and macromolecules but permeable to low molecular weight chemical products (such enzymes are called fixed

PH2270_book.fm Page 67 Tuesday, May 16, 2006 9:37 AM

Enzymes

enzymes and their immobilization makes it possible for some processes to be carried out at higher temperatures and thereby faster). For example, this method is used for protecting urease in artificial kidneys. Protein cleavage or carbohydrate linkages at specific locations are achieved by an enzyme that works only at that specific location. Peptide and nucleic acid syntheses require enzymes to establish linkages between specific amino acids. These cleavages and linkages have been applied to both DNA and RNA recombination, replication, and repair processes as well as to the incorporation of markers in DNA and RNA molecules.

ENZYME PROTEIN ENGINEERING The biotechnology industry uses enzymes in various ways: to break down components, to improve food flavor and texture, and to synthesize additives. Enzyme characteristics impose restrictions on the processes in which they can be used effectively (e.g., enzyme activity may be limited at the required process temperature, pH, or microenvironment). Many research programs are targeted at modifying enzyme activity and increasing their efficiency to improve processing, safety, and quality. This work is essentially multidisciplinary, involving changes to protein structure that alter physical or catalytic properties, making the enzyme more suited to a particular purpose. Modifying enzymes in a rational, directed manner requires detailed structural information, such as that derived from X-ray analysis of single crystals coupled with computer predictions of the probable effects of single amino acid changes on enzyme function. This is a starting point for an enzyme protein engineering cycle that allows for testing of the previous design stage and incorporation of any new information into the next stage. This cycle is common to the successful design and synthesis of all biocatalysts: 1. 2. 3. 4. 5.

Protein crystallography of the pure enzyme to determine its 3-D structure. Computer-aided design of the enzyme protein expression model. Production of a novel gene by DNA synthesis or site-directed mutagenesis. Assessment of enzyme activity after expression in E. coli or another cell line. Preparative biochemistry and refolding of the modified enzyme protein.

Use of these techniques enables the generation of enzymes with improved stability, enhanced activity, and modified specificity for effective industrial catalysis. Work is concentrating on papain, an example of a protease in current use, and phospholipase A2, a lipase of potential interest. Genetic models have been designed to direct the changes needed to modify the pH dependence and stability of papain, such as a papain molecule that retains its normal wide specificity and activity below pH 4. Protein hydrolysis at low pH is attractive commercially since potential problems associated with microbial spoilage or the growth of pathogenic microorganisms are greatly reduced. Fungal acid proteinases are not seen as alternatives, since they cause the accumulation of peptides and are too specific in action. In parallel with the work

PH2270_book.fm Page 68 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

on papain, modifications to phospholipase A2 are being pursued to increase heat stability and alter specificity. Phospholipase A2 represents a class of enzymes that has lipids and phospholipids as substrates and is the only enzyme in this class for which the structure is known. It has a very particular specificity for the C2 position of phospholipids. The native enzyme requires phosphate for this reaction to take place. There is also a need for calcium binding. If the enzyme could be modified to show activity without the need for Ca++ binding and the presence of a phosphate group, the route would be open for upgrading cheap oils and fats by interesterification of the C2 position of the triglyceride. Currently, only lipases with specificities at the one and three positions of the triglycerides are readily available. Over the last 10 years, molecular dynamic procedures have been developed for biomolecules in an aqueous environment, including the computation of free energy by thermodynamic cycle integration. Using these methods, structural change effects on function can be defined in precise quantitative terms. Molecular dynamics techniques, combined with experience in predicting structural activity, and expertise in site-directed mutagenesis and cellular expression, are very powerful conduits for providing a deeper insight into the mechanisms and action of phospholipase. This understanding provides a basis for applying these methods to industrial situations. A working understanding of the factors that govern protein folding and related errors helps predict changes needed to produce correct folding of a modified enzyme protein where there is increased risk of folding error and loss of biological activity. Minor changes might be sufficient to rectify the problem if its cause was identified and understood. In practical terms, it would be advantageous, for example, to convey the high activity of phospholipase A2 from an unsuitable source, such as snake venom, to the porcine enzyme.

TRANSFORMATION IN NONAQUEOUS SYSTEMS The ability to use enzymes in nonaqueous systems offers numerous opportunities for syntheses that are impractical in aqueous solution because of equilibrium constraints or poor solubility of substrate and/or product. The range of possible biotransformations, however, is severely limited by the restricted number of solvent/enzyme combinations that actually function and the slow mass transfer rate of substrate and product between organic and aqueous phases. Many feasible catalyses have involved enzymes in nonpolar organic solvents in which the enzymes are exceedingly stable. In water-miscible polar organic solvents, the extent and nature of the interactions between protein and solvent (residual water or organic solvent) determine stability. Poor understanding of these factors and the way they influence enzyme conformation and stability represents an obstacle to rapid progress in manipulating enzymes for the synthesis and modification of chemical substances. The practical key to success is the development of computer-based molecular designs for enhancing the stability of organic solvents. Computer programs are being developed that will take a protein structure, apply all available information on its structure and stability, and generate a list of modifications that will enhance its stability. In practice, this involves an interplay of design, testing, and logical program development that can then be used to guide the stability modification of any enzyme.

PH2270_book.fm Page 69 Tuesday, May 16, 2006 9:37 AM

Enzymes

Investigators are currently concentrating on two enzymes for which there are adequate three-dimensional structural data and for which DNA has been cloned. The choice is also based on the enzymes’ established and potential use in multiphase biotransformations (peptide synthesis, ester synthesis, phospholipid acyl exchange) relevant to pharmaceuticals and fine chemicals. Phospholipase and cysteine proteinase also present a contrast between single- and double-domain enzymes whose response to solvents is likely to be different at the molecular level; information from both should yield a high-quality data base. Molecular design considerations are focused on the protein-solvent interface and how modification of both the solventaccessible groups and the solvent affects enzyme structure and stability. A polar organic solvent can separate water molecules from the surface of an enzyme, leading to a loss of its conformational integrity. To increase stability, their molecular binding is tightened; or else the water molecules are replaced by redesigning the protein surface that involves developing methods for increasing the binding affinity of water molecules and for replacing water molecules by proteins. In a polar solvent, stability of some proteins can be compromised because the hydrophobic interaction that drives protein folding in an aqueous environment is absent. In such systems, surface ion pairs take on a major role in stabilization and methods are necessary to optimize the selection of residues to be modified. To test the effectiveness of computer-based design, the structure and stability of modified proteins have been analyzed by determining structural changes using X-ray analysis, and by correlating these structural changes with measured changes in stability. New insights into protein stability can then be written into programs for the next cycle of design. The design base will also be used to extend catalytic efficiency by stabilizing transition states using the longest distance electrostatic interactions made possible by organic solvents. An opportunity exists, then, for the development of new enzyme products and processes, where industrial collaboration is a major contributory factor.

RESTRICTION ENZYMES Restriction endonucleases are endodeoxyribonucleases that digest double-stranded DNA, after recognizing specific nucleotide sequences, by cleaving two phosphodiester bonds — one bond within each strand of the duplex DNA. Restriction enzymes form part of the restriction-modification system of bacterial cells to provide protection against invasion of the cell by foreign DNA. Protection against self-digestion is achieved by the presence of specific DNA methyl transferases that transfer methyl groups to adenine or cytosine residues to produce N-6-methyladenine or 5-methylcytosine. Unmodified foreign DNA entering the cell is degraded by the host’s restriction enzyme system.

RESTRICTION-ENZYME CLASSES All restriction endonucleases and their corresponding DNA modification methyl transferases have been classified into classes I, II, or III, according to genetic and protein structure, cofactor dependence, binding specificity, and cleavage.

PH2270_book.fm Page 70 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

CLASS I RESTRICTION ENZYMES Class I restriction enzymes exhibit both restriction and DNA modification activities located on different subunits of multifunctional enzyme complexes, requiring Mg+2 ions, ATP, and sadenosyl methionine (SAM) as cofactors. These enzymes cleave DNA at unspecific sites, usually 100–1,000 base pairs downstream of their recognition sequence.

CLASS II RESTRICTION ENZYMES Class II restriction enzymes and their corresponding modifications, the methyl transferases, act as separate proteins. This enzyme class is site-specific and hydrolyzes specific phosphodiester bonds within or in close proximity to their recognition sequence, and only require Mg++ ions as a cofactor.

CLASS III RESTRICTION ENZYMES Class III restriction enzymes, like class I enzymes, combine restriction and modification activities in a single enzyme complex composed of different subunits. They recognize specific sequences, but cleave twenty-five to twenty-seven base pairs in a direction outside the recognition sequence. They require Mg++ ions for activity but lack both the ATPase activity of class I enzymes and the absolute requirement for SAM. Of these three categories, class II restriction endonucleases are by far the most useful in biomolecular and recombinant DNA technology because of their absolute sequence specificity in binding and cleavage reactions. All commercially available restriction enzymes fall into class II.

RESTRICTION-ENZYME SPECIFICITY Class II restriction endonuclease specificity is defined by three parameters: 1. The recognition sequence 2. The cleavage position 3. The methylation sensitivity The majority of enzymes have a palindromic recognition sequence characterized by the common structural property of a twofold rotational symmetry axis. Most palindromic recognition sequences are tetra-, penta- or hexanucleotides (i.e., Hae III, Hinf I, and Eco RI), but there are also a number of enzymes that recognize sequences longer than six bases (i.e., Not I), interrupted palindromes (i.e., Mam I), or nonpalindromic sequences (i.e., Bsm I). A further subclass of enzymes is the class II S enzyme that cleaves at precise distances from its recognition sequence (i.e., Ksp 6321).

STAR ACTIVITY Under less than optimal conditions, a phenomenon occurs that is exhibited by a few restriction endonucleases; this star activity is caused by a relaxation of sequence specificity.

PH2270_book.fm Page 71 Tuesday, May 16, 2006 9:37 AM

Enzymes

ENZYME FAMILIES AND COMPATIBLE ENDS Restriction endonucleases cleave to generate either a 5′-cohesive, 3′-cohesive, or blunt end. The lengths of the cohesive (also called sticky or overhanging) ends depends upon the position of the cleavage sites in the recognition sequence. With one exception (Nci I, which produces 3′-phosphorylated fragments), all restriction endonucleases generate fragments with 5′-phosphate and 3′-hydroxyl groups. Restriction enzymes producing the same single-strand fragment ends form enzyme families. Although individual enzymes have different recognition sequences, their overhanging ends are complementary — their cleavage fragments can be ligated with fragments produced by any other member of the same family (e.g., the GATC family) of restriction enzymes.

ISOSCHIZOMERS Restriction enzymes that are isolated from different organisms but recognize identical sequences are known as isoschizomers. Isoschizomeric enzymes possess different cleavage sites to one another in a particular recognition sequence (partial isoschizomers, i.e., Kpn 1, Asp 718) or may show different sensitivities to methylation (i.e., Dpn I, Nde II, and Sau 3A). About a thousand restriction endonucleases and around a hundred DNA modification methyltransferases of known sequence specificity from about 900 different microorganisms have been described. Of these, about 130 are commercially available.

REVERSE TRANSCRIPTASES Reverse transcriptases are used for in vitro synthesis of cDNA transcripts of specific RNA sequences for the preparation of cDNA libraries. The cDNA transcripts are used for the analysis of prokaryotic and eukaryotic gene structure, organization, and expression. Comparison of cDNA and genomic DNA sequences accounts for the existence of intervening sequences, splicing events, and genomic rearrangements in eukaryotic genes. Reverse transcriptases are also used in dideoxy sequencing reactions in place of Klenow enzyme, since reverse transcriptases often synthesize through GC-rich regions where Klenow enzyme is hindered. Thus, reverse transcriptase and Klenow enzyme ideally complement each other. In addition, reverse transcriptases can be used for RNA sequencing, dideoxy DNA sequencing, 3′-end labeling of DNA fragments, and the generation of ss probes for genomic footprints.

DNA AND RNA NUCLEASES NUCLEASES DNase I DNase I is used in isolating proteins (e.g., membrane proteins).

PH2270_book.fm Page 72 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

RNase-free DNase I RNase-free DNase I is used in all assays that are sensitive to the presence of RNase, such as in isolating DNA-free RNA produced by in vitro synthesis using the SP6 or T7 RNA polymerases. It is also used for radioactive labeling by nick translation, plasmid construction, sequencing, and for mapping DNase-sensitive regions in eukaryotic DNA. Exonuclease III Exonuclease III is principally used for the generation of single-stranded regions of dsDNA by acting as a 3′–5′-exonuclease. The modified DNA is used as a substrate for labeling DNA with Klenow enzyme, or as a substrate in sequencing studies. Exonuclease III is also used for the construction of deletion mutants by removing nucleotides from 3′-termini of blunt or 5′-overlapping ends. Endonuclease from Neurospora crassa Endonuclease from Neurospora crassa (viper) is only available from Boehringer Mannheim Biochemicals and is used in hybridization to hydrolyze single-stranded DNA. The enzyme is also used for sequencing RNA. It also cleaves uridine-N, adenosine-N, and guanosine-N bonds, but not cytidine-N bonds in the presence of 7 M urea. Nuclease Bal 31 Bal 31 nuclease is used for mapping DNA with restriction enzymes; the activity of Bal 31 can be stopped by specifically complexing Ca++ with EDTA without affecting subsequent cleavage with restriction endonucleases requiring Mg++ for activity. Mung Bean Nuclease With a restrictive, single-stranded specificity for DNA and RNA, mung bean nuclease under mild denaturing conditions extends its nuclease specificity to the A/T-rich regions in double-stranded DNA. This property makes mung bean nuclease ideal for many applications, including transcript mapping, precise excision of gene coding sequences from genomic DNA, removal of single-stranded sticky ends produced by restriction enzymes, removing single-stranded regions of DNA hybrids, analysis of DNA conformation, and hairpin loop cleavage in classical cDNA synthesis. Nuclease P1 Nuclease P1 is used for the large-scale production of 5′-mononucleotides. Nuclease S7 Nuclease S7 (micrococcal nuclease) preferentially cleaves single-stranded substrates, but will also cleave double-stranded DNA or RNA. Nuclease S7 cleaves the

PH2270_book.fm Page 73 Tuesday, May 16, 2006 9:37 AM

Enzymes

5′-phosphodiester bonds of DNA and RNA to yield 3′-mononucleotides and oligonucleotides. Its activity is totally dependent upon Ca++, so it is completely inactivated by adding EDTA. If present in the incubation reaction, Mg++ is not chelated and the Mg++-concentration remains unaffected. The enzyme manifests A/U specificity in RNA digestion at pH 7.5. Incubation at 50ºC at pH 7.5, in the presence of 10 mM CaCl2 and 8 M urea, results in cleavage of the Np/(A/U) bonds upstream of A/U. In sequencing experiments with nuclease S7, various RNA fragments are shifted downstream by one nucleotide in contrast to RNase T1 or RNase U2 fragments of identical length. Its single-stranded specificity has also proven useful for removing endogenous RNA from in vitro translation systems. For this application, the nuclease is dissolved in a buffer containing 50 mM glycine, 5 mM Ca++-acetate, at pH 9.2 before use. The mRNA is digested without affecting its ribosomal and transfer RNA integrity, and the absence of residual nuclease activity in the cell lysate ensures synthesis of full-size translation products. Nuclease S1 Nuclease S1 is used to nick or linearize supercoiled plasmid DNA, cleave mismatches in double-stranded DNA, and to hydrolyze short single-strand protruding ends. Nuclease S1 is used for S1-mapping, a technique used to generate transcript maps of cloned genomic fragments or for the construction of deletion mutants. Uracil-DNA Glycosylase Uracil-DNA glycosylase is used to cleave DNA at any position where a deoxyuridylate residue has been incorporated. The generated AP-DNA is hydrolyzed by either AP-endonuclease or an alkali treatment. U-DNA can be prepared by in vitro or in vivo methods and, depending upon how it is prepared, the uracil-DNA glycosylase can be used for general, site-specific, or strand-specific U-DNA cleavage. The method has been used to increase efficiency of site-directed mutagenesis procedures, and to produce highly labeled oligonucleotide probes.

RIBONUCLEASES DNase-free RNase DNase-free RNase can be used in any isolation technique to remove RNA that might interfere with subsequent procedures (e.g., in plasmid preparations and genomic DNA isolation). RNase A RNase A is an endonuclease that cleaves RNA but not DNA, and shows optimal activity at pH 7.0–7.5. The enzyme specifically attacks pyrimidine nucleotides by cleaving the phosphodiester bond 3′-adjacent to Py/pN. Cyclic 2′:3′-pyrimidine nucleotides are obtained as intermediates, and pyrimidine-3′-phosphates and oligonucleotides with pyrimidine-3′-phosphate terminal groups are the final products. For isolating DNA, the RNase A should be boiled before use.

PH2270_book.fm Page 74 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

RNase CL3 RNase CL3 cleaves RNA predominantly at Cp/N bonds and produces fragments with 3′-terminal cytidine phosphate. Ap/N and Gp/N bonds are cleaved less frequently and Up/N bonds are cleaved to a minimal extent. RNase CL3 is therefore used as a minus U-specific RNase in RNA sequencing. The Cp/N cleavage is more efficient in the presence of urea and at a temperature of 50°C (i.e., under conditions in which the secondary structure of the RNA is destroyed). In this case an approximately five- to tenfold larger amount of enzyme is necessary. RNase H RNase H is used for investigating in vivo RNA-primed initiation of DNA synthesis, and in the synthesis of cloned DNA. It is also used for the site-specific cleavage of RNA and to detect RNA-DNA regions in the naturally occurring dsDNA. Another application is removing mRNA poly (A) sequences to improve mRNA hom*ogeneity in gel electrophoresis. RNase T RNase T cleaves RNA, but not DNA, showing optimal activity at pH 7.4. The endonuclease specifically attacks the phosphodiester bond 3′-adjacent to GpN. Cyclic 2′–3′-guanosine nucleotides are obtained as intermediates. Guanosine-3′phosphates and oligonucleotides with guanosine-3′-phosphate terminal groups are the final products. RNase T is used for RNA sequencing and RNA fingerprinting. RNase U2 RNase U2 cleaves RNA but not DNA, and shows optimal activity at pH 4.5. The endonuclease specifically attacks purine nucleotides by cleaving the phosphodiester bond 3′-adjacent to Pu/pN. Cyclic 2′–3′-purine nucleotides are obtained as intermediates. The final products are purine-3′-phosphates and oligonucleotides with purine 3′-phosphate terminal groups. Reversal of the final step can be used for the synthesis of ApN or GpN.

PROTEASES Pronase Depending on the composition of the preparation, Pronase shows proteolytic activity on both denatured and native protein, typically proceeding to the level of single amino acids. Not a uniform enzyme, pronase is composed of a spectrum of endopeptidases (e.g., serine and metallo proteases, and carboxy and amino peptidases). Proteinase K Proteinase K is extremely effective on native proteins and rapidly inactivates endogenous nucleases such as RNase and DNase, thus, making proteinase K especially

PH2270_book.fm Page 75 Tuesday, May 16, 2006 9:37 AM

Enzymes

TABLE 5.3 Tissue Dissociation and Cell Isolation Enzymes Name

Source

Substrate

Chymotrypsin Clostripain Elastase Papain Pronase Proteinase K

Pancreas C. histolyticum Pancreas Papaya S. griseus T. album

Proteins Proteins Elastin, Proteins Proteins Proteins Keratin, Proteins

Subtilisin V8 Protease Trypsin Chondroitinase

B. subtilis S. aureus Pancreas P. vulgaris

Proteins Proteins Proteins Chondroitin sulfate

Hyaluronidase

Testes

Neuraminidase

C. perfringens

Hyaluronic Acid, Chondroitin sulfate Oligosaccharides, Glycoproteins, Glycolipids

DNase I

Pancreasv DNA

Specificity Carboxyl side of tyr, trp, phe, leu Carboxyl side of arg Carboxyl side of gly, ala, val, leu, ile Carboxyl side of arg, lys Mixture of activities; broad cleavage Broad cleavage on carboxyl side of aromatic or hydrophic residues Broad cleavage Carboxyl side of glu, asp Carboxyl side of arg, lys 1,4 fl-D hexosaminyl; 1,3 fl-D Dermatan sulfate glucuronosyl; and 1,3 a-L iduronosyl linkages 1,4 linkages between N-acetyl glucosamine and glucuronate residues 2,3-; 2,6-; and 2,8-glycosidic linkages

Yields 5′-phosphodi- & oligonucleotides

useful for isolating native RNA and DNA from tissues or cell lines. In addition, Proteinase K is used for the analyzing membrane structure by cell-surface protein and glycoprotein modification. It also promotes cell lysis by activating bacterial autolytic factors. Recognition Sequence of Factor Xa The tetrapeptide recognition sequence of protease factor Xa rarely occurs in natural protein moieties. This makes this endoprotease very useful for the cleavage of recombinant fusion proteins generated from vectors designed to express such products. The commercial product does not typically contain contaminating proteases and is not stabilized with bovine serum albumin (BSA) to ensure the highly specific cleavage of fusion proteins. Sequencing-Grade Proteases If endoproteases are to be successful sequence proteins, compare DNA-fingerprint analyses, detect amino acid exchange in site-directed mutagenesis, or attain partial proteolysis for the study of protein domain structures, the principal requirements are purity, specificity, and freedom from impurities that can interfere in peptide separation by reverse-phase HPLC. Each batch of sequencing-grade protease should be checked by SDS gel electrophoresis with silver stain or by HPLC to guarantee

PH2270_book.fm Page 76 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

consistent purity, and the specificity of each batch should be determined with glucagon, insulin B, or melittin. Modifying Enzymes DNA- and RNA-modifying enzymes are the foundation of recombinant DNA technology, providing the means to manipulate and alter DNA and RNA structure in a defined way. A wide range of modifying enzymes are available for performing almost all molecular biology applications. Ligases, DNA and RNA polymerases, DNA phosphorylating and dephosphorylating enzymes, nucleases and phosphodiesterases are enzymes used to perform these processes — including cloning, sequencing, transcription, dephosphorylation, labeling, and mutagenesis. All these applications require modifying enzymes that are available in a highly pure and stable form to ensure the required activity is present and has specific function.

DNA PHOSPHORYLATION

AND

DEPHOSPHORYLATION

Alkaline Phosphatase Alkaline phosphatase is used for the dephosphorylation of 5′-phosphorylated ends of DNA or RNA. Dephosphorylated vector DNA is thereby prevented from self-annealing prior to inserting DNA fragments. The generated 5′-hydroxylated ends can be labeled effectively with [g32P]-ATP and T4 polynucleotide kinase. The labeled nucleotides can be used for DNA and RNA sequencing by degradation of the end-labeled RNA with base-specific RNases, as well as for DNA-mapping and -fingerprinting. T4 Polynucleotide Kinase T4 polynucleotide kinase is used to label 5′-termini of DNA and RNA with [g32P]ATP by direct phosphorylation of 5′-hydroxyl groups or by the exchange reaction. This 5′-terminal labeling is used in mapping restriction sites by partial digestion, DNA or RNA fingerprinting, DNA footprinting by DNase or methylation protection, hybridization studies, DNA- or RNA-ligase substrate synthesis, and DNA sequence analysis. T-4 polynucleotide kinase can also be used as a specific 3′-phosphatase.

DNA LIGASES E. coli DNA Ligase E. coli DNA ligase is a single polypeptide bacterial enzyme that catalyzes the formation of phosphor diester bonds between neighboring 3′-hydroxyl and 5′-phosphate ends in double-stranded DNA, but requires NAD as a cofactor. It also closes single-stranded nicks in double-stranded DNA. The E. coli enzyme cannot act on nucleotides containing RNA primer sequences and needs longer overlapping sticky ends for ligation than the T4 enzyme. E. coli ligase is used for full-length cDNA synthesis because its more restrictive properties are advantageous for synthesis.

PH2270_book.fm Page 77 Tuesday, May 16, 2006 9:37 AM

Enzymes

T4 DNA Ligase T4 DNA ligase catalyzes the formation of phosphor-diester bonds between neighboring 3′-hydroxyl and 5′-phosphate ends in double-stranded DNA, and closes single-stranded nicks in double-stranded DNA.

DNA

AND

RNA POLYMERASES

Taq DNA Polymerase With its excellent processing ability, absence of exonuclease activity, and optimal function at high temperatures, Taq DNA polymerase is commonly used in DNA sequencing, especially where resolving secondary structures plays a major role. Taq DNA polymerase also accepts modified deoxyribonucleoside triphosphates as substrates, and can be used to label DNA fragments either with radionucleotides, digoxigenin, or biotin. E. coli DNA Polymerase I E. coli DNA polymerase I synthesizes DNA complementary to the intact strand in a 5′→3′-direction using the 3′-OH termini of the nick as a primer. The 5′→3′exonucleolytic activity of DNA polymerase I simultaneously removes nucleotides in the direction of synthesis. The polymerase activity sequentially replaces nucleotides removed by exonucleolytic activity with radioactively or chemically labeled deoxyribonucleoside triphosphates. At low temperatures (i.e., 15ºC) the unlabeled DNA in the reaction is replaced by newly synthesized radioactively or chemically labeled DNA. The method is frequently used for the preparation of sequence-specific probes for screening libraries, as well as for genomic DNA and RNA blots. A special formulation containing a standardized amount of DNase I is used for nick translation. The nick translation method is based on the ability of DNase I to introduce randomly distributed nicks into DNA at low enzyme concentrations in the presence of M92+. Endonuclease-free E. coli DNA polymerase I is also used in second strand synthesis of cDNA with a special formulation of DNase I for nick translation. Labeling-Grade Klenow Enzyme Labeling-grade Klenow enzyme is used for random primer labeling of DNA fragments, using random oligonucleotides as primers together with either radioactively labeled or nonradioactively-labeled nucleotides (e.g., DIG-11-dUTP, Biotin-16dUTP). The enzyme is also used for fill-in reactions and 3′-end labeling of 3′recessed DNA fragments, using labeled deoxynucleotides. Sequencing-Grade Klenow Enzyme The sequencing-grade Klenow enzyme is used for the second-strand synthesis of cDNA, DNA sequencing according to Sanger’s dideoxy chain termination method, and for elongation of oligonucleotides in site-directed mutagenesis.

PH2270_book.fm Page 78 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

T4 DNA Polymerase T4 DNA polymerase is used to label 3′-termini of DNA. Extensive labeling is achieved by the replacement reaction, in which the 3′-exo-nuclease activity of the enzyme first digests dsDNA to produce molecules with recessed 3′-termini. On subsequent addition of labeled dNTPs, the polymerase activity of T4 DNA polymerase extends the 3′-ends along the length of the template. Exonuclease III from E. coli Exonuclease III from E. coli is typically used to create partially single-stranded dsDNA for subsequent polymerization reactions. Molecules labeled to a high specific activity are mainly used as hybridization probes and have two advantages over nicktranslation-prepared probes: they lack artificial hairpin structures and can easily be converted into strand-specific probes by cleavage with suitable restriction endonucleases. In combination with T4 gene 3′ protein, T4 DNA polymerase is used for gap-filling in site-directed mutagenesis. T4 Gene 32 Enzyme T4 gene 3′ enzyme is used for in vitro DNA synthesis stimulation and for RNA or single-stranded DNA region stabilization. It has also been used in site-specific mutagenesis, acting in parallel with T4 DNA polymerase and T4 DNA ligase. T4 gene 3′ is also useful in restriction-enzyme digests of large-scale DNA preparations that do not reach completion because the newly formed digested-DNA ends inhibit product formation. Addition of T4 gene 3′ helps drive these restriction digests to completion. Terminal Transferase Terminal transferase is used to add hom*opolymer tails to DNA fragments. This tailing reaction is used in cDNA cloning where DNA fragments without suitable overhanging ends must be ligated. Terminal transferase is also used to tag the 3′ends of single-stranded and double-stranded DNA with labeled nucleotides (e.g., digoxigenin, biotin). E. coli RNA Polymerase When it is not possible to clone DNA into a vector containing a phage-RNA-polymerase promoter, E. coli RNA polymerase can be used for synthesizing transcripts. RNA Polymerases (SP6, T3, T7) RNA polymerases SP6, T3, and T7 require a DNA template and Mg++ as a cofactor for RNA synthesis; they are powerfully stimulated by either BSA or spermidine and are not inhibited by the antibiotic rifampicin, in contrast to bacterial RNA polymerases. RNA polymerases SP6, T3, and T7 are tremendously promoter-specific, and they can only transcribe the DNA downstream of their appropriate promoter. The T3 and T7 promoter sequences differ only by three base pairs from one another,

PH2270_book.fm Page 79 Tuesday, May 16, 2006 9:37 AM

Enzymes

and each transcribes only the DNA cloned downstream of its corresponding promoter. SP6, T3, and T7 are used to transcribe DNA that is cloned into vectors having two phage promoters in diametric opposition, flanking a multilinker region (e.g., pSPTBM 20, pSPTBM 21, and pT3T7 lac). RNA can be selectively synthesized from either strand of the DNA insert, depending upon the polymerase used, and consistently labeled single-stranded RNAs can also be generated with this system. Transcripts can be labeled with biotin, DIG11-UTP, or labeled to a high specific activity with [32P]- or [32S]-labeled nucleotides. The resulting RNA probes are highly strand-specific, and principally useful for blotting techniques, in situ hybridization, and genomic sequencing. For RNA sequencing, the cloned DNA is transcribed with the appropriate RNA polymerase in the presence of 3′-dNTPs, which act as chain terminators in a similar manner as the ddNTP’s in Sanger sequencing. Another application of RNA probes generated by SP6, T7, or T3 RNA polymerases is in RNase protection studies. Polynucleotide Phosphorylase Polynucleotide phosphorylase catalyzes the reversible polymerization of ribonucleoside diphosphates with the release of inorganic phosphate. The highest synthesis level is obtained with a polyribonucleotide primer with a 3′-terminal OH group with ribonucleoside-5′-pyrophosphates and the divalent cations Mg+2 and Mn+2. Unlike RNA polymerases, polynucleotide phosphorylase does not require a template and cannot copy one. Polymerization can either start by elongation of a primer or from two molecules of nucleoside diphosphate. Labeled RNA fingerprints can be obtained with the enzyme by labeling the 3′-end. This is accomplished by the addition of [32P]dNDP, which is subsequently removed by treatment with nuclease P1. Polynucleotide phosphorylase is used to synthesize a variety of high-molecular-weight polynucleotides and oligonucleotides with defined sequences. RNase Inhibitor RNase inhibitor is isolated from placenta and is active over a broad pH range between 5–8 with a maximum activity at pH 7–8. The isoelectric point of the protein is reached at pH 4.7. A minimum concentration of 1 mM dithiothreitol is required to maintain RNase inhibitor in a fully active form. Under severe denaturing conditions (7 M urea or at temperatures above 65°C) the inhibitor is inactivated, which means bound RNases are released again at this temperature. RNase inhibitor inactivates RNase by noncovalently binding to the enzyme molecule. The binding ratio to RNase A is 1:1. RNase inhibitor is applied to improve cDNA synthesis, protein translation performance, in vitro RNA synthesis, in vitro virus replication, and to prepare RNase-free antibodies.

PHOSPHODIESTERASES Phosphodiesterase from Calf Spleen Phosphodiesterase from calf spleen attacks the 5′-terminal OH group and releases 3′-mononucleotides. The exonuclease is extremely sensitive to the secondary struc-

PH2270_book.fm Page 80 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

ture of the substrate. Native calf-thymus DNA and long deoxyribonucleotides still having complementary double-stranded structures are relatively resistant to enzymatic attack. Phosphodiesterase from calf spleen is used particularly in sequence studies of oligonucleotides derived from both ribonucleic and deoxyribonucleic acids. Phosphodiesterase from Snake Venom Phosphodiesterase from snake venom is an exonuclease that attacks the 3′-terminal OH group, releasing 5′-mononucleotides; its optimal pH is ≈8.9–9.8. The enzyme can also attack double-stranded high-molecular-weight DNA. Phosphodiesterase from snake venom is available exclusively from Boehringer Mannheim Biochemicals, and also contains single-stranded-specific endonuclease to convert supercoiled DNA to the open-circular form. The enzyme is primarily used in DNA and RNA sequencing studies.

MISCELLANEOUS MODIFYING ENZYMES Methylase Hpa II Methylase Hpa II is used for studying the in vivo effect of distinct methylated 5′mCpG3′ residues on gene expression within eukaryotic cells. The 5-methylcytosine residues are predominantly located in CpG sequences at a level of about 1% of total nucleotides. Protoplast-Forming Enzyme This enzyme is used for forming protoplasts in a variety of yeasts and fungi (e.g., Saccharomyces cerevisiae, Aspergillus nidulans, Neurospora crassa, etc.), which can then be transformed with exogenous DNA or used for fusion studies. Protoplastforming enzyme is functionally tested by the transformation of 1 × 1,010 cells/ml Saccharomyces cerevisiae (DBY 746, ATCC 44773) with 1 μg pBT (-1 DNA). The regeneration rate is typically greater than 20% with transformation rates of 1–5 × 104 transformed yeast cells/μg DNA. This enzyme is also useful for the isolation of high-molecular-weight DNA or RNA from lower fungi.

MANUFACTURERS’ DIRECTORY ENDONUCLEASES, RESTRICTION, DNA Accurate Chemical and Scientific (800) 645-6264 American Biochemicals, Inc. (619) 597-6050 Appligene, Inc. (800) 955-1274 Bio/Can Scientific (416) 828-2455 Bio-Rad Laboratories (510) 741-1000 Boehringer Mannheim Corp. (317) 845-2000 Bresatec, Ltd. (618) 234-2644 Carolina Biological Supply Co. (800) 334-5551

PH2270_book.fm Page 81 Tuesday, May 16, 2006 9:37 AM

Enzymes

Fisher Scientific (800) 562-1729 Gibco BRL Life Technologies (800) 828-6686 ICN Biomedicals, Inc. (800) 854-0530 Lifecodes Corp. (800) 543-3263 Molecular Biology Resources, Inc. (800) 626-7873 New England Biolabs, Inc. (800) 632-5227 Pharmacia Biotech, Inc. (800) 526-3593 Promega Corp. (800) 356-9526 Sigma Chemical Co. (800) 325-3010 Stratagene (800) 424-5444 United States Biochemical Corp. (800) 321-9322 Wako Chemicals USA (800) 992-WAKO

REVERSE TRANSCRIPTASES Accurate Chemical & Sci. (800) 645-6264 Appligene, Inc. (800) 955-1274 Bio/Can Scientific (416) 828-2455 Bio-Rad Laboratories (510) 741-6891 Boehringer Mannheim Corp. (317) 845-2000 Bresatec, Ltd. (618) 234-2644 Epicentre Technologies (800) 284-8474 Fisher Scientific (800) 562-1729 5 Prime 3 Prime, Inc. (800) 533-5703 Gibco BRL Life Technologies (800) 828-6686 New England Biolabs, Inc. (800) 632-5227 The Perkin-Elmer Corp. (800) 762-4000 Sigma Chemical Co. (800) 325-3010 United States Biochemical Corp. (800) 321-9322 Worthington Biochemical Corp. (800) 445-9603

NUCLEASES Ambion, Inc. (800) 888-8804 Appligene, Inc. (800) 955-1274 Bio/Can Scientific (416) 828-2455 Boehringer Mannheim Corp. (317) 845-2000 Bresatec, Ltd. (618) 234-2644 Carolina Biological Supply Co. (800) 334-5551 Fisher Scientific (800) 562-1729 Gibco BRL Life Technologies (800) 828-6686 Molecular Biology Resources, Inc. (800) 626-7873 Promega Corp. (800) 356-9526 Qiagen, Inc. (800) 426-8157 Sigma Chemical Co. (800) 325-3010

PH2270_book.fm Page 82 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

United States Biochemical Corp. (800) 321-9322 Wako Chemicals USA (800) 992-WAKO Worthington Biochemical Corp. (800) 445-9603

RIBONUCLEASES Accurate Chemical and Sci. (800) 645-6264 Ambion, Inc. (800) 888-8804 American Biochemicals, Inc. (619) 597-6050 American Biorganics, Inc. (800) 648-6689 Appligene, Inc. (800) 955-1274 Atomergic Chemetals Corp. (516) 694-9000 Bio/Can Scientific (416) 828-2455 Boehringer Mannheim Corp. (317) 845-2000 Fisher Scientific (800) 562-1729 5 Prime 3 Prime, Inc. (800) 533-5703 Gibco BRL Life Technologies (800) 828-6686 Promega Corp. (800) 356-9526 Qiagen, Inc. (800) 426-8157 Rowley Biochemical Inst. (508) 948-2067 Sigma Chemical Co. (800) 325-3010 United States Biochemical Corp. (800) 321-9322 Wako Chemicals USA (800) 992-WAKO Worthington Biochemical Corp. (800) 445-9603

PROTEASES AMRESCO, Inc. (800) 829-2805 Atomergic Chemetals Corp. (516) 694-9000 Boehringer Mannheim Corp. (317) 845-2000 Calzyme Laboratories, Inc. (800) 523-9127 Clontech Labs, Inc. (800) 662-2566 P.J. Cobert Assoc. (314) 993-2390 Fisher Scientific (800) 562-1729 Gibco BRL Life Technologies (800) 828-6686 Intergen Co. (800) 431-4505 Marcor Development (201) 489-5700 Pierce Chemical Co. (800) 874-3723 Promega Corp. (800) 356-9526 Sigma Chemical Co. (800) 325-3010 United States Biochemical Corp. (800) 321-9322 Viobin Corp. (608) 849-5944 Wako Chemicals USA (800) 992-WAKO Worthington Biochemical Corp. (800) 445-9603

PH2270_book.fm Page 83 Tuesday, May 16, 2006 9:37 AM

Enzymes

MODIFYING ENZYMES, DNA Accurate Chemical and Sci. (800) 645-6264 Ambion, Inc. (800) 888-8804 Appligene, Inc. (800) 955-1274 Bio/Can Scientific (416) 828-2455 Boehringer Mannheim Corp. (317) 845-2000 Bresatec, Ltd. (618) 234-2644 Epicentre Technologies (800) 284-8474 Fisher Scientific (800) 562-1729 5 Prime 3 Prime, Inc. (800) 533-5703 Fluka Chemical Co. (800) 358-5287 Gibco BRL Life Technologies (800) 828-6686 Lifecodes Corp. (800) 543-3263 Midland Certified Reagent (800) 247-8766 Molecular Biology Resources, Inc. (800) 626-7873 New England Biolabs, Inc. (800) 632-5227 Pharmacia Biotech, Inc. (800) 526-3593 Promega Corp. (800) 356-9526 Sigma Chemical Co. (800) 325-3010 Stratagene (800) 424-5444 United States Biochemical Corp. (800) 321-9322 Worthington Biochemical Corp. (800) 445-9603

PH2270_book.fm Page 84 Tuesday, May 16, 2006 9:37 AM

PH2270_book.fm Page 85 Tuesday, May 16, 2006 9:37 AM

DNA-Amplification Reagents OVERVIEW

Currently, PCR, the standard method of DNA amplification, is the oldest and most developed procedure. Several newer approaches for nucleic acid amplification have been developed, having some advantages over PCR for particular purposes — greater amplification per cycle (TAS, 3SR, Q-replicase), isothermal reaction (3SR), or coupled amplification-mutation detection (LAR, LCR, LAS) — and may eventually gain more widespread use after further development. Long before the advent of in vitro nucleic-acid amplification systems, the use of nucleic acid hybridization for amplification had long been a problem; the major impediment was the sensitivity of the detection method. Rapid, nonisotopic, nucleic acid-based systems could only be formulated for target sequences that were highly abundant. The detection of rare sequences required the use of long assays, highly radioactive probes, and large tissue samples, since the most sensitive hybridization probes — those continuously labeled with 32P can detect a sequence only when at least 104–105 target molecules are present in a sample. A major breakthrough was the development of PCR. With PCR, rare sequences in minute tissue samples could be amplified 106-fold or more in several hours, enabling the use of nonisotopic detection systems and oligonucleotide probes for rapid hybridization-based immunoassays. Since the first PCR study appeared in 1985, a multitude of modifications to the basic protocol have appeared, extending the procedure to a diverse range of applications. PCR also stimulated the development of several other in vitro nucleic-acid amplification systems such as: the transcription-based amplification system (TAS); the self-sustained sequence replication (3SR) system; the ligation amplification reaction (LAR), ligase chain reaction (LCR) or ligase-based amplification system (LAS); and an RNA replication system based on Q-replicase. PCR, LCR, Q-replicase, and nucleic acid sequence-based amplification (NASBA) all repeatedly copy a nucleic acid sequence to emphasize its presence. Another approach to DNA or RNA identification, amid the billions of genomic base pairs, is to tag the target gene chemically and detect the tag.

AMPLIFICATION BY DNA SYNTHESIS (PCR) The PCR system is simple in concept yet powerful in application, involving repeated cycles of DNA polymerase-mediated primer extension (see Figure 6.1). Synthesis 85

PH2270_book.fm Page 86 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

5ʹ 3ʹ

3ʹ ds DNA Target 5ʹ Heat Denature + Primers A and B

5ʹ

3ʹ 5ʹ

Cycle 1

3ʹ

5ʹ

3ʹ

5ʹ DNA Polymerase

5ʹ

3ʹ 3ʹ 5ʹ

5ʹ 3ʹ

3ʹ

5ʹ Heat Denature + DNA Polymerase

5ʹ

3ʹ 3ʹ

5ʹ B

A 5ʹ 3ʹ 5ʹ 3ʹ

Cycle 2

3ʹ 5ʹ 3ʹ 5ʹ A

5ʹ

3ʹ

5ʹ

Repeated Cycle Lead to Amplification of the Target Sequence

FIGURE 6.1 Polymerase chain reaction (PCR) amplification system; PCR amplification involves repeated cycles of DNA polymerase-mediated primer extension.

of the target DNA sequence is directed by two oligonucleotides that bracket the target on opposite DNA strands. By using a thermostable DNA polymerase, the reaction is repeatedly cycled through alternating heat denaturation, primer hybridization, and primer extension steps, until sufficient target sequence amplification has been achieved. Typically, a 106-fold amplification can be achieved in 2–4 hours utilizing thirty thermal cycles. Target sequence amplification using PCR has several advantages: (1) The reaction uses a single thermostable enzyme and can be performed in a single tube without repeated reagent additions; (2) thermal cycling can be easily automated by using commercially available programmable temperature cycling devices; (3) RNA target sequence amplification can be easily accomplished by synthesizing cDNA with reverse transcriptase prior to PCR; (4) since the reaction

PH2270_book.fm Page 87 Tuesday, May 16, 2006 9:37 AM

DNA-Amplification Reagents

product generally has a discrete number of base pairs, the success of the reaction can frequently be estimated by the mobility of the product during gel electrophoresis. PCR does have several problems: (1) the number of thermal cycling steps involved in the reaction (e.g., for thirty PCR cycles, ninety temperature transfers must be performed and the incubation time at each temperature must be monitored, thus, using automated equipment is mandatory when PCR is run routinely); (2) the number of product molecules only doubles with each PCR cycle, and even then, this low efficiency per cycle is seldom achieved in execution — as a result, cycles must be increased to reach a specific amplification level; (3) PCR cannot distinguish RNA from DNA in mixed nucleic acid samples when the RNA and DNA sequences are colinear, so direct RNA detection requires the removal of all DNA; (4) all primer sequences are not equally efficient, and unfortunately, the art of selecting primer sequences has not as yet been fully mastered — each primer pair’s appropriateness must be determined experimentally, and the reaction conditions for each target and primer pair frequently requires optimization; (5) extraneous sequences can be coamplified due to nonspecific hybridization of the primers, and the severity of this problem increases with amplification level; (6) false positives can result from reagent or equipment contamination with the end products of previous reactions or from extraneous DNA contamination through aerosols generated by uncovered coughs or sneezes. This contamination represents the greatest problem in performing any target amplification reaction. Since PCR and other amplification reactions can generate more than 106 copies of the target sequence, starting with just one initial molecule, accidental introduction of previous reaction end product or external contamination easily produces false positives when high amplification levels are necessary. Consequently, by amplifying DNA in other than the desired sequence, PCR can be too powerful for its own good; follow-up electrophoresis or employing a second probe of the amplified sequence is usually necessary to assure the desired target sequence has been amplified. As previously mentioned, automation can typically deal with contamination by not distributing splashes and aerosols, since contamination can be introduced in the lab and even in sample handling. Complexity of the PCR method is a hindrance to its routine use in immunoassay. The method is significantly labor-intensive even with automation, and in view of these limitations, PCR is currently in competition with other nucleic acid amplification methods such as LCR, Qβ-replicase, and NASBA. In addition, the hom*ogeneous gene assay may also replace PCR for some applications.

AMPLIFICATION BY RNA TRANSCRIPTION (TAS AND 3SR) TAS was the first amplification system based on transcription. Figure 6.2 shows the basic TAS protocol that involves generating duplex cDNAs that contain the bacteriophage T7 transcription promoter, followed by T7 RNA polymerase-mediated transcription of the cDNA to achieve target sequence amplification. First strand cDNA synthesis is accomplished by primer extension of oligonucleotides containing the polymerase binding sequence (PBS) at their 5′-end, using avian myeloblastosis

PH2270_book.fm Page 88 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

Step A 5' Step B cDNA Synthesis II

5' RNA 3' A TCS

3' 5'

5' RNA 3' DNA Δ + RT + Primer B 5' RNA 3' DNA + B 5' DNA 3' DNA + T7, T3, or SP6 Polymerase 3' RNA 3' RNA 3' RNA 3' RNA 3' RNA 3' RNA 3' RNA 3' RNA 3' RNA 3' RNA + B 5' DNA 3' RNA

A 3'

Step C

A 3' 5' Cycle I of TAS 5' 5' 5' 5' 5' 5' 5' 5' 5' 5'

Step D

RNA Transcription I

5' Δ + RT

Step E

5' DNA 3' DNA A 3' 5'

B 5' DNA 3' RNA Δ + RT

cDNA Synthesis II

5' DNA 3' DNA

Step F A

B 5' DNA 3' DNA

5' A

B 5' DNA 3' DNA

Cycle II of TAS

Step G

RNA Transcription II

A 3' 5'

B 5' DNA 3' RNA + T7, T3, or SP6 Polymerase

5' 5' 5' 5' 5' 5' 5' 5' 5' 5'

3' RNA 3' RNA 3' RNA 3' RNA 3' RNA 3' RNA 3' RNA 3' RNA 3' RNA 3' RNA +

3' 5'

B 5' DNA 3' RNA

FIGURE 6.2 Transcriptional Amplification System (TAS); basic protocol involves generating duplex cDNAs containing the bacteriophage T7 transcription pro-mater, followed by T7 RNA polymerase-mediated transcription of the cDNA to amplify the target sequence; additional target sequence amplifications are completed with added cDNA synthesis cycles and transcription of end-product RNA.

PH2270_book.fm Page 89 Tuesday, May 16, 2006 9:37 AM

DNA-Amplification Reagents

virus (AMV) reverse transcriptase and denatured or target RNA as a template. Following thermal denaturation of the RNA-DNA duplex, another primer directs second strand synthesis of the cDNA. The duplex cDNA, containing a promoter, is then transcribed with T7 RNA polymerase to produce the antisense RNA amplification product. Additional target sequence amplification can be completed by added cycles of cDNA synthesis and transcription of the RNA product. In PCR, only two target copies can be produced per cycle. TAS, for example, produces between 10 and 100 copies per cycle, yielding a 106-fold amplification of the target in only four to six cycles (3–4 hours). The main drawback to TAS is the number of enzyme additions required, since the reverse transcriptase and RNA polymerase lack thermostability. With 3SR the thermostability problem is avoided with a modification of the TAS protocol — 3SR, which utilizes enzymatic RNA degradation in the RNA-DNA heteroduplexes in place of thermal denaturation (see Figure 6.3). Scheme A Step 1. 3'

Scheme B 5'

5' A'

3' A

2. B

RNase H

3. 4. 3'

5. B 6.

A RT

RT 7. RNase H

RNase H

10.

RT A'

9. 3'

FIGURE 6.3 3SR TAS; the thermostability problem is avoided with a modified TAS protocol, 3SR, utilizing enzymatic RNA degradation in the RNA-DNA heteroduplexes instead of thermal denaturization; adding RNase H to the protocol permits the reaction to be performed in a single tube at a single temperature without reagent addition once the reaction is initiated. With 3SR, sequences are amplified 106–109 times faster in a 1-hour incubation cycle at 42°C. As with TAS, the majority of the amplified product is single-stranded RNA, having either the antisense sequence of the target (scheme A), the same sequences (scheme B), or both, depending on which primers have the T7 PBS. Approximately 1% of the amplified product exists as double-stranded cDNA and as RNA-cDNA heteroduplex.

PH2270_book.fm Page 90 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

The addition of RNase H to the TAS protocol permits the reaction to be performed in a single tube at a single temperature without reagent additions, once it is initiated. With 3SR, target sequences are amplified 106- to 109-fold in a 1-hour incubation cycle at 42°C. As with TAS, the majority of the amplification product is single-stranded RNA, which can have either the antisense sequence of the target (Figure 6.3, scheme A), the same sequences (scheme B), or both, depending on which primers have the T7 PBS. Approximately 1% of the amplification product exists as double-stranded cDNA and as RNA-cDNA heteroduplex. The 3SR system has a number of advantages: (1) The reaction is simple to perform. The procedure has no manipulative cycles. All reagents are pipetted into a single tube and incubated at a single temperature, so that no thermal cycling, repeated reagent additions, or tube transfers are required. (2) 3SR is rapid;104-fold amplification is achieved in 15 minutes and the reaction is complete (a 106-fold amplification) in less than 1 hour. In contrast, the same level of amplification by either the PCR or TAS systems requires 3–4 hours. (3) The majority of the 3SR product is single-stranded RNA, which can be measured by a quantitative hybridization system. Hybridization systems, such as the bead-based sandwich system, do not perform well with the double-stranded DNA product from PCR; the single-stranded product can also be used directly for nucleic acid sequencing. (4) The method will specifically amplify RNA in mixed nucleic acid samples. Application of the 3SR system to DNA target sequences requires the use of thermal denaturation during the initial synthesis of cDNA incorporating the T7 promoter sequence. Without the use of these thermal denaturation steps, duplex DNA will not serve as a substrate for the 3SR reaction. Disadvantages of the 3SR system mainly involve the enzymes used in the reaction. Thermostable enzymes for this system are not yet available and, therefore, the maximum incubation temperature for efficient amplification is only 42°C. As a result, the amount of specific product from 3SR must be quantitated by hybridization, since 3SR may produce significant levels of nonspecific RNA.

AMPLIFICATION BY LIGATION (LAR, LCR, LAS) DNA ligase is used to amplify target sequences through the repeated joining of oligonucleotides that hybridize to the target. This technique, LAR, LCR, and LAS, uses four oligonucleotides — two per target strand (see Figure 6.4). After hybridization of the two oligonucleotides to adjacent sequences of the target strand, the two oligonucleotides are joined by DNA ligase to form the product, which is then separated from the target sequence by a heat denaturation. Then both the ligation product and the target serve as a substrate for the next cycle of hybridization and ligation. Ligase is an enzyme that normally knits together small, newly synthesized pieces of DNA to build the daughter strand. An LCR probe consists of two halves of a DNA sequence complementary to a target sequence. If the sample contains the targeted sequence, the half probes zero in on it and are stitched together by the ligase. As with PCR, the initial product serves as a template for the next round of replication, with millions of copies generated after 20–30 cycles. A thermostable ligase is currently available. LCR handles target sequences up to 50 base pairs,

PH2270_book.fm Page 91 Tuesday, May 16, 2006 9:37 AM

DNA-Amplification Reagents

ds DNA

Denature Hybridize A, B, A', B'

Step A 5'

3' 3'

B' 5'

Step B

Ligate 5'

3' 3'

5' B

5' Denature Hybridize

Step C 5'

3' 3'

B' 3'

A' 3'

5' 3'

5' Ligate

Step D

Repeated Cycles of Steps C and D

FIGURE 6.4 Ligase-based amplification system; DNA ligase can be used to amplify a target sequence through the repeated joining of oligonucleotides that hybridize to the target. This technique, LAR, LCR, and LAS, uses four oligonucleotides — two per target strand. After hybridization of the two oligonucleotides to adjacent sequences of the target strand, the two oligonucleotides are joined by ligase to form the product, which is then separated from the target sequence by a heat denaturation; then both the ligation product and the target serve as a substrate for the next cycle of hybridization and ligation.

PH2270_book.fm Page 92 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

whereas PCR goes up to 2,000 base pairs. While PCR is currently used for cloning and sequencing, LCR is not appropriate for those applications at this stage of development. A tremendous advantage of using LCR in immunoassays is speed. Another is the easy assay requirements. With PCR, much depends on operator skill. LCR can be used by the inexperienced operator and still produce comparable results. LCR is also more reliable. With PCR, there is always the possibility that when an extraneous sequence is multiplied, a confirmatory test is required. The presence of a PCR product is not evidence that the target sequence is present, although with LCR, the presence of the amplified product is proof of the target sequence being present. Because LCR detection does not require characterization and confirmation, as does PCR, LCR is more easily automated. The greatest advantage of LCR is its potential to combine amplification with detecting genetic mutations. Mispairing of oligonucleotides to the target sequence at the bases to be joined prevents ligation. As a result, amplification success is dependent upon whether the probe oligonucleotides contain either a normal or a mutant sequence. The ligase-based systems suffer from two major drawbacks at present: (1) the E. coli DNA ligase and the bacteriophage T4-DNA ligase can produce blunt-end ligation of duplex oligonucleotides, as well as joining single-stranded oligonucleotides; while the efficiency of these two activities is low, the large excess of free oligonucleotides, relative to oligonucleotides hybridized to the target sequence in the ligation reaction, can result in high background levels due to template-independent ligation — reaction conditions that inhibit formation of templateindependent product have been discovered, but these lower the efficiency of the template-dependent ligase activity; (2) the ligase-based system has limited amplification efficiency. Like PCR, each LCR cycle should yield a twofold amplification of the target sequence. However, the kinetic properties of the ligases make this difficult, especially when genomic DNA is used as the target. In order to obtain a 98+% efficiency, the reaction must be incubated for 5 hours per cycle; shorter cycle times result in lower ligation efficiency and require additional cycles that must be completed.

AMPLIFICATION BY RNA REPLICATION (Qβ REPLICASE) The replicase enzyme polymerase from the RNA bacteriophage Qβ amplifies hybridization signals (see Figure 6.5). The technique is as intense for one sample as another — the first sample has 10,000 times as much propagated DNA as the second. The calibrations made possible by the kinetics, plus the fact that unlike PCR and LCR, Q-replicase does not require temperature shifts, make it a good candidate for immunoassay. It can also be run under the same conditions for a variety of tissue types, which is not true of PCR. Despite its originality, however, Qβ-replicase amplification has a tarnished reputation, since the enzyme is extremely unstable and therefore difficult to control; when it copies RNA, it does so unevenly, and when it is probed, mutations typically occur. Thus, if a single nonspecific molecule comes through, it might result in a false positive. Steps have been taken to increase the number of

PH2270_book.fm Page 93 Tuesday, May 16, 2006 9:37 AM

DNA-Amplification Reagents

5ʹ

Step 1

3ʹ (+) RNA Target Qβ 5ʹ Sequence

3ʹ A

T7 Promoter Sequence

Denaturation Step + RT and Primer B 5ʹ 3ʹ Step 2

5ʹ Qβ 3ʹ Sequence

3ʹ (+) RNA DNA A

3ʹ DNA 5ʹ DNA

3ʹ + T7 RNA Polymerase

Step 3

3ʹ

5ʹ (−) RNA + Qβ Replicase

Step 4

5ʹ

3ʹ (+) RNA

3ʹ

5ʹ (−) RNA

FIGURE 6.5 Q-replicase amplification system; substrate RNA for the replicase is synthesized by performing one cycle of the TAS protocol using primers that contain the Q-replicase 5′and 3′-recognition sequences; replicase enzyme polymerase from the RNA bacteriophage Q amplifies the hybridization signal. Q-replicase requires no temperature shifts, which makes it a good candidate for immunoassay. Q-replicase, however, gets high false positives through contamination.

washes to develop the system’s reliability, but Q-replicase still has an unacceptably high number of false positives due to contamination. For signal amplification, a probe rRNA sequence containing both the required 5′- and 3′-replicase recognition sequences is hybridized to immobilized target sequences. After extensive washing to remove unhybridized probe RNA, the probe RNA is eluted from the target sequence and incubated with the Q-replicase. A 107fold amplification of the probe sequence was observed in the 20-minute incubation. However, significant background problems were encountered. Any small amount of probe RNA that is nonspecifically retained also becomes highly amplified due to the high efficiency of the Q-replicase. The Q-replicase method is useful in target amplification as well. In the protocol shown in Figure 6.5, substrate RNA for the replicase is synthesized by performing one cycle of the TAS protocol using primers that contain the Q-replicase 5′- and 3′-recognition sequences. The TAS product is further amplified by replication of the RNA with the Q-replicase. Such a system has the potential to produce a 106- to 109-fold amplification of a target sequence in one hour, allowing thirty minutes for one cycle of TAS and thirty minutes for Qreplication. Since formation of the replicase substrate in this protocol is dependent

PH2270_book.fm Page 94 Tuesday, May 16, 2006 9:37 AM

Producing Biomolecular Substances

upon successful synthesis of the TAS product, background problems are much less likely than with signal amplification systems.

SEQUENCE-BASED AMPLIFICATION NASBA employs cycling DNA and RNA syntheses (Figure 6.6). The reactions are hom*ogeneous and do not require the addition of enzymes or temperature cycles. Templates are added to the medium with the enzymes and primers, and the reaction is allowed to proceed at 42°C for 1–2 hours, depending upon the amount of template. This yields 102–104 target molecule copies; the amplification can be confirmed by agarose gel electrophorresis, a slot blot, or a dot blot. NASBA can detect as few as 10 molecules in a mixture of 1010 molecules. Although simpler than LCR and PCR, in that the entire reaction is conducted at one temperature and can be performed on RNA or DNA, the complexity of the procedure may be more than a lab technician can handle. It’s unlikely that the procedure can be automated efficiently because there are too many steps and reagents involved in the procedure. However, the method is valuable in research, where all variables can be controlled. Scheme A Step 1. 3ʹ

Scheme B 5ʹ

5ʹ Aʹ

3ʹ A

2. B

RNase H

3. 4. 5.

3ʹ

5ʹ

3ʹ

B 6.

A RT

RT 7. RNase H

Bʹ

RNase H

8. RT

Aʹ

5ʹ

9. 10.

3ʹ

FIGURE 6.6 Sequence-Based Amplification (NASBA).

5ʹ

3ʹ

PH2270_book.fm Page 95 Tuesday, May 16, 2006 9:37 AM

DNA-Amplification Reagents

MANUFACTURERS’ DIRECTORY DNA-AMPLIFICATION REAGENTS Advanced Polymer Systems (415) 366-2626 Appligene, Inc. (800) 955-1274 BIO 101, Inc. (800) 424-6101 Boehringer Mannheim Corp. (317) 845-2000 Clontech Labs, Inc. (800) 662-2566 Fisher Scientific (800) 562-1729 Midwest Scientific (800) 227-9997 National Biosciences, Inc. (800) 747-4362 The Perkin-Elmer Corp. (800) 762-4000 Qiagen, Inc. (800) 426-8157 Stockwell Scientific (800) 722-8920 (in CA 800-248-0426) Synthetic Genetics, Inc. (800) 562-5544 TosoHaas (215) 283-5000 Tri-Continent Scientific, Inc. (800) 937-4738 United States Biochemical Corp. (800) 321-9322 Zymed Labs (800) 874-4494

PH2270_book.fm Page 96 Tuesday, May 16, 2006 9:37 AM

PH2270_C007.fm Page 97 Tuesday, May 23, 2006 10:01 AM

Cell-Culture Media OVERVIEW

Cell-culture media generally support the in vitro existence and growth of living organisms or their parts such as viruses, microbes, fungi, and plant, and animal cells and tissues. The media may be either liquid or thickened by the addition of a neutral gel former such as agarose. There is no single universal growth medium since individual growth requirements vary widely; however, all media generally consist of the following components: • • •

• • •

Water Energy (food) carbon source (e.g., glucose) Nitrogen sources (e.g., speciﬁc amino acids, since organisms differ in ability to synthesize amino acids and essential amino acids sometimes must be added to the media as nutrients) Inorganic salts used as trace mineral sources and buffers Vitamins Oxygen; sometimes CO2 and/or other gases

MEDIA PROKARYOTIC CELL-CULTURE MEDIUM Prokaryotic cell-culture medium is composed of salts, buffers, vitamins, hormones, amino acids, sugars or other carbon sources for use as food, and other deﬁned compounds. Each cell line requires its own particular formula for optimal growth, and is available sterilized and bottled, ready for use, or in packaged and powdered form, as illustrated by recombinant E. coli medium in Table 7.2.

EUKARYOTIC CELL-CULTURE MEDIUM Unlike prokaryotic cells, eukaryotic cells, and especially mammalian cells, require specialized cultivation techniques since they have delicate cell walls, do not grow well in solution, and have particular nutritional and environmental requirements. For example, eukaryotic cells may require: • •

Small quantities of antibiotics to prevent contamination; Special enzymes and hormones;

PH2270_C007.fm Page 98 Tuesday, May 23, 2006 10:01 AM

Producing Biomolecular Substances

TABLE 7.1 New Developments in Cell-Culture Media

• •

Development

Company

Comments

Serum — 40 nm ﬁltration Magnetic microcarriers Chambered slides Insect culture media

Hyclone Dynal Nunc JR Scientiﬁc

All FBS ﬁltered Improves harvesting Glass or plastic slides Baculovirus applications

Other proteins and peptides; and Fatty acids and vitamins.

Eukaryotic cell media may include a pH indicator and might either contain natural blood products or be synthetically composed of biochemicals. The culture system must be temperature regulated with fresh nutrients replenished regularly. For example, if protein recovery is the purpose, the media must be designed to maximize both cell proliferation and antibody or recombinant protein production (see Figure 7.1); the medium might even be serum-free to reduce the concentration of material that might copurify with the end-product proteins or antibodies; also, the medium might be specially buffered to compensate for lack of intrinsic buffering. Newer developments in eukaryotic cell-culture media formulation are summarized in Table 7.1.

DEFINED MEDIUM Cell-culture medium is the nonserum portion of a cell-culture nutrient solution composed of salts, buffers, vitamins, hormones, amino acids, sugars, and other deﬁned compounds (thus, the term defined medium). Each cell line is known to require its own particular formula of ingredients for optimal growth. Media are available in sterilized bottled form, ready for use, or in packaged and powdered form (Table 7.2).

SERUM Serum must be of both low toxicity and high purity and, regardless of origin, lotto-lot variation must be minimal. Both sera obtained from bovine fetuses of 16–36 weeks gestation and sera obtained from calves about 15 weeks postpartum are commonly used, and may be supplemented with iron. Natural variation occurs on a lot-to-lot basis due to variations in diet, time of day, season of slaughtering, age of calf or fetus, body weight of parent, geographic location, slaughtering stress, veterinary management of the animals, and so forth. Serum for biopharmaceutical purposes is either domestically collected and processed under strict guidelines or imported under certiﬁcation. Sera not processed under guidelines are used only for research purposes and not for biopharmaceutical applications. Sera are generally added to culture media in a ratio of about 7% to 93% medium — by far, the serum

PH2270_C007.fm Page 99 Tuesday, May 23, 2006 10:01 AM

Cell-Culture Media

Daily Antibody Concentration

Concentration (μg/ml)

120 100 80 60

Fed-Batch Perfusion

40 20 10 0

Days

Daily Production (grams)

Daily Antibody Production

30 20

Fed-Batch Perfusion

10 0 0

Cumulative Production (grams)

Days Cumulative Production

100 80 60 40

Fed-Batch Perfusion

20 10 0 0

Days Production Per Liter of Media Consumed

Production Per Liter of Media Consumed (mg)

120 100 80 60

Fed-Batch Perfusion

40 20 10 0

30 Days

FIGURE 7.1 Fed-batch vs. perfusion cell-culture systems.

PH2270_C007.fm Page 100 Tuesday, May 23, 2006 10:01 AM

100

Producing Biomolecular Substances

TABLE 7.2 Typical Recombinant E. coli Production Medium Presterilization Medium (added to fermenter before sterilization) Yeast Extract 5.0 g/l Ammonium Sulfate 4.0 g/l Potassium Phosphate Monobasic 7.0 g/l Potassium Phosphate Dibasic 8.0 g/l L-6 0.1 ml/l Poststerilization Medium (added aseptically to fermenter after sterilization from separate, sterile stock solution) Glucose 5.0 g/l Magnesium Sulfate-7-Hydrate 1.0 g/l 2.0 ml/l Trace Metals*,** 2.0 ml/l Vitamins*,*** Tetracycline* 15.0 mg/l * Filter sterilized ** Trace Metals Ferric Chloride-1-Hydrate 27.0 g/l Zinc Chloride-4-Hydrate 2.0 g/l Sodium Molybdate-2-Hydrate 2.0 g/l Cupric Sulfate-5-Hydrate 1.9 g/l Calcium Chloride-6-Hydrate 2.0 g/l Hydrochloric Acid 100.0 ml/l Boric Acid 0.5 g/l *** Vitamins Riboﬂavin 0.42 g/l Pyridoxine 1.4 g/l Pantothenic Acid 5.4 g/l Biotin 0.06 g/l Niacin 6.0 g/l Folic Acid 0.04 g/l

is the most expensive component in the mixture. Because of its high cost and unpredictable availability, there is a trend toward replacing animal sera in culture medium with deﬁned medium.

BOVINE SERUM Bovine serum is a key component of mammalian cell-culture medium. While the factors promoting eukaryotic cell growth in vitro are not completely understood, it has been determined that bovine serum does enhance cell viability in culture.

PH2270_C007.fm Page 101 Tuesday, May 23, 2006 10:01 AM

Cell-Culture Media

101

TABLE 7.3 Reported Bovine Fetal and Placental Infections Herpesvirus group

Adenovirus group Parvovirus (2)

Picomavirus group Reovirus group Paramyxovirus group Rinderpest group Togavirus group

Retroviridae

DNA Viruses Pseudorabies Infectious bovine rhinotracheitis Malignant catarrhal fever (African) Bovine adenovirus

RNA Viruses Bovine enterovirus Bovine rhinovirus (foot & mouth disease) Bluetongue Parainﬂuenza (3) Rinderpest Wesselsbron virus Rift Valley fever Bovine viral diarrhea Bovine leukemia virus (4) Unclassified Viruses Bovine syncytial virus (probably RNA) Tickborne fever

OTHER SERA Other sera generally available are fetal calf and equine. Fetal serum, used for mammalian tissue and cell culture, is particularly effective since it does not contain antibodies that are found in postpartum sera.

SERUM CONTAMINATION At least seventeen viruses, plus various bacteria and parasites, pass the bovine placenta and infect the fetus (Table 7.3). Some of the most common viral contaminants of FBS are infectious bovine rhinotracheitis (IBR) and parainfluenza 3 (PI3); all commercially produced lots of FBS are contaminated with bovine viral diarrhea (BVD). At best, pools of raw FBS probably contain at least 104 infectious BVD viral particles per milliliter, and various antibodies are usually bound to most of these viruses. Screening for BVD usually yields false negatives, however, and investigators using FBS must generally determine if BVD and its associated antibodies inﬂuence the cells they want to cultivate. Other potential FBS contaminants are teratogenic substances, toxins, pesticides, herbicides, heavy metals, antibiotics, and aflatoxins from the fodder of pregnant cows, microorganisms that contaminate the blood during collection and processing, and cellular components released from ruptured blood cells. Manufacturers’ testing has shown that most contaminating substances in FBS are so diluted in the large serum pools that they generally do not inﬂuence cell

PH2270_C007.fm Page 102 Tuesday, May 23, 2006 10:01 AM

102

Producing Biomolecular Substances

growth. Various cell lines may require changes in media (e.g., a move from FBS to iron-supplemented calf serum, or the addition of a particular growth or attachment factor for optimal results). Endotoxin and hemoglobin levels are commonly used to evaluate serum quality. Endotoxin concentration indicates the level of microbial contamination prior to ﬁlter sterilization. Serum is a natural detoxicant, and toxic pooled serum indicates that the fetal blood has not been properly collected and processed. Hemoglobin concentration indicates the degree of blood cell hemolysis during collection and processing. FBS shelf-life is inﬂuenced by enzymatic activity, cross-linking of proteins with sugars (browning), autooxidation, photo oxidation, and salt precipitation. In the past, shortages and increased demand for FBS led to high prices that tended to promote illegal activity, speciﬁcally the smuggling of low-cost foreign serum, with far-reaching implications — infection of cell cultures used for biopharmaceutical manufacture with exotic disease agents, and threat to the domestic beef industry by contamination of the serum used to produce attenuated animal vaccines or other biologics with foot and mouth disease virus (one of the smallest viruses known not signiﬁcantly retained by current ﬁltration systems, and sometimes resistant to gamma-ray sterilization). Currently, the USDA lacks funds to inspect all meat-packing plants in approved countries to ascertain that serum is collected according to regulations. Smuggled serum is transported directly into the U.S. or indirectly through another country (i.e., Canada or Mexico). Serum with 0.1 μ ﬁltration has been available in the industry for some years, and triple-serial ﬁltration has recently reduced concerns about FBS as a source of mycoplasma contamination. One vendor, Hyclone™, uses a dual-ﬁlter system for all its FBS products. In practical terms, this means that the ﬁlter system will trap all particles larger than 75 nm, none smaller than 24 nm, and particles between those sizes according to a normal Gaussian (bell-curve) distribution. With the intense scrutiny placed on components used in vaccines and bioengineered products, however, serum-free media has become the industry standard.

SERUM-FREE (DEFINED) MEDIUM Deﬁned medium, or serum-free medium, is a nutrient solution that retains the viability and growth of cells in vitro, but in which animal sera has been replaced with a proprietary mix of growth factors, hormones, carrier proteins, and other biological modiﬁers (see Table 7.2). Deﬁned medium is not yet available for the growth of all cell lines but its use is growing. Deﬁned medium is generally preferred because of its consistency and because its use simpliﬁes subsequent product puriﬁcation steps. Given the supply problems and infectious disease problems of FBS, and also considering its lot-to-lot variation, there is a deﬁnite need for serum-free media in bioprocessing. Formulations designed for high-density hybridoma cell-line growth and optimized for the production of monoclonal antibodies require a reduction in substances that may copurify with the secreted antibody, promoting improvements in both the basic nutrient medium and the deﬁned protein supplements, hormone and/or growth factor supplementation, carrier proteins, addition of more

PH2270_C007.fm Page 103 Tuesday, May 23, 2006 10:01 AM

Cell-Culture Media

103

Relative Cell Number 128

1 Culture Time

FIGURE 7.2 Principle of microcarrier cell culture.

nutrient and cells during growth instead of more serum, and passage of the culture through a series of nutrient media formulations progressively decreasing in serum content. Hybridoma cell growth parameters vary widely. In general, rapidly proliferating cells usually exhibit greater requirements for energy substrates and require a more highly enriched medium. High-density cell cultures with low serum supplementation and deﬁned protein supplement levels require more frequent medium exchange or subculture.

CELL-CULTURE GELS Cell culture and cloning cell lines in agar is over a century old. Certain cell lines that do not grow in agar will grow in agarose. Some cell lines grow ten to twenty times better in agarose, and some that lose their mutagenicity in agar retain it in agarose. Gel melting and congealing temperatures are important in applications involving the recovery of living cells or physiologically active factors. Agarose has applications in plant, insect, and mammalian cell culture. Tissue-culture methodology evolved from early twentieth century embryology work that conﬁrmed a reproducible technique for in vitro cell growth with the continuation of normal cell function. Traditional cell-culture techniques began to evolve with the discovery of the growth-promoting effects of embryo extracts on cell cultures. Modern methods of tissue and cell culture have progressed to culturing cells directly on glass, producing cultures from individual cells, and purposely propagating speciﬁc cells. In order to cultivate eukaryotic cells in vitro, conditions such as those in vivo must be reproduced as closely as possible.

PH2270_C007.fm Page 104 Tuesday, May 23, 2006 10:01 AM

104

Producing Biomolecular Substances

TABLE 7.4 Amino Acids Nonessential

Essential

Alanine Asparagine Aspartic acid Cysteine Glutamine Glutamic acid Glycine Proline Serine Tyrosine

Arginine Histidine Isoleucine Leucine Lysine Methionine Phenylalanine Threonine Tryptophan Valine

Amino Acid Derivative 4-Hydroxyproline 5-Hydroxylysine 6-N-Methyllysine Desmosine γ-carboxyglutamic acid

MEDIA SUPPLEMENTS In many applications using deﬁned medium for in vitro cell cultivation, the addition of supplemental nutrients and reagents is required. Antibiotics, buffers, and reagents are frequently added to prevent bacterial contamination, control pH, and visibly monitor medium conditions (see Table 7.2). Other supplements such as amino acids and vitamins are useful for enriching media beyond normal concentrations. Attachment factors and transport factors are often added to growth media to facilitate rapid cell propagation and enhance proper cell metabolism. Still other media supplements may be added to create conditions that will elicit speciﬁc cellular responses and effects (e.g., altered protein biosynthesis, accelerated antibody production and secretion, toxic response mechanisms, etc.). As a result, a wide variety of media supplements is available to those designing cell-culture bioprocesses. In the following section, the broad range of media supplements frequently used in cell-culture work is discussed.

AMINO ACIDS Amino acids are organic compounds containing one or more basic amino groups and one or more acidic carboxyl groups that are polymerized to form peptides and proteins (Table 7.4). Only 10 of the over 80 amino acids found in nature serve as building blocks for cellular proteins; these are termed essential amino acids. The rest are termed nonessential amino acids.

ANTIBIOTICS Antibiotics are routinely added to cell-culture growth media to control bacterial contamination. Antifungal and antiyeast compounds are also used in cell cultures to prevent contamination. Flawed aseptic technique and overconﬁdence in antibiotics are some of the fundamental causes of contamination in cell cultures; dependence on antibiotics can conceal imperfect technique. Many antibiotics are extremely toxic to cells, although toxicity is generally dependent on cell-line sensitivity. Since anti-

PH2270_C007.fm Page 105 Tuesday, May 23, 2006 10:01 AM

Cell-Culture Media

105

TABLE 7.5 Typical Luria Broth Formula with Tetracycline Medium Tryptone Yeast Extract Sodium Chloride Tetracycline*

10.0 5.0 5.0 15.0

g/l g/l g/l mg/l

* Filter sterilized

biotics are generally used in routine cell-culture work, care must be exercised not only in choosing the antibiotic, but in selecting the concentration to be used. The following antibiotics are common additions to cell-culture media (see also Table 7.5).

AMPHOTERICIN-B Amphotericin-B is an antifungal agent produced by Streptomyces sp., and is an effective against yeasts and fungi at 2.5 mg/l but not against Mycoplasma, sp.

AMPICILLIN Ampicillin is effective against both gram-positive and gram-negative bacteria at a concentration of 100 mg/l.

TYLOSIN SOLUTION Tylosin solution (5 mg/ml) is effective against Mycoplasma at a concentration of 2 ml/l.

ATTACHMENT FACTORS Attachment factors are essential for the binding and spreading of anchor-dependent cells on culture-vessel surfaces, and are particularly important when culturing cells that are incapable of providing their own biomatrix. Anchor-dependent cell-nutrient requirements are indicated by the type of substrate on which they are grown. For anchor-dependent cell production, additional surface area is generally added to the culture vessel by mixing in small polymer microcarrier beads so that the resulting cell-microcarrier aggregates increase the effective surface area of the culture vessel. Other variations encapsulate cells in a gel matrix, utilize ceramic microcarrier particles, or grow cells in hollow-ﬁber cartridges.

BUFFERS Buffers are mixtures of inorganic salts known as a physiological or balanced-salt solutions (see Table 7.6). The functions of these salt solutions in the cell-culture medium are as follows:

PH2270_C007.fm Page 106 Tuesday, May 23, 2006 10:01 AM

106

Producing Biomolecular Substances

TABLE 7.6 Sample Cell-Disruption Buffers Trizma Base EDTA pH NaOAc EDTA pH

• • • •

50.0 5.0 8.0 50.0 1.0 5.5

mM mM mM mM

Maintain pH Maintain osmotic pressure Provide a source of energy Disrupt cells to release desired intracellular biomolecules

Mammalian cells can survive over the wide pH range of 7.8–11.6, but generally optimal cell growth is accomplished between pH 7.2–7.4, and it is permissible to allow the pH to deviate outside the limits of pH 6.8–7.6 to function effectively. The pKa of the chosen buffer should be as close as possible to the required pH. No buffer, however, is capable of holding the pH constant in a system in which acids or bases are being produced; buffers only slow the rate of pH change and cells in culture produce acidic products that act to lower the pH of the medium. Most media utilize phosphates and a bicarbonate system for buffering. The bicarbonate ion can be converted to gaseous carbon dioxide and lost from the medium, resulting in a pH rise. Carbon dioxide can be maintained by supplying a gas phase (5% CO2 to 95% air in a CO2 incubator) or by sealing the vessel so that the CO2 produced by metabolic processes is retained in the vessel and reabsorbed by the medium.

SODIUM BICARBONATE The most commonly used buffer in cell-culture media is sodium bicarbonate, which dissociates according to the following equation: NaHCO 3 + H 2 O ↔ Na + +HCO -3 +H 2 O ↔

(7.1)

Na + +H 2 CO 3 +OH - ↔ Na + +OH - +H 2 O+CO 2

(7.2)

This buffer, however, has two important disadvantages: (1) the pKa1 of sodium bicarbonate is 6.35 at 37ºC and the pKa2 is 10.3 at 37ºC, resulting in less than ideal buffering throughout the physiological pH range, and (2) from the above equation it can be seen that carbon dioxide is released into the atmosphere with a resultant increase in solution alkalinity, with the number of produced hydroxyl ions increasing according to the amount of sodium bicarbonate added to the solution. It is, however,

PH2270_C007.fm Page 107 Tuesday, May 23, 2006 10:01 AM

Cell-Culture Media

107

possible to control this reaction by artiﬁcially supplying carbon dioxide to the atmosphere and preventing H2 gas from leaving the liquid, thereby reducing the hydroxyl-ion concentration in the solution.

HANKS’ BALANCED SALT SOLUTION Hanks’ Balanced Salt Solution is intended to equilibrate with air in a closed system that contains a low concentration of sodium bicarbonate.

EARLE’S BALANCED SALT SOLUTION Earle’s Balanced Salt Solution is intended to equilibrate with a gaseous phase containing approximately 5% CO2, and contains a higher concentration of sodium bicarbonate in the solution. It has superior buffering ability because it contains a greater amount of sodium bicarbonate, although it is more difﬁcult to use because it requires a special gaseous mixture of 5% CO2 to 95% air, supplied above the culture medium. If this procedure is not followed, the pH rapidly increases and cell growth is inhibited at normal incubation temperature. Another related method is to use a medium that produces adequate buffering capacity but does not require the 5%-CO2-to-95%-air mixture to be added to the environment above the culture medium. In some cases this can be achieved by using a medium containing Earle’s salts with the sodium bicarbonate concentration reduced to 0.85 g/l.

L-15 MEDIUM An entirely different approach is utilized with L-15 medium, which uses the buffering capacity of free-base amino acids, omits sodium bicarbonate, substitutes galactose for glucose, and adds pyruvate. The pH of this medium is approximately 7.8, which is higher than that of most others. Since there is no production or loss of CO2, the pH will not change further. This medium makes cell growth possible in open-culture vessels regardless of the CO2 of the atmosphere.

HEPES BUFFER The most commonly utilized alternative to bicarbonate is N-2-Hydroxyethylpiperazine-N′2-ethanesulphonic acid (HEPES). This buffer acts as a zwitterion and has proven superior to conventional buffers in comparative biological assays with cellfree preparations. It has many properties that make it an ideal buffer for cell culture, since it does not require an enriched atmosphere to maintain pH. HEPES does not bind divalent cations, is soluble at 2.25 M at 0°C, and is competent as a buffer for growing many different cells and viruses in culture — although it may exhibit toxicity at concentrations greater than 40 mM. Sodium bicarbonate should also be added as a nutritional requirement so that the sodium bicarbonate concentration does not exceed 10 mM (0.85 g/l) with a HEPES concentration of 20 mM. Whenever sodium bicarbonate is used to buffer

PH2270_C007.fm Page 108 Tuesday, May 23, 2006 10:01 AM

108

Producing Biomolecular Substances

TABLE 7.7 HEPES Buffered Media pH at Various Temperatures Temp.ºC

Adjusted pH

Temp.ºC

Adjusted pH

37 30 29 28 26 25 24 23

7.30 7.38 7.40 7.42 7.43 7.44 7.46 7.47

22 21 20 18 17 10 15 5

7.48 7.44 7.50 7.53 7.54 7.55 7.50 7.58

cell-culture media at a concentration of greater than 1.0 g/l, a carbon dioxideenriched atmosphere is required. Note: Since the delta pKa/ºC of the HEPES buffer is -0.014, the pH reading recorded in HEPES buffered medium varies inversely with the medium temperature. Table 7.7 gives pH levels to be expected at various temperatures. HEPES will have a pronounced effect on the ﬁnal pH, and it is necessary to measure its buffering capacity at working temperatures because of the contributions by other buffers (see Table 7.8). HEPES may be steam sterilized and adjusted to desired pH with sodium hydroxide, and concentrations of 10 mM (2.38 g/l) to 25 mM (5.96 g/l) have been employed with no apparent toxicity. Note: Since the delta pKa/ºC of HEPES buffer is -0.014, the pH reading recorded in a HEPES buffered medium will vary inversely with the temperature of the medium. Table 7.8 gives pH levels that should be expected at various temperatures.

TABLE 7.8 HEPES Buffered Media pH at Various Temperatures Temp.ºC

Adjusted pH

Temp.ºC

Adjusted pH

37 30 29 28 27 26 25 24 23

7.30 7.38 7.40 7.41 7.42 7.43 7.44 7.46 7.47

22 21 20 19 18 17 10 15 5

7.48 7.44 7.50 7.52 7.53 7.54 7.55 7.50 7.58

PH2270_C007.fm Page 109 Tuesday, May 23, 2006 10:01 AM

Cell-Culture Media

109

GROWTH FACTORS Cell-culture methods have evolved from early work in embryology in which reproducible cell-growth techniques demonstrated the in vitro continuation of normal cell function. Traditional cell-culture techniques continued to improve with the discovery of the growth-promoting effects of mitogenic proteins (called growth factors because they control the proliferation and survival of normal cells in vitro and stimulate DNA synthesis in microconcentrations). Growth factors are polypeptide molecules that transmit their biological information via speciﬁc membrane receptors. They differ from other biological modiﬁers (e.g., hormones) because they perform by stimulating DNA synthesis. Growth factors have been isolated from a number of biological sources and recombinantly produced in quantity. Growth factors have been identiﬁed in extracts of blood, tissue, and cultured cells from both fetal and adult sources. They have also been identiﬁed as essential regulators of functions ranging from the stimulation of metabolite transport to chemotaxis and prostaglandin synthesis, and they have demonstrated profound stimulatory and inhibitory effects on proliferation and differentiation in a wide variety of cell types. Classical serum-carried hormones (e.g., adrenocorticotropic hormone and calcitonin) are synthesized in the adenohypophysis and thyroid and are released to act on more distant target cells such as those of the adrenal cortex or bone. Growth factors, however, can be produced by many different classes of cells. Furthermore, these factors, although detectable in circulation, appear to act mostly as autocrines and paracrines by causing effects within the cell or in the local cell environment in which they are produced. These microenvironmental functions encompass an elaborate cell-to-cell communication and mediation function that underlies normal growth, differentiation, and the integration of tissues and organ systems in both embryonic and adult organisms. It has been difﬁcult to utilize in vitro paradigms to study the growth effects of classical systemic hormones, since, with the exception of insulin, they do not generally elicit mitogenesis in cultured cells. In contrast, the characteristics of growth factors greatly facilitate the use of cell-culture models and rDNA technology to better understand their complex integrative effects. Although the presence of growth-promoting factors was detected in serum, tissues, and embryonic extracts in the early twentieth century, none of them were isolated and fully characterized until the 1970s.

EPIDERMAL GROWTH FACTOR (EGF) Epidermal growth factor (EGF) is a small peptide found in body ﬂuids. Also known as urogastrone, EGF induces the proliferation of basal skin cells and is highly expressed in A-431 human carcinoma cells. EGF acts by both low- and highafﬁnity binding through a speciﬁc cell surface receptor glycoprotein. EGF receptors have been found in most classes of cells, excluding those of hematopoietic origin. Of ancient evolutionary origin, these receptors have been detected in ﬁsh, amphibians, fruit ﬂies, and mammals. EGF is mitogenic in many epithelial and ﬁbroblast cell cultures.

PH2270_C007.fm Page 110 Tuesday, May 23, 2006 10:01 AM

110

Producing Biomolecular Substances

FIBROBLAST GROWTH FACTORS (FGF) Fibroblast growth factors (FGFs) are the most potent angiogenic agents known. Occurring in two forms, acidic and basic, the FGFs are associated with a small family of heparin-binding growth factors that also include three products of tumorderived oncogenes. Fibroblast growth factors are known by more than twenty synonyms, and incorporate a proliferative response by means of saturable high-afﬁnity receptors in cells of neuroectodermal and mesodermal origin (e.g., endothelial cells, smooth muscle cells, ﬁbroblasts, etc.). Acidic FGF Acidic-FGF-induced mitogenesis is enhanced if heparin is present, although the interaction of basic FGF and heparin is unclear. Basic FGF Basic FGF is common in many classes of cells, while acidic FGF is found only in the central nervous system, the bone matrix, and in osteosarcomas. Like TGF-β, basic FGF can also promote soft agar growth in anchorage-dependent cells. Basic-FGF in vitro cell treatment stabilizes the phenotype across repeated cell passages, essentially delaying the aging process. Paradoxically, reports have established that FGF, like TGF-β, is a potent inhibitor of differentiation (independent of its mitogenic effects).

OTHER GROWTH FACTORS Other growth factors have been extensively characterized, including two insulin-like factors known as IGF-I and IGF-II (single-chain polypeptides that exhibit both endocrine and paracrine/autocrine actions). Although the liver is a major source of these factors, they are synthesized in most tissues of the body. The structure of IGFs and insulin are similar enough to allow cross-reactivity with heterologous receptors such as tyrosine kinase. Both IGFs are known to have complex dependent interactions with circulating levels of growth hormone (GH) and display sulfation factor activity in cartilage. Some growth factors regulate the production of other factors. The expansion of such factors is a complex process depending on the state of cell differentiation and on cell-growth conditions. The growth-factor response to cellular receptors generally involves tyrosine-speciﬁc protein kinases (i.e., they phosphorylate Tyr-OH groups in their target proteins), which can be altered by various stimuli. Cellular responsiveness may be governed by either the cell-differentiation stage or the extracellular matrix on which the cells reside. Cell response to an individual growth factor can be multifunctional, either within an individual cell type or a different developmental lineage.

LECTINS Lectins are a unique class of carbohydrate-binding proteins. Different lectins bind to speciﬁc carbohydrate residues (e.g., jack bean lectin attaches to mannose). Lectins

PH2270_C007.fm Page 111 Tuesday, May 23, 2006 10:01 AM

Cell-Culture Media

111

ﬁnd use in the study of glycoproteins, polysaccharides, cell-surface proteins, and cellular membranes. Lectins also agglutinate certain blood cells, are used to stimulate mitosis in lymphocytes, can be conjugated with enzymes and dyes, and are immobilized on chromatographic supports to purify certain cells and proteins. Lectins are extracted from a variety of sources — soybeans, peanuts, tomatoes, castor beans, wheat germ, jack beans, and horseshoe crabs.

TOXINS Toxins are peptides and polypeptides that act as biological poisons. Commonly produced by certain plants, animals, and microorganisms, they have also been synthesized in the laboratory. Current research focus is on conjugated toxins bound to lectins or antibodies, which can then be carried to a speciﬁc site in a targeted cell or tissue. Some newer biopharmaceutical products involve coupling a cytotoxin to a site-speciﬁc monoclonal antibody molecule, thereby delivering the cytotoxin to the exact biological target (e.g., a malignant tumor or a particular strain of malignant plasma cell). A number of toxins are commercially available including aﬂatoxin, diphtheria toxin, ricin from castor oil beans, snake venoms, and pokeweed extracts. Most toxins are extremely dangerous to produce, purify, and characterize, and pose serious laboratory safety problems.

TRANSPORT FACTORS Transport factors are proteins that enable essential micrometabolytes to be transported to cells in usable form. Since many substances such as lipids and other molecules are only slightly soluble, they are not absorbed very well by cells and can even be toxic in free form. Transport factors help the conveyance of such metabolytes to the cells and are crucial when culturing in a deﬁned medium. The iron-binding protein transferrin, for example, is an extremely useful transport factor in cell-culture work, conveying essential metabolytes to the cells and making iron available to cells in a recognizable form. Cellular iron is used as an enzymatic cofactor in key metabolic pathways such as in the generation of ATP. Transferrin also displays bacteriostatic and fungiostatic properties. Transferrin’s metabolic transport functions make it a key growth promoter in cell-culture work and it is frequently used in serum-free media.

VITAMINS The ﬁrst preparation of a vitamin, an essential food factor, was a concentrate of a potent anti-beriberi substance from rice polishings. Since the active factor was an amine, and was necessary for life, the substance was designated a vita amine. The term has been retained to designate accessory food factors that are neither amino acids nor inorganic elements. Since we now know that not all of these substances are amines, the a’s have been combined and the terminal e has been dropped. The earliest laboratory demonstrations of these food factors were descriptions of a lipid-

PH2270_C007.fm Page 112 Tuesday, May 23, 2006 10:01 AM

112

Producing Biomolecular Substances

soluble essential food factor in butterfat and egg yolk, designated fat-soluble A, and of a heat-labile water-soluble factor in wheat germ, designated water-soluble B.

WATER Distillation and ion-exchange puriﬁcation do not necessarily remove the trace organics that can inhibit cell growth, and water is still a major source of endotoxin contamination in cell culturing. Endotoxin (a lipopolysaccharide) is released from the cell walls of gram-negative bacteria. Approximately 104 bacteria yield 1 ng of lipopolysaccharide (LPS), equal to more than ten endotoxin units. Scientists often do not perceive that their cell cultures are contaminated with LPS, and as a ramiﬁcation, conclusions drawn from these studies are not always valid. Subnanogram quantities of endotoxin can alter cell function. Endotoxins have also been found in lots of viral vaccines. The Limulus amebocyte lysate (LAL) test is currently the approved method for endotoxin testing. A UV sterilizing light can partially control microorganisms in continuous culture systems, although microorganisms can propagate in almost any portion of the water puriﬁcation system. Microorganisms retained on ﬁlters can still reproduce, release endotoxin, and even grow through ﬁlters to contaminate downstream processes.

MANUFACTURERS’ DIRECTORY MEDIA, CELL CULTURE, BASIC, LIQUID Acumedia Manufacturers, Inc. (800) 783-3212 J. Brooks Laboratories, Inc. (800) 682-1095 Celox Corp. (800) 552-3569 Clonetics Corp. (800) 852-5663 Fisher Scientiﬁc (800) 562-1729 Gibco BRL Life Technologies (800) 828-6686 Hyclone Laboratories (800) 492-5663 ICN Biomedicals, Inc. (800) 854-0530 Intergen Co. (800) 431-4505 Irvine Scientiﬁc (800) 437-5706 JRH Biosciences (800) 255-6032 Mediatech, Inc. (800) 235-5476 Sigma Chemical Co. (800) 325-3010 Stratagene (800) 424-5444 Viobin Corp. (608) 849-5944 Whittaker Bioproducts (800) 654-4452

MEDIA, CELL CULTURE, BASIC, POWDERED American Biorganics, Inc. (800) 648-6689 Amicon Inc. (800) 4-AMICON

PH2270_C007.fm Page 113 Tuesday, May 23, 2006 10:01 AM

Cell-Culture Media

Biocell Labs. (800) 222-8382 BIO 101 Inc. (800) 424-6101 J. Brooks Laboratories Inc. (800) 682-1095 Fisher Scientiﬁc (800) 562-1729 Gibco BRL Life Technologies (800) 828-6686 Hyclone Laboratories (800) 492-5663 ICN Biomedicals Inc. (800) 854-0530 Irvine Scientiﬁc (800) 437-5706 JRH Biosciences (800) 255-6032 Mediatech Inc. (800) 235-5476 Quality Reviews Inc. (800) 658-1397 Quest Intl. Inc. (800) 231-6777 Sigma Chemical Co. (800) 325-3010 TA Biochemicals Inc. (716) 685-4390 Viobin Corp. (608) 849-5944 Whittaker Bioproducts (800) 654-4452

MEDIA, CELL CULTURE, CONDITIONED Amicon, Inc. (800) 4-AMICON Biocell Labs. (800) 222-8382 Clonetics Corp. (800) 852-5663 Colorado Serum Co./Western Inst. Co. (303) 295-7527 Fisher Scientiﬁc (800) 562-1729 ICN Biomedicals, Inc. (800) 854-0530 Igen Inc. (800) 336-IGEN Irvine Scientiﬁc (800) 437-5706 Research Organics (800) 321-0570 Sepracor, Inc. (800) 752-5277 Whittaker Bioproducts (800) 654-4452

MEDIA, CULTURE, RECOMBINANT-BACTERIAL Acumedia Manufacturers, Inc. (800) 783-3212 American Biorganics, Inc. (800) 648-6689 Atomergic Chemetals Corp. (516) 694-9000 Becton Dickinson Microbiological Systems (800) 638-8663 BIO 101, Inc. (800) 424-6101 Difco Laboratories, Inc. (800) 521-0851 Lallemand, Inc. (514) 522-2133 Marcor Development (201) 489-5700 PML Microbiologicals (800) 547-0659 Quest Intl., Inc. (800) 833-8308 REMEL (800) 255-6730 Sigma Chemical Co. (800) 325-3010 Viobin Corp. (608) 849-5944

113

PH2270_C007.fm Page 114 Tuesday, May 23, 2006 10:01 AM

114

Producing Biomolecular Substances

MEDIA, SERA Accurate Chemical and Sci. (800) 645-6264 The Binding Site, Inc. (800) 633-4484 Bio/Can Scientiﬁc (416) 828-2455 Biocell Labs (800) 222-8382 Bioreclamation, Inc. (516) 483-1196 P.J. Cobert Assoc. (314) 993-2390 Colorado Serum Co./Western Inst. Co. (303) 295-7527 Fitzgerald Industries International (800) 370-2222 Gen Trak, Inc. (800) 221-7407 Gibco BRL Life Technologies (800) 828-6686 Hilltop Lab Animals, Inc. (800) 245-6921 Hyclone Laboratories (800) 492-5663 Immunocorp (800) 363-8803 Intergen Co. (800) 431-4505 Irvine Scientiﬁc (800) 437-5706 Jackson ImmunoResearch Labs. (800) 367-5296 JRH Biosciences (800) 255-6032 Lampire Biological Laboratories (215) 795-2838 Organon Teknika-Cappel (800) 523-7620 Pel-Freez (800) 643-3426 PML Microbiologicals (800) 547-0659 Rockland, Inc. (215) 369-1008 The Salzman Corp. (800) 553-8903 Sigma Chemical Co. (800) 325-3010 Solomon Park Research Laboratories (206) 821-7005 United States Biochemical Corp. (800) 321-9322 Vector Laboratories, Inc. (800) 227-6666 Whittaker Bioproducts (800) 654-4452

MEDIA, AMINO ACIDS American Biorganics, Inc. (800) 648-6689 Atomergic Chemetals Corp. (516) 694-9000 BACHEM Bioscience (800) 634-3183 Boehringer Mannheim Corp. (317) 845-2000 Difco Laboratories, Inc. (800) 521-0851 Eastern Chemical (800) 645-5566 Fluka Chemical Co. (800) 358-5287 Gibco BRL Life Technologies (800) 828-6686 ICN Biomedicals, Inc. (800) 854-0530 Irvine Scientiﬁc (800) 437-5706 Isotec, Inc. (800) 448-9760 Millipore Corp. (800) 225-1380

PH2270_C007.fm Page 115 Tuesday, May 23, 2006 10:01 AM

Cell-Culture Media

Peninsula Laboratories, Inc. (800) 922-1516 Peptides International (800) 777-4779 Peptide Technologies (202) 234-3384 Pfaltz and Bauer, Inc. (800) 225-5172 Promega Corp. (800) 356-9526 Research Organics (800) 321-0570 Research Plus, Inc. (800) 341-2296 Sigma Chemical Co. (800) 325-3010 Sithean Corp. (804) 346-4230 Spectrum Chemical (800) 772-8786 Synthetech, Inc. (503) 967-6575 TA Biochemicals, Inc. (716) 685-4390 Toronto Res. Chemicals (416) 638-9696

MEDIA, ANTIBIOTICS Accurate Chemical and Sci. (800) 645-6264 AMRESCO, Inc. (800) 829-2805 Bioreclamation, Inc. (516) 483-1196 Boehringer Mannheim Corp. (317) 845-2000 J. Brooks Laboratories, Inc. (800) 682-1095 Carolina Biological Supply Co. (800) 334-5551 Celox Corp. (800) 552-3569 Clonetics Corp. (800) 852-5663 Colorado Serum Co./Western Inst. Co. (303) 295-7527 Fisher Scientiﬁc (800) 562-1729 Gibco BRL Life Technologies (800) 828-6686 Hyclone Laboratories (800) 492-5663 ICN Biomedicals, Inc. (800) 854-0530 Intergen Co. (800) 431-4505 Irvine Scientiﬁc (800) 437-5706 JRH Biosciences (800) 255-6032 Mediatech Inc. (800) 235-5476 Research Organics (800) 321-0570 Sigma Chemical Co. (800) 325-3010 Spectrum Chemical (800) 772-8786 Thomas Scientiﬁc (800) 345-2100 Whittaker Bioproducts (800) 654-4452

MEDIA, CELL-ATTACHMENT FACTORS Accurate Chemical and Sci. (800) 645-6264 Biomedical Technologies (617) 344-9942 Boehringer Mannheim Corp. (317) 845-2000 Cellon Sarl. (352) 495-5985

115

PH2270_C007.fm Page 116 Tuesday, May 23, 2006 10:01 AM

116

Producing Biomolecular Substances

Clonetics Corp. (800) 852-5663 Fisher Scientiﬁc (800) 562-1729 Gibco BRL Life Technologies (800) 828-6686 Intergen Co. (800) 431-4505 JRH Biosciences (800) 255-6032 Mallinckrodt Spec. Chem. (800) 354-2050 Sigma Chemical Co. (800) 325-3010 Telios Pharmaceuticals (619) 622-2650 York Biologicals (516) 751-6553

MEDIA, MATRICES & MICROPARTICLES Bellco Glass, Inc. (800) 257-7043 Biomedical Technologies (617) 344-9942 Bio-Tek Instruments, Inc. (800) 451-5172 JRH Biosciences (800) 255-6032 Millipore Corp. (800) 225-1380 Sigma Chemical Co. (800) 325-3010 Unisyn Technologies, Inc. (800) 735-4035 Verax Corp. (603) 448-4445

MEDIA, BALANCED SALT SOLUTION Accurate Chemical and Sci. (800) 645-6264 American Biorganics, Inc. (800) 648-6689 Biocell Labs (800) 222-8382 J. Brooks Laboratories, Inc. (800) 682-1095 Celox Corp. (800) 552-3569 Clonetics Corp. (800) 852-5663 Colorado Serum Co./Western Inst. Co. (303) 295-7527 Difco Laboratories, Inc. (800) 521-0851 Fisher Scientiﬁc (800) 562-1729 Gibco BRL Life Technologies (800) 828-6686 Hyclone Laboratories (800) 492-5663 ICN Biomedicals, Inc. (800) 854-0530 Irvine Scientiﬁc (800) 437-5706 JRH Biosciences (800) 255-6032 Mediatech, Inc. (800) 235-5416 Sigma Chemical Co. (800) 325-3010 Spectrum Chemical (800) 772-8786 TA Biochemicals, Inc. (716) 685-4390 Thomas Scientiﬁc (800) 345-2100 Whittaker Bioproducts (800) 654-4452

PH2270_C007.fm Page 117 Tuesday, May 23, 2006 10:01 AM

Cell-Culture Media

MEDIA, HEPES BUFFER Accurate Chemical and Sci. (800) 645-6264 Aldrich Chemical Co., Inc. (800) 558-9160 American Biorganics, Inc. (800) 648-6689 American Intl. Chem. (800) 238-0001 AMRESCO, Inc. (800) 829-2805 Atomergic Chemetals Corp. (516) 694-9000 Bio-Rad Laboratories (510) 741-6891 Boehringer Mannheim Corp. (317) 845-2000 Celox Corp. (800) 552-3569 Clonetics Corp. (800) 852-5663 Electron Microscopy Sci. (800) 523-5874 Fisher Scientiﬁc (800) 562-1729 Gallard-Schlesinger (800) 645-3044 Genzyme Corp. (800) 332-1042 Gibco BRL Life Technologies (800) 828-6686 Irvine Scientiﬁc (800) 437-5706 JRH Biosciences (800) 255-6032 J. T. Baker, Inc. (800) 582-2537 Mallinckrodt Spec. Chem. (800) 354-2050 Marcor Development (201) 489-5700 MTM Research Chemicals (800) 238-2324 Research Organics (800) 321-0570 Sigma Chemical Co. (800) 325-3010 Solutions Plus, Inc. (314) 349-4922 Spectrum Chemical (800) 772-8786 TA Biochemicals, Inc. (716) 685-4390 Ted Pella, Inc. (800) 237-3526 Tousimis Research (800) 638-9558 Turner Designs (408) 749-0994 United States Biochemical Corp. (800) 321-9322 Wako Chemicals USA (800) 992-WAKO Whittaker Bioproducts (800) 654-4452 York Biologicals (516) 751-6553

CELL-CULTURE MEDIA, GROWTH FACTORS Accurate Chemical and Sci. (800) 645-6264 American Biorganics, Inc. (800) 648-6689 Amicon, Inc. (800) 4-AMICON Biocell Labs. (800) 222-8382 Boehringer Mannheim Corp. (317) 845-2000

117

PH2270_C007.fm Page 118 Tuesday, May 23, 2006 10:01 AM

118

Producing Biomolecular Substances

Celox Corp. (800) 552-3569 Clonetics Corp. (800) 852-5663 Colorado Serum Co./Western Inst. Co. (303) 295-7527 Fisher Scientiﬁc (800) 562-1729 Gibco BRL Life Technologies (800) 828-6686 Hyclone Laboratories (800) 492-5663 ICN Biomedicals, Inc. (800) 854-0530 Intergen Co. (800) 431-4505 Irvine Scientiﬁc (800) 437-5706 JRH Biosciences (800) 255-6032 Mallinckrodt Spec. Chem. (800) 354-2050 Marcor Development (201) 489-5700 Pierce Chemical Co. (800) 874-3723 Research Diagnostics (201) 584-7093 Sepracor, Inc. (800) 752-5277 Sigma Chemical Co. (800) 325-3010 Thomas Scientiﬁc (800) 345-2100 Wako Chemicals USA (800) 992-WAKO Whittaker Bioproducts (800) 654-4452

CELL-CULTURE MEDIA, LECTINS Accurate Chemical and Sci. (800) 645-6264 American Biorganics, Inc. (800) 648-6689 Atomergic Chemetals Corp. (516) 694-9000 Boehringer Mannheim Corp. (317) 845-2000 P.J. Cobert Assoc. (314) 993-2390 DAKO (800) 235-5743 EY Laboratories, Inc. (800) 821-0044 List Biological Labs, Inc. (800) 726-3213 Pharmacia Biotech, Inc. (800) 526-3593 Pierce Chemical Co. (800) 874-3723 Sigma Chemical Co. (800) 325-3010 United States Biochemical Corp. (800) 321-9322 Vector Laboratories, Inc. (800) 227-6666 Wako Chemicals USA (800) 992-WAKO Worthington Biochemical Corp. (800) 445-9603

CELL-CULTURE MEDIA, TOXINS Accurate Chemical and Sci. (800) 645-6264 BACHEM Bioscience (800) 634-3183 Peptides International (800) 777-4779 Research Biochemicals (800) 736-3690 Sigma Chemical Co. (800) 325-3010 Wako Chemicals USA (800) 992-WAKO

PH2270_C007.fm Page 119 Tuesday, May 23, 2006 10:01 AM

Cell-Culture Media

119

CELL-CULTURE MEDIA, TRANSPORT FACTORS Accurate Chemical and Sci. (800) 645-6264 Biocell Labs. (800) 222-8382 Celox Corp. (800) 552-3569 Clonetics Corp. (800) 852-5663 Diagnostic Chemicals (800) 565-0265 Gen Trak, Inc. (800) 221-7407 Gibco BRL Life Technologies (800) 828-6686 Hyclone Laboratories (800) 492-5663 ICN Biomedicals, Inc. (800) 854-0530 Igen, Inc. (800) 336-IGEN Irvine Scientiﬁc (800) 437-5706 JRH Biosciences (800) 255-6032 Molecular Probes (503) 465-8300 Research Diagnostics (201) 584-7093 Sepracor, Inc. (800) 752-5277 Wako Chemicals USA (800) 992-WAKO

CELL-CULTURE MEDIA, VITAMIN SUPPLEMENTS Accurate Chemical and Sci. (800) 645-6264 American Biorganics, Inc. (800) 648-6689 Gibco BRL Life Technologies (800) 828-6686 ICN Biomedicals, Inc. (800) 854-0530 Irvine Scientiﬁc (800) 437-5706 JRH Biosciences (800) 255-6032 Mediatech, Inc. (800) 235-5476 Sigma Chemical Co. (800) 325-3010 Spectrum Chemical (800) 772-8786 Whittaker Bioproducts (800) 654-4452

CELL-CULTURE MEDIA, WATER

FOR INJECTION

AMSCO Scientiﬁc (800) 444-9009 Aqua-Chem, Inc. (414) 577-2825 Consolidated Stills & Sterilizers (617) 782-6072 Irvine Scientiﬁc (800) 437-5708 IWT (815) 877-3041 Osmonics, Inc. (800) 848-1750 Paul Mueller Co. (800) 641-2830 Regional Scientiﬁc Assoc. (301) 948-9383 Xentex, Inc. (913) 764-3808

PH2270_C007.fm Page 120 Tuesday, May 23, 2006 10:01 AM

PH2270_book.fm Page 121 Tuesday, May 16, 2006 9:37 AM

Fermenters INTRODUCTION TO FERMENTATION

Fermentation utilizes metabolic and enzyme activity of microorganisms to transform organic compounds into different products. With genetic engineering technology, scientists can manipulate genes controlling cellular protein expression to produce structured proteins or other biomolecules. Cell-derived biopharmaceuticals are currently licensed for human use around the world, including interferons, human enzymes, and hormones — all produced through recombinant DNA technology. Ordinary fermentation encompasses processes and products that are not covered in this study. Solid substrate fermentation that produces bread and cheese is not covered here. Fermentation is deﬁned as the process of growing a culture of microorganisms in a nutrient media, thereby converting feed material into desired end products. Sometimes described as a biochemical reaction where microorganisms (bacteria or fungi) act as biocatalysts, fermentation has been around for a long time, beginning with herb doctors, brewers, bakers, cheese makers, and vintners of ancient civilizations who used empirical methods to create their desired products — such as the ancient Egyptians brewing beer and fermenting wine. Mining by microbial leaching dates back to Roman times and has been utilized in Spain since the 1700s.

HISTORICAL PERSPECTIVE 1700S: DAWN

OF THE

SCIENTIFIC APPROACH

FERMENTATION

The dawn of the scientiﬁc approach to fermentation appeared with the quantiﬁcation of ethanol fermentation by Gay-Lussac in the 1700s. By the early 1800s, Pasteur, Kutzing, Schwann, and Cagniard-Latour demonstrated that fermentation was caused by living yeast organisms. Today’s biotechnology industry, run on a mixture of pure research and practical application, really began with Louis Pasteur, who demonstrated that fermentation was caused by living cells, and not by decomposition.

EARLY 1800S: PASTEUR In his early studies, Pasteur worked on the separation of dextro- and levo-tartaric acids, and concluded that only living things could produce such asymmetrical, and optically active substances. He refuted the idea of spontaneous generation, and proved that fermentation was the product of living cells and that “life was necessary to create new life” (Virchow, in 1858, subsequently deﬁned this concept as, omnis cellulae, ex cellulae… — “all cells come from cells”). It took Pasteur twenty years of labo121

PH2270_book.fm Page 122 Tuesday, May 16, 2006 9:37 AM

122

Producing Biomolecular Substances

ratory experimentation and bitter conﬂict with the European scientiﬁc community to ﬁnally convince them of the veracity of his theories. His understanding of microbes, their behavior, and Pasteur’s association of biological activity with molecular structure, was as signiﬁcant to his generation as recombinant DNA concepts are to ours. Pasteur was fascinated by the interaction of practice and theory and became involved in developing a long series of industrial applications. He discovered the cause of milk souring and the fermenting of wine and beer, providing enormous savings for France — and later England and Germany. He identiﬁed the parasite devastating the silk industry, and developed techniques for attenuating viruses in his study of chicken cholera. Pasteur also studied anthrax, demonstrating how to protect sheep from the disease in his classic experiment making use of the scientiﬁc method. His study of rabies resulted in saving the life of a young boy after injecting him with a sequence of attenuated viral material obtained from infected animals.

MID-1800S–1858: VIRCHOW, SCHWANN,

AND

FLEMMING

Rudolf Virchow’s work at Berlin’s Pathological Institute helped establish a basis for cellular biotechnology, as did the earlier work of Schwann and the later work of Strasburger, Flemming, and of Astbury in Great Britain. Gregor Mendel’s midnineteenth century work developing the laws of heredity was rediscovered thirty years later, and subsequently followed by the discovery of chromosomes and, later on, genes. The landmark DNA studies of Crick and Watson set the stage for today’s biotechnology explosion, spurred on by the commercial exploitation of industrial fermentation.

1916–1918 WWI: IMPETUS FERMENTATION

FOR THE

DEVELOPMENT

OF INDUSTRIAL

World War I gave a major impetus to the development of industrial fermentation. Food shortages led to large-scale production of brewers’ yeast and glycerol in Germany, and bulk chemical shortages led to Great Britain’s large-scale production of acetone and glycerol. As an industry, fermentation actually commenced in the early part of the twentieth century. Bacteria, yeast, and ﬁlamentous fungi were ideal organisms for research, but also for producing a wide variety of products given the numerous types and quantities of biological reactions.

THE ROARING TWENTIES: FERMENTATION SPREADS; MOST SOLVENTS ARE FERMENTATION PRODUCTS By 1925, most manufactured solvents were fermentation products. Fermentation, on the other hand, was beginning to be used on a larger scale by the pharmaceutical industry. In 1944, the mass production of penicillin for the Normandy landings was regarded as a deﬁning moment in pharmaceutical manufacture. Even so, by the 1940s a growing sophistication in synthetic organic chemistry, coupled with readily available and inexpensive feed stocks, led to the supplanting of fermentation by synthetic production.

PH2270_book.fm Page 123 Tuesday, May 16, 2006 9:37 AM

Fermenters

123

TABLE 8.1 Major Industrial Fermentation Products • • • • • •

Food and beverages Antibiotics Vitamins Amino acids Organic acids Solvents

• • • • • •

Alcohols Carbohydrates Enzymes Single cell proteins Biological pesticides Miscellaneous

THE 1930S AND 1940S: SYNTHETIC CHEMICAL PROCESSES REPLACE FERMENTATION Abundant supplies of raw materials, the exploitation of petroleum, and the technological advances associated with World War II led to the replacement of fermentation by synthetic chemical processing.

1941–1949 WWII: FERMENTATION USED PHARMACEUTICALS

FOR

PENICILLIN

AND

By the 1940s, bacteria, yeast, and ﬁlamentous fungi were used to produce a wide variety of products, with numerous types and quantities of biological reactions. Fermentation was being used on a larger scale by the pharmaceutical industry, and the mass production of penicillin in 1944 for the Normandy landings was regarded as a deﬁning moment for the pharmaceutical industry.

THE 1950S: FERMENTATION

AS A

DISCRETE SCIENCE

By the 1950s only 1% of industrial solvents used in the U.S. were fermentation products. The intensiﬁed research into the fundamental processes of life that occurred during the ’50s and into the ’60s, gave new life to fermentation as a discrete science. Microbes were designed to produce bioactive proteins such as human insulin, enzymes, vitamins, amino acids, hormones, and speciﬁc antibiotics at industrial scale. Table 8.1 lists some industrial fermentation products. Metabolic functions of fungal and microbial organisms are capable of producing a wide variety of products. In the food and beverage industries, fermentation changes grape juice into wine, milk into yogurt and cheese, cabbage into sauerkraut, soybeans into soy sauce, and, in recent applications, produces glycerol, ethanol, and amino, citric, and lactic acids, plus a myriad of other materials for the food and chemical industries.

THE 1960S: MICROBIAL METAL MINING REVIVED; DESIGNER MICROBES ASSEMBLE PROTEINS Special microbes with an afﬁnity for metals are used to mine low-grade ores for uranium, copper, and other metals, and to remove or recover metals from industrial efﬂuents. In the United States, the Ofﬁce of Technology Assessment (OTA) estimated

PH2270_book.fm Page 124 Tuesday, May 16, 2006 9:37 AM

124

Producing Biomolecular Substances

that by the beginning of the 1990s, copper produced by microbial leaching would be almost 18% of total production. Since the early 1960s Canada has utilized microbial uranium mining. Other fermentation applications include removing sulfur from high-sulfur coal, renovating low-yield oil reserves, and using microbes to degrade plastic waste. Thus, the cultivation of microbes for these industries required suitably designed fermenters.

THE 1970S: FERMENTATION

FOR

BIOPHARMACEUTICAL MANUFACTURING

By the 1970s, the biopharmaceutical industry uses fermentation for manufacturing hormones, biologicals, and biopharmaceuticals. Microorganisms convert waste and raw materials into energy and/or industrial chemicals, or degraded waste into either recyclables or harmless materials — a major environmental solution. Special microbes are also used to enhance oil recovery.

THE MID-1970S: CREATING CHIMERAS FRAGMENTS INTO PLASMIDS

BY INSERTING

FOREIGN DNA

By 1973, the insertion of foreign DNA fragments into plasmids to create chimeras, and the subsequent functional reinsertion of these plasmids into E. coli, requires the development of improved containment vessels and process controllers employing computer technology. Today, fermentation is a key factor in biotechnology. It is notable that some of the products listed in Table 8.1 are produced in units deﬁned by size, such as research- or pilot-scale fermenters. Most commercially important fermentations are aerobic (requiring oxygen), but anaerobic fermentations are increasing in importance.

MID-1980S: FERMENTATION

FOR

PRODUCING BIOWEAPONS

Much detailed information on Iraq’s bioweapons program is now unclassiﬁed and illustrates many of the technical facets of using fermentation for the manufacture of bio-weapons of mass destruction (bio-WMDs). Utilizing a comprehensive range of delivery agents and munitions in the bioweapons program including weaponized anthrax, botulinum, and ricin toxins, and incapacitating agents such as aflatoxin, mycotoxin, hemorrhagic conjunctivitis virus, ebola virus, and rotavirus, the scope of Iraqi bioweapons agents included antipersonnel and antifacility weapons, which employed a wide variety of delivery systems, from 122mm rockets and artillery shells to aerial bombs and missile warheads ﬁlled with anthrax, botulinum toxin and aﬂatoxin, ranging to “economic-type” weapons such as wheat cover smut. Given Iraqi claims that only 5 years elapsed from the program’s inception in 1985, to its conclusion at the beginning of the Gulf War, its achievements were quite remarkable, including the production and weaponization of large quantities of bacterial agents, aﬂatoxin, and applications research on a variety of other biological agents. The bioweapons program was production-oriented, and in addition to laboratoryscale equipment, it utilized a pilot plant with a 150-liter fermenter. After bacterial strains were received from overseas, initial work focused on literature study. Then subsequent research and development concentrated on characterizing bacillus anthra-

PH2270_book.fm Page 125 Tuesday, May 16, 2006 9:37 AM

Fermenters

125

cis (for anthrax disease) and clostridium botulinum (for botulinum toxin) to establish their pathogenicity, growth, sporulation conditions, and storage parameters. Anthrax, an acute bacterial disease of both animals and humans, is acquired by ingestion or inhalation of the bacterial spores, or contact with them through skin lesions. Untreated, it produces fatal infections. Botulinum toxin, on the other hand, produces acute muscular paralysis resulting in death. In the beginning there was no actual production of bioweapon agents, and imported fermenters were not used, however, since the program looked ahead to production, a Ministry of Defense report was drafted recommending that a single-cell protein plant be acquired for producing Botulinum. The Iraqi government agreed, and the program was subsequently transferred. After a slow beginning, equipment, including large fermenters, was also transferred to the facility. New equipment was acquired, and new staff added to the bioweapons group. Research then shifted to application of the biological agents as WMDs. They studied the agents’ effects on larger animals in the laboratory, in inhalation chambers, and in the ﬁeld. After initial ﬁeld trials, production scale-up was initiated for botulinum toxin and anthrax. Botulinum toxin production commenced with a 950-liter fermenter in ﬂasks and in a small bench-top laboratory fermenter. Initial anthrax batches employed 7- and 19-liter bench-top fermenters. The 150-liter fermenter was used to produce bacillus subtilis, a simulant for anthrax as a bioweapons agent. After ﬁve or six runs of subtilis, production of anthrax was inaugurated. About 15–16 production runs were completed, producing 1,500 liters of the anthrax toxin, which was then concentrated to 150 liters. Additional production with laboratory fermenters was also completed. A report on the Bioweapons Program successes was then submitted, resulting in a decision to enter full-scale production. A new site for production was selected and given the designator “329.” This plant’s design philosophy was taken from that of their chemical weapons production facility (i.e., the buildings were well separated, research areas were segregated from production areas, and architectural features of the buildings were copied where appropriate). New facility plans envisioned both production R&D and weapons agent storage, but not munitions ﬁlling. Subsequently a search for production equipment for the program was conducted and two 1,850-liter and seven 1,980-liter fermenters were located in Iraq. The 950-liter fermenter line used for producing botulinum at Taji was transferred, and some larger fermenters were also procured from abroad. After Iraq concluded a contract for a 5,000-litre fermenter, however, their export license was denied. Much of the anthrax fermentation capacity was initially used for producing the anthrax simulant for bioweapon ﬁeld trials. Anthrax production totaled 8,925 liters, and about 6,000 liters of concentrated botulinum were produced. From the program’s inception, there was interest in other bioweapons agents than anthrax and botulinum, and it became Iraqi policy to expand the program into other areas. Thus, there were plans for program diversiﬁcation, which included research facilities for viruses and genetic engineering. In addition to anthrax and botulinum, a new agent, Clostridium perfringens, which produces gas gangrene, common in war casualties and typically resulting in amputation of infected limbs, was added to the research. Generally considered nonlethal in humans, but of serious medical concern because of its carcinogenic activity, the work on perfringens was initiated with aflatoxin, a toxin commonly associated with fungally contaminated food grains and known for pro-

PH2270_book.fm Page 126 Tuesday, May 16, 2006 9:37 AM

126

Producing Biomolecular Substances

ducing liver cancer. Subsequently, research was also conducted on trichothecene mycotoxins, which produce nausea, vomiting, diarrhea and skin irritation but, unlike most microbial toxins, can be absorbed through the skin. Research was also conducted into the effects of aflatoxin as a bioweapons agent when combined with chemicals. The aflatoxin was produced in 5-liter ﬂasks, and a total of 1,850 liters was produced. Another fungal agent investigated for weapons potential was wheat cover smut, which produces a black growth so that the contaminated grain cannot be used as food. Following small-scale production, production was scaled-up and a considerable quantity of contaminated grain was harvested; however this idea was not further developed and the contaminated grain was destroyed. Another toxin considered was ricin, a protein derived from the castor bean and highly lethal to both humans and animals. When inhaled, ricin produces lung tissue breakdown, resulting in hemorrhagic pneumonia and death. Using artillery shells for delivery, a ricin weapons trial was a failure, and the project was scrapped. The investigation of viruses as weapons was begun and then taken over by the Iraqi government for weapons production. It was decided that within the overall bioweapons program, there would be facilities for bacteriology, virology, and genetic engineering. Three viral agents required for the weapons program were obtained in Iraq: Hemorrhagic conjunctivitis virus (an acute disease that causes extreme pain and temporary blindness), Rotavirus (a disease that causes acute diarrhea and can lead to dehydration and death), and Camel-Pox virus (which causes fever and skin rash in camels and sometimes infection in humans), but were not produced in any appreciable quantity. Efforts were then concentrated on weaponizing both the biological and toxic agents. Further anthracis with subtilis simulant, botulinum, and aflatoxin trials were conducted employing 122mm rockets as delivery systems and were considered a success. Live ﬁrings of 122mm rockets ﬁlled with those agents and trials of R900 aerial bombs with the subtilis simulant were performed, and ﬁnal R900 aerial bomb trials using the subtilis simulant, botulinum, and aflatoxin then followed. After Iraq’s invasion of Kuwait, the program was drastically intensiﬁed, with emphasis shifting to production and weaponization. Six fermenters with ancillary equipment were used for botulinum and 5,900 liters of concentrated toxin was produced. It was then decided that they needed more anthrax, and the fermenters used for producing botulinum were then modiﬁed for increased anthrax production. Perfringens production began using the 150-liter fermenter, and a total of 390 liters of concentrated perfringens was produced. In the mid-1980s Iraq produced 19,000 liters of concentrated botulinum toxin using fermentation technology, with nearly 10,000 liters loaded into munitions, 8,500 liters of concentrated anthrax with 6,500 liters loaded into munitions, and 2,200 liters concentrated aflatoxin with 1,580 liters loaded into munitions. Iraq stated that deactivation procedures were used for the stored bulk bacterial agents, where the detoxiﬁed liquid was emptied into a facility septic tank and eventually dumped and buried at the test site. Accordingly, Iraq claimed that 8,000 liters of concentrated botulinum, more than 2,000 liters of concentrated anthrax, 390 liters of concentrated perfringens, as well as an unspeciﬁed quantity of aflatoxin, were subsequently destroyed. Although Iraq stated that they destroyed all their bioweapons and biomaterials, they later changed their story, claiming they could not identify the exact location of the destruction.

PH2270_book.fm Page 127 Tuesday, May 16, 2006 9:37 AM

Fermenters

127

In April 1995, Kurdish forces seized several specialized tractor-trailers near Mosul that were subsequently turned over to the U.S. Military. The design, equipment, and layout of these trailers was, according to a CIA report, “strikingly similar to descriptions” provided by an Iraqi chemical engineer who had managed a mobile plant used for the production of biological weapons. A second mobile facility was discovered by U.S. forces in Mosul. The report further stated, “Although the second trailer appeared to have been looted, the remaining equipment, including a fermenter, was in a conﬁguration similar to that of the ﬁrst mobile plant.” In addition, U.S. forces found second- or possibly third-generation designs of fermentation plants, which included system improvements such as cooling units, “apparently engineered to solve production problems encountered with older designs.” Manufacturer I.D. plates on the fermenters exhibited manufacturing dates of 2002 and 2003, indicating that Iraq produced these units as late as the year they were recovered. An investigation into what other industrial processes might require a fermenter, refrigeration unit, and gas capture system concurred with the accepted expert view that “bioweapons agent production is the only consistent, logical purpose for the use of these mobile plants.”

THE FERMENTATION PROCESS EXPRESSION SYSTEMS Recent advances enabled various methods of controlling cell expression to produce recombinant proteins. Since many genes are expressed in multiple systems, the system with the greatest applicability for producing a particular recombinant protein must be ascertained. As expression system manipulation, fermentation technology, and downstream processing techniques advance, selecting genes that have the best applicability for producing a particular end product must also change. Two key factors affecting this choice are: (1) the quantity of the target protein necessary, and (2) the target protein’s structural complexity. Additional key factors that must be determined include the ﬁnal product’s estimated market in volume and dollars, whether or not speciﬁc modiﬁcations are required to preserve the end product biological activity, and whether or not the end product is chemically stable. Therefore, an optimal expression system would be one that yields the maximum quantity of properly folded, bioactive protein. Nevertheless, there are situations in which the system expressing the highest level of target protein is not the system that produces the most bioactive or properly folded product, so the optimal performance of an expression system should sometimes be weighed against its maximum production capability. Unfortunately, there is no optimal system for expressing and commercially producing all recombinant proteins, so each recombinant product presents its own unique challenge. A detailed discussion of recombinant DNA methods and expression vectors utilized for each heterologous expression system is beyond the scope of this work; however, the subject has been brieﬂy reviewed in Chapters 4 and 10, and is treated at length in some of the references appearing at the end of this volume. Specialized vectors enabling efﬁcient introduction of heterologous genetic material into the host cell have been developed for each expression system. The heterologous gene is

PH2270_book.fm Page 128 Tuesday, May 16, 2006 9:37 AM

128

Producing Biomolecular Substances

integrated into the host genome in either single or multiple copies, or located in an extrachromosomal DNA fragment (plasmid) that autonomously replicates after incorporation into a cell and producing many genetic copies. The gram-negative bacterium E. coli is the best characterized, most easily grown, and most frequently used microorganism for industrially producing recombinant protein. E. coli’s superior growth characteristics are well documented — a colony can double in about 20–30 minutes when grown in enriched medium, and one bacterium is capable of generating almost 150 grams of dry-cell-weight per liter in a deﬁned medium. Yeasts such as S. cerevisiae offer advantages for large-scale recombinant protein production. Unlike E. coli and other bacteria, yeasts lack endotoxins and are therefore considered ultrasafe for the production of food, health care, and pharmaceutical products. Yeasts maintain a well-established fermentation technology and extensively characterized genetics, and culturing them is simple, cost-effective, and rapid, with populations doubling in around 90 minutes when grown in enriched medium containing glucose as a carbon source. S. cerevisiae, for example, grows on minimal medium and can exploit a variety of non-glucose carbon sources. In addition to S. cerevisiae, pichia pastoris is also used for production since its genetic protein expression is tightly regulated by methanol — suggesting pastoris produces simple and cost-effective large-scale fermentation. Although a heterologous protein may be toxic to the host cell, selection against heterologous gene-containing cells can be reduced by growing P. pastoris to initially high densities on carbon sources, such as glucose or glycerol, that repress heterologous gene expression when this is regulated by a methanol-induced promoter. Subsequently switching to methanol as carbon source initiates product expression. Inducible promoters are also often used in E. coli and S. cerevisiae expression systems. Pastoris strains select for efﬁcient growth with methanol in a deﬁned minimal medium at high cell density. From a process standpoint, then, these strains appear to be ideal for hosting heterologous gene expression, producing concentrations of 130 grams dry-cell-weight per liter for continuous culture fermentation. Since commercial products must typically be produced in large scale to be cost effective, the protein expression level in a host system is generally a major factor for choosing a heterologous expression system. Historically, bacterial expression systems have been extensively employed, and decades of use have led to the accrual of a large body of relevant literature. To protect heterologous proteins from cellular proteases, bacterial cells frequently concentrate these proteins in the cell as inclusion bodies, in which they accumulate as insoluble complexes. High E. coli expression levels probably result from this protection, and levels as high as 30% total cell protein have been reported. Nevertheless, although produced in large quantities in bacteria, recombinant proteins may not be properly folded, consequently reducing the amount of recoverable bioactive product. Regardless, a large number of mammalian gene products have been synthesized in E. coli. Although structured proteins have successfully been produced in bacteria, protein levels are typically much lower than those produced through intracellular expression, and the product is trapped in the periplasmic space between the plasma membrane and cell wall. Bacterial expression systems have been developed with a fusion protein approach whereby a synthetic fragment of staphylococcal protein-A, typically

PH2270_book.fm Page 129 Tuesday, May 16, 2006 9:37 AM

Fermenters

129

expressed by gram-positive bacteria, is linked with heterologous genes directing more than 80% of the fusion product expressed into the bacterial medium. Thus, in deﬁned medium with optimized fermentation, the protein yield, consisting of a staphylococcal protein-A fragment bound to a recombinant cytokine, has been reported at nearly 1 gram per liter. The fusion protein is then recovered using IgG afﬁnity chromatography, and the product subsequently separated from the protein-A fragment by chemical cleavage. Compared to bacteria, yeasts synthesize heterologous proteins at much lower levels. Nevertheless, since an end product with biological activity is a primary consideration, and yeasts typically produce biologically active proteins, yeasts have been utilized by Europeans to a greater extent than bacteria for recombinant production. Superoxide dismutase (SOD) is produced in S. cerevisiae at levels of 25–30% total cellular protein, and, when grown to high cell-densities, P. pastoris can synthesize hTNF at levels approaching 30–35% soluble protein.

EXPRESSION MODES There are two principal expression modes: (1) intracellular, where the heterologous protein accumulates within the host cell’s cytoplasm either as a soluble protein or an insoluble aggregate; and (2) extracellular, wherein the heterologous protein’s genetic sequence is manipulated so that the protein is expressed into the culture medium. Selecting either mode affects the level of expression, end-product recovery, and protein puriﬁcation. Expression systems also govern the biomolecular nature of the ﬁnal product, inﬂuence protein folding, disulﬁde bond formation, and posttranslational modiﬁcation. High-synthesis levels are typical of intracellular expression systems, particularly in E. coli, and high expression of a heterologous protein in the cytoplasm of E. coli often results in the compartmentalizing of the protein into inclusion bodies. Accordingly, isolating inclusion bodies from bacteria by centrifugation is a signiﬁcant ﬁrst step in purifying a recombinant protein. Inclusion bodies are not typically found in mammalian cell expression systems, although insect expression systems exhibit bodies of aggregated protein when large quantities of a viral nucleoprotein is expressed. Despite this, all expression systems exhibit disadvantages associated with intracellular protein expression. Protein synthesis begins with adding a methionine to the NH2-terminus of the protein. Since target proteins are typically processed from larger precursors, they have speciﬁc NH2-termini other than methionine. And, in view of the fact that special cellular mechanisms are required to eliminate the methionine from rudimentary intracellular proteins, there is the possibility that intracellularly expressed heterologous protein may have an incorrect NH2-terminus. In both prokaryotic and eukaryotic cells, enzymes used to remove these NH2-terminal methionines, i.e., methionine aminopeptidases, have the same general speciﬁcity. Thus, over-expression of the protein can result in an incomplete processing of the NH2-terminal methionine, yielding some improperly processed protein in the endproduct mixture. Consequently, if the native protein form has a speciﬁc NH2-terminus other than methionine, it is necessary to determine whether or not the variant containing the NH2-terminal methionine retains all desired biological properties and does not exhibit increased or altered immunogeneity.

PH2270_book.fm Page 130 Tuesday, May 16, 2006 9:37 AM

130

Producing Biomolecular Substances

Other approaches have been employed to avoid NH2-terminii heterogeneity: (1) the native protein’s genetic sequence is combined with that of a polypeptide. The target recombinant protein can then be separated from the complex either by chemical or endopeptidase cleavage; (2) enzymatic treatment is utilized to process polypeptide hormones, which are frequently expressed endogenously since longer precursor proteins subsequently encounter posttranslational processing. For example, an hGH precursor produced in E. coli at 20% total cell protein was converted by DAP I into authentic human growth hormone (hGH). Intracellular expression might not yield easily isolated, soluble, and properly folded bioactive products since the host cell cytoplasm-reducing conditions might be quite different from those encountered by the native protein in its indigenous environment, causing the protein to either aggregate or improperly fold. When this happens, the isolated protein requires additional processing to resolubilize and refold it. Adding downstream processing adds cost to the puriﬁcation strategy and scale up. Extracellular expression in this situation offers some advantages: (1) An NH2terminus corresponding to that of the native protein can be produced, since the heterologous protein’s NH2-terminal signal peptide is removed during expression. Nevertheless, incorrect or incomplete signal peptide processing occasionally occurs. (2) If the heterologous protein is an expressed cellular protein, extracellular expression in the heterologous host will typically yield a properly folded product with correct disulﬁde bonds. For example, mapping a peptide via a tryptic digest demonstrates that recombinant interferon secreted by S. cerevisiae has the same disulﬁde bond structure as the natural human protein. (3) The secretion pathway of the eukaryotic system is a means to obtain correctly glycosylated end product, since the expression process protects them from degradation by intracellular proteases. Recombinant expression into the medium also simpliﬁes downstream puriﬁcation since it removes the target products from contaminating intracellular proteins. The host’s protein expression into the medium can be immobilized and used in high cell-density production, although optimal conditions for protein expression must be maintained and medium conditions adjusted to preserve protein’s activity. There are also some disadvantages associated with extracellular expression. Production levels may be less than those obtained with a comparable intracellular system, and in E. coli the heterologous protein is often trapped in the periplasmic space and must be released by methods suitable for scale up. Posttranslational modiﬁcations, such as acetylation, phosphorylation, acylation, and carboxylation also ultimately affect choosing an expression system. Thus, each modiﬁcation must be evaluated to determine if it is speciﬁcally required for the biological activity or stability of the end product. hSOD typiﬁes the way in which posttranslational modiﬁcations determine an expression system. hSOD, an N-acetylated enzyme that prevents oxidation by scavenging superoxide radicals, is sometimes expressed at high levels in both bacteria and yeasts. E. coli does not normally acetylate the N-termini of its proteins; therefore hSOD is expressed in the unacetylated form with the N-terminal methionine removed. In contrast, the hSOD expressed in S. cerevisiae is N-acetylated and therefore identical to that of the hSOD expressed in human erythrocytes. Therefore, for pharmaceutical purposes, it is preferable to utilize the yeast product that is identical to the hSOD, thereby eliminating an adverse immune response. The spec-

PH2270_book.fm Page 131 Tuesday, May 16, 2006 9:37 AM

Fermenters

131

iﬁcity of both the yeast and human acetylation enzymes are characteristically the same, and yeast is sometimes a useful alternative host for producing a heterologous protein that is N-acetylated. Phosphorylation is another example of a host-cell-speciﬁc posttranslational modiﬁcation. E. coli expresses the human c-myc protein, a cellular hom*olog of the myc oncogene, yielding a 60–69 kDa protein. Apparently, there are no modiﬁcations in E. coli-expressed proteins, since they comigrate with proteins derived from an in vitro coupled transcription-translation system. The myc proteins expressed in S. cervisiae are modiﬁed by phosphorylation, however, and are not as extensively modiﬁed as c-myc phosphoproteins synthesized with the baculovirus insect-cell system; despite this, the yeast’s inducted phosphate-free gene expression limits this modiﬁcation. Thus, if large quantities of end product are required, and if the target is unmodiﬁed with few disulﬁde bonds, bacterial and yeast expression systems should be considered ﬁrst. Therefore, bacterial and yeast expression systems can offer some important advantages over insect or mammalian expression systems, since they require less time to develop transformed product-expressing cell lines, and produce typically higher levels of crude product. In addition, the cells have simpler growth and culture requirements with well-established fermentation strategies. Although microbial systems can typically produce large quantities of crude material, they are ill-adapted to produce properly folded end products. Microbial systems cannot be used if the end product requires a particular modiﬁcation that is processed by insect or mammalian expression systems or, if the prokaryotically derived product is biologically inactive. Thus, if in-house expertise is available, simultaneous investigations of various expression systems can typically speed initial product development. Aseptic conditions are critical for bioprocessing since foreign substances and foreign organisms, along with their metabolic byproducts, contaminate cell cultures and complicate downstream puriﬁcation. To avoid contamination, the fermentation system should be made of noncorrosive and nontoxic material that can be repeatedly sterilized. Smaller fermenters are usually lined with or constructed of borosilicate glass, while pilot and production units are typically constructed of #310 stainless steel. And while it is not difﬁcult to produce milligram amounts of product from recombinant bacterial cell lines in a laboratory, it is quite another thing to generate the gram and kilogram amounts of cell-produced products needed for biopharmaceutical applications. Using standard laboratory techniques to manufacture such quantities would require thousands of ﬂasks, hundreds of liters of media, and would present purity, growth, and contamination problems. Various proprietary systems designed to meet high-volume commercial production demands were developed to grow cultures to high cell density and effectively produce multikilogram amounts of biomolecular end product in a pure and active form. Cost considerations are vital when evaluating expression systems. Unfortunately, actual examples of detailed industrial cost analyses are not available because of numerous legal safeguards employed by biopharm manufacturers to protect their proprietary products and processes. Manufacturing cost, capital investment requirements, and return on investment are typically calculated with basic engineering and accounting principles, company production and market research experience, and/or

PH2270_book.fm Page 132 Tuesday, May 16, 2006 9:37 AM

132

Producing Biomolecular Substances

all of those combined. A thorough cost analysis should be completed early in the product development cycle, covering all available options for product expression, synthesis, and large-scale production. This is especially important in situations where an expression system has not been proven at production levels, or when bioprocesses or product recovery strategies differ signiﬁcantly from established procedures. Initial values that should be evaluated to determine capital investment and associated operating costs are the market value in dollars per unit or unit dose, potential market size, share in units necessary to yield the anticipated production volume, and the manufacturing cost as a function of production efﬁciency, protein produced per unit volume, and production cost per unit volume. Production cost is further broken down into manufacturing cost and required capital investment, a function of both the scale of the manufacturing plant and equipment cost. Choosing an appropriate expression system can therefore dramatically reduce a product’s manufacturing cost, despite downstream processing being a major expense. For example, production cost is affected by operating expenses such as medium and consumables costs necessary to promote expression of the protein product, and choosing an expression system which requires more complex and expensive media will typically amplify production costs., Mammalian and insect cells often require additional media supplements. Bioprocess length, downtime between runs, and potential lost production time as a result of contamination are all important in determining costs when comparing production systems, and labor costs still represent a signiﬁcant proportion of production overhead. Expression systems with high production efﬁciencies are generally preferred because a smaller number of batches is necessary to produce a speciﬁc product volume. Expression systems also determine the design and complexity of the fermenter and its associated equipment. Cell concentration (expressed as dry cell weight/liter for a microbial fermentation), total intracellular protein concentration (expressed as percentage of total cell protein), and fermentation time must be considered when estimating intracellular product manufacturing efﬁciency. In contrast, extracellular efﬁciency is evaluated by protein expression level per unit volume, total extracellular protein concentration, and fermentation time. If the downstream processes necessary to obtain pure, bioactive product are inefﬁcient and/or costly, an alternative production system should be selected. Two signiﬁcant values necessary for conducting a downstream product recovery and puriﬁcation cost analysis are: (1) the level of purity required in the ﬁnal product, and (2) the necessary degree of purity of the starting material. Puriﬁcation is affected by variables including starting material source (is the protein deposited in intracellular inclusion bodies or is it secreted into the culture medium?), cost, complexity, number of puriﬁcation steps, protein refolding efﬁciency, and the length of time required for puriﬁcation. In evaluating a production system, the initial step should be to deﬁne the product’s critical requirements, such as properties of the protein and its potential market size. Nevertheless, if the native material is in limited supply, information regarding its properties may be unavailable at project initiation. For example, it may not be known whether or not the protein’s oligosaccharides play a role in the product’s in vivo bioactivity or if the unfolded puriﬁed protein can be subsequently refolded. Likewise, the optimal treatment duration and,

PH2270_book.fm Page 133 Tuesday, May 16, 2006 9:37 AM

Fermenters

133

thus, market size, may not be determined accurately until after the completion of phases I–III clinical trials.

ENZYME FERMENTATION IN THE ORGANIC PHASE Useful synthetic products can be produced at sufﬁciently high yield by hydrolytic enzymes if the reaction equilibrium is shifted sufﬁciently toward synthesis in an organic solvent. A further increased yield can be obtained if the water produced during hydrolytic reactions is continuously removed to control the water’s hydrolytic activity. Salt hydrates and saturated salt solutions are among a number of different techniques evaluated to remove the water generated by the enzyme reaction. While prior work demonstrated some successful techniques for removing water, there had not as yet been a design that sufﬁciently controlled water activity during enzymatic reactions — until a twin-core packed-bed fermenter and a packed-bed hollow-ﬁber unit incorporating salt-hydrate pairs and salt solutions, respectively, were reported. The novel twin-core packed-bed fermenter, which consisted of a removable inner core of salt hydrate separated from an outer core of lipase immobilized on a polypropylene support, was reported as successfully enabling separation of the inner salthydrate core from the enzymic core, through which the substrate mixture was pumped, allowing for recovery and reuse of both the enzyme and the salt hydrate. Complete esteriﬁcation was accomplished with this design. Previously, this was not achievable without a salt hydrate in fermenters. In this fermenter, employing a saturated salt solution, the enzyme and solution were separated with a membrane. The enzyme was immobilized on a microporous polypropylene matrix and placed on the shell side of the membrane while the salt solution was circulated on the lumen side. Diluted by water formed by the enzymatic reaction, the salt solution was resaturated after passing through the salt bed. Complete esteriﬁcation with controlled water activity was characteristic for this system; this reaction was previously not achievable in fermenters without water activity controls. Another example of a two-phase enzyme fermenter utilizes hollow-ﬁber cartridges to separate the organic from the aqueous phases. Since there is a low aqueous solubility, substrates are dissolved in the organic phase. On the side in contact with the organic phase, the enzyme is immobilized by entrapping it in the hollow ﬁbers that maintain separation with a slightly positive pressure, preventing the aqueous phase from penetrating the membrane into the organic phase. Such units were used for interesterifying triglycerides and fatty acids, and for producing optically active 2R- and 3S-(3-9-methoxyphenyl) glycidic acid methyl esters. In the ester system, the membrane was joined to a crystallizer in order to recover optically pure product. These conﬁgurations have the advantages of enzyme reusability, longer stability, and ease of reloading the enzyme when activity declines. Some newer enzyme fermenter designs include a packed bed where feed is introduced as a square wave using an elastic membrane pulsator. When compared to a nonpulsed fermenter, the unit enables Saccharomyces cerevisiae to increase the production of ethanol by 18%, depending upon the frequency of the pulses and the overall hydraulic residence time. This system’s increased performance is probably caused by better degassing, less

PH2270_book.fm Page 134 Tuesday, May 16, 2006 9:37 AM

134

Producing Biomolecular Substances

back-mixing, and improved mass transfer. A second development was the continuously-aerated plug-ﬂow fermenter. The unit is assembled from a rotating stainless steel spiral screw in a stainless steel container partially ﬁlled with culture broth medium. As the screw rotates, the culture moves along the length of the vessel, where mixing is achieved by aeration, and true plug-ﬂow is achieved where limited mixing occurs between the two adjacent loops. A possible disadvantage here is that the unit functions like a series of batch reactors, so inoculation is required for each new loop. Large-scale operation and scale up, however, has yet to be successfully demonstrated. Attempts at fermenter designs that address existing limitations and new processing avenues still require improvement and signiﬁcantly better performance over current designs. Furthermore, developments in fermenter design remain of high priority, since single continuously stirred tank fermenters (CSTFs) do not provide a universal process solution. Fermenter design improvements require understanding bioprocess limitations, so that more rational, creative, and focused approaches can be developed.

CELL GROWTH AND PRODUCTION KINETIC GROWTH Understanding cell growth is necessary for better fermenter design and improved operation. Cell-production kinetics involves the cell growth rate and how it is affected by various chemical and physical conditions. Cell kinetics is the consequential interaction of numerous complicated biochemical reactions and transport phenomena, involving multiple stages of multicomponent systems. During growth, a heterogeneous mixture of young and old cells is continuously transforming and adapting to a changing environment. As a result, the completely accurate mathematical modeling the system’s growth kinetics is virtually impossible. Thus, in order to derive simpler operating and performance models that can accommodate mathematical representation, assumptions must be made regarding various cellular components and cell population dynamics, as presented in Table 8.2. The simplest operation and performance fermenter model is the unstructured, distributed model, based upon two assumptions: (1) cells can be represented by a single component, such as cell mass;

TABLE 8.2 Kinetic Growth and Production: Cellular Models Cell Components Population Distributed

Segregated

Unstructured

Structured

Cells are represented by a single component, Multiple cell components are uniformly which is uniformly distributed throughout distributed throughout the culture, interact the culture. with each other. Cells are represented by a single component, Cells are composed of multiple components but form a heterogeneous mixture. and form a heterogeneous mixture.

PH2270_book.fm Page 135 Tuesday, May 16, 2006 9:37 AM

Fermenters

135

number; or their concentration of protein, DNA, or RNA. This is true for a balanced growth only, since doubling cell mass for balanced growth is accompanied by doubling all the other measurable properties of a cell population; and (2) cellular mass is distributed uniformly throughout the cell culture; and therefore the cell suspension is regarded as hom*ogeneous. The cell’s heterogeneous nature is ignored and cell concentration is expressed as wet or dry weight per unit volume. In addition to assumptions about cells, a medium must be formulated so that only one component limits growth. Other components should be present at sufﬁciently high concentration so that minor changes do not signiﬁcantly affect growth and proliferation. The fermentation environment must also be controlled so that parameters such as pH, temperature, and DO2 concentration are maintained at a constant level. In this section, cell kinetic growth equations are developed and later applied to the analysis and design of the perfect fermenter. Structured, more realistic models that consider the multiplicity of cellular components will be introduced later. Accordingly, growth rate can be deﬁned in various ways. While dCX /dt and rX appear to be the same, this is actually not correct — the two quantities are only equivalent during batch operation. The expression dCX /dt is change in cell concentration that takes into account the effect of in and out media ﬂow, cell recycling, and other operating conditions, while rX is the actual cellular-growth rate. The growth rate based upon the number of cells, and growth rate based upon the aggregate cell weight are also not essentially the same in view of the fact that cell size varies considerably from one growth stage to another. To illustrate, when individual cell mass increases without cell division, the growth rate based on the cell weight increases, while the growth rate based upon the number of cells remains constant. During exponential growth, however, the growth rate based upon the cell number and the growth rate based upon cell weight are proportional. The growth rate is sometimes confused with the division rate (the rate of cell division per unit time). If all cells in a fermenter vessel at time t = 0 (Cn = Cn0) divided once after a certain period of time, the cell population will have increased to Cn0 × 2. If cells are divided N times after the time t, the total number of cells will be: Cn = Cn0 × 2N

(8.1)

The average division rate is: δ=

(

1 log2 Cn − log2 C n 0 t

)

(8.2)

The division rate at time t is: δ=

d log 2 Cn dt

(8.3)

Therefore, the growth rate, expressed as the change in cell number with time, is the slope of the Cn-vs.-t curve, while the division rate is the slope of the log2-Cn-vs.-t

PH2270_book.fm Page 136 Tuesday, May 16, 2006 9:37 AM

136

Producing Biomolecular Substances

Death Stage

Cell Concentration

Static Stage

Exponential-Growth Stage

Latent Stage

Time

Death Phase

Stationary Phase

Log Phase

Lag Phase

Log cell number

FIGURE 8.1 Batch fermentation stages.

Time

FIGURE 8.2 Batch fermentation growth curve.

curve. Since the division rate is constant during exponential growth but the growth rate is not, these terms are mutually exclusive and should not be confused. If fresh, sterile medium is inoculated and cell density is measured during subsequent growth and then plotted against time, the results distinguish six stages in the batch cellgrowth cycle, as seen in Figure 8.1 and Figure 8.2. 1. Latent: the period during which the change in cell number is null. 2. Accelerated growth: the period during which the cell number begins to increase and the rate of cell division accelerates.

PH2270_book.fm Page 137 Tuesday, May 16, 2006 9:37 AM

Fermenters

137

3. Exponential growth: the period during which the cell number increases exponentially and the growth rate increases, while the division rate remains constant at its maximum. 4. Decelerated growth: the period subsequent to the point where the growth rate reaches maximum during which both growth and division rates decrease. 5. Static: the period during which the cell population reaches maximum for the given conditions and proliferation stops. 6. Death: the period after the limiting growth substance is depleted where cells begin to die and the number of viable cells decreases. Kinetic growth data and kinetic models can be used to predict the length of the various fermentation stages in order to estimate the necessary size for the fermenter before considering other and more complex factors. In addition, the models used in designing batch fermentation processes can be used to predict the fermenter size necessary for continuous culture fermentation. The main objective of fermentation is to support a speciﬁc culture’s growth and proliferation and to promote a high end-product yield. In certain circ*mstances excessive essential nutrient concentration can inhibit growth or even kill off the culture, so essential nutrients should not always be supplied in excess. It is a common practice to limit the concentration of at least one essential substance, keeping all the others in excess, so that growth increases exponentially until this essential (limiting) substance is depleted. For single cells, growth rate can be expressed in terms of cell concentration X, and the speciﬁc growth rate μ (time) is deﬁned by: μ=

1 dX X dθ

(8.4)

If the value of μ is constant, the culture is growing exponentially. Typically, batch fermentation cell growth and metabolic end product synthesis will demonstrate exponential increase until limiting nutrient concentration decreases to a level at which the cells begin to die. The perfect time to terminate a fermenter batch is ordinarily when the end product is formed at the same rate as cell proliferation. Sometimes, when end product formation lags behind cell proliferation, the fermentation is allowed to proceed until desired product concentration is attained even though the cells are in the death stage. Mass doubling time (μd) is usually determined, and the speciﬁc growth rate μ for single-cell growth is related to mass doubling time by the formula μd = ln(2)/μ. Customarily in the case of a limiting substance, the Monod relationship (see Figure 8.3) is used: μ=

μ m Ci K i + Ci

(8.5)

where μm is the maximum speciﬁc growth rate, Ki is a saturation constant, and Ci is the limiting nutrient concentration i. The maximum speciﬁc growth rate μm is determined when Ci >> Ki.

PH2270_book.fm Page 138 Tuesday, May 16, 2006 9:37 AM

138

Producing Biomolecular Substances

mmax

Specific Growth Rate (m)

Substrate Concentration (S)

FIGURE 8.3 Monod growth curve.

Ki is the concentration of nutrient i when μ = μm/2. The Monod equation was derived empirically and is a simpliﬁed model of complex cell growth. Determination of μm and Ki produces a measure of population growth that can be used to predict large-scale culture population dynamics, particularly when nutrient uptake is also determined. With a more complex environment (e.g., two limiting nutrients, toxin formation, product concentration inhibition, etc.), variations of the basic Monod formula can determine speciﬁc growth rate μ at any concentration and, at the same time, predict maximum cell population for a culture under speciﬁc circ*mstances such as with two limiting nutrients: ⎛ C1 ⎞ ⎛ C2 ⎞ μ = μm ⎜ ⎝ C1 + K1 ⎟⎠ ⎜⎝ C2 + K 2 ⎟⎠

(8.6)

⎛ Ci ⎞ ⎛ C P ⎞ μ = μm ⎜ ⎝ Ci + K i ⎟⎠ ⎜⎝ CP + K P ⎟⎠

(8.7)

or with product inhibition:

where KP is the saturation constant for the product Ki, and CP is the product concentration. When a stirred-tank fermenter (STF) (see Figure 8.4) is inoculated, it will begin to proliferate exponentially after the latent stage, so that the change in cell concentration in the batch culture is equal to its cellular growth rate: dC X = rX = μC X dt

(8.8)

PH2270_book.fm Page 139 Tuesday, May 16, 2006 9:37 AM

Fermenters

139

FIGURE 8.4 Stirred-tank fermenter.

Then, Eq. 8.8 is integrated to develop the performance equation for the batch culture:

∫

CX0

dC X = rX

∫

CX0

dC X = μC X

∫ dt = t − t

(8.9)

This only applies when rX is larger than zero. Thus, t0 is not the time that the culture was inoculated, but the time that the cells actually started proliferating, which is also the beginning of the accelerated growth stage. Accordingly, the batch growth time t – t0 is the area under the 1/rX vs. CX curve between CX and CX0. Batch growth time is estimated by the CX -vs.-t curve, which is a more direct determination. A U-shaped curve is characteristic of autocatalytic reactions: S+X→X+X

(8.10)

The autocatalytic reaction rate is slow at the start because the X concentration functioning as a biocatalyst is low. The rate increases as cells multiply and reach maximum population density. As the substrate is depleted and toxic metabolic waste products accumulate, the reaction rate slows and eventually approaches null. Monod satisfactorily represents growth rate during exponential growth:

∫

C X0

( K S + CS )dC X = μ maxCS C X

∫ dt

(8.11)

ΔC X C X − C X0 = − ΔCS −(CS − CS0 )

(8.12)

and growth yield (YX/S) is: YX /S =

PH2270_book.fm Page 140 Tuesday, May 16, 2006 9:37 AM

140

Producing Biomolecular Substances

Change in cell concentration with respect to time is: ⎞ C ⎛ K SYX /S K SYX /S C (t − t 0 )μ max = ⎜ + 1⎟ ln X + ln S0 ⎝ C X0 − CS0 YX /S ⎠ C X0 C X0 + CS0 YX /S CS

(8.13)

Monod parameters μmax and KS cannot be approximated by a series of batch fermentations as easily as Michaelis-Menten parameters are determined by kinetic enzyme reactions. With Michaelis-Menten, the initial reaction rate in batch fermentations can be measured as a function of substrate concentration. However, in cell growth the initial rate is always zero as a result of the latent period during which Monod does not apply. Although the Monod equation has the same general form as the MichaelisMenten, the respective rate-reaction components are different. In Michaelis-Menten: dCP r C = max S dt K M + CS

(8.14)

dC X μ maxCS C X = dt K S + CS

(8.15)

and in Monod:

The latent stage is the initial growth period where the growth rate for the cell population is either null or negligible, although the cells are still able to increase in size; it occurs while the cells adjust to a new environment and new medium before initiating accelerated growth. Factors such as cell type, age, inoculum size, and culture conditions determine this stages’ length. For example, if a culture is inoculated from a low-nutrient concentration medium to a higher-nutrient concentration, the length of the latent stage increases. If, on the other hand, the culture is inoculated from a high- to a low-nutrient medium, there is typically no latent stage. Another important factor affecting latent stage length is inoculum size (i.e., if a small number of cells are inoculated into a large volume, they will generally experience an extended latent stage). In large-scale fermentation, a primary objective is to shorten the latent stage as much as possible, so when inoculating a large fermenter, a series of progressively larger seed lots are created to minimize the latent-stage effect. Batch culture involves inoculating sterile medium with a seed culture. After inoculation, apart from air addition for aerobic fermentations and the removal of waste gasses, usually nothing additional is required to be added or removed. The rapid change to a new environment (the sterile fermenter) affects four key variables: (1) inoculation into a medium with high nutrient concentration causes delayed cell growth until the culture adapts to the new environment; (2) essential molecules synthesized by the cell to promote growth (vitamins, activators) may be lost by diffusion out of the cells and may take time for replenishment; (3) the inoculum size and viable cell percentage greatly affect the duration of the latent

PH2270_book.fm Page 141 Tuesday, May 16, 2006 9:37 AM

Fermenters

141

stage; and (4) the maturity of the inoculum is important because newer cells have not stored the same quantity of required metabolic substances as cells already in exponential growth. The latent stage ends once a key cellular component reaches critical level (c) within the cell, therefore: c = aV + bN 0θ1 + dθ1

(8.16)

where V is the volume of inoculum, N0 is the number of cells/new volume, a is the key cellular limiting component concentration/old volume x, b is the increase in key cellular component concentration/time per cell for older cells, d is the internal cell production of the key cellular component for newer cells, and θ1 is the time of the latent stage. Since θ1 is dependent on V, the larger V is, the shorter θ1 will be, so θ1 is proportional to 1/N0 for large inoculum volumes. Fermenter design should minimize the length of the latent stage in order to develop maximum functionality. Thus, the following three points are signiﬁcant: (1) the inoculum should be as active as possible — preferably in exponential growth; (2) the inoculum medium should correspond as closely as possible to that of the fermenter; and (3) a reasonably large inoculum, at least 5% total fermenter volume, should be used to minimize losing key metabolic intermediates through diffusion. When the accelerated growth cycle begins, it increases gradually, reaching maximum rate during the exponential growth stage. This transitional period is frequently called the accelerated growth stage and is sometimes deﬁned as part of the latent stage. With single-cell organisms a culture undergoing balanced growth emulates a ﬁrstorder autocatalytic reaction (e.g., the progressive doubling of cell number results in a continually increasing rate of growth). Thus, the rate of cell increase at any particular time is proportional to the number (Cn) of cells present at that time: rn =

dCn = μCn dt

(8.17)

where the constant μ is known as the speciﬁc growth rate (hr–1). The specific growth rate should not be confused with growth rate, which has different units and meaning. The growth rate is the change in cell number with time while the specific growth rate is: μ=

1 dCn d ln Cn = Cn dt dt

(8.18)

which is change in the log of the cell number with time. Comparing Eq. 8.18 and Eq. 8.19 shows that the speciﬁc growth rate μ is equal to ln 2 times the division rate μ. If μ is constant with time during the exponential growth period, the expression can be integrated from time t0 to t yielding:

PH2270_book.fm Page 142 Tuesday, May 16, 2006 9:37 AM

142

Producing Biomolecular Substances

Cn = Cn0 exp[μ(t − t0 )]

(8.19)

where Cn0 is the cell number concentration at t0 when the exponential growth starts. Thus, Eq. 8.20 shows exponential increase in the number of cells with respect to time. The time required to double the population, called the doubling time (td), can be determined by establishing Cn = 2 Cn0 and t0 = 0, and solving for t: td =

ln 2 1 = μ δ

(8.20)

The doubling time is inversely proportional to the speciﬁc growth rate and is equal to the reciprocal of the division rate. The exponential growth rate where X0 is the initial population at time 0 and X is the population at time τ for a key cellular limiting component μ for cell population X at time τ is: ⎛ X ⎞ μ Cθ ln ⎜ ⎟ = m i ⎝ X 0 ⎠ Ci + K i

(8.21)

this predicts the culture’s exponential growth if Monod applies. However, Monod is not precise, and if growth is extremely rapid, the two following equations can be used (1) where C0 is the initial key cellular component concentration: μ=

μ m Ci K i + C0

(8.22)

and (2), where B is a constant and N is the cell population: μ=

μ m Ci Ci + BN

(8.23)

The consumption rate of a key cellular limiting component (dA/dμ) is proportional to the number of viable cells reaching static stage where X is the number of cells and KA is a the proportionality constant for a key cellular limiting component A. If the culture is in exponential growth, and if the initial concentration of A = A0 and X = X0 (initial concentration of cells, when A = 0, X = XS (the static population), then this relationship yields a maximum population (XS) for initial concentration of key cellular component A (A0) and initial cell population (X0) at the start of exponential growth: XS = X0 +

μ A0 KA

(8.24)

Integrating the equation between X = X0 and X = XS, yields the time taken to reach static stage.

PH2270_book.fm Page 143 Tuesday, May 16, 2006 9:37 AM

Fermenters

143

In normal practice, fermentation is terminated either before the death stage is reached or just as the cell population starts to decrease as a result of total depletion of a key cellular component A. An exception occurs, however, when product formation exhibits a sizable lag behind cell proliferation; so, if product concentration is still increasing, fermentation could be taken to the death stage.

FACTORS AFFECTING SPECIFIC GROWTH RATE Previously, it was stated that one of the most widely employed equations for the effect of substrate concentration on μ is Monod, an empirical expression based on that normally associated with enzyme kinetics: μ=

μ maxCS K S + CS

(8.25)

where CS is the concentration of the limiting substrate and KS is a system coefﬁcient. The value of KS is equal to the concentration of the key metabolic limiting substance when the speciﬁc growth rate is half of maximum (μmax). And, while Monod is an oversimpliﬁcation of cell growth, it frequently describes fermentation systems with low cell-growth-inhibiting component concentrations (see Figures 8.5 and 8.6). According to Monod, further increases in limiting substances after μ reaches μmax does not affect speciﬁc growth rate. Also, speciﬁc growth rate decreases as substrate concentration is increased beyond a certain level. The following expressions improve Monod: μ=

μ maxCS K l1 + CS + ( K l2 CS )2

(8.26)

0.5

μ, 1/h

0.4 0.3 0.2 0.1 0 0

S, g/L FIGURE 8.5 Monod kinetics with growth rate limited by one substrate; speciﬁc growth rate reaches a maximum value of 0.5 h–1; value of KS here is 0.5 gL–1; note that when S = 0.5 gL–1, μ is half of its maximum.

PH2270_book.fm Page 144 Tuesday, May 16, 2006 9:37 AM

144

Producing Biomolecular Substances

0.5 DO = 8 mg/L

μ, 1/h

0.4

DO = 0.2 mg/L

0.3 DO = 0.1 mg/L

0.2 0.1 0 0

S, g/L FIGURE 8.6 Monod kinetics with growth rate limited by two substrates; speciﬁc growth rate reaches maximum value of 0.5 h–1; value of KS here is 0.5 gL–1; value of KDO is 0.1 mgL–1. Note that when CDO = 0.1 mgL–1, μ is half of its maximum at values of S >> KS.

μ = μ max (1 − e− CS / K s ) μ=

(8.27)

μ max (1 + K S CS− λ )

(8.28)

μ maxCS βn + CS

(8.29)

μ=

If several limiting substances are used, the following expression can be employed: μ = μ max

C1 C2 K1 + C1 K 2 + C2

(8.30)

If the limiting substance is the culture’s energy source, a certain amount of substrate is used for other than growth, and some expressions include the term, ke, for cell subsistence: μ=

μ maxCS − ke K S + Cs

(8.31)

When CS is so low that the ﬁrst term of the right-hand side of Eq. 8.31 is μke, the speciﬁc growth rate is null. Alternative models appear to be better as growth models for certain microorganisms, although the solutions are more difﬁcult than Monod’s.

PH2270_book.fm Page 145 Tuesday, May 16, 2006 9:37 AM

Fermenters

145

Waste Medium & Product Fresh Medium

FIGURE 8.7 Continuously stirred-tank fermenter (CSTF).

Cell growth produces metabolic wastes that collect in the medium and inhibit cell growth; their effect can also be added to Monod, as in: ⎛ Cs ⎞ ⎛ K P ⎞ μ = μ max ⎜ ⎝ K S + CS ⎟⎠ ⎜⎝ K P + CP ⎟⎠ C ⎞ ⎛ CS ⎞ ⎛ μ = μ max ⎜ 1− P ⎟ ⎟ ⎜ ⎝ K S + CS ⎠ ⎝ CPm ⎠

(8.32)

(8.33)

CPm is the maximum metabolic waste concentration allowed in the medium for uninhibited cell growth. The microorganism-speciﬁc growth rate is directly affected by medium pH, temperature, O2 supply, etc., since optimal pH and temperature generally differ from one organism to another. When a culture’s numerical growth is limited by the exhaustion of the available limiting substance(s) and the accumulation of metabolic waste products, the growth rate declines and eventually stops. At this point the culture is said to be in the static stage. Transition from exponential to static growth involves an erratic growth period during which various cellular components are synthesized at unequal rates and consequently have chemical compositions different from those of the cells in exponential growth. The static stage is usually followed by a death stage. Death occurs because of cellular energy reserve depletion and/or metabolic waste accumulation. Like the preceding growth stages, the death stage is exponential. In some cases, organisms not only die but also decompose.

KINETIC MODELS The material balance for a culture in a CSTF can be shown where rX is the fermenter cell growth rate and dCX/dt represents the change in cell concentration with time. For valid fermenter operation at steady-state, the cell-concentration change over time

PH2270_book.fm Page 146 Tuesday, May 16, 2006 9:37 AM

146

Producing Biomolecular Substances

1 rx 2

Cxopt

FIGURE 8.8 CSTF at maximum productivity.

must equal zero (dCX/dt = 0), and the culture must grow quickly enough to replace any cells lost through the outlet: τm =

V C X − C Xi = F rX

(8.34)

This reﬂects the required residence time as equal to CX – CXi times 1/rX on the 1/rXvs.-CX curve. Graphing cell residence time measures up fermenter effectiveness; the shorter the residence time in reaching a particular cell concentration, the more effective the fermenter. If the input stream is sterile (CX = 0), and the cells are growing exponentially (rX = μCX), for a CSTF at steady state with a sterile feed, the speciﬁc growth rate μ is equal to the dilution rate D and is equal to the reciprocal of the residence time μm since the speciﬁc growth rate of a culture can be controlled by changing media ﬂow rate (see Figures 8.8 through 8.10). 300

800

Flux (1/m3 hr.)

600 200 400 100 200

0 0

40 Elapsed Time (min.)

FIGURE 8.9 Flux and throughput vs. time for CSTF.

Throughput (l/hr.)

Flux (1/m3 hr.) Throughput (l/hr.)

PH2270_book.fm Page 147 Tuesday, May 16, 2006 9:37 AM

Fermenters

147

mmax

Specific Growth Rate (m)

D = F/V

Substrate Concentration (S)

FIGURE 8.10 Relationship of dilution to speciﬁc growth rate for a CSTF at steady state.

So, if growth rate can be expressed by Monod, then: D=μ=

1 μ C = max S τ m K S + CS

(8.35)

CS can be calculated with known residence time and Monod as: CS =

KS 1 τ mμ −max

(8.36)

It should be noted that this is only valid when μmμmax > 1. If μmμmax < 1, the growth rate of the cells is less than the rate of cells leaving the outlet stream, all the cells in the fermenter will be washed out and Eq. 8.36 is invalid. If the growth-yield (YX/S) is constant, then: C X = YX /S (CSi − CS )

(8.37)

⎛ KS ⎞ C X = YX / S ⎜ CSi − 1 ⎟ τ mμ −max ⎝ ⎠

(8.38)

⎛ KS ⎞ CP = CPi + YP / S ⎜ CSi − 1 ⎟ τ mμ −max ⎝ ⎠

(8.39)

CX is:

Similarly, CP is:

PH2270_book.fm Page 148 Tuesday, May 16, 2006 9:37 AM

148

Producing Biomolecular Substances

Again, the above equations are valid only when μmμmax > 1; the same values for CSTF cultures can also be calculated by determining material balances for substrate and product concentrations. Equality of the speciﬁc growth and dilution rates in steady state fermentation is helpful in studying the effects of media components on the speciﬁc growth rate. By measuring steady state substrate concentration at various ﬂow rates, kinetic models can be evaluated and parameters can be approximated. If the particular culture follows Monod, the plot of 1/μ-vs.-1/CS yields μmax and KS from the intercept and slope of a plot similar to that of Michaelis-Menten, with the advantage that it shows the relationship between the independent variable CS and the dependent variable μ. However, since 1/μ approaches inﬁnity, μ as the substrate concentration decreases, undue importance is given to measurements made at low-substrate concentrations and insufﬁcient importance is given measurements made at high-substrate concentrations. For a better parameter match, the following two linear relationships can be employed: CS K C = S + S μ μ max μ max

(8.40)

μ CS

(8.41)

μ = μ max − K S

Since input and output ﬂow must be precisely controlled, separate sterile nutrient and spent medium reservoirs are necessary for proper steady state CSTF operation. Sometimes foaming and ﬁlter clogging by cell aggregates make controlling outlet ﬂow quite difﬁcult, and since it typically takes several days or even weeks to reach steady state due to cellular mutation and new environment adaptation, there is also a high risk of contamination. The speciﬁc growth rate during CSTF fermentation can be estimated by measuring the slope of cell concentration vs. time. Substrate concentrations are determined at the same points, and plots are constructed to determine particular kinetic parameters. The unit cost of biopharmaceuticals manufactured by continuous culture bioprocess has come down over the years, and it will continue to do so as products become more exotic. Thus, a majority of biopharmaceuticals currently in the regulatory pipeline will conceivably be produced by continuous culture. Key to both the biologic and ﬁnancial aspects of biopharmaceutical manufacturing is the relationship between continued cell growth and product formation. The cost of production is proportional to the cost of producing cells, which can be dramatically reduced if productive cell life can be extended. When employing continuous culture in a CSTF, the process typically maintains high densities of product-expressing cells for long periods of time. In such systems, cells are harvested, medium containing expressed product is removed, and new medium is either added or perfused at regular intervals. Products maintained in incubated batch reactors are at greater risk of degradation, aggregation, oxidation, inactivation, and contamination by proteolytic enzymes and other cell lysis products

PH2270_book.fm Page 149 Tuesday, May 16, 2006 9:37 AM

Fermenters

149

than those produced by continuous culture — for the reason that the product residence time in the CSTF is shorter and the removed end product can be chilled to 9°C to minimize degradation. Furthermore, costs can all be reduced by reducing fermenter and facility size, utility requirements, media handling problems, and downstream puriﬁcation, thereby enabling the use of conventional clean-room technologies. Continuous culture can reduce the capital requirements for bioprocess facilities and put much less highvalue product at risk during a particular time interval, as compared with batch systems where everything can be lost if contamination, mechanical failure, or incubation failure should occur. For the design and analysis of continuous culture systems based on the CSTF, four assumptions are required: (1) mixing takes place so that exit stream contents have the same composition as the rest of the fermenter vessel; (2) the concentration in the fermenter vessel of all components is the same in all areas of the vessel; (3) if the process is aerobic, the D2O concentration is the same in all parts of the vessel; and (4) heat transfer characteristics of the system are constant (i.e., heat generated by the fermentation is continuously removed). When concentration, cell population, and temperature in any area of the vessel do not change with time, a steady state has been reached and mass balance can be applied to any component so that: ⎡ rate ⎤ ⎡ rate ⎤ ⎡ rate ⎤ ⎢ of ⎥ ⎢ of ⎥ ⎢ ⎥ of ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ addition ⎥ − ⎢ removal ⎥ + ⎢ production ⎥ = 0 ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ to ⎥ ⎢ from ⎥ ⎢ within ⎥ ⎢⎣ system ⎥⎦ ⎢ system ⎥ ⎢ system ⎥ ⎣ ⎦ ⎣ ⎦

(8.42)

Steady state balance (neglecting cell death) is depicted by: ⎛ F⎞ ⎛ F⎞ ⎜⎝ ⎟⎠ X0 = ⎜⎝ ⎟⎠ X − r V V

(8.43)

If X is cell concentration in both the fermenter vessel and exit stream, X0 is the cell concentration of the feed, F is the ﬂow rate of the feed and exit stream, V is the fermenter vessel volume, r is the rate of cell formation (cells/unit time/unit volume) equal to dX/dτ, then D (= F/V), known as the dilution rate, is the number of fermenter vessel volumes passing through the system per unit time — the inverse of the mean residence or holding time. The dilution rate is universally employed throughout the biotechnology industry. With a single CSTF, the feed stream is normally sterile medium. Therefore X0 is zero, and X(Dτ) = 0. Since either X = 0 or (D – τ) = 0, a cell population > 0 can be maintained only if the speciﬁc growth rate μ is balanced by the dilution rate D (i.e., a nonzero cell population can only be maintained when D = τ). Once the culture has adjusted its speciﬁc growth rate to the dilution rate, the expression DX0 = X(Dτ) can be satisﬁed by any X value greater than zero. This

PH2270_book.fm Page 150 Tuesday, May 16, 2006 9:37 AM

150

Producing Biomolecular Substances

Oxygen Supply Expended Medium and Products

Fresh Medium Carbon Dioxide Removal

Gas Permeable Tubing Distributed Lung

FIGURE 8.11 Static maintenance fermenter (chemostat).

condition is only reliable when: (1) speciﬁc growth rate is independent of cell population; and (2) growth is exponential and not affected by decreasing concentration of the growth-limiting key cellular component. For continuous culture, a static maintenance fermenter, or chemostat (see Figure 8.11), can be used to maintain cell densities at as high as 100 times the density of conventional systems. Stocked with cells, porous tubes constantly input nutrients and remove waste products, gas permeable tubes continuously add oxygen and remove carbon dioxide, while computer controls monitor these functions and maintain a hom*oeostatic environment. During continuous culture, the cell population is constant both biochemically and physiologically, and maintained over an indeﬁnite period. Major chemostatic continuous culture advantages include the ability to (1) manipulate the speciﬁc growth rate over a full range without the need to vary medium composition or environmental conditions, (2) strategically evaluating the effects of the culture’s physico-chemical environment while maintaining constant growth rate, (3) analyzing both the culture’s steady state and its deﬁned transitory states, (4) setting unique culture conditions by changing the nature of the growth-limiting substrate(s), and (5) maintaining culture growth over long periods. Chemostat theory forecasts that it has signiﬁcantly higher productivity than the batch culture. Chemostat productivity can be maximized by careful manipulation of the environmental parameters, reducing downtime for cleaning, and eliminating the time-consuming, repetitive operations (e.g., sterilization, cleaning, etc.) associated with batch fermenters. Continuous culture is an important bioprocess tool that relates not only to productivity, but to concentration, substrate conversion efﬁciency, duration of biosynthetic activity, and product stability. In a batch culture, if one of the key cellular components is at growth-limiting concentration, the cell balance at

PH2270_book.fm Page 151 Tuesday, May 16, 2006 9:37 AM

Fermenters

151

steady state can be written using the Monod expression for speciﬁc growth rate μ, and if the yield factor Y is deﬁned as: Y=

Mass of cells formed Nutrient consumed

(8.44)

The steady-state balance on the limiting component is expressed by a Monod chemostat model; where X0 = 0 (sterile feed). D(C0 − Ci ) −

μ m Ci X =0 Y (Ci + K i )

(8.45)

Once the dilution rate has exceeded the maximum possible growth rate, the only solution is X = 0. The situation where D exceeds Dmax, and all the cells are lost from the system, is known as washout. When X = 0 and Ci = C0, then: Dmax =

μ m C0 C0 + K i

(8.46)

OPTIMAL CONDITIONS The cell production rate per fermenter volume is DX (FX/V), and the maximum cell production rate is: d ( DX ) =0 dD

(8.47)

The rate for optimal cell production is: 1/ 2

Dopt

⎡ Ki ⎤ = μm − ⎢ ⎥ ⎣ (C0 + K i ) ⎦

(8.48)

Thus, if C0 >> Ki (which is usually the case), Dopt, approaches μm near washout. In terms of controlling a CSTF, since X and Ci are sensitive near washout, optimal (maximum) cell production is not permitted so that fermenter operation will remain stable. Typically, fermentations are carried out at a dilution rate of around 80% of Dmax. Determining batch cycle residence time and the processing time for continuous cultures facilitates a close estimation of the required fermenter’s capacity. Moreover, since a fermenter is really a multipurpose processor and, in practice, performs a number of functions, to ensure proper operation it must: be of sufﬁcient size to provide the required production capacity as a bioprocessor; be designed to ensure that media and cells are well dispersed as mass transfer equipment; maintain ade-

PH2270_book.fm Page 152 Tuesday, May 16, 2006 9:37 AM

152

Producing Biomolecular Substances

quate available dissolved oxygen concentration for individual cells in aerobic cultures; provide for temperature control, and for the rapid dispersion of control chemicals as a control device (i.e., pH and foaming and ensure representative sampling of important parameters such as cell population, pH, dissolved oxygen concentration, etc.); and, ensure constant temperature is maintained during the growth cycle by both cooling and heating to sterilize the system in situ as a heat transfer device. Having determined the required fermenter volume, consideration must next be given to including other key functions in the fermenter model.

MAXIMIZING CONTINUOUS CULTURE PRODUCTIVITY Continuous culture is typically the method of choice for acquiring relevant information about a bioprocess, although maximizing productivity during continuous culture is often quite difﬁcult. Contamination and natural selection of spontaneous mutations must be constantly monitored. Also, small environmental deviations typically cause signiﬁcant reductions in cell population. Continuous culture parameters are based upon the growth kinetics of single-cell organisms in the exponential growth stage. Although in theory growth should go to inﬁnite density, in practice the essential growth-rate-limiting substance that is depleted and the toxic metabolic waste products that are formed result in limited density. A good continuous culture maximizes viable cells and minimizes dead and dying cells. Both biochemical and physical techniques typically detect and quantify living cells in the population.

CONTINUOUS CULTURE DILUTION RATE The dilution rate for continuous culture is equal to the ﬂow rate of fresh medium into the fermenter divided by fermenter volume. During continuous culture, the rate of change in vessel organism concentration equals the rate of growth minus organisms removed. At steady state, the biomass and growth-rate-limiting substrate concentrations are constant and the speciﬁc growth rate can be controlled by substrate concentration. At the critical dilution rate loss of cell population (washout) will occur. Practically, maximum speciﬁc growth rate can be determined by increasing the dilution rate until washout occurs. Substrate consumption during continuous culture follows ﬁrst order Monod kinetics. In continuous culture, biomass concentration and growth-limiting substance concentration can be determined from supply-vessel growth-limiting substrate concentration and the biomass yield of that substrate.

RECOMBINANT CULTURE KINETICS PLASMID-CARRYING CELL GROWTH-RATE KINETICS Assuming p is the probability of plasmid-carrying cells (X+) producing plasmid-free cells (X–) after one division, then N plasmid-carrying cells will produce N(1 – p) plasmid-carrying cells and Np plasmid-free cells after one division; the total number of X+ cells will be N(2 – p). During exponential growth, the plasmid-carrying cell growth-rate will be:

PH2270_book.fm Page 153 Tuesday, May 16, 2006 9:37 AM

Fermenters

153

dC X + = (1 − p )μ +C X + dt

(8.49)

where μ+ is the speciﬁc growth rate of the plasmid-carrying cells, and C X + is the number of plasmid-carrying cells per unit volume. If the mass of the cells is approximately proportional to the number of cells, the preceding rate equation can be also applicable when C X + is cell mass per unit volume, and the growth rate of plasmidfree cells will be: dC X − = pμ +C X + + μ −C X − dt

(8.50)

If we assume that μ+ and p are constant, the integration of Eq. 8.50 yields: C X + = C X + exp[(1 − p )μ +t ] 0

(8.51)

where C X + is the initial concentration of the plasmid-carrying cells. For the plasmidfree cells, substituting Eq. 8.51 into Eq. 8.50 yields: dC X − − μ −C X − = pμ +C X + exp[(1 − p )μ +t ] 0 dt

(8.52)

Solving Eq. 8.52 for the constant μ+ yields

CX− =

pμ +C X + 0

(1 − p )μ + − μ −

{exp[(1 − p )μ +t ] − exp(μ −t )} + C X0− exp(μ −t )

(8.53)

Therefore, Eqs. 8.51 and 8.53 predict how C X + and C X − change with time for given values of p, μ+, and μ–. The fraction of the plasmid-carrying cells in the total population (ƒ) can be expressed as: f=

CX+ CX+ + CX−

(8.54)

Substituting Eqs. 8.51 and 8.53 into Eq. 8.54 yields: f=

exp[(1 − p )μ +t ] pμ +C X + 0 exp[(1 − p )μ +t ] + {exp(1 − p )μ +t − exp(μ −t )} (1 − p )μ + − μ −

(8.55)

PH2270_book.fm Page 154 Tuesday, May 16, 2006 9:37 AM

154

Producing Biomolecular Substances

which shows the change in the fraction of the plasmid-carrying cells with time during the exponential batch fermentation growth. During exponential growth, the number of generations (n) of the plasmid-carrying cells can be calculated from: n=

μ +t ln 2

(8.56)

So, combining Eqs. 8.55 and 8.56 will result in f for the nth generation: fn =

1− α − p 1 − α − p[ 2 n (α + p−1) ]

(8.57)

where α is the ratio of the speciﬁc growth rates: α=

μ− μ+

(8.58)

Eq. 8.57 predicts change in ƒ for the number of generations required for a series of batch fermentations if it is assumed the culture propagated exponentially during each batch. Assuming it could take 25 generations to scale up from a tube culture to a 33,000-liter fermenter, the effects of p and η on ƒ25, which decreases as η and p, increase. When p is 0.01 and η < 1, ƒ25 is close to 1 (i.e., the plasmid-carrying cells are quite stable). However, as α approaches 2.0, ƒ25 becomes zero. Practically, values of α generally range from 1.0–2.0. The change in ƒ is a function of η and ρ, where ρ was constantly adjusted to be 0.01. The value of ƒ decreases rapidly with the increase of η and ρ, and when α is 1.4, all plasmid-carrying cells will lose their plasmids in about 33 generations. 1.0 0.8 p = 0.1 0.04

0.6

0.01

f25

0.001 0.0001

0.4 0.2 0

0.5

1.0 α

1.5

2.0

FIGURE 8.12 Fraction plasmid-carrying cells after 25 generations (ƒ25) as a function of η and ρ.

PH2270_book.fm Page 155 Tuesday, May 16, 2006 9:37 AM

Fermenters

155

The material balance for plasmid-carrying recombinant cells in a CSTF is: − DC X + + (1 − p )μ +C X + =

dC X + dt

(8.59)

Similarly, the material balance for plasmid-free cells is: − DC X − + pμ +C X + + μ +C X − =

dC X − dt

(8.60)

Adding Eqs. 8.59 and Eq. 8.60 yields the total cell concentration: (μ +C X + + μ −C X − ) − D(C X + + C X − ) =

d (C X + + C X − ) dt

(8.61)

If a CSTF is operated so that total cell concentration is constant with time, then: μ + (C X + + αC X − ) = D(C X + + C X − )

(8.62)

and if μ = 1, then Eq. 8.62 simpliﬁes to: μ+ = D

(8.63)

Thus, speciﬁc cell growth rate in a fermenter vessel is constant and ordained by the dilution rate, so, solving Eqs. 8.59 and 8.60 after substituting Eq. 8.63 yields: C X + = C X + exp(− pDt )

(8.64)

C X − = C X0− + C X + [1 − exp(− pDt )]

(8.65)

and

During continuous fermentation, then, plasmid-carrying cell concentration is reduced and the concentration of plasmid-free cells increases. If α does not equal 1, μ+ is no longer constant during steady-state CSTF operation at a constant dilution rate, but depends on the cell concentrations (C X + , C X − ) and α, according to Eq. 8.62. As a result, μ+ changes with time. C X + and C X − changes with time can be estimated by simultaneously solving Eqs. 8.59, 8.60, and 8.62. So, substituting Eq. 8.62 into Eq. 8.59 and dividing by C X + + C X − yields:

PH2270_book.fm Page 156 Tuesday, May 16, 2006 9:37 AM

156

Producing Biomolecular Substances

α=0

1.0

1.0 0.8 1.1

0.6 f 1.2

0.4 1.4

0.2 2.0 0

FIGURE 8.13 Fraction plasmid-carrying cells (f) vs. number of generations (n).

df (1 − p )Df = − Df + dt α + (1 − α ) f

(8.66)

The Eq. 8.66 solution shows how the fraction of plasmid-carrying cells decreases with time (change in ƒ with time). For the Eq. 8.66 solution, the initial value of ƒ is approximated from Eq. 8.57. Initial fraction changes of plasmid-carrying cells with time during steady-state CSTF operation (initial ƒ value), for example, is approximated by assuming 20 generations are required for the stepwise inoculation, the initial batch, and any subsequent unsteady-state continuous fermentation. The fraction of plasmid-carrying cells, therefore, is reduced with both time and any increase in dilution rate. When ρ and D are sufﬁciently low and α approaches 1, the CSTF can be economically operated over long periods. However, if α increases to 1.4, the CSTF will lose almost all its plasmid-carrying cells within 100 hours of steady-state operation. The following methods can be used to stabilize recombinant culture systems that tend to drop their plasmids during fermentation: 1. Formulate the medium to favor growth of plasmid-carrying cells over plasmid-free cells. 2. Put selective pressure against plasmid-free cells using auxotrophic mutants or antibiotic-resistant plasmids. 3. Use a temperature-dependent mutant plasmid or strain. 4. Use a temperature-dependent control of gene expression. Plasmid is more likely to be kept unchanged when the gene expression is repressed. The higher the genetic expression, the more segregants tend to appear. 5. Use a plasmid containing no transposable elements. 6. Use a recombination-deﬁcient strain.

FERMENTER OPERATING MODES Fermenters typically operate in either batch or continuous mode. In batch mode, the fermentation proceeds for a deﬁned period and ends when the product concentration

PH2270_book.fm Page 157 Tuesday, May 16, 2006 9:37 AM

Fermenters

157

reaches a preselected level. Fermenter contents are then harvested, separated, recovered, and puriﬁed. Continuous-culture fermentation requires a steady state environment with sterile medium continuously entering and medium containing the desired end product continuously exiting, the vessel at the same rate. Continuous-culture fermentation frequently maximizes productivity, although the aseptic steady-state environment is difﬁcult to maintain. Strictly speaking, a fermenter grows prokaryotic cells such as bacteria and fungi, and a bioreactor grows eukaryotic cells such as insect and mammalian cells. In practice, many manufacturers and bioprocess scientists still refer to both as fermenters. Distinguishing between fermenters and bioreactors is frequently based on their method of agitation. Some systems are designed for multipurpose use, with interchangeable stirring systems, and in many cases differentiation has become a matter of intended use as much as a matter of design difference. Fermenters are designed to maintain an environment suitable for the controlled growth of microorganisms, so regulating the environment and maintaining aseptic conditions are their principal functions, regardless of size. Growth medium must be regulated with respect to agitation, temperature, aeration, pH, DO2, foam control, and other parameters. Aseptic conditions are critical for production since foreign organisms and their byproducts contaminate cultures, disturbing cell growth, altering product expression, and making it difﬁcult to purify biomolecular end products. To avoid this, fermenters must be made of noncorrosive and nontoxic material that can be repeatedly sterilized. Smaller fermenters are usually constructed of borosilicate glass while pilot-size systems are typically constructed of stainless steel.

STRUCTURE

AND

CONFIGURATION

Basically, fermenters consist of three parts: (1) the culture vessel, (2) their associated supply and environmental systems, and (3) their measurement and control systems. Commercial fermenter systems range over a wide complexity spectrum to meet various end-user needs. Table 8.3 lists some of the factors that should be considered in designing fermenters, and Table 8.4 lists additional operational characteristics. Within the bench-, pilot-, and production-scale groupings most manufacturers offer similar equipment. At the low end are the simple bench-top fermenters used for preliminary development work or educational purposes. These are simply closed, temperature-controlled vessels with minimum instrumentation, accessory equipment, and sophisticated features (see Table 8.4).

TABLE 8.3 Important Fermenter Design Considerations • • • • • •

Maintain sterility Oxygen transfer Heat transfer Instrumentation and controls Biological kinetics Fluid hydraulics

• • •

Mass transfer of substrate to microorganism Mass transfer of product from microorganism Safety Failsafe control Pressure buildup Microorganism escape

PH2270_book.fm Page 158 Tuesday, May 16, 2006 9:37 AM

158

Producing Biomolecular Substances

TABLE 8.4 Typical Fermenter Characteristics • • • • • • • •

SIZE

AND

Ease of installation Integrated piping Computer controlled Multiple ports Suitable surfaces Programmable inputs In situ sterilization Heating and cooling

• • • • • • • •

Convenient aeration Foam control Sterilizable peripherals Interchangeable peripherals Batch vs. continuous Failsafe system Physical integrity Reliability

SCALE

As bioprocess complexity increases, signiﬁcant accessory equipment and instrumentation are added, particularly measurement and control systems. Despite manufacturer claims to the contrary, there are few design features signiﬁcant enough to rank one make of fermenter over another. Essential features to look for in selecting fermenters are product reliability, multipurpose ﬂexibility, interchangeability, level of sophistication, compatibility, and range of monitoring instrumentation available. Fermenters are frequently categorized according to size: laboratory, or research, or bench-top fermenters typically range from 1–50 liters (see Figures 8.14 and 8.15), pilot plant fermenters typically range from 50–1,000 liters, and production scale fermenters are usually larger than 1,000 liters. Although this division is somewhat arbitrary, it has generally been adopted by industry. Nevertheless, some productionscale fermenters are equivalent in volume to pilot-scale fermenters, and some pilotscale fermenters, correspondingly, are equivalent in volume to research- or benchtop fermenters. The large number of fermenters currently available covers a spectrum that blurs distinction between one class and another, especially when producing smaller pro-

FIGURE 8.14 Stirred-tank fermenter.

PH2270_book.fm Page 159 Tuesday, May 16, 2006 9:37 AM

Fermenters

159

FIGURE 8.15 3-liter bench-top stirred-tank fermenter.

duction quantities of high-value-added substances such as cytokines or interferons. Bench-top fermenters range from basic models suitable for undergraduate experimentation, to those with highly sophisticated monitoring and control systems, armed with formidable computing capability, which monitors and controls numerous variables. Table 8.5 gives a feature comparison, typical of most stirred-tank fermenters.

TABLE 8.5 Stirred-Tank Fermenter (STF) Comparison Feature Size: Sterilize: Mixing: Drive: Fittings: Seals: Jacketing: Ports: Surfaces: Parts: Other:

Small Bench

Large Bench

Pilot-Scale

1–10 L 10–100 L 100–1,000 L autoclave in situ in situ airlift bladed turbine impellers ←⎯⎯⎯ direct drive or magnetically coupled ⎯⎯⎯→ autoclavable ←⎯⎯⎯⎯⎯⎯⎯ sterilized in situ ⎯⎯⎯⎯⎯⎯⎯→ autoclavable ←⎯⎯⎯⎯⎯⎯⎯⎯ O-ring seals ⎯⎯⎯⎯⎯⎯⎯⎯→ ←⎯⎯⎯⎯⎯⎯⎯⎯⎯ yes, with internal bafﬂes ⎯⎯⎯⎯⎯⎯⎯⎯→ 4–10 4–10 10–20 ←⎯⎯⎯⎯⎯⎯⎯ electro-polished stainless steel, Pyrex ⎯⎯⎯⎯⎯⎯⎯→ ←⎯⎯⎯⎯⎯⎯⎯ interchangeable impellers and ﬁttings ⎯⎯⎯⎯⎯⎯⎯→ bottom aeration by sparging; maximized oxygen transfer; remote valve process control; view window; sterile compressed air between agitator seals; distinctive safety features

PH2270_book.fm Page 160 Tuesday, May 16, 2006 9:37 AM

160

Producing Biomolecular Substances

CONTAINMENT VESSEL Regardless of fermenter size, certain elements are common to all. Most bench-top fermenters are made of borosilicate glass (Pyrex™/Kimax™) and #440 stainless steel. Both materials meet surface requirements and they are nontoxic, noncorrosive, easily cleaned, and can be steam sterilized. Glass vessels are usually composed of a glass cylinder with stainless steel head and bottom plates; they tend to be less expensive, provide easy viewing of the fermentation, and are simple to maintain. Some glass units resemble exotic beakers with head plates. The head plate provides ports for nutrient media and gas input as well as waste product removal. Inert silicone rubber o-rings generally provide a seal between the vessel and head plate, and sterilization for bench-scale units is generally performed following disassembly by autoclaving. Stainless steel vessels are more expensive and, unless a viewing port is included, severely restrict visual examination of the fermentation. Their added strength makes them far more durable, and double-wall steam sterilization systems are frequently standard, even on smaller units. Interestingly enough, more contamination problems have been attributed to glass vessels than to stainless, since selfsterilizable stainless vessels offer greater protection against contamination than the glass vessels’ silicone rubber seals.

MIXING

AND

AERATION

Agitation systems are integral to all fermenters. On smaller units, mixing is generally accomplished by direct-drive mechanical stirring through a seal in the head plate. Some models offer either magnetically coupled agitators, or air-lift systems to eliminate mechanical seals, which are usually limited to the smaller volume fermenters where low torque produces effective agitation. At the lower end, LSL BioLaﬁtte, for example, offers a 6-liter fermenter with an overlay of pressurized sterile air between the two impeller shaft seals and an automatic seal wear monitor. The transfer of energy, nutrients, substrate, and metabolite within the bioreactor must be brought about by a suitable mixing device. The efﬁciency of any one nutrient may be crucial to the efﬁciency of the whole fermentation. For the three phases, mixing fermenter contents causes the following: • • • •

Air dispersion in the medium hom*ogenization, which equalizes temperature and nutrient concentration throughout the fermenter Suspension of culture and nutrients Dispersion of immiscible liquids

One of the most critical factors for operating a fermenter is providing for adequate gas exchange. Typically, oxygen is the most important gaseous substrate for cell metabolism, and carbon dioxide is the most important gaseous metabolic product. For oxygen to be transferred from an air bubble to an individual microbe (see Figure 8.16) several independent and partial resistances must be overcome:

PH2270_book.fm Page 161 Tuesday, May 16, 2006 9:37 AM

Fermenters

161

Gas-fluid phase boundary

Fluid-cell phase boundary

CG CX 1 2 4

CY 3 Gas bubble Gas Fluid Film Film

5 CF

Fluid

Fluid Film

CZ Cell

FIGURE 8.16 Oxygen transfer resistances from medium air bubble to microbe; a critical factor for a bioreactor is adequate gas exchange. Oxygen is typically the most important gaseous substrate for cell metabolism, and carbon dioxide is the most important gaseous metabolic product. For oxygen to be transferred from an air bubble to an individual microbe, some resistances for oxygen transfer from air bubbles in the medium to cells include: (1) resistance within the gas ﬁlm to the phase boundary; (2) resistance to penetration of the phase boundary between the gas bubble and the liquid, (3) resistance to transfer from the phase boundary to the liquid, (4) resistance to movement within the nutrient solution, and (5) resistance to transfer to the surface of the cell.

1. Resistance within the gas ﬁlm to the phase boundary. 2. Resistance to penetration of the phase boundary between the gas bubble and the liquid. 3. Resistance to transfer from the phase boundary to the liquid. 4. Resistance to movement within the nutrient solution. 5. Resistance to transfer to the surface of the cell. Oxygen transfer rate and the volumetric oxygen transfer coefﬁcient are dependent upon the following parameters: 1. 2. 3. 4. 5. 6.

Vessel geometry: diameter, capacity. Mixing properties: power, impeller conﬁguration and size, bafﬂes. Aeration: sparger rate, geometry, location. Nutrient medium: composition, density, viscosity. Microorganism: morphology, concentration. Antifoam agent temperature.

INCREASED SIZE

AND

COMPLEXITY

As fermenter size increases, its special and more complex features tend to multiply. Just as some small fermenters now offer a range of features once limited to the larger equipment, the largest bench-top units now offer all the features typical of

PH2270_book.fm Page 162 Tuesday, May 16, 2006 9:37 AM

162

Producing Biomolecular Substances

TABLE 8.6 Examples of Fermenter Control System Functions • • • • • • • • • • • • • • • • •

Pressure measurement and control Shaft power measurement and control RPM measurement and control Oxygen measurement and control Flow rate measurement and control Flow rate measurement and control change Exit gas analysis Exit gas composition changes Medium composition changes Carbon source feed rate Feed rate change Foam control pH changes Acid or base addition Additives to change redox Temperature measurement and control Weight

pilot-scale units. Since the basics of fermenter design are the same, regardless of size, distinctions occur mainly in additional devices incorporated into the larger units, such as the additional entry and exit ports that are used to accommodate more monitoring probes required by the larger vessels — although every so often extra ports are employed for separating medium and other components before adding them to the culture vessel. Typically, larger vessels enable greater control (e.g., the increased thermal mass of larger vessels permits better temperature control, and their larger volume facilitates closer pH regulation and medium buffering precision). Larger size also allows utilizing process equipment not available for smaller units. For example, the Chemap Chemcell™ system enables efﬁcient, bubble-free medium aeration and cell separation, and thereby avoids growth constraints from the buildup of metabolic waste products. Pilot-scale vessels are from time to time used as largescale research vessels by increasing the amount of their ancillary equipment. Nevertheless, increased agitation effects and other key elements must be carefully considered when they apply to large research fermenters so they can also apply in the same modality to pilot-scale units. In addition, larger vessels are normally a part of an integrated bioprocess system with piping, electrical conduits, exotic monitoring and control elements, and centralized system controllers. Table 8.6 details some of the various functions that can be incorporated into a fermenter system.

PROCESS MONITORING

AND

CONTROL

Prior to discussing speciﬁc process monitoring and control applications, some general features of bioreactors relevant to control should be discussed. A computer

PH2270_book.fm Page 163 Tuesday, May 16, 2006 9:37 AM

Fermenters

163

automated system with process monitoring and controls consist of monitoring (1) liquid ﬂow, (2) dissolved gas concentrations such as O2, CO2, N2, and (3) other process parameters. Two main characteristics important to be aware of prior to designing a fermenter control system are: 1. Fermenters are multivariate systems; as one would anticipate, bioreactor control involves many variables. 2. Fermenters exhibit nonlinear dynamics; fermenter control is complicated by the fact that nonsteady-state behavior is nonlinear, which has some consequences: (a) Hysteresis, e.g., a step increase in bioreactor feed rate in a stirred-tank bioreactor (STB) results in a different transient than a correspondingly equivalent step decrease in feed rate back to the initial conditions. (b) Multiple steady states are often observed with identical feed conditions and in some cases exotic dynamics such as limit cycles, oscillatory transients, or long time lags may occur; the reasons for these behaviors are ultimately a result of the complexities of the living cell. (c) Monitoring and control of many important desirable variables are only measurable with large time lags or not at all. This gives opportunity for accurate mathematical models and/or state-of-the-art estimation techniques. Fortunately, simple models with single feedback loops are available and perform very well. Process monitoring and control stands at the leading edge of fermenter development. Progress in bioanalytical chemistry resulted in a vast array of probes, biosensors, and detectors, which enable the process engineer to screen a myriad of growth parameters. Table 8.7 lists some of the instrumentation parameters typically checked by process control systems. A bioprocess’s ﬁnancial success is directly related to its level of process control sophistication, therefore it is expected that many future bioreactor developments will be in the area of process monitoring and control. Process monitoring and control systems can be readily adapted to any size bioreactor, the only limitation being the number of such devices the vessel can physically accommodate. Currently, the modular approach is quite popular and offers a wide selection of products of varying capability that can be selected according to the project’s

TABLE 8.7 Typical Fermenter Control Ranges Temperature: Agitator speed: Stability: pH range: Pressure: DO2 range: Air flow:

8ºC above coolant to 60ºC ± 1ºC 0–1,000 rpm >98% 2–12 ± 0.1 2,000 mbar 0–100% 0–6 liters/minute

PH2270_book.fm Page 164 Tuesday, May 16, 2006 9:37 AM

164

Producing Biomolecular Substances

TABLE 8.8 Fermenter Process Control Monitoring Air flow: Coolant: Power: Temperature: Rheology: Redox: Culture: Cell concentration: Imncytmtry: Gas analysis:

ﬂow meter ﬂow meter V.O.M.; torque resistance thermocouples, thermistors, diodes tube viscometers, cone and plate viscometers, concentric cylinder viscometers, inﬁnite sea viscometers, foam control pH, dissolved oxygen (DO2), polarographic probes, galvanic probes, permeable tube/oxygen analyzer, dissolved carbon dioxide (DCO2) enzyme probes, metabolic heat, substrate analysis gravimetric dry weight, turbidity cell number, impedence, carbon dioxide production, oxygen production, DNA content, cell particle size distribution oxygen analyzers, carbon dioxide analyzers, mass spectrometry, gas chromatography

speciﬁc needs, so an imposing array of conﬁgurations are possible. A good example is the B. Braun BIOSTAT™, that offers seven basic conﬁgurations, with culture vessel volumes of 2, 2.75–20, 25, 29, 72, 92, 150, and 300 liters, as well as associated measurement and control modules, comprising either digital or analog displays for pH control, additive measurement, antifoam addition, level control, pO2, pCO2, redox, temperature, and real-time recording of results during cellculture. The antifoam control employs a conductance-regulated adjustable sensor, which is linked through the controller to a pump for adding antifoam chemicals, and with a variable adjustment to eliminate extra antifoam addition that may result from splashes. The pO2 measurement is carried out by either galvanic or polarographic electrodes, regulated by the system controller that governs agitator speed and/or airﬂow-rate adjustment. Actual pH measurement and control is accomplished by means of a pH electrode linked through the system’s controller to discrete pumps for adding either acid or base. While regular pH electrodes degrade with repeated sterilization, the newer gel-ﬁlled electrodes are extremely stable. Inlet air is controlled by a simple pump and ﬂow meter, while temperature is monitored by a thermocouple linked to a discrete temperature control system. At the core of the bioreactor is the system controller. Today, in addition to purchasing complete units and modular assemblies, there is a growing market trend to retroﬁt older bioreactors with new, state-of-the-art control systems. For example, Wheaton offers a user-friendly Macintosh-based controller, their Proteus 2000™, which can be retroﬁtted to nearly all bench-top units, and some of their pilot-scale units. Astra Scientiﬁc offers a monitoring and control package for IBM PCs, along with the interface hardware necessary for systems producing analog output. For systems without computer control, few indeed these days, data acquisition and real-time logging can still be acquired with I/O hardware packages.

PH2270_book.fm Page 165 Tuesday, May 16, 2006 9:37 AM

Fermenters

165

(a)

3 39°

37°

42° Growth Rate h–1

45°

30°

1 21°

47°

17° 0.5

48° 13.5°

0.1 3.1 3.15 3.2 3.25 3.3 3.35 3.4 3.45 3.5 3.55 1,000 / T(K)

Dimentionless Specific Growth Rate μ/μm

(b) 1.1 1 0.9

With Adaptation

0.8 0.7 0.6 0.5

Without Adaptation

0.4 0.3 0.2 0.1 0 2

FIGURE 8.17 (a) Effect of temperature on growth rate; maximum growth rate is at 39°C. Plot is a function of inverse absolute temperature; the declining line from 39°C to 21°C to 13°C suggests that the growth rate behaves similar to chemical reaction rate constants. (b) Effect of pH on growth rate; typical pH ranges over which reasonable growth can be expected are about 1–2 units; with adaptation broader ranges can be achieved. (c) Effect of dissolved oxygen concentration (DO2) on growth rate; critical dissolved oxygen concentration refers to the vaue od DO2, below which the growth rate is lower than maximum. Continued.

PH2270_book.fm Page 166 Tuesday, May 16, 2006 9:37 AM

166

Producing Biomolecular Substances

(c)

Dimensionless Growth Rate

μ/μm

1 E-Coli 0.95 0.9 0.85 0.8 0.75 0.7 0

0.1

0.2 0.3 D.O (mg/l)

0.4

FIGURE 8.17 Continued.

PROCESS MONITORING: SENSOR LOCATION

AND

FUNCTION

1. 2. 3. 4.

Many sensors penetrate into the unit interior. Some sensors operate on continuously withdrawn samples. Some sensors do not come into contact with either the medium or gasses. Inline sensors; part of fermentation equipment where the measured value is used directly for process control without the operator intervention. 5. Online sensors; part of fermentation equipment, but the measured value is not directly available for process control. An operator must intervene to enter the measured value into the fermentation system for process control. 6. Offline sensors; not a part of fermentation equipment, and the sensor’s measured value is not directly available for process control. The intervention of an operator is essential for the actual measurement and for entering the measured value into the system for process control.

PROCESS CONTROL Process control is concerned with making adjustments to the process, based upon measuring one or more variables that change as a result of the process function, as shown in Table 8.9.

FEEDBACK CONTROLLERS Feedback controllers compare the measured value of the process variable that must be controlled with its set point and adjust an actuator in order to suppress the deviation between the measuring value and the set point.

PH2270_book.fm Page 167 Tuesday, May 16, 2006 9:37 AM

Fermenters

167

TABLE 8.9 Fermenter Process Control System Components • • • • • • •

Sensors Controllers Actuators (ﬁnal element) Process controllers Manual controls Automated controllers Feedback controllers

TABLE 8.10 Automatic Control Systems • • • • • • • • •

On/Off controllers Modulated controllers Proportional controllers Integral controllers Derivative controllers Combination controllers: PI, PD, and PID Two-position on/off controllers Modulated controllers Proportional controllers

AUTOMATIC CONTROL SYSTEMS The controller output change is proportional to the input signal created by the environmental change error that was detected by the sensor: M = M o + KcΣ

(8.67)

where M = output signal, and Mo = controller output signal when there is no error, Kc = controller gain or sensitivity, Σ = the error signal. Various control systems are shown in Table 8.10.

INTEGRAL CONTROLLERS The integral controller output signal is determined by the integral of the error over operating time:

∫

M = M o + 1 / Ti Σdt where Ti is the integral time.

(8.68)

PH2270_book.fm Page 168 Tuesday, May 16, 2006 9:37 AM

168

Producing Biomolecular Substances

DERIVATIVE CONTROLLERS Derivative controllers sense the rate of change of an error signal and contributes an output signal component that is proportional to the derivative of the error signal. M = M o + Td d Σ / dt

(8.69)

where Td is the integral time.

FEED-FORWARD CONTROLLERS Feed-forward controllers employ a measured variable(s) other than the process variable that must be controlled to carry out an action.

ADAPTIVE BIOREACTOR CONTROLS Mostly applied in systems where process variables or characteristics aren’t known and can’t be measured directly, or when the bioprocesses’ static or dynamic behavior changes with time. The adaptive controller can either use online measured process data, a theoretical model, or a combination of these to predict the change in the static or dynamic behavior of the process.

COMPLEX BIOREACTOR CONTROLS Computer applications in fermentation technology: data logging, data analysis, process control, process data logging, data analysis.

CLEANING

AND

STERILIZATION

A bioprocess facility’s ability to be cleaned is crucial to its proper design and construction. Piping, inlet-outlet ports, valves, sensors, regulators, and other components must be designed to eliminate dead spaces, ridges, and crevices where material can accumulate, and must be constructed to resist the wear and corrosion that can produce areas that are preferential sites for microorganism growth and substance contamination. The cleaning process begins when a fermenter is emptied following completion of a culture run. Clean-in-place (CIP) and sterilize-in-place (SIP) systems minimize disassembly and down time. CIP and SIP systems, sanitary design veriﬁcation, and sterile operation are critical parts of fermenter validation — if a facility can’t satisfy validation protocols, it can’t operate. In view of the FDA’s increased emphasis on aseptic and sterile bioprocess validation, coupled with the motivation of individual companies to develop cost-effective facilities that can be validated, there is a need in bioprocess engineering to incorporate validation considerations in initial designs.

STERILIZATION Rather than being viewed as ancillary, CIP and/or SIP systems should be designed as integral fermenter components with validation maintained from the start. If clean-

PH2270_book.fm Page 169 Tuesday, May 16, 2006 9:37 AM

Fermenters

169

ing and sterilization procedures do not validate after a fermenter has been built and is ready to begin operation, system disassembly, redesign, and/or component reconﬁguration could be necessary — time-consuming and costly remedies. Discovering a contaminated batch after production could be expensive, necessitating extensive downstream processing or perhaps even dumping of the entire batch. Much of today’s leading-edge sterilization technology came about through the proprietary developmental efforts of biopharm producers and fermenter manufacturers. For example, years ago, threaded fittings were almost exclusively used in clean steam systems. When it became apparent that these ﬁttings could not be adequately sterilized, they were replaced by sanitary fittings. Then came the perceived need for higher-quality welds between the ﬁttings and the tubing to eliminate dead spaces, ridges, and crevices. Since the FDA examines the integrity of the whole system, the facility must be not only designed to support the bioprocess, but validated as an integrated system. For example, if the facility has an excellent steam distribution system to a process room with the wrong ﬂoor ﬁnish and poor HVAC conditions, the installation won’t be validated. Most industrial fermentations are carried out in pure culture. If foreign organisms proliferate in the medium or on any of the equipment, culture organisms are forced to compete with the contaminating organisms for nutrients. Foreign organisms also typically produce metabolic byproducts that limit a producing culture’s growth. Therefore, before starting a fermentation, the medium, additives, and all equipment must be completely sterile, and aseptic conditions must be maintained. Steam sterilization is desirable because it not only destroys contaminants but cleanses the system if it reaches all system contact sites, process liquids, and gases. In addition, there must be a mechanism for collecting the contaminated condensate that forms as the steam condenses, and a regulatory system is required to verify that all critical areas have been sterilized at a sufﬁciently high temperature for an adequate time. Continuous steam ﬂow during processing also provides a sterile barrier, with process liquids on the one side and steam on the other so that in the event of leakage at a harvest port, for example, the steam would minimize contamination. Medium and equipment sterilization are generally completed by destroying all contaminating organisms with either moist or dry heat, chemical agents, ultraviolet radiation, ultrasound (which mechanically disrupts contaminating organisms), or by ultraﬁltration.

MEANS

STERILIZATION

Heat: Heat is the most widely-used means of sterilization, and it is employed for both medium and sterilizable fermenter parts. It can be applied as either dry or moist (steam) heat. Moist heat is more effective for sterilization for the reason that bacterial heat resistance is greater in the dry state, resulting in death kinetics that are much lower for dry cells than for moist. Conduction of heat in dry air is also less rapid than in steam, so dry heat is usually used only for sterilizing glassware and other heatable solid materials. Under pressure, steam temperature can increase signiﬁcantly above the boiling point of water. Laboratory autoclaves are usually operated at a pressure of

PH2270_book.fm Page 170 Tuesday, May 16, 2006 9:37 AM

170

Producing Biomolecular Substances

around 30 psia, corresponding to 121°C, where bacteria and their spores are rapidly destroyed. Antiseptic chemicals: The oxidizing or alkylating properties of various antiseptic chemicals typically kill microorganisms; however, they are not used for medium or fermenter sterilization because residual antiseptics typically inhibit fermentation. Some of the major antimicrobial chemical agents are phenol and phenolic compounds (phenol, cresol, orthophenylphenol), alcohol (ethyl, methyl), halogens (iodine, hypochlorites, chloramines), detergents, dyes, quaternary ammonium compounds, acids, alkalies, and gaseous sterilizing chemicals (i.e., ethylene oxide, propiolactone, formaldehyde, etc.). Ultraviolet (UV) light: UV is absorbed by cells causing DNA damage and consequent cell death. The best bactericidal efﬁciency is at wavelengths around 265 nm, although UV has little ability to penetrate nonbiotic matter and its use has been limited to reducing microbial populations in areas such as clean rooms or sterile chambers where asepsis must be maintained. Ultrasound: Ultrasound of sufﬁcient intensity can disrupt microorganisms, although the technique is usually employed for extracting cellular components rather than for sterilization. Ultrafiltration: Ultraﬁltration can be effectively used for removing microorganisms from the air or from other gases; with liquids, ultraﬁltration is used primarily for thermolabile products such as sera and enzymes. However, since moist heat is the most cost-effective and efﬁcient sterilization method for fermentation, this section only covers cell death kinetics and sterilization by steam. Medium sterilization within the fermenter vessel is performed in the batch mode by direct steam sparging, by electrical heaters, or by constant-pressure steam inside the vessel walls. Sterilization cycles are composed of heating, holding, and cooling stages.

THE DEL FACTOR (∇) The total Del factor (i.e., a measure of the size of a task to be accomplished) required for a complete sterilization cycle should be equal to the sum of the Del factor for heating, holding, and cooling: ∇ total = ∇ heat + ∇ hold + ∇ cool

(8.70)

The values of ∇heat and ∇cool are predetermined by the methods employed for the heating and cooling stages, and the value of ∇hold is determined by the length of the controlled holding period. Holding time can be estimated by: (1) calculating the total sterilization standard factor, ∇total, and (2) measuring the temperature versus time proﬁle during the heating, holding, and cooling stages of the sterilization cycle. If experimental measurements are not practical, theoretical equations for heating and cooling can be employed, which are of linear, exponential, or hyperbolic form depending on the mode of heating and cooling. For batch heating by the direct sparging of steam into the medium, the hyperbolic form should be used:

PH2270_book.fm Page 171 Tuesday, May 16, 2006 9:37 AM

Fermenters

171

T = T0 +

Hmst c( M + m s t )

(8.71)

For constant-rate heat ﬂow, the linear form should be used: T = T0 +

qTt cM

(8.72)

For isothermal heating, the exponential form should be used: ⎛ UAt ⎞ T = TH + (T0 − TH )exp ⎜ − ⎝ cM ⎟⎠

(8.73)

and for batch cooling using a continuous nonisothermal heat sink (e.g., passing cooling water through the vessel jacket), the linear form should be used: ⎧⎪ ⎡ ⎛ UA ⎞ ⎤ mct ⎫⎪ T = TC0 + (T0 − TC0 )exp ⎨ ⎢1 − exp ⎜ − ⎬ ⎝ mcC ⎟⎠ ⎥⎦ M ⎭⎪ ⎩⎪ ⎣

(8.74)

(3) then plotting the values of kd as a function of time, (9) integrating the areas under the kd-vs.-time curve for the heating and cooling periods to estimate ∇heat and ∇cool, respectively (if using theoretical equations integrate numerically after substituting in the proper temperature proﬁles): ∇ = ln

n0 = n

∫

kd dt = kd0

⎛

Ed ⎞

∫ exp ⎜⎝ − RT ⎟⎠ dt

(8.75)

and (5), calculating the holding time from the expression: t hold =

∇ hold ∇ total − ∇ heat − ∇ cool = kd kd

(8.76)

CONTINUOUS STERILIZATION Continuous sterilization simpliﬁes production planning, allows maximum facility utilization and minimum delay, provides reproducible conditions, can be operated at higher temperatures than batch sterilization (190°C vs. 121°C) so that the sterilization time can be shortened to only one or two minutes), requires less steam and cooling water since it recovers heat from the sterilized medium, and is less laborintensive and easier to automate. A continuous sterilization system consists of three main units: (1) heating, (2) holding, and (3) cooling. Heating is generally categorized into two types — direct heating (e.g., steam injection) and indirect heating (e.g.,

PH2270_book.fm Page 172 Tuesday, May 16, 2006 9:37 AM

172

Producing Biomolecular Substances

shell-and-tube or plate-and-frame heat exchangers). Direct heating is generally more effective because there is no barrier between the medium and the heat source. Steam injection rapidly heats the medium to the ultimate sterilization temperature, so sterilization during the duration of heating is negligible. For indirect heating, the plate-and-frame heat exchangers are generally more effective for heat transfer than the shell-and-tube exchangers because of their larger heat-transfer areas, and they are also better for sterilizing high-viscosity media since plate-and-frame heat exchangers are limited to lower pressures (less than 20 atm) because of their weaker structural strength. Sterilization can be expressed as: In – Out – Killed by Sterilization = Accumulation

(8.77)

with accumulation equal to zero at steady state. Cell input and output both have a bulk ﬂow and an axial diffusion phase, and efﬁciency increases with increasing particle diameter and/or air ﬂow velocity. Thermal death of microorganisms is described by ﬁrst-order kinetics: dn = − kd n dt

(8.78)

where kd is the speciﬁc death rate that depends on both the species and physiological form (e.g., the value of kd for bacterial spores at 121°C is on the order of 1 min–1 and for vegetative cells it can vary from 10 to 101 min–1 depending upon the particular organism). Exponential decrease in the cell population is shown either by:

n =− n0

∫ k dt

(8.79)

or, by ⎛ n − n0 exp ⎜ − ⎝

⎞

∫ k dt ⎟⎠ d

(8.80)

The dependence of the speciﬁc death rate kd on temperature follows Arrhenius: ⎛ E ⎞ kd = kd0 exp ⎜ − d ⎟ ⎝ RT ⎠

(8.81)

where Ed is activation energy obtained from the slope of the (ln)kd-vs.-1/T plot (e.g., the activation energy of E. coli is 127 kcal/gmole and that of Bacillus stearothermophilus is only 68.7). Thus, the sterilization design standard (∇), can be deﬁned as a Del factor that increases as the ﬁnal number of cells approaches null. However, if even one organism survives, the entire culture is still contaminated. The Del factor

PH2270_book.fm Page 173 Tuesday, May 16, 2006 9:37 AM

Fermenters

173

to reduce the number of contaminating microorganisms to zero is inﬁnity (i.e., it is theoretically impossible to ensure the total destruction of all viable cells). Practically, however, the ﬁnal number of contaminating microorganisms is expressed as a fraction equal to the statistical probability of contamination (e.g., with n = 0.001, the chance for a contaminant surviving the sterilization is 1 in 1,000), therefore, the Del factor to reduce the number of cells in a fermenter from 1,010 viable microorganisms to 0.001 is: ∇ = ln

1010 = 30 0.001

(8.82)

with a sterilization unit design based on the calculated Del factor.

STEAM STERILIZATION Steam sterilization is most commonly used in SIP systems, and sterilization system design must ensure that the steam comes in contact with all sites exposed to process materials and gasses, culture material, or product. All inlet ports, outlet ports, supply lines, harvest lines, sensors, regulators, the reactor vessel itself, and any stretch of piping that carries materials critical to the process must be sterilized. Steam should reach each of these sites at a temperature and pressure adequate to destroy contaminating organisms and must remain at these levels for a speciﬁc period of time. Process engineers generally design STFs, bioreactors, and their associated system components to be compatible with steam sterilization (e.g., the pitch of the piping and the positioning of ports, feed lines, and regulatory devices can affect their accessibility to the steam), and as previously discussed, it is essential that the designs eliminate dead spaces, ridges, and crevices where microorganisms can evade sterilization and ﬂourish.

STEAM SYSTEM COMPONENTS Steam system components consist of the steam generator, ﬁlters, steam traps, temperature and pressure gauges and regulators, isolation valves, and condensate recovery system. The steam generator boils the feed water to produce steam; bafﬂes within the generator (usually made of stainless steel mesh) remove any remaining water droplets and ﬁlters remove large contaminants such as chemicals in the feed water, rust, or pipe scale; steam traps keep the steam dry once it is in the process line by continually evaporating condensate as it accumulates; pressure-reducing valves maintain energy-efﬁcient operation; and isolation valves control steam access to the fermenter vessel, distribution lines, accessory vessels, and ports. In continuous culture, valves enable isolation of the fermenter vessel while the inlet and outlet lines are sterilized. Recovery pumps facilitate reuse of the hot condensate in a heat exchanger. Steam sterilization has applications in a variety of industries, although in FDA- and/or USDA-regulated industries (e.g., pharmaceuticals, dairy, food, and brewing), the steam must be clean to minimize the possibility of product contamination. Unfortunately, the terms plant steam, clean steam, and pure steam have different meanings for process engineers from different engineering schools, from

PH2270_book.fm Page 174 Tuesday, May 16, 2006 9:37 AM

174

Producing Biomolecular Substances

different companies, and even for those who attended different fermenter or bioreactor manufacturer training programs. This lack of standardization has plagued the regulated industries, and with different equipment manufacturers using different deﬁnitions relative to their systems’ material, design, and size speciﬁcations, there is a somewhat limited compatibility between systems.

PLANT

UTILITY STEAM

Plant or utility steam is mainly used to boil feed water in a steam generator and should not come in contact with sites that are exposed to cells, process liquids, gasses, or product. This type of steam also ﬂows through heat exchangers, around vessels to warm liquids, or around gasses being added to the reaction vessel. Plant steam, produced from common tap water, generally contains particulate matter and other contaminants and is usually ﬁltered to prevent large impurities from damaging the heat exchanger. Plant steam condensate can be recirculated to produce more utility steam. Plant steam is the heat source for generating clean steam or pure steam, which differ according to the purity of the feed water. For clean steam production the feed source is deionized water in which organic contaminants and bacteria may still be present. Although heating deionized water kills bacteria, bacterial pyrogens and other organic contaminants can still remain. In bioprocesses where a certain level of pyrogen is acceptable, it may be more cost-effective to remove pyrogens and other organic contaminants during downstream processing. Clean steam/pure steam (i.e., organic-contaminant/pyrogen-free steam) systems employ water produced from deionized water that subsequently undergoes reverse osmosis and distillation. Pure steam contains no metals, bacteria, or pyrogens. Typical applications of pure steam include mammalian cell culture (in which cell growth is highly sensitive to any level of impurity), and parenteral pharmaceutical manufacture (where even trace levels of pyrogen could provoke a severe immunologic response).

CLEAN STEAM

PURE STEAM

Using clean steam or pure steam (usually referred to as clean steam) places speciﬁc demands on both the fermentation and sterilization systems. Clean steam is ionhungry (and therefore highly corrosive) and any piping, vessel linings, valves, ﬁttings, or traps that come in contact with this steam must be corrosion-resistant since corrosion introduces impurities and leaves rough surfaces that provide ideal breeding sites for microorganisms. Stainless steel is the material of choice for piping and associated components since it does not react with clean steam. Grade 316 stainless steel is acceptable for areas not exposed to welding. During welding, however, a carbide precipitate forms and subsequently, strong acids or bases dissolve the carbide, leaving striations in the metal. Grade 316L (low carbon) stainless steel minimizes this problem. Although steam generator design is fairly consistent between the various industries, generators used with clean steam systems are made of stainless steel and require a high polish. Threaded connections should be avoided, and all welded sites and internal surfaces in contact with process materials must be smooth and highly polished. Ball valves, rather than diaphragm valves, are used

PH2270_book.fm Page 175 Tuesday, May 16, 2006 9:37 AM

Fermenters

175

wherever there is exposure to product and process liquid, since ball valves are easier to clean and are less susceptible to extremes of temperature and pressure. Vaponics (Rockland, MA) and Electro-Steam Generator Corp. (Alexandria, VA) manufacture clean steam generator units.

STEAM SYSTEM DESIGN Steam system design usually relies on gravity for complete drainage of condensate. Steam traps are designed to detect condensate in the process line, open in order to drain the condensate, and close again when they sense steam. Traps should be located at regular intervals along the piping and at the end of distribution lines since steam traps, by eliminating water from the conduit, help remove potential contaminants and also prevent water hammer that could loosen the lines. Thermodynamic steam traps sense velocity differences between steam and condensate, while thermostatic traps respond to temperature differences. For example, the internal bellows of a thermostatic trap expands to capacity at steam temperature, closing the trap; when condensate is present the trap temperature falls and the bellows contracts, allowing condensate to drain into a collection vessel. Opening and closing of the trap is a continuous process during steam sterilization. Ideally, the condensate drains near the steam temperature so as to ensure minimal backup into the steam line. It is important to keep in mind that the condensate produced in a clean steam system is not sterile, should not be reintroduced into the system, and is generally transferred to a kill tank prior to disposal. Spirax-Sarco™ and Nicholson™, two major producers of steam traps for the biotechnology and biopharmaceutical industries, manufacture traps composed of 316L stainless steel speciﬁcally designed for clean steam applications that vary only with respect to size, surface ﬁnish, and type of connections — threaded, butt-welded, ﬂanged, or sanitary. Sanitary connections incorporating quick-release Triclover™ clamps simplify removal and off-line steam trap sterilization, thereby reducing downtime. Less expensive sealed traps are preferred for sites requiring limited take-down. Sanitary ﬁttings eliminate threads and crevices and assure a smooth, polished ﬁnish, minimizing the potential for contamination. Both Spirax-Sarco™ and Nicholson™ offer traps with sanitary ﬁttings that cost approximately six times more than standard traps, and with approximately 15–20 traps required on a pilot-scale fermenter, this can dramatically increase its cost. In designing fermenters, for example, engineers cut costs by installing steam traps off the direct process lines wherever possible, with valves separating the traps from contact with process liquids. Since the valves are only open during steam sterilization, they prevent condensate from backing up into the process lines, and since there is no necessity for maintaining sterility downstream of the valve, standard, threaded traps are generally used there. For biomanufacturing applications in which steam traps are placed directly on the process line, they can be attached using sanitary TriClamp™ connections. Although the sanitary connections are substantially more costly, they provide easy access to the trap for off-line cleaning, direct visual inspection, and maintenance. Steam traps take a lot of abuse, and with the Tri-Clamp design they can be easily removed from service, broken-down, and examined to verify cleanliness.

PH2270_book.fm Page 176 Tuesday, May 16, 2006 9:37 AM

176

Producing Biomolecular Substances

SUPPLEMENTARY FERMENTER EQUIPMENT All fermenter manufacturers offer supplementary equipment such as air pumps, nutrient and medium pumps and reservoirs, shakers, autoclavable ﬁttings, valves, probes, etc., which are often readily interchangeable so that they allow choices based on culture requirements. In addition, some manufacturers such as Virtis™ sell addon packages comprised of a sidearm adaptor, product reservoir, medium reservoir, and ﬂow control device, to convert their batch STFs into either CSTFs or to CSTFs with cell recycling (chemostats). To retroﬁt its laboratory STFs Chemap offers its Chemcell™ system, consisting of a head plate with a vibrating meshed cylinder through which the harvested material, but not the cells, can pass for continuous culture operation. Oxygen, or a controlled gas mixture, is efﬁciently infused into the medium by the meshed cylinder’s vibrations, and frothing is suppressed. Three ﬁltered gas ports, an inlet for headspace aeration, an inlet for Chemcell™ aeration, and an outlet are also included.

OTHER FERMENTER CONSIDERATIONS Two other considerations worthy of mention are: (1) the immobilized cell fermenter in which whole cells are trapped in some matrix (i.e., gel or hollow ﬁbers), which keeps the cells in place and permits the entry of oxygen, nutrients, and the removal of end product, or where the cells are immobilized on ceramic or polymer beads; or (2) the necessity of high containment levels. When working with genetically engineered microorganisms, stringent safety measures are a legal requirement. The leader in this ﬁeld is Chemap, whose units are built as fully integrated systems — preassembled, and pretested. Failsafe mechanisms and alarms are incorporated wherever possible, ﬂanges and unions are provided with double o-rings, an absolute air ﬁlter system and air incinerator is provided, inoculation and sampling employ special valves, and foam detectors are redundant for security.

FERMENTER DESIGN CELL KINETICS In designing a fermenter to carry out a particular process, various factors are essential to achieve satisfactory results. Normally, fermenter productivity is expressed as quantity of product produced per unit of time and volume. If the inlet stream is sterile (CX = 0), the productivity of the cell mass is equal to C Xi / μm, which is equal to the slope of the straight line OAB of the CX-vs.-μm curve. As the slope of the line increases, productivity increases, and the slope will have its maximum value when tangent to the CX-curve. Thus, maximum productivity is equal to the slope of the line OC, and is attained when μm = (μm)opt, and CX = (CX)opt. The operating conditions for maximum CSTF productivity can be graphically approximated using the 1/rX-vs.-CX curve, where maximum productivity is accomplished with minimum residence time. Cell productivity for a CSTF with a sterile feed is:

PH2270_book.fm Page 177 Tuesday, May 16, 2006 9:37 AM

Fermenters

177

μ CC CX = rX = max S X μm K S + CS

(8.83)

Optimal cell concentration for maximum productivity can be expressed as: CSi α +1

CS ,opt =

(8.84)

Furthermore, optimal residence time can be expressed as: τ m ,opt =

α μ max (α − 1)

(8.85)

As discussed earlier, the residence time required for a fed batch STF to reach a particular cell concentration level is: τb = t0 +

∫

C X0

dC X rX

(8.86)

where t0 is the time required to reach exponential growth, and the area under the 1/rXvs.-CX curve between CX and CX0 is equal to τb – t0. Now, since the 1/rX-vs.-CX curve is U-shaped, the following two conclusions can be determined for a single fermenter unit: (1) the most productive fermenter system is a CSTF operated at a cell concentration where the value of 1/rX is minimum, since it requires the smallest residence time; and (2) if the ﬁnal cell concentration to be reached is in the static growth stage, the batch STF is a better choice than the CSTF because the batch residence time required is smaller than that of the CSTF. Whether or not one chooses to select the minimum fermenter system for maximum productivity depends on the shape of the 1/rX-vs.-CX curve and the speciﬁc bioprocess requirements (i.e., ﬁnal product conversion). If ﬁnal cell concentration is less than C Xopt , and since two CSTFs in series require more residence time than one, a single CSTF is better than two CSTFs connected in series. Now, if the ﬁnal cell concentration is much larger than C Xopt , the best blend of two fermenters for a minimum total residence time is a CSTF operated at C Xopt followed by either a packed-bed fermenter (see Figure 8.18) or a chemostat. A CSTF operated at C Xopt followed by another CSTF connected in series is also better than one CSTF. Cellular productivity in a CSTF increases with increase in the dilution rate and reaching a maximum value (see Figure 8.9). But, if the dilution rate is increased beyond this point, productivity will decrease abruptly and the number of cells will start to decrease because their generation rate is less than that of fermenter cell washout (i.e., the productivity of the fermenter is limited by the loss of cells through the outlet stream). One way to improve this is to recycle the culture, separating the cells from the product stream using a cross-ﬂow ﬁlter unit (see Figure 8.19).

PH2270_book.fm Page 178 Tuesday, May 16, 2006 9:37 AM

178

Producing Biomolecular Substances

Fresh Medium

Waste Medium and Product

FIGURE 8.18 Packed-bed fermenter.

Filtrate

Recirculating Pump Fresh Medium

Bleed

FIGURE 8.19 CSTF with cell recycling.

The high cell concentrations maintained by cell recycling increases productivity since growth rate is proportional to cell concentration. However, there is a limit to increasing cell concentration to increase productivity because there is a point where nutrient transfer is decreased below the necessary minimum due to overcrowding and cell aggregation. Also, maintaining extremely high cell concentrations is not practical because ﬁlter units typically fail at the higher cell concentrations. And, if all the cells are recycled back into the fermenter, cell concentration increases continuously with time and steady state is never reached. Thus, to operate a CSTF with cell recycling in steady state requires a bleed stream. The material balance for cells in a steady state CSTF with cell recycling is:

βD =

β =μ τm

(8.87)

It should be noted that the actual rates of ﬂow in and out of the ﬁlter units are not important as far as overall material balance is concerned, and in a CSTF with cell recycling, the bleed ratio is deﬁned as:

PH2270_book.fm Page 179 Tuesday, May 16, 2006 9:37 AM

Fermenters

179

β=

B F

(8.88)

In this case D instead of D is equal to the speciﬁc growth rate, and when β = 1, the cells are not recycled. Therefore, D = β, and if cell growth is expressed by Monod: CS =

βK S τ mμ max − β

(8.89)

which is valid when μmμmax > β, cell concentration in the fermenter can therefore be calculated as: CX =

YX / S (CSi − CS ) β

(8.90)

Various other types of fermenters have been designed and tested to improve either the disadvantages of the stirred-tank (i.e., high power consumption and shear damage), or to meet the speciﬁc requirements of a particular fermentation process (e.g., better aeration, effective heat removal, cell retention, immobilization of cells, etc.). Fermenters are usually classiﬁed by vessel type, such as tank or column, and with or without pressure loops. Tank and column fermenters are cylindrical vessels generally distinguished by their height-to-diameter ratios (H/D): H/D < 3 for the tank fermenter H/D > 3 for thhe column fermenter Tank or column fermenters come equipped with either internal or external liquid circulation loops. Another means of classifying fermenters is by their manner of mixing (e.g., compressed air, mechanical agitation, external liquid pumping, etc.). Representative fermenters of the three basic fermenter types and each of the above categories are listed in Table 8.3; their advantages and disadvantages are listed in Table 8.4.

BUBBLE-COLUMN FERMENTER The simplest fermenter is the bubble-column type, typically a long cylindrical vessel with a sparging device at the bottom. The contents are mixed by the rising bubbles, which also provide the oxygen supply (see Figure 8.20). Without moving parts, it is quite efﬁcient in transferring oxygen per unit of energy consumption. And, as the cells settle out, higher concentrations can be maintained in the lower portion of the column without a separation device. Bubblecolumn units are, as a rule, limited to aerobic fermentations. The rising bubbles do

PH2270_book.fm Page 180 Tuesday, May 16, 2006 9:37 AM

180

Producing Biomolecular Substances

(a)

(b)

(c)

Air

FIGURE 8.20 Column fermenters; (a) bubble column, (b) stirred bubble column, (c) bubble column with external loop.

not provide enough mixing for optimal growth, and only the lower part of the column attains a high cell concentration, which results in rapid initial fermentation that noticeably slows and ultimately produces less desirable products. And, as cell concentration increases, progressively higher air ﬂow rates are required to maintain cell dispersion that can cause excessive foaming and high bubble retention in the column, thereby decreasing productivity. Additionally, since the bubbles rising in the column rapidly coalesce, oxygen transfer is somewhat decreased.

STF DESIGN ADDITIONS Design additions have attempted to overcome these weaknesses. Sieve plates have been installed for effective gas/liquid contact and dispersion of coalesced bubbles. A bubble-column fermenter can be equipped with agitators and, to enhance mixing without increasing shear on the culture, the media can be recirculated with an external pump. In addition to standard versions, STFs sometimes have internal or external liquid circulation loops with or without mechanical agitation. Depending on how the media is circulated, STFs by and large can be classiﬁed in four categories: (1) air-lift (see Figure 8.21), (2) pressure cycle, with an external loop [Figure 8.22(a)], stirred-tank, with an internal loop [Figure 8.22(b)], and pressure cycle, with an internal loop [Figure 8.22(c)]. Analogous to the mechanism of the bubblecolumn unit, the air-lift fermenter circulates media by means of sparged air, creating density differences between the foam-rich part of the media and the denser bubbledepleted part. A pressure-cycle fermenter is an air-lift unit with an outer circulation loop especially developed for aerobic fermentation requiring heat removal. Both medium and air are introduced into both the upper and lower parts of the loop. A heat exchanger that cools the liquid medium is integral to the loop, and the rising air provides both oxygen for the culture and a gentle agitation. Media circulation and mixing can sometimes be enhanced by installing a marine impeller or by externally circulating the liquid by using a pump. However, adding a propeller or pump tends to diminish the two real advantages of air-lift fermenters: their gentle handling of delicate cells and their effectiveness.

PH2270_book.fm Page 181 Tuesday, May 16, 2006 9:37 AM

Fermenters

181

Steam to Sterilize Sterile Medium Filter

Inoculum

Exhaust Gasses

Filter Air Inlet

FIGURE 8.21 Air-lift fermenter. (b) (c) (a)

Air

FIGURE 8.22 Tank fermenters; (a) pressure cycle with external loop, (b) stirred-tank with internal loop, (c) pressure cycle with internal loop.

BASIC PHYSICAL ELEMENTS The initial phase of fermenter design must establish the physical volume necessary for the equipment and whether the fermentation will operate with a single fermenter on a batch basis, multiple fermenters on a batch basis, or in a continuous culture system. To obtain the required volume for the vessel, the time of the fermentation cycle can be calculated by using kinetic models of the growth rate matched to the required production rate. The time cycle and the availability of the vessel for processing can then be used to determine the required fermenter volume (e.g., a single fermenter employing a speciﬁc culture, working an eight-hour day, ﬁve days a week should be a great deal larger than a single fermenter used for the same culture, working a 168-hour, seven-day week in order to generate the same amount of product). In general, transport mechanisms for the growth of microorganisms can be modeled. In the simple case of aerobic bacteria, the major problem is to get oxygen to the organism in sufﬁcient quantities to enable exponential growth. The types of resistance to the mass transfer from the medium to the cell can be categorized as:

PH2270_book.fm Page 182 Tuesday, May 16, 2006 9:37 AM

182

Producing Biomolecular Substances

(1) diffusion from bulk gas to the gas/medium interface; (2) solution of the gas in the medium at the interface; (3) diffusion of the dissolved gas into the medium; (4) transport of the dissolved gas to the immediate region of the cell; (5) diffusion through the static region of the medium surrounding the cell; (6) diffusion into the cell; and (7) consumption by the cell depending on its growth kinetics. The rate of mass transfer of a gas into a liquid can be written as: N A = kG (C1 − C2 ) = k L (C3 − C4 )

(8.91)

Here NA is the rate of mass transfer of gas A (kmol/s m2), C1 and C2 are gas concentrations in the gas phase, C3 and C9 are gas concentrations in the liquid phase, and kG and kL are gas and liquid mass transfer coefﬁcients [with dimensional resistance–1]. The simplest explanation of mass transfer is the twin-ﬁlm theory in which conditions at the gas/liquid interface can be represented by a plot of concentration against distance from the interface. This model enables the simple rate law to be written as: N A = kG ( pA − pI ) = k L (C I − C A ) where p is the partial pressure of gas in the gas mixture (e.g., oxygen in air); C is the concentration of gas dissolved in the liquid; the subscript I refers to interface conditions, and the subscript A refers to bulk conditions. The regions near the interface bounded by the interface and where the bulk concentration begins to change are known as boundary layers. The mass-transfer rate equation N A = kG (δP) = k L (δC ) assumes a straight-line relationship over the boundary layers. The mass-transfer coefﬁcients (kG, kL) are actually ﬁlm masstransfer coefﬁcients related only to conditions at the interface. Because it is difﬁcult to measure values of pI and CI, the overall mass-transfer coefﬁcients related to concentration and conditions in the bulk gas and liquid are used. Overall coefﬁcients are deﬁned by N A = K G ( pA − p*) = K L (C * −C A ) where p* is the partial pressure of the gas in equilibrium with a solution of the gas at concentration CA and C* is the equilibrium concentration of the gas in solution with a partial pressure of pA above the solution. Henry’s Law applies to equilibrium data for many gasses that are soluble in water (medium): p* = HCA where H is the Henry constant. If Henry applies, then the following relationships are valid: pA = HC*; p* = HCA; pI = HCI. These equilibrium relationships can be used to derive interrelationships between ﬁlm coefﬁcients and overall coefﬁcients. For sparingly soluble gases like oxygen (10 ppm @ 1 atmosphere pressure), the numerical value of H is quite large (at 30°C, H = 8.85 × 109 atmospheres/mol fraction). Thus, the most important resistance to mass transfer lies in the boundary layer. Because the rate of mass transfer NA is the rate per unit area perpendicular to the direction of transfer, the total mass transfer of gas is depicted by: rate = driving force/resistance

(8.92)

PH2270_book.fm Page 183 Tuesday, May 16, 2006 9:37 AM

Fermenters

183

option is not suitable for aseptic fermenter operation, except in the cases of packedbed fermenters or hollow-ﬁber bioreactors, and bubble formation is the primary aeration technique of aseptic fermentation. The most commonly used fermenter to perform aerobic fermentations is the mechanically agitated fermenter, using an agitator with a sparger ﬁtted below the impeller blades to admit air. Typically fermenter dimensions are expressed in terms of ratios so that, provided these ratios are maintained (independent of the volume of the fermenter), the fermenter scale up is simpliﬁed in terms of power requirements, aeration, agitator speed, etc., particularly if geometrical and dynamic similarities are maintained (see Batch Scale up and Continuous-Culture Scale up, below). Agitation keeps bubbles of air circulating through the medium. The faster the rotation speed of the impeller, the longer the bubbles take to pass through the medium in the fermenter before exiting, and the more time is therefore available for mass transfer. The overall mass transfer coefﬁcient (KL) for such vessels is related to the impeller diameter Di and the speed of revolution N for constant impeller design. Oxygen absorption into fermentation media is established by: ⎛Π ⎞ K Lα = KV = 0.002 ⎜ G ⎟ ⎝ V ⎠

0.7

VS0.2

(8.93)

where KV is the overall volumetric oxygen transfer coefﬁcient (1/s); ΠG/V is the power input per unit volume of media in a gassed system (W/m3), and VS is the superﬁcial air velocity based on the cross-sectional area of the fermenter (m/s). This expression only applies for a liquid height (L) to vessel diameter (Dτ) ratio of 1.0. For L/Dτ ratios between 2 and 9, values of KV are about 50% higher than those produced by the above expression. Total transfer rate is equal to KV (C*CA), and the C units will determine the KV units (e.g., if C is measured in kmol/m3, then KV is also in kmol/s m3 volume). There are two important criteria for maintaining cell growth in aerobic fermentations: (1) maintaining a dissolved oxygen concentration above the critical concentration required for exponential growth (typical values range from 0.003 to 0.05 mol/m3), and (2) supplying oxygen at a rate matching metabolic utilization. Although no air supply is necessary for anaerobic fermentation, the need for medium agitation applies to both aerobic and anaerobic fermentation in order to provide necessary dispersion functions for materials other than oxygen. When liquid is placed in a cylindrical vessel with a centrally mounted impeller and no bafﬂes, the main liquid motion is circular and overall mixing is poor. Bafﬂes increase mixing speed with a penalty of higher power requirements. To design an agitated fermenter it is important to predict the power requirements for aerobic fermentations since oxygen transfer is a function of power input per unit volume (ΠG/V). Practical problems in biotechnology and biochemical engineering fall into two categories: (1) the system behavior and material components are well deﬁned mathematically; (2) system behavior and material components are either only partly deﬁned or are interdependent and cannot be deﬁned in precise mathematical terms (the classic three-body problem). The ﬁrst type of problem can be solved by mathematical calculation even though the solution might require complicated procedures such as integrating simultaneous differential equations. An

PH2270_book.fm Page 184 Tuesday, May 16, 2006 9:37 AM

184

Producing Biomolecular Substances

example of this type of problem is in liquid sterilization (see Cleaning and Sterilization, above) where thermal-death kinetics are well deﬁned. It is also possible to express temperature proﬁles for heating and cooling mathematically, and a combination of these equations can enable a mathematical solution to be obtained for the value ∇. Problems of the second type, however, where no mathematical solution is possible, can still be solved empirically by dimensional analysis.

COMPLEX PHYSICAL ELEMENTS In order to use dimensional analysis, sufﬁcient information about the system’s physics must be known so that incorrect results will not develop if a key variable is omitted. The procedure is performed by initially determining all the possible behavioral relationships of the variable of interest: Y = function (other variables; physical properties, etc.) = K(AaBbCc … etc.), where K is a dimensionless constant. Next, the variables and physical properties are expressed in terms of basic dimensions (e.g., mass, length, time, etc). For instance, velocity (m/s) can be expressed dimensionally as length/time and acceleration (m/s2) as length/time2. After rewriting the basic expression in terms of dimensions, the exponents of the dimensions (a, b, c, … etc.) are equated, and the resultant simultaneous equations are then solved. Thus, for all practical purposes, ⎛ g⎞ P⎜ ⎟ ⎝l⎠

1/ 2

= 2π

(8.94)

or ⎛l⎞ P = 2π ⎜ ⎟ ⎝ g⎠

1/ 2

(8.95)

Dimensionless groups are collections of physical properties and variables such that irrespective of the units used for the calculation the value of the group (number) will always be the same provided consistent units are used (i.e., all SI, all metric, all engineering, etc.). Bioprocess engineering problems associated with liquid ﬂow, mass transfer, heat transfer, mixing etc., can only be practically solved using dimensional analysis. Not all relationships use dimensionless groups, however. An example of a dimensional expression is Equation 8.93, where the groups (ΠG /V) and (VS) are not dimensionless. In such cases, it is most important that the deﬁned units are strictly adhered to; otherwise a wrong result will be obtained.

KINETIC MODELING OF THE FERMENTATION PROCESS AGITATION DIMENSIONAL ANALYSIS Dimensional analysis can be applied to agitation to obtain a relationship of the power required (π) with variables and physical properties:

PH2270_book.fm Page 185 Tuesday, May 16, 2006 9:37 AM

Fermenters

185

Π = KDia N bρcμ d g e

(8.96)

where DI is the impeller diameter, N is the speed of rotation, D is the liquid density, μ is the liquid viscosity, and g is the acceleration due to gravity. A relationship containing three dimensionless groups is obtained: ⎛ Di2 N ρ ⎞ ⎛ Di N 2 ⎞ Π k = ⎜⎝ μ ⎟⎠ ⎜⎝ g ⎟⎠ ρN 3 Di5

(8.97)

where Π Power number (Po) (ρN 3 Di5 )

(8.98a)

Di2 N ρ Reynolds number (Re) μ

(8.98b)

Di N 2 Froude number (Fr) g

(8.98c)

This relationship applies to a particular agitator in a particular fermenter vessel with a particular bafﬂe conﬁguration, although the same relationship can be used for different size vessels and agitators provided that there is both geometric and dynamic similarity. For geometric similarity, the following ratios should be the same for both vessels: Dτ / Di vessel diameter/impeller diameter

(8.99a)

L / Di media height/impeller diameter

(8.99b)

I / Di impeller height from bottom/impeller diameter

(8.99c)

W / Dτ baffle width/vessel diameter

(8.99d)

For dynamic similarity there are two possible methods: (1) the scale up can be based on establishing that the Reynolds number (Re) is the same in both size fermenter vessels; with constant physical properties this means Di3N of fermenter vessel #1 should equal Di3N of fermenter vessel #2; this method is suitable for anaerobic fermentation processes; and (2) for aerobic processes, since we are dealing with a

PH2270_book.fm Page 186 Tuesday, May 16, 2006 9:37 AM

186

Producing Biomolecular Substances

two-phase system (media/gas bubbles), a better scale up method employs the Weber number (We) with constant physical properties so that Di3N for fermenter vessel #1 equals Di3N for fermenter vessel #2. For a marine impeller at a Reynolds number greater than 5,000, for example, the value of the power number (Po) would be 5.8; this only applies to a single impeller placed at distance Di from the base of the fermenter vessel (i.e., I/Di = 1). For multiple impellers: Po = J(Po)1, where J is the number of impellers and (Po)1 is the power number for a single impeller. In aerobic operation, the power requirement is less if air is fed to the underside of the agitator (gassing system). General correlations covering a single impeller are: ⎛ Po 2 NDi3 ⎞ (Po )G = 0.354 ⎜ ⎝ Q 0.56 ⎟⎠

0.45

(Po )G = 0.7 exp(−0.9Q )

(8.100a)

(8.100b)

where N is the impeller speed (rev/s), Di is the impeller diameter (m), and Q is the volumetric air ﬂow rate (m3/s). The subscripted G applies to a gassed system. These correlations are dimensional, can be applied to large-scale fermenters, and only deﬁned units are used. For calculation of the mass transfer coefﬁcient the power requirement per unit volume (PG/V) is also required. In designing an aerobic fermenter it is essential to determine the ungassed power requirement, otherwise an under-speciﬁed motor could burn out if the air ﬂow is stopped or fails while the system is running. Unlike most chemical processes that frequently operate at high temperatures, bioprocesses usually function best at near-ambient temperatures. Nevertheless, heat transfer is still important and both the application and removal of heat from any given segment of the bioprocess must be considered, since heat transfer here is a deliberate attempt to alter or regulate any temperature changes generated by the bioprocess. Isothermal heat transfer is important in controlling the fermenter environmental temperature and for processes such as cell disruption that generate large quantities of heat. Biocatalysts also have a well-deﬁned temperature range over which they are active. If the temperature is too low, biological activity could be insigniﬁcant, and if it is too high for an extended period of time, thermal degradation could result. For organisms that normally grow near ambient, as well as their enzymes, this deﬁned temperature range is usually below 90ºC. Large fermenters often show net heat generation due to the friction of rotating impellers, metabolic activity, or exothermic reactions. In all such circ*mstances the fermenter requires cooling. In contrast to the requirements of deliberate temperature alteration, the rate of heat transfer during isothermal operations is usually quite low. Thus, the size of the heating/cooling system in the fermenter will be inﬂuenced by its predicted uses — larger for rapid heating, smaller for constant temperature maintenance. There are two ways of transferring heat to and from agitated vessels: (1) using jackets, and (2) using internal coils. The heat transfer coefﬁcient at the inside vessel wall is:

PH2270_book.fm Page 187 Tuesday, May 16, 2006 9:37 AM

Fermenters

187

⎛ D2 N ρ ⎞ hi Dτ = 0.36 ⎜ i ⎟ k ⎝ μ ⎠

0.67

⎛ μ⎞ (Pr)0.33 ⎜ ⎟ ⎝ μS ⎠

0.14

(8.100c)

where Dτ is the vessel’s inside diameter, Di is the agitator diameter, N is the agitator speed in revolutions, k is the liquid’s thermal conductivity, and Pr is the liquid’s Prandtl number. With jackets either ﬁtted or integral to the fermenter vessels, the surface area per unit volume falls as the vessel’s diameter and volume increase. Once a certain volume has been exceeded the surface area becomes too small to remove fermentation-generated heat and the only alternative is to use internal coils. Fermentation-generated heat also dictates the maximum volume of a fermenter vessel that is suitable for jacketing. The heat-transfer coefﬁcient at the outside surface of a coil (h0) ﬁtted inside an agitated fermenter vessel is: ⎛ D2 N ρ ⎞ h0 Dτ = 0.90 ⎜ i ⎟ k ⎝ μ ⎠

0.67

⎛ μ⎞ (Pr)0.33 ⎜ ⎟ ⎝ μs ⎠

0.14

(8.101)

Note that the vessel diameter Dτ appears in the Nusselt number, not in the coil diameter. These expressions for jackets and coils can be applied for both laminar and turbulent conditions and are valid for Reynolds numbers over the range likely to be encountered (from Re 500 to Re 500,000). In the double-pipe annular heat exchanger with a hot liquid ﬂowing in the outer pipe to heat a colder liquid ﬂowing in the inner pipe, there are two ways to arrange the liquid ﬂow: (1) concurrently (parallel, in the same direction), or (2) counter-currently (parallel, in the opposite direction). Analysis of heat balances based on heat-transfer characteristics and thermal transfer between liquids can be carried out to demonstrate that: Q = UAΔt

(8.102)

Temperature difference (Δt) used is the log mean temperature difference (ΔtLM) as deﬁned by: Δt LM =

( ΔT1 − ΔT2 ) ln( ΔT1 / ΔT2 )

(8.103)

ΔT1 and ΔT2 are the terminal temperature differences. As illustrated, countercurrent operation yields a higher temperature difference compared with concurrent operation, and offers a number of advantages: (1) a higher value for ΔtLM means that the same task can be performed using a smaller heat exchanger [the area required (A) will be smaller if U is the same], (2) the temperature difference throughout the heat exchanger does not exhibit as large a variation as the concurrent ﬂow, and (3), the hot liquid could have vacated the heat exchanger at a lower temperature with countercurrent ﬂow. For practicality, heat exchangers are always operated counter

PH2270_book.fm Page 188 Tuesday, May 16, 2006 9:37 AM

188

Producing Biomolecular Substances

currently to get maximum advantage from the above points. Heat transfer to and from a fermenter vessel is somewhat different than from a normal heat exchanger. When heating a batch of medium in a fermenter the temperature of the contents continually varies and, therefore, the log mean temperature difference (ΔtLM) also varies. In the case of convection heat transfer with the contents initially at ti and the heating liquid temperature tS, it can be shown that: ⎛ UAθ ⎞ (t − t s ) = exp ⎜ − (ti − t s ) ⎝ mC p ⎟⎠

(8.104)

where t is the temperature at time t, m is the mass of the medium in the vessel, CP is the speciﬁc heat capacity of the medium, and U is the heat transfer coefﬁcient. When the fermenter contents are maintained at constant temperature the standard heat exchanger expression can be used. The log mean temperature difference (ΔtLM) is calculated using terminal temperature differences even though the vessel contents are at constant temperature. Unlike countercurrent heat exchangers, the exit temperature of the heating/cooling liquid in the fermenter jacket or coil should not approach the contents temperature closer than 5ºC.

KINETIC FERMENTATION MODELING So far, the kinetic models described have been unstructured, distributed models based on two false assumptions: (1) cells can be represented by a single component such as cell mass or cell number per unit volume, and (2) cellular mass is distributed uniformly throughout the culture. Unstructured, distributed models do not acknowledge changes in cell composition during growth and do not account for latent stages, sequential uptake, or changes in mean cell size during batch culture growth, although the unstructured, distributed models described by Monod can satisfactorily predict cell growth in many situations. Structured models, on the other hand, recognize the multiplicity of cell components and their interactions, and many such models have been created based on assumptions about speciﬁc cell components and their interactions. In order to create a simple structured model that can be used for pragmatic bioprocess scale ups, stipulate that a system is composed of either a single or multiple cells that do not contain abiotic material, and carry mass m on a dry basis at a speciﬁc volume vˆ . Postulate that there are c components in the cell and the mass of the zth component per unit volume of the system is Cˆ XZ. Also, postulate kinetic rate expressions for p reactions occurring in the system and the rate of zth component formed from the yth reaction per unit volume of the system is rˆX y,z . Then, assuming the speciﬁc volume vˆ is constant with time during batch cultivation, the change of the zth component in the system with respect to time can be expressed as: dCˆ XZ = dt

∑r

Xy,z

y =1

−

1 dm ˆ C XZ m dt

(8.105)

PH2270_book.fm Page 189 Tuesday, May 16, 2006 9:37 AM

Fermenters

189

where the second term of the right-hand side of Eq. 8.105 represents the dilution of intracellular components by biomaterial growth, since all variables denoted by circumﬂexes are intracellular properties. Since the structural models acknowledge cellcomponent multiplicity and multi-component intracellular interactions, the model should be expressed with intrinsic variables. The relationship between the cell growth rate and the kinetic rate expressions of all intracellular reactions is dCˆ X = dt

∑∑r

−

Xy,z

z =1 y =1

1 m ˆ CX m dt

(8.106)

Concentration terms can be expressed as mass per unit culture volume vˆ instead of mass per biotic system volume mv. Concentration based on culture volume can be expressed as: dC XZ mvˆ = dt V

∑r

Xy,z

(8.107)

y =1

Although the concentration terms are based on total culture volume, kinetic parameters still remain on a biotic-phase basis. One of the simplest structured models, the two-compartment model, is based on the following premises: (1) the cell comprises two basic compartments — a synthetic portion A, such as precursor molecules and RNA, and a structural portion B such as protein and DNA: dCˆ X =0 dt

(8.108)

(2) the synthetic portion A is fed by uptake from substrate S and the structural portion B is in turn fed from the synthetic portion: S→A→B

(8.109)

(3) the ﬁrst reaction rate in Eq. 8.109 is proportional to the product of substrate and cell concentration: rX1, A = k1CS Cˆ X

(8.110)

here CS is the mass of substrate per unit abiotic volume, v − mvˆ. The rate of the second reaction in Eq. 8.99 is proportional to the product of the concentrations of the synthetic portion and the structural portion: rX2 , B = − rX2 , A = k2Cˆ X A Cˆ X B

(8.111)

PH2270_book.fm Page 190 Tuesday, May 16, 2006 9:37 AM

190

Producing Biomolecular Substances

Transcribing Eq. 8.111 for each component, and substituting Eqs. 8.112 and 8.113 for the reaction rates: dCˆ X A 1 dm ˆ = k1CSCˆ X − k2Cˆ X A Cˆ X B − CX A dt m dt

(8.112)

dCˆ X B 1 dm ˆ C XB = k2Cˆ X A Cˆ X B − dt m dt

(8.113)

change in substrate concentration is shown by: dCS k = − 1 CS Cˆ X dt YX /S

(8.114)

where YX/S is the yield constant. Thus, Eqs. 8.100c, 8.104, 8.105, and 8.106 can be solved simultaneously to obtain the change of Cˆ X B , Cˆ X A , Cˆ X , Cˆ S , and m; (4) cell ˆ ˆ X B ) has doubled its initial value, division only occurs if the structural portion ( mvC and dividing cells apportion each component equally to their daughter cells. Therefore, the total number of cells in the system will be proportional to the structural portion, and the average cell mass is equal to the total cell mass divided by total cell number. Therefore:

Average mass of cell =

ˆ ˆX m mvC Cˆ α = ˆX ˆ ˆ XB C XB n mvC

(8.115)

This model can also predict the change:

C X = C X A + C X B′

(8.116)

Eqs. 8.117, 8.118, and 8.119 can be expressed with concentrations in terms of mass per unit culture volume: dC X A ⎛ V ⎞ ⎛ V ⎞ =⎜ ⎟⎠ k1CS C X − ⎜⎝ ˆ ⎟⎠ k2C X A C X B ˆ ⎝ dt V − mv mv

(8.117)

dC X B ⎛ V ⎞ kC C =⎜ ⎝ mvˆ ⎟⎠ 2 X A X B dt

(8.118)

PH2270_book.fm Page 191 Tuesday, May 16, 2006 9:37 AM

Fermenters

191

1 dCS =− dt YX /S

⎛ V ⎞ ⎟ k1CS C X ⎜⎝ V − mvˆ ⎠

(8.119)

where C X = (mvˆ / V )C X

(8.120a)

C X A = (mvˆ / V )Cˆ X A

(8.120b)

C X B = (mvˆ / V )Cˆ X B

(8.120c)

CS = [(V − mvˆ ) / V ]CS

(8.120d)

m = C XV

(8.120e)

and

The mass m is related to CX

The total number of cells is proportional to C X B and the average mass of a cell to C X / C XB . The simulation curve of a batch culture (see Figure 8.23) shows changes in mass (C X / C Xmax ), number (C X B / C X Bmax ), and size of cells (C X / C X B ), and substrate con-

Cell Concentration

Static Stage

Exponential-Growth Stage

Latent Stage

Time

FIGURE 8.23 Batch fermentation growth curve.

Death Stage

PH2270_book.fm Page 192 Tuesday, May 16, 2006 9:37 AM

192

Producing Biomolecular Substances

centration (CS /CS0) in dimensionless form, and illustrates the following: (1) during latent stage, cells grow in size but not in number; (2) during exponential growth, cells are largest; (3) during the static stage, cells no longer grow or divide. Even though this model provides many features that unstructured models are not able to predict, it requires only two parameters, which is the same number of parameters required for Monod.

STF BATCH SCALE UP SCALE-OF-AGITATION METHOD A practical, yet scientiﬁcally sound, means of performing elementary scale up of standard STFs is based upon the scale-of-agitation method. In the STF system, the primary scale-up criterion is equal media mixing when comparing bench- or pilotscale batches to production-scale batches. Although scale-of-agitation analysis has its limitations, especially in the mixing of heterogeneous media (consisting of suspensions, non-Newtonian liquids, and gas dispersions), this simpliﬁed analysis can be applied to these systems provided the right system variables are used. Such variables include superﬁcial gas velocity, dimensionless aeration numbers for gas systems, and terminal settling velocities as functions of particle size for suspensiontype media containing microcarriers. The scale-of-agitation approach is useful for quick “down-and-dirty” batch scale ups of many STFs and STBs. One of the most important fermentation techniques is mixing. For liquid media, mixing can be deﬁned as a transport process that occurs simultaneously during which constituent materials (i.e., solutes, suspensions, gasses, etc.) achieve uniform concentration or dispersion in the liquid. At macroscopic scale, mixing occurs by bulk transport where constituents are blended by the vortex action of an impeller. At microscopic scale, substances in proximity are blended by eddy currents that impose drag where local velocity and shear differences act on the medium. At molecular scale, blending occurs via molecular diffusion and mass transfer is considered independent of mechanical mixing. Large-volume mixing primarily depends on ﬂow within the fermenter vessel, while small-volume mixing is principally dependent on shear force. The following section covers a simpliﬁed STF scale up utilizing the scale-of-agitation method.

GEOMETRIC METHOD

FOR

STF SCALE UP

The ﬁrst component operation of the scale-of-agitation process depends on the geometric similarity of the smaller and larger fermenters and employs proportional scale up of the geometric parameters of the fermenter vessels, along with geometric ratios such as D/T, where D is the diameter of the impeller and T is the diameter of the vessel (assumed to be cylindrical), or as Z/T, where Z is the depth of the media in the vessel. Similar ratios are compared for both the smaller fermenters (D1/T1) and the larger production units (D2/T2). For example, R = D1/T1 = D2/T2 where R is the geometric scaling factor. After R has been determined, other required parameters such as the rotational speed N of the scaled-up fermenter’s agitator can

PH2270_book.fm Page 193 Tuesday, May 16, 2006 9:37 AM

Fermenters

193

TABLE 8.11 Some Fermenter Suppliers and Products Company

Range of Sizes (Liters)

Applikon Astra Scientiﬁc Bellco Biotech B. Braun Biotech Chemap Cole Parmer Enprotech L–H Fermentation LSL Biolaﬁtte New Brunswick Sci. Schlegel Associates Sulzer Biotech Verax Virtis Wheaton Scientiﬁc

1–1,200 2–100 1.5–50 1–14,000 2–500,000 1–15 2.5–90 1–1,000 1–65,000 1–80; 1–10,000 ind 2–500 1.5–100,000 16 ml Fluidized Bed 1–20 0.5–45

Latest Product (Liters) 5 1.5 & 50 14,000 CMF Series (2–35) 1 L Round Bottom System MCT-25 (3 & 6) Series 2000 (12–100) Maestro™ Series (1–10) BioFlo™ IV (1–10) SpinFirm™ (3.5–1,000) 1.6 & 24 OmniCulture™ (1–2) Turbo-lift™ (19 & 45)

then be calculated by power law relationships as N2 = N1(1/R)n (as with Power Requirements, above). Rotational speeds are expressed either in rpm or in terms of s–1. The power law exponent, n, has a deﬁnite physical signiﬁcance; the value of n and its physical signiﬁcance being determined either empirically or theoretically means. Table 8.11 lists some of the most common power law exponent values assigned to n. The scale up can be completed by using predicted values of N2 to determine the horsepower requirements of the scaled-up system. In most systems, D/T will be in the range of 0.15_ ≥ D/T ≥ 0.6, and Z/T will be in the range of 0.3 ≥ Z/T ≥ 1.5. These values, in conjunction with N and horsepower requirements, completely deﬁne the major agitation system parameters.

DIMENSIONLESS NUMBERS METHOD

FOR

STF SCALE UP

The second method uses dimensionless numbers to predict scale-up parameters that simplify design computation by reducing the number of variables. The dimensionless number approach can also be used successfully in heat and mass transfer scale-up calculations. Usually, the primary independent variable in a dimensionless number correlation is the Reynolds number:

N RE =

D 2 pN μ

(8.121)

PH2270_book.fm Page 194 Tuesday, May 16, 2006 9:37 AM

194

Producing Biomolecular Substances

where N D p μ

= = = =

shaft speed (s–1) propeller blade diameter (cm) the density of the solution/dispersion (g/cm3) the viscosity of the solution/dispersion (g/[cm/s]).

FROUDE NUMBER

FOR

STF SCALE-UP APPLICATIONS

Dimensionless numbers are extensively used for scale up applications. One example is the Froude number: N FR =

DN 2 g

(8.122)

where g is acceleration due to gravity in cm/s–1. The Froude number compares inertial forces to gravitational forces inside the system.

POWER NUMBER METHOD FOR STF GEOMETRIC SCALE UP Another example is the power number, which is a function of both the Reynolds and the Froude numbers: NP =

Pgc pN 3 D 5

(8.123)

where P is power and gc, is a gravitational conversion factor. The number relates density, viscosity, rotational speed, and impeller diameter. The power number correlation has been used successfully for STF geometric scale up. Approximately six additional dimensionless numbers are involved in the various aspects of mixing, heat and mass transfer, etc. Both described scale-up methods are based upon dimensional analysis, a traditional ﬂuid mechanics approach. Unfortunately, these methods do not always achieve accurate results in many of the varied fermenter designs and environments. There is, however, a third method (actually a combination of the ﬁrst two) that can be applied to various design and production situations.

SCALE-OF-AGITATION METHOD

FOR

GEOMETRIC SCALE UP

The basis of this method is geometric scale up with the power law exponent n equal to 1. This provides for equal media velocities in both large- and small-scale equipment. Additionally, several dimensionless factors are used to relate media properties to the physical properties of the equipment. In particular, bulk liquid velocity comparisons are made in relation to the largest impeller blade in the system. The method is best suited for turbulent ﬂow agitation in fermenter vessels that are vertical

PH2270_book.fm Page 195 Tuesday, May 16, 2006 9:37 AM

Fermenters

195

cylinders. Although this method has been successfully applied with marine propeller agitation systems, its original development was based on low-rpm, axial, or radial impellers. Because the most intensive mixing occurs in the volume immediately around the impeller, this method focuses on that particular region. The terminology used to develop the principle behind this approach has been previously described: (1) Determine the D/T ratio of the tank, based upon the largest impeller with which bench-scale or pilot batches have been successfully fermented. It is also necessary to determine the rotational speed and the horsepower of the agitator. Although medium is a difﬁcult material for theoretical prediction since it is non-Newtonian and somewhat compressible, for simplicity only two of the medium’s properties are utilized: point density and viscosity. (2) Calculate the impeller Reynolds number for the original compounding in the smaller system using Equation 8.123; mixing performed in the initial fermentation must be in the turbulent range. If the impeller Reynolds number is 2,000), the NQ curves ﬂatten out and are therefore independent of the Reynolds number. The terminal pumping number, NQ/RE > 2,000, plotted against the D/T ratio, results in Equation 8.124. The cross-sectional area of the pilot-size tank is determined by using Equation 8.125. A=

πT 2 cm 2 4

(8.125)

Then, the value of effective pumping capacity for the smaller-size fermenter is calculated using Equation 8.126: Q = NQ ND 3 cm 3 /s

(8.126)

Finally, by inserting the values derived in Equations 8.124 and 8.125 into Equation 8.126, the value for bulk media velocity is obtained: VB =

Q cm/s A

(8.127)

PH2270_book.fm Page 196 Tuesday, May 16, 2006 9:37 AM

196

Producing Biomolecular Substances

The bulk liquid velocity is then used to determine the level of agitation achieved in the original fermenter batch. Subsequently, the scaled-up production fermenter and its agitator are then designed or selected so that the scale of agitation produced in the larger vessel matches that required for the smaller-size fermenter. Mixing is an equivalent process if the calculated bulk media velocity for the production-size fermenter lies within ±1 unit of the scale of agitation determined from an analysis of the R&D or pilot batches. It is also relatively easy to match the scale of agitation by adjusting the production fermenter agitator rpm, since fermenters are usually equipped with variable-speed impellers capable of multiple agitation levels.

CONTINUOUS-CULTURE SCALE UP THEORETICAL BASIS Since a scale up is only as good as the input data provided, scale up of a continuousculture bioprocess requires detailed data gathered with great care from process development and optimization studies. Continuous culture is usually justiﬁed when productivity is a major consideration. Normally, comparisons are made before scale up between batch and continuous-culture methods. Critical parameters for batch cultures are the doubling time and total downtime; the inoculum size is of much less importance. In 1975, Pirt presented an excellent theoretical basis for continuous culture with cell recycle in which the dilution rate exceeds the speciﬁc growth rate of the microorganism. Ideally, a system should be employed that recycles cells quickly, removing both end product and waste. The advantages of cell-recycle systems with regard to product formation have been experimentally demonstrated. Generally, labscale continuous culture is used to screen strains and to optimize media and culture conditions. Pilot-scale continuous culture is used to verify optimal process conditions and to produce sufﬁcient data for appraisal of large-scale production. Eventually, to commercialize a product the process must be transferred to a production facility. Before making the decision to build such a facility it is necessary to know whether the continuous culture process will actually work in large scale and whether it will function predictably. Although engineering ﬁrms are hired periodically to carry out bioprocess scale up, design, and construction, scaling techniques should begin in either the company’s laboratory or pilot facility or in an independent bioprocess contract facility experienced in scale up technique. Usually multiple progressions (scalings) of continuous-culture volumes are required to develop a product.

LABORATORY

PILOT PLANT

PRODUCTION

The movement of a process from laboratory to pilot plant to production facility is known as scale up. If a scale up is to be successful, the resulting projections for yield, cost, and efﬁciency must be accurate. Scale-up processes must also overcome technical difﬁculties and the limiting features of bioprocesses must be identiﬁed. Common scale-up parameters in continuous-culture systems are generally based on tank geometry, oxygen supply, chemostat volume, substrate concentration, critical dilution rate, cooling capacity, mixing effect, loss of cell population (washout), and agitator power input. As the scale increases, certain parameters change because of

PH2270_book.fm Page 197 Tuesday, May 16, 2006 9:37 AM

Fermenters

197

either increased fermenter size or the changed nature of its operation. These features must be considered when scaling up or down. Chemical features include pH control agents, growth medium, and water quality. Physical features include tank conﬁguration, chemostat volume, aeration, agitation, pressure, medium sterilization, and temperature control. Finally, biological features include mutation probability, contamination probability, and selection pressure.

KEY BIOLOGICAL ASPECT A key biological aspect inﬂuenced by scale up is the total number of generations necessary for the cultivated strain to produce a given amount of product. For example, production volumes 1,000 times larger than laboratory- or pilot-scale might hypothetically require 10 additional generations to reach end stage. The stability of the plasmids harboring the production gene in genetically engineered strains must be able to weather this larger generation number. Thus, a recombinant strain and its unique bioprocess developed under laboratory- and pilot-scale conditions may become neither optimal nor practical at production scale. Therefore, if a fermentation process at production scale is known to require 22 culture generations, then subcultures should be used to simulate this generation number in the laboratory fermenter. A bioprocess may have differing facets when implemented at different scales. Prudent planning requires process analysis, veriﬁcation in a pilot plant, and then a ﬁnal scale up for production. Certain biological features that seem unimportant at smaller scales can have signiﬁcant inﬂuence on the function, design, and operation of production-scale processes, and optimal laboratory conditions may not correspond to those in a production environment. The key to developing an optimal production protocol is simulation of the particular scalar production environment during process development.

MAINTAINING

THE

STEADY STATE

Maintaining culture volume by weight control with a pressure transducer is the most common method used in pilot-scale fermenters, and overﬂow dams provide a simple means of changing the working volume of the vessel. Cylinder-supplied gasses with mass ﬂow regulators are used to keep gassing rates constant, while tachogenerator feedback systems on the agitators maintain constant stirring rates. At steady state, a culture’s oxygen demand is constant and, thus, the agitator speed must be constant since DO2 concentration is affected by agitator speed. On peristaltic pumps, tubing can wear and become contorted so the medium input ﬂow rate must be checked regularly. The medium exit ﬂow rate should also be constant, so exit tubes must be large enough to overcome blockage from aggregates or mycelial growth. And at steady state, pH additions, waste gas composition and levels, etc., are also constant. Process control typically relies on sensing devices within the vessel, so high quality sensors for pH, DO2, etc., should be used since such sensors will prove cost-effective over many fermentations. Removing large samples from the culture vessel requires about 4–5 culture volumes for the fermentation to reequilibrate to steady state. Since direct sampling from the culture vessel is unacceptable, a T-piece with a sampling device can be inserted into the culture exit ﬂow-line close to the vessel. Monitoring unicellular

PH2270_book.fm Page 198 Tuesday, May 16, 2006 9:37 AM

198

Producing Biomolecular Substances

organism biomass levels using spectrophotometry at 590 nm gives a quick, reasonably accurate assessment that only requires small sample volumes. Magnetic stirring bars placed in the holding reservoirs are useful in producing a hom*ogeneous medium during extended cultivation. Accurately buffering the medium can prevent the inconvenience of continual acid or alkali additions.

CONTAMINATION

AND

MUTANT SELECTION

As the result of enforced environmental burden, microbial adaptation will select a population most adapted to that particular environment, which may reduce fermentation productivity. Contamination will deﬁnitely reduce productivity, and can alter the culture’s physiological characteristics and the composition and/or biological activity of the end product. Before inoculating the sterile fermenter, the system should be run under operational conditions in order to ensure asepsis, since microscopic evaluations may not determine a contaminant in time as a result of either sampling error or low contaminant frequency in the total population. Provided that the possibility of contamination is reduced enough so that continuous culture can become an attractive option, recombinant technologies now make it possible to cost-effectively produce commercial products from rapidly growing, genetically altered organisms capable of growing in extreme environments or on unusual carbon sources. Nevertheless, plate assays should still be performed regularly at several pH values, on different media formulations, and over a wide range of incubation temperatures.

ASSESSING PRODUCTIVITY LOSS Continuous culture with cell recycle should normally increase productivity; however, productivity loss should always be assessed since it is generally a result of mutant selection and often accompanied by changes in cell metabolism and morphology. Should productivity loss occur, all parameters must be examined including both contamination and environmental change (e.g., pH, temperature, alteration of limiting nutrient concentration, etc.), since changing the concentration of a critical (but not rate-limiting substance, for example) could negate the productivity loss.

OTHER PROBLEMS If the target culture is known to produce foam, low levels of an antifoam agent can be added to the medium below the surface level of the culture, since this is more effective than surface addition. Regular gasket and seal replacement (without damaging the seals by over-tightening) is important. Wall growth in a glass or glass-lined vessel can be reduced by siliconizing the internal surfaces. Where high carbohydrate concentrations are present in the medium, temperature control circulating heated water or steam should prevent the continuous buildup of burnt material that typically occurs from the use of cartridge heaters. In vessels where cartridge heaters are inserted into pockets, removing the heater during autoclaving generally prolongs heater life. Holding reservoirs should be capable of both sterile addition and medium extraction during a fermentation. When working anaerobically, it is advisable to sparge the media reservoir with nitrogen to maintain anaerobic conditions throughout

PH2270_book.fm Page 199 Tuesday, May 16, 2006 9:37 AM

Fermenters

199

the system. Anti-grow back tubes provide a break in the medium addition lines to prevent organism growth up the lines, and a hooded overﬂow dam prevents variations in medium level due to high-speed surface agitation.

FERMENTER MANUFACTURERS’ DIRECTORY Anspec Co., Inc. Applikon Dependable Instruments B.V. Applikon, Inc. Artisan Industries, Inc. B. Braun Biotech International GmbH Bellco Glass, Inc. Bioengineering AG BioPro International, Inc. Biotech Instruments, Ltd. Broadley-James Corp. Cellex Biosciences, Inc. Chemap AG Chemap, Inc. Alfa-Laval Group Cole-Parmer Instrument Co. Curtin Matheson Scientiﬁc, Inc. Electrolab, Ltd. Eppendorf-Netheler-Hinz GmbH Fisons Instruments, Inc. Genetic Research Instrumentation, Ltd. Heraeus Instruments Labortechnik GmbH Incelltech UK, Ltd. Infors AG Ingold Electrodes, Inc. Jaeger Biotech Engineering, Inc. Kinetek Systems, Inc. LH Fermentation LSL Biolaﬁtte, Inc. Matachana S.A. Antonio MBR Bio Reactor AG MKS Instruments, Inc. MoBiTec GmbH NBS Biologicals New Brunswick Scientiﬁc Co., Inc. “NPO” “Biopribor” “Scientiﬁc-Production Corp. for Bioinstrument” Paul Mueller Co. Pegasus Scientiﬁc Inc. Phoenix Electrode Co. Pope Scientiﬁc, Inc. Prochem Mixing Equipment, Ltd.

PH2270_book.fm Page 200 Tuesday, May 16, 2006 9:37 AM

200

Producing Biomolecular Substances

Sanyo Electric Co., Ltd. SciLog, Inc. Setric Genie Industriel Sulzer Biotech Systems Supelco, Inc. Techne (Cambridge), Ltd. TKA-Teknolabo A.S.S.I Srl Virtis Co. Waters Corp. Wheaton, Inc.

PH2270_book.fm Page 201 Tuesday, May 16, 2006 9:37 AM

Bioreactors INTRODUCTION

Cell culture is a generalized term describing controlled eukaryotic cell growth, as opposed to fermentation, which describes controlled prokaryotic (i.e., microorganism) cell growth. The term cell culture can be further characterized as the controlled, maintained growth of mammalian or insect cells in nutrient medium under controlled conditions, thereby converting feed stock into a desired end product. Such cellular production may be regarded as a biochemical reaction in which eukaryotic cells act as biocatalysts in producing structured proteins. This section discusses bioreactors and the technologies associated with them. From the simple unglazed clay amphorae in which the ancient Greeks fermented their wines to the computer-controlled multifunctional complex bioreactor units we see today, great progress has been made in their design — producing more end product faster, with higher yield, and with more reliability. In past decades, considerations such as shear force acting on fragile culture cells, the speciﬁc conditions for waste biotreatment, and improving methods for product and by-product recovery, have inspired new bioreactor designs. This section discusses and evaluates bioreactor designs developed in recent years, with innovations typically occurring in three areas: (1) design improvements that increase oxygen transfer and decrease shear stress; (2) designs for two-phase reactions in water; (3) designs for the better culture activity control; and (4) bioreactor designs for environmental waste processing. Since cell production utilizes cellular metabolic and enzyme activity to transform biomaterials into a wide assortment of biomolecular end products, scientists can manipulate protein expression to create various structured proteins by utilizing genetic engineering technology. Cell-derived biopharmaceuticals have been licensed for human use around the world, including interferons, monoclonal antibodies, human enzymes, and human hormones — all produced through recombinant DNA technology. The solid-substrate growth modes typically used in making bread and cheese have little application in biopharm technology, so they are not discussed here. Regardless of size, however, bioreactors are used to maintain environments suitable for the controlled growth of eukaryotic cells; so, regulating the culture environment and maintaining aseptic conditions are the bioreactor unit’s key functions. Growth media must be regulated and controlled through agitation or mixing, temperature, aeration, pH, DO2, and antifoam control, as well as through maintaining and controlling other critical parameters. Aseptic conditions are critical here, since foreign organisms and their toxic byproducts can, by contaminating a culture, disturb cell growth, alter product expression, and generally make it difﬁcult to purify the end product. To avoid these problems, bioreactors are typically made of a noncorrosive and nontoxic material that can be repeatedly 201

PH2270_book.fm Page 202 Tuesday, May 16, 2006 9:37 AM

202

Producing Biomolecular Substances

sterilized. Smaller units are typically constructed of borosilicate glass (Pyrex™, or Kimax™), while pilot-sized systems are generally made of #440 stainless steel.

HISTORICAL PERSPECTIVE The biopharm industry uses living cell cultures to manufacture hormones, biologicals, and biopharmaceutical products (a.k.a., biodrugs). Eukaryotic cells convert feedstocks to structured proteins or other biopharm end products. By 1973, foreign DNA fragments inserted into eukaryotic cells for protein and nucleic acid expression required improved containment vessels and computerized process controllers. Today, cell culture is a key element in biotechnology, with many biotechnology products produced in bioreactors that are typically deﬁned by size as bench-top, pilot-scale, or production-scale units.

FERMENTERS

VS.

BIOREACTORS

Strictly speaking, as we’ve previously discussed, a bioreactor grows eukaryotic cells, and a fermenter (as covered in Chapter 7) grows prokaryotic cells, although many manufacturers and scientists still refer to both as fermenters. Distinguishing between them, from time to time, has been based on their method of agitation since eukaryotic cells and particularly mammalian ones are delicate and extremely vulnerable to shear. Some bioreactors are designed for multipurpose use with interchangeable mixing or impeller systems, and delineation between them is typically a matter of intended use rather than design differences. Regardless of size, however, regulating the environment and maintaining aseptic conditions are, as previously emphasized, key functions. Aseptic conditions are critical, since foreign organisms and their toxic byproducts can contaminate cultures, disturbing cell growth, altering product expression, and making it sometimes difﬁcult to separate and purify end products. To avoid this, bioreactors are typically made of noncorrosive and nontoxic material that can be repeatedly sterilized. Smaller units are typically constructed of borosilicate (Pyrex™ or Kimax™) glass, while pilot-sized and production-size systems are generally made of stainless steel with Pyrex™ or Kimax™ ported windows.

BIOREACTOR OPERATING MODES Bioreactors can operate in either a batch or continuous mode. When operating in the batch mode, cell growth proceeds for a deﬁned period, and is ended when the end-product concentration reaches a predetermined level. Subsequently, the contents are harvested, separated, recovered, and puriﬁed. Continuous culture production requires maintaining a steady state environment with sterile medium continuously entering and exiting the vessel at the same rate. Continuous culture frequently maximizes productivity, although the obligatory aseptic steady-state environment is somewhat difﬁcult to maintain.

STRUCTURE

AND

CONFIGURATION

Basically, like fermenters, bioreactors consist of four components: (1) a containment vessel; (2) a media and reagent supply system; (3) an environmental system; and (4)

PH2270_book.fm Page 203 Tuesday, May 16, 2006 9:37 AM

Bioreactors

203

TABLE 9.1 Important Bioreactor Design Considerations • • • • • • • • •

Maintenance of sterility Oxygen transfer rate Heat transfer Instrumentation and control protocol Biological kinetics Fluid hydraulics Mass transfer of substrate to microorganism Mass transfer of product out of microorganism Safety • Failsafe control • Pressure buildup • Microorganism escape

TABLE 9.2 Typical Bioreactor Characteristics • • • • • • • •

Ease of installation Integrated piping system Computer controlled operation Multiple entry and exit ports Suitable contact surfaces Programmable inputs In situ sterilization Heating and cooling

• • • • • • • •

Convenient aeration Foam control Sterilizable peripherals Interchangeable peripherals Batch or continuous Alarm and failsafe systems Physical integrity Reliability

related measurement and control systems. Commercial bioreactors range over a wide spectrum of complexity to meet diverse user needs. Table 9.1 lists some of the factors to be considered in designing a bioreactor, and Table 9.2 lists additional bioreactor operational characteristics. Within the bench-, pilot-, or production-scale groupings, most manufacturers offer basically the same equipment. At the low end of complexity are the bench-top units that are typically used for preliminary development work or educational purposes. These have closed, temperature-controlled containment vessels with minimum instrumentation, accessories, and complex features.

SIZE

AND

SCALE

As the complexity of the bioprocess increases, signiﬁcant peripheral equipment and instrumentation are typically added to the bioreactor, particularly complex measurement and control systems. Despite manufacturers’ claims to the contrary, there are few design features signiﬁcant enough to rank any bioreactor unit over another. Essential features to look for in selecting bioreactors are reliability, multipurpose ﬂexibility, accessory interchangeability, level of sophistication, compatibility with the bioprocess, and the range of monitoring instrumentation available for the system.

PH2270_book.fm Page 204 Tuesday, May 16, 2006 9:37 AM

204

Producing Biomolecular Substances

TABLE 9.3 Stirred-Tank Bioreactor (STBR) Comparison Feature Size: Sterilize: Mixing: Drive: Fittings: Seals: Jacketing: Ports: Surfaces: Parts: Other:

Small Bench

Large Bench

Pilot-Scale

1–10 liters 10–100 liters 100–1,000 liters autoclave in situ in situ Airlift bladed turbine Impellers direct drive or magnetically coupled autoclavable sterilized in situ autoclavable O-ring seals yes, with internal bafﬂes 4–10 4–10 10–20 electro-polished stainless steel, Pyrex interchangeable impellers and ﬁttings bottom aeration by sparging; maximized oxygen transfer; remote valve process control; view window; sterile compressed air between agitator seals; distinctive safety features.

The wide variety of currently available bioreactor designs covers a spectrum that blurs the differences between one class and another, especially when the units are used for producing small quantities of high-value-added biomolecular substances such as cytokines or interferons. Bench-top units range from basic models suitable for undergraduate experimentation to those with highly sophisticated monitoring and control systems coupled with formidable computational capability that can monitor and control countless variables. Table 9.3 points out features that are typical of most stirred-tank bioreactors in various sizes. Bioreactors are frequently categorized according to size: laboratory or research or bench-top bioreactors generally range from 1–50 liters, pilot plant bioreactors typically range from 50–1,000 L, and production scale units are usually larger than 1,000 L. Although the division is somewhat arbitrary, it has been embraced by the bioprocess industries. Nevertheless, some production-scale units are equivalent in volume to pilot-scale bioreactor units, and some pilot-scale units are correspondingly equivalent in function to some bench-top units. To illustrate: a 300 L batch of a monoclonal antibody could typically be construed as a pilot-sized batch, while a 300 L batch of a would-healing cytokines (with its functional dose in fractions of a μg), would be equal to a production-sized batch.

MIXING

AND

AGITATION

Agitation systems are integral to most bioreactors. On many smaller units, agitation is consummated via direct-drive mechanical stirring through the head-plate seal. Some models offer either magnetically coupled agitators or employ air-lift “bubble” systems rather than risking contamination through mechanical seals. These are more often than not limited to smaller volume bioreactors where low torque can still produce effective agitation. At the lower end, LSL BioLaﬁtte, offers a 6-liter biore-

PH2270_book.fm Page 205 Tuesday, May 16, 2006 9:37 AM

Bioreactors

205

actor with an overlay of pressurized sterile air between the impeller shaft seals, and an automatic monitor for seal wear.

CONTAINMENT VESSELS Regardless of bioreactor size, certain elements are common to all. Bench-top units are typically constructed of borosilicate (Pyrex™) glass and/or #440 stainless steel; both materials meeting the vessel’s surface requirements, are nontoxic, noncorrosive, easily cleaned, and can be steam sterilized. Glass vessels are typically composed of a borosilicate (Pyrex™) glass cylinder with #440 stainless-steel head and bottom plates. Bench-top units tend to be less expensive, provide easy viewing of the culture medium, and are usually simple to maintain. In fact, some of the simpler glass units resemble exotic beakers with head plates that provide ports for nutrient, medium, and gas input, and for waste removal. Inert, silicone-rubber O-rings typically provide a seal between the vessel, the head plate, and the bottom plate. Sterilization of benchtop units is typically completed by autoclaving the bioreactor’s components after disassembly and cleaning. Stainless steel bioreactors are typically more expensive than the borosilicate glass units, and, unless a viewing port is incorporated into their design, they severely restrict visual examination of the medium along with the inoculated culture. Stainless steel’s added strength makes them far more durable than the glass units, and double-wall steam sterilization systems are from time to time incorporated as standard equipment, even in the smaller bench-top units. Interestingly, more contamination problems have been attributed to glass than to stainless vessels, probably because the double-wall steam sterilizable stainless units offer greater contamination protection than the glass units’ silicone O-ring rubber seals.

INCREASED SIZE

VS.

COMPLEXITY

As bioprocess complexity increases, signiﬁcant accessory equipment and instrumentation are added to the bioreactor units, particularly to the measurement and control systems. Despite manufacturers’ claims to the contrary, there are few design features signiﬁcant enough to rank any one bioreactor over another. In selecting bioreactors, essential features to seek out are reliability, ﬂexibility, degree of sophistication, and the assortment of monitoring instrumentation equipment available for the system. As bioreactor size increases, the special and more complex features naturally tend to proliferate. Just as some of the smaller units currently offer a range of features once limited to the larger ones, the largest bench-top bioreactors now offer most of the features typical of pilot-scale units. Since the fundamentals of bioreactor design are the same regardless of unit size, distinctions typically occur with the accessories built into the larger units, such as additional entry and exit ports, frequently used to accommodate additional monitoring probes — although sometimes extra ports are used to separate media and culture components before adding them to the bioreactor vessel. As previously discussed, bioreactors are frequently categorized according to overall size and volume: bench-top bioreactor units typically range from 1–50 liters, pilot-plant units typically range from 50–1,000 liters, and production-scale bioreac-

PH2270_book.fm Page 206 Tuesday, May 16, 2006 9:37 AM

206

Producing Biomolecular Substances

tors have volumes that are typically larger than 1,000 liters. Although this division is somewhat arbitrary, it has nevertheless been embraced by industry. Some pilotscale units are equivalent in function to production-scale units, and some pilot-scale units, correspondingly, are equivalent in function to bench-top (research) units. To illustrate: a 300-liter batch of monoclonal antibody produced in a pilot-scale bioreactor for mutagenesis or bioprocess research might be considered analogous to a smaller batch produced in a bench-top research unit, while a 300 L batch of a wouldhealing cytokine produced in a pilot-scale unit (which has an end product functional dose of fractions of a microgram) might be considered analogous to a productionsized batch. Larger vessels typically enable greater control because the increased thermal mass of the larger systems enables better temperature control and the larger volumes facilitate closer pH and buffering precision. The larger volume also allows sophisticated process control equipment currently unavailable with smaller units. For example, Chemap’s Chemcell™ enables an efﬁcient bubble-free medium aeration and cell separation, thereby avoiding growth restriction by built up metabolic wastes. Additionally, production units and sometimes pilot-scale units are typically part of integrated bioprocess systems with hard piping, electrical conduits, sophisticated controls, real-time monitoring, and additional regulatory elements.

ASEPTIC CONDITIONS Aseptic conditions are critical for bioprocessing, given that foreign organisms, their byproducts, and other contaminants may greatly complicate downstream antibody puriﬁcation. To avoid such contamination, bioreactors are made of noncorrosive, nontoxic material that can be repeatedly sterilized. Small units are typically lined with, or constructed of borosilicate (Pyrex™ or Kymax™) glass, while pilot and production units are generally made of #440 stainless steel, with Pyrex™ or Kymax™ ported windows. While it is not difﬁcult to produce milligram amounts of end product from hybridomas in a laboratory, it is quite another thing to generate the gram and kilogram amounts of the protein needed for in vitro diagnostics and biopharmaceuticals. Using standard laboratory techniques to manufacture these quantities could require thousands of ﬂasks, hundreds of liters of media, and can present a myriad of purity, growth, media, and contamination problems. Various proprietary systems, designed to meet the demands of high-volume commercial antibody production, have been developed to grow hybridomas to high cell density, and thereby, effectively produce multikilograms of antibody in the pure and active form.

MONOCLONAL ANTIBODY PRODUCTION INTRODUCTION Monoclonal antibody (mAB) technology was the ﬁrst demonstrable success of the biotech revolution. The rapid-paced transition from research processes to substantial commercial applications occurred in just slightly more than a decade from the establishment of the Milstein-Kohler hybridoma procedure in 1974, through the ﬁrst

PH2270_book.fm Page 207 Tuesday, May 16, 2006 9:37 AM

Bioreactors

207

CX VH

Hinge Region

Fab

CH2

CH3

COO−

N Chain

L +H

ai Ch

H Chain

COO−

Pepsin Papain

N +H 3

FIGURE 9.1 Simpliﬁed model of human antibody molecule with 4-chain basic structure and domains.

commercial mAB production by Hybritech in 1979, to the technology’s ﬁnal recognition in 1984 when Milstein and Kohler were awarded the Nobel prize in medicine. β-lymphocytes and their progeny, the plasma cells, are responsible for the functions of humoral immunity, articulated through the production of certain plasma proteins called antibodies or immunoglobulins. Immunoglobulins are protein molecules that carry antibody activity that is the property of exclusive combination with the substance that elicited their formation (an antigen). With the possible exception of natural antibodies present from birth, antibodies generally arise as a consequence of foreign substances introduced into the body. Immunoglobulins are glycoproteins composed of 82–96% polypeptide, and 4–18% carbohydrate, conﬁgured in equal numbers of heavy and light polypeptide chains. Forming a bilaterally symmetrical structure held together by noncovalent forces and covalent interchain disulﬁde bridges, these chains possess nearly all the biologic properties associated with antibody molecules. All normal immunoglobulins have this structure (see Figure 9.1), though some are composed of more than one 4-chain unit. Each polypeptide chain is composed of a number of domains formed by the interchain disulﬁde bonds. The terminal domains of each chain show greater variation in amino acid sequence, and are therefore designated as variable regions to distinguish them from the relatively constant domains.

ANTIGEN-ANTIBODY BINDING Antigen binding activity is associated with the variable VH and VL domains. As previously noted, immunoglobulins comprise a family of proteins with the same basic molecular structure but with an extensive array of antigen binding speciﬁcities and biological activities that reﬂect their structural differences in the variable domains of their polypeptide chains. This structural heterogeneity also produces a multispeciﬁc response by the immunoglobulin. However, there are mABs, highly speciﬁc antibodies derived from a single cell line, which recognize only one deﬁned

PH2270_book.fm Page 208 Tuesday, May 16, 2006 9:37 AM

208

Producing Biomolecular Substances

antigen. Continuous culture of these cell lines produces identical antibodies against the deﬁned antigens. Thus, a mAB-expressing cell line can be thought of as a continuously cultured, immortal line of cells that produces identically constructed immunoglobulins against a speciﬁc antigen. Since creation of the ﬁrst hybridoma — a cloned mammalian cell line produced by combining the antibody-producing ability of β-lymphocytes with the immortality of murine myeloma cells — the use of monoclonal antibodies has burst out on the scene. Their use in IVD immunoassays, biopharm therapeutics, and in afﬁnity puriﬁcation techniques are just a few commercial applications. High-volume commercial production systems have been developed for growing hybridomas to yield antibodies in a pure and active form. Antibodyantigen binding is a reversible process characterized by an association rate constant (k1) and a dissociation rate constant (k2), with the rate of antibody/analyte complex (AbAn) formation dependent on the concentration of the bound and free species: k1

Ab + An AbAn

(9.1)

d[ AbAn] = dt

(9.2)

formation rate of Ab An = k1[Ab][An] − k 2 [ AbAn]

(9.3)

EQUILIBRIUM DIALYSIS Equilibrium dialysis is the accepted reference method for determining antibody afﬁnity. A dialyzable antigen or hapten and the test antibody are placed in chambers on opposite sides of a dialysis membrane. The system is left until the concentration of free antigen is the same on either side of the membrane (equilibrium); the solutions are then sampled. The average afﬁnity (Ko) is the reciprocal of the free antigen (Agfree) concentration when half of the antibody’s combining sites are occupied. Thus, for Immunoglobulin G (IgG) with two occupied sites: K o = 1 / [ Agfree ]

(9.4)

DETERMINING ANTIBODY AFFINITY The Scatchard equation is the accepted general form for determining antibody afﬁnity: r / [ Agfree ] = nK − rK

(9.5)

Where r is the number of antibody sites occupied (i.e., concentration of bound antigen), n is the antibody valency, and K the afﬁnity constant. Plotting r/[Agfree]

PH2270_book.fm Page 209 Tuesday, May 16, 2006 9:37 AM

Bioreactors

209

against r yields a curve of slope – K that intercepts the r axis at the point where all the antibody combining sites are occupied, that is n, the antibody valency. The Langmuir equation is another formula that can be used to derive the afﬁnity (Ko), and the number of available antibody binding sites (n), by plotting 1/r against 1/[Agfree]: 1 / r = (1 / nK o 1 / [ Agfree ] + 1 / n)

(9.6)

The Sips equation can be used to determine afﬁnity (Ko) and the heterogeneity of a population of antibody molecules. Plotting log r/(n – r) against log[Agfree] yields a line with a slope equal to a, which measures the heterogeneity of binding afﬁnity: (log r / (n − r ) + a log[ Agfree ] + a log K o )

(9.7)

IMMUNOASSAY Whenever detection of a speciﬁc antigen is desired, monoclonal antibodies are the vehicle of choice, since they are completely hom*ogeneous and directed towards a single antigenic determinant. Prior to the advent of hybridoma technology, monoclonal antibodies with special properties would have been impossible to maintain (e.g., bispeciﬁc monoclonal antibodies produced by the fusion of two antibodyproducing hybridoma lines, antiidiotypic antibodies used in place of antigen, and interspecies chimeric antibodies). Furthermore, recombinant DNA technology enabled the production and use of single-domain antibodies, antibody fragment (Fab)-like proteins, and hybrid proteins with both antigen-binding site and speciﬁc enzyme activity. Antibody-producing cells from other animals have been used for fusion with murine myeloma cells, and although the efﬁciency of generating an interspecies hybridoma is low, few of them, if any, secrete speciﬁc antibody. One possible solution is to back fuse the antibody-producing mortal cells with an already generated heterohybridoma cell line (e.g., sheep spleen cells with sheep crossed with mouse heterohybridoma cells). Some of these hetero-hybridoma mABs have been prepared to a very high titer. Immunoglobulins are often conjugated to enzymes that have the ability to effect a color change with the appropriate substrate. This allows their use in the following: 1. Sensitive immunoassays for biomolecular substances; the most widely used are peroxidase and alkaline phosphatase because they catalyze reactions producing colored products, 2. Enzyme-linked immunosorbent assays (ELISAs) are popular test methods that employ enzyme-antibody conjugates. 3. Immunostaining for light microscopy. 4. Hybridoma screening for monoclonal antibody production.

PH2270_book.fm Page 210 Tuesday, May 16, 2006 9:37 AM

210

Producing Biomolecular Substances

ANTIBODY-BASED IMMUNOASSAYS Most immunoassays are either limited reagent (popular methods because they make use of limited amounts of antibody) or reagent excess, noncompetitive methods in which the primary reagents are in excess. Currently, an increasing number of variations appear in research reports and papers. Most assay system complexes can readily be represented by one-line formulas. This notation system will be used throughout this section to help make the different assay systems understandable. Standard abbreviations will be used when the class (IgG, IgM) or type of antibody (monoclonal mAB) or antibody fragment [Fab’, F(ab’)2] or its origin (mouse, M; rabbit, R; goat, G; sheep, S) is relevant.

OPTIMIZING

AND

VALIDATING IMMUNOASSAYS

The Scatchard and Langmuir equations are derivatives of chemical kinetics laws. These related mass action equations are useful for estimating antibody activity, deriving models for standard curve ﬁtting, illuminating fundamental principles relevant to the performance of different assay designs, and assay optimization. The basic limitations and advantages associated with competitive and noncompetitive immunoassays are relevant to developers of IVDs for new analytes, along with higher sensitivities achieved by reagent-excess assays and their general unsuitability for small-molecule, nonprotein analytes. Strategies for the optimization and evaluation of such assays are also essential, and should be based on estimates of precision proﬁles taken at crucial stages in the assay’s development, as part of its validation process.

POLYCLONAL ANTIBODIES Polyclonal antibodies are produced by periodically injecting the host animal with a puriﬁed antigen, and subsequently over time it produces antibodies to the antigen, which are separated from the animal’s serum. The term polyclonal refers to the fact that the collected antibodies are usually a mixture of all the immunoglobulins that were circulating in the animal at the time the blood was collected. Consequently, these antibodies will recognize many different antigens and will probably contain many immunoglobulins that have no avidity for the injected antigen. Polyclonal antibodies are adequate for many immunoassays, although their mixed speciﬁcity sometimes limits their applications.

AFFINITY-PURIFIED ANTIBODIES Afﬁnity-puriﬁed antibodies are prepared by passing polyclonal antibodies through a chromatographic column in which a target antigen was immobilized on the column support. As the immunoglobulins pass through the column, the immobilized antigen’s corresponding speciﬁc antibodies bind to it and are thereby held on the column to be subsequently recovered from the column by changing the eluent conditions (e.g., pH, ionic strength, etc). This is much less expensive than preparing mABs, so the procedure has broad application in preparing antibody conjugates where it is

PH2270_book.fm Page 211 Tuesday, May 16, 2006 9:37 AM

Bioreactors

211

cost-effective to label only antibodies that are intended to function in the test, since afﬁnity puriﬁcation should remove all nonfunctional immunoglobulins.

MONOCLONAL ANTIBODIES An mAB-expressing cell line is a continually cultured, immortal line of cells producing identically constructed immunoglobulins against a speciﬁc antigen. Since the advent of the hybridoma, the use of monoclonal antibodies in immunoassays in commercial applications (e.g., therapeutics, and in biopharmaceutical afﬁnity puriﬁcation techniques) has proliferated. Many high-volume commercial production systems have been developed for growing mAB-producing hybridomas, yielding antibodies in their pure and active forms. Since mABs are completely hom*ogeneous and are directed towards single antigenic determinants, they are the vehicle of choice whenever speciﬁc antigen detection is sought. Prior to the advent of hybridoma technology, it would have been impossible to obtain mABs with special properties (e.g., bispeciﬁc mABs produced after the fusion of two antibody-producing hybridoma lines — antiidiotypic antibodies that can be used in place of antigens — and interspecies chimeric antibodies). Furthermore, recombinant DNA technology enables the production and use of single-domain antibodies, Fab-like proteins, and hybrid proteins, with both antigen-binding site and site-speciﬁc enzyme activities. Antibody-producing cells from other animals have been used for fusions with murine carcinoma cells, although the efﬁciency of generating interspecies hybridoma cells is low, and few of them secrete speciﬁc antibody. One possible solution is to back fuse the antibody-producing mortal cells with an already generated heterohybridoma cell line (e.g., sheep-spleen cells with sheep-crossed-with-mouse heterohybridoma cells). Some of these hetero-hybridoma mABs have been prepared to very high titer.

HYBRIDOMAS Hybridomas are cloned mammalian cell lines produced by fusing two cells of different origin: an immortal cell (one that divides continuously, i.e., a murine myeloma) and an antibody-producing cell (i.e., a human β-lymphocyte). Before the advent of mAB technology, antibody production in animals was stimulated by the injection of various antigens with the resulting polyclonal antibodies extracted from the blood and puriﬁed. However, these methods had some limitations since the blood contains many other antibodies with multiple speciﬁcities. One of the main early uses of antibodies was in clinical diagnostic binding assays. Natural antibodies vary signiﬁcantly from batch to batch and harvest to harvest in the same animal, and from animal to animal. It was this variation that made binding assay standardization so difﬁcult (i.e., if several types of antibody bind to the deﬁned target molecule with different strengths, binding curves are extremely complicated and the assays are not accurate). Although the antigen-antibody reaction is one of the most precise in nature, the possibility still exists of a cross-reaction between the antibody and a foreign antigen, or one of the body’s own substances; with a polyclonal antibody mixture, there is an even greater possibility of a cross-reaction.

PH2270_book.fm Page 212 Tuesday, May 16, 2006 9:37 AM

212

Producing Biomolecular Substances

Antigen PEG

Cell Fusion Spleen Cells (HAT Resistant)

Nonsecretor Myeloma (HAT Sensitive) Culture in HAT

Test for Positive Wells

Clone Antibody Producers

FIGURE 9.2 Creation of hybridoma cell lines for mAB production.

Hybridoma clones, however, can produce mABs (see Figure 9.2) that have extremely high speciﬁcities for deﬁned antigens — much higher than those of polyclonal antibodies. Gel afﬁnity chromatography have been utilized to increase speciﬁcity of naturally derived polyclonal antibodies, which, in some cases, approaches that of monoclonals; however, the method is somewhat expensive and generally does not duplicate monoclonal speciﬁcity. In addition, the quantity of antiserum obtained is low (only a few grams from each animal) and the antibody quality varies considerably with each lot. The original hybridoma fusion process was accomplished with the Sendai virus, however, fusions are now performed in other ways such as in polyethylene glycol (PEG). As a result of this fusion, nuclei combine and mutations occur. In spite of this, many mutants are unstable and die or are selectively killed, but a viable number usually stabilize and begin to produce antibody; Figure 9.3 illustrates the creation of an antitumor mAB for use in in vitro and in vivo diagnostics or in in vivo therapeutics.

CREATING MONOCLONAL ANTIBODIES There are two distinct ways of producing mABs from a hybridoma cell line: 1. Grow them in suspension culture. 2. Grow them in animals from ascites-producing tumors. While cell-culture produced proteins are not contaminated by circulating immune complexes or other antibodies, they are generally produced in lower concentration than is found in ascites-produced ﬂuid. And, although animal produced antibodies may be contaminated with other immunoglobulins and circulating immune com-

PH2270_book.fm Page 213 Tuesday, May 16, 2006 9:37 AM

Bioreactors

213

Human Immune Spleen Cell Culture

Cells Fused Using 50% Polyethylene Glycol

Cultured in HAT Medium (Hypoxanthine, Aminopterin, Thymidine)

Mouse Hybridoma Cell Culture

Cells Diluted Gradually with Culture Medium

Spleen cells do not grow. Hybridoma cells are killed. Hybrid cells grow.

Supernatants Screened for Antitumor Antibody Cells from Positive Wells Cloned Clone Supernatant Tested for Antitumor Antibody Selection of Cells Producing Antitumor Antibody

FIGURE 9.3 Producing mABs against a speciﬁc tumor.

plexes, they can be produced in much higher titers. Monoclonal antibodies have been produced commercially by both methods because each has its own particular market niche. Today, both cell-culture material and ascites ﬂuid are used as IVD immunoassay components, for use in immunochemical characterization, and for antigen isolation. Today, cell culture is regarded as more cost-effective than ascites production, and is considered essential for material to be used as components of in vivo diagnostics or therapeutics. Some commercial IVD antibody production still employs ascites ﬂuid, since this has traditionally been the least expensive way of generating large quantities of high-titer antibody. Ascites ﬂuid contains approximately 1,000 times the antibody concentration of cell culture medium and, since one BALB/c mouse generally produces 3–5 ml of ascites, this volume is equivalent in antibody generation to 3–5 liters of cell culture medium.

MAB PRODUCTION IN ASCITES FLUID OVERVIEW Ascites ﬂuid is most readily produced in animals when the myeloma parent line, the spleen cell donor, and the animal used to produce the ascites ﬂuid are genetically identical (isogeneic), at least at major histocompatibility loci. In most work, this requires the use of BALB/c mice both for immunization and hybridoma production since the myelomas most commonly used for fusion derive from this strain of mouse. If the strain or species of the hybridoma differs from the animal being used to produce ascites it is necessary to prevent immunological host rejection of the tumor by using congenitally athymic (nude) mice or by immunosuppressing the animals. Even with

PH2270_book.fm Page 214 Tuesday, May 16, 2006 9:37 AM

214

Producing Biomolecular Substances

isogeneic animals, it is still necessary to pretreat the peritoneal cavity with the hydrocarbon compound pristane to maximize the probability of hybridoma growth. Although pristane’s mechanism is not completely known, the material most likely acts as an irritant rather than as an immunosuppressive. If a hybridoma repeatedly fails to produce ascites under standard conditions, it might be necessary to immunosuppress the host animal with antilymphocyte serum, immunosuppressive drugs, or radiation — either singly or in combination, even though both mice and hybridoma are thought to be isogeneic (e.g., a combination of hydrocortisone and radiation facilitates rat and human ascites ﬂuid production in mice; approximately 0.5 ml of rabbit- or horse-antimouse lymphocyte serum, and sublethal doses of radiation (≈600 rads) are also effective, although some host animals can succumb to infection).

HOST ANIMALS Ascites ﬂuid is most readily produced in animals when the myeloma parent line, the spleen cell donor, and the animal used to produce the ascites ﬂuid are isogeneic, at least at major histocompatibility loci. In most work, this requires the use of BALB/c mice both for immunization and hybridoma production since the myelomas most commonly used for fusion derive from this strain of mouse. If the strain or species of the hybridoma differs from the animal being used to produce ascites, it is necessary to prevent immunological host rejection of the tumor by using congenitally athymic (nude) mice or by immunosuppressing the animals. Even with isogeneic animals, it is still necessary to pretreat the peritoneal cavity with the hydrocarbon compound pristane to maximize the probability of hybridoma growth. Although pristane’s mechanism is not completely known, the material most likely acts as an irritant rather than as an immunosuppressive. If a hybridoma repeatedly fails to produce ascites under standard conditions, it might be necessary to immunosuppress the host animal with antilymphocyte serum, immunosuppressive drugs, or radiation — either singly or in combination, even though both mice and hybridoma are thought to be isogeneic (e.g., a combination of hydrocortisone and radiation facilitates rat and human ascites ﬂuid production in mice; approximately 0.5 ml of rabbit- or horseantimouse lymphocyte serum, and sublethal doses of radiation (≈600 rads) are also effective, although some host animals can succumb to infection). The BALB/c mice should be 6–8 weeks old when ﬁrst injected with pristane and can be either male or female, although male mice are signiﬁcantly better antibody producers than females. The mice are injected with 0.5 ml pristane (2,6,10,14-tetramethylpentadecane) 1 to 2 weeks before injecting the hybridoma cells; the optimal time interval between pristane and hybridoma injections is 10–14 days. Mice that have been pristane-treated and not subsequently used for more than 4 weeks must be re-injected with pristane and injected with myeloma cells after a 1-week waiting period in order to produce a reasonable volume of ascites ﬂuid.

ANTIBIOTICS Gentamycin (0.2 mg/ml) can be routinely employed in place of penicillin and streptomycin or as an emergency replacement if a contaminating microorganism is

PH2270_book.fm Page 215 Tuesday, May 16, 2006 9:37 AM

Bioreactors

215

resistant to penicillin and streptomycin. Routine use of gentamycin can be particularly advantageous in that it also acts against mycoplasmas. Other antibiotics that are of use against resistant organisms are tylosinate (25 μg/ml) or novobiocin (200 μg/ml) for gram-positive organisms and neomycin (100 μg/ml) or kanamycin (500 μg/ml) for gram-negative organisms. Kanamycin is also effective against mycoplasmas. Adding amphotericin-B (fungizone) at 5–10 μg/ml or mycostatin at 100 μg/ml can help eradicate fungal contaminants. Amphotericin B is supplied as a suspension and cannot be ﬁlter-sterilized without losing activity. A more complete listing and discussion of the antibiotics suitable for cell culture can be found in Chapter 7, CellCulture Media.

INJECTING HOSTS When hybridoma cells are injected into an appropriate animal, they usually grow as a transplantable myeloma, secreting large amounts of immunoglobulin. If the myeloma is introduced as a subcutaneous tumor the antibody is recoverable from the serum, while if the tumor is injected intraperitoneally, it forms ascites ﬂuid. In intraperitoneal-injected mice both the serum and the ascites ﬂuid are rich in immunoglobulin. Most ascites ﬂuid is secreted by the myeloma although ﬂuid is secreted by the host. This is signiﬁcant because antibodies against viral or other animal-host antigens can result in reactions with these circulating immune complexes. Serum is not generally taken with the mouse ascites ﬂuid, although there is no reason (other than perhaps convenience) why it should not also be recovered. The amount of ascites obtainable from an animal varies; in mice the amount is typically 3–5 ml — sometimes as high as 10 ml. Delaying harvest to obtain a high ascites yield often results in premature death of the mice before the ascites can be collected. Further, since the mice become greatly distressed as the ascites volume increases, ethics mandate targeting a modest yield of about 5 ml. The optimal number of cells required to establish active culture growth varies from hybridoma to hybridoma. Injections of 106 cells will typically produce ascites; the optimal range, however, is 6–32 × 106 cells per mouse, with larger doses leading to shorter survival time and smaller ascites volumes (e.g., excess cell injection will not increase the probability of initiating tumor growth, but it is probable that the mice will die early because of systemic effects). Ascites-producing cells are originally produced in culture. However, if large acites-ﬂuid volumes are required, myeloma cells obtained from ascites can be serially routed through the host mice. Ascites-produced cells can also be cryo-preserved to initiate later production. Although some hybridomas have a tendency to produce solid tumors, even with solid tumors an occasional mouse will still produce ascites, and cells from this ﬂuid can be utilized based on the assumption that they were naturally selected for growth in ascites form (many solid tumor-producing hybridomas can produce high-serum antibody titers). Ascites production can take from 7–21 days and, as previously mentioned, it is less distressing to the animals to aim for a modest yield of around 5 ml. The degree of distention associated with a 5 ml yield is usually learned through experience; however, as a crude guide, the mouse looks as if it were in late-term pregnancy. Host mice must be checked at least three times a week — and daily in the late tumor growth stages.

PH2270_book.fm Page 216 Tuesday, May 16, 2006 9:37 AM

216

Producing Biomolecular Substances

SACRIFICING HOSTS The host mice are typically sacriﬁced by carbon dioxide gassing or by cervical dislocation. The abdominal skin is swabbed with disinfectant and a small incision is made over the peritoneum and slightly to the side. The skin is retracted and the peritoneal cavity wall is grasped with a forceps and stretched slightly in order to form a pouch into which a Pasteur pipette is inserted to collect the ascites ﬂuid. There is some risk at this point that the ﬂuid will leak out if the forceps is not properly manipulated. If harvesting and processing are performed under aseptic technique the ascites should be relatively sterile, which facilitates their later use without further processing since sterilizing small liquid volumes is not cost-effective. MAB

PRODUCTION

CELL CULTURE

Research mABs are frequently produced using mouse ascites as the antibody source for both puriﬁcation and characterization. This technique is reasonably rapid, and produces ascites containing from 1–20 mg/ml deﬁned antibody. Ascites-produced mABs are typically puriﬁed by afﬁnity chromatography or biochemical methods, typically resulting in a good yield of high-purity antibody. Hybridoma cell growth medium is also a rich source of antibodies. And, although cell culture secretes considerably less antibody into the liquid than secreted into ascites, it has distinct advantages if the end products are intended for in vivo diagnostic or therapeutic uses: (1) cell culture produces antibodies under much greater environmental control and controlled conditions than possible in animals and, (2) cell culture is less likely to produce end products contaminated with foreign antibodies and with circulating immune complexes, which are difﬁcult and expensive to completely remove during downstream puriﬁcation. Antibody can be produced simply by maintaining the hybridoma culture and splitting it every 2–3 days. Even a 10 ml culture will produce amounts of antibody sufﬁcient for many purposes, since many procedures require only a few μLs of antibody at the concentration secreted by many cell cultures. Cultures can readily be scaled up to 50 or to 100 ml without any specialized equipment by using larger ﬂasks. To allow for adequate gas exchange, the ﬂasks should not be completely ﬁlled with medium and should have their lids loosened when in the oxygen incubator to allow gas exchange. Larger volumes can be produced in simple STBs and other specialized equipment available for larger-scale production. Large-scale hybridoma growth can be facilitated in bioreactors designed for large-scale suspension culture of mammalian cell lines. Bioreactor scale up is described later. An alternative to using larger culture vessels is the hollow-ﬁber cell culture. In this system the medium is continuously circulated past a semipermeable membrane that encloses the cells, low-molecular-weight medium components diffuse across the membrane, and low-molecular-weight waste products are removed. The cells are thereby effectively growing in serum-free media, free from interaction with foreign serum proteins or immune complexes, but having access to the lowmolecular-weight growth factors in the medium. The secreted antibody is then collected periodically.

PH2270_book.fm Page 217 Tuesday, May 16, 2006 9:37 AM

Bioreactors

217

EUKARYOTIC GROWTH MEDIUM Eukaryotic cultures are typically grown in solutions of deﬁned basal medium supplemented with 5–20% serum. Sometimes, costs have been reduced by using less expensive sera, since most established hybridoma lines do not absolutely require 10% fetal calf serum (FCS). Various sera have been used, including mixtures of horse and FCS, adult bovine serum, newborn calf serum, and calf serum at various concentrations. For 1–2-week short culture periods, serum concentration has been sometimes reduced to only 5 percent. A few years ago, the serum supply, especially FBS, was limited, and the FBS market has ﬂuctuated wildly. Both worldwide demand and price have been high; even raw bulk serum can exceed $150–$250 per liter (depending on the type), and processing can typically add another $35 per liter. Assuming relatively high cell productivity, serum may add from $200–$1,000 per gram to the manufacturing cost of the end product. Nevertheless, key serum-free media components can be purchased in bulk at low cost, yielding lower cost for serum substitute formulations. Those anticipating higher cost for serum-free medium are well advised to develop their formulations in-house, even if their eventual media preparation is contracted out. For example, an incremental investment in media development of less than $500,000 could easily be recovered in anywhere from a year to even a few months, and for many cell lines the cost of in-house media development can be even less. Recently, data derived from serum-supplemented media containing undeﬁned substances such as immune complexes and hormones known to affect cell systems has been questioned, and bioprocess applications have demonstrated not only variability, but increased downstream puriﬁcation costs when unreﬁned products are contaminated by extraneous proteins and/or circulating immune complexes. The possibility of introducing adventitious agents such as viruses and mycoplasma sp. into culture systems via the serum also favors culturing hybridoma cells in totally deﬁned medium. Medium is a key reason for the optimization of hybridoma cell culture environments. In using mammalian hybridomas to produce monoclonal antibodies, a modiﬁed medium, such as RPMI 1640, Ham’s F12, Dubecco’s modiﬁcation of Eagle’s Minimal Essential Medium (DMEM), and similar formulations may be employed by combining them with various amounts of carbohydrates, glutamines, or amino acids, and trace metals, inorganics, vitamins, lipids, steroids, attachment and growth factors. Originally bovine or other serum was commonly added to provide the protein factors necessary for regulating and stimulating growth — factors such as IL-2, EGF, and the attachment factors laminin and ﬁbronectin. Conversely, serum-free deﬁned medium can also provide these key ingredients without the disadvantages of serum-based media. Efﬁcient bench-scale bioreactors with computerized process controllers and systems for determining cell number, gas concentration, expression products, waste products, and feed stocks (i.e., amino acids, sugars, vitamins, trace minerals, etc.) have spurred on effective and rapid media development as a bioprocess tool. Serum harbors numerous adventitious agents such as pyrogens, circulating immune complexes, viruses, foreign DNA, and foreign antibodies. Replacement of serum-based media with deﬁned media is therefore desirable for producing in vitro

PH2270_book.fm Page 218 Tuesday, May 16, 2006 9:37 AM

218

Producing Biomolecular Substances

diagnostics (IVDs) and in vivo therapeutics. While totally serum-free cell media is available, it is still common to grow cells in serum-based medium, then subsequently replace it with serum-free medium or, every so often, even protein-free medium, allowing for product expression. Various serum-free media have been developed, and hybridoma cell growth is best in a basal medium supplemented with additives such as growth and attachment factors, trace metals, vitamins, etc.

SERUM-FREE MEDIUM (SFM) Serum-free medium is preferred to serum-supplemented medium to facilitate producing purer antibodies and reducing their downstream puriﬁcation costs. Serumsupplemented media usually contains combinations of BSA, insulin, transferrin, assorted growth factors, and hormones — with protein levels typically ranging from 10–150 μg/ml. Many of the problems with serum-supplemented media occur when using serum-derived media, especially during mAB production, but to a lesser extent than with serum-supplemented media. A number of serum-free media preparations have been employed for growing hybridomas and many are commercially available. Published serum-free media formulae for murine hybridomas include the Murakami formula (the medium is supplemented with insulin, transferrin, ethanolamine, and selenite). The advantages of serum-free long-term growth media extend far beyond process stability, ease of validation, operational consistency, higher puriﬁcation yield, and signiﬁcant cost reduction, since media cost is a signiﬁcant factor in total production cost. Basal medium (without serum or serum substitute) costs about $1–$4 per liter in quantity and prepared from components, and from $4–$10 per liter in preformulated powder or supplied in solution.

PROTEIN-FREE MEDIUM (PFM) Protein-free medium (PFM) supports the growth and long-term mAB production in static cultures of murine hybridomas, in hollow-ﬁber cultures, and in STB cultures. PFM offers advantages over serum-containing media, since it is a totally-deﬁned eukaryotic cell culture medium containing no polypeptide growth or attachment factors, and in its cell growth kinetics and actual mAB production, it compares quite favorably with serum-containing media. With the proper formulation, conversion time is either nonexistent or minimal, and mAB levels are equal or superior to those of cultures grown in serum-supplemented media, with signiﬁcantly less contamination. Protein-free media growth rates typically average in excess of 90% of serumcontaining media (e.g., RPMI-1640 + 10% FCS). PFM successfully supports high-density hollow-ﬁber, and STB cultures, producing yields comparable to those of STB cultures in serum-containing medium. Although cellular adaptation not typically required, delicate or sensitive cultures can be readily converted to PFM without loss of biological productivity by using conventional cell culture adaptation techniques. PFM growth rates are also acceptable for many recombinant applications, resulting in less exogenous protein contaminants in the unreﬁned product, that signiﬁcantly reduces the downstream puriﬁcation cost. As much as 90% of all mABproducing murine hybridomas may be successfully cultured in PFM.

PH2270_book.fm Page 219 Tuesday, May 16, 2006 9:37 AM

Bioreactors

SUSPENSION

219 VS.

ANCHORAGE-DEPENDENT CELLS

Although hybridomas, myelomas, transformed, and lymphoid (circulatory-system) cells grow in suspension culture similar to bacteria, most mammalian cells are anchorage-dependent and unable to prosper unless anchored to a surface. Typically, smaller research labs grew these cells on specially treated culture-vessel-surfaces as are found in treated Petri dishes and T-ﬂasks. However, since these small vessels cannot accommodate large-scale cell cultures, alternative methods were developed. The major technical challenge here was to increase the cell-attachment area within a given culture vessel. Roller bottles, porous solid-supports, and more recently, hollow-ﬁber systems, have been used for this purpose. For anchorage-dependent cells, traditional conﬁgurations incorporating microcarriers appear to offer high surface to volume ratios. Most microcarriers are about 150–180 μ in diameter, and when they are utilized in 3-10% concentrations, yield an effective surface area of about 10–30 centimeters. Microcarriers employ a wide range of surface coatings including tertiary- and quaternaryamine groups, polystyrene, collagens, and gelatins. Cells growing on solid bead surfaces, however, remain exposed, and can still be seriously damaged by collisions with other beads, impeller blades, or even the vessel wall. Eddy currents and other turbulences created by the rotating impeller also strip cells from microcarrier surfaces.

POROUS MICROCARRIERS Porous microcarriers help overcome, or at least signiﬁcantly reduce, the above problems since hybridoma cells can grow inside the microcarrier where they are safe from the detrimental effects previously discussed. Porous collagen beads, are also a good substrate for culturing murine hybridoma cells, since the beads’ pores are large enough to permit interior cell growth, that will protect the cells growing inside the beads from hydrodynamic stresses. Murine hybridoma growth on solid microcarriers is inhibited at agitation speeds over 100 rpm. However, when the cells are cultured within the porous microcarriers at the same level of agitation, cell growth is unaffected by the increased hydrodynamic stress. The porous conﬁguration is particularly signiﬁcant because cells adhering to the larger beads should be more susceptible to shear damage, yet this was not the case. The most serious problems with microcarriers are: (1) the susceptibility of the cells growing on the outside of the microcarrier to the agitation; (2) the cell damage resulting from the interaction of cells with the turbulence vortex; (3) collisions of the microcarriers with themselves; or (4) the microcarriers impacting on the impeller blade or other parts of the bioreactor unit. At low concentrations, when the length of the smallest vortices approach that of the microcarrier diameter, cell growth is sharply curtailed. Apparently, such vortices are too small to move the microcarriers and thereby transmit energy to the cells immobilized on the microcarrier surface. Bead-bead interaction also increases with microcarrier concentration and is the predominant cause of cell death at high carrier concentration. Cell concentration decreases immediately after agitation rate increase (e.g., from 35 rpm to 100 rpm, to 150 rpm, to 225 rpm). This temporary decrease results partly

PH2270_book.fm Page 220 Tuesday, May 16, 2006 9:37 AM

220

Producing Biomolecular Substances

from loss of the cells growing on the microcarrier surface where they are unprotected by the matrix, and partly from a lower growth rate resulting from increased hydrodynamic stress at the higher agitation. For example, if a stop-stir-type inoculation is used instead of inoculating during continuous agitation, hybridoma cells can achieve higher attachment densities with greater inoculation consistency among the porous microcarriers than if insufﬁcient inoculation contributes to a lower cell growth rate. Satisfactory porous microcarrier performance offers both a protective cell environment and the ability to supply nutrients to cells growing in the support interior. Murine hybridoma cells closely ﬁll the microcarrier void volume, so protective effects are maximized by increasing the ratio of void volume to particle surface area. In addition to porous collagen microcarriers, cross-linked gelatin porous beads are also available as Percell™ beads (Perstorp Biolytica, AB, Lund, Sweden). About 220 μm in diameter with a 50% void volume, they have been productively used for murine hybridoma cell cultivation. Porous microcarriers with 99% void volumes and 500–800 μ diameters are at the upper size limit able to support oxygen transfer to interior cells coupled together with maximum protection ability. Verax (Lebanon, NH) offered porous, sponge-like microspheres composed of native bovine collagen, 500–600 μm in diameter with interconnecting 20–40 μm pores and channels, with an internal volume that is 85% cell accessible. Karyon (Norwood, MA) protected cells by ﬁrst attaching them to gelatin fragments and then entrapping them in solid microbeads of calcium-chelated sodium alginate. Damon Biotech (Needham Heights, MA) has also successfully grown large quantities of anchorage-independent cells by microencapsulating them with sodium alginate in solid microbeads and then coating them with a synthetic amino acid polymer. The alginate is subsequently liqueﬁed to generate a capsule containing the cells; membrane permeability depends on the choice and concentration of the amino acid polymer, which is adjusted to allow the passage of nutrients into, and cellular waste out of, the capsule. Permeability can also be adjusted so that the desired end product accumulates within the microcapsule or diffuses into the culture medium. Signiﬁcant growth of murine hybridomas has been demonstrated in microcarriers containing insoluble denatured collagen (gelatin) either as shards or nonporous microbeads (commercially available as GelibeadsTM). The cells are usually attached to the insoluble gelatin matrix the evening before encapsulation, and subsequently the entire mixture of cells, gelatin, and sodium alginate is passed through bead-forming heads. Many shards or GelibeadsTM can be contained within a single microcapsule because the gelatin shards and GelibeadsTM are much smaller than the average diameter of microcapsules (i.e., 10–50 μm for the shards, 50–100 μm for the GelibeadsTM and about 700 μm for the entire microcapsule). Gelatin shards are too small and diverse in size to be suitable as conventional microcarriers without microencapsulation. A solid substratum in the microcapsules allows for vigorous and healthy growth of murine hybridoma cells on GelibeadsTM. Because only the exterior surface of the GelibeadsTM is accessible for cell attachment, densities of 8 × 107 viable cells per ml of microcapsule (equaling 1.6 × 107 viable cells per ml culture volume) have been attained. The percentage of viable cells

PH2270_book.fm Page 221 Tuesday, May 16, 2006 9:37 AM

Bioreactors

221

typically declines, however, after 25 days because the high number of cells causes the formation of necrotic areas in the capsule. When growing murine hybridoma cells in microcapsules, the cells enter a kind of stationary phase (the cell number stabilizes and DNA synthesis drops precipitously); however, when fresh serumcontaining medium is added to the culture, growth resumes. Production of antibody, therefore, depends on serum, fresh medium, and cell growth, given that antibody production initially lags during early exponential growth, ceases during stationary phase and resumes again when fresh serum-containing medium is added. Gelatin-shard-inverted-microcarriers offer advantages for anchorage-dependent cells in suspension: (1) gelatin is inexpensive; (2) the semipermeable polyaminoacid membrane protects the cells from damage while permitting nutrients intake and waste outﬂow; (3) the membrane’s permeability can be controlled enabling two product harvesting methods: (a) when the product’s molecular size is below the exclusion size of the membrane, the product diffuses into the culture medium, facilitating separating the product from the cells and cellular debris, and (b) when the product’s size is too great to pass through the membrane it accumulates in the microcapsules concentrating the product; (4) entrapped product can then be harvested by simply collecting the microcapsules and subsequently disrupting them. This makes inverted microcarriers an economically attractive method for large-scale anchorage-dependent cell cultivation.

ADVENTITIOUS AGENTS Microbial contamination is a constant threat to cell culture, since the medium supporting eukaryotic cell growth also supports the growth of various bacteria and fungi. Many novel products produced in animal cell substrates may theoretically transmit various bacteria, viruses, and fungi from animals to humans through these products. The challenge of identifying potential adventitious agents closely parallels the challenge of identifying agents causing particular emerging infectious diseases. Thus, the major focus of this discussion will be the approaches used in regulatory settings to ensure that mAB products are devoid of adventitious agents. Knowing mABs are free from adventitious agents is a large part of consumer conﬁdence in these products. Thus, ensuring that mAB products administered to the public do not contain adventitious agents is a key regulatory goal. Of course, the potential for the presence of adventitious agents in any protein product must also be evaluated in terms of its overall beneﬁt. In the past, biological products served as vectors for various viral diseases. Examples include the contamination of vaccines with hepatitis B virus in the 1940s (because a human-derived excipient contained hepatitis B virus), contamination of early polio and adenovirus vaccines with simian virus 40 in the late 1950s and early 1960s, contamination of blood products with hepatitis viruses and HIV, and contamination of dura mater grafts with the Creutzfeldt-Jakob disease agent. In these examples, either human or animal materials used in production usually caused the contamination. mAB production generally involves inoculation of a cell substrate with a hybridoma seed and subsequent puriﬁcation of the bulk antibody product after a sufﬁcient time for protein production. Thus, adventitious agents could theo-

PH2270_book.fm Page 222 Tuesday, May 16, 2006 9:37 AM

222

Producing Biomolecular Substances

retically enter a hybridoma culture through any of the raw ingredients. Close control of the manufacturing environment (by producing mABs in sophisticated modern equipment and facilities), appropriate testing of the raw materials, and the testing of both the bulk and ﬁnal products can help make certain adventitious agents do not enter the mAB end product. Most mABs are subjected to puriﬁcation steps that reduce the likelihood of end product contamination. Some excellent techniques for reducing contamination to acceptable levels are utilizing culture environments under a ﬂow of sterile air and incorporating antibiotics in the medium. These useful techniques can also help manage contamination: sterility testing, contaminant identiﬁcation, and the use of special antibiotics in emergency situations. Experience has shown that media quality is rewarded by avoiding delays and repeat batches. Constituents purchased from reputable manufacturers are sterile, and manufacturer quality programs are usually quite good. Constituents prepared and sterilized in the laboratory or plant, however, should be sterility tested. Sterility testing should be performed for fungi using Sabouraud medium (Code #CM147, Oxoid), and with trypticase/soy broth for aerobic bacteria. Anaerobe testing is usually not required since they are not typically a source of contamination in aerobic cell culture. The test media should be passed through a vented 0.22 μm micropore ﬁlter, such as the Millipore IVEX-HP, which is designed to collect both microorganisms and spores. The ﬁlter is ﬁrst washed with sterile water followed by the medium being ﬁltered. Excess medium is removed by washing three times with sterile water; the ﬁlter assembly is ﬁlled with the appropriate growth medium, stoppered, and incubated to allow microbial growth. The test for fungi is incubated at 32°C for 14 days, while aerobic cultures are incubated for 7 days at 23°C. Aseptic technique is veriﬁed by preparing additional ﬁlters, not challenged by test medium; if these show contamination, either the sterile technique or one of the assay components (water, media, syringes, etc.) is contaminated. Contaminant identity often provides insight to its origin. Yeast contamination typically originates on the skin of workers, suggesting the need for more careful technique and the use of surgical gloves. Bovine mycoplasma sp. typically originates from FCS; certain types of bacteria that are typically waterborne suggest inadequate media ﬁltration or possible contamination from the water bath. Fungal contamination is typically ascertained by examining the culture with an inverted microscope; on the other hand, low-level bacterial contamination is difﬁcult to identify microscopically since many cell fragments are difﬁcult to distinguish from bacteria while demonstrating Brownian movement. A culture suspected of bacterial contamination should be gram-stained and characterization of the suspected contaminating agents should be determined by a competent bacteriologist. In addition, mycoplasma sp. and viral detection procedures are quite specialized and usually require the use of contract laboratories for accurate testing. mAB intended for in vivo therapeutic use must undergo rigorous screening for viral contaminants, and thus, always require testing by contract laboratories. The best thing to do with a contaminated culture is to autoclave it. Then again, there are occasions when it is worthwhile to attempt to cure a contaminated culture, such as when an important cell line is contaminated and there are no seed cultures available, or when your liquid-nitrogen-banked culture has become contaminated. Antibiotic sensitivity testing may be of some value here,

PH2270_book.fm Page 223 Tuesday, May 16, 2006 9:37 AM

Bioreactors

223

although numerous antibiotics have adverse effects on murine hybridoma cell growth, not only by inhibiting it, but by not destroying the contaminating organisms.

BIOREACTOR

VS.

ASCITES-FLUID PRODUCTION

OKT3 antibody has been approved for therapeutic use since 1986, and probably represents one of the better characterized mABs. Producing OKT3 by large-scale cell culture reveals some minor biochemical differences between bioreactor-produced and ascites-produced mABs, which probably result from posttranslational peptide chain modiﬁcation. Nearly identical modiﬁcations occur when mABs are produced in vitro, illustrating a protein processing modiﬁcation common to many in vitro cell-culture-produced and bioreactor-produced mABs. Either the posttranslational antibody modiﬁcation occurs extracellularly, or the intracellular carboxypeptidases responsible for these protein modiﬁcations function in an altered way, depending upon environmental conditions affecting the culture. A number of carboxypeptidases have been shown to be responsible for terminal processing bradykinins, enkephalins, renins, erythropoetins, and calmodulins, and are intracellularly present in serum — either associated with secretory granules or with the cell membranes. Incubation of mABs with ascites ﬂuid removes the heterogeneity, indicating that posttranslational antibody modiﬁcation can result from enzymes in the ascites ﬂuid. Since most OKT3 antibody secreted in serum-free media lacks a C-terminal lysine, some of the carboxypeptidases might be associated with the hybridoma as well. A modiﬁcation identical to that which results from carboxypeptidase-B treatment transforms high-molecular isoforms into isoforms typical of the majority of bioreactor-produced OKT3, or ascites-produced antibody. Isoelectric focusing and subsequent Western blotting indicate an additional isoform produced by differently charged heavy chains. Carboxy-terminus analysis indicates that most ascites-produced OKT3 heavy chain molecules terminate in glycine, and bioreactor-produced OKT3 heavy chain molecules terminate in lysine. Comparison of these results with the amino acid sequence of OKT3 derived from cDNA indicates that glycine, the C-terminus residue of ascites-produced OKT3, is the residue predicted from the cDNA sequence, and that the C-terminus residue released from the extra isoform was lysine, as predicted from the cDNA. In the case of OKT3, this alteration does not affect the antibody’s in vitro function. AOKT3, TC-OKT3, and extra isoforms all induced mitosis and mitogenic responses that required the interaction of OKT3 and the monocyte Fc-receptor, which is unaffected by the charge difference at the C-terminus of the extraisoform. Differences between bioreactor- and ascites-produced OKT3 also occur in the carbohydrate moieties, although carbohydrate differences do not necessarily result in evident charge or functional differences. Murine IgG2a antibodies contain approximately 5–10% carbohydrate, and studies of OKT3 suggest that it contains 3% carbohydrate attached at a single site on the heavy chain in close proximity to the hinge region, so that any difference in glycosylation between ascites- and bioreactorproduced OKT3 probably doesn’t result in appreciable charge differences. Both biochemically and functionally, culture-produced OKT3 is identical to the ascitesproduced product. Consequently, if both antibody production techniques employ the

PH2270_book.fm Page 224 Tuesday, May 16, 2006 9:37 AM

224

Producing Biomolecular Substances

same hybridoma clones, charge differences will typically occur in 5–10% of the bioreactor-produced antibodies, resulting from the possible alteration of the heavy chain C-terminus processing, and will exhibit differences in assays sensitive to charge, such as isoelectric focusing and ion exchange LC. The carboxy-terminus is part of the constant region of the immunoglobulin molecule and is highly protected. C-terminus amino acid sequence analysis of 13 murine IgG heavy chains of various subclasses reveals that they all end in either pro-gly-lys-COOH or leu-gly-lys-COOH. This sequence is also conserved at the Cterminus of all human IgG heavy chains. In vivo, the C-terminus lysine is regularly removed by carboxypeptidases, and evidence of this type of immunoglobulin processing can be derived from known protein sequences since sequences identiﬁed in secreted protein appear to lack lysine, while sequences identiﬁed in nucleic acids usually show it. Since many mABs are originally characterized from ascites ﬂuid and subsequently scaled up in bioreactors, this difference between the end products should be recognized.

BIOREACTOR ENGINEERING AND DESIGN DESIGN

AND

ENGINEERING CONSIDERATIONS

Basic Physical Elements The initial phase of bioreactor design must establish the physical volume necessary for the equipment; and whether the cell cullture will operate with a single bioreactor on a batch basis, with multiple bioreactors on a batch basis, or as a continuous culture system. To obtain the required volume for the vessel, the time of the fermentation cycle can be calculated by using kinetic models of the growth rate matched to the required production rate. The time cycle and the availability of the vessel for processing can then be used to determine the required bioreactor volume (e.g., a single bioreactor employing a speciﬁc culture working an eight-hour day, ﬁve days a week should be a great deal larger than a single bioreactor used for the same culture, working a 168-hour, seven-day week in order to generate the same amount of product). In general, transport mechanisms for the growth of microorganisms can be modeled. In the simple case of aerobic bacteria, the major problem is to get the oxygen to the organism in sufﬁcient quantities to enable exponential growth. The types of resistance to the mass transfer from the medium to the cell can be categorized as: (1) diffusion from bulk gas to the gas/medium interface, (2) solution of the gas in the medium at the interface, (3) diffusion of the dissolved gas into the medium, (4) transport of the dissolved gas to the immediate region of the cell, (5) diffusion through the static region of the medium surrounding the cell, (6) diffusion into the cell, and (7) consumption by the cell depending on its growth kinetics. The rate of mass transfer of a gas into a liquid can be written as: N A = kG (C1 − C2 ) = k L (C3 − C4 )

(9.8)

PH2270_book.fm Page 225 Tuesday, May 16, 2006 9:37 AM

Bioreactors

225

Here NA is the rate of mass transfer of gas A (kmol/s m2), C1 and C2 are gas concentrations in the gas phase, C3 and C4 are gas concentrations in the liquid phase, and kG and kL are gas and liquid mass-transfer coefﬁcients [w/dimensional resistance–1]. The simplest explanation of mass transfer is the twin-ﬁlm theory in which conditions at the gas/liquid interface can be represented by a plot of concentration against distance from the interface. This model enables the simple rate law to be written as: N A = kG ( pA − pI ) = k L (C I − C A )

(9.9)

Where p is the partial pressure of gas in the gas mixture (e.g., oxygen in air); C is the concentration of gas dissolved in the liquid; the subscript I refers to interface conditions, and the subscript A refers to bulk conditions. The regions near the interface bounded by the interface and where the bulk concentration begins to change are known as boundary layers. The mass-transfer rate equation assumes a straightline relationship over the boundary layers: N A = kG (δP) = k L (δC )

(9.10)

The mass-transfer coefﬁcients (kG, kL) are actually ﬁlm mass-transfer coefﬁcients related only to conditions at the interface. Because it is difﬁcult to measure values of pI and CI, the overall mass-transfer coefﬁcients related to concentration and conditions in the bulk gas and liquid are used. Overall coefﬁcients are deﬁned by: N A = K G ( pA − p*) = K L (C * − C A )

(9.11)

Where p* is the partial pressure of the gas in equilibrium with a solution of the gas at concentration CA and C* is the equilibrium concentration of the gas in solution with a partial pressure of pA above the solution. Henry’s Law applies to equilibrium data for many gasses that are soluble in water (medium): p* = HCA

(9.12)

Where H is the Henry constant. If Henry applies, then the following are valid: pA = HC*;

p* = HCA;

pI = HCI

(9.13)

These equilibrium relationships can be used to derive interrelationships between ﬁlm coefﬁcients and overall coefﬁcients. For sparingly soluble gases like oxygen (10 ppm at 1 atmosphere pressure), the numerical value of H is quite large (at 30°C, H = 4.85 × 104 atmospheres/mol fraction). Thus, the major resistance to mass transfer lies in the boundary layer. Because the rate of mass transfer (NA) is the rate per unit area perpendicular to the direction of transfer, the total mass transfer of gas is depicted by:

PH2270_book.fm Page 226 Tuesday, May 16, 2006 9:37 AM

226

Producing Biomolecular Substances

rate = driving force/resistance

(9.14)

where A is the transfer surface area per unit volume of medium. In order to provide a large surface area for mass transfer, either the gas is broken up into small bubbles, or the medium (substrate) must be spread out over a large surface area. The second option is not suitable for aseptic bioreactor operation, except in the cases of packedbed bioreactors or hollow-ﬁber bioreactors, and bubble formation is the primary aeration technique of aseptic fermentation. The most commonly used bioreactor to perform aerobic fermentations is the mechanically agitated bioreactor, using an agitator with a sparger ﬁtted below the impeller blades to admit air. Typically bioreactor dimensions are expressed in terms of ratios so that, provided these ratios are maintained (independent of the volume of the bioreactor), the bioreactor scale up in terms of power requirements, aeration, agitator speed, etc., is simpliﬁed, particularly if geometrical and dynamic similarities are maintained (see Batch Scale up and Continuous-Culture Scale up, below). Agitation keeps bubbles of air circulating through the medium. The faster the rotation speed of the impeller, the longer the bubbles take to pass through the medium in the bioreactor before exiting, and the more time is therefore available for mass transfer. The overall masstransfer coefﬁcient (KL) for such vessels is related to the impeller diameter Di and the speed of revolution N for constant impeller design. Oxygen absorption into fermentation media is established by: ⎛Π ⎞ K Lα = KV = 0.002 ⎜ G ⎟ ⎝ V ⎠

0.7

VS0.2

(9.15)

where KV is the overall volumetric oxygen transfer coefﬁcient (1/s); πG/V is the power input per unit volume of media in a gassed system (W/m3), and VS is the superﬁcial air velocity based on the cross-sectional area of the bioreactor (m/s). This expression only applies for a liquid height (L) to vessel diameter (Dτ) ratio of 1.0. For L/Dτ ratios between 2 and 4, values of KV are about 50% higher than those produced by the above expression. Total transfer rate is equal to KV(C*CA), and the C units will determine the KV units (e.g., if C is measured in kmol/m32 then KV is also in kmol/s m3 volume). There are two important criteria for maintaining cell growth in aerobic fermentations: (1) maintaining a dissolved oxygen concentration above the critical concentration required for exponential growth (typical values range from 0.003 to 0.05 mol/m3), and (2) supplying oxygen at a rate matching metabolic utilization. Although no air supply is necessary for anaerobic fermentation, the need for medium agitation applies to both aerobic and anaerobic fermentation in order to provide necessary dispersion functions for materials other than oxygen. When liquid is placed in a cylindrical vessel with a centrally mounted impeller and no bafﬂes, the main liquid motion is circular and overall mixing is poor. Bafﬂes increase mixing speed with a penalty of higher power requirements. To design an agitated bioreactor it is important to predict the power requirements for aerobic fermentations since oxygen transfer is a function of power input per unit volume (μG/V).

PH2270_book.fm Page 227 Tuesday, May 16, 2006 9:37 AM

Bioreactors

227

Practical problems in biotechnology and biochemical engineering fall into two categories: (1) the system behavior and material components are well-deﬁned mathematically; and, (2) system behavior and material components are either only partly deﬁned or are interdependent and cannot be deﬁned in precise mathematical terms (the classic three-body problem). The ﬁrst type of problem can be solved by mathematical calculation even though the solution might require complicated procedures such as integrating simultaneous differential equations. An example of this type problem is in liquid sterilization with thermal-death kinetics that are well deﬁned. It is also possible to express temperature proﬁles for heating and cooling mathematically and a combination of these equations can enable a mathematical solution to be obtained for the value of ∇. Problems of the second type, however, where no mathematical solution is possible, can still be solved empirically by dimensional analysis. Complex Physical Elements In order to use dimensional analysis, sufﬁcient information about the system’s physics must be known so that incorrect results will not develop if a key variable is omitted. The procedure is performed by initially determining all the possible behavioral relationships of the variable of interest: Y = function (other variables; physical properties, etc.) = K (AaBbCc…etc.), where K is a dimensionless constant. Next, the variables and physical properties are expressed in terms of basic dimensions (e.g., mass, length, time, etc). For instance, velocity (m/s) can be expressed dimensionally as length/time and acceleration (m/s2) as length/time2. After rewriting the basic expression in terms of dimensions, the exponents of the dimensions (a, b, c, etc.) are equated, and the resultant simultaneous equations are then solved. Thus, for all practical purposes: ⎛ g⎞ P⎜ ⎟ ⎝l⎠

1/ 2

= 2π

(9.16)

or ⎛l⎞ P = 2π ⎜ ⎟ ⎝ g⎠

1/ 2

(9.17)

Dimensionless number groups are collections of physical properties and variables such that irrespective of the units used for the calculation the value of the group (number) will always be the same provided consistent units are used (i.e., all SI, all metric, all engineering, etc.). Bioprocess engineering problems associated with liquid ﬂow, mass transfer, heat transfer, mixing etc., can only be practically solved using dimensional analysis. Some of the more common dimensionless groups are given in Table 9.4.

PH2270_book.fm Page 228 Tuesday, May 16, 2006 9:37 AM

228

Producing Biomolecular Substances

TABLE 9.4 Dimensionless Number Sets Name

Group 10 n 0.001

Meaning number of cells probability of contamination

Del (sterilization)

∇ = ln

Froude (Fr) (agitation)

DN 2 v2 or g Lg

inertia forces gravitation forces

Grashof (Gr)

3 D32 pC g∇p ν2 C

gravitation forces viscous forces

Nusselt (Nu)

hD k

gradient at boundary fluid gradient

Peclét) (Pe)

DvpC p k

convection heat transfer conduction heat trannsfer

Dv D

momentum mass transfer diffusion mass transfeer

Power (Po)

Pmo pDI5 N 3

drag force on impeller inertial force

Prandtl (Pr)

CP ν k

diffusion of momentum diffusion of heat

Dvp μ

inertia forces viscous forces

Reynolds (Re) Fluid ﬂow Agitation

Bubble

Impeller

DN 2 p μ DBM Vt pc μc DI2 Npc μc

Schmidt (Sc)

μ pD

diffusion of momentum molecular diffusion

Sherwood (Sh)

kL D32 DAB

total mass transfer diffusive mass transfer

Weber (We)

D3N 2 p σ

inertia forces rising bubble forces

Note: D is characteristic length (e.g., pipe diameter, sweep diameter of agitator); v is ﬂuid velocity; ν is ﬂuid viscosity; Cp is ﬂuid speciﬁc heat capacity; k is ﬂuid thermal conductivity; D is ﬂuid diffusivity coefﬁcient; ƒ is ﬂuid surface tension; L is characteristic length; N is speed of rotation; ηc and pc are the viscosity and density of the liquid phase, respectively. The Sauter-mean diameter (D32) is determined from the measured droplet size distribution as: D32 = Σ in=1ni Di3 / Σ in=1ni Di2

PH2270_book.fm Page 229 Tuesday, May 16, 2006 9:37 AM

Bioreactors

229

Not all relationships use dimensionless groups, however. Dimensional analysis can be applied to agitation to obtain a relationship of the power required (Π) with variables and physical properties: Π = KDia N bρcμ d g e

(9.18)

where DI is the impeller diameter, N is the speed of rotation, is the liquid density, μ is the liquid viscosity, and g is the acceleration due to gravity, a relationship containing three dimensionless groups is obtained: a

⎛ D 2 N ρ ⎞ ⎛ Di N 2 ⎞ Π = k⎜ i ⎟ ⎜ ⎟ 3 5 ρN Di ⎝ μ ⎠ ⎝ g ⎠

(9.19)

where Π Power number (Po) (ρN 3 Di5 )

(9.20)

Di2 N ρ Reynolds number (Re) μ

(9.21)

Di N 2 Froude number (Fr) g

(9.22)

This relationship applies to a particular agitator in a particular bioreactor vessel with a particular bafﬂe conﬁguration, although the same relationship can be used for different size vessels and agitators provided that there is both geometric and dynamic similarity. For geometric similarity, the following ratios should be the same for both vessels: Dτ / Di vessel diameter/impeller diameter

(9.23)

L / Di media height/impeller diameter

(9.24)

I / Di impeller height from bottom/impeller diameter

(9.25)

W / Di baffle width/vessel diameter

(9.26)

PH2270_book.fm Page 230 Tuesday, May 16, 2006 9:37 AM

230

Producing Biomolecular Substances

For dynamic similarity there are two possible methods: (1) the scale up can be based on establishing that the Reynolds number (Re) is the same in both size bioreactor vessels; with constant physical properties this means Di3N of bioreactor vessel #1 should equal Di3N of bioreactor vessel #2; this method is suitable for anaerobic fermentation processes; and (2), for aerobic processes, since we are dealing with a two-phase system (media/gas bubbles), a better scale up method employs the Weber number (We) with constant physical properties so that Di3N for bioreactor vessel #1 equals Di3N for bioreactor vessel #2. For a marine impeller at a Reynolds number greater than 5,000, for example, the value of the power number (Po) would be 5.8; this only applies to a single impeller placed at distance Di from the base of the bioreactor vessel (i.e., I/ Di = 1). For multiple impellers: Po = J (Po)1, where J is the number of impellers and (Po)1 is the power number for a single impeller. In aerobic operation, the power requirement is less if air is fed to the underside of the agitator or gassing system. General correlations covering single impellers are: ⎛ Po 2 NDi3 ⎞ (Po )G = 0.354 ⎜ ⎝ Q 0.56 ⎟⎠

0.45

(Po )G = 0.7 exp(−0.9Q )

(9.27)

(9.28)

where N is the impeller speed (rev/s), Di is the impeller diameter (m), and Q is the volumetric air ﬂow rate (m3/s). The subscripted G applies to a gassed system. These correlations are dimensional, can be applied to large-scale bioreactors, and only deﬁned units are used. For calculation of the mass-transfer coefﬁcient the power requirement per unit volume (G/V) is also required. In designing an aerobic bioreactor it is essential to determine the ungassed power requirement, otherwise an underspeciﬁed motor could burn out if the air ﬂow is stopped or fails while the system is running. Unlike most chemical processes that frequently operate at high temperatures, bioprocesses usually function best at near-ambient temperatures. Nevertheless, heat transfer is still important, and both the application and removal of heat from any given segment of the bioprocess must be considered since heat transfer here is a deliberate attempt to alter or regulate any temperature changes generated by the bioprocess. Isothermal heat transfer is important in controlling a bioreactor’s environmental temperature, and for processes such as cell disruption that generate large quantities of heat. Biocatalysts also have a well-deﬁned temperature range over which they are active. If the temperature is too low, biological activity could be insigniﬁcant, and if it is too high for an extended period of time, thermal degradation could result. For organisms that normally grow near ambient, as well as their enzymes, this deﬁned temperature range is usually below 40°C. Heat Generation Large fermenters and bioreactors often show net heat generation due to the friction of rotating impellers, metabolic activity, or exothermic reactions. In all such circum-

PH2270_book.fm Page 231 Tuesday, May 16, 2006 9:37 AM

Bioreactors

231

stances the bioreactor requires cooling. In contrast to the requirements of deliberate temperature alteration, the rate of heat transfer during isothermal operations is usually quite low. Thus, the size of the heating/cooling system in the bioreactor will be inﬂuenced by its predicted uses — large for rapid heating, smaller for constant temperature maintenance. There are two ways of transferring heat to and from agitated vessels: (1) using jackets, and (2) using internal coils. The heat transfer coefﬁcient at the inside vessel wall is ⎛ D2 N ρ ⎞ hi D_ = 0.36 ⎜ i ⎟ k ⎝ μ ⎠

0.67

⎛ μ⎞ (Pr)0.33 ⎜ ⎟ ⎝ μS ⎠

0.14

(9.29)

Where Dτ is the vessel’s inside diameter, Di is the agitator diameter, N is the agitator speed in revolutions, k is the liquid’s thermal conductivity, and Pr is the liquid’s Prandtl number. With jackets either ﬁtted or integral to the bioreactor vessels, the surface area per unit volume falls as the vessel’s diameter and volume increase. Once a certain volume has been exceeded the surface area becomes too small to remove fermentation-generated heat and the only alternative is to use internal coils. Cell-culture-generated heat also dictates the maximum volume of a bioreactor vessel that is suitable for jacketing. The heat-transfer coefﬁcient at the outside surface of a coil (h0) ﬁtted inside an agitated bioreactor vessel is: ⎛ D2 N ρ ⎞ h0 D_ = 0.90 ⎜ i ⎟ k ⎝ μ ⎠

0.67

⎛ μ⎞ (Pr)0.33 ⎜ ⎟ ⎝ μs ⎠

0.14

(9.30)

Note that the vessel diameter Dτ appears in the Nusselt number, not in the coil diameter. These expressions for jackets and coils can be applied for both laminar and turbulent conditions and are valid for Reynolds’ numbers over the range likely to be encountered (from Re 500 to Re 500,000). In the double-pipe annular heat exchanger with a hot liquid ﬂowing in the outer pipe to heat a colder liquid ﬂowing in the inner pipe, there are two ways to arrange the liquid ﬂow: (1) concurrently (parallel, in the same direction), or (2) counter-currently (parallel, in the opposite direction). Analysis of heat balances based on heat transfer characteristics and thermal transfer between liquids can be carried out to demonstrate that: Q = UAΔt

(9.31)

The temperature difference (Δt) used is the log mean temperature difference (ΔtLM) as deﬁned by: Δt LM =

( ΔT1 − ΔT2 ) ln( ΔT1 / ΔT2 )

(9.32)

PH2270_book.fm Page 232 Tuesday, May 16, 2006 9:37 AM

232

Producing Biomolecular Substances

ΔT1 and ΔT2 are the terminal temperature differences. As illustrated, counter-current operation yields a higher temperature difference compared with concurrent operation, and offers a number of advantages: (1) a higher value for ΔtLM means that the same task can be performed using a smaller heat exchanger [the area required A will be smaller if U is the same]; (2) the temperature difference throughout the heat exchanger does not exhibit as large a variation as the concurrent ﬂow; and (3) the hot liquid could have vacated the heat exchanger at a lower temperature with countercurrent ﬂow. For practicality, heat exchangers are always operated countercurrently to get maximum advantage from the above points. Heat transfer to and from a bioreactor vessel is somewhat different than from a normal heat exchanger. When heating a batch of medium in a bioreactor, the temperature of the contents continually varies, and therefore the log mean temperature difference (ΔtLM) also varies. In the case of convection heat transfer with the contents initially at ti and the heating liquid temperature tS, it can be shown that: ⎛ UAθ ⎞ (t − t s ) = exp ⎜ − (ti − t s ) ⎝ mC p ⎟⎠

(9.33)

Where t is the temperature at time Δt, m is the mass of the medium in the vessel, CP is the speciﬁc heat capacity of the medium, and U is the heat transfer coefﬁcient. When the bioreactor contents are maintained at constant temperature the standard heat exchanger expression can be used. The log mean temperature difference ΔtLM is calculated using terminal temperature differences even though the vessel contents are at constant temperature. Unlike counter-current heat exchangers, the exit temperature of the heating/cooling liquid in the bioreactor jacket or coil should not approach the contents’ temperature closer than 5°C.

PROCESS CONTROL SYSTEM DESIGN Bioreactor Dynamics Bioreactors are multivariable systems with nonlinear dynamics. Prior to discussing speciﬁc control applications, some general bioreactor features relevant to control will be covered (see Table 9.5). Two key eccentricities that must be known before designing a bioreactor control system are: 1. Bioreactors are multivarite systems: bioreactor control systems will involve many variables. 2. Bioreactors have nonlinear dynamics: controlling bioreactors is complicated by the fact that the nonsteady-state behavior is also nonlinear, which results in several problems: (a) hysteresis is observed; a step increase in reactor feed rate in STBs results in transients far different than when correspondingly equivalent step decreases in feed rate are taken backwards towards the initial conditions. Moreover, multiple steady states present for identical feed conditions and, in certain cases, exotic dynamics, i.e.,

PH2270_book.fm Page 233 Tuesday, May 16, 2006 9:37 AM

Bioreactors

233

TABLE 9.5 Examples of Process Control Functions • • • • • • • • • • • • • • • • •

limit cycles, oscillatory transients, and long time lags may be exhibited; (b) the reasons for these sometimes unpredictable behaviors are the complexities of living cells. 3. Many key variables desirable for monitoring and control are only measurable after large time lags, or not at all. This gives incentive for creating more accurate mathematical models or newer estimation techniques. Fortunately, simple models and single-input-single-output feedback loops are available, and work quite well for these purposes.

PROCESS MEASUREMENT

AND

CONTROL

Prior to discussing speciﬁc process monitoring and control applications, some general features of bioreactors relevant to control should be discussed. A computer automated system has process monitoring systems and controls for (1) liquid ﬂow, (2) dissolved gas concentrations: O2, CO2, N2, and (3) other process parameters. Two main characteristics that are important to be aware of prior to designing a bioreactor control system are: 1. Bioreactors are multivariate systems: as one would anticipate, bioreactor control involves many variables. 2. Bioreactors exhibit nonlinear dynamics: Bioreactor control is complicated by the fact that nonsteady-state behavior is nonlinear, which have some consequences: (a) hysteresis, e.g., a step increase in bioreactor feed rate in a stirred-tank bioreactor (STB) results in a different transient than a

PH2270_book.fm Page 234 Tuesday, May 16, 2006 9:37 AM

234

Producing Biomolecular Substances

TABLE 9.6 Typical Bioreactor Control Ranges Temperature: Agitator speed: Stability: pH range: Pressure: DO2 range: Air flow:

8ºC above coolant to 60ºC ± 1ºC 0–1,000 rpm >98% 2–12 ± 0.1 2,000 mbar 0–100% 0–6 liters/minute

correspondingly equivalent step-decrease in feed rate back to the initial conditions. (b) multiple steady states are often observed with identical feed conditions, and in some cases, exotic dynamics such as limit cycles, oscillatory transients, or long time-lags may occur; the reasons for these behaviors are ultimately a result of the complexities of the living cell. 3. Monitoring and control of many important desirable variables are only measurable with large time lags or not at all. This gives opportunity for accurate mathematical models and/or state-of-the-art estimation techniques. Fortunately, simple models with single feedback loops are available and perform very well. Process monitoring and control stands at the leading edge of bioreactor development. Progress in bioanalytical chemistry resulted in a vast array of probes, biosensors, and detectors, which enable the process engineer to screen a myriad of growth parameters (see Table 9.6 for typical parameter control ranges). Table 9.7 lists some of the instrumentation parameters typically checked by process control systems. A bioprocess’s ﬁnancial success is directly related to its level of process control sophistication, and it is expected, therefore, that many future bioreactor developments will be in the area of process monitoring and control. Process monitoring and control systems can be readily adapted to any size bioreactor, the only limitation being the number of such devices the vessel can physically accommodate. Currently, the modular approach is quite popular and offers a wide selection of products of varying capability that can be selected according to the project’s speciﬁc needs, so an imposing array of conﬁgurations are possible. A good example is the B. Braun BIOSTAT™ that offers seven basic conﬁgurations having culture vessel volumes of 2, 2.75–20, 25, 29, 72, 92, 150, and 300 liters, as well as associated measurement and control modules comprising either digital or analog displays for pH control, additive measurement, antifoam addition, level control, pO2, pCO2, redox, temperature, and real-time recording of results during cell culture. The antifoam control employs a conductance-regulated adjustable sensor, which is linked through the controller to a pump for adding antifoam chemicals, and with a variable adjustment to eliminate extra antifoam addition that may result from splashes. The pO2 measurement is carried out by either galvanic or polaro-

PH2270_book.fm Page 235 Tuesday, May 16, 2006 9:37 AM

Bioreactors

235

TABLE 9.7 Typical Bioreactor Process Control Monitoring Air flow: Coolant: Power: Temperature: Rheology: Redox: Culture: Cell concentration: Immunocyticity: Gas analysis:

Flow meter Flow meter V.O.M.; torque Resistance thermocouples, thermistors, diodes Tube viscometers, cone and plate viscometers, concentric cylinder viscometers, inﬁnite sea viscometers, foam control pH, dissolved oxygen (DO2), polarographic probes, galvanic probes, permeable tube/oxygen analyzer, dissolved carbon dioxide (DCO2.) Enzyme probes, metabolic heat, substrate analysis Gravimetric dry weight, turbidity Cell number, impedence, carbon dioxide production, oxygen production, DNA content, cell particle size distribution Oxygen analyzers, carbon dioxide analyzers, mass spectrometry, gas chromatography

graphic electrodes, regulated by the system controller that governs agitator speed and/or airﬂow-rate adjustment. Actual pH measurement and control is accomplished by means of a pH electrode linked through the system controller to discrete pumps for adding either acid or base. While regular pH electrodes degrade with repeated sterilization, the newer gel-ﬁlled electrodes are extremely stable. Inlet air is controlled by a simple pump and ﬂow meter, while temperature is monitored by a thermocouple linked to a discrete temperature control system. At the core of the bioreactor is the system controller. Today, in addition to purchasing complete units and modular assemblies, there is a growing market trend to retroﬁt older bioreactors with new, state-of-the-art control systems. For example, Wheaton offers a userfriendly Macintosh-based controller, their Proteus 2000™ that can be retroﬁt to nearly all bench-top units and some of their pilot-scale units. Astra Scientiﬁc offers a monitoring and control package for IBM PCs, along with the interface hardware necessary for systems producing analog output. For systems without computer control, few indeed these days, data acquisition and real-time logging can still be acquired with I/O hardware packages. Process Monitoring: Sensor Location and Function 1. 2. 3. 4.

PH2270_book.fm Page 236 Tuesday, May 16, 2006 9:37 AM

236

Producing Biomolecular Substances

TABLE 9.8 Typical Process Control Components • • • • • • •

Sensors Controllers Actuators (Final Element) Process controllers Manual controls Automated controllers Feedback controllers

6. Off-line sensors; not a part of fermentation equipment, and the sensor’s measured value is not directly available for process control. The intervention of an operator is essential for the actual measurement and for entering the measured value into the system for process control. Process Control Process control is concerned with making adjustments to the process based upon measuring one or more variables that change as a result of the process function using the components listed in Table 9.8. Feedback Controllers Feedback controllers compare the measured value of the process variable that must be controlled with its set point and adjust an actuator in order to suppress the deviation between the measuring value and set point. Automatic Control Systems The controller output change is proportional to the input signal created by the environmental change error that was detected by the sensor: M = Mo + K cΣ

(9.34)

where M = output signal, and Mo = controller output signal when there is no error, Kc = controller gain or sensitivity, and Σ = the error signal. Automatic control systems are listed in Table 9.9. Integral Controllers The integral controller output signal is determined by the integral of the error over operating time:

PH2270_book.fm Page 237 Tuesday, May 16, 2006 9:37 AM

Bioreactors

237

TABLE 9.9 Automatic Control Systems • • • • • • • • •

On/off controllers Modulated controllers Proportional controllers Integral controllers Derivative controllers Combination controllers: PI, PD, and PID Two-position on/off controllers Modulated controllers Proportional controllers

∫

M = M o + 1 / Ti Σ dt

(9.35)

where Ti is the integral time. Derivative Controllers Derivative controllers sense the rate of change of an error signal and contribute an output signal component that is proportional to the derivative of the error signal: M = M o + Td dΣ / dt

(9.36)

where Td is the integral time. Feed-Forward Controllers Feed-forward controllers employ a measured variable(s) other than the process variable that must be controlled to carry out an action. Adaptive Bioreactor Controls Mostly applied in systems where process variables or characteristics aren’t known and can’t be measured directly, or when the bioprocesses’ static or dynamic behavior changes with time. The adaptive controller can either use online measured process data, a theoretical model, or combination of these to predict the change in the static or dynamic behavior of the process. Complex Bioreactor Controls Computer applications in fermentation technology include data logging, data analysis, process control, process data logging, and data analysis.

PH2270_book.fm Page 238 Tuesday, May 16, 2006 9:37 AM

238

Producing Biomolecular Substances

BIOREACTORS II: KINETIC MODELING AND BIOREACTOR DYNAMICS AND BIOREACTOR DESIGN PROGRAM MODELING CELL GROWTH

AND

PRODUCT FORMATION

Overview Bioreactors are multivariate systems with nonlinear dynamics. A bioreactor’s nonsteady-state behavior is nonlinear. This has several consequences; hysteresis is often observed, where a step increase in bioreactor feed rate in an STB results in transients that are different than when a corresponding equivalent step decrease is effected in feed-rate, back to the initial conditions. Moreover, multiple steady states are often observed for identical feed conditions, and in some cases, exotic dynamics such as limit cycles, oscillatory transients, and long time lags may be evident. The reasons for this behavior are ultimately results of the variable complexities of living cells. Thus, many important monitoring and control variables are only measurable with large time lags or are not measurable at all. This gives hope for more accurate mathematical models or estimation techniques. Fortunately, the simpler models, and single-input-single-output feedback loops are becoming available and work in many cases. Growth Reaction Eukaryotic cell growth can be represented as: Cell + {C-source, N-source, others} + O 2 → More Cells + Extracellular + CO2 + H 2O

(9.37)

where the cell culture medium is food for the cell (i.e., it serves as a source for all the elements needed by the cell for growth and biosynthesis, and for protein/endproduct formation). The end-product compound’s carbon dioxide and water, on the product side of the reaction, result from the oxidation of feedstocks in the medium. Since the cellular material contains C, N, P, S, K, Na, Ca, etc., the medium must be formulated to supply those elements in appropriate form. This growth reaction is useful for interpreting laboratory data reported in the literature, since many early cell growth studies were reported by microbiologists; their terms are typically used to describe cell-growth stoichiometry, and can relate the above reaction to commonly reported cell properties. Structured Models Unstructured, distributed kinetic models do not acknowledge changes in cell composition during growth and do not account for latent stages, sequential uptake, or

PH2270_book.fm Page 239 Tuesday, May 16, 2006 9:37 AM

Bioreactors

239

changes in mean cell size during batch culture growth, although Monod’s unstructured, distributed models can satisfactorily predict cell growth in many situations. Structured models, on the other hand, recognize the multiplicity of cell components and their interactions, and many such models have been created based upon assumptions about speciﬁc cell components and their interactions. In order to create a simple structured model that can be used for realistic bioprocess scale ups, it must be stipulated that a system is composed of multiple cells that do not contain abiotic material, and carry mass m on a dry basis at a speciﬁc volume vˆ . If one postulates there are c components in the cell and the mass of the zth component per unit volume of the system is Cˆ XZ , and that kinetic rate expressions for p reactions occurring in the system and that the rate of the zth component formed from the yth reaction per unit volume of the system is rˆX y,z — then, assuming the speciﬁc volume vˆ is constant with time during the batch cultivation, the change of the zth component in the system with respect to time can be expressed as: dCˆ X Z = dt

∑r

X y ,z

−

y =1

1 dm ˆ CXZ m dt

(9.38)

where the second term of the right side of Eq. 9.38 represents the dilution of intracellular components by the growth of biomaterial, since all variables denoted by circumﬂexes are intracellular properties; and since the structural models acknowledge cell-component multiplicity and multicomponent intracellular interactions, the model is expressed with intrinsic variables. The relationship between the cell growth rate and the kinetic rate expressions of all intracellular reactions is: dCˆ X = dt

∑∑r

Xy,z

−

z =1 y =1

1 m ˆ CX m dt

(9.39)

Concentration terms can be expressed as mass per unit culture volume vˆ instead of mass per biotic system volume mv. Concentration based on culture volume can be expressed as: dC XZ mvˆ = dt V

∑r

Xy,z

(9.40)

y =1

And, although concentration terms are based on total culture volume, kinetic parameters still remain on a biotic-phase basis. One of the simplest structured models, the two-compartment, is based on the following premises. (1) The cell comprises two basic compartments — the synthetic portion A, such as precursor molecules and RNA, and the structural portion B, such as protein and DNA:

PH2270_book.fm Page 240 Tuesday, May 16, 2006 9:37 AM

240

Producing Biomolecular Substances

dCˆ X =0 dt

(9.41)

(2) The synthetic portion A is fed by uptake from the substrate S and the structural portion B is fed from the synthetic portion: S→A→B

(9.42)

(3) The rate of the ﬁrst reaction in Eq. 9.43 is proportional to the product of both the substrate and the cell concentration: rX1, A = k1CS Cˆ X

(9.43)

where CS is mass of substrate per unit of abiotic volume, v − mvˆ. The rate of the second reaction in Eq. 9.44 is proportional to the product of the concentrations of the synthetic portion and the structural portion: rX2 , B = − rX2 , A = k2Cˆ X A Cˆ X B

(9.44)

Transcribing Eq. 9.44 for each component, and substituting Eqs. 9.45 and 9.46 for the reaction rates: dCˆ X A 1 dm ˆ = k1CS Cˆ X − k2Cˆ X A Cˆ X B − C XA dt m dt

(9.45)

dCˆ X B 1 dm ˆ = k2Cˆ X A Cˆ X B − C XB dt m dt

(9.46)

where the change in substrate concentration is shown by: dCS k = − 1 CS Cˆ X dt YX /S

(9.47)

and, where YX/S is the yield constant. Thus, Eqs. 9.48, 9.52, 9.53, and 9.54 can be solved simultaneously to obtain the change of Cˆ X B , Cˆ X A , Cˆ X , Cˆ S , and m. (4) Cell ˆ ˆ X B ) has doubled in initial value division only occurs if the structural portion ( mvC and the dividing cells apportion each component equally to their daughter cells. Therefore, the total number of cells in the system will be proportional to the structural

PH2270_book.fm Page 241 Tuesday, May 16, 2006 9:37 AM

Bioreactors

241

portion, where the average cell mass is equal to the total cell mass divided by total number of cells. Consequently,

Average mass of cell =

ˆ ˆX m mvC Cˆ α = ˆX ˆ ˆ XB C XB n mvC

(9.48)

The model can also predict the change: IC X = C X A + C X B′

(9.49)

Eqs. 9.50, 9.51, and 9.52 can be expressed with concentrations in terms of mass per unit culture volume: dC X A ⎛ V ⎞ ⎛ V ⎞ =⎜ ⎟⎠ k1CS C X − ⎜⎝ ˆ ⎟⎠ k2C X A C X B ˆ ⎝ dt V − mv mv

(9.50)

dC X B ⎛ V ⎞ =⎜ kC C ⎝ mvˆ ⎟⎠ 2 X A X B dt

(9.51)

1 dCS =− dt YX /S

⎛ V ⎞ ⎟ k1CS C X ⎜⎝ V − mvˆ ⎠

(9.52)

where C X = (mvˆ / V )C X

(9.53a)

C X A = (mvˆ / V )Cˆ X A

(9.53b)

C X B = (mvˆ / V )Cˆ X B

(9.53c)

CS = [(V − mvˆ ) / V ]CS

(9.53d)

and

and, mass m is related to CX

PH2270_book.fm Page 242 Tuesday, May 16, 2006 9:37 AM

242

Producing Biomolecular Substances

m = CXV

(9.54)

Thus, the total number of cells is proportional to C X B and the average mass of a cell to C X / C X B . The batch culture simulation curve shows changes in mass (C X / C Xmax ), number (C X B / C X Bmax ), and size of cells (C X / C X B ), and substrate concentration (CS/CS0) in dimensionless form, and illustrates the following: (1) during latent stage, cells grow in size but not in number; (2) during exponential growth, cells are largest; (3) during static stage, cells no longer grow or divide. And, even though this model provides many features that unstructured models are not able to predict, it requires only two parameters…the same number of parameters required for Monod! Cell Yield and Stoichiometric Coefficients Now, if we convert this growth reaction to a mass basis: Slope = 0.046 g of Cells per Gram Glucose Consumed

(9.55)

Then the above value is designated as cell yield, growth yield, or yield coefﬁcient. So, if the previously stated growth reaction is examined, the slope (in mass units) can be equated as: ⎛ MW of Cell ⎞ Cell Yield = α ⎜ ⎝ MW of Substrate ⎟⎠

(9.56)

The numerator contains the number (amount) of cells created, while the denominator contains the amount of carbon substrate consumed. This measurement enables us to calculate the stoichiometric coefﬁcient, α, or: ⎛ 180 ⎞ ⇒ 0.33 α = (0.046 ) × ⎜ ⎝ 24.6 ⎟⎠

(9.57)

Invariably, the amount of protein produced is proportional to the number (amount) of cells, and higher levels of biosynthesis are possible only with aerobic cultures. Cell Yield Mathematically, cell yield is deﬁned as: YX /S =

Amount of Cell Produced ΔX = Amount of Substrrate Consumed ΔS

(9.58)

where ΔX represents change in cell concentration and ΔS represents change in substrate concentration. The subscript X/S indicates the basis of cell yield on substrate. This notation comes in handy when it’s necessary to calculate the yield based

PH2270_book.fm Page 243 Tuesday, May 16, 2006 9:37 AM

Bioreactors

243

on more than one substrate. Examining the above and comparing it with the growth reaction, it is evident that the yield deﬁned here corresponds to a mass-based stoichiometric coefﬁcient. Taking the limit of Eq 9.59 as DS approaches zero: YX /S =

dX dS

(9.59)

The absolute sign eliminates the negative value of the derivative. Note that dS is negative, because substrate is consumed. Yield is always a positive value. The above deﬁnition of yield can be applied to product P on the basis of substrate consumed. Thus, YP /S =

dP dS

(9.60)

Similarly, product yield based on cells (amount) is expressed as: YP / X =

dP dX

(9.61)

Typically, yield of species i, based on species j, is calculated from: Yi / j =

di dj

(9.62)

From this, it is clear that two different yields that have a common species may be combined so that: Yi / j =

Yi / k Yk / j

(9.63)

Stoichiometric Coefficient Measurements For the growth reaction given in Eq. 9.59, the ratio g/b is called the respiratory quotient, often abbreviated as RQ, and is easily measured in large scale bioreactors. In Eq. 9.59, if the nature of extracellular product is known (i.e. x, y, z), then it is possible to calculate w, x, y, and z from experimental measurement of RQ and one other measurement.

CELL GROWTH THERMODYNAMICS Heat Release Due to Growth Cell growth is the function of a complex network of metabolic reactions. Coupled anabolic and catabolic reactions take place so that the energy released in the former,

PH2270_book.fm Page 244 Tuesday, May 16, 2006 9:37 AM

244

Producing Biomolecular Substances

and is efﬁciently used to drive the latter. Some energy, however, is lost as heat. The purpose of this section is to quantify heat release due to cell growth. In large-scale bioprocesses it is mandatory to remove heat so that the culture is maintained at an optimum physiological temperature. In small bioreactors metabolic heat is removed quite easily, while in very large bioreactor units (>10,000 liters), in which rapidly growing cells are cultivated, it is necessary to design adequate heat-transfer systems for heat dispersal. Bioreactor temperatures must be maintained within ± 0.5°C to maintain optimal growth conditions. Consider a growth reaction with no signiﬁcant amount of extracellular product being formed. Under such conditions, Eq. 9.59 simpliﬁes to: C6 H12O 6 + α NH 3 + βO 2 → α CH1.8O 0.5 N 0.2 + γ CO 2 + δH 2O

(9.64)

Since nitrogen consumption is typically quite small compared to the amount of consumed carbon, and since nitrogen does not oxidize (while carbon, of course, does), we can approximate: C6 H12O 6 + βO 2 → α CH1.8O 0.5 N 0.2 + γ CO 2 + δ H 2O

(9.65)

Consider this reaction’s heat balance using one mole of feedstock consumed as a basis: Heat released = (α ) × [(MW biomass) × ( − ΔH C )] − (− − ΔH S ) × (MW substrate)

(9.66)

where (–ΔHC) and (–ΔHS) are heat of combustion per gram of cell per gram of substrate, respectively. Rearranging: Heat released ⎛ MW biomass ⎞ = (α ) × ⎜ × (− ΔH C ) − (− ΔH S ) ⎝ MW substrate ⎟⎠ MW substrate

(9.67)

The left side is heat released per gram of substrate consumed and the coefﬁcient of the ﬁrst term on the right is the growth yield, that is: YΔ /S = (YX /S ) × (− ΔH C ) − (− ΔH S )

(9.68)

where YΔ/S is the heat yield, which is determined on the basis of substrate consumed. So, dividing the above by YX/S yields: YX /S = (− ΔH C ) −

(− ΔH S ) (YX /S )

(9.69)

PH2270_book.fm Page 245 Tuesday, May 16, 2006 9:37 AM

Bioreactors

245

Both Eq. 9.70 and 9.71 are useful in determining heat release resulting from growth, YΔ/X and substrate consumption, YΔ/S. Heat Release from Extracellular Products When signiﬁcant product is present, Eq. 9.59 is modiﬁed to: Heat released = (α) × (MW biomass) × (− ΔHc ) + (β) ( MW product) × (− ΔH p ) − (− ΔH s ) × ( MW substrate)

(9.70)

where (–ΔHp) is heat of combustion per gram of extra cellular product(s). Dividing the above by MW of substrate gives: YΔ /S = (YX /S ) × (− ΔH c ) + (YP /S ) × (− ΔH p ) − (− ΔH S )

(9.71)

Dividing by the above yields: YΔ / X = (− ΔHc ) + (YP / X ) ∗ (− ΔH P ) −

(− ΔH S ) (YX / S )

(9.72)

Modeling Bioreactor Sterilization Bioreactor sterilization cycles are composed of heating, holding, and cooling stages. The total Del factor (∇) (i.e., measure of the size of the task to be accomplished) required for a complete cycle should, therefore, be equal to the sum of the heating, holding, and cooling Del factors: ∇ total = ∇ heat + ∇ hold + ∇ cool

(9.73)

Values of ∇Heat and ∇Cool The values of ∇heat and ∇cool are predetermined by the methods employed for the heating and cooling stages and the value of ∇hold is determined by the length of the controlled holding period. Holding time is estimated by: (1) calculating the total sterilization standard factor ∇total; (2) measuring the temperature versus time proﬁle during the heating, holding, and cooling stages of the sterilization cycle (if experimental measurements are not practical, theoretical equations for heating and cooling can be employed, which can be of linear, exponential, or hyperbolic form, depending upon the mode of heating and cooling). For batch heating by direct steam sparging into the medium, the hyperbolic form should be used: T = T0 +

Hmst c( M + m s t )

(9.74)

PH2270_book.fm Page 246 Tuesday, May 16, 2006 9:37 AM

246

Producing Biomolecular Substances

for constant-rate heat ﬂow, the linear form should be used T = T0 +

qTt cM

(9.75)

for isothermal heating, the exponential form should be used ⎛ uAt ⎞ T = TH + (T0 − TH )exp ⎜ − ⎝ cM ⎟⎠

(9.76)

and for batch cooling using a continuous non isothermal heat sink (e.g., passing cooling water through the vessel jacket), the linear form should be used ⎧⎪ ⎡ ⎛ UA ⎞ ⎤ mct ⎫⎪ T = TC0 + (T0 − TC0 )exp ⎨ ⎢1 − exp ⎜ − ⎬ ⎝ mc c ⎟⎠ ⎥⎦ M ⎭⎪ ⎩⎪ ⎣

(9.77)

(3) then plotting the values of kd as a function of time; (4) integrating the areas under the kd-vs.-time curve for the heating and cooling periods, to estimate ∇heat and ∇cool, respectively (if using theoretical equations integrate numerically after substituting the proper temperature proﬁles): ∇ = ln

n0 = n

∫

kd dt = kd0

⎛

Ed ⎞

∫ exp ⎜⎝ − RT ⎟⎠ dt

(9.78)

and (5), calculating the holding time from the expression t hold =

∇ hold ∇ total − ∇ heat − ∇ cool = kd kd

(9.79)

Continuous sterilization simpliﬁes production planning, allows maximum facility utilization and minimum delay, provides reproducible conditions, can be operated at higher temperatures than batch sterilization (140°C vs. 121°C so that the sterilization time can be shortened to only one or two minutes), requires less steam and cooling water since it recovers heat from the sterilized medium, and is less laborintensive (easier to automate). A continuous sterilization system consists of three main units: heating, holding, and cooling. Heating Heating is generally of two types: (1) direct heating (e.g., steam injection); and (2) indirect heating (e.g., shell-and-tube or plate-and-frame heat exchangers). Direct heating is usually more effective since there is no barrier between the medium and

PH2270_book.fm Page 247 Tuesday, May 16, 2006 9:37 AM

Bioreactors

247

the heat source. Steam injection rapidly heats the medium to the ultimate sterilization temperature, so sterilization during heating is negligible. For indirect heating, plateand-frame heat exchangers are generally more effective for heat transfer than the shell-and-tube type because of their larger heat-transfer areas; they are also better for sterilizing high-viscosity media since plate-and-frame heat exchangers are limited to pressures less than 20 atm because of their weaker structural strength. Sterilization can be expressed as: In – Out – Killed by Sterilization = Accumulation

(9.80)

with accumulation equal to zero at steady state. Cell input and output both have a bulk-ﬂow and an axial-diffusion phase, and efﬁciency increases with increasing particle diameter and/or air ﬂow velocity. Thermal death of microorganisms is described by ﬁrst-order kinetics: dn = − kd n dt

(9.81)

where kd is the speciﬁc death rate that depends on both the species and physiological form (e.g., the value of kd for bacterial spores at 121°C is on the order of 1 minute and for vegetative cells it can vary from 10–101 minutes (depending on the particular organism). Exponential decrease in cell population is shown by:

n =− n0

∫ k dt

(9.82)

or by ⎛ n = n0 exp ⎜ − ⎝

⎞

∫ k dt ⎟⎠ d

(9.83)

The dependence of the speciﬁc death rate kd on temperature follows Arrhenius: ⎛ E ⎞ kd = kd0 exp ⎜ − d ⎟ ⎝ RT ⎠

(9.84)

where Ed is activation energy obtained from the slope of the (ln) kd-vs.-1/T plot. Del Factor (∇) Thus, the sterilization design standard (∇), can be deﬁned as a Del factor that increases as the ﬁnal number of cells approaches null. However, if even one organism survives, the entire culture is still contaminated. The Del factor to reduce the number

PH2270_book.fm Page 248 Tuesday, May 16, 2006 9:37 AM

248

Producing Biomolecular Substances

of contaminating microorganisms to zero is inﬁnity (i.e., it is theoretically impossible to ensure the total destruction of all viable cells). Practically, though, the ﬁnal number of contaminating microorganisms is expressed as a fraction equal to the statistical probability of contamination (e.g., with n = 0.001, the chance for a contaminant surviving the sterilization is 1:1,000), therefore, the Del factor to reduce the cell number in a bioreactor from 1010 viable microorganisms to 0.001 is: ∇ = ln

1010 = 30 0.001

(9.85)

with the sterilization unit’s design based on the calculated Del factor. Steam sterilization is most commonly used in SIP systems and sterilization system design should ensure that the live steam comes in contact with all sites exposed to process materials, gasses, culture material, and/or product. All inlet and outlet ports, supply lines, harvest lines, sensors, regulators, the reactor vessel itself, and any piping that carries materials critical to the process must be sterilized. Steam should reach all of these sites at a temperature and pressure adequate to destroy contaminating organisms and must remain at these levels for the speciﬁed period of time. Process engineers generally design STBs and associated system components to be compatible with steam sterilization (e.g., the pitch of the piping and the positioning of ports, feed lines, and regulatory devices can affect their contact with the steam) and, as previously discussed, it is essential that designs eliminate dead spaces, ridges, and crevices where microorganisms can evade sterilization.

GROWTH KINETICS

AND

PRODUCT FORMATION

Growth Kinetics In this section bioreactor analysis is limited to batch-culture systems. If a viable inoculum is introduced into a medium that contains a carbon source, a suitable nitrogen source, other nutrients necessary for growth, and an appropriate temperature and pH are maintained, the culture will grow with the rate of biomass synthesis proportional to the biomass present. That is: rx = μX

(9.86)

where rx is the number (amount) of cells produced in gL–1 h–1, and X is the cell concentration in gL–1. The parameter μ is called the specific growth rate, which is analogous to the speciﬁc rate constant in chemical reactions. The treatment of chemical reactions is recalled and summarized below for easy reference: Reaction:

A→ B

Rate Expression: − rA = kC A

(9.87)

PH2270_book.fm Page 249 Tuesday, May 16, 2006 9:37 AM

Bioreactors

249

In the above CA is concentration of A (mol AL–1), –rA is reaction rate (mol AL–1 h–1) and k is rate constant (h–1). The negative sign in front of –rA is to comply with the deﬁnition of rA, which is the rate of generation of A. In Eq. 9.88, the negative sign is not necessary since X increases with time. Consider the cell balance in a batch bioreactor: Cells in–Cells out + Generation of cells = Accumulation of cells in Bioreactor: 0 − 0 + (rx ) ∗ (V ) =

d (Vx ) dt

(9.88)

Substituting for rx from Eq. 9.88 and noting that volume of reactor is constant gives: dX = μX dt The above can be expressed as μ=

1 dX ⎛ ΔX ⎞ ⎛ 1 ⎞ =⎜ ⎟ ×⎜ ⎟ X dt ⎝ X ⎠ ⎝ Δt ⎠

(9.89)

The term, DX/X, represents the fractional increase in cell amount and Dt is the time over which this fractional increase was generated. That is, μ is a fraction of the biomass formed per unit of time. So, if μ is 0.3h–1, the biomass will approximately increase by 30% every hour. We use the term approximately because we use ﬁnite quantities to describe a rule applied at inﬁnitesimal scale. Treating μ as a constant, Eq. 9.89 can be integrated to show: X = X 0 Exp(μt )

(9.90)

where X0 is the initial (inoculum) cell concentration. The time t refers to the time since the inoculum emerged from lag phase. Eq. 9.90 can be rearranged, setting the ⎛ x⎞ conditions for doubling the biomass. That is ⎜ ⎟ = 2 and t = the doubling time, td: ⎝ x0 ⎠ ⎛ ln(2 ) ⎞ ⎛ 0.693 ⎞ = td = ⎜ ⎝ μ ⎟⎠ ⎜⎝ μ ⎟⎠

(9.91)

Doubling time and speciﬁc growth rate have been reported by many researchers; typically the values for murine hybridomas average a growth rate value [μ[h–1]] of 0.05 and the doubling time value td of 13.9 (see Figure 9.4).

PH2270_book.fm Page 250 Tuesday, May 16, 2006 9:37 AM

250

Producing Biomolecular Substances

mmax

Specific Growth Rate (m)

Substrate Concentration (S)

FIGURE 9.4 Monod growth curve. 0.5

μ, 1/h

0.4 0.3 0.2 0.1 0 0

S, g/L

FIGURE 9.5 Monod kinetics with a one substrate limitation.

The Specific Growth Rate Value (μ) The speciﬁc growth rate value μ depends upon a number of factors such as growth medium composition, temperature, pH, and numerous others. Experimental studies demonstrate that one cannot increase growth rate beyond a certain maximum, μmax, because of inherent metabolic reaction rate limitations. In general, when substrate S is limiting growth, Monod (1949; see Figure 9.5) reported that growth rate variations can be expressed as: μ=

μ mS KS + S

(9.92)

where KS is called the substrate saturation constant or simply the Monod constant. The signiﬁcance of KS is that when substrate concentration is equal to KS, the growth

PH2270_book.fm Page 251 Tuesday, May 16, 2006 9:37 AM

Bioreactors

251

0.5 DO = 8 mg/L

μ, 1/h

0.4

DO = 0.2 mg/L

0.3 DO = 0.1 mg/L

0.2 0.1 0 0

S, g/L

FIGURE 9.6 Monod kinetics with growth rate limited by two substrates.

rate is exactly half of the maximum growth rate. The speciﬁc growth rate reaches a maximum value of 0.5 h–1; the value of KS here is 0.5 gL–1. Note that when S = 0.5 gL–1, μ is half maximum; this form can be used to describe dependence of μ on more than one limiting nutrient. In many practical applications lowered oxygen availability often limits growth. When both substrate S and dissolved oxygen concentration CDO2 both limit growth, the speciﬁc growth rate can be mathematically described as: μ=

μ mS C DO × K S + S K DO + C DO

(9.93)

Figure 9.6 illustrates the behavior of the maximum growth rate when two substrates limit growth. Where parameters KS and KDO are cell-speciﬁc, and KS is typically on the order of 10 mg/L of nutrient substrate, KDO is more than 1 mg/L of oxygen in the case of mammalian and insect cells. Now, considering growth conditions of substrate limitation in batch bioreactors only: incorporating the substrate’s limited condition, the bioreactor material balance equation, Eq. 9.93, can be modiﬁed to use: dX μ S = m X dt K S + S

(9.94)

In order to integrate the above, one of the variables, S, needs to be replaced in terms of X. The yield relationship, so Eq. 9.94, can be integrated as:

∫ This simpliﬁes to:

X X0

dX = −YX /S

∫

(9.95)

PH2270_book.fm Page 252 Tuesday, May 16, 2006 9:37 AM

252

Producing Biomolecular Substances

⎛ X − X0 ⎞ S = S0 − ⎜ ⎝ YX /S ⎟⎠

(9.96)

where subscript 0 refers to initial concentration and, substituting for S from Eq. (9.95) in Eq. (9.96) and integrating, yields: ⎛ K SYX /S + S0YX /S + X0 ⎞ ⎛ X ⎞ ⎛ K SYX /S ⎞ ⎛ YX /S S0 + X0 + X ⎞ ⎜⎝ ⎟⎠ ln ⎜⎝ X ⎟⎠ − ⎜⎝ Y S + X ⎟⎠ ln ⎜⎝ ⎟⎠ = μ mt (9.97) YX /S S0 + X0 YX /S S0 0 X /S 0 0 For analyzing batch systems, the above is used to calculate cell concentration and thereafter calculate substrate concentration using Eq. 9.98. Metabolic Quotient and Rate Expression Rate expressions for cell growth were previously covered. So now let us examine rate expressions for other medium components in the growth reaction. Consider that on the basis of one g of substrate consumed, the growth reaction can be written as: 1 g S + YO2 /S g of O 2 + YNH 3 /S g of NH 3 = YX/S g of Biomass + YCO2 /SCO 2 + others

(9.98)

The stoichiometric coefﬁcients then become yield coefﬁcients on the basis of substrate. Thus, the general rate expression is then: rS r r r r = O2 = NHS = X = CO2 −1 −YO2 /S −YNHS /S YX /S YCO2 /SS

(9.99)

where ri is expressed in g of i L–1 h–1. Since rx is the most fundamental of the various rates, it is conventional to write the stoichiometric coefﬁcient in terms of it, that is:

rS =

rX −YX /S

(9.100)

Y r rO2 = rX O2 /S ⇒ X YX /S Y X /O 2 Following the above, the rate expression for species i can then be written as: ri =

rx Y X /i

(9.101)

PH2270_book.fm Page 253 Tuesday, May 16, 2006 9:37 AM

Bioreactors

253

Metabolic quotients are rate expressions on the basis of biomass unit mass; that is: qi =

μ ri ⎛ r ⎞ ⎛ 1⎞ ⇒⎜ x ⎟ ×⎜ ⎟ ⇒ ⎝ Y X /i ⎠ ⎝ X ⎠ X Y X /i

(9.102)

The metabolic quotient for oxygen is of special interest, since this property determines the upper limit of cell concentration possible in many cell growth systems. Typical metabolic coefﬁcient values of a murine hybridoma for glucose, g glucose (g cell)–1 is 0.2 m and for h–1 qO2, g glucose (g cell)–1 h–1 is 0.02. Factors Affecting Growth Rate Nutrients, pH, temperature, DO2 concentration, and other environmental conditions affect the growth rate. Temperature and pH dependence are illustrated in Figures 9.7a and 9.7b. In Figure 9.7a the maximum growth rate is observed at 39°C for E. coli. Product formation kinetics for insulin, product yield (YP/S), and cell yield (YX/S) are also affected by temperature — and cell yield typically decreases with temperature. Nevertheless, similar deﬁning relationships for end products have not been reported. It should be noted that optimal growth temperature may well be different from the optimal temperature for end product formation. The optimal eukaryotic cell growth pH range is 6.2–7.2 for mammalian cells. As previously noted, optimal pH for end product formation may be different from that required for optimal growth. Many cells produce a different end product mix when the pH is altered. Hybridomas, for example, produce antibodies at a higher rate at pH 6.2 than at pH 7.2, and because of the difference in conditions for growth and product formation, optimization is often necessary. Oxygen is an important substrate for aerobic organisms since the cellular production of metabolic energy is directly related to the oxygenation rate, and oxygen concentration is very strongly coupled to the cell growth rate that typically depends on the DO2 concentration (see Figure 9.8). The critical dissolved oxygen concentration refers to values below that where the growth rate is lower than the maximum value, and the growth rate sharply rises to its maximum value along with the dissolved oxygen concentration. The concentration at which maximum growth is attained is referred to as the critical –1 oxygen concentration, CCRIT O2 , which is typically between 1–2 mg L for hybridoma cells. Remember, of course, that these values are going to be signiﬁcantly lower than the 37°C air saturation value of 6.7 mg L–1. Product Formation Kinetics Product formation falls into one of the following categories: 1. Growth-associated product formation (see Figure 9.9a). 2. Nongrowth-associated product formation (see Figure 9.9b). 3. Mixed-mode product formation (see Figure 9.9c).

PH2270_book.fm Page 254 Tuesday, May 16, 2006 9:37 AM

254

Producing Biomolecular Substances (a) 3 39°

37°

42° Growth Rate h–1

45°

30°

1 21°

47°

17° 0.5

48° 13.5°

0.1 3.1 3.15 3.2 3.25 3.3 3.35 3.4 3.45 3.5 3.55 1,000 / T(K) 1.1 Dimentionless Specific Growth Rate μ/μm

(b)

1 0.9

With Adaptation

0.8 0.7 0.6 0.5

Without Adaptation

0.4 0.3 0.2 0.1 0 2

FIGURE 9.7 (a) Effect of temperature on growth rate; maximum growth rate is at 39°C. Plot is a function of inverse absolute temperature; the declining line from 39°C to 21°C to 13°C suggests that the growth rate behaves similar to chemical reaction rate constants. (b) Effect of pH on growth rate; typical pH ranges over which reasonable growth can be expected are about 1–2 units; with adaptation broader ranges can be achieved.

Typical time proﬁles of these situations are illustrated above. In Type I, the product is simultaneously formed with the growth of cells. That is, product concentration increases with cell concentration. The metabolic quotient for P can be expressed as a function of μ, or: rP = qP X ⇒ αμX qP = YP / X μ

(9.103)

PH2270_book.fm Page 255 Tuesday, May 16, 2006 9:37 AM

Bioreactors

255

Dimensionless Growth Rate μ/μm

1 E-Coli 0.95 0.9 0.85 0.8 0.75 0.7 0

0.1

0.2 0.3 D.O (mg/l)

0.4

FIGURE 9.8 Effect of dissolved oxygen concentration (DO2) on growth rate; critical dissolved oxygen concentration refers to the vaue od DO2, below which the growth rate is lower than maximum.

It is clear from the above, the proportionality constant, is the yield coefﬁcient, YP/X. In Type II, product formation is unrelated to growth rate but is a function of cell concentration, and is expressed as: rP = qP X ⇒ βX

(9.104)

Hybridomas forming antibody exhibit type II behavior. In the third category, product formation is a combination of growth rate and cell concentration: rP = qP X ⇒ (αμ + β) X

(9.105)

Therefore, for a simpler analysis, lets consider a general case. In a batch reactor, product turnover is accomplished by running a mass balance on the product, i.e.: Rate of Product Formation = Accumulation of Product

(rp ) × (V ) =

d (VP ) dt

(9.106)

So, for a constant V: dP = rp = (αμ + β) X dt

(9.107)

PH2270_book.fm Page 256 Tuesday, May 16, 2006 9:37 AM

256

Producing Biomolecular Substances

Cell or Product Concentration

(a)

Cell, X

Product, P

Time

Cell or Product Concentration

(b)

Cell, X

Product, P

Time

Cell or Product Concentration

(c)

Cell, X

Product, P

Time

FIGURE 9.9 (a) Growth-associated product formation. (b) Nongrowth-associated product formation. (c) Mixed-mode product formation.

PH2270_book.fm Page 257 Tuesday, May 16, 2006 9:37 AM

Bioreactors

257

And, if we consider the exponential phase only, X = X0 Exp (μmt), substituting in the above yields: dP = (αμ + β) X 0 Exp(μ mt ) dt

(9.108)

Integrating from t = 0, P = P0 we get: P − P0 =

(αμ + β) X 0 (Exp (μ mt ) − 1) μm

(9.109)

This expression may be used to calculate product concentration at the end of a bioreactor run.

OXYGEN TRANSFER

BIOREACTORS

Overview Oxygen is needed by all cells for respiration, whether the cells are in suspension or on microcarriers, and it must be available as dissolved oxygen (DO2). Since oxygen solubility is quite small (≅6–7 mg/L) under normal cultivation conditions, the metabolic oxygen requirement is usually supplied by continuous aeration of the culture medium. Thus, a continuous supply of oxygen must be maintained for aerobic bioprocesses. In this section we develop a quantitative appreciation for the metabolic oxygen demand, followed by methods that can be used for calculating the rates at which the oxygen can be transferred from sparged air and methods useful for characterizing the oxygen mass-transfer coefﬁcient. Finally bioreactor operation and design are evaluated based upon their oxygen transfer capabilities. Metabolic Oxygen Demand An organism’s metabolic oxygen demand depends on the biochemical nature of its cells and their growth conditions. The need for oxygen is satisﬁed for most cells by keeping the dissolved oxygen concentration in the medium at about 1 mg/L. If the oxygen level falls far below this value, oxygen consumption decreases, with a concomitant decrease in cellular energy production and, as a result, cell growth. This behavior has already been described mathematically in the Fermenter chapter of this book. The oxygen concentration value above which growth rate is at maximum CRIT has been described as the critical oxygen concentration CO2 . The oxygen requirement for cell growth is expressed best in the parameter, yield coefﬁcient YX/O2, and represents the amount of oxygen required to grow one gram of cells. Volumetric Oxygen Mass-Transfer Coefficient In a typical aeration system, oxygen from the dissolved air bubble is transferred through the gas-liquid interface followed by liquid phase diffusion/bulk transport to

PH2270_book.fm Page 258 Tuesday, May 16, 2006 9:37 AM

258

Producing Biomolecular Substances

Gas-fluid phase boundary

Fluid-cell phase boundary

CG CX 1 2 4

CY 3 Gas bubble Gas Fluid Film Film

5 CF

Fluid

Fluid Film

CZ Cell

FIGURE 9.10 Oxygen transfer resistances from medium air bubble to microbe; a critical factor for a bioreactor is adequate gas exchange. Oxygen is typically the most important gaseous substrate for cell metabolism, and carbon dioxide is the most important gaseous metabolic product. For oxygen to be transferred from an air bubble to an individual microbe, some resistances for oxygen transfer from air bubbles in the medium to cells include: (1) resistance within the gas ﬁlm to the phase boundary; (2) resistance to penetration of the phase boundary between the gas bubble and the liquid, (3) resistance to transfer from the phase boundary to the liquid, (4) resistance to movement within the nutrient solution, and (5) resistance to transfer to the surface of the cell.

the cells. Although this is a multistep serial transport in a well-dispersed system, the major resistance to oxygen transfer is the liquid ﬁlm encasing the gas bubble (see Figure 9.10). Consider oxygen concentration proﬁles in the region near the interface (see Figure 9.11), where oxygen concentration at the air bubble-medium interface is equal at steady state to the transport of oxygen through both the gas and liquid ﬁlms. This can be expressed by: NO2G = kG A(C DOG − C DOGi )

(9.110a)

NO2 L = k L A(C DOLi − C DOL )

(9.110b)

NO2G = NO2 L

(9.110c)

where subscript G and L refer to the gas and liquid phases respectively; the terms NO2G and NO2L are the oxygen transfer expressed in g O2h–1, where A is the interfacial area and CDO is the oxygen concentration expressed in g O2 per unit volume. At the interface, equilibrium between liquid and gas phase oxygen is reached. That is: C DOGi = mC DOLi

(9.111)

PH2270_book.fm Page 259 Tuesday, May 16, 2006 9:37 AM

Bioreactors

259

CDOG

CDOGi Liquid

Gas CDOLi

CDOL

FIGURE 9.11 O2 concentration proﬁle at air-bubble-medium interface; a critical factor for a bioreactor is adequate gas exchange.

In view of low oxygen solubility, and therefore that kG is much higher than kL: C DOG ≈ C DOGi

(9.112)

⎛C ⎞ NO2 = k L A ⎜ DOG − C DOL ⎟ ⎝ m ⎠

(9.113)

Eq. 9.112 can be rewritten as:

The subscript L in NO2 has been dropped since the above represents overall oxygen transfer. The driving force in this consists of the difference between bulk oxygen concentration in the two phases, where the ﬁrst term represents the oxygen concentration of the liquid that is in equilibrium with the gas-phase oxygen. If air is the gas medium, it equals 7mg/L at 35°C. When oxygen transfer is applied to the entire bioreactor volume, A represents the total interfacial area and kL represents the average mass-transfer coefﬁcient as bulk-gas and liquid-phase oxygen concentrations. If we divide this by the volume of the liquid phase V, the resulting term represents the amount of oxygen transferred per unit volume per unit time, which falls in the same units as the rate expressions developed in the last section. And, since the rate is the result of a physical phenomenon, let’s distinguish it here by the symbol, RO2. That is: ⎛ A⎞ ⎛ C ⎞ RO2 = k L ⎜ ⎟ ⎜ DOG − C DOL ⎟ ⎝V⎠⎝ m ⎠

(9.114)

The term kL A represents the mass-transfer coefﬁcient and interfacial area available for mass transfer in the bioreactor, where the air is sparged and the liquid agitated

PH2270_book.fm Page 260 Tuesday, May 16, 2006 9:37 AM

260

Producing Biomolecular Substances

to disperse bubbles so that the interfacial area can be maintained high to enhance the oxygen transfer rate. In these systems, the area, A, is not easily measured or even estimated. But, a term consisting of the product-mass-transfer coefﬁcient and the interfacial area is much more easily measured. Furthermore, it is convenient to use the interfacial area per unit volume a rather than total area, A, since the oxygen transfer rate is expressed per bioreactor unit volume, similar to that of the rate of cell growth, which is reported on a volumetric basis. Thus, the area per unit volume a is combined with the mass-transfer coefﬁcient kL and known by the term kLa. In Eq. 9.114 the term

C DOG can be replaced by oxygen solubility in the conditions m

∗ maintained in a bioreactor, C DOL :

* RO2 = k La (C DOL − C DOL )

(9.115)

Eq. 9.115 is the working equation for describing oxygen transfer from gas phase to growth medium, and, in order to calculate the oxygen transfer rate (OTR), we need * , and the the mass-transfer coefﬁcient kLa, the oxygen solubility of the medium C DOL actual DO2 concentration in the medium CDOL. In the last section we used the term CDO to describe dissolved oxygen concentration. In this discussion, there is a need to make a distinction between gas- and liquid-phase concentrations. In Eq. 9.115, both concentrations are expressed on the basis of the liquid phase. Therefore, from here on we will drop the subscript L. Also it will be reintroduced in situations where it is necessary to make a distinction between the two phases. Bioreactor Oxygen Balance In the case of bioreactor oxygen balance, where its cells are growing and consuming oxygen, there is a continuous inﬂow of air at constant volumetric ﬂow rate, and the liquid medium containing the inoculated cell culture is mixed by a marine impeller — if the metabolic oxygen uptake rate is qO2, the cell concentration is X, and the bioreactor system function is calculated over a sufﬁciently short period — X can be treated as a constant. Consider the oxygen balance in the liquid phase of the bioreactor: the O2 transferred from the gas phase minus the O2 consumed by the cells equals the Oxygen Accumulation: * [ k La (C DO − C DO )] × V − qO2 XV =

d (VC DO ) dt

(9.116)

For a constant liquid-phase volume, this can be simpliﬁed to: d (C DO ) * = k La (C DO − C DO ) − qO2 X dt

(9.117)

PH2270_book.fm Page 261 Tuesday, May 16, 2006 9:37 AM

Bioreactors

261

The concentration, CDO is readily measured using a dissolved oxygen electrode. If the oxygen being supplied is in exact balance with the oxygen consumed by the cells, the dissolved oxygen concentration remains a constant; that is, the derivative in Eq. (9.117) will vanish: * qO2 X = k La (C DO − C DO )

(9.118)

One useful application of this is in estimating the maximum hybridoma cell concentration that a particular bioreactor is capable of supporting in terms of its oxygen supply. Example: A bioreactor has an oxygen mass-transfer capability coefﬁcient of 400h–1. What is the maximum concentration that a mammalian hybridoma cell culture grown in this bioreactor can attain if the respiration rate of the cell culture is 0.35g O2 (g Cell)–1h–1, the critical oxygen concentration is 0.2mg/L, and the oxygen saturation of the air is 6.7mg/L. Solution: From Eq. 9.118, we have:

* k La (C DO − C DO ) qO2

(9.119)

The maximum expected oxygen concentration is equal to:

(6.7 – 0.2) = 6.5 mg/L

(9.120)

Therefore, the maximum cell concentration that can be grown at a maximum growth rate is:

Xmax =

* k La (C DO − C DO )max ( 400 h −1 ) × (6.5 mgO 2 L−1 ) ⇒ 7.4 gCell L−1 (9.121) ⇒ qO2 0.35 gO 2 (gCell)−1h −1

Factors Affecting Mass-Transfer Coefficient The mass-transfer coefﬁcient KLa is strongly affected by both agitation speed and the air-ﬂow rate. In general: K La = k ( Pg / VR )0.4 (VS )0.5 ( N )0.5

(9.122)

where k is a constant, Pg is the power required for aerated bioreactor, VR is the bioreactor volume, VS is air ﬂow rate, and N is agitator speed.

PH2270_book.fm Page 262 Tuesday, May 16, 2006 9:37 AM

262

Producing Biomolecular Substances

Note: the mass-transfer coefﬁcient increases with agitation speed and air-ﬂow rate. Example: You are part of a tech service team asked to evaluate if a 10,000 L bioreactor unit is adequate to produce 10 kg/day of antibody protein using a hybridoma strain that expresses the protein as 5% cellular protein. In order to enhance stability, nutrients are manipulated to give a low speciﬁc growth rate of 0.2 h–1. The oxygen demand is 0.15 g O2/g cell Δ–h. Assume that the end-product protein formation is cell-growth related. Data: Assume a lag phase of 4 hours, the typical cleaning time following a batch and the preparation for the next batch is 8 hours, and the plant runs three shifts. Cell yield on the substrate is 0.55 g cell/g of substrate. Available support services can supply a maximum inoculum of 6 kg of cells every 24 hours. Maximum KLa for the bioreactor is 500h–1. The bioreactor’s accessories are capable of handling cell concentrations of 60g/L. Assume any other parameters you need to complete the calculation, and assume that critical oxygen concentration is 0.2 mg/L and the DO2 at air saturation is 6.4mg/L.

MODELING EUKARYOTIC CELL GROWTH

AND

PRODUCTION

Kinetic Growth Understanding cell growth kinetics is necessary for the correct design and operation of a bioreactor and bioproduction. Cell kinetics can be described as the consequential interaction of numerous complicated biochemical reactions and transport phenomena, involving multiple stages of multicomponent systems. During growth, a heterogeneous mixture of young and old cells is continually transforming and adapting to its changing environment. Thus, a completely accurate mathematical modeling of the system’s growth kinetics is virtually impossible. Therefore, to derive simpler bioreactor operation and performance models that can be expressed mathematically, some assumptions must be made regarding cell components and population dynamics, as presented in Table 9.10. The simplest model, the unstructured, distributed model, is based on two assumptions: (1) Cells can be represented by a single component, such as mass, number, or concentration of protein, DNA, or RNA. This is particularly true for balanced growth, because in balanced growth, doubling cell

TABLE 9.10 Kinetic Growth and Production of Cellular Models Population Distributed

Segregated

Cell Components Unstructured Cells are represented by a single component, which is uniformly distributed throughout the culture. Cells are represented by a single component but form a heterogeneous mixture.

Structured

Multiple cell components are uniformly distributed throughout the culture interact with each other. Cells are composed of multiple components and form a heterogeneous mixture.

PH2270_book.fm Page 263 Tuesday, May 16, 2006 9:37 AM

Bioreactors

263

mass is accompanied by doubling all other measurable population properties. (2) The cell mass is distributed uniformly throughout the culture (i.e., the culture can be regarded as hom*ogeneous; the heterogeneous nature of the cells is ignored, and the cell concentration is expressed as either wet or dry weight per unit volume). In addition to such assumptions, the medium must be formulated so that only one component limits the growth rate and all the other components are present at sufﬁciently high concentrations so minor changes do not signiﬁcantly affect cell growth and proliferation. The growth environment is also controlled so that parameters such as pH, temperature, and DO2 concentration are constant. Hybridoma cell kinetic growth equations will now be developed in this section, and subsequently will be applied to the analysis and design of the “perfect” mABproducing bioreactor. Accurate, structured models (more realistic models that consider the multiplicity of cellular components) are discussed later on, and growth rate is variously deﬁned. While dCX/dt and rX appear to be the same, the two terms are only equivalent in batch operation. The expression dCX/dt is the change in the bioreactor’s cell concentration, which includes the consequence of in and out media ﬂow, cell recycling, and other operating conditions, while rX is the actual cell growth rate. The growth rate based on the number of cells and cell weight are also not equivalent since the average cell size varies considerably from one stage to another. To illustrate, when an individual cell mass increases without division, the growth rate based on cell weight increases, while the growth rate based on the number of cells remains constant. During exponential growth, however, the growth rate based on cell number and growth rate based on cell weight are proportional. Growth rate is sometimes confused with division rate (the rate of cell division per unit time). If all the cells in a vessel at time t = 0 (Cn = Cn0 ) have divided once after a certain period of time, the cell population will have increased to Cn0 × 2. If cells are divided N times after the time t, the total number of cells will be: Cn = Cn0 × 2 N

(9.123)

1 δ = (log2 Cn − log2 Cn0 ) t

(9.124)

and the average division rate is:

and the division rate at time t is:

δ=

d log2 Cn dt

(9.125)

Thus, growth rate can be expressed as the change in cell number with time or the slope of the Cn-vs.-t curve (see Figure 9.12), while the division rate is the slope of the log2-Cn-vs.-t curve. Since the division rate is constant during exponential growth,

PH2270_book.fm Page 264 Tuesday, May 16, 2006 9:37 AM

Death Phase

Stationary Phase

Log Phase

Lag Phase

Producing Biomolecular Substances

Log cell number

264

Time

FIGURE 9.12 Cell-culture growth curve.

and the growth rate is not, these two terms should not be confused. If fresh, sterile medium is inoculated and cell density is measured during subsequent growth and plotted against time, the results will differentiate six stages in the batch growth cycle: 1. Latent: the period during which the change in cell number is null. 2. Accelerated growth: the period during which the cell number begins to increase and the rate of cell division accelerates. 3. Exponential growth: the period during which the cell number increases exponentially, the growth rate increases, while the division rate remains constant at its maximum. 4. Decelerated growth: the period subsequent to the point where the growth rate reaches maximum during which both growth and division rates decrease. 5. Static: the period during which the cell population reaches maximum for the given conditions and proliferation stops. 6. Death: the period after the limiting growth substance is depleted where cells begin to die and the number of viable cells decreases. Kinetic growth data and models can predict growth stage lengths to estimate the required bioreactor size before considering the purchase of more complex models. In addition, the same models used for designing batch cell growth processes can also be used to predict the bioreactor size necessary for continuous-culture cell growth. The key objective of kinetic cell growth modeling is to support the growth of a speciﬁc culture and to promote elevated end product yield. In certain circ*mstances excessive essential nutrient concentration can inhibit growth, or even kill off the culture, so essential nutrients should not be supplied in excess. It is a common practice to limit the concentration of at least one of the essential nutrients, keeping all others in excess, so that growth increases exponentially until the essential (lim-

PH2270_book.fm Page 265 Tuesday, May 16, 2006 9:37 AM

Bioreactors

265

iting) substance is depleted. For single cells, growth rate can be expressed in terms of cell concentration X, and the speciﬁc growth rate where μ is time is deﬁned by: μ=

1 dX X dθ

(9.126)

If the value of μ is constant, the culture is growing exponentially. Typically, batch cell growth and metabolic end product synthesis demonstrate exponential increases until limiting nutrient concentration decreases to a level at which the cells begin to die. Thus, the perfect time to terminate bioreactor batches is when cell proliferation at the same rate as end product formation. At times, when product formation lags behind cell proliferation, cell growth is allowed to proceed until desired product concentration is reached even though cells are in the death stage. Mass doubling time (μd) is determined, and the speciﬁc growth rate for single-cell growth is related to mass doubling time by the formula μd = ln (2)/μ. Customarily, in the case of a single limiting substances, Monod is used: μ=

μ m Ci K i + Ci

(9.127)

where μm is the maximum speciﬁc growth rate, Ki is a saturation constant, and Ci is the limiting nutrient concentration i. The maximum speciﬁc growth rate μm is determined when Ci >> Ki and Ki is the concentration of nutrient i when μ = μm/2. Monod was derived empirically and is a simpliﬁed model of complex cell growth. Determination of μm and Ki produces a measure of population growth that can be used to predict culture population dynamics on a large scale, particularly when nutrient uptake is also established. With a more complex environment (e.g., two limiting nutrients, toxin formation, inhibition by product concentration, etc.), variations of the basic Monod formula can determine the speciﬁc growth rate μ at any concentration and at the same time predict maximum cell population for a culture under speciﬁc circ*mstances such as with two limiting nutrients: ⎛ C1 ⎞ ⎛ C2 ⎞ μ = μm ⎜ ⎝ C1 + K1 ⎟⎠ ⎜⎝ C2 + K 2 ⎟⎠

(9.128)

⎛ Ci ⎞ ⎛ C P ⎞ μ = μm ⎜ ⎝ Ci + K i ⎟⎠ ⎜⎝ CP + K P ⎟⎠

(9.129)

or, with product inhibition:

where KP is the saturation constant for the product (Ki), and CP is the product concentration.

PH2270_book.fm Page 266 Tuesday, May 16, 2006 9:37 AM

266

Producing Biomolecular Substances

When an STB is inoculated, it will begin to proliferate exponentially after the latent stage, so that change in cell concentration in a batch culture is equal to its cellular growth rate: dC X = rX = μC X dt

(9.130)

Then, Eq. 9.130 is integrated to develop a performance equation for the batch culture:

∫

CX0

dC X = rX

∫

CX0

dC X = μC X

∫ dt = t − t

(9.131)

This only applies when rX is larger than zero. Therefore, t0 in Eq. 9.131 is not the time that the culture was inoculated, but the time that the cells actually start proliferating, which is also the beginning of the accelerated growth stage. According to Eq. 9.131, the batch growth time t – t0 is the area under the 1/rXvs.-CX curve between CX and CX0. Batch growth time is estimated by the CX-vs.-t curve, which is a more direct determination. The U-shaped curve is characteristic of autocatalytic reactions: S+X→ X+X

(9.132)

The autocatalytic reaction rate is slow at the start because the concentration of X functioning as a biocatalyst is low. The rate increases as cells multiply and reach maximum population density. As substrate is depleted and toxic metabolic waste products accumulate, the reaction rate approaches null. Monod kinetics satisfactorily represents the growth rate during exponential growth:

∫

CX0

( K S + CS )dC X = μ maxCSC X

∫ dt

(9.133)

C X − C X0 ΔC X = − ΔCS −(CS − CS0 )

(9.134)

where growth yield (YX/S) is: YX /S =

and, change in cell concentration with respect to time is: ⎞ C ⎛ CS K SYX / S K SYX / S (t − t0 )μ max = ⎜ + 1⎟ ln X + ln 0 ⎠ C X0 C X0 + CS0 YX / S CS ⎝ C X0 + CS0 YX / S

(9.135)

PH2270_book.fm Page 267 Tuesday, May 16, 2006 9:37 AM

Bioreactors

267

Monod parameters μmax and KS cannot be projected by a series of batch runs as easily as Michaelis-Menten parameters are determined with a series of kinetic enzyme reactions. With Michaelis-Menten, the initial reaction rate in batch cell growths can be measured as a function of substrate concentration. However, in cell growth where the initial rate is always zero because of the latent period during which Monod doesn’t apply. Although Monod has the same general form as Michaelis-Menten, their respective rate-reaction components are different; in Michaelis-Menten: dCP r C = max S dt K M + CS

(9.136)

dC X μ maxCSC X = dt K S + CS

(9.137)

and in Monod:

The latent stage is an initial growth period where cell population growth is either null or negligible, although cells are still able to increase in size. This period usually occurs while cells adjust to their new medium and environment before accelerating their growth rate. Factors such as cell type, age, inoculum size, and culture conditions determine a particular stage’s length. For example, if a culture is inoculated from a low-nutrient medium to a higher-nutrient medium, the length of the latent stage increases. If, on the other hand, the culture is inoculated from a high- to a lowernutrient medium, there is typically no latent stage. Another important factor affecting the length of the latent stage is inoculum size, therefore, if a small number of cells are inoculated into a larger volume, they typically experience an extended latent stage. In large-scale cell growth, a primary objective is to shorten the latent stage as much as possible; so, to inoculate a large bioreactor, a series of progressively larger seed lots are used to minimize any latent stage effect. Batch culture involves inoculating sterile medium with a seed culture of the cells to be grown. After inoculation, apart from air addition and the removal of waste gasses, nothing else is usually required. Rapid change to the sterile bioreactor, a new environment, can affect four important variables: (1) inoculation into a high-nutrient concentration medium can cause delayed cell growth until the culture adapts; (2) essential molecules synthesized by the cell to promote growth (vitamins, activators) may be lost by diffusion and may take time for replenishment; (3) inoculum size and the percentage of viable cells greatly affect the latent stage duration; (4) the maturity of the inoculum is important because newer cells have not stored the same quantity of required metabolic substances as cells already in exponential growth. The latent stage ends once a key cellular component reaches a critical level c within the cell, therefore: c = aV + bN 0θ1 + d θ1

(9.138)

PH2270_book.fm Page 268 Tuesday, May 16, 2006 9:37 AM

268

Producing Biomolecular Substances

where V is the volume of inoculum, N0 is the number of cells/new volume, a is the key cellular limiting component concentration/old volume x, b is the increase in key cellular component concentration/time per cell (for older cells), d is the internal cell production of the key cellular component (for newer cells), and θ1 is the time of the latent stage. Since θ1 is dependent on V, the larger V is, the shorter θ1 will be, so θ1 is proportional to 1/N0 for large inoculum volumes. To get maximum utilization, the bioreactor’s design should minimize the length of the latent stage. The following three points are important: (1) the inoculum should be as active as possible, preferably in exponential growth; (2) the medium of the inoculum should correspond as closely as possible to that of the bioreactor; and (3) a reasonably large inoculum volume, at least 5% of the total volume of the bioreactor, should be used to minimize the loss of key metabolic intermediates by diffusion. When accelerated growth begins, it gradually increases, reaching maximum growth during the exponential growth stage. This period is frequently called the accelerated-growth stage, and occasionally it is deﬁned as part of the latent stage. With single-cell organisms, a culture undergoing balanced growth emulates a ﬁrst-order autocatalytic reaction (e.g., the progressive doubling of cell number results in a continually increasing rate of growth). Thus, the rate of cell increase at any particular time is proportional to the number Cn of cells present at that time: rn =

dCn = μCn dt

(9.139)

where the constant μ is known as the speciﬁc growth rate (hr–1). The speciﬁc growth rate should not be confused with the growth rate which has different units and meaning. The growth rate is deﬁned as the change in cell number with time, while the speciﬁc growth rate is: μ=

1 dCn d ln Cn = Cn dt dt

(9.140)

or change in log of the cell number with time. Comparing Eqs. 9.139 and 9.140 shows that the speciﬁc growth rate μ is equal to ln2 times the division rate μ. If μ is constant with time during exponential growth, the expression can be integrated from time t0 to t yielding: Cn = Cn0 exp[μ(t − t0 )]

(9.141)

where Cn0 is the cell number concentration at t0 when the exponential growth starts. Thus, Eq. 9.141 shows exponential increase in the number of cells with respect to time. The time required to double the population, called the doubling time (td), can be determined by establishing Cn = 2 Cn0 and t0 = 0, and solving for τ:

PH2270_book.fm Page 269 Tuesday, May 16, 2006 9:37 AM

Bioreactors

269

td =

ln 2 1 = μ δ

(9.142)

The doubling time is inversely proportional to the speciﬁc growth rate and is equal to the reciprocal of the division rate. The exponential growth rate where X0 is the initial population at time 0 and X is the population at time μ for a key cellular limiting component μ, for cell population X at time μ is: ⎛ X ⎞ μ Cθ ln ⎜ ⎟ = m i ⎝ X 0 ⎠ Ci + K i

(9.143)

This predicts the culture’s exponential growth if Monod applies. However, Monod is not precise and if growth is extremely rapid the two following equations can be used: (1) where C0 is the initial key cellular component concentration: μ=

μ m Ci K i + C0

(9.144)

or (2), where B is a constant and N is the cell population: μ=

μ m Ci Ci + BN

(9.145)

When the rate of consumption of a key cellular limiting component dA/dμ is proportional to the number of viable cells reaching static stage, where X is the number of cells, and KA is a proportional constant for key cellular limiting component A. If the culture is in exponential growth, and if the initial concentration of A = A0 and X = X0 (initial concentration of cells, when A = 0, X = Xs (the static population), then this relationship yields the maximum population Xs for initial concentration of key cellular component A (A0) and initial cell population X0 at the start of exponential growth: XS = X0 +

μ A0 KA

(9.146)

Integrating the equation between X = X0 and X = Xs yields the time taken to reach static stage. In normal practice, cell growth is terminated either before the death stage is reached or just as the cell population starts to decrease as a result of total depletion of key cellular component A. An exception would occur, however, when product formation exhibits a sizable lag behind cell proliferation; thus, if product concentration is still increasing, cell growth could be taken into the death stage.

PH2270_book.fm Page 270 Tuesday, May 16, 2006 9:37 AM

270

Producing Biomolecular Substances

Factors Affecting Specific Growth Rate As previously stated, one of the most widely-employed equations for the effect of substrate concentration on μ is the Monod, an empirical expression based on the type normally associated with enzyme kinetics: μ=

μ maxCS K S + CS

(9.147)

where CS is the concentration of the limiting substrate and is a system coefﬁcient. The value of KS is equal to the concentration of the key metabolic limiting substance when the speciﬁc growth rate is half of maximum μmax. And, while Monod is an oversimpliﬁcation of the cell growth mechanism, it frequently describes cell growth systems with low cell–growth-inhibiting component concentrations. According to Monod, further increase in limiting substance concentration after μ reaches μmax does not affect the speciﬁc growth rate. Also, the speciﬁc growth rate decreases as substrate concentration is increased beyond a certain level. The following expressions have been used to improve Monod: μ=

μ maxCS K I1 + CS + ( K I 2 CS )2

μ = μ max (1 − e− CS / K S ) μ=

(9.148)

(9.149)

μ max (1 + K S CS− λ )

(9.150)

μ maxCS βn + CS

(9.151)

μ=

If several limiting substances are used, the following expression should be employed: μ = μ max

C1 C2 K1 + C1 K 2 + C2

(9.152)

If the limiting substance is the culture’s energy source, a certain amount of substrate is used for other purposes than growth, and some expressions include the term, ke, for cell subsistence: μ=

μ maxCS − ke K S + CS

(9.153)

PH2270_book.fm Page 271 Tuesday, May 16, 2006 9:37 AM

Bioreactors

271

When CS is so low that the ﬁrst term of the right side of Eq. 9.153 is –Ke, the speciﬁc growth rate is null. The alternative models appear to be better growth models for certain microorganisms, although their solutions are more difﬁcult than Monod’s. Cell growth produces metabolic wastes that collect in the medium and usually inhibit cell growth; their effect can also be added to Monod, as in: ⎛ CS ⎞ ⎛ K P ⎞ μ = μ max ⎜ ⎝ K S + CS ⎟⎠ ⎜⎝ K P + CP ⎟⎠ C ⎞ ⎛ CS ⎞ ⎛ μ = μ max ⎜ 1− P ⎟ ⎟ ⎜ ⎝ K S + CS ⎠ ⎝ CPm ⎠

(9.154)

(9.155)

CPm term is the maximum metabolic waste concentration allowed in the medium for uninhibited cell growth. The speciﬁc growth rate of cells or microorganisms is directly affected by medium pH, temperature, oxygen supply, etc., since optimal pH and temperature generally differ from one organism to another. When a culture’s numerical growth is limited by the exhaustion of the available limiting substance(s) and/or the accumulation of metabolic waste products, the rate of growth declines and eventually stops. At this point the culture is said to be in the static stage. Transition from exponential to static growth involves an erratic growth period during which various cellular components are synthesized at unequal rates and consequently have chemical compositions different from those of cells in exponential growth. The static stage is usually followed by a death stage. Death occurs either because of cellular energy reserve depletion and/or as a result of metabolic waste accumulation. Like the preceding growth stages, the death stage is exponential. In some cases, the organisms not only die but also decompose. Kinetic Models The material balance for a culture in a continuous stirred-tank bioreactor (CSTB) can be shown where rX is bioreactor cell growth rate and dCX/dt represents change in cell concentration with time. For valid steady-state bioreactor operation, the cell concentration change over time must equal zero (dCX/dt = 0), and since the culture grows quickly enough to replace cells lost through the outlet: τm =

V C X − C Xi = F rX

(9.156)

This reﬂects the required residence time as equal to CX- C xi times 1/rX on the 1/rX vs.-CX curve. Graphing cell residence times compares bioreactor effectiveness; the shorter the residence time in reaching a particular cell concentration, the more effective the bioreactor. If the input stream is sterile (CX = 0) and the cells are growing exponentially (rX = μ CX), for a CSTB at steady state with a sterile feed

PH2270_book.fm Page 272 Tuesday, May 16, 2006 9:37 AM

272

Producing Biomolecular Substances

the speciﬁc growth rate μ is equal to the dilution rate D and is equal to the reciprocal of the residence time μm since the speciﬁc growth rate of a culture can be controlled by changing the media ﬂow rate. So, if the system’s growth rate can be expressed by Monod, then: D=μ=

1 μ C = max S τ m K S + CS

(9.157)

CS can be calculated with known residence time and Monod kinetics as: CS =

KS τ mμ max − 1

(9.158)

It should be noted, however, that the preceding is only valid when μmμmax > 1. If μmμmax < 1, the growth rate of the cells is less than the rate of cells leaving the outlet stream, all the cells in the bioreactor will be washed out and Eq. 9.158 is invalid. If the growth yield (YX/S) is constant, then: C X = YX /S (CSi − CS )

(9.159)

⎞ ⎛ KS C X = YX /S ⎜ CSi − τ mμ max − 1 ⎟⎠ ⎝

(9.160)

⎛ ⎞ KS C = CPi + YP /S ⎜ CSi − τ mμ max − 1 ⎟⎠ ⎝

(9.161)

CX is:

and similarly, CP is:

Again, the above equations are valid only when μmμmax > 1, and the same values for CSTB cultures can also be calculated by determining material balances for substrate and product concentrations. Equality of the speciﬁc growth and dilution rates in steady state cell growth is helpful in studying the effects of media components on the speciﬁc growth rate. By measuring steady state substrate concentration at various ﬂow rates, kinetic models can be evaluated and parameters can be approximated. If the particular culture follows Monod, the plot of 1/μ-vs.-1/CS yields μmax and KS from the intercept and slope of a plot similar to that of Michaelis-Menten, with the advantage that it shows the relationship between the independent variable CS and the dependent variable μ. However, since 1/μ approaches inﬁnity μ as the substrate concentration decreases, undue importance is given to measurements made

PH2270_book.fm Page 273 Tuesday, May 16, 2006 9:37 AM

Bioreactors

273

at low-substrate concentrations and insufﬁcient importance is given measurements made at high-substrate concentrations. For a better parameter match, the following two linear relationships can be employed: CS K C = S + S μ μ max μ max

(9.162)

μ CS

(9.163)

μ = μ max − K S

Since input and output ﬂow must be precisely controlled, separate sterile nutrient and spent medium reservoirs are necessary for proper steady-state CSTB cell growths. Sometimes foaming and ﬁlter clogging by cell aggregates make control of outlet ﬂow quite difﬁcult, and since it typically takes several days (or even weeks) to reach steady state due to cellular mutation and adaptation to the new environment, there is also a high risk of contamination. The speciﬁc growth rate during CSTB cell growth can be estimated by measuring the slope of cell concentration vs. time. Substrate concentrations are determined at the same points, and plots can be constructed to determine particular kinetic parameters. The unit cost of biopharmaceuticals manufactured by continuous bioprocesses has come down over the years, and will continue to do so as products become more exotic. Thus, a majority of biopharmaceuticals currently in the regulatory pipeline will conceivably be produced by continuous culture. Key to both the biologic and ﬁnancial aspects of biopharm manufacturing is the relationship between continued cell growth and product formation. The cost of producing product is proportional to the cost of producing cells, which can be dramatically reduced if productive cell life can be extended. When employing continuous culture in a CSTB, the process typically maintains high densities of product-expressing cells for long periods of time. In these systems cells are harvested, medium containing expressed product is removed, and new medium is either added or perfused at regular intervals. Products maintained in incubated batch reactors are at greater risk of degradation, aggregation, oxidation, inactivation, and contamination by proteolytic enzymes and other cell lysis products than those produced by continuous culture because the product residence time in the CSTB is shorter and the removed end product can be chilled to 4°C to minimize degradation. By further reducing bioreactor vessel size, the facility dimensions, utility requirements, media handling problems, and downstream puriﬁcation costs can all be reduced, permitting the use of conventional clean-room technology. Continuous culture can reduce capital requirements for bioprocess facilities and put much less high-value product at risk during a particular time interval as compared with batch systems where everything can be lost if contamination, mechanical failure, or incubation failure should occur. The design and analysis of continuous culture systems are based on the CSTB. Four assumptions required for CSTB system analysis are:

PH2270_book.fm Page 274 Tuesday, May 16, 2006 9:37 AM

274

Producing Biomolecular Substances

(1) mixing takes place so that the exit stream contents have the same composition as the rest of the bioreactor vessel contents, (2) the concentration of all components in the bioreactor vessel is the same in all areas of the vessel, (3) if the process is aerobic, the concentration of dissolved oxygen is the same in all parts of the vessel, and (4), the heat transfer characteristics of the system are constant (i.e., heat generated by cell growth is continuously removed). When concentrations, cell population, and temperature in any area of the vessel do not change with time, a steady state has been reached and a mass balance can be applied to any component such that: ⎡ rate ⎤ ⎡ rate ⎤ ⎡ rate ⎤ ⎢ of ⎥ ⎢ of ⎥ ⎢ ⎥ of ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ addition ⎥ − ⎢ removal ⎥ + ⎢ production ⎥ = 0 ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ to ⎥ ⎢ from ⎥ ⎢ within ⎥ ⎢⎣ system ⎥⎦ ⎢ system ⎥ ⎢ system ⎥ ⎣ ⎦ ⎣ ⎦

(9.164)

A steady-state balance (neglecting cell death) is depicted by: ⎛ F⎞ ⎛ F⎞ ⎜⎝ ⎟⎠ X0 = ⎜⎝ ⎟⎠ X − r V V

(9.165)

If X is the cell concentration in the bioreactor vessel and exit stream), X0 is the cell concentration of the feed, F is the ﬂow rate of the feed and exit stream, V is the bioreactor vessel volume, r is the rate of cell formation (cells/ unit time/unit volume) equal to dX/dq, then D (= F/V), known as the dilution rate, is the number of bioreactor vessel volumes passing through the system per unit time (the inverse of the mean residence or holding time). The dilution rate is universally employed throughout the biotechnology industry. With a single CSTF, the feed stream is normally sterile medium. Therefore X0 is zero, and X(D – μ) = 0. Since either X = 0 or (D – μ) = 0, a cell population > 0 can be maintained only if the speciﬁc growth rate μ is balanced by the dilution rate D (i.e., a nonzero cell population can only be maintained when D = μ). Once the culture has adjusted its speciﬁc growth rate to the dilution rate, the expression DX0 = X(D – μ) can be satisﬁed by any X value greater than zero. This condition is only reliable when: (1) speciﬁc growth rate is independent of cell population, and (2) growth is exponential and not affected by decreasing concentration of the growth-limiting key cellular component. For continuous culture, a static maintenance bioreactor or chemostat can be used to maintain cell densities at as high as 100 times the density of conventional systems. Stocked with cells, porous tubes constantly input nutrients and remove waste products, gas permeable tubes continuously add oxygen and remove carbon dioxide, while computer controls monitor these functions and maintain a hom*oeostatic environment. During continuous culture the cell population is maintained both biochemically and physiologically constant over an indeﬁnite period. Major advantages of

PH2270_book.fm Page 275 Tuesday, May 16, 2006 9:37 AM

Bioreactors

275

chemostatic continuous culture include the ability to: (1) manipulate the speciﬁc growth rate over a full range without the need to vary medium composition or environmental conditions, (2) strategically evaluate the effects of the physico-chemical environment while maintaining a constant growth rate, (3) analyze both the culture’s steady state and deﬁned transitory states, (4) set unique culture conditions by changing the nature of the growth-limiting substrate(s), and (5) maintain culture growth over long periods. Chemostat theory predicts signiﬁcantly higher productivity than batch cultures. Productivity is maximized by careful manipulation of environmental parameters, reduced downtime for cleaning, and elimination of the time-consuming, repetitive operations (e.g., sterilization, cleaning, etc.) associated with batch cultures. Continuous culture is an important bioprocess optimization tool that relates not only to productivity, but to concentration, substrate conversion efﬁciency, duration of biosynthetic activity, and product stability. In a batch culture, if one of the key cellular components is at growth-limiting concentration, the cell balance at steady state can be written using the Monod expression for speciﬁc growth rate μ, and if the yield factor Y is deﬁned as: Y=

mass of cells formed nutrient consumed

(9.166)

the steady state balance on the limiting component is expressed by the Monod chemostat model. Where X0 = 0 (sterile feed): D(C0 − Ci ) −

μ m Ci X =0 Y (Ci + K i )

(9.167)

μ m C0 C0 + K i

(9.168)

Optimal Conditions The cell production rate per bioreactor volume is DX (FX/V), then the maximum cell production rate is: d ( DX ) =0 dD and the rate for optimal cell production is:

(9.169)

PH2270_book.fm Page 276 Tuesday, May 16, 2006 9:37 AM

276

Producing Biomolecular Substances

1/ 2

⎤ ⎡ Ki Dopt = μ m − ⎢ ⎥ C K ( ) + i ⎦ ⎣ 0

(9.170)

•

As a bioprocessor, the bioreactor must be sized to provide the required production capacity; As mass-transfer equipment, the bioreactor must be designed to ensure that media and cells are well dispersed and, for aerobic cultures, that adequate dissolved oxygen concentration is maintained and available for individual cells; As a control device, the bioreactor must provide for temperature control for the rapid dispersion of control chemicals (i.e., for pH and foaming) and ensure representative sampling of important parameters (i.e., cell population, pH, dissolved oxygen concentration, etc.); and As a heat transfer device, the bioreactor must ensure that constant temperature is maintained during the growth cycle by cooling, and heating to sterilize the system in situ.

Maximizing Productivity Continuous culture is sometimes used for hybridomas, although maximizing productivity with this culture mode is often quite difﬁcult. Contamination and selection of spontaneous mutations must be constantly monitored. Also, small environmental deviations attributable to bioreactor problems can now and then cause signiﬁcant cell population reduction. Continuous culture parameters for hybridomas are based on the kinetics of single cells in exponential growth. Although growth should proceed, in theory, to inﬁnite density, in practice, either a growth-rate-limiting substance is depleted and/or toxic metabolic waste product(s) are formed, resulting in limited population density. A good continuous culture maximizes viable cells and minimizes dead and dying cells. Both biochemical and physical techniques are typically used to detect and quantify living cells in the population (see Fermenters). The dilution rate for a continuous culture is equal to the ﬂow rate of fresh medium into the bioreactor divided by bioreactor volume. During continuous culture, the rate of change in hybridoma cell concentration equals the rate of growth minus the cells removed. At steady state the biomass and growth-rate-limiting concentrations are constant and the speciﬁc growth rate can be controlled by substrate concentration. At a critical dilution rate the loss of cell population (washout) will occur.

PH2270_book.fm Page 277 Tuesday, May 16, 2006 9:37 AM

Bioreactors

277

Practically, the maximum speciﬁc growth rate can be determined by increasing the dilution rate until washout. Substrate consumption during continuous culture follows ﬁrst-order Monod kinetics. In continuous culture, the biomass concentration and growth-limiting substance concentration can be determined from the supply vessel concentration of the growth-limiting substrate and the biomass yield of the substrate. Measurement of Stoichiometric Coefficients A good starting point for a discussion of cell growth is to examine what cells are made of — their chemical composition. Although there are many different prokaryotic microorganism and eukaryotic-cell species, a very large part of their biomass is composed of just a few elements — carbon, hydrogen, oxygen, and nitrogen, which are, interestingly enough, among the most common elements on earth. Cell Composition In this section we introduce the growth reaction that characterizes the material balance associated with cell growth, deﬁne the chemical formula that represents approximately 95% of the dry biomass, and also deﬁne yield that enables us to derive useful engineering information from literature that reports the typical cell yield. Determining material balance in a bioreactor typically requires the respiratory quotient value and one other measurement. Cells primarily contain water; cell mass is 70% water, so, it is conventional to express the biomass cell composition on a dry-weight basis. Growth Reaction In the above example, we assume that all the carbon found in the substrate is incorporated into the cell mass. This does not actually happen, since the cell needs to oxidize (respire) some of the carbon to produce energy for biosynthesis and for maintenance of the cell’s metabolic machinery. In addition, cells may produce extracellular products that accumulate in the liquor. Hence we can represent growth as: Cell + {C-source, N-source, others} + O 2 → More Cells + Extracellular + CO 2 + H 2O

(9.171)

The medium is the “food” for the cell. It serves as a source for all the elements needed by the cell for growth or biosynthesis, and for protein (end product) formation. The compound’s carbon dioxide and water on the product side of the reaction, result from the oxidation of glucose in the medium. Since the cellular material contains C, N, P, S, K, Na, Ca, etc., the medium must be formulated to supply these elements in an appropriate form. If we neglect the others and assign a stoichiometric coefﬁcient for each of the species in the above equation, on the basis of one mole of glucose (C-source) consumed, we can rewrite the above as:

PH2270_book.fm Page 278 Tuesday, May 16, 2006 9:37 AM

278

Producing Biomolecular Substances

C6 H12O 6 + aNH 3 + bO 2 → αCH1.8O 0.5 N 0.2 + βCH xO y N z + γCO 2 + δH 2O (9.172) where ammonia represents the nitrogen source. We can refer to this reaction as the growth reaction. Please note that whatever nitrogen is supplied in the medium is expressed here as equivalent nitrogen in the form of ammonia. Cells require nitrogen in both the organic and inorganic form, and we typically supply the inorganic nitrogen as salts of ammonium (e.g., ammonium phosphate) while the organic nitrogen is usually supplied as amino acids or proteinacious extracts, which are rich in nitrogen. In most production processes employing recombinant cells, glucose is used as the carbon source. However, in producing low-value products, less expensive carbon sources such as molasses or corn meal are used. Compare this against pure glucose at $1.00/lb! The growth reaction shown is useful in interpreting the laboratory data reported in the literature; since early cell growth studies were reported by microbiologists, we typically employ their terms to describe growth stoichiometry, and can then relate the above reaction equation to commonly reported cell properties. Cell Yield and Stoichiometric Coefficients Consider the experimental Pseudomonas lindneri growth data originally reported by Bauschop and Elsden. Their experiment consisted of inoculating ﬁve test tubes containing growth medium with the bacterium Pseudomonas lindneri. Each of the test tubes contained different carbon source concentrations — glucose at levels from about 4 mM–36 mM. The cultures were anaerobic ally incubated at 30°C until growth ceased (about two days). The resulting cells were ﬁltered, dried, and weighed. The biomass obtained was plotted against the starting glucose concentration. An important observation illustrated by their data, is the straight-line relationship between the carbon source reactant concentration and the end product cell concentration, where the slope of the line represents the amount of cells obtained per unit amount of glucose consumed.

Slope =

270 μg ml−1 270 μg ml−1 ⇒ ⇒ 8.2 μg μmol−1 33 mM 33 μmol mol−1

(9.173)

Now, if we convert the above to a mass basis: Slope = 0.046 g of cells per g of glucose conssumed

(9.174)

The above value is then termed cell yield, growth yield, or yield coefﬁcient. And, if you examine the growth reaction given in the previous section, the slope (in mass units) calculated above can be equated as:

PH2270_book.fm Page 279 Tuesday, May 16, 2006 9:37 AM

Bioreactors

279

⎛ MW of Cell ⎞ Cell Yield = α ⎜ ⎝ MW of Substrate ⎟⎠

(9.175)

The numerator contains the amount of cells created; the denominator contains the amount of substrate consumed. Thus, the measurements reported by Bauschop and Elsden enable us to calculate the stoichiometric coefﬁcient α or, ⎛ 180 ⎞ ⇒ 0.33 α = (0.046 ) × ⎜ ⎝ 24.6 ⎟⎠

(9.176)

Considering the data set, cell yield depends on growth conditions, and under anaerobic conditions, the slope (yield) is 58.2 g (mol substrate)–1, or 0.32 g cell (g substrate)–1. Similarly, under anaerobic conditions, yield is 22 g (mol substrate)–1, or 0.21 g cell (g substrate)–1. Invariably, the yield under anaerobic conditions will be smaller than under aerobic conditions because the cell derives signiﬁcantly more metabolic energy under aerobic conditions. It is important to note that not all cells can grow both aerobically and anaerobically. From the practical viewpoint, an aerobic organism is preferred because the amount of produced protein is proportional to the amount of cells, and higher levels of biosynthesis are possible with aerobic cultures than with anaerobic ones. Cell Yield Mathematically, cell yield can be deﬁned as:

YX /S =

Amount of Cell Produced ΔX = Amount of Substrrate Consumed ΔS

(9.177)

where ΔX represents change in cell concentration and ΔS represents change in substrate concentration. The subscript X/S indicates the basis of yield–cell on the basis of substrate. This notation comes in handy when it is necessary to calculate yield based on more than one substrate. Examining the above and comparing it with the growth reaction, one notes that the yield deﬁned here corresponds to a massbased stoichiometric coefﬁcient. Taking the limit of Eq. 9.177 as dS approaches zero:

YX /S =

dX dS

(9.178)

The absolute sign eliminates the negative value of the derivative. Note that dS is always negative, because substrate is consumed. Yield is always a positive value. The above deﬁnition of yield can be applied to product, P on the basis of substrate consumed. Thus,

PH2270_book.fm Page 280 Tuesday, May 16, 2006 9:37 AM

280

Producing Biomolecular Substances

YP /S =

dP dS

(9.179)

Similarly, product yield based upon cells can be expressed as YP / X =

dP dX

(9.180)

In general, yield of the species i, based on species j, can be calculated from Yi / j =

di dj

(9.181)

From the above it is clear that we can combine two different yields that have a common species as Yi / j =

Yi / k Yk / j

(9.182)

Example: Batley, in 1979, reported the aerobic growth of yeast on ethanol as:

C2 H 5OH + 0.153 NH 3 + 1.851 O 2 → 1.03 CH1.704 O 0.408 N 0.149 + 0.970 CO 2 + 2.346 H 2O You may calculate YX/E, YX/O2, YX/NH3 on a mass basis, and the MW of a cell = 12 +1.704 + (14) (0.149) + (16) (0.408) = 22.32. The MW of ethanol = (2) (12) + 5 + 16 + 1 = 46, thus:

(1.03) × (MW of Cell) (1.03) × (22.32 ) = ⇒ 0.5 (1) × (MW of Ethanol) ( 46 )

(9.183)

(1.03) × (MW of Cell) (1.03) × (22.32 ) = ⇒ 0.388 (1.851) × (MW of Oxygen) (1.851) × ( 32 )

(9.184)

YX / E =

Y X /O 2 =

YX / NHS =

(1.03) × (22.32 ) (1.03) × (MW of Cell) = ⇒ 8.839 (9.185) monia) (0.153) × (17 ) (0.153) × (MW of Amm

A harvest of yeast on ethanol of about 0.5 is consistent with the observation that about half of the substrate is aerobically converted to cell mass. Thus, if the yield

PH2270_book.fm Page 281 Tuesday, May 16, 2006 9:37 AM

Bioreactors

281

on a carbon source is signiﬁcantly less than 0.5, it is likely that medium formulation is inadequate to support good growth. Example: Yeast grown on glucose can be described by:

C6 H12O 6 + 0.48 NH 3 + 3 OH 2 → 0.48 C6 H10O 3N + 3.12 CO 2 + 4.32 H 2O To calculate the following parameters for a design requiring 50g/L of yeast in a batch reactor of 100,000 liters: For the nutrient concentration of the media for glucose and ammonium sulfate: (1) calculate YX/S and YX/O2; (2) calculate total oxygen required; (3) determine oxygen uptake rate (g O2L–1h–1) when cell concentration increases at a rate of 0.7 g L–1h–1. The total cell mass to be produced is = (105 L) (50 g L–1) = 5,000 kg:

YZ /S =

(0.48 ) × (MW of Cell) (0.48 ) × (144 ) = ⇒ (1) × (MW of Glucose) (180 )

0.384 g cell (g substratte)

Y X /O 2 =

(0.48 ) × (MW of Cell) (0.48 ) × (144 ) ⇒ = (3) × (MW of O 2 ) ( 3) × ( 32 )

0.72 g cell (g O 2 ) Glucose needed =

Ammonia needed =

(9.186)

−1

(9.187)

−1

(Cell Mass) (5, 000) = ⇒ 13, 020 kg (YX / S ) (0.384)

(Cell Mass) (5000 ) = ⇒ 590 kg ⇒ 130 g L−1 (YX /S ) (8.4771)

(9.188)

(9.189a)

1 × ( MW of ( NH 4 )2 SO 4 ) × ( Ammonia needed) (9.189b) ( NH 4 )2 SO 4 needed = 2 MW NH 3 =

1 132 × × 590 ⇒ 2, 292 kg ⇒ 22.9 g L−1 2 17

Total Oxygen Required =

(Cell Mass) 5, 000 ⇒ = 6, 944 kg (YZ / O2 ) 0.72

(9.189c)

(9.189d)

PH2270_book.fm Page 282 Tuesday, May 16, 2006 9:37 AM

282

Producing Biomolecular Substances

Oxygen Consumption Rate =

Cell Mass Generation Rate YX / O2

(9.189e)

0.7 = ⇒ 0.972 g L−1 h 0.72 Stoichiometric Coefficient Measurements For the growth reaction given in Eq. 9.189e, the ratio μ/b is called the respiratory quotient, often abbreviated as RQ, and is easily measured in large scale bioreactors. In Eq. 9.189e, if the nature of extracellular product is known (i.e., x, y, z), then it is possible to calculate w, x, y, and z from experimental measurement of RQ and one other measurement. If no signiﬁcant amount of extracellular product is formed, as in simply growth processes, then only RQ or one other measurement is needed to compute stoichiometric coefﬁcients. The example given below illustrates this idea: for the reaction equation representing E. coli growth, RQ was measured as 0.85. Calculate w, x, y, and z. C6 H12O 6 + aNH 3 + bO 2 → αCH1.8O 0.5 N 0.2 + γCO 2 + δH 2O Solution: Carry out elemental balances and then solve them; here, we can write four elemental balances: C, H, O and N. We have six unknowns: w, x, y, z, a, and b. One additional relationship is obtained from the given RQ value, to enable the solution:

C balance:

6=α+γ

H balance:

12 + 3a = 1.8α + 2δ

N balance:

a = 0.2α

O balance:

6 + 2b = 0.5α + 2 γ + δ

RQ =

γ = 0.85 b

Rearrange the above equations:

⎡ 1 ⎢1.8 ⎢ ⎢ 0.2 ⎢ ⎢ 0 ⎢⎣ 0.5 Solution yields:

1 0 0 1 2

0 2 0 0 1

0 −3 −1 0 0

0 ⎤ 0 ⎥⎥ 0 ⎥ ⎥ −0.85 ⎥ −2 ⎥⎦

⎡α ⎤ ⎡ 6 ⎤ ⎢ γ ⎥ ⎢12 ⎥ ⎢ ⎥ ⎢ ⎥ ⎢δ ⎥ = ⎢ 0 ⎥ ⎢ ⎥ ⎢ ⎥ ⎢a⎥ ⎢ 0 ⎥ ⎢⎣ b ⎥⎦ ⎣⎢ 6 ⎥⎦

(9.190)

PH2270_book.fm Page 283 Tuesday, May 16, 2006 9:37 AM

Bioreactors

283

⎡α ⎤ ⎡ 6 ⎤ ⎢ γ ⎥ ⎢12 ⎥ ⎢ ⎥ ⎢ ⎥ ⎢δ ⎥ = ⎢ 0 ⎥ ⎢ ⎥ ⎢ ⎥ ⎢a⎥ ⎢ 0 ⎥ ⎢⎣ b ⎥⎦ ⎢⎣ 6 ⎥⎦

⎡ 1 ⎢1.8 ⎢ ⎢ 0.2 ⎢ ⎢ 0 ⎢⎣ 0.5

1 0 0 1 2

0 2 0 0 1

0 −3 −1 0 0

0 ⎤ 0 ⎥⎥ 0 ⎥ ⎥ −0.85 ⎥ −2 ⎥⎦

(9.191)

CELL-GROWTH THERMODYNAMICS Heat Release Due to Growth Cell growth consists of a complex network of metabolic reactions. Coupled catabolic and anabolic reactions take place so that energy released in the former is efﬁciently used to drive the latter. However, some energy is always lost as heat. The purpose of this chapter is to quantify the heat release due to growth. In largescale processes it is necessary to remove this heat so that the culture is maintained at physiological temperature. In small reactors metabolic heat is removed quite easily, while in very large bioreactors (>10,000 liters) in which rapidly growing cells are cultivated, it is necessary to design an adequate heat transfer system for heat removal. Bioreactor temperatures must be maintained within ± 0.5°C to maintain physiologic conditions conducive to optimal growth. Consider the growth reaction when no signiﬁcant amount of extracellular product is formed. Under these conditions Eq. 9.191 simpliﬁes to: C6 H12O 6 + aNH 3 + bO 2 → αCH1.8O 0.5 N 0.2 + γCO 2 + δH 2O

(9.192)

Since nitrogen consumption is usually small compared to the amount of carbon consumed, and because nitrogen does not go through oxidation (while C does!), we can approximate the above as: C6 H12O 6 + bO 2 → αCH1.8O 0.5 N 0.2 + γCO 2 + δH 2O

(9.193)

Consider heat balance of this reaction using one mole of glucose consumed as the basis. Heat released = (α ) × [(MW biomass) × (− ΔH C )] − (− ΔH S ) × (MW substrate)

(9.194)

where (–ΔHC) and (–ΔHS) are heat of combustion per gram of cell and per gram of substrate, respectively. Rearranging: Heat released ⎛ MW biomass ⎞ × (− ΔH C ) − (− ΔH S ) = (α ) × ⎜ ⎝ MW substrate ⎟⎠ MW substrate

(9.195)

PH2270_book.fm Page 284 Tuesday, May 16, 2006 9:37 AM

284

Producing Biomolecular Substances

The left side is the heat released per gram of substrate consumed and the coefﬁcient of the ﬁrst term on the right is growth yield. That is: YΔ /S = (YX /S ) × (− ΔH C ) − (− ΔH S )

(9.196)

where Y/S is “heat yield” on the basis of substrate consumed. Dividing the above by YX/S gives: YX /S = (− ΔH P ) −

(− ΔH P ) (YX /S )

(9.197)

Both Eqs. 9.196 and 9.197 are useful in determining heat release due to growth, YΔ/X and substrate consumption, YΔ/S. Extracellular Product Heat Release When signiﬁcant amount of product is present, Eq. 9.194 will be modiﬁed to: Heat released = (α ) × (MW biomass) × (− ΔH C ) + (β) × (MW product ) × (− ΔH P ) − (− ΔH S ) × (MW substrate)

(9.198)

where (–ΔHp) is heat of combustion per gram of extra cellular product(s). Dividing the above by MW of substrate yields: YΔ /S = (YX /S ) × (− ΔH C ) + (YP /S ) × (− ΔH P ) − (− ΔH S )

(9.199)

Dividing yields:

YΔ / X = (− ΔHC ) + (YP / X ) × (− ΔH P ) −

(− ΔH S ) (YX / S )

(9.200)

Example: In citric acid cell growth, 2,000 kg glucose is fermented over a 50 h period resulting in 1,300 kg citric acid and 400 kg biomass. The process required 800 kg O2. Calculate the total heat released. Heat of combustion of citric acid (C6H8O7) is –1,960 kJ gmol–1. This is a case of extracellular product is formed. Use Eq. 9.199 or 9.200 to ﬁrst obtain the heat yield:

PH2270_book.fm Page 285 Tuesday, May 16, 2006 9:37 AM

Bioreactors

285

YX / S = (4, 000) × (2, 000)−1 = 0.2 (− ΔH P ) = (1, 960) × (192)−1 = 10.21 kJ g−1 YP / S = (1, 300) × (2, 000)−1 = 0.65 (− ΔH S ) = 15.6 kJ g−1 for glucose

(9.201)

(− ΔHC ) = 22 kJ g−1 for biomass YΔ / S = (YX / S )(− ΔHC ) + (YP / S ) × (− ΔH P ) − ( ΔH S ) YΔ / S = (0.2) × (22) + (0.65) × (10.21) − (15.6) = −4.564 kJ g−1 The negative value indicates that heat is rejected or being generated due to growth. The interpretation here is similar to that for a chemical reaction. Recall heat of reaction is negative for an exothermic reaction. Ignoring the sign as we label it as heat generated:

Heat Generated = (YΔ /S ) × (Substrate Consumed) = ( 4.564 ) × (2, 000, 000 )

(9.202)

= 9.13 × 10 6 kJ Alternatively: we can calculate heat generated from the oxygen-uptake data:

Heat Generated = (YΔ /O2 ) × (Oxygen Consumed) = (166.21 kJ per g O 2 ) × (800 × 10 3 g O 2 )

(9.203)

=12.97 × 10 6 kJ These methods differ by about 30%; the second method is obviously more accurate! In the ﬁrst method, we assume that all extracellular products are accounted for in citric acid; but the cell produces other products not included in the energy balance.

GROWTH KINETICS

AND

PRODUCT FORMATION

Growth Kinetics In this chapter we limit the bioreactor analysis to batch systems. If a viable inoculum is introduced into a medium that contains a carbon source, suitable nitrogen source, other nutrients necessary for growth, and the physiological temperature and pH are maintained, it will grow. The rate of biomass synthesis is proportional to biomass present. That is:

PH2270_book.fm Page 286 Tuesday, May 16, 2006 9:37 AM

286

Producing Biomolecular Substances

rx = μX

(9.204)

where rX is the amount of cells synthesized in g L–1 h–1, X is cell concentration in g L–1. The parameter μ is called specific growth rate, analogous to the speciﬁc rate constant in chemical reaction rate expressions. Recall the treatment of chemical reactions, summarized below for easy reference: Reaction:

A→ B

Rate Expression: − rA = kC A

(9.205)

In the above CA is concentration of A (mol A L–1), –rA is reaction rate (mol A L–1 h–1) and k is rate constant (h–1). The negative sign in front of –rA is to comply with the deﬁnition of rA, which is the rate of generation of A. In Eq. 9.205, the negative sign is not necessary as X increases with time. Consider cell balance in a batch bioreactor: Cells in – Cells out + Generation of Cells = Accumulation of Cells in Bioreactor 0 − 0 + (rx )* (V ) =

d (Vx ) dt

(9.206)

(9.207)

Substituting for rX from Eq. 9.207 and noting that volume of reactor is constant gives: dX = μX dt

(9.208)

The above can be expressed as μ=

1 dX ⎛ ΔX ⎞ ⎛ 1 ⎞ =⎜ ⎟ ×⎜ ⎟ X dt ⎝ X ⎠ ⎝ Δt ⎠

(9.209)

The term, DX/X, represents fractional increase in cell amount and dt is the time over which the fractional increase was accomplished. That is, μ can be interpreted as a fraction of biomass formed per unit time. For example if μ is 0.3 h–1, every hour the biomass will approximately increase by 30%. We use the term “approximately” because we are using ﬁnite quantities to describe the rule which applied at inﬁnitesimal scale. Treating μ as a constant for now, Eq. 9.209 can be integrated to give: X = X0 exp(μt )

(9.210)

PH2270_book.fm Page 287 Tuesday, May 16, 2006 9:37 AM

Bioreactors

287

where X0 is the initial (inoculum) cell concentration. The time, t, refers to the time since the inoculum emerged from lag phase. Eq. 9.210 can be rearranged setting the conditions for doubling of biomass. That is X/X0= 2 and t = the doubling times td: ⎛ ln(2 ) ⎞ ⎛ 0.693 ⎞ = td = ⎜ ⎝ μ ⎟⎠ ⎜⎝ μ ⎟⎠

(9.211)

The doubling time and speciﬁc growth rate have been reported by many researchers. Given below is a sample of typical values: Organism

Growth Rate, μ [h–1]

Doubling time, td [h]

Hybridoma cells Insect cells

0.05 0.06

13.9 11.6

On What Does the Specific Growth Rate (μ) Depend? The speciﬁc growth rate μ depends on a number of factors such as growth medium composition, temperature, pH, and others. Experimental studies have shown that one cannot increase growth rate beyond a certain maximum value, μm due to inherent metabolic reaction rate limitations. In general, when substrate S is limiting growth, Monod (1949) reported that growth rate variations can be expressed as: μ=

μ mS KS + S

(9.212)

where KS is called Monod constant or simply the substrate saturation constant. The signiﬁcance of KS is, when substrate concentration is numerically equal to KS, growth rate is exactly half of maximum growth rate. In Monod Kinetics, the speciﬁc growth rate reaches a maximum value of 0.5 h–1; The value of KS here is 0.5 g L–1. Note that when S = 0.5 g L–1, μ is half of its maximum; this form can describe dependence of μ on more than one limiting nutrient. In many practical applications O2 availability for respiration often limits growth. When substrate S and dissolved oxygen concentration CDO2 are both limiting growth, speciﬁc growth rate can be mathematically described as: μ=

μm S C DO × K S + S K DO + C DO

(9.213)

The behavior of maximum growth rate when two substrates are limiting where the parameters KS and KDO are cell speciﬁc. KS is typically in the order of 10 mg/L for glucose and KDO is less than 1 mg/L for oxygen in the case of bacteria and yeast. KDO has been reported to be higher for mammalian and insect cells. When two substrates are limiting, the speciﬁc growth rate reaches a maximum value of 0.5 h–1;

PH2270_book.fm Page 288 Tuesday, May 16, 2006 9:37 AM

288

Producing Biomolecular Substances

and the value of KS here is 0.5 g L–1; Value of KDO is 0.1 mgL–1; and note that when CDO = 0.1 mg L–1, μ is half of its maximum at values of S >> KS. Let us now consider growth under conditions of only substrate limitations in a batch bioreactor. Incorporating the substrate limited condition and bioreactor material balance equation, Eq. 9.213, can be modiﬁed and we may write: dX μ S = m X dt K S + S

(9.214)

To integrate the above, one of the variables, S, must be replaced in terms of X. The yield relationship, Eq. 9.214, can be integrated as:

∫

dX = −YX /S

∫

(9.215)

This simpliﬁes to: ⎛ X − X0 ⎞ S = S0 − ⎜ ⎝ YX /S ⎟⎠

(9.216)

where subscript 0 refers to initial concentration. Substituting for S from Eq. 9.215 in Eq. 9.216 and integrating: ⎛ K SYX /S + S0YX /S + X0 ⎞ ⎛ X ⎞ ⎛ K SYX /S ⎞ ⎛ YX /S S0 + X0 + X ⎞ ⎜⎝ ⎟⎠ ln ⎜⎝ X ⎟⎠ − ⎜⎝ Y S + X ⎟⎠ ln ⎜⎝ ⎟⎠ = μ mt (9.217) YX /S S0 + X0 YX /S S0 0 X /S 0 0 For analyzing batch systems, use the above to calculate cell concentration and then calculate substrate concentration using Eq. 9.217. Metabolic Quotient and Rate Expression We have already discussed rate expressions for cell growth in Eq. 9.212. Let us now examine rate expressions for other medium components in the growth reaction in Eq. 9.212. Consider the growth reaction on the basis of one g of substrate consumed. It can be written as, 1 g S + YO2 / S g of O 2 + YNH3 / S g of NH 3 = YX / S g of Biomass + YCO2 / S CO 2 + others (9.218) The stoichiometric coefﬁcients in growth reaction become yield coefﬁcients on the basis of substrate. (See Equation 9.218.) The general rate expression is then:

PH2270_book.fm Page 289 Tuesday, May 16, 2006 9:37 AM

Bioreactors

289

rS r r r r = O2 = NHS = X = CO2 −1 −YO2 /S −YNHS /S YX /S YCO2 /SS

(9.219)

where ri is expressed in g of i L–1 h–1. Since rx is the most fundamental of the various rates, it is conventional to write the stoichiometric coefﬁcient in terms of it. That is: rS =

rX −YX /S

(9.220)

Y r rO2 = rX O2 /S ⇒ X YX /S Y X /O 2 Following the examples above, the rate expression for species i can be written as: ri =

rX Y X /i

(9.221)

Metabolic quotients are rate expressions on the basis of unit mass of biomass. That is: qi =

ri μ ⎛ r ⎞ ⎛ 1⎞ ⇒⎜ X ⎟ ×⎜ ⎟ ⇒ ⎝ Y X /i ⎠ ⎝ X ⎠ X Y X /i

(9.222)

The metabolic quotient for oxygen is of special interest. This single property determines the upper limit of cell concentration that can be achieved in many bacterial cell growth systems. We will see further analysis in the next chapter. Typical values of metabolic coefﬁcients are shown in Table 9.11. Example: if the speciﬁc growth rate of a bacterium is 0.35h–1 and cell yield is 0.6, calculate glucose consumption rate.

qG =

μ 0.35 = = 0.48 gG (g cell)−1 h −1 Y X /G Y X /G

TABLE 9.11 Typical Values of Metabolic Coefficients Organism

Glucose g Glucose (g cell)–1 h–1

qO2 g Glucose (g cell)–1 h–1

Yeast Hybridoma

0.5 0.2

0.2 0.02

PH2270_book.fm Page 290 Tuesday, May 16, 2006 9:37 AM

290

Producing Biomolecular Substances

Factors Affecting Growth Rate Nutrients, pH, temperature, DO2 concentration, and other environmental conditions all affect the growth rate. Temperature and pH dependence are illustrated in Figures 9.7a and 9.7b. In Figure 9.7a the maximum growth rate is observed at 39°C for e. coli. The product formation kinetics (e.g., for insulin, product yield YP/S and cell yield YX/S) are also affected by temperature. Typically, cell yield decreases with temperature; similar deﬁning relationships for end products have not been reported. It’s important to note here that optimal growth temperature may be different from optimal temperature for end product formation. In the effect of temperature on the growth rate of cells, the maximum growth rate is at 39°C, and the plot is a function of inverse absolute temperature. The decline from 39°C to 21°C and then to 13°C suggests that the growth rate constant behaves similar to a chemical reaction rate constant. In the effect of pH on growth rate, the typical pH range over which reasonable growth can be expected is about 1–2 units. With adaptation, broader ranges can be achieved. Optimal pH values for cell growth range from 6.2–7.2 for eukaryotic insect and mammalian cells. As previously noted, the optimal pH for end product formation may be different from that for growth. Many cells produce a different mix of end products when pH is altered. However, in the case of cells expressing recombinant proteins, pH usually affects the kinetics of recombinant protein generation rather than the end product mixture. Hybridomas produce antibodies at a higher rate at pH 6.2 than at pH 7.2. Because of the difference in conditions for growth and product formation, optimization is often necessary. Oxygen is an important substrate for aerobic organisms. Since metabolic energy production by cells is directly related to the oxygenation rate, oxygen concentration is very strongly coupled to the rate of cell growth. As illustrated in Figure 9.4, the growth rate depends on dissolved oxygen concentration. The critical dissolved oxygen concentration refers to value of DO below which growth rate is lower than the maximum value, and the growth rate sharply rises to its maximum value with dissolved oxygen concentration. The concentration at which maximum growth rate is attained is often referred to as critical oxygen concentration CCRIT O2 . This value is typically less than 0.5mg L–1 for bacteria and yeast, and about 1–2 mg L–1 for animal and insect cells. Note that these values are signiﬁcantly lower than the air saturation value of 6.7 mg L–1 at 37°C. Product Formation Kinetics Product formation kinetics fall into one of the following three types: 1. Growth-associated product formation. 2. Nongrowth-associated product formation. 3. Mixed-mode product formation. Typical time proﬁles of these three cases are illustrated in Figures 9.9a, 9.9b, and 9.9c, respectively. In Type I, product is formed simultaneously with growth of cells.

PH2270_book.fm Page 291 Tuesday, May 16, 2006 9:37 AM

Bioreactors

291

That is, product concentration increases with cell concentration. The metabolic quotient for P can be expressed as a function of μ: rP = qP X ⇒ αμX qP = YP / X μ

(9.223)

From the above equation, it is clear the proportionality constant a is the yield coefﬁcient YP/X. Type II product formation is unrelated to growth rate, but is a function of cell concentration. This is expressed as: rP = qP X ⇒ βX

(9.224)

Antibody formation by a hybridoma and some antibiotic cell growth exhibits this type of behavior. In the third category, product formation is a combination of growth rate and cell concentration. That is: rP = qP X ⇒ (αμ + β) X

(9.225)

Many biochemical processes fall into this category. Note that if b is zero and a is YP/X, this case reduces to Type I. If a = 0, it reduces to a nongrowth-associated case. Therefore let us consider this more general case for further analysis. In a batch reactor, product accumulation can be obtained by carrying out mass balance on the product: Rate of Product Formation = Accumulation of Product (rP ) × (V ) =

d (VP ) dt

(9.226)

(9.227)

For a constant V: dP = rP = (αμ + β) X dt

(9.228)

If we consider the exponential phase only, X = X0 Exp (μmt). That is, substituting in the above gives: dP = (αμ + β) X 0 Exp(μ mt ) dt

(9.229)

PH2270_book.fm Page 292 Tuesday, May 16, 2006 9:37 AM

292

Producing Biomolecular Substances

Integrating from t = 0, P = P0 we get: P − P0 =

(αμ + β) X 0 (Exp(μ mt ) − 1) μm

(9.230)

The above expression can be used to calculate the amount of product concentration at the end of a growth cycle.

OXYGEN TRANSFER

BIOREACTORS

Overview Oxygen is needed by all cells for respiration. Oxygen is used by cells in suspension and must be available as DO2. Since oxygen solubility is quite small, about 6–7 mg/L under normal cultivation conditions, the metabolic oxygen requirement is supplied as needed by continuous aeration of the culture medium. Actively respiring yeast requires about 0.15 g O2 (g cell)–1 h. At a cell concentration of 10 g L–1, the medium saturated with air can support less than 30 seconds worth of metabolic oxygen. Thus, a continuous supply of oxygen must be maintained in any viable aerobic manufacturing process. In this section we get a quantitative appreciation for metabolic oxygen demand, followed by methods used in calculating rates at which oxygen is transferred from sparged air, and methods useful in characterizing the oxygen mass-transfer coefﬁcient. Finally, we evaluate bioreactor operation and design based upon oxygen transfer capability. Metabolic Oxygen Demand Metabolic oxygen demand of an organism depends on the biochemical nature of the cell and cultivation conditions. Oxygen need is usually satisﬁed in most cells if the dissolved oxygen concentration in the medium is kept at about 1 mg/L. If the oxygen level is allowed to fall far below this value, oxygen consumption rate decreases with concomitant decrease in biochemical energy production, and as a result cell growth rate also decreases. The value of oxygen concentration above which growth rate is at the maximum was described as the critical oxygen concentration CCRIT O2 . The oxygen requirement for growth is expressed best in the yield coefﬁcient parameter YX/O2. It represents the amount of oxygen required to grow one gram of cells. Volumetric Oxygen Mass-Transfer Coefficient In a typical aeration system, oxygen from the air bubble is transferred through the gas-liquid interface followed by liquid phase diffusion/bulk transport to the cells. Although this is a multistep serial transport, in a well-dispersed system the major resistance to oxygen transfer is in the liquid ﬁlm surrounding the gas bubble. Consider the oxygen concentration proﬁles in the region near the interface and the oxygen concentration proﬁle at the air bubble medium interface, where the transport

PH2270_book.fm Page 293 Tuesday, May 16, 2006 9:37 AM

Bioreactors

293

of oxygen through the gas and liquid ﬁlms are equal at steady state. They can be expressed by: NO2G = kG A(C DOG − C DOGi )

(9.230a)

NO2 L = k L A(C DOLi − C DOL )

(9.230b)

NO2G = NO2 L

(9.230c)

where subscript G and L refer to gas and liquid phases, respectively. The terms, NO2G and NO2L are oxygen transfer expressed in g O2h–1, A is interfacial area and CDO is oxygen concentration expressed in g O2 per unit volume. At the interface, equilibrium between the liquid and gas phase oxygen is reached. That is: C DOGi = mC DOLi

(9.231)

Because of low oxygen solubility and the fact that kG is much higher than kL: C DOG ≈ C DOGi

(9.232)

Consequently, Eq. 9.230a can be written as: ⎞ ⎛C NO2 = k L A ⎜ DOG − C DOL ⎟ ⎠ ⎝ m

(9.233)

The subscript L in NO2 has been dropped to note that the above represents overall transfer of oxygen. The driving force in the above consists of the difference between bulk oxygen concentrations in the two phases; the ﬁrst term represents the concentration of oxygen in the liquid, which is in equilibrium with the bulk gas-phase oxygen. If air is the gas medium, this term will equal to 7 mg/L at 35°C. When oxygen transfer is applied to an entire volume of a bioreactor, A represents the total interfacial area and kL represents an average mass-transfer coefﬁcient. The concentrations will be bulk gas- and liquid-phase oxygen concentrations. If we divide this equation by the volume of liquid phase V, the resulting term will represent the amount of oxygen transferred per unit volume per unit time — which is in the same units as the rate expressions we saw in the last section. Since the rate is due to a physical phenomena, let us distinguish it by the symbol, RO2. That is: ⎛ A⎞ ⎛ C ⎞ RO2 = k L ⎜ ⎟ ⎜ DOG − C DOL ⎟ ⎝V⎠⎝ m ⎠

(9.234)

PH2270_book.fm Page 294 Tuesday, May 16, 2006 9:37 AM

294

Producing Biomolecular Substances

The term, kL A represents the product of mass-transfer coefﬁcient and interfacial area available for mass transfer. In a bioreactor, air is sparged and the liquid is agitated to break up the bubbles so that interfacial area can be kept high to enhance the rate of oxygen transfer. In such systems the area A is not easily measured or estimated. But the term consisting of the product — mass-transfer coefﬁcient and interfacial area — is more readily measured. Furthermore, it is convenient to use interfacial area per unit volume a rather than total area A because the oxygen transfer rate is expressed per unit volume of bioreactor — similar to rate of cell growth, which is reported on a volumetric basis. Hence, area per unit volume a is combined with the mass-transfer coefﬁcient kL and is given by the term kLa. In Eq. 9.234 the * term CDOG/m can be replaced by oxygen solubility at bioreactor conditions C DOL . * RO2 = k La (C DOL − C DOL )

(9.235)

The above will be our working equation for describing transfer of oxygen from gas phase to growth medium. In order for us to calculate oxygen transfer rate (OTR), we need the mass-transfer coefﬁcient kLa, solubility of oxygen in the mediu* m C DOL , and the dissolved oxygen concentration in the medium CDOL. In the last chapter we had used the notation CDO to describe dissolved oxygen concentration. In the discussion above, there was a need to make a distinction between gas- and liquid- phase concentration. In Eq. 9.235, one notes that both concentrations are expressed on the basis of liquid phase. Hence, from here on we will drop the subscript L. In situations where we need to make a distinction between the two phases, we will reintroduce the subscript L and G. Bioreactor Oxygen Balance Now, let us consider the case of oxygen balance within a bioreactor in which cells are growing and, during that process, consuming oxygen. There is a continuous inﬂow of air at a constant volumetric ﬂow rate. The liquid medium and culture it contains are agitated by a marine impeller. If the metabolic oxygen uptake rate is qO2 and the cell concentration is X, because the function of the bioreactor system is calculated over a sufﬁciently short period, X can be treated as a constant. Consider the oxygen balance over the liquid phase of the bioreactor as: O2 transferred from the Gas Phase – the O2 consumed by the Cells = Oxygen Accumulation: * [ k La (C DO − C DO )] × V − qO2 XV =

d (VC DO ) dt

(9.236)

For constant liquid-phase volume, this can be simpliﬁed to d (C DO ) * = k La (C DO − C DO ) − qO2 X dt

(9.237)

PH2270_book.fm Page 295 Tuesday, May 16, 2006 9:37 AM

Bioreactors

295

(9.238)

One useful application of this is in estimating the maximum cell concentration that a particular bioreactor is capable of supporting in terms of its oxygen supply. Example: A bioreactor has an oxygen mass-transfer capability coefﬁcient of 400 h–1. What is the maximum concentration that a mammalian cell culture grown aerobically in this bioreactor can attain if the respiration rate of the cell culture is 0.35 g O2 (gCell)–1h–1, and the critical oxygen concentration is 0.2 mg/L, and the oxygen saturation of the air is 6.7 mg/L. Solution: From Eq. 9.238, we have:

* k La (C DO − C DO ) qO2

(9.239)

The maximum expected oxygen concentration is equal to

(6.7 – 0.2) = 6.5 mg/L

(9.240)

Therefore, the maximum cell concentration that can be grown at a maximum growth rate is:

Xmax =

* k La (C DO − C DO )max ( 400 h −1 ) × (6.5 mg O 2 L−1 ) ⇒ ⇒ qO2 0.35 g O 2 (g Cell)−1 h −1

(9.241)

−1

7.4 g Cell L

Factors Affecting Mass-Transfer Coefficient The mass-transfer coefﬁcient (KLa) is strongly affected by both the agitation speed and the air-ﬂow rate. In general: K La = k ( Pg / VR )0.4 (VS )0.5 ( N )0.5

(9.242)

where k is a constant, Pg is the power required for aerated bioreactor, VR is the bioreactor volume, VS is air ﬂow rate, and N is agitator speed. Note: the mass-transfer coefﬁcient increases with both agitation speed and airﬂow rate.

PH2270_book.fm Page 296 Tuesday, May 16, 2006 9:37 AM

296

Producing Biomolecular Substances

Example: You are part of a tech service team asked to evaluate if a 10,000-L Bioreactor is adequate to produce 10 kg/day of a protein using a strain cell that expresses the protein as 20% cellular protein. In order to enhance stability, nutrients are manipulated to give a low speciﬁc growth rate of 0.2 h–1. The oxygen demand is 0.15 g O2/g cell – h. Assume that the end-product protein formation is cell-growth related. Data: Assume the lag phase is 4 hours, the typical cleaning time following a batch and preparation for the next batch is 8 hours, and the plant runs three shifts. Cell yield on the substrate is 0.55 g cell/g of substrate. Available support services can supply a maximum inoculum of 6 kg of cells every 24 hours. The maximum KLa for the bioreactor is 500 h–1. Bioreactor accessories are capable of handling cell concentrations of 60 g/L. Assume any other parameters you need to complete the calculation. Assuming that the critical oxygen concentration. is 0.2 mg/L and DO2 at air saturation is 6.4 mg/L: * CRIT Max. Oxygen Transfer Rate = K La (C DO − C DO ) = (500 ) ×

(6.4 − 0.2 ) × 10 −3 g L−1 h −1 Therefore,

Max. Cell Conc. Sustainable =

3.1 Max OTR = = 20.6 g/L 0.15 qO2

(9.243)

Solution A: The lag phase and cleaning/prep time is 12 hr. If a batch is completed within each 24-hr period, production is limited to 12 hr per day. If this is not a limitation, one can optimize production by varying the batch time. Lets ﬁrst evaluate assuming a 12 hr batch time. If a maximum cell concentration of 20.6 g/L is obtained, the amount of end product protein produced is = (0.2)(0.5)(20.6) = 2.06 g/L, since 50% of the cell dry matter is assumed to be protein. Hence in 10,000 liters, we can produce 20.6 kg. Next: determine the inoculum level. The maximum batch growth phase is 12 hr. Substitute in growth equation n and assume enough nutrients are present to support exponential growth during the 12 hr period, and:

X = X0 exp(μ mt ) (20.6 ) = X0 exp((0.2 )(12 )) or

X0 = 1.87 g/L

(9.244)

For 10,000 liters, we need 18.7 kg every 12 hr. Since only 6 kg is available, the maximum protein that can be produced is

(0.2)(0.5)[0.6 Exp.(0.2)(12)]10,000 = 6.61 kg

(9.245)

Solution B: Now if we allow batch times to be longer than 12 hr — meaning that there might not be a harvest every day — and since it is advantageous to use the maximum inoculum concentration, then X0 = 0.6 g/L. This value is obtained by concentrating 6 kg of cells in 10,000 L. Since maximum cell concentration is ﬁxed due to aeration

PH2270_book.fm Page 297 Tuesday, May 16, 2006 9:37 AM

Bioreactors

297

requirements, use the batch growth equation n to ﬁnd a batch growth time of 17.7 hr. Hence 20.6 kg of end product protein will be produced every 29.7 hr, which gives a 24 hr production rate of 16.6 kg of end product protein.

CELL GROWTH

AND

PRODUCTION

Kinetic Growth Understanding the growth kinetics of cells is necessary for the correct design and operation of a bioreactor; in bioproduction, cell kinetics is the consequential interaction of numerous complicated biochemical reactions and transport phenomena involving multiple stages of multicomponent systems. During growth, a heterogeneous mixture of young and old cells is continuously transforming and adapting to the changing environment. As a result, completely accurate mathematical modeling of such system growth kinetics is virtually impossible. In order to derive simpler models of bioreactor operation and performance that can be expressed by mathematical models, some assumptions must be made regarding various cell components and cell population dynamics, as presented in Table 9.1. The simplest model is the unstructured, distributed model, that is based on two assumptions: (1) Cells can be represented by a single component, such as cell mass, cell number; or the concentration of protein, DNA, or RNA. This is particularly true for balanced growth, since the doubling of cell mass for balanced growth is accompanied by a doubling of all other measurable properties of the cell population. (2) Cell mass is distributed uniformly throughout the culture (i.e., the cell suspension is regarded as hom*ogeneous, the heterogeneous nature of cells is ignored, and cell concentration is expressed as wet or dry weight per unit volume). In addition to such assumptions about cells, the medium must be formulated so that just one component limits the reaction rate, and that all the other components are present at sufﬁciently high concentrations so that minor change does not significantly affect cell growth and proliferation. The growth environment is also controlled so that parameters such as pH, temperature, and DO2 concentration are constant. In this section, cell kinetic growth equations are developed and subsequently applied to the analysis and design of the “perfect” bioreactor. Structured models (more realistic models that consider the multiplicity of cellular components) are discussed later on. The growth rate is deﬁned in various ways. While dCX/dt and rX appear to be the same, the two terms are only the same in batch operation. The expression dCX/dt is the change in bioreactor cell concentration, which includes the effect of media ﬂow in and out, cell recycling, and other operating conditions, while rX is the actual cell growth rate. A growth rate based on the number of cells and a growth rate based on cell weight are also not necessarily the same since average cell size may vary considerably from one stage to another. To illustrate, when the individual cell mass increases without division, the growth rate based on cell weight increases, while the growth rate based on the number of cells remains constant. During exponential growth, however, the growth rate based on cell number and growth rate based on cell weight are proportional. Growth rate can sometimes be confused with division rate (the rate of cell division per unit time). If all the cells in a vessel at time t = 0

PH2270_book.fm Page 298 Tuesday, May 16, 2006 9:37 AM

298

Producing Biomolecular Substances

(Cn = Cn0) have divided once after a certain period of time, the cell population will have increased to Cn0 × 2. If cells are divided N times after the time t, the total number of cells will be: Cn = Cn0 × 2 N

(9.246)

1 δ = (log2 Cn − log2 Cn0 ) t

(9.247)

and the average division rate is:

and the division rate at time t is: δ=

d log2 Cn dt

(9.248)

Thus, growth rate is expressed as the change in cell number with time is the slope of the Cn-vs.-t curve, while the division rate is the slope of the log2-Cn-vs.-t curve. Since the division rate is constant during exponential growth and the growth rate is not, these two terms should not be confused. If fresh, sterile medium is inoculated and cell density is measured during subsequent growth and plotted against time, the results differentiate six stages in the batch growth cycle: 1. Latent: the period during which the change in cell number is null. 2. Accelerated growth: the period during which the cell number begins to increase and the rate of cell division accelerates. 3. Exponential growth: the period during which the cell number increases exponentially, the growth rate increases, while the division rate remains constant at its maximum. 4. Decelerated growth: the period subsequent to the point where the growth rate reaches maximum, during which both growth and division rates decrease. 5. Static: the period during which the cell population reaches maximum for the given conditions and proliferation stops. 6. Death: the period after the limiting growth substance is depleted where cells begin to die and the number of viable cells decreases. Kinetic growth data and models can be used to predict the length of these growth stages in order to estimate required bioreactor size before considering other, more complex models. In addition, the same models used in designing batch cell growth processes can be used to predict the bioreactor size necessary for continuous-culture cell growth. The main objective of kinetic cell growth modeling is to support the growth of a speciﬁc culture and to promote an elevated end product yield. In certain

PH2270_book.fm Page 299 Tuesday, May 16, 2006 9:37 AM

Bioreactors

299

circ*mstances excessive essential nutrient concentration can inhibit growth (or even kill off the culture), so essential nutrients should not at all times be supplied in excess. It is a common practice to limit the concentration of at least one of the essential nutrients, keeping all others in excess, so that growth increases exponentially until the essential (limiting) substance is depleted. For single cells, growth rate can be expressed in terms of cell concentration X, and the speciﬁc growth rate μ is deﬁned by: μ=

1 dX X dθ

(9.249)

where μ is time. If the value of μ is constant, the culture is growing exponentially. Typically, batch cell growth and metabolic end product synthesis demonstrate exponential increase until limiting nutrient concentration decreases to a level at which the cells begin to die. Thus, the perfect time to terminate a bioreactor batch is when end product formation is the same rate as cell proliferation. Sometimes, when product formation lags behind cell proliferation, cell growth is allowed to proceed until desired product concentration is reached even though cells are in the death stage. Mass doubling time μd is usually determined, and the speciﬁc growth rate μ for single-cell growth is related to mass doubling time by the formula μd = ln(2)/μ. Customarily in the case of a limiting substance, the Monod relationship is used: μ=

μ m Ci K i + Ci

(9.250)

where μm is the maximum speciﬁc growth rate, Ki is a saturation constant, and Ci is the limiting nutrient concentration i. The maximum speciﬁc growth rate μm is determined when Ci >> Ki and Ki is the concentration of nutrient i when μ = μm/2. The Monod equation was derived empirically and is a simpliﬁed model of complex cell growth. Determination of μm and Ki produces a measure of population growth that can be used to predict culture population dynamics in large scale, particularly when nutrient uptake is also determined. With a more complex environment (e.g., two limiting nutrients, toxin formation, inhibition by product concentration, etc.), variations of the basic Monod formula can determine the speciﬁc growth rate μ at any concentration and at the same time predict maximum cell population for a culture under speciﬁc circ*mstances such as with two limiting nutrients: ⎛ C1 ⎞ ⎛ C2 ⎞ μ = μm ⎜ ⎝ C1 + K1 ⎟⎠ ⎜⎝ C2 + K 2 ⎟⎠ or with product inhibition:

(9.251)

PH2270_book.fm Page 300 Tuesday, May 16, 2006 9:37 AM

300

Producing Biomolecular Substances

⎛ Ci ⎞ ⎛ C P ⎞ μ = μm ⎜ ⎝ Ci + K i ⎟⎠ ⎜⎝ CP + K P ⎟⎠

(9.252)

where KP is the saturation constant for the product (Ki), and CP is the product concentration. When an STB is inoculated, it will begin to proliferate exponentially after the latent stage, so that change in cell concentration in a batch culture is equal to its cellular growth rate: dC X = rX = μC X dt

(9.253)

Then, Eq. 9.253 is integrated to develop a performance equation for the batch culture:

∫

C X0

dC X = rX

∫

C X0

dC X = μC X

∫ dt = t − t

(9.254)

but, that only applies when rX is larger than zero. Therefore, t0 in Eq. 9.254 is not the time that the culture was inoculated, but the time that the cells actually start proliferating, which is also the beginning of the accelerated growth stage. According to Eq. 9.254, the batch growth time t – t0 is the area under the 1/rX-vs.-CX curve between CX and CX0. Batch growth time is estimated by the CX-vs.-t curve, which is a more direct determination. The U-shaped curve is characteristic of autocatalytic reactions: S+X→ X+X

(9.255)

The autocatalytic reaction rate is slow at the start because the concentration is low of X functioning as a biocatalyst. The rate increases as cells multiply and reach maximum population density. As substrate is depleted and toxic metabolic waste products accumulate, the reaction rate approaches null. Monod kinetics satisfactorily represents the growth rate only during exponential growth:

∫

C X0

( K S + CS )dC X = μ maxCS C X

∫ dt

(9.256)

ΔC X C X − C X0 = − ΔCS −(CS − CS0 )

(9.257)

and the growth yield (YX/S) is: YX /S =

and change in cell concentration with respect to time is:

PH2270_book.fm Page 301 Tuesday, May 16, 2006 9:37 AM

Bioreactors

301

⎞ C ⎛ K SYX /S K SYX /S C (t − t 0 )μ max = ⎜ + 1⎟ ln X + ln S0 ⎝ C X0 + CS0 YX /S ⎠ C X0 C X0 + CS0 YX /S CS

(9.258)

Monod kinetic parameters μmax and KS cannot be approximated by a series of batch cell growths as easily as Michaelis-Menten parameters are determined by kinetic enzyme reactions. With Michaelis-Menten, the initial reaction rate in batch cell growths can be measured as a function of substrate concentration. However, in cell growth the initial rate is always zero as a result of a latent period during which Monod kinetics does not apply. Although Monod has the same general form as the Michaelis-Menten, the respective rate-reaction components are different. In Michaelis-Menten: dCP r C = max S dt K M + CS

(9.259)

dC X μ maxCS C X = dt K S + CS

(9.260)

and in Monod:

The latent stage is the period where the cell population growth rate is either null or negligible, although cells are still able to increase in size. This occurs while cells adjust to their new medium and environment before initiating accelerated growth. Factors such as cell type, age, inoculum size, and culture conditions determine the stage length. For example, if a culture is inoculated from a low-nutrient concentration medium to that of higher-nutrient concentration, the length of the latent stage increases. If, on the other hand, the culture is inoculated from a high- to a lowernutrient medium, there is typically no latent stage. Another important factor affecting latent stage length is inoculum size (i.e., if a small number of cells are inoculated into a large volume, they will generally experience an extended latent stage). In large-scale cell growth, a primary objective is to shorten this stage as much as possible; in order to inoculate a large bioreactor a series of progressively larger seed lots are created to minimize latent stage effect. Batch Culture Batch culture involves inoculating sterile medium with a seed culture of the cells to be grown. After inoculation, apart from air addition to aerobic cell growths and the removal of waste gasses, usually nothing else is required to be added or removed from the batch. The rapid change to a new environment (the sterile bioreactor) can affect four important variables: (1) inoculation into a medium with high nutrient concentrations can cause delayed cell growth until the culture adapts to the new environment; (2) essential molecules synthesized by the cell to promote growth (vitamins, activators) may be lost by diffusion out of the cells and may take time

PH2270_book.fm Page 302 Tuesday, May 16, 2006 9:37 AM

302

Producing Biomolecular Substances

for replenishment; (3) the inoculum size and viable cell percentage greatly affect the duration of the latent stage; (4) the maturity of the inoculum is important because newer cells have not stored the same quantity of required metabolic substances as cells already in exponential growth. The latent stage ends once a key cellular component reaches a critical level c within the cell, therefore: c = aV + bN 0θ1 + d θ1

(9.261)

where V is the volume of inoculum, N0 is the number of cells/new volume, a is the key cellular limiting component concentration/old volume x, b is the increase in key cellular component concentration/time per cell (for older cells), d is the internal cell production of the key cellular component (for newer cells), and θ1 is the time of the latent stage. Since θ1 is dependent on V, the larger V is, the shorter θ1 will be, so θ1 is proportional to 1/N0 for large inoculum volumes. Bioreactor Design Bioreactor design should minimize the length of the latent stage to get maximum utilization out of the bioreactor. The following three points are important: (1) the inoculum should be as active as possible (preferably in exponential growth); (2) the inoculum medium should correspond as closely as possible to that of the bioreactor; and (3) a reasonably large volume of inoculum (at least 5% of the total bioreactor volume) should be used to minimize losing key metabolic intermediates by diffusion. When accelerated growth begins, it increases gradually, reaching its maximum rate during the exponential growth stage. This transitional period is frequently called the accelerated growth stage and is sometimes deﬁned as part of the latent stage. With single-cell organisms, a culture undergoing balanced growth emulates a ﬁrstorder autocatalytic reaction (e.g., the progressive doubling of cell number results in a continually increasing rate of growth). Thus, the rate of cell increase at any particular time is proportional to the number (Cn) of cells present at that time: rn =

dCn = μCn dt

(9.262)

where the constant μ is known as the speciﬁc growth rate (hr–1). The speciﬁc growth rate should not be confused with the growth rate which has different units and meaning. The growth rate is change in cell number with time while the speciﬁc growth rate is: μ=

1 dCn d ln Cn = Cn dt dt

(9.263)

which is change in the log of the cell number with time. Comparing Eqs. 9.262 and 9.263 shows that the speciﬁc growth rate μ, is equal to ln 2 times the division rate,

PH2270_book.fm Page 303 Tuesday, May 16, 2006 9:37 AM

Bioreactors

303

μ. If μ is constant with time during the exponential growth period, the expression can be integrated from time to t0 yielding Cn = Cn0 exp[μ(t − t 0 )

(9.264)

where Cn0 is the cell number concentration at t0 when the exponential growth starts. Thus, Eq. 9.264 shows exponential increase in the number of cells with respect to time. The time required to double the population, called the doubling time (td), can be determined by establishing Cn = 2Cn0 and t0 = 0, and solving for t: td =

ln 2 1 = μ δ

(9.265)

The doubling time is inversely proportional to the speciﬁc growth rate and is equal to the reciprocal of the division rate. The exponential growth rate where X0 is the initial population at time 0 and X is the population at time μ for a key cellular limiting component μ, for cell population X at time, μ is: ⎛ X ⎞ μ Cθ ln ⎜ ⎟ = m i ⎝ X 0 ⎠ Ci + K i

(9.266)

This predicts the culture’s exponential growth if Monod applies, but Monod is not precise and if growth is extremely rapid the two following equations can be used: (1) where C0 is the initial key cellular component concentration: μ=

μ m Ci K i + C0

(9.267)

or (2), where B is a constant and N is the cell population. μ=

μ m Ci Ci + BN

(9.268)

When the rate of consumption of a key cellular limiting component (dA/dμ) is proportional to the number of viable cells reaching static stage, where X is the number of cells, and KA is a proportionality constant for key cellular limiting component A, if the culture is in exponential growth, and if the initial concentration of A = A0 and X = X0 (initial concentration of cells, when A = 0, X = Xs, the static population), then this relationship yields the maximum population (XS) for initial concentration of key cellular component A (A0) and initial cell population X0 at the start of exponential growth.

PH2270_book.fm Page 304 Tuesday, May 16, 2006 9:37 AM

304

Producing Biomolecular Substances

XS = X0 +

μ A0 KA

(9.269)

Integrating the equation between X = X0 and X = XS generates the time taken to reach static stage. In normal practice, cell growth is terminated either before the death stage is reached or just as the cell population starts to decrease as a result of total depletion of key cellular component A. An exception would occur, however, when product formation exhibits a sizable lag behind cell proliferation; thus, if product concentration is still increasing, cell growth could be taken into the death stage. Factors Affecting Specific Growth Rate As previously stated, Monod is the most widely employed equation for the effect of substrate concentration on μ. It is an empirical expression based on the type of equation normally associated with enzyme kinetics: μ=

μ maxCS K S + CS

(9.270)

where CS is the concentration of the limiting substrate and KS is a system coefﬁcient. The value of KS is equal to the concentration of the key metabolic limiting substance when the speciﬁc growth rate is half of maximum (μmax). And, while Monod is an oversimpliﬁcation of the cell growth mechanism, it frequently describes cell growth systems with low cell growth inhibiting component concentrations. According to Monod, further increase in limiting substance concentration after μ reaches μmax does not affect the speciﬁc growth rate. Also, the speciﬁc growth rate decreases as substrate concentration is increased beyond a certain level. The following expressions have been created to improve Monod: μ=

μ maxCS K I1 + CS + ( K I 2 CS )2

μ = μ max (1 − e− CS / K S ) μ=

(9.271)

(9.272)

μ max (1 + K S CS− λ )

(9.273)

μ maxCS βn + CS

(9.274)

μ=

If several limiting substances are used, however, the following expression can be employed:

PH2270_book.fm Page 305 Tuesday, May 16, 2006 9:37 AM

Bioreactors

305

μ = μ max

C1 C2 K1 + C1 K 2 + C2

(9.275)

If the limiting substance is the culture’s primary energy source, a certain amount of substrate is used for other purposes than growth, and some expressions include the term, ke, for cell subsistence: μ=

μ maxCS − ke K S + CS

(9.276)

When CS is so low that the ﬁrst term of the right side of Eq. 9.276 is μ ke, the speciﬁc growth rate is null. The alternative models appear to be better growth models for certain microorganisms, although their solutions are more difﬁcult than Monod’s. Cell growth produces metabolic wastes that collect in the medium and usually inhibit cell growth; their effect can also be added to Monod, as in: ⎛ CS ⎞ ⎛ K P ⎞ μ = μ max ⎜ ⎝ K S + CS ⎟⎠ ⎜⎝ K P + CP ⎟⎠ C ⎞ ⎛ CS ⎞ ⎛ μ = μ max ⎜ 1− P ⎝ K S + CS ⎟⎠ ⎜⎝ CPm ⎟⎠

(9.277)

(9.278)

The CPm term is the maximum metabolic waste concentration allowed in the medium for uninhibited cell growth. The speciﬁc growth rate of cells or microorganisms is directly affected by medium pH, temperature, oxygen supply, etc., since optimal pH and temperature generally differ from one organism to another. When a culture’s numerical growth is limited by the exhaustion of the available limiting substance(s) and/or the accumulation of metabolic waste products, the rate of growth declines and eventually stops. At this point the culture is said to be in the static stage. Transition from exponential to static growth involves an erratic growth period during which various cellular components are synthesized at unequal rates and consequently have chemical compositions different from those of cells in exponential growth. The static stage is usually followed by a death stage. Death occurs either because of cellular energy reserve depletion and/or as a result of metabolic waste accumulation. Like the preceding growth stages, the death stage is exponential. In some cases, the organisms not only die but also decompose. Kinetic Models The material balance for a culture in a CSTB can be shown where rX is bioreactor cell growth rate and dCX/dt represents change in cell concentration with time. For valid steady-state bioreactor operation, the cell-concentration change over time must

PH2270_book.fm Page 306 Tuesday, May 16, 2006 9:37 AM

306

Producing Biomolecular Substances

equal zero (dCX/dt = 0), and since the culture grows quickly enough to replace cells lost through the outlet: τm =

V C X − C Xi = F rX

(9.279)

this reﬂects the required residence time as equal to CX – C xi times 1/rX on the 1/rXvs.-CX curve. Graphing cell residence times compares bioreactor effectiveness; the shorter the residence time in reaching a particular cell concentration, the more effective the bioreactor. If the input stream is sterile (CX = 0), and the cells are growing exponentially (rX = μCX), for a CSTF at steady state with a sterile feed, the speciﬁc growth rate μ is equal to the dilution rate D and is equal to the reciprocal of the residence time μm since the speciﬁc growth rate of a culture can be controlled by changing the media ﬂow rate. So, if the system’s growth rate can be expressed by Monod, then: D=μ=

1 μ C = max S τ m K S + CS

(9.280)

CS can be calculated with known residence time and Monod kinetics as: CS =

KS 1 τ mμ −max

(9.281)

It should be noted, however, that the preceding is only valid when μmμmax > 1. If μmμmax < 1, the growth rate of the cells is less than the rate of cells leaving the outlet stream, all the cells in the bioreactor will be washed out, and Eq. 9.281 is invalid. If the growth yield (YX/S) is constant, then: C X = YX /S (CSi − CS )

(9.282)

⎛ KS ⎞ C X = YX / S ⎜ CSi − 1 ⎟ τ mμ −max ⎝ ⎠

(9.283)

⎛ KS ⎞ CP = CPi + YP / S ⎜ CSi − 1 ⎟ τ mμ −max ⎝ ⎠

(9.284)

CX is:

similarly, CP is

PH2270_book.fm Page 307 Tuesday, May 16, 2006 9:37 AM

Bioreactors

307

Again, the above equations are valid only when μmμmax > 1; the same values for CSTB cultures can also be calculated by determining material balances for substrate and product concentrations. Equality of the speciﬁc growth and dilution rates in steady-state cell growth is helpful in studying the effects of media components on the speciﬁc growth rate. By measuring steady-state substrate concentration at various ﬂow rates, kinetic models can be evaluated and parameters can be approximated. If the particular culture follows Monod, the plot of 1/μ-vs.-1/CS yields μmax and KS from the intercept and slope of a plot similar to that of Michaelis-Menten, with the advantage that it shows the relationship between the independent variable (CS) and the dependent variable (μ). However, since 1/μ approaches inﬁnity μ as the substrate concentration decreases, undue importance is given measurements made at low-substrate concentrations and insufﬁcient importance is given measurements made at high-substrate concentrations. For a better parameter match, the following two linear relationships can be employed: CS K C = S + S μ μ max μ max

(9.285)

μ CS

(9.286)

μ = μ max − K S

PH2270_book.fm Page 308 Tuesday, May 16, 2006 9:37 AM

308

Producing Biomolecular Substances

oxidation, inactivation, and contamination by proteolytic enzymes and other cell lysis products than those produced by continuous culture because the product residence time in the CSTB is shorter, and the removed end product can be chilled to 4°C to minimize degradation. Further, by reducing bioreactor vessel size, the facility dimensions, utility requirements, media handling problems, and downstream puriﬁcation costs can all be reduced, permitting the use of conventional clean-room technology. Continuous culture can reduce capital requirements for bioprocess facilities and put much less high-value product at risk during a particular time interval, as compared with batch systems where everything can be lost if contamination, mechanical failure, or incubation failure should occur. Continuous culture bioreactor design and analysis is based upon the CSTB. Four assumptions required for a CSTB system analysis are: (1) mixing takes place so that the exit stream contents have the same composition as the rest of the bioreactor vessel contents, (2) the concentration of all components in the bioreactor vessel is the same in all areas of the vessel, (3) if the process is aerobic — the concentration of dissolved oxygen is the same in all parts of the vessel, and (4), the heat transfer characteristics of the system are constant (i.e., heat generated by cell growth is continuously removed). When concentrations, cell population, and temperature in any area of the vessel do not change with time, a steady state has been reached and a mass balance can be applied to any component such that: ⎡ rate ⎤ ⎡ rate ⎤ ⎡ rate ⎤ ⎢ of ⎥ ⎢ of ⎥ ⎢ ⎥ of ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ addition ⎥ − ⎢ removal ⎥ + ⎢ production ⎥ = 0 ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ to ⎥ ⎢ from ⎥ ⎢ within ⎥ ⎢⎣ system ⎥⎦ ⎢ system ⎥ ⎢ system ⎥ ⎣ ⎦ ⎣ ⎦

(9.287)

A steady state-balance (neglecting cell death) is depicted by: ⎛ F⎞ ⎛ F⎞ ⎜⎝ ⎟⎠ X0 = ⎜⎝ ⎟⎠ X − r V V

(9.288)

If X is the cell concentration in the bioreactor vessel and exit stream, X0 is the cell concentration of the feed, F is the ﬂow rate of the feed and exit stream, V is the vessel volume, r is the rate of cell formation (cells/unit time/unit volume) equal to dX/dμ, then D (= F/V), known as the dilution rate, is the number of bioreactor vessel volumes passing through the system per unit time (the inverse of the mean residence or holding time). The dilution rate is universally employed throughout the biotechnology industry. With a single CSTF, the feed stream is normally sterile medium. Therefore, X0 is zero, and X(D – μ) = 0. Since either X = 0 or (D – μ) = 0, a cell population > 0 can be maintained only if the speciﬁc growth rate μ is balanced by the dilution rate

PH2270_book.fm Page 309 Tuesday, May 16, 2006 9:37 AM

Bioreactors

309

D (i.e., a nonzero cell population can only be maintained when D = μ). Once the culture has adjusted its speciﬁc growth rate to the dilution rate, the expression DX0 = X (D – μ) can be satisﬁed by any X value greater than zero. This condition is only reliable when: (1) speciﬁc growth rate is independent of cell population, and (2) growth is exponential and not affected by decreasing concentration of the growthlimiting key cellular component. For continuous culture a static maintenance bioreactor, or chemostat, can be used to maintain cell densities at as high as 100 times the density of conventional systems. Stocked with cells, porous tubes constantly input nutrients and remove waste products and gas permeable tubes continuously add oxygen and remove carbon dioxide while computer controls monitor these functions and maintain a hom*oeostatic environment. During continuous culture the cell population is maintained both biochemically and physiologically constant over an indeﬁnite period. Major advantages of chemostatic continuous culture include the ability to: (1) manipulate the speciﬁc growth rate over a full range without the need to vary medium composition or environmental conditions, (2) strategically evaluate the effects of the physico-chemical environment while maintaining a constant growth rate, (3) analyze both the culture’s steady state and deﬁned transitory states, (4) set unique culture conditions by changing the nature of the growth-limiting substrate(s), and (5) maintain culture growth over long periods. Chemostat theory predicts signiﬁcantly higher productivity than batch cultures. Productivity is maximized by careful manipulation of environmental parameters, reduced downtime for cleaning, and elimination of the time-consuming, repetitive operations (e.g., sterilization, cleaning, etc.) associated with batch cultures. Continuous culture is an important bioprocess optimization tool that relates not only to productivity, but to concentration, substrate conversion efﬁciency, duration of biosynthetic activity, and product stability. In a batch culture, if one of the key cellular components is at growth-limiting concentration, the cell balance at steady state can be written using the Monod expression for speciﬁc growth rate μ, and if the yield factor Y is deﬁned as: Y=

mass of cells formed nutrient consumed

(9.289)

the steady-state balance on the limiting component is expressed by the Monod chemostat model. Where X0 = 0 (sterile feed): D(C0 − Ci ) −

μ m Ci X =0 Y (Ci + K i )

(9.290)

Once the dilution rate has exceeded the maximum possible growth rate, the only solution is X = 0. The situation where D exceeds Dmax and all the cells are lost from the system is known as washout. When X = 0 and Ci = C0, then: Dmax =

μ m C0 C0 + K i

(9.291)

PH2270_book.fm Page 310 Tuesday, May 16, 2006 9:37 AM

310

Producing Biomolecular Substances

Optimal Conditions The cell production rate per bioreactor volume is DX (FX/V), and the maximum cell production rate is: d ( DX ) =0 dD

(9.292)

and the rate for optimal cell production is 1/ 2

⎤ ⎡ Ki Dopt = μ m − ⎢ ⎥ ⎣ (C0 + K i ) ⎦

(9.293)

Thus, if C0 >> Ki (which is usually the case), Dopt approaches μm near washout. In terms of controlling a CSTB, since X and Ci are sensitive near washout, optimal (maximum) cell production is not permitted so that bioreactor operation will remain stable. Typically, cell growths are carried out at a dilution rate of around 80% of Dmax. Determining batch cycle residence time and the processing time for continuous cultures facilitates a close estimation of the required bioreactor volumetric capacity. Moreover, since the bioreactor is really a multipurpose processor and in practice performs a number of functions, to ensure proper operation: (1) as a bioprocessor, the bioreactor must be sized to provide the required production capacity; (2) as masstransfer equipment, the bioreactor must be designed to ensure that media and cells are well-dispersed and, for aerobic cultures, that adequate dissolved oxygen concentration is maintained and available for individual cells; (3) as a control device, the bioreactor must provide for temperature control for the rapid dispersion of control chemicals (i.e., for pH and foaming) and ensure representative sampling of important parameters (i.e., cell population, pH, dissolved oxygen concentration, etc.); (4) as a heat-transfer device, the bioreactor must ensure that constant temperature is maintained during the growth cycle by cooling and heating to sterilize the system in situ. Having determined the required bioreactor volume, consideration must next be given to including other necessary functions in the bioreactor model. Maximizing Productivity Continuous culture is most often the method of choice for obtaining relevant information about a bioprocess, although maximizing productivity during continuous culture is often quite difﬁcult. Contamination and natural selection of spontaneous mutations must be constantly monitored. Also, small environmental deviations attributable to bioreactor operational problems typically cause signiﬁcant reductions in cell population. Continuous culture parameters are based on the growth kinetics of single cells in exponential growth. Although growth should go to inﬁnite population density in theory, in practice an essential growth-rate-limiting substance is depleted and/or toxic metabolic waste product(s) are formed, resulting in limited population density. A good continuous

PH2270_book.fm Page 311 Tuesday, May 16, 2006 9:37 AM

Bioreactors

311

culture maximizes viable cells and minimizes dead and dying cells. Both biochemical and physical techniques are typically used to detect and quantify living cells in the population (see Table 9.1). The dilution rate for a continuous culture is equal to the ﬂow rate of fresh medium into the bioreactor divided by bioreactor volume. During continuous culture the rate of change in organism concentration in the vessel equals the rate of growth minus organisms removed. At steady state the biomass and growthrate-limiting substrate concentrations are constant and the speciﬁc growth rate can be controlled by substrate concentration. At the critical dilution rate a loss of cell population (washout) will occur. The maximum speciﬁc growth rate can be practically determined by increasing the dilution rate until washout. Substrate consumption during continuous culture follows Monod ﬁrst-order kinetics. In continuous culture, the biomass concentration and growth-limiting substance concentration can be determined from the supply vessel concentration of the growth-limiting substrate and the biomass yield of that substrate. Recombinant Culture Kinetics Assuming the probability of unmutated cells (X+) to produce mutated cells (X–) is p after one division, then N unmutated cells will produce N(1 – p) unmutated cells and Np mutated cells after one division, and the total number of X+ cells will be N(2 – p). During exponential growth, the unmutated cell growth rate will be: dC X + = (1 − p )μ + C X + dt

(9.294)

where μ+ is the speciﬁc growth rate of the unmutated cells, and CX+ is the number of unmutated cells per unit volume. If the mass of the cells is approximately proportional to the number of cells, the preceding rate equation can be also applicable when CX+ the mass of cells per unit volume is, and the growth rate of the mutated cells will be: dC X − = pμ +C X + + μ −C X − dt

(9.295)

If we assume μ+ and p are constant, the integration of Eq. 9.295 yields: C X + = C X + exp[(1 − p )μ +t ] 0

(9.296)

where C X + is the initial concentration of the unmutated cells. For the mutated cells, substituting Eq. 9.296 into Eq. 9.294 yields: dC X − − μ −C X − = pμ +C X + exp[(1 − p )μ +t ] 0 dt

(9.297)

PH2270_book.fm Page 312 Tuesday, May 16, 2006 9:37 AM

312

Producing Biomolecular Substances

Solving Eq. 9.297 for the constant μ+ yields:

CX− =

pμ +C X + 0

(1 − p )μ + − μ −

{exp[(1 − p )μ +t ] − exp(μ −t )}+ C X0− exp(μ −t )

(9.298)

Therefore, Eqs. 9.297 and 9.298 predict how C X + and C X − change with time for given values of ρ, μ+, and μ–. The fraction of the unmutated cells in the total population ƒ can be expressed as: f=

CX+ CX+ + CX−

(9.299)

Substituting Eqs. 9.297 and 9.298 into Eq. 9.299 yields: f=

exp[(1 − p )μ +t ] pμ +C X + 0 {exp(1 − p )μ +t − exp(μ −t )} exp[(1 − p )μ +t ] + (1 − p )μ + − μ −

(9.300)

which shows the change in the fraction of the unmutated cells with time during the exponential growth of a batch cell growth. During exponential growth, the number of generations η of the unmutated cells can be calculated from: n=

μ +t ln 2

(9.301)

Combining Eqs. 9.300 and 9.301 will result in f for the nth generation, fn =

1− α − p 1 − α − p[ 2 n (α + p−1) ]

(9.302)

where α is the ratio of the speciﬁc growth rates: α=

μ− μ+

(9.303)

Eq. 2.302 predicts the change in f for the number of generations required for a series of batch cell growths if it is assumed the culture propagated exponentially during each batch. Assuming that it could take about 25 generations to scale up from a in vitro culture to a 33,000-liter bioreactor, this shows the effects of p and μ on ƒ25, which decreases as μ and ρ increase. When is 0.01 and μ < 1, ƒ25 is close to 1 (i.e.,

PH2270_book.fm Page 313 Tuesday, May 16, 2006 9:37 AM

Bioreactors

313

the unmutated cells are quite stable). However, as μ approaches 2.0, ƒ25 becomes zero. Practically, values of μ generally range from 1.0–2.0. The change in ƒ as a function of η and μ, where ρ was constantly adjusted to be 0.01. The value of f decreases rapidly with the increase of n and μ, and when μ is 1.4, all unmutated cells will mutate in about 33 generations. The material balance for unmutated recombinant cells in a CSTB is: − DC X + + (1 − p )μ +C X + =

dC X + dt

(9.304)

Similarly, the material balance for unmutated cells is: − DC X − + pμ +C X + + μ −C X − =

dC X − dt

(9.305)

and, adding Eqs. 9.299 and 9.301 yields the total cell concentration: (μ +C X + + μ −C X − ) − D(C X + + C X − ) =

d (C X + + C X − ) dt

(9.306)

If the CSTB is operated so that the total cell concentration is constant with time, then: μ + (C X + + αC X − ) = D(C X + + C X − )

(9.307)

and if μ = 1, then Eq. 9.307 simpliﬁes to: μ+ = D

(9.308)

Thus, the speciﬁc cell growth rate in a bioreactor is constant and ordained by the dilution rate, and solving Eqs. 9.306 and 9.307 after substituting Eq. 9.308 yields C X + = C X + exp(− pDt )

(9.309)

C X − = C X0− + C X + [1 − exp(− pDt )]

(9.310)

and

During continuous cell growth, then, the concentration of unmutated cells is reduced and the concentration of mutated cells increases. If μ does not equal 1, μ+ is no longer constant during steady-state CSTB operation at a constant dilution rate, but depends on the cell concentrations (C X + , C X − ) and on μ, according to Eq. 9.307. As

PH2270_book.fm Page 314 Tuesday, May 16, 2006 9:37 AM

314

Producing Biomolecular Substances

a result, μ+ changes with time. Just how C X + and C X − change with time can be estimated by simultaneously solving Eqs. 9.307, 9.308, and 9.310. Substituting Eq. 9.310 into Eq. 9.307 and dividing by C X + + C X − yields: df (1 − p) Df = − Df + dt α + (1 − α) f

(9.311)

The solution of Eq. 9.311 shows how the fraction of unmutated cells decreases with time (change in f with time). The initial value of ƒ for the solution of Eq. 9.311 is approximated from Eq. 9.309. Change in the initial fraction of unmutated cells with time during steady-state CSTB operation (initial f value), for example, is approximated by assuming that 20 generations are required for the stepwise inoculation, the initial batch, and any subsequent unsteady-state continuous cell growth. The fraction of unmutated cells, therefore, is reduced with both time and any increase in the dilution rate. When ρ and D are sufﬁciently low and μ approaches 1, the CSTB can be economically operated over long periods. However, if μ increases to 1.9, the CSTB will lose almost all its unmutated cells within 100 hours of steadystate operation. The following six methods can be used to stabilize recombinant culture systems that tend to mutate during cell growth: 1. 2. 3. 4.

Medium should favor growth of unmutated cells over mutated cells. Maintain selective pressure against mutated cells. Employ temperature-dependent strains employed. Employ temperature-dependent gene expression control; mutation is less likely when gene expression is repressed; the higher the genetic expression, the more segregants tend to appear. 5. Cultures should contain no transposable elements. 6. Employ a recombination-deﬁcient strain. Cell-Growth Modeling in a Batch Bioreactor The simplest way to model cell growth will be to consider an unstructured, unsegregated model for cell growth. For this kind of model: rx = dX/dt = mX

(9.312)

where, rX = rate of cell generation (g/l-hr), X = cell concentration (g/l), and m = speciﬁc growth rate (hr–1) The most commonly used expression that relates the speciﬁc growth rate of the cell to the substrate concentration is the Monod equation, which is: m = mmaxS/(Ks + S)

(9.313)

where, m = speciﬁc growth rate (hr–1), mmax = maximum speciﬁc growth (hr–1), S = substrate concentration (g/l), and KS = saturation constant for substrate (g/l).

PH2270_book.fm Page 315 Tuesday, May 16, 2006 9:37 AM

Bioreactors

315

mmax

D = F/V

Specific Growth Rate (m)

Substrate Concentration (S)

FIGURE 9.13 The Monod growth curve.

Figure 9.13 depicts Monod’s equation. One should note that Monod is empirical and does not have any mechanical basis. The equation is only valid for an exponentially growing a culture under balanced growth conditions. The equation does not represent transient conditions well. Despite the simplicity and no apparent fundamental basis, it still works surprisingly well in a large number of steady-state and dynamic situations. This characteristic has important implications in control of bioreactors. Continuous Bioreactor Dynamics For a continuously fed bioreactor, the cells are continuously supplied substrate at growth-limiting level, and hence they remain in the exponential phase. Since the cells remain in the exponential phase, Monod can be applied. Bioreactor cell balance can be written as: FX – FXƒ + V(dX/dt) = rx

(9.314)

where, F = volumetric ﬂow rate (l/hr), X = cell concentration inside the reactor and in the outlet stream (g/l), Xƒ = cell concentration in the feed (g/l), V = reactor volume (l), and rX = rate of cell generation (g/l-hr). For a sterile feed (Xƒ = 0), and noting that the reaction rate can be written in terms of the speciﬁc growth rate (rx = mX), Eq. 9.314 can be reduced to: dX/dt = (m – D)X

(9.315)

where D = dilution rate = F/V (hr–1). And, balance on the substrate yields: FS – FSƒ + V(dS /dt) = rsV

(9.316)

PH2270_book.fm Page 316 Tuesday, May 16, 2006 9:37 AM

316

Producing Biomolecular Substances

…where, F = volumetric ﬂow rate (l/hr), S = cell concentration inside the bioreactor and in the outlet stream (g/l), Sf = substrate concentration in the feed (g/l), V = reactor volume (l), rs = rate of substrate consumption (g/l-hr). A yield parameter (Yx/s) is deﬁned that relates the amount of cell mass produced per amount of substrate consumed, and is mathematically represented as: Yx/s = mass of cells produced/mass of substrate consumed = rx/–rs (9.317) Combining Eqs. 9.317, 9.321, and 9.322 yields: dS/dt = D(Sf – S) – mX/Yx/s

(9.318)

The CSTB is now completely described by Eqs. 9.320 and 9.323 with m given by Eq. 9.318. At steady state (with ﬁxed Sf and D), the following are the values for m (speciﬁc growth rate), S (substrate concentration), and cell concentration (X): m=D

(9.319)

S = DKS/(mmax – D)

(9.320)

X = Yx/s(Sf – S)

(9.321)

There are a few characteristics of an open-loop CSTB, which are conceptually different from that of a chemical reactor, that are important to know before any control system for a bioreactor can be designed. D must be less than mmax to achieve a realistic value of S. The same conclusion can be derived by looking at the steady state solution in Eq. 9.320. The two solutions are Eq. 9.321 and X=0

(9.322)

The corresponding substrate concentration is S = Sf

(9.323)

Eqs. 9.322 and 9.323 deﬁne a situation called washout. This situation is encountered whenever the value of dilution rate equals or exceeds mmax. A rigorous discussion of washout would point to the fact that whenever m (Sƒ), i.e., m evaluated at Sƒ is less than mmax, then the critical dilution rate for washout will occur at D = m (Sƒ), and not at D = mmax. The control algorithm should be completely aware of this unproductive state. For the given set of equations, numerical solution is required since the system is described by two coupled nonlinear differential equations, i.e., Eqs. 9.320 and 9.323. Linear control theory can be applied in only a limited sense, i.e., only near the steady state when the system model is linearized.

PH2270_book.fm Page 317 Tuesday, May 16, 2006 9:37 AM

Bioreactors

317

Start up is an important consideration as well. The general procedure in the start up avoiding washout would be to initiate cell growth in a batch mode until the exponential phase is reached. At this point, the sterile feed would be started with a dilution rate such that D < m (Sƒ). A nonwashout steady state could be reached after a transient phase. Multiplicity and Steady-State Stability in a Continuous-Culture Bioreactor Although the control loop of a CSTB is simple, the system is complicated by the presence of multiple steady states and the stability considerations of these steady states. The next discussion will highlight these problems. As already implied, the control design of a biological reactor described by Eq. 9.320 and 9.323 should take into account the nonlinear nature of these differential equations. Multiple critical points are common with nonlinear systems. This has been shown earlier in the discussion of washout. A systematic approach to an efﬁcient control design will involve: 1. Calculating the number of steady states. 2. Characterizing the nature of the steady states with respect to their stability. 3. Designing appropriate control loops based upon steps 1 and 2. Calculating Multiple Steady States Once the governing equations describing the system are in place, the steady states are found by replacing all time derivatives by zero. This can be done by inspection and algebraic solution. For high order or complex models, a nonlinear root-ﬁnding technique should be used. Steady State Stability A steady state is stable if, for initial conditions near the steady state, all transients converge to it. If the transients diverge, steady state is called unstable. The diverging transients always end at some other stable state. Stability analysis of the steady state involves whether or not the steady state under construction is stable, and the information about state-to-state transitions, in the case of an unstable steady state, is obtained near the steady state. Information about steady-state stability and local dynamics is accomplished through linear stability analysis. Results of this analysis are only valid near the steady state, and generation of this phase plane is only suitable for general (nonlocalized) behavior and provides information only about state-to-state transitions. Proportional Control of STBR with Monod Before designing the closed-loop continuous bioreactor, one should understand the open-loop STBR fully since the scope of closed-loop STBR will be given only by

PH2270_book.fm Page 318 Tuesday, May 16, 2006 9:37 AM

318

Producing Biomolecular Substances

the knowledge about the open-loop STBR. Linear stability analysis and phase plane analysis for open-loop CSTB and closed-loop STBR are detailed below. Linear Stability Analysis of an Open-Loop STBR As already discussed, for an open-loop STBR with Monod kinetics, there exist two steady states, i.e., a nontrivial steady state (deﬁned by Eqs. 9.328, 9.329, and 9.330) and washout steady state (deﬁned by Eqs. 9.331 and 9.332). The Jacobian J for the system deﬁned by Eqs. 9.320, and 9.327 with m given by Eq. 9.318 is: J=

δX ′ / δX δS ′ / δS

δX ′ / δS δS ′ / δS

(9.324)

where X′ = dX/dt. S′ = dS/dt

(9.325)

Substituting for X′ and S′ from Eqs. 9.320 and 9.325 yields the value of the Jacobian for this system: J=

m−D − m / YX /S

Xm′ − D − m′X / YX /S

(9.326)

where m′ = δm/δs. For a nontrivial steady state, stability is guaranteed if the following equations are satisﬁed: Trace J < 0

(9.327)

Det J > 0

(9.328)

–D – m′X/ Yx/s < 0

(9.329)

Xm′m/ Yx/s > 0

(9.330)

This yields:

In the case of Monod, m′ > 0 for all S. Therefore, the nontrivial state is always stable. For the washout state, conditions for stability derived from a similar procedure are: m(Sƒ) – D < 0

(9.331)

(m(Sƒ) – D)(–D) > 0

(9.332)

PH2270_book.fm Page 319 Tuesday, May 16, 2006 9:37 AM

Bioreactors

319

washout

1.0

S / SR .6 .4 β

1.0

X / SR

FIGURE 9.14 Open-loop phase plane for bioreactor with Monod kinetics; where SR is feedsubstrate concentration and and are the steady-state cell and substrate concentrations.

Eqs. 9.331 and 9.332 indicate that D must be greater than m (Sƒ) for the washout steady state to be stable. Thus, any dilution rate which gives any realistic solution (X > 0, and S > 0) will result in washout being unstable. The same conclusion can be derived by the phase plane analysis that is also discussed in the next section. Phase Plane Analysis for Open-Loop CSTBR Construction of the phase plane for open-loop CSTBR be achieved via integrating equations 9.321 and 9.328, selecting several time points, plotting the values of S and X at each point, then repeating for new initial conditions or sketched directly from the results of the linear analysis. As shown in Figure 9.14, all initial conditions result in achievement of the desired steady state. The motive for controlling this reactor would be to maintain a closed-loop system such that washout could be avoided regardless of ﬂow ﬂuctuations. An easy approach to achieve this would be to measure the cell concentration and manipulate the ﬂow rate to force the reactor to nontrivial steady state. This can easily be accomplished with a simple proportional controller whose stability analysis is discussed in next subsection. Stability Analysis of Closed-Loop Bioreactor The governing equation for the proportional controller that manipulates the ﬂow rate as a response to changing cell concentration inside the reactor is given by: D = Dss + Kc(X – Xsp)

(9.333)

Where, D = dilution rate that is manipulated by the controller (hr–1), Dss = dilution rate corresponding to the nontrivial steady state for X = Xsp in open loop CSTBR

PH2270_book.fm Page 320 Tuesday, May 16, 2006 9:37 AM

320

Producing Biomolecular Substances

(hr–1), Kc = controller gain (l/g-hr), X = cell concentration in the reactor (g/l), and Xsp = controller set point and the desired cell concentration in the reactor (g/l). Substituting the value of D from Eq. 9.333 into Eqs. 9.320 and 9.327 shows that X = 0, S = Sƒ is no longer a steady-state solution. A rigorous analysis for this system will show that the worst case for this system as X approaches 0 corresponds to D = 0. This is equivalent to saying that the CSTB will approach the behavior of a batch reactor. The conditions for the stability of the system under consideration according to linear stability analysis are: Kc – Dss – m′/ Yx/s < 0

(9.334)

Kc + m′/ Yx/s > 0

(9.335)

Any positive value of Kc is sufﬁcient to satisfy Eqs. 9.334 and 9.335, and hence guarantee stability. This is not surprising, keeping in mind the stability of nontrivial steady state in open-loop CSTB. It seems fair to expect the closed-loop phase plane similar to open-loop phase plane for reasonable values of Kc. The whole discussion can be summarized as follows. Since the nontrivial state is always stable for realistic D values, there is little incentive for closed-loop operation other than to prevent washout from large ﬂow disturbances. The incentive for closed-loop operation increases signiﬁcantly if the growth kinetics are more complex (e.g., substrate inhibited growth kinetics). This is discussed in the next section. Controlling Continuous-Culture Bioreactors with Substrate Inhibition Although Monod makes a nice substrate model of limited cases, it does not approximate real situations well, since biological systems are inhibited by high substrate concentration. Thus, understanding STBs is important: STB dynamics with substrate inhibition kinetics points to an interesting control problem. Those interested can ﬁnd a detailed analysis by Dibiasio in the bibliography for open- and closed-loop STBs with substrate inhibition. A number of reports have appeared regarding control issues relative to continuous-culture bioreactors. Fed Batch Reactor Dynamics Though CSTBs are an excellent tool to study bacterial metabolism, they are not extensively used in the biotechnology industry. The most widely used cell-culture mode for industrial production of proteins is fed-batch cell culture. Fed batch is an interesting system, since it does not have a true steady state; evaluation of the steadystate variables will position the system all the way through the operational cycle. Since steady-state variables can not be measured online, their estimation becomes an important element of reactor optimization and control. There are many reports on the theoretical and functional issues relative to fed-batch culture.

PH2270_book.fm Page 321 Tuesday, May 16, 2006 9:37 AM

Bioreactors

321

FIGURE 9.15 Bioreactor design window.

Bifurcation Analysis Bifurcation analysis is a convenient method for classifying the various dynamic behaviors possible with a cell culture model and is extensively discussed by Razon and Schmitz.

BIOREACTOR DESIGN PROGRAM With this bioreactor design program, you can design your own bioreactor or calculate different parameters for the bioreactor already in use. This section gives only a few program screens to provide an idea of the possibilities. When the program starts, the following window appears, and you can ﬁll in all the known parameters of a prospective bioreactor design (see Figure 9.15). After completing this, the drawing, “Bioreactor Conﬁguration,” shows how the bioreactor will look; the program stores 50 different bioreactor plans in its databank. Behind the power number Np is a question mark. Click it and the next window appears (see Figure 9.16). Here you can choose an impeller, and by clicking “OK” this value is transposed into the power number in the ﬁrst window, although any number may be used. Click the calculate button in the “calculation” block, and the next window appears with all the necessary calculations (see Figure 9.17).

PH2270_book.fm Page 322 Tuesday, May 16, 2006 9:37 AM

322

Producing Biomolecular Substances

FIGURE 9.16 Power number window.

FIGURE 9.17 Newtonian liquid calculations window.

If the Reynolds number is smaller then 10,000, the value is printed in red along with a note. In this window it is possible to either print and/or save the results. Selecting “Explanations” in the main window, one can chose “Drawing” or “Text.” Click on “Drawing” and this next window appears (see Figure 9.18). By pointing at one of the parameters, the explanation is shown in white. The next ﬁgures are to be used for another bioreactor with the appropriate calculations. To obtain this program or get more information, e-mail the author, P.T.E. van Santen. at: [emailprotected]. On this site, a demo version of this program (≈ 589 kb) can be downloaded or, if you want to buy a registered, fully functional copy, visit https://secure.shareit.com/shareit/checkout.html.

PH2270_book.fm Page 323 Tuesday, May 16, 2006 9:37 AM

Bioreactors

323

FIGURE 9.18 Bioreactor drawing window.

TABLE 9.12 Some Bioreactor Manufacturers and Products Company

Range of Sizes (L)

Applikon Astra Scientiﬁc Bellco Biotech B. Braun Biotech Chemap Cole Parmer Enprotech L–H Fermentation LSL Biolaﬁtte New Brunswick Schlegel Assoc. Sulzer Biotech Wheaton Sci.

1–1,200 2–100 1.5–50 1–14,000 2–500,000 1–15 2.5–90 1–1,000 1–65,000 1–10,000 Ind’l 2–500 1.5–100,000 0.5–45

Latest Product (L) 5 1.5 and 50 14,000 CMF Series (2–35) 1 L Round Bottom System MCT-25 (3 and 6) Series 2000 (12–100) Maestro™ Series (1–10) BioFlo™ IV (1–10) SpinFirm™ (3.5–1,000) Turbo-lift™ (19 and 45)

PH2270_book.fm Page 324 Tuesday, May 16, 2006 9:37 AM

324

Producing Biomolecular Substances

BIOREACTOR MANUFACTURERS’ DIRECTORY A/G Technology Corp. Applikon, Inc. Artisan Industries, Inc. Asgard Biotechnology Astra Scientiﬁc International, Inc. B. Braun Biotech International GmbH Baxter Diagnostics Scientiﬁc Products Bellco Glass, Inc. BioPro International, Inc. Bio-Recovery, Inc. Cellco, Inc. Cellex Biosciences, Inc. Chemap, Inc. Alfa-Laval Group Columbus Instruments International CPG, Inc. Curtin Matheson Scientiﬁc, Inc. Eppendorf-Netheler-Hinz GmbH FMC BioProducts Hoechst Celanese Separations Products Div. ICN Biochemicals, Inc. INTEGRA Biosciences, Inc. Jaeger Biotech Engineering, Inc. Kimble/Kontes Kinetek Systems, Inc. Lab-Line Instruments, Inc. LH Fermentation LSL Biolaﬁtte, Inc. Microgon, Inc. New Brunswick Scientiﬁc Co., Inc. Paul Mueller Co. Pope Scientiﬁc, Inc. Regis Technologies, Inc. SciLog, Inc. Sulzer Biotech Systems Techne (Cambridge) Ltd. Unisyn Technologies, Inc. Ventrex Laboratories, Inc. Verax Corp. VirTis Co. Wheaton, Inc. Wheaton Scientiﬁc

PH2270_book.fm Page 325 Tuesday, May 16, 2006 9:37 AM

Biomolecular Synthesizers DNA SYNTHESIZERS

DNA SYNTHESIS Approximately 60% of DNA synthesis is performed for PCR use with the remaining 40% performed for general sequencing, mutagenesis studies, gene construction, special antisense constructions, and for physical studies of certain species. There is also a growing use of DNA probes in the determination of bacterial and viral contaminants, genetic abnormalities, and for the study of gene receptor sites involved in the action of various drugs. Antisense applications are substantially increasing, especially with synthesizers capable of producing modified nucleic acids. In antisense work, research oligonucleotides are designed and synthesized to compete with specific DNA and subsequently injected into infected cells. This competition alters protein production, thereby interfering with the production of bacterial toxins and viral replication. Formation of the desired DNA chain begins with anchoring the first nucleotide base to a suitable solid support. Then selected purine-pyrimidine pairs (adenine, thymidine, guanine, and cytosine) are added in stages until the desired DNA structure is achieved, each cycle linking one nucleotide to the DNA chain. The base DNA sequence is its most characteristic property, since DNA’s genetic code is based upon nucleotide arrangement or sequence that determines the amino acid sequence of proteins found in the cell. Synthetic single-stranded DNA is used in many fundamental biotechnology operations. The DNA molecule is in the form of a double helix composed of two complementary DNA strands. All DNA molecules are composed of repeating deoxyribonucleotide units (composed of the sugar 2-deoxyribose, phosphate, and either a purine or a pyrimidine linked by a phosphate group joining the 3′,2′ position of one sugar to the 5′,2′ position of the next). Most native DNA molecules are doublestranded and antiparallel, resulting in a right-handed helix structure kept together by hydrogen bonds between a purine on one chain and a pyrimidine on the other. DNA is a carrier of genetic information that is encoded in the arrangement or sequence of the bases. This genetic code determines the amino acid sequence of proteins found in the cell. DNA is present in chromosomes and the chromosomal material of cell organelles (e.g., mitochondria) and is also present in viruses. Automated DNA synthesis involves the linking of the bases or constructing the genetic 325

PH2270_book.fm Page 326 Tuesday, May 16, 2006 9:37 AM

326

Producing Biomolecular Substances

TABLE 10.1 DNA Synthesizer Coupling Chemistries Synthesizer Coupling Chemistry Systems Phosphate and phosphite triesters (phospho-triester) Phosphoramidite methyl phosphoramidite (MED) cyanoethyl phosphoramidite (CED) Beta-cyanoethyl phosphoramidite (BED) Hydrogen-phosphonate (H-phosphonate)

Company

Manufacturers’ Coupling Chemistry Systems System

ABI Cruachem Milligen/Biosearch Pharmacia/LKB Biotech Vega Technologies

beta-cyanoethyl phosphoramidite, H-phosphonate and RNA cyanoethyl phosphoramidite cyanoethyl phosphoramidite, H-phosphonate beta-cyanoethyl phosphoramidite cyanoethyl phosphoramidite

code. The resultant synthetic single-stranded DNA is used in fundamental biotechnology research operations.

COUPLING CHEMISTRIES The first solid-phase DNA synthesizers introduced deoxyribonucleotides in either a phosphate or phosphite triester form, with each cycle linking one nucleotide base to the DNA chain. These original systems had unstable products and/or slow cycling times. A synthesis chemistry based on phosphoramidites and derived from phosphite triester coupling, was developed by Beaucage and Caruthers at the University of Colorado at Boulder. The phosphoramidite chemistry, patented by Applied Biosystems and Beckman, improved cycle times and the quality of the yield. In DNA synthesis, four chemical methods are now typically employed: phosphate triester, phosphoramidite, hydrogen-phosphonate (or H-phosphonate), and β-cyanoethyl phosphoramidite (BED). BED appears to be the most common method used in commercial DNA instruments (see Table 10.1). In this method, a derivatized nucleoside base is attached to a controlled-pore glass or polyamide resin solid support that is packed in the column. DNA synthesis occurs in the column and on the support material. Phosphoramidites are generally used as the monomer building blocks. A DNA synthesizer can have one to four columns where DNA fragments are synthesized simultaneously. The amount of amidite that is attached to the support typically ranges from 0.2 mmol to 15 mmol in concentration. The DNA is synthesized in five component steps per synthesis cycle: (1) detritylation, (2) activation, (3) coupling, (4) capping the unreacted chains, and (5) oxidation. A delivery system dispenses reagents into the columns on instructions from the system microprocessor-based controller. After cycle completion, the DNA

PH2270_book.fm Page 327 Tuesday, May 16, 2006 9:37 AM

Biomolecular Synthesizers

327

fragment is eluted from the solid support to a collection vial and subsequently purified using HPLC or electrophoresis. Synthetic oligonucleotides, the products of DNA synthesis, synthetic oligonucleotides, generally have five primary uses in biotechnology research: 1. 2. 3. 4. 5.

Hybridization probes for gene cloning. Links and adapters in rDNA methods. Primers in sequencing techniques. Site-specific base alterations in mutagenesis studies. Precursors to larger DNA fragments for entire gene synthesis.

Developing coupling chemistries for the added deoxyribonucleotides has been an important step in the development of automated DNA synthesis. Unreliable coupling chemistry methods were the principal reasons for dissatisfaction with some of the first units, and current manufacturers are continually working on improvements in this area. When carrying out any multistage chemical reaction, it is essential that the yield at each stage be maximized, otherwise very little product is recovered at the end of a sequence. These considerations are even more important in the case of DNA synthesis because it is typically a 20–30-stage process. Thus, in order to create a practical synthesizer, it was first necessary to develop chemistries that produced a high product yield at each coupling step. Also desirable was that the coupling reactions were capable of quickly reaching endpoint at room temperature.

ADVANTAGE

MULTIPLE CHEMISTRIES

Different chemical coupling schemes are utilized in DNA synthesizers, and manufacturers seem to have developed one or more preferred chemistries for their systems. In one of the first synthesizers, the nucleotides were utilized in the phosphate triester form, and another instrument, the phosphite triester form. At first, both systems had problems; the phosphate triester coupling was rather slow and required heating, and although the phosphite triester coupling was fast, the reagents tended to be unstable. Since their first introduction, however, substantial improvements have been made in both chemistries. Following introduction of the first phosphate and phosphite triester systems, an alternative system based upon phosphoramidites was introduced. The coupling was fast at room temperature and the yield was high, although the reagents were unstable when exposed to air and moisture. This could be overcome, however, by careful handling of the reagents. Unreacted chains could be capped by the addition of a chemical blocking group to prevent their reaction with the incoming nucleotide of the next sequence. Thus, the only chains that would continue to be built would be those with the correct sequence. A variation on the phosphoramidite chemistry [cyanoethyl phosphoramidite (CED)] was also developed, eliminating the need for toxic blocking agents at the end of the sequence. The final reaction product is a complex mixture of the desired DNA segments and all of the smaller chains that were capped at each stage of the process. Separation of the desired product is usually accomplished by liquid chromatography or gel

PH2270_book.fm Page 328 Tuesday, May 16, 2006 9:37 AM

328

Producing Biomolecular Substances

electrophoresis. After it has been determined that a particular chemistry is fast, stable, and efficient, the cost per coupling cycle is an important consideration. Since the reagents for these systems are high-value-added products, they provide a continuing profit source for the manufacturers and suppliers throughout the life of the instrument. The cost of the reagents consumed by a typical DNA synthesizer, for example, can be well in excess of US$20,000 per year.

DNA SYNTHESIZERS An automated DNA synthesizer is basically an instrument for the controlled, sequential dispensing of reagents into a reaction chamber. Scientists use DNA synthesizers for various PCR applications, antisense drug research, genetic probes, comparison studies dealing with mutations and disease, and all kinds of species-specific analysis from insects to humans.

AUTOMATED UNITS Automated DNA synthesizers predominantly produce primers for sequencing, gene amplification, and making hybridization probes for cloning, links and adapters for rDNA methods, and DNA and RNA for antisense drug research, and DNA for sitespecific mutagenesis studies. DNA synthesizers can also be used for genetic mapping when primers are used to analyze sequences in both directions from a site. RNA synthesizing capability is important for in situ hybridization studies. DNA synthesizer prices currently range from about US$12,000 for a semiautomated unit to around US$75,000 for a fully-loaded, automated, top-of-the-line system.

THROUGHPUT CAPABILITY DNA synthesizers can generally be differentiated by their throughput capability, which is based on the number of columns on the instrument. Large core facilities that supply custom nucleic acids to pharmaceutical development or university labs generally use high-throughput synthesizers with multiple columns. Smaller labs doing individual syntheses for research generally require only one- or two-column units and are discovering the cost advantages of having their own small-scale automated DNA synthesizers. Sometimes three or four labs will choose to work together to synthesize their own oligonucleotides rather than rely on a core facility. Many synthesizers provide extremely low cost per synthesis cycle. Milligen-Biosearch Division of Millipore offers a full range of DNA synthesizers to accommodate users’ varied needs — from a small one-column model made specifically for small users, to their model #8800, an automated batch reactor that synthesizes large quantities of DNA in a single run. The time required to produce DNA with an automated synthesizer has become shorter since the first synthesizer was introduced in 1979. For example, ABI’s newer units produce oligonucleotides in less than a day, whereas in 1982 the same syntheses usually required three to four days. DNA-sequencing chemistries are also evolving to meet the increasing need for speed, accuracy, and new capabilities. Pharmacia LKB’s DNA synthesizer, for example, provides linkers for nonradiographic primers through a range of modified ami-

PH2270_book.fm Page 329 Tuesday, May 16, 2006 9:37 AM

Biomolecular Synthesizers

329

dites. Fluorescein primers for sequencing can be labeled in one step and biotin primers can be made for solid-phase methods with specific linker amidites. Although DNA synthesizers are evaluated primarily on throughput, other factors such as chemistries and innovative labeling processes are emerging as important features.

SCALE UP DNA synthesizers are typically differentiated by their throughput capability, based on the number of columns on the instrument. Large core production facilities that supply custom nucleic acids to pharmaceutical development or university labs generally use high throughput synthesizers with multiple columns. Milligen-Biosearch’s Model #8800 is a batch-reactor that synthesizes large quantities at one time. The time required to produce DNA with large-scale automated synthesizers has been diminishing, and recombinant bacterial processes are well-suited for producing large quantities of DNA because, after the initial investment, production output is high and operating cost is low. Bacteria is typically used for production yields of recombinant DNA. Yeast and mammalian systems also express DNA, but with much lower yields than bacterial systems, and therefore they are not as yet cost-effective for production scale up.

MANUFACTURERS’ DIRECTORY DNA SYNTHESIZERS Applied Biosystems (800) 345-5ABI Cruachem, Inc. (800) EASY DNA Milligen/Biosearch (Millipore) (800) 225–1380 Omnifit USA Corp. (800) 864–0135 Pharmacia Biotech, Inc. (800) 526-3593 Vega Biotechnologies (800) 528-4882

DNA SYNTHESIZER ACCESSORIES

AND

REAGENTS

Applied Biosystems (800) 345-5ABI BioGenex (800) 421-4149 Burdick and Jackson (800) 368-0050 Clontech Labs., Inc. (800) 662-2566 CPG, Inc. (800) 362-2740 Cruachem, Inc. (800) EASY DNA Dynal Inc. (800) 638-9416 Eppendorf North America (800) 421-9988 Glen Research Corp. (800) 327-4536 Millipore Corp. (800) 225-1380 Peninsula Laboratories Inc. (800) 922-1516 Pharmacia Biotech, Inc. (800) 526-3593 SynChem, Inc. (800) 882-9267

PH2270_book.fm Page 330 Tuesday, May 16, 2006 9:37 AM

330

Producing Biomolecular Substances

TABLE 10.2 Peptides Utilized in Biotechnology Biologically active peptides Other peptides and enzyme substrates Synthetic enzyme inhibitors Synthetic chromogenic enzyme substrates α-nitroanilides [αNA] fl-naphthylamides [flNA] 4-methoxy-fl-naphthylamides [4MflNA] 7-amido-4-methylcoumarin [AMC] 7-amido-4-trifluoromethyl coumarins [AFC] 3-(2-furylacryloyl)-substrates [FA-substrates] Other chromogenic and fluorogenic enzyme substrates Standard chromophoric compounds

PEPTIDE TECHNOLOGY OVERVIEW Over 50% of the dry weight of most organisms is made up of proteins that are macromolecular and composed of amino acid molecules linked together by peptide bonds (so called because the digestive enzyme pepsin breaks such linkages). Amino acids and their polymers are used for studying protein structure, antigenicity, and for a variety of other applications (see Table 10.2). Two or more linked peptides form polymeric structures called polypeptides; proteins are linked polypeptides. In vivo, many peptides also occur in free form, and are not necessarily associated with proteins; many of these free peptides have intense biological activity (see Tables 10.2 and 10.3). A peptide may occur naturally (e.g., many hormones and other biologically active substances are peptides), or may be a fragment of a larger natural peptide, polypeptide, or protein molecule (as are many subunit vaccines). It may also be a synthetic reproduction of a natural peptide (e.g., biopharmaceutical preparations of atrial-natriuretic factor), or it may be a synthetic construct with no exact counterpart in nature (e.g., many peptide analogs tested for improved stability, potency, or safety over original peptide molecules). The amino acids common in proteins combine to form different peptides consisting of one amino acid residue. Dipeptides consist of two amino acid residues, with 400 possible dipeptide combinations. The number of polypeptide combinations with three or more peptide chains (e.g., pentameric, hexameric, duodecameric peptides, etc.) is formidable. Certain proteins also contain special amino acid derivatives of a standard type. A chain of identical amino acid residues — all glycine, for example, is called a hom*opolymer. A chain of different amino acids in a precisely known order is called a sequential polymer. If there is no particular order in a chain of known amino acids, then the peptide is a random polymer; peptide chains that are internally cross-linked form multichain polymers. The various combinations of amino acids determine the bioactivity of a protein or a free polypeptide (i.e., whether it is an enzyme, hormone, or antibody). Some examples of polypeptide hormones are insulin with two peptide chains of thirty and twenty-one amino acid residues, respectively, secreted by B

PH2270_book.fm Page 331 Tuesday, May 16, 2006 9:37 AM

Biomolecular Synthesizers

331

Table 10.3 Biologically Active Peptides AdipokinetichHormones Collagen-derived octapeptide Adjuvant peptides Corticotropin-releasing factors [CRF] Adrenocorticotropic hormones [ACTH] Delta sleep-inducing peptide [DSIP] Alfa bag cell peptides Dermorphins Alfa mating factors Diazepam binding inhibitors and related peptides Allatostatins Diprotins Allatotropins Dynorphins Amyloid plaque core-A4 polypeptide Endorphins Angiotensin II receptor binding protein Endothelins and related peptides Angiotensins and related peptides Enkephalins and pro-enkephalin-related peptides Apolipoprotein-E amide Epidermal growth factor receptor peptides Aspartame Epidermal mitosis-inhibiting pentapeptide Galanin Calmodulin binding domain Gastric inhibitory peptides [GIP] Calpeptin [Z-Leu-Nle-H] Gastrin-releasing peptides [GRP] Cardioactive peptides Gastrins Casomorphins Glucagons and related peptides Cerebellins Gonadotropin-releasing hormones [GnRH, LHRH] CCK antagonists Granuliberin-R Cecropins Oxytocin and related peptides Chemotactic peptides Pancreastatin Cholecystokinin-pancreocymin sequences

Atrial natriuretic peptides [ANF, ANP] Experimental allergic encephalitogenic peptides [EAE] Bombesin and related peptides Fibrinogen fragments and related peptides Bombesin antagonists Fibrinolysis-Inhibiting factors Serum thymic factor and related peptides Leumorphins Splenopentin Lipocortin I Somatostatin and related peptides Liver cell-growth factor Bradykinin antagonists Fibronectin-fragments and related peptides Bradykinin-potentiating peptides Fibronectin function inhibitors Buccalin FMRF amide and related peptides Bursin and related peptides Follicular gonadotropin-releasing peptide Calcitonins Formyl-peptides Calcitonin gene-related peptides [CGRP] Kemptide Protein kinase-related peptides Kinetensin PYY Kinins [Bradykinins] Renin substrates and inhibitors/ACE inhibitors Laminin peptides Sarafotoxin Leuco*kinins Sarcophagine Leucomyosuppressin Secretins Leucopyrokinin Steroidogenesis activator polypeptide Lymphocyte activating pentapeptide Substance-P and antagonists Magainins Tachykinins Melanin-concentrating hormone Continued

PH2270_book.fm Page 332 Tuesday, May 16, 2006 9:37 AM

332

Producing Biomolecular Substances

Table 10.3 (Continued) Biologically Active Peptides Growth hormone-releasing factors [GRF] and related Parathyroid hormones [PTH] Helodermin and related peptides Peptide antibiotics Hemorphins Peptide-T and related reptides Herpesvirus ribonucleotide reductase inhibitors PHI, PHM Human seminal plasma inhibin Pressinoic acid and related peptides Insulin B Proctolin Interleukin sequences and antagonists Proinsulin C-peptide (human) Neurotensin and related peptides

Melanotropin potentiating factor TRH-analog Metorphamide Tryptophyllin Motilins Urodilatin MSH and MSH-release-inhibiting factor Valosin Myomodulin Vasoactive intestinal peptides [VIP] and antagonists Neoendorphins Vasopressins Neuropeptide Y [NPY] Vasotocins

cells of the pancreas; corticotropin with thirty-nine amino acid residues, secreted by the anterior pituitary gland; and oxytocin with nine residues, secreted by the posterior pituitary gland. Many extremely potent toxins (e.g., mushroom poisons) are polypeptides, as are many antibiotics and synthetic vaccines. The twenty natural amino acids are each structurally unique. Most naturally occurring peptides consist of L-amino acids, and D-amino acids are often used in synthetic constructs (an exception is S. aureus peptidoglycan that naturally contains D-Glu and D-Ala). The structural variations between amino acids and between the L- and D-forms determine differences in biochemical activity that must be taken into account in designing synthesis protocols. Chemically, amino acids have much in common, and their differences are sufficiently subtle for the art of peptide sequencing and synthesis to have evolved into a specialized and complex branch of biochemistry (Table 10.4). It was not until the 1950s that protein chemistry had advanced to the point where a peptide could be accurately sequenced and synthesized. The active development of peptide therapeutic agents by biotechnology and biopharmaceutical companies has coincided with a considerable increase in the market for highly purified peptides. Similarly, there is a growing interest in certain purified proteins, coinciding with the development of recombinant biologically active proteins.

PEPTIDES BY SOLID-PHASE SYNTHESIS PEPTIDE SYNTHESIS Advances in peptide chemistry, genetic engineering and other biotech-related techniques have resulted in a sharp rise in new peptide applications. Peptide linkages

PH2270_book.fm Page 333 Tuesday, May 16, 2006 9:37 AM

Biomolecular Synthesizers

333

TABLE 10.4 Symbols and Abbreviations: Amino Acids and Derivatives Alanine β-Alanine Alloisoleucine 2-Aminoadipic acid 3-Aminoadipic acid 2-Aminobutyric acid 4-Aminobutyric acid 6-Aminohexanoic acid 3-Aminopropionic acid Arginine Asparagine Aspartic acid Caproyl 4-Carboxyglutamic acid Cyclohexylalanine Cysteine Cystine 2 2,4-Diaminobutyric acid 2,2-Diaminopimelic acid Glutamic acid Glutamine Glycine Histidine hom*ocitrulline hom*ocysteine hom*oserine Hydroxylysine Hydroxyproline Isoleucine Lauroyl Leucine Lysine Methionine Norleucine Norvaline Ornithine Palmitoyl Phenylalanine Phenylglycine Proline Pyroglutamic acid Sarcosine Serine Statin Stearoyl Thienylalanine Threonine Tryptophan Tyrosine Valine

Ala A β-Ala alle Aad β-Aad Abu t-Abu e-Ahx β-Ala Arg R Asn N Asp D Cap Gla Cha Cys C (Cys) A2bu A2pm Glu E Gln Q Gly G His H Hci Hcy Hse Hyl Hyp Ile I Lau Leu L Lys K Met M Nle Nva Orn Pal Phe F Phg Pro P Pyr Sar Ser S Sta Ste Thi Thr T Trp W Tyr Y Val V

PH2270_book.fm Page 334 Tuesday, May 16, 2006 9:37 AM

334

Producing Biomolecular Substances

occur by eliminating a water molecule through the direct condensation of the alphaamino group of one amino acid with the alpha-carboxyl group of the another amino acid. Peptide bonds can be formed chemically in solution, on a solid support, and biologically using the transcription-translation mechanism of a cell or cellular enzyme. Efforts to utilize peptide hormones made by chemical synthesis date back to the mid-1950s when the first peptide biologic, oxytocin, was developed by Sandoz. By 1977, 15 commercial peptide biodrugs had been introduced, including corticotropin, thyrotropin-releasing hormone, and calcitonin.

SOLID-PHASE PEPTIDE SYNTHESIS Solid-phase peptide synthesis (SPPS) was first introduced by Merrifield in 1962. This original protocol was used in manual peptide synthesis and later incorporated into commercial instruments introduced in the early 1980s, and has remained relatively unchanged as applied by current commercial instrumentation. Whether manual or automated methods are employed, amino acids must have reactive functional groups protected in order to join the proper amino group of one amino acid with the carboxyl group of another to form the amide bond between the two amino acids. Each amino acid has at least one carboxyl and one amino group. Some amino acids have these functional groups in side chains as well. During peptide synthesis, unwanted byproducts occur if reactions are not selectively limited to the desired functional group. Synthesizing peptides starts with the carboxyl terminus, or C-terminus, amino acid. The second amino acid that is added must have the amine component of its molecule, or the N-terminus, protected or blocked. This allows the proper formation of the amide bond. Subsequent amino acid additions require the N-terminus to be blocked, usually using tertbutyloxycarbonyl (tBoc) to protect the amine group. Material costs have traditionally been higher in solid-phase synthesis due to the necessity of using large quantities of solvents to wash the resins between synthesis steps and the large excess of amino acid derivatives required to achieve high yield at each coupling stage. The recent introduction of very high titer resins can lower solvent usage costs dramatically and current coupling methods using various additives have, in several instances, allowed reduction of the amount of excess amino acid required for coupling completion. SPPS uses an insoluble polymer for support of organic synthesis with the C-terminal residue of the peptide attached to the polymer, and grows the peptide chain toward the amino end of the peptide. After the desired sequence has been assembled on the support, the chain is cleaved and the finished peptide released into the supernatant. The great advantage of using a polymer-supported peptide chain is speed, due to the elimination of laborious purifications at intermediate stages in the system. It is now considered standard for SPPS to use tBoc and a 1% cross-linked polystyrene resin as an alpha-amino protecting group. Side-chain functional groups are mostly masked by benzyl-derived blocking groups that are cleaved by strong acids. In the tBoc procedure, the selectivity of cleavage of the alpha-protecting group, side-chain blocking groups, and the peptide-resin link depends upon differences in the rate of acidolytic cleavage. Orthogonal systems, such as 9-fluorenylmethyloxycarbonyl (Fmoc), where the side-chain blocking groups and the peptide-resin

PH2270_book.fm Page 335 Tuesday, May 16, 2006 9:37 AM

Biomolecular Synthesizers

335

link are completely stable to the reagent used to remove the alpha-protecting group, have been introduced. The alpha-protecting Fmoc group is removed by base, whereas the side-chain protections and the peptide-resin link are cleaved by mild acid. Major advances for Fmoc SPPS include the introduction of the trityl-protected derivatives of asparagine and glutamine and the pentamethyl chroman-sulfonyl (Pmc) protecting group for arginine. They can be easily removed by TFA and facilitate synthesis as well as de-protection of peptides connecting these amino acids. Using more than one chemistry allows us to take advantage of their different properties. An example of dual-chemistry-mixed-strategy is synthesis of branched peptides. Starting with an Fmoc-Lys (tBoc) amino acid derivative, tBoc strategy is used during amino-acid couplings of the first chain, and Fmoc for the second chain. Peptide synthesis can also be achieved by a combination of solution- and solid-phase synthesis, convergent peptide synthesis. Protected peptide segments built by solid phase are cleaved, purified, analyzed and further used for fragment condensation, either in solution or onto a peptide chain attached to a resin. Standard SPPS results in a crude product containing deletion peptides that are difficult to remove unequivocally by normal purification procedures. In combination synthesis, suitably blocked peptides are synthesized separately and then incorporated into the desired sequence. The deletion sequences will therefore differ from the desired sequence by more than a single amino acid residue, and their removal will be simpler. Several resins have been described that yield protected peptides upon cleavage. Kaiser et al., describe an oxime support that can be used in tBoc SPPS. Sheppard and Williams synthesized an acid-labile anchor, and Mergler et al., developed Sasrin resin, which is of higher loading and easier handling than the Sheppard resin. Condensation of protected fragments made by SPPS combines the ease and speed of solid phase with the purification advantage of solution phase.

SOLID-PHASE PEPTIDE SYNTHESIZERS Peptide synthesizers were developed before DNA synthesizers, but have only lately been adapted to use in genetic engineering. Today most companies marketing DNA synthesizers also market peptide synthesizers or plan to do so in the near future. Peptides are of greater use than DNA in research, health care, and biotechnology products, but are more complicated to produce. The basic building blocks of peptides are 20 amino acids, each having individualized coupling characteristics. Early instruments produced peptides 17–20 bases long. Today’s instruments can produce peptides of over 50 bases and advances continue. Generally, one can produce peptides faster than they can be purified. For example, a 10-amino-acid peptide can be synthesized in about 30 hours, but purifying it could take weeks. Both HPLC and gel electrophoresis are used in peptide purification. Beckman produced the first commercial peptide synthesizer in 1969. The next generation of instrument was introduced in 1980 by Vega Biotechnologies. Computers were, and continue to be, integral to these systems. Manufacturers interface their peptide synthesizers with PCs and specially designed software that can direct synthesis of a particular peptide from various amino acid combinations. Extensive data bases have been developed for peptide synthesis and the PC provides easy access via CD-ROM.

PH2270_book.fm Page 336 Tuesday, May 16, 2006 9:37 AM

336

Producing Biomolecular Substances

ABI, Advanced Chemtech, Milligen/Biosearch, Protein Technologies, and Pharmacia/LKB Biotechnology market automated peptide synthesizers; some manufacturers such as DuPont market semiautomated units with reagents added by the user. The trend in biotechnology, however, is toward fully-automated systems, although one large custom peptide manufacturer claims to produce all their custom peptides on semiautomated systems at a higher profit than they claim they could realize on automated units! In automated peptide synthesis, the C-terminus of an amino acid is embedded in a solid support to protect it. Then, amino acids with blocked amine terminii are added one at a time. The various reagents required for peptide synthesis are added sequentially and removed by filtration and washing. Once synthesis is complete, the peptide is removed from the resin, isolated, and purified by HPLC. In the Merrifield protocol, the solid support attached to the C-terminus amino acid is polystyrene. The step-by-step addition of each amino acid occurs, sequentially, as follows: 1. Cleavage of the amino protecting group using trifluoroacetic acid (TFA). 2. Neutralization using triethylamine dichloromethane (IDCM). 3. Acylation/amino acid addition usually using t-butyloxycarbonyl (tBoc) amino acids. There are certain limitations addressed in modified techniques including the use of harsh acids for both de-protection of the amino acid and removal of the finished peptide fragment from the resin. The Medical Research Council (Cambridge, England) developed a better method for peptide synthesis that eliminated t-Boc solidphase synthesis disadvantages. In this technique a polyamide gel, rather than polystyrene, supports the C-terminus amino acid rather than polystyrene, and the protocol uses fluorenyl-methoxycarbonyl (Fmoc) protected amino acids instead of tBoc. The Fmoc amino acids do not require repeated acid washings for de-protection. Some commercially available peptide synthesizers have the ability to operate with either the tBoc or Fmoc protocols (Table 10.5). The goal of peptide synthesis is to have high-yield production of high-purity peptides, and the coupling reaction is critical for the purity and percent yield. Less than 100% success of amino acid coupling will produce a peptide of varying purity. For instance, a 55-residue (amino acid) peptide synthesized at a 99.9% coupling yield will actually result in only 94.6% of the total peptide content of the final product (the 55-mer peptide). Thus, the final product also consists of 5.1% of a 54mer peptide. If the same 55-mer peptide is synthesized at 98% coupling yield, then the final peptide consists of only 33.6% of the 55-mer peptide and an even greater percent (37%) of the 54-mer peptide. Therefore, a decrease in the production yield of less than 2% will result in the final content being 63% contaminated by peptides other than the desired product. Most commercial peptide synthesizer coupling efficiencies of commercial peptide synthesizers are >99.5%.

SOLUTION SYNTHESIS Classical solution synthesis routes are similar in methodology and equipment to typical pharmaceutical organic synthesis. Depending on the molecule’s complexity,

PH2270_book.fm Page 337 Tuesday, May 16, 2006 9:37 AM

Biomolecular Synthesizers

337

TABLE 10.5 Peptide Synthesizer Coupling Chemistries Peptide Coupling Chemistries System tBoc (tertiary-butyloxycarbonyl) Fmoc (9-fluorenyl-methoxycarbonyl) H-phosphonate DDZ (2,3,5, dimethoxyphenylpropyl-2,oxycarbonyl) Fmoc-PEP (pentafluorophenyl) (active ester) Fmoc-DHBT (3-hydroxy-4-oxo-1,2,3-benzotriazine (active ester) Fmoc or t-Boc w/NMP/HOBt (N-methylpyrrolidone/1-hydroxy-benzotriazole) Fmoc-BOP (Castro’s reagent)/HOBt Fmoc w/DIPCDI (Diisopropylcarbodiimide)

Company

Manufacturers’ Coupling Systems System

ABI Advanced Chemtech Applied Protein Technol DuPont Labortec AG (Swiss) Bachem, (U.S. distr.) Milligen/Biosearch: Pharmacia/LKB Biotech Vega Industries

Fmoc, t-Boc, NMP/HOBt t-Boc, Fmoc, DDZ t-Boc, Fmoc manual system t-Boc, Fmoc Fmoc-HOBt, Fmoc/BOP/HOBt, Fmoc w/DIPCDI, t-Boc or Fmoc active esters Fmoc active esters/HOBt t-Boc, Fmoc

solution synthesis can be approached step-wise, or by fragment condensations. Current economics favors the preparation of three to eight residue small molecules by solution synthesis. The major advantages of syntheses performed in solution are a plethora of coupling methods, a wide variety of protecting groups, opportunities for intermediate purification, and linear scale-up potential. In solution synthesis, the many purification steps lead to long production times and costs. Attempts at performing solution syntheses sequentially with minimal purification between steps usually results in impurity profiles as similarly seen in solid-phase synthesis — obviating the advantages of solution synthesis over solid-phase synthesis in obtaining easily purified material at the final step of the synthesis. Solution-phase synthesis is generally useful for producing large amounts of short-chain peptides. There is no question whether 5kg of a five-amino-acid peptide or 1 ton of aspartame will be prepared by solution phase. Peptides synthesized in solution range in size from 2–70 amino acids, and in amounts from milligrams to tons. The largest current synthetic production of any peptide is that of aspartame (L-aspartyl-L-phenylalanine methyl ester), trade-named NutraSweet™, with an annual output of approximately 4,000 tons. Merck Sharp and Dohme has developed scale-up synthesis of several thyrotropin-releasing hormone analogs in 10–100kg quantities. Hoechst produces peptide hormones, such as oxytocin, IHRH and its analog Buserelin, TRH, and ACTH in

PH2270_book.fm Page 338 Tuesday, May 16, 2006 9:37 AM

338

Producing Biomolecular Substances

kilogram quantities. Hoffmann-La Roche scaled up solution synthesis of thymosin alpha-1, and Bachem (Switzerland) produces kilogram amounts of peptides ranging in size from 3–32-mer.

SEQUENTIAL PEPTIDE SYNTHESIS In the sequential synthesis method, the peptide chain is elongated one amino acid at a time from the carboxyl terminus. Ideally, at each step in this type of synthesis the intermediate products are purified and characterized. This is the method of synthesis best understood by classical organic chemists. Typical problems using sequential solution synthesis methods include poor solubility of intermediates, physical handling problems, filtration difficulties, occasionally increased racemization potential, handling losses, high cost of many purifications, and long development times. The advantages include definite intermediate characterization, well-understood reaction optimization methods, and relatively small organic solvent requirements.

FRAGMENT CONDENSATION SYNTHESIS When the molecules approach molecular weights of 1,500, the solution fragment condensation method starts to have advantages over the sequential synthesis method. Each fragment is usually synthesized by the sequential solution method and purified prior to condensation. The advantages of performing a fragment condensation synthesis over sequential addition synthesis include parallel synthesis of intermediates (shortening overall development time), higher solubility of the smaller fragments over the nearly-completed sequentially synthesized peptide, and ease of purification of products from intermediate fragments. Because of the expense of both components in the fragment coupling, they are usually reacted in equimolar amounts. Often, this results in incomplete reaction and lower yields than in sequential synthesis, where the amino acid to be added — much less costly than the growing peptide chain — can be added in large excess.

RECOMBINANT PEPTIDE SYNTHESIS Systems in use today for expression of peptides and proteins are bacteria, yeast, insect cells, mammalian tissue culture cells, and whole transgenic animals. In the most common recombinant method, millions of peptides are expressed on the surfaces of a bacteriophage. They are selected and then amplified in E. coli. Methods for generating peptides in phage follow the same strategy of cloning randomly synthesized oligonucleotides into the Gene III region of a filamentous bacteriophage (e.g., M13). Because the phage expresses the adsorption protein coded for by the Gene III region on its surface, the peptides are expressed along with the protein as part of its N-terminus.

SCREENING

FOR

PEPTIDES

WITH

MONOCLONAL ANTIBODIES

Peptides are thus readily accessible to screening with mABs and to subsequent affinity purification. Selected peptides can then be amplified by phage expansion in

PH2270_book.fm Page 339 Tuesday, May 16, 2006 9:37 AM

Biomolecular Synthesizers

339

E. coli. For example, a library of 4 × 107 distinct hexapeptide epitopes, expressed on the surface of a phage, was screened for binding to two mABs against a known epitope of the expressed protein. Phages that display reactive peptides are selected by binding to a biotinylated antibody, and then plated on streptavidin-coated dishes. Following washing to remove unbound phages, bound phages are removed from the dish, used to infect E. coli, and thereby amplified. Following successive repetition of this technique, peptides binding both antibodies could be identified, all of which had sequence similarities to the epitope. These peptides can be generated even though no specific information about the sequence was introduced into the epitope library. Sometimes, despite the many peptides generated, a peptide duplicating the exact native epitope is not found because the library was too small to represent the full repertoire of possible peptides or that expression of certain sequences does not occur in bacteriophage systems. The phage method is particularly useful for epitope mapping, i.e., identifying the parts of proteins specifically recognized by mABs. Foreign peptides can also be expressed on the Gene VIII coat protein of the phage surface in thousands of copies, and phages coated with foreign peptides incorporated into the Gene VIII protein make excellent antigens. Protein expression on phage surfaces are also used to make phage antibody libraries as alternatives to mABs. Phage antibodies are phages that display antigen-binding domains of antibodies on either Gene II or Gene VIII coat protein, and they offer the prospect of obtaining mABs without animals or animal cells. Instead, any antigen of interest could be used to make an affinity selection from a phage-antibody library (hopefully commercially available) of those phages that display antigen-binding antibody. In a similar strategy, a phage-generated peptide library of 3 × 108 recombinant phages was screened, bearing hexa-mAB against norpeptide sequences for binding a portion of Beta endorphin, and clones with binding activity were identified. Repetitive peptides containing common terminal residues did not resemble any known ligands for the antibody, although the peptide-bound antibody with relatively low affinity compares to known high-affinity ligands, though this screening method does not discriminate between low- and high-affinity peptides or select nonbinding peptides. To augment selection of high-affinity peptides, one can first select low-affinity peptides by using relatively high concentrations of antibody, and subsequent selections can be performed with lower concentrations, so that only high-affinity peptides will bind antibody. This system is particularly useful in identifying and isolating cytokine receptor antagonists.

PHAGE-GENERATED PEPTIDE LIBRARIES There are several limitations or flaws in the phage-generated library techniques for drug discovery. Biological systems inherently select certain peptide sequences, restrict placement of critical amino acids, and not incorporate other sequences. Also, a significant limitation of this technique for drug selection is its inability to incorporate D-amino acids, novel amino acids, or to allow incorporation of secondary configurations into peptides using cyclic amino acids. D-amino acids could be used to prolong the half-life of peptide-based therapeutics, reduce potential side effects,

PH2270_book.fm Page 340 Tuesday, May 16, 2006 9:37 AM

340

Producing Biomolecular Substances

enhance biological activity, increase solubility, and hasten secretion. Also, phagebased synthesis does not lend itself to assays carried out in solution.

SPLIT PEPTIDE SYNTHESIS

ON A

BEAD LIBRARY

Another system creates a peptide library on millions of beads, each bead containing a single peptide. In this split-synthesis approach, standard solid-phase peptide synthesis with Fmoc or Boc chemistry is used to make pentapeptides directly on the beads. The split-synthesis approach is used by reason of the differing coupling rates for each amino acid. In the first step, a pool of beads is distributed into separate reaction vessels containing a single amino acid. Following the reaction, the beads are re-pooled, and the cycle is then repeated until a peptide chain of desired length is formed on each bead. The beads are then screened for binding using either alkaline phosphatase or fluorescein-conjugated ligands. Labeled beads are manually selected using a micromanipulator and forceps; after this the peptide is analyzed on a glass filter in an automated sequencer. With this technique, 19 amino acids can be randomly incorporated into pentapeptides — which can theoretically yield 19 individual peptides of varying sequences, with any one sequence being present on at least one bead; and in screening for peptide binding to a high-affinity monoclonal antibody, six reactive beads were recovered from a total of two million — one binding with an affinity nearly identical to that of the native epitope.

CLEAVABLE LINKERS Cleavable linkers have also been incorporated onto each bead so peptides can be removed from the beads into solution for assay at any stage during synthesis. The ability to release peptides into solution permits testing for biological activity, and not just for binding. Since Selectide Corp. (Tucson, AZ) first introduced this technology, they have introduced conformational restrictions and have incorporated novel amino acids into peptides, permitting the rapid identification of initial targets from which workers can further develop peptide therapeutics. Houghton Pharmaceuticals generated a Synthetic Peptide Combinatorial Library (SPCL) of hexapeptides in free solution, with the first two positions in each peptide individually and specifically designed, and the last four positions consisting of equimolar mixtures of L-amino acids. Each of the 324 distinct peptide mixtures were assayed using competitive ELISA. The synthesis strategy employed methyl benzhydrylamine polystyrene resin in porous polypropylene packets, the so-called teabag method, and tBoc chemistry combined with the multiple-peptide synthesis technique to develop an interactive process by which, following the addition of successive amino acids in specific positions, growing peptides could be screened and deleted from the library if they did not bind a given ligand. In this manner, both the amino acid requisite for binding and its location in the peptide sequence were selected, and the number of peptide sequences requiring screening reduced 20 times with each amino acid addition cycle. With this method, the ability was determined of each of the original 324 dipeptide mixtures to inhibit the binding of a mAB to a larger, 13-residue peptide. And, by the time the 5th positions of all the peptide sequences were defined, the concentration

PH2270_book.fm Page 341 Tuesday, May 16, 2006 9:37 AM

Biomolecular Synthesizers

341

of the peptides required to inhibit binding of the monoclonal by the 13-mer peptide was 3,500-times lower than the required concentration of the preceding peptide mixture. Among the 20 peptides in which the sixth position was defined, they identified a sequence matching the original antigenic determinant recognized by the monoclonal. The method is extremely flexible, since free peptides generated in solution can be used in all types of assay systems.

ENZYMATIC SYNTHESIS Carboxypeptidase Y (a catabolic enzyme that degrades peptides or proteins from the carboxy-terminus) was the first enzyme to be used in enzymatic synthesis. Under appropriate conditions, many catabolic enzymes can be forced to reverse their activity, take on a synthetic function, and catalyze the step-wise synthesis or the fragment condensation of peptides.

CELL-FREE TRANSLATION SYSTEMS Cell-free translation systems provide a new method for the commercial production of peptides such as hormones, growth factors, lymphokines, and a host of other powerful agents with valuable medical applications. Cell-free systems are mixtures of everything needed to synthesize peptides and polypeptides except the intact cell. Up to now, these systems have been confined to bench research mostly in tracking gene expression in the cell. The limitation stems from their inability to run for more than a couple of hours and so their productivity is low. Researchers have discovered a way of running cell-free systems ten times longer and producing polypeptides at more than 100 times the previous record. It is speculated that these systems may soon assume a role in the commercial production of peptides. The key process development is a continuous-flow buffer system containing the peptide building blocks with mRNA, ATP, and GTP. An Amicon microultrafiltration system feeds the buffer through the mixture, allowing the newly synthesized polypeptides to be continuously harvested. The system employs two different cellular extracts routinely used in cell-free systems to supply the factors needed for synthesis — one from e. coli and the other from the wheat embryo. Both extract systems run continuously for 20 hours with a yield of 100 phage or viral-coat protein copies per molecule of mRNA. Production was maintained in the wheat embryo extract system for 40 hours, manufacturing about 300 copies of calcitonin per mRNA. The mRNAs in the system escape the enzymatic degradation that usually brings peptide synthesis to a halt. This apparent protection of the mRNAs comes from their constant engagement with the translation process because of the continuous-flow system. A problem with other cell-free systems that is lacking in this system is the loss of small, crucial protein-initiation factors through pores in the filtration membrane. Again, the continuous involvement of these proteins creates dynamic multiprotein units incapable of passing through the membrane’s pores. The system is employed to produce unstable, low-yield proteins that might be translated in sufficient quantity for use in functional assays. The method is expected to increase utilization of cellfree methods, although there are still reservations about using wheat embryos, since

PH2270_book.fm Page 342 Tuesday, May 16, 2006 9:37 AM

342

Producing Biomolecular Substances

many researchers believe the reticulocyte system gives the best mRNA translation when the desired gene product is known. One major advantage of the cell-free system is its ability to synthesize toxic protein. Cytotoxic proteins, such as those that bind to DNA, are produced either in low quantities or transiently in live cells, and increased production is cytotoxic. Cytotoxic protein can be produced in quantity in cell-free systems. In contrast, wheat embryo extracts do not promote many critical polypeptide modifications after translation (such as protein folding and glycosylation), although it may be possible to supplement the embryo system with enzymes to overcome this disadvantage; the system still needs more work, and an alternate natural system should be devised to do the job since ingredients in this system are too esoteric and costly. Beyond that, the possibilities are virtually limitless. Theoretically, anything that can be done with recombinant DNA and organic synthesis can be done in a cell-free system. For instance, instead of taking 18 months producing recombinant DNA, a cell-free system might do the job in 6–12 months. It is also a much simpler system and potentially quite competitive. A window of opportunity exists even with the problem of posttranslational modification since unfolded molecules could simplify protein analog production by sustaining the incorporation of unnatural amino acids. These analogs are generally used to block the action of natural polypeptide hormones. Although this cell-free system may expand the applications range of cell-free processes, the system might also be useful for screening peptides like tPA. When searching for active peptides on a protein molecule, it is difficult to express them in mammalian cells, so a cell-free system might be appropriate for proteins with exceptional potency, such as lymphokines or proteins with very low mRNA. Cell-free systems have inherent drawbacks. A current goal in mammalian cellculture is a serum-free medium that produces a fairly clean intermediate from with which to purify the protein of interest. Cell-free translation, however, typically uses lysed cells that adds lots of debris to the intermediate — almost a step backwards. Still, the ratio of desired protein to impurity might be so large that it could make purification easier.

PEPTIDE SYNTHESIS EQUIPMENT Commercialization of the equipment needed for large-scale, cost-effective, reliable synthesis is still in early development, although some extremely sophisticated custom equipment is now available. Equipment used in upscaling peptide synthesis ranges from standard chemical process equipment for solution-phase syntheses to highly automated equipment for solid-phase synthesis. Solid-phase methods depend upon high reaction efficiencies, which require highly specialized equipment, suitable for handling sticky resin slurries that must be fully whetted at every step of the process and must be composed of materials that can withstand the effects of trifluoroacetic acid and other corrosive solvents. Such equipment must also have high reliability and fault tolerance, since one error in a long synthesis could result in the loss of hundreds of thousands of dollars worth of product. Since one of the major benefits of solid-phase synthesis is reduction of labor through process automation, the equipment should also employ modern process-control technologies. Solid-phase synthe-

PH2270_book.fm Page 343 Tuesday, May 16, 2006 9:37 AM

Biomolecular Synthesizers

343

sis is inherently a multiloop batch process where, each step must be optimized. Although many laboratory-scale peptide synthesizer control systems seem to meet these requirements, none are suitable for plant and process optimization at pilot scale. There are no commercial pilot-scale solid-phase synthesizers that meet control requirements, follow GMP instrument guidelines, are designed with appropriate regard for plumbing reliability and reagent delivery assurance, and that can be used in a pilot plant GMP operations. Currently, large-scale solid-phase syntheses are either done manually, which lessens process reliability, or performed on custom equipment designed by experienced engineers and chemists familiar with the many subtleties of scaling up solid-phase syntheses. And, in order to develop an automated process-scale solid-phase synthesizer, expertise in mechanical and chemical engineering, electronics, and programming software for process control must be combined with an in-depth understanding of solid-phase peptide synthesis and process automation. Simply enlarging bench-scale units will not do.

PROCESS-SCALE PEPTIDE SYNTHESIZER An important part of designing a process-scale synthesizer is the development and optimization of a plumbing system that meets bioprocess requirements and avoids the pitfalls of large-scale solid-phase synthesis. Cross-contamination, internal volume of valves and tubing, and housekeeping must engender appropriate regard for plumbing reliability and reagent delivery assurance at required flow rates. Errors in system design can result in costly adjustments and can jeopardize entire production batches, since mechanical problems have frequently plagued systems in the past. Currently, there are manufacturers that understand the requirements of biomass equipment and produce high-quality components necessary for project success. In addition, practical considerations such as mounting numerous components must also be considered. Electrical components must be appropriately isolated from solvents, and improper mounting of plumbing components can result in cross-contamination, short valve life, and wasted reagents. It is especially important to keep in mind how spills can affect the unit’s performance and user safety.

SCALING-UP PEPTIDE SYNTHESIS Synthetic peptides and bioactive proteins have rapidly entered pharmaceutical production. Peptides, like proteins, are produced in smaller quantities than classical organic pharmaceuticals since their dosages are typically small. Literature reports of large-scale peptide syntheses report 100–500-gram syntheses as opposed to the tens of kilograms typical for other classes of pharmaceuticals. Scaling up peptide synthesis requires skilled technical personnel with experience both in biochemistry and biomolecular purification. Peptides, depending on size and composition, may require synthesis and purification typical of organic molecules or synthesis resembling the cellular protein assembly and purifications typically found in bioprocess operations. Peptide properties require those engaged in their scale up to be extremely flexible and multidisciplinary; scaling up peptide synthesis is undergoing intense activity as more pharmaceutical companies bring peptides to market. Synthetic

PH2270_book.fm Page 344 Tuesday, May 16, 2006 9:37 AM

344

Producing Biomolecular Substances

methodologies currently meet the need for high yield processes. The advantages and disadvantages of the several technologies used for peptide synthesis — from solution synthesis to solid-phase synthesis to recombinant expression in a host system — must be considered together with their limitations. Synthesizing large-quantities of biologically active peptides and proteins can be achieved by using various chemical, biochemical, and recombinant production strategies, although peptides have their own unique characteristics and require unique scale up approaches. Advances most relevant to large-scale peptide synthesis include enhanced flexibility in both solution- and solid-phase synthesis, developing resins for synthesizing protected fragments, and improving fabrication condensations either in solution or on the solid-phase resin. Calcitonin production demonstrates how scale-up and production approaches have evolved over the years. Calcitonin, a 32mer peptide hormone, is used to treat Paget’s disease and osteoporosis where insufficient calcium is deposited in the bone. Sandoz, for example, manufactured calcitonin by solution phase, whereas Armour produced it by solid phase. Bachem produced it by solution phase until they switched to convergent synthesis. Calcitonin fragments were selected according to their glycine and proline distribution and synthesized by solid phase on Sasrin resin. The first fragment containing two cysteine residues is oxidized and purified by counter-current distribution. Two other protected fragments are also purified, and the three fragments coupled in solution. The final product is then extracted and purified. Recombinant processes for the production of calcitonin have been developed by a number of companies, since they would tend to be more cost-effective than other synthesis methods for large-scale production. Due to small molecular size, however, calcitonin is degraded by bacterial proteases. Successful cloning strategy, then, involves expression of calcitonin as a fusion protein with subsequent cleavage, purification, and amidation. Large-scale peptide synthesis can be carried out either in solution, in solid-phase, in combination synthesis, or by recombinant methods, although there are vast technical and operational differences between the various strategies. To obtain 1–10 kg of a relatively pure peptide by solution synthesis, large amounts of protected amino acids (around 50 kg) and protected fragments (around 10 kg) must be prepared, requiring many hours of work. That level of operation has yet to be achieved with the solid phase, however. Typically, solid-phase synthesis has not progressed beyond kilogram production. The most attractive approach for 40-mer peptide, scale up is the combination method, that is, making fragments by solid-phase synthesis and then coupling them by fragment condensation. For peptides longer than 50-mer, the recombinant approach offers an excellent combination of quality and cost-effectiveness.

SOLUTION PEPTIDE SYNTHESIS SCALE-UP Large-scale peptide syntheses development usually follows one of two pathways. Depending upon the size of the molecule, economics currently favors producing small molecules of three to eight residues by classical solution synthesis. These syntheses are similar in methodology and equipment to normal organic pharmaceutical synthesis. Depending upon the complexity of the molecule, solution synthesis can be managed either in steps or by fragment condensation.

PH2270_book.fm Page 345 Tuesday, May 16, 2006 9:37 AM

Biomolecular Synthesizers

345

SOLID-PHASE PEPTIDE SYNTHESIS SCALE-UP Peptides greater in length than 8–10 mer can be conveniently scaled-up using solidphase synthesis. Solid-phase synthesis of kilogram quantities of small peptides (5–8 residues) is cost-effective when compared to solution methods with modern synthesis and purification techniques. The tedious and expensive development effort for developing a solution-synthesis process is rapidly becoming impractical in light of the necessity for reducing development risk by getting large quantities of material (i.e., hundreds of grams) on the market in a very short time. The time it takes to add an amino acid to the peptide chain (cycle time) in large-scale solid-phase synthesis is very close to that in the laboratory where a single amino acid may be easily added every day. Speed, minimal racemization, few insolubility problems, and convenient documentation of good manufacturing practices by a computer printout are the advantages of large-scale solidphase synthesis. The disadvantages are awkward monitoring, capping of incomplete reactions, and the strong hydrofluoric-acid cleavage required for tBoc synthesis. And, since there is no intermediate purification in solid-phase synthesis, the rate-limiting step becomes the final purification. Solid-phase synthesis is more cost-effective in terms of labor than solution synthesis. Since many peptides never reach typical pharmaceutical production scale because of their high biological activity, labor costs remain a significant portion of cost. Although solid-phase synthesis has many advantages, ranging from easy process automation to a single purification sequence, there are some significant drawbacks in using solid-phase synthesis for large-scale production. Two of the most significant drawbacks are the methods for cleaving the peptide from the resin and the removal of trace impurities from the final product. Cleavage of the resin at bench-scale is usually accomplished by using anhydrous hydrogen fluoride (HF) — very dangerous to handle. Scaling up HF reactions for peptide synthesis are only possible in highly specialized facilities, and scale ups beyond 1kg of resin, which only represents about four moles of crude peptide, have not been accomplished to date. Most scale-up methods avoid HF by using protecting groups that are removable under milder conditions. Many peptide hormones and their active analogs are carboxyl terminal amides allowing ammonolysis for cleavage of the peptide from the resin. Ammonolysis is a well-understood procedure that is readily scaled up using standard process equipment, although it cannot be used for peptides containing aspartic and glutamic acid derivatives. Several methods, such as the using Fmoc amino acid protection combined with an acid labile resin, allow the use of milder conditions for cleaving the peptide from the resin. To date these methods have not been examined at the kilogram level but might be viable methodologies if Fmoc amino acid derivative cost of is reduced.

RECOMBINANT PEPTIDE SYNTHESIS SCALE-UP Recombinant bacterial processes are well suited for the production of large quantities of peptides because after initial investment, the production output is high and the cost is low. Production yields of recombinant growth hormones in bacteria have exceeded 20 grams per liter of medium. Yeast and mammalian systems can express short peptide sequences, but at lower yields than bacterial systems.

PH2270_book.fm Page 346 Tuesday, May 16, 2006 9:37 AM

346

Producing Biomolecular Substances

ENZYMATIC PEPTIDE SYNTHESIS SCALE-UP Scaling up of enzymatic peptide synthesis can be accomplished with immobilized carboxypeptidase Y, and almost 500 mg of the 99% pure peptide Bzl-Arg-MetNH2, for example, can be produced within 24 hours. There are, as yet, no data to establish that really long peptides have been successfully synthesized enzymatically at large-scale.

MANUFACTURERS’ DIRECTORY PEPTIDE SYNTHESIZERS Applied Biosystems (800) 345-5ABI Advanced Chemtech (502) 636-5604 Applied Protein Technologies (800) 782-4224 DuPont (800) 551-2121 Labortec AG (distr. by Bachem Bioscience) (800) 634-3183 Milligen/Biosearch (Millipore) (800) 225-1380 Pharmacia/LKB Biotech (201) 457-8000 Vega Biotechnologies (800) 528-4882

PEPTIDE SYNTHESIZER REAGENTS Applied Biosystems (800) 345-5ABI Bachem Bioscience (800) 634-3183 Beckman Instruments in CA (800) 742-2345 Cambridge Research Biochemicals (800) 327-0125 Fluka Chemical Co. (800) 358-5287 MTM Research Chemicals (800) 238-2324 Peninsula Laboratories, Inc. (800) 922-1516 Peptides International (800) 777-4779 Pfaltz and Bauer, Inc. (800) 225-5172 Pierce Chemical (800) 874-3723 Research Organics (800) 321-0570 Research Plus, Inc. (800) 341-2296 Sigma Chemical Co. (800) 325-3010 Synthetech, Inc. (503) 967-6575 Toronto Research Chemicals (416) 638-9696 Wako Chemical (800) 992-WAKO

PH2270_book.fm Page 347 Tuesday, May 16, 2006 9:37 AM

Biotechnology Glossary Aerobic Requiring oxygen for growth. Affinity chromatography Technique used to isolate pure antibodies. A column is prepared from antigen covalently coupled to an inert solid phase, such as cross-linked dextran beads. The antibody-containing solution is run into the column with a neutral buffer. Specific antibody binds to the antigen, while unbound antibody and other proteins are washed through. The specific antibody is eluted using a buffer that dissociates the antigen/antibody bond, which is either of a high or low pH, or a denaturing agent. By using antibody bound to a solid phase, the technique can be used to isolate antigens. Affinity chromatography is used in bioprocess engineering for the separation and purification of almost any biomolecule on the basis of either biological function or chemical structure. The substance of interest is first bound to the immobilized ligand, and then dissociated and recovered by changing experimental conditions, i.e., the molecule is specifically and reversibly absorbed by the complementary binding substance (ligand) and immobilized on a matrix. Allogenic Of the same species but with a different genotype. Amino acids Building blocks of proteins; there are twenty amino acids: alanine, arginine, aspargine, aspartic acid, cysteine, glutamic acid, glutamine, glycine, histidine, isoleucine, leucine, lysine, methionine, phenylalanine, proline, serine, threonine, tryptophan, tyrosine, and valine. Amplification The process of increasing the number of copies of a particular gene or chromosomal sequence. Anaerobic Able to grow in the absence of oxygen. Antibiotic Substance typically formed as a metabolic byproduct in bacteria or fungi and used to treat bacterial infections. Antibiotics can be produced either naturally, using microorganisms, or synthetically. Antibody Protein produced by higher animals and humans in response to the presence of a specific antigen. Anticodon A triplet of nucleotide bases (codon) in transfer RNA that pairs with (is complementary to) a triplet in messenger RNA. For example, if the codon is UCG, the anticodon is AGC. See also Base; Base pair; Complementarity. Autosome A chromosome other than a sex chromosome. Auxotrophic mutant A cell mutated so that it requires a specific growth substance beyond the minimum required for normal metabolism and reproduction.

347

PH2270_book.fm Page 348 Tuesday, May 16, 2006 9:37 AM

348

Producing Biomolecular Substances

Bacillus subtilis A bacterium commonly used as host in recombinant DNA experiments. Important because of its ability to secrete protein. Bacteriophage A virus that lives in and destroys bacteria. Also called a phage. Base One of four chemical units on a DNA molecule that, according to their order and pairing, represent various amino acids: adenine (A), cytosine (C), guanine (G), and thymine (T). In RNA, uracil (U) substitutes for thymine. Base pair Two nucleotide bases on different strands of the nucleic acid molecule that can bond together in only one way: adenine with thymine (DNA) or uracil (RNA) and guanine with cytosine. Batch processing Growth in a closed system with a specific amount of nutrient medium. In bioprocessing, defined amounts of nutrient material and living matter are placed in a bioreactor and removed when the process is completed. See also Continuous processing. Biocatalyst An enzyme that activates or speeds up a biochemical reaction. Bioconversion Chemical restructuring of raw materials through the use of a biocatalyst. Biological response modulator A substance that alters the growth of functioning of the cell, including hormones and other compounds (e.g., cytokines, lymphokines) that affect the nervous and immune systems. Biomass The total amount of biological matter in a given area. As commonly used in biotechnology, biomass refers to the total cell mass in a fermenter or bioreactor. Bioprocess Process in which living cells or their components are used to produce a desired end product. Biosynthesis Production of a chemical by means of living organisms. Biotechnology Development of products by a biological process using intact organisms such as yeasts and bacteria or by using natural substances (e.g., enzymes) from organisms. β) lymphocytes (B cells) A class of lymphocyte released from bone marrow B (β that produce antibodies. Catalyst An agent such as an enzyme or a metallic complex that facilitates a reaction but is not changed by the reaction. Cell culture Growing cells under laboratory conditions; also those cells grown by cell culture. Cell fusion See fusion. Cell line Cells removed from a living organism that grow and continuously replicate in vitro. Chemostat A growth chamber that keeps a bacterial culture at a specific volume and rate of growth by continually adding fresh nutrient medium while removing spent medium and waste products from the culture. Chimera An animal or lower organism produced by grafting an embryonic part of one species onto the embryo of either the same or a different species. Chromosomes Threadlike components in a cell that contain DNA and various associated proteins; genes are carried on chromosomes. Cistron The length of chromosomal DNA representing the smallest functional heredity unit, essentially a gene.

PH2270_book.fm Page 349 Tuesday, May 16, 2006 9:37 AM

Biotechnology Glossary

349

Clone A group of cells or organisms derived from a common ancestor. Because there is no combining of genetic material (as in sexual reproduction), the clones are genetically identical to their parents; also the process of producing a cell line of clones. Codon A sequence of three nucleotide bases that specifies production of an amino acid or represents a signal to stop or start a function. Coenzyme An organic compound that is necessary for the functioning of an enzyme; coenzymes are typically smaller than enzymes, and sometimes are separable from them. Cofactor A nonprotein substance required for some enzymes to function; cofactors can either be coenzymes or metallic ions. Colony-stimulating factors (CSFs) Lymphokines that induce the maturation and proliferation of white blood cells from the embryonic cells present in bone marrow. Complementarity The relationship of nucleotide bases on two different DNA or RNA strands; when these bases are paired properly (adenine with thymine in DNA or with uracil in RNA, and guanine with cytosine), the strands are said to be complementary. Complementary DNA (cDNA) DNA synthesized from a messenger RNA, rather than from a DNA template; this DNA is typically used for cloning or as a DNA probe for locating specific genetic sites in DNA hybridization studies. Continuous processing A bioprocessing method in which new materials are added and products continuously removed at a rate that maintains a specific volume. See also Batch processing. Culture Living organisms cultivated in prepared medium; also to grow such organisms in prepared medium. Culture medium A complex mixture of organic and inorganic materials used as a nutrient system for the artificial cultivation of bacteria or other cells. Cyto- Referring to the cell or the cell plasm. Cytogenetics The study of the cell and its heredity-related components, especially chromosomes. Cytoplasm Cellular material within the cell membrane that typically surrounds the nucleus. Cytotoxic Causing cell death. Deoxyribonucleic acid (DNA) A molecule that carries genetic information for living systems and consists of four bases (adenine, cytosine, guanine, and thymine) attached to a sugar-phosphate backbone, and arranged in two connected strands forming a double helix. See also Complementary DNA (cDNA); Double helix; Recombinant DNA (rDNA). Diploid A cell with two complete sets of chromosomes. See also Haploid. DNA See Deoxyribonucleic acid (DNA). DNA probe A nucleic acid molecule that has been labeled with a radioactive isotope, dye, or enzyme and is used to locate a particular nucleotide sequence (gene) on a DNA molecule. DNA sequencing Determining the order of nucleotide bases in a DNA molecule.

PH2270_book.fm Page 350 Tuesday, May 16, 2006 9:37 AM

350

Producing Biomolecular Substances

Double helix A term often used to describe the configuration of a DNA molecule, consisting of two spiraling strands of nucleotides (a sugar, a phosphate, and a base) joined crosswise by the specific pairing of the bases. See also Deoxyribonucleic acid (DNA); Base; Base pair. Downstream processing Bioprocess stages that take place after the fermentation or bioconversion stage, and which include the separation, purification, and packaging of the end product. Endonuclease An enzyme that breaks nucleic acids at specific interior bonding sites, producing specific nucleic acid fragments of various lengths. See also Exonucleases. Enzyme A protein catalyst that facilitates the specific chemical and/or metabolic reactions necessary for cell growth and reproduction. Enzyme-linked immunosorbent assay (ELISA) An analytical method used for detecting antibodies, where an antigen is adsorbed onto a solid phase and the test antibody is added. Similar to a radioimmunoassay, the ELISA ligand used to detect an antibody is an enzyme linked to a molecule specific for that bound antibody. Enzymes such as peroxidase and phosphatase are often used. In the final stage a chromogenic substrate is typically added, generating a colored end product in the presence of the enzyme portion of the ligand. The optical density of this solution is measured after a defined period, and is directly proportional to the amount of enzyme, which is, in turn, related to the amount of test antibody. Erythropoietin A protein that boosts production of red blood cells; clinically useful in treating certain types of anemias. Escherichia coli (E. coli) A bacterium that inhabits the intestinal tract of many vertebrates. Many of the recombinant DNA techniques have been carried out with this organism because it has been well-characterized genetically, and has about a 30% protein turnover. Eukaryote A solitary cell or one from an organism containing a nucleus with a well-defined membrane surrounding it. All organisms except bacteria, viruses, and blue-green algae are eukaryotic. See also Prokaryote. Exon Part of the gene in a eukaryotic cell that is transcribed into messenger RNA, and encodes a protein. See also Intron; Splicing. Exonuclease An enzyme that breaks down nucleic acids only at the ends of polynucleotide chains releasing one nucleotide at a time in sequential order. See also Endonuclease. Expression In genetics, the manifestation of a characteristic specified by a gene. In some hereditary diseases, a person can carry the gene for the disease but not actually have it since the gene present is not expressed. In industrial biotechnology, the term is used to mean the production of a protein through instruction of a gene inserted into a new host. Fermentation The anaerobic process of growing microorganisms for the production of various chemical or pharmaceutical compounds. Microbes are typically incubated under specific conditions in the presence of nutrients in large tanks called fermenters.

PH2270_book.fm Page 351 Tuesday, May 16, 2006 9:37 AM

Biotechnology Glossary

351

Frameshift Insertion or deletion of one or more nucleotide bases such that the incorrect base triplets are read as codons. Fusion Joining the membrane of two cells, thereby creating daughter cells containing the nuclear material from both parent cells used in forming hybridomas. Gene A chromosome segment that either directs the syntheses of proteins, or has regulatory functions. See also Operator gene; Regulatory gene; Structural gene; Suppressor gene. Gene machine A computerized device for synthesizing genes by combining nucleotides (bases) in the proper order; also used by Pharmacia/LKB Biotechnology as the trade name for their automated DNA synthesizer unit. Gene mapping Determination of relative gene locations on a chromosome. Gene sequencing Determining the sequence of nucleotide bases in a strand of DNA; also called DNA sequencing. Gene therapy Replacement of a defective gene in an organism suffering from a genetic disease where recombinant DNA techniques are used to isolate properly functioning genes and insert them into cells. More than 300 single-gene genetic disorders have been identified in humans, and a significant percentage of these may be amenable to gene therapy. Genetic code A mechanism by which genetic information is stored in living organisms. The code uses sets of three nucleotide bases (codons) to make the amino acids that, in turn, constitute proteins. Genetic engineering Technology used to alter the genetic material of living cells to make them capable of producing new substances or performing new functions. Genetic screening Use of specific biological tests to screen for inherited diseases or medical conditions. Such testing can be conducted prenatally to check for metabolic defects and congenital disorders in the developing fetus, or postnatally to screen for carriers of hereditary disease. Genome The total hereditary material of a cell, comprising the entire chromosomal set found in the nucleus of a given species. Genotype The genetic makeup of an individual or group. See also Phenotype. Growth hormones Proteins produced by the pituitary gland and involved in cell growth that are sometimes added to mammalian cell cultures in vitro to increase cell or end product yield in a given time; also called somatotropins. Haploid A cell with half the usual number of chromosomes, i.e., only one chromosome set. Sex cells are haploid. See also Diploid. hom*ologous Corresponding or alike in structure, position, or origin. Hormone A chemical that acts as a messenger or stimulator signal, relaying instructions to stop or start certain physiological activities. Hormones are synthesized in one type of cell and then released to direct the function of other cell types. Host A cell or organism used for growth of a virus, plasmid, or other form of foreign DNA, or for the production of cloned substances.

PH2270_book.fm Page 352 Tuesday, May 16, 2006 9:37 AM

352

Producing Biomolecular Substances

Host-vector system The combination of DNA-receiving cells (host) and DNAtransporting substance (vector) used for introducing foreign DNA into a DNA receptor cell. Hybridization The production of offspring, or hybrids, from genetically dissimilar parents. The process can be used to produce hybrid plants (by crossbreeding two different varieties) or hybridomas (hybrid cells formed by fusing two unlike cells, used in producing monoclonal antibodies). The term is also used to refer to the binding of complementary strands of DNA or RNA. Hybridoma A cell line produced by fusing two cells of different origin. In monoclonal antibody technology, hybridomas are formed by fusing an immortal cell (one that divides continuously) and an antibody-producing cell. See also Monoclonal antibody; Myeloma. Immunoglobulin A generic term for proteins that function as antibodies, differ somewhat in structure, and which are grouped into five categories on the basis of these differences: Immunoglobulin G (IgG), Immunoglobulin M (IgM), Immunoglobulin A (IgA), Immunoglobulin D (IgD), and Immunoglobulin E (IgE). Immunomodulators A diverse class of proteins that boost the immune system; many are cell growth factors (e.g., cytokines, lymphokines) that accelerate the production of specific cells important in mounting an immune response in the body. Immunotoxins Specific monoclonal antibodies that have a protein toxin molecule attached; the monoclonal antibody is targeted against a tumor cell and the toxin is designed to kill that cell when the antibody binds to it. Immunotoxins have also been called magic bullets. Inducer Any molecule or substance that increases the rate of enzyme synthesis, usually by blocking the action of the corresponding repressor. Interferon A class of lymphokine important in the immune response. There are three major types of interferon: alpha (leukocyte), beta (fibroblast), and gamma (immune). Interferons inhibit viral infections and may have anticancer properties. Interleukin A type of lymphokine whose role in the immune system is being extensively studied. Two types of interleukin have been identified. Interleukin 1 (IL-1), derived from macrophages, is produced during inflammation and amplifies the production of other lymphokines, notably interleukin 2 (IL-2). IL-2 regulates the maturation and replication of T lymphocytes. Intron A DNA sequence in eukaryotic cells that is contained in the gene but does not encode for protein. The intron “splits” the coding region of the gene into segments called exons. See also Exon; Splicing. In vitro Literally, in glass. Operations performed in a glass apparatus. In vivo Operations performed in living organisms. Isogeneic Of the same genotype; also isogenic. Library A set of cloned DNA fragments.

PH2270_book.fm Page 353 Tuesday, May 16, 2006 9:37 AM

Biotechnology Glossary

353

Ligase An enzyme used to join DNA or RNA segments together; also called DNA ligase or RNA ligase, respectively. Linkage The tendency for certain genes to be inherited together due to their physical proximity on the chromosome. Linker A DNA fragment with a restriction site that can be used to join DNA strands. Lymphokine A class of soluble protein produced by white blood cells that plays a role, not as yet not fully understood, in the immune response. See also Interferon; Interleukin. Lysis Literally, breaking cells apart. Lysozyme Enzymes present in tears, saliva, egg whites, and some plant tissues, that destroy the cells of some bacteria. Macrophage A type of white blood cell produced in blood vessels and loose connective tissues that ingests dead cells and tissue and is involved in producing interleukin-1 (IL-1). When exposed to the lymphokine macrophageactivating factor, macrophages also kill tumor cells. See also Phagocyte. Macrophage-activation factor (MAF) An agent that stimulates macrophages to attack and ingest cancer cells. Magic bullets See Immunoglobulin. Messenger RNA (mRNA) A nucleic acid that carries instructions to a ribosome for the synthesis of a particular protein. Monoclonal antibody A highly specific, purified antibody derived from only one clone of cells and that recognizes only one antigen. See also Hybridoma; Myeloma. mRNA See Messenger RNA. Multigenic Of hereditary characteristics, one that is specified by several genes. Mutagen A substance that induces mutations. Mutant A cell that manifests new characteristics due to a change in its DNA. Mutation A change in the genetic material of a cell. Muton The smallest element of a chromosome whose alteration can result in mutation or a mutant organism. Myeloma A type of tumor cell that is used in monoclonal antibody technology to form hybridomas. Nuclease An enzyme that, by cleaving chemical bonds, breaks down nucleic acids into their constituent nucleotides. See also Exonuclease. Nucleic acid Large molecules generally found in the cell’s nucleus and/or cytoplasm that are made up of nucleotide bases. The two kinds of nucleic acid are DNA and RNA. Nucleotide Building blocks of nucleic acids. Each nucleotide is composed of sugar, phosphate, and one of five nitrogen bases (one base is for RNA). The sequence of the bases within the nucleic acid determines what proteins will be made. Nucleotide base See Base. Nucleus The structure within eukaryotic cells that contains chromosomal DNA. Oligonucleotide A polymer consisting of a small number (about 2–10) of nucleotides.

PH2270_book.fm Page 354 Tuesday, May 16, 2006 9:37 AM

354

Producing Biomolecular Substances

Oncogene A gene thought to be capable of producing cancer. Oncogenic Cancer-causing. Oncology Study of tumors. Operator gene A region of the chromosome, adjacent to the operon, where a repressor protein binds to prevent transcription of the operon. Operon A gene sequence responsible for synthesizing the enzymes needed for biosynthesis of a molecule. An operon is controlled by an operator gene and a repressor gene. Opsonin An antibody that renders bacteria and other antigenic material susceptible to destruction by phagocytes. Paratope A site on an antibody molecule that attaches to an antigen. Peptide Two or more amino acids joined by a linkage called a peptide bond. Phage See Bacteriophage. Phagocyte A type of white blood cell that can ingest invading microorganisms and other foreign material. See also Macrophage. Phenotype Observable characteristics resulting from interaction between an organism’s genetic makeup and the environment. See also Genotype. Plasmid A small circular form of DNA that carries certain genes and is capable of replicating independently in a host cell. Polyclonal Derived from different types of cells. Polymerase A general term for enzymes that carry out the synthesis of nucleic acids. Polypeptide A long chain of amino acids joined by peptide bonds. Probe See DNA probe. Prokaryote An organism (e.g., bacterium, virus, blue-green algae) whose DNA is not enclosed within a nuclear membrane. See also Eukaryote. Promoter A DNA sequence that is located in front of a gene and controls gene expression. Promoters are required for binding of RNA polymerase to initiate transcription. Prophage A phage nucleic acid that is incorporated into the host’s chromosome but does not cause cell lysis. Protein-A A protein produced by the bacterium staphylococcus aureus that specifically binds antibodies. It is useful in the purification of monoclonal antibodies. Recombinant DNA (rDNA) DNA formed by combining segments of DNA from different types of organisms. Regulatory gene A gene that acts to control the protein-synthesizing activity of other genes. Replication Reproduction or duplication (e.g., as in replication of an exact copy of a strand of DNA). Replicon A DNA segment (e.g., chromosome or plasmid) that can replicate independently. Repressor A protein that inhibits transcription of a gene by binding to an operator adjacent to a structural gene. Restriction enzyme An enzyme that breaks DNA in highly specific locations, creating gaps into which new genes can be inserted.

PH2270_book.fm Page 355 Tuesday, May 16, 2006 9:37 AM

Biotechnology Glossary

355

Retrovirus An animal virus that contains the enzyme reverse transcriptase. This enzyme converts viral RNA into DNA that can combine with the DNA of the host cell to produce more viral particles. Ribonucleic acid (RNA) A molecule similar to DNA that functions primarily to decode the instructions for protein synthesis that are carried by genes. See also Messenger RNA (mRNA); Transfer RNA (tRNA). Ribosome A cellular component containing protein and RNA that is involved in protein synthesis. Scale up The transition from small-scale to large industrial-scale production. Selective medium Nutrient material constituted such that it will support the growth of specific organisms while inhibiting the growth of others. Single-cell proteins Cell or protein extracts from microorganisms grown in large quantities for use as protein supplements. Splicing The removal of introns and joining of exons to form a continuous coding sequence in RNA. Structural gene A gene that codes for a protein, such as an enzyme. Substrate Material acted on by an enzyme. Suppressor gene A gene that can reverse the effect of a mutation in other genes. Template A molecule that serves as the pattern for synthesizing another molecule. T lymphocytes (T cells) White blood cells produced in the bone marrow but maturing in the thymus; important in the body’s defenses against certain bacteria and fungi, helping B lymphocytes make antibodies, and helping in the recognition and rejection of foreign tissues. T lymphocytes are also important in the body’s defense against cancers. Transcription Synthesis of messenger (or any other) RNA on a DNA template. Transduction The transfer of genetic material from one cell to another by means of virus or phage vector. Transfection The infection of a cell with nucleic acid from a virus, resulting in replication of the complete virus. Transfer RNA (tRNA) RNA molecules that carry amino acids to sites on ribosomes where proteins are synthesized. Transformation A change in the genetic structure of an organism by the incorporation of foreign DNA. Transgenic organism An organism formed by the insertion of foreign genetic material into the germ cell lines of organisms. Recombinant DNA techniques are commonly used to produce transgenic organisms. Translation A process by which the information on a messenger RNA molecule is used to direct the synthesis of a protein. Transposable element See Transposon. Transposon A DNA segment that can be moved around and inserted at several sites in bacterial DNA or in a phage to alter the host DNA. Transposons can also cause mutations. tRNA See Transfer RNA. Vaccine A preparation that contains an antigen consisting of whole diseasecausing organisms (killed or weakened), or parts of such organisms, used to confer immunity against the disease that the organisms cause. Vaccine

PH2270_book.fm Page 356 Tuesday, May 16, 2006 9:37 AM

356

Producing Biomolecular Substances

preparations can be natural, synthetic, or derived by recombinant DNA technology. Vector An agent (e.g., plasmid or virus) used to carry new DNA into a cell. Virion An elementary viral particle consisting of genetic material and a protein covering. Virus A submicroscopic organism that contains genetic information but cannot reproduce itself. To replicate, it must invade another cell and use part of that cell’s reproductive machinery. Yeast The general term for single-celled fungi that reproduce by budding. Some yeasts can ferment carbohydrates (starches and sugars).

PH2270_book.fm Page 357 Tuesday, May 16, 2006 9:37 AM

Bibliography and Recommended Reading REFERENCES 1. Acton, R.T., P.A. Barstad, and R.K. Zwerner. 1979. Propagation and scaling up of suspension cultures in Methods in Enzymology, Vol. 58, 211, edited by W.B. Jakoby and I.H. Pastan. New York: Academic Press. 2. Adams, R.L.P. 1980. Cell culture for biochemists, in Laboratory Techniques in Biochemistry and Molecular Biology, edited by R.H. Burdon and P.H. von Kippenberg. Amsterdam: Elsevier. 3. Aebersold, R.H., et al. 1987. N-terminal amino acid sequence analysis of proteins separated by one or two dimensional gel electrophoresis after in situ digestion on nitrocellulose. Proc. Natl. Acad. Sci. USA 84:6970–6974. 4. Aiba, S., M. Shoda, and M. Nagatani. 1968. Kinetics of product inhibition in alcohol fermentation. Biotechnol. Bioeng. 10:845. 5. Aiba, S., A.E. Humphrey, and N.F. Millis. 1973. Biochemical Engineering. 2nd ed., 242–246. Tokyo: University of Tokyo Press. 6. Al-Hakim, A.H., and R. Hull. 1986. Studies towards the development of chemically synthesized non-radioactive biotinylated nucleic acid hybridization probes. Nucl. Acids Res. 24:9965–9976. 6. Ansari, A.A., N.S. Hattikudur, S.R. Joshi, et al. 1985. ELISA solid phase: Stability and binding characteristics. J. Immunol. Methods 84:117–124. 7. Antonarakis, S.E. and H.H. Kazazian. 1983. Polymorphism and molecular pathology of the human β-globin gene, in Progress in Histology, XIII, 49–73, edited by E.B. Brown. New York: Grune and Stratton. 8. Ausubel, F.M., R. Brent, R.E. Kingston, et al. 1987. Current Protocols in Molecular Biology. New York: Greene Publishing/Wiley-Interscience. 9. Aziz, C.E., M.W. Fitch, L.K. Linquist, J.G. Pressman, G. Georgiou, and G.E. Speitel, Jr. Methanothrophic biodegradfation of trichloroethylene in a hollow-fiber membrane bioreactor. Environ. Sci. Technol. 1995, 29:2574–2583. This study reports biodegradation and transfer experiments of a combination of a fed-batch bioreactor and a hollow-fiber membrane bioreactor for TCE elimination from waste water. Metabolism was spatially separated from cometabolism avoiding the usual competition between substrates. TCE mass balances were closed by using radiolabeled TCE and a mathematical model of the hollow-fiber bioreactor was presented. 10. Bailey, J.E. and D.F. Ollis. 1986. Biochemical Engineering Fundamentals, 2nd ed. New York: McGraw Hill. 11. Balazs, I., M.I. Baird, and K. Wexler, et al. 1986. Characterization of the polymorphic DNA fragments detected with a new probe derived from the D14S1 locus. Am. J. Hum. Genet. 39–A229. 12. Barnes, D. 1987. Serum-free animal cell culture. BioTechniques 5:534–542. 357

PH2270_book.fm Page 358 Tuesday, May 16, 2006 9:37 AM

358

Producing Biomolecular Substances

13. Basheer, S., K.I. Mogi, and M. Nakajima. 1996. Development of a novel hollow-fiber membrane reactor for the interesterification of tryglycerides and fatty acides modified lipase. Process Biochem. 30:531–536. 14. Bell, G.I., M.J. Selby, and W.J. Rutter. 1982. A highly polymorphic region near the human insulin gene is composed of simple tandemly repeating sequences. Nature 295:31–35. 15. Berger, S.L. and A.R. Kimmel, eds. 1987. Guide to molecular cloning techniques, in Methods in Enzymology, Vol. 152. San Diego: Academic Press/Harcourt Brace Jovanovich. 16. Botstein, D., R.L. White, M. Skolnick, et al. 1980. Construction of a genetic linkage map in man using restriction fragment length polymorphisms. Am. J. Hum. Genet. 32:314–331. 17. Brodeur, B.R., P. Tsang, and Y. Larose. 1984. Parameters affecting ascites tumour formation in mice and monoclonal antibody production. J. Immunol. Methods 71:265. 18. Bridgman, P.W. 1946. Dimensional Analysis. New Haven: Yale University Press. 19. Brindle, K. and T. Stephenson. 1996. The application of membrane biological reactors for the treatment of wastewaters. Biotechnol. Bioeng. 49:601–610. A well-documented review on the various ways to involve membranes for wastewater biotreatment. 20. Brown, D.E. 1981. Power requirements in a production-scale fermenter, in Fluid Mixing, N1. Inst. Chem. Engrs. Symp. Ser., No. 64, Rugby. 21. Butt, W.R., ed. 1984. Practical Immunoassay: The State of the Art. New York: Marcel Dekker. 22. Cahn, F., I.V. Yannas, and A.F. Steuer. 1986. Porous microcarrier particles for mammalian cell culture. Presented at the 192nd National Meeting of the American Chemical Society, Anaheim, CA. 23. Capon, D.J., E.Y. Chen, A.I. Levinson, et al. 1983. Complete nucleotide sequence of the T24 human bladder carcinoma oncogene and its normal hom*ologue. Nature 302:33–37. 24. Carstensen, J.T., et al. 1982. Scale up factors in the manufacturing of solution dosage forms. Pharm. Tech. 6(11):67–77. 25. Central Intelligence Agency and Defense Intelligence Agency Report 28 May, 2003: Iraqi Mobile Biological Warfare Agent Production Plants. 26. Cepko, C. 1988. Immortalization of neural cells via oncogene transduction. Trends Neurosci. 11:6–9. 27. Chandler, H.M., J.C. Cox, K. Healey, A. MacGregor, R.R. Premier, and J.G.R. Hurrell. 1982. An investigation of the use of urease-antibody conjugates in enzyme immunoassays. J. Immunol. Methods 53:187. 28. Chang, T., Z. Steplewski, and H. Koprowski. 1980. Production of monoclonal antibodies in serum-free medium. J. Immunol. Methods 39:369–375. 29. Chilton, T.H., T.B. Drew, and R.H. Jebens. 1944. Heat transfer coefficients in agitated vessels. Ind. Eng. Chem. 36:510. 30. Chirgwin, J.M. 1990. Molecular biology for non-molecular biologists. Diabetes Care 13:188–198. 31. Cohen, S.A., A. Chang, S. Boyer, and R. Helling. 1973. Construction of biologically functional bacterial plasmids in vitro. Proc. Natl. Acad. Sci. USA 70:3240–3244. 32. Colburn, A.P. 1933. A method of correlating forced convection heat transfer data and a comparison with fluid friction. TransAm. Inst. Chem. Eng. 29:174.

PH2270_book.fm Page 359 Tuesday, May 16, 2006 9:37 AM

Bibliography and Recommended Reading

359

33. Colowick, S.P. and N.O. Kaplan. 1979. Methods in Enzymology, Vol. 58, edited by W.B. Jakoby and I.H. Pastan. New York: Academic Press. 34. Connell, C., S. Fung, C. Heiner, et al. 1987. Automated DNA sequence analysis. BioTechniques 5:342–348. 35. Coulson, J.M., J.F. Richardson, and D.G. Peaco*ck. 1979. Chemical Engineering. Vol. 3, 2nd ed. Oxford: Pergamon Press. 36. Coulson, J.M., J.F. Richardson, J.R. Backhurst, and J.H. Harker. 1978. Chemical Engineering, Vol. 1, 3rd ed. Oxford: Pergamon Press. 37. Cummings, G.H., and A.S. West. 1950. Heat transfer data for kettles with jackets and coils. Ind. Eng. Chem. 42:2303. 38. Crick, F.H.C. 1966. The genetic code: III. Sci. Am. 215:55–62. 39. Crick, F.H.C., L. Barnett, S. Brenner, and R.J. Watts-Tobin. 1961. General nature of the genetic code for proteins. Nature 192:1227–1232. 40. Connell, C., S. Fung, C. Heiner, et al. 1987. Automated DNA sequence analysis. BioTechniques 5:342–348. 41. Darfler. 1990. A protein-free medium for the growth of hybridomas and other cells of the immune system. In Vitro Cell Dev. Biol. 26:769–778. 42. Darfler, F., and P. Insel. 1984. Growth of lymphoid cells in serum-free medium, in Media for Serum-Free Culture of Neuronal and Lymphoid Cells, 187–196. New York: Alan R. Liss. 43. Deen, K.C., T.A. Landers, and M. Berniger. 1983. Use of T4 DNA polymerase replacement synthesis for specific labeling of plasmid-cloned inserts. Anal. Biochem. 135:456–465. 44. Deindoerfer, F.H., and A.E. Humphrey. 1959a. Analytical method for calculating heat sterilization time. Appl. Micro. 7:256–264. 45. Deindoerfer, F.H., and A.E. Humphrey. 1959b. Principles in the design of continuous sterilizers. Appl. Micro. 7:264–270. 46. De Martinville, B., A.R. Wyman, R. White, et al. 1982. Assignment of the first random restriction fragment length polymorphism (RFLP) locus (D14S1) to a region of human chromosome 14. Am. J. Hum. Genet. 34:216–226. 47. Dierks, S.E., J.E. Butler, and H.B. Richerson. 1986. Altered recognition of surfaceadsorbed compared to antigen-bound antibodies in the ELISA. Mol. Immunol. 23:403–411. 48. Drew, T.B., H.C. Hottel, and W.H. McAdams. 1936. Heat transmission. TransAm. Inst. Chem. Eng. 32:271. 49. Elder, J.K., E.M. Southern. 1987. Computer-aided analysis of one-dimensional restriction fragment gels, in Nucleic Acid and Protein Sequence Analysis: A Practical Approach, 165–172, edited by M.J. Bishop and C.J. Rawlings. Oxford: IRL Press. 50. Elibol M. and F. Mavituna. 1996. Use of perfluorocarbon for oxygen supply to immobilized Streptomyces coelicolor A3(2). Process Biochem. 5:507–512. 51. Erlich, H.A. 1989. PCR Technology. New York: Stockton Press. 52. Ey, P., S. Prowse, and C. Jenkin. 1978. Isolation of pure IgG1, IgG2a, and IgG2b, immunoglobulins from mouse serum using Protein A-Sepharose. Immunochemistry 15:429. 53. Feinberg, A.P. and B. Vogelstein. 1984. A technique for radio labeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 137:266–267. 54. Felder, R.M., and R.W. Rousseau. 1986. Elementary Principles of Chemical Processes, 2nd ed., 630–635. New York: John Wiley and Sons. 55. Finn, R.K. and R.E. Wilson. 1954. Population dynamic of a continuous propagator for micro-organisms. J. Agric. and Food Chem. 2:66.

PH2270_book.fm Page 360 Tuesday, May 16, 2006 9:37 AM

360

Producing Biomolecular Substances

56. Fletcher, S.M., and M.J. Davie. 1980. Monoclonal antibody: A major new development in immunology. J. Forensic Sci. Soc. 20:163–167. 57. Freshney, R.I. 1983. Culture of Animal Cells: A Manual of Basic Technique. New York: Alan R. Liss. 58. Fletcher, S.M., and M.J. Davie. 1980. Monoclonal antibody: A major new development in immunology. J. Foren. Sci. Soc. 20:163–167. 59. Forster, A.C., J.L. McInnis, D.C. Skingle, et al. 1985. Non-radioactive hybridization probes prepared by the chemical labeling of DNA and RNA with a novel reagent, photobiotin. Nucl. Acids Res. 13:745–761. 60. Freshney, R.I. 1983. Culture of Animal Cells: A Manual of Basic Technique. New York: Alan R. Liss. 61. f*ckushima, K., T. Okada, and K. Ochial. 1993. Production, characterization, and application of monoclonal antibodies which distinguished three glucosyltransferases from Streptococcus mutans. Inf. Immunol. 61(1):323–328. 62. Furui M., T. Furutani, T. Shibatani, Y. Nakamoto, and T. Mori. 1996. A membrane bioreactor combined with crystallizer for production of optically (2R, 3S)-3-(4methoxyphenyl)-glycidic acid methyl ester. J. Ferment. Bioeng. 81:21–25. 63. Gardner, I.D. 1985. An enzyme immunoassay for rapid isotyping of monoclonal antibodies. Pathology 17(64). 64. Gilbert, W. 1985. Genes-in-pieces revisited. Science 228:823. 65. Gillam, I.C. and G.M. Tener. 1986. N4-(6-aminohexyl)-cytidine and deoxycytidine nucleotides can be used to label DNA. Anal. Biochem. 157:199–207. 66. Gitler C.J. and W.L. Hochfeld. 1982. Macrophage processing of antigens for activation and suppression. Israel J. Med. [HaRefuah], 49(7):130–135. 67. Glassy, M., J. Tharakan, and P. Chau. 1988. Serum-free media in hybridoma culture and monoclonal antibody production. Biotech. Bioeng. 32:1015–1028. 68. Gluzman, Y., and T. Shenk. 1983. Enhancers and Eukaryotic Gene Expression. New York: Cold Spring Harbor Laboratory. 68. Goldfarb, M., K. Shimizu, M. Perucho, et al. 1982. Isolation and preliminary characterization of a human transforming gene from T24 bladder carcinoma cells. Nature 296:404–409. 69. Goodbourn, S.E.Y., D.R. Higgs, J.B. Clegg, et al. 1983. Molecular basis of length polymorphism in the human zeta-globin gene complex. Proc. Natl. Acad. Sci. USA 80:5022–5026. 70. Goodman, R.H. 1990. Regulation of neuropeptide gene expression. Annu. Rev. Neurosci. 13:111–127. 71. Green, M., T. Maniatis, and D. Melton. 1983. Human β-globin pre-mRNA synthesized in vitro is accurately spliced in Xenopus oocyte nuclei. Cell 32:681–694. 72. Griffiths, J. 1987. Serum and growth factors in cell culture media: An introduction and review. Dev. Biol. Stand. 66:155–160. 73. Gusakov, A.V., A.P. Sinitsyn, I.Y. Davydkin, V.Y. Davydkin, and O. Protas. 1996. Enhancement of enzymatic cellulose hydrolysis using a novel type of bioreactor with intensive stirring induced by electromagnetic field. Appl. Biochem. Biotech. 56:141–153. 74. Gusella, J., R.E. Tanzi, M.A. Anderson, W. Hobbs, K. Gibbons, R. Raschtchian, T.C. Gilliam, M.R. Wallace, N.S. Wexler, and P.M. Conneally. 1984. DNA markers for nervous system diseases. Science 225:1320–1325. 75. Hanley, M.R. 1987. Analysis of receptor-coupled events in neuropeptide action using clonal cell lines. Prog. Brain Res. 72:189–196. 76. Hochfeld, W.L. 1985. Monoclonal Antibodies. Middlefield, CT: Theta Corporation.

PH2270_book.fm Page 361 Tuesday, May 16, 2006 9:37 AM

Bibliography and Recommended Reading

361

77. Hochfeld, W.L. 1991. Downsizing techniques for achieving efficient and cost-effective scale up. Genet. Eng. News, July-August. 78. Hochfeld, W.L. 1992. Liquid Chromatography and Amino Acid Analyzer Systems. Oakbrook, IL: Biotech Publishing. 79. Hochfeld, W.L. 1992. Scale up of continuous culture fermentation for the production of biodrugs. Genet. Eng. News, December. 80. Hochfeld, W.L., 1994. The Biotech 2000 Series — Production and Downstream Processing of Biomolecular Materials, Vol. 1, Building Blocks: Biotechnology Reagents and Consumables, Buffalo Grove, IL: Interpharm Press. 81. Hochfeld, W.L. 1994. The Biotech 2000 Series — The Production and Downstream Processing of Biomolecular Materials, Vol. 2, Growth and Synthesis: Fermenters, Bioreactors, and Biomolecular Synthesizers, Buffalo Grove, IL: Interpharm Press. 82. Hochfeld, W.L. Market and Technology Report: Biotech electrophoresis and DNAsequencing systems. #9803, January, 1998, Streamwood, IL: Biotech Consulting Int., Publ. Div. 83. Hochfeld, W.L. Market and Technology Report: Biotechnology equipment and instrumentation. #9804, January, 1998, Streamwood, IL: Biotech Consulting Int., Publ. Div. 84. Hochfeld, W.L. Market and Technology Report: Fermenters, bioreactors, and biomolecular synthesizers. #9811, January, 1998, Streamwood, IL: Biotech Consulting Int., Publ. Div. 85. Hochfeld, W.L. Market and Technology Report: Immunoassay systems #9813, February, 1998, Streamwood, IL: Biotech Consulting Int., Publ. Div. 86. Hochfeld, W.L. Market and Technology Report: Capillary electrophoresis. #9806, March, 1998, Streamwood, IL: Biotech Consulting Int., Publ. Div. 87. Hochfeld, W.L. Market and Technology Report: Liquid chromatography and amino acid analyzer systems #9809, March, 1998, Streamwood, IL: Biotech Consulting Int., Publ. Div. 88. Hochfeld, W.L. Market and Technology Report: Enzymes and peptides for biotechnology. #9810, May, 1998, Streamwood, IL: Biotech Consulting Int., Publ. Div. 89. Hochfeld, W.L. Market and Technology Report: Specialty amino acids for biopharmaceuticals. #9801, May, 1998, Streamwood, IL: Biotech Consulting Int., Publ. Div. 90. Hochfeld, W.L. 1994. Market and Technology Report: Biopharmaceutic peptides. #9802, June, 1998, Streamwood, IL: Biotech Consulting Int., Publ. Div. 91. Hochfeld, W.L. Market and Technology Report: Biotechnology reagents and consumables. #9805, June, 1998, Streamwood, IL: Biotech Consulting Int., Publ. Div. 92. Hochfeld, W.L. Market and Technology Report: Biotechnology in Western Europe. #9817, September, 1998, Streamwood, IL: Biotech Consulting Int., Publishing Div. 93. Hochfeld, W.L. April, 2000. The anthrax vaccine controversy: Questioning its efficacy, safety, and strategy. Military Medicine, Bethesda, MD: Assn. of Military Surgeons of the United States (AMSUS). 94. Hochfeld, W.L., ed. Panel summary from the 2000 Emerging Infectious Diseases Conference on Viral Vaccines and Adventitious Agents, Atlanta, Georgia, June, 2000. 95. Hoffman, D.R., A.L. Grossberg, and D. Pressman. 1972. Antihapten antibodies of restricted heterogeneity: Studies on binding properties and component chains. J. Immunol. 108:18. 96. Hollman, P.R. Consistent mixing: The key to uniform quality. Paper presented at conference in May, 1991, College of Engineering, Department of Engineering Professional Development, University of Wisconsin-Madison.

PH2270_book.fm Page 362 Tuesday, May 16, 2006 9:37 AM

362

Producing Biomolecular Substances

97. Hoogenraad, N., T. Helman, and J. Hoogenraad. 1983. The effect of pre-injection of mice with pristane on ascites tumour formation and monoclonal antibody production. J. Immunol. Methods 61:317. 98. Jablonski E., E.W. Moomaw, R.H. Tullis, et al. 1986. Preparation of oligodeoxynucleotide-alkaline phosphatase conjugates and their use as hybridization probes. Nucl. Acids Res. 14:6115–6128. 99. Jackson, A.T., and J. Lamb. 1981. Calculations in Food and Chemical Engineering. London: Macmillan Press 100. Jackson, A.T. 1985. Some problems of industrial scale up. J. Biol. Ed. 19:11–48. 101. Jackson, A.T. 1991. Process Engineering in Biotechnology (Int. Series in the Physical and Chemical Engineering Science), Englewood Cliffs, NJ: Prentice-Hall. 102. Johnson, R. 1990. Quality assurance of tissue culture media used in the biotechnology industry. BioPharm. 3(2):40–44. 103. Johnson, J.L. 1981. Genetic characterization, in Manual of Methods for General Bacteriology, 450–472, edited by P. Gerhardt, et al. Washington, D.C.: American Society for Microbiology. 104. Jeffreys, A.J. 1987. Highly variable mini-satellites and DNA fingerprints. Biochem. Soc. Trans. 15:309–317. 105. Jeffreys, A.J., N.J. Royle, V. Wilson, et al. 1988. Spontaneous mutation rates to new length alleles at tandem-repetitive hyper variable loci in human DNA. Nature 332:278–281. 106. Jeffreys, A.J., V. Wilson, and S.L. Thein. 1985. Hyper variable “mini-satellite” regions in human DNA. Nature 314:67–73. 107. Jeffreys, A.J., V. Wilson, and S.L. Thein. 1985. Individual specific “fingerprints” of human DNA. Nature 316:76–79. 108. Jeffreys, A.J., V. Wilson, S.L. Thein, et al. 1986. DNA “finger-prints” and segregation analysis of multiple markers in human pedigrees. Am. J. Hum. Genet. 39:11–24. 109. Jeffreys, A.J., V. Wilson, and S.L. Thein. 1985. Hyper variable “mini-satellite” regions in human DNA. Nature 314:67–73. 110. Jeffreys, A.J., V. Wilson, and S.L. Thein. 1985. Individual-specific “fingerprints” of human DNA. Nature 316:76–79. 111. Johnson, E.L. and R. Stevenson. 1978. Basic Liquid Chromatography. Palo Alto, CA: Varian. 112. Johnson, R. 1990. Quality assurance of tissue culture media used in the biotechnology industry. BioPharm. 3(2):40–44. 113. Johnson, J.L. 1981. Genetic characterization, in Manual of Methods for General Bacteriology, edited by P. Gerhardt, et al. Washington, D.C.: American Society for Microbiology, 450–472. 114. Johnstone, A. and R. Thorpe. 1982. Immunochemistry in Practice. Oxford: Blackwell Scientific. 115. Joner, E. and G. Christiansen. 1988. Hybridoma technology products: required virus testing. BioPharm. 1(8):50–55. 116. Kazazian, H.H., et al. 1982. Linkage of beta-thalassemia mutations and betaglobin gene polymorphisms with DNA polymorphisms in human betaglobin gene cluster. Nature 296:627–631. 117. Keller, G.H., C.U. Cumming, D-P. Huang, et al. 1988. A chemical method for introducing haptens onto DNA probes. Anal. Biochem. 170:441–450. 118. Kilmartin, J.V., B. Wright, and C. Milstein. 1982. Rat monoclonal antitubulin antibodies derived by using a new non-secreting rat cell line. J. Cell Biol. 93:576.

PH2270_book.fm Page 363 Tuesday, May 16, 2006 9:37 AM

Bibliography and Recommended Reading

363

119. Kohler, G. and C. Milstein. 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256:495–496. 120. Kumar, A., P. Tchen, F. Roullet, et al. 1988. Non-radioactive labeling of synthetic oligonucleotide probes with terminal deoxynucleotidyl transferase. Anal. Biochem. 169:376–382. 121. Laemmli, U.K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685. 122. Leary, J.J., D.J. Brigati, and D.C. Ward. 1983. Rapid and sensitive colorimetric method for visualizing biotin-labeled DNA probes hybridized to DNA or RNA immobilized on nitrocellulose: Bio-blots. Proc. Natl. Acad. Sci. USA 80:4045–4049. 123. Lewin, B. 1987. Genes, 3rd ed. New York: Wiley and Sons. 124. Lee, James H. 1991. Biochemical Engineering (Intl. Series in the Physical and Chemical Engineering Science), Englewood Cliffs, NJ: Prentice-Hall. 125. Levenspiel, O. 1958. Longitudinal mixing of fluids flowing in circular pipes. Ind. Eng. Chem. 50:343–346. 126. Levenspiel, O. 1972. Chemical Reaction Engineering. 2nd ed. New York: John Wiley and Sons. 127. Livingston, A.G., L.M.F. Dos Santos, P. Pavasant, E.N. Pistikopoulos, and L.F. Strachan. 1996. Detoxification of industrial wastewaters in an extractive membrane bioreactor. Wat. Sci. Tech. 33:1–8 Detoxification of industrial wastewater effluents are inefficient and costly treatments. The use of aerosols is presented here for the first time, and the overall process shows high treatment efficiency. 128. Lounes, M., J. Audet, J. Thibault, and A. LeDuy. 1995. Description and evaluation of reciprocating plate bioreactors. Bioprocess. Eng. 13:1–11. A unique impeller design that provides hom*ogeneous mixing without introducing high shear rate. 129. Maniatis, T., E.F. Friuch, and J. Sambrook. 1982. Molecular Cloning: A Laboratory Manual. New York: Cold Spring Harbor Laboratory. 130. Marino, M. 1989. Expression systems for heterologous protein transduction. BioPharm 2(7):1833. 131. Martin, J. and J.F. Gusella. 1986. Huntington’s disease: Pathogenesis and management. New Engl. J. Med. 315:1267–1276. 132. Martin, S., P. Soucaille, and J.S. Condoret. 1995. Bubble-free gaseous transfer in bioreactors using perfluorocarbons. Bioprocess. Eng. 13:293–300. 133. Matsudaira, P. 1987. Sequences from picomole quantities of proteins electro-blotted onto polyvinylidene difluoride membranes. J. Bio. Chem. 262:10035–10038. 134. Maxam, A.M. and W. Gilbert. 1977. A new method of sequencing DNA. Proc. Natl. Acad. Sci. USA 74:5463–5467. 135. McCabe, W.L., J.C. Smith, and P. Harriott. 1985. Unit Operations of Chemical Engineering. 4th ed. New York: McGraw-Hill. 136. McCracken, S. 1985. Preparation of RNA transcripts using SP6 RNA polymerase. Focus 7(2):5–8. 137. McGarrity, G.J. 1979. Detection of contamination, in Methods in Ecology, Vol. 58, No. 18, edited by W.B. Jakoby and I.H. Pastan. New York: Academic Press. 138. McNair, H.M. 1978. High-performance liquid chromatography equipment IV. J. Chromatogr. Sci. 588. 139. Michaels, J.D., A.K. Mallik, and E.T. Papoutsakis. 1996. Sparging and agitationinduced injury of cultured animal cells — Do cell-to-bubble interactions in the bulk liquid injure cells. Biotech. Bioeng. 51:399–409. 140. Michel, B.J. and S.A. Miller. 1962. Power requirements of gas-liquid agitated systems. Am. Inst. Chem. Eng. J. 8:262.

PH2270_book.fm Page 364 Tuesday, May 16, 2006 9:37 AM

364

Producing Biomolecular Substances

141. Milner, R.J. and J.G. Sutcliffe. 1988. Molecular biological strategies applied to the nervous system. Disc. Neurosci. 5(2):11–63. 142. Mishell, B.B. and S.M. Shiigi. 1980. Selected Methods in Cellular Immunology, 368. San Francisco: W.H. Freeman. 143. Millward, H.R., B.J. Bellhouse, I.J. Sobey, and R.W.H. Lewis. 1995. Enhancement of plasma filtration using the concept of the vortex wave. J. Membrane Sci. 100:121–129. 144. Millward, H.R., B.J. Bellhouse, and I.J. Sobey. 1996. The vortex wave membrane bioreactor:hydrodynamics and mass transfer. Chem. Eng. J. 62:175–181. 145. Mizrahi, A. and A. Lazar. 1988. Media for cultivation of animal cells: An overview. Cytotechnology 1:199–214. 146. Monaco, A.P. and L.M. Kunkel. 1987. A giant locus for the duch*enne and Becker muscular dystrophy gene. Trends Genet. 3:33–37. 147. Monod, J. 1942. Recherches sur des Croissances des Cultures Bacteriennes. 2nd ed. Paris: Hermann et Cie. 148. Morton, J.R., R.W. Hicks, and J.G. Fenic. 1976. How to design agitators for desired process response. Chem. Eng. 26:April. 149. Murakami, H., H. Masui, G. Sato, M. Sueoka, T. Chow, and T. Kano-Suloka. 1982. Growth of hybridoma cells in serum-free medium: Ethanolamine is an essential component. Proc. Natl. Acad. Sci. USA 79:1158. 150. Nathans, D. and H.O. Smith. 1975. Restriction endonucleases in the analysis and restructuring of DNA molecules. Annu. Rev. Biochem. 44:273–293. 151. Nakamura, Y., M. Leppert, P. O’Connell, et al. 1987. Variable number of tandem repeat (VNTR) markers for human gene mapping. Science 235:1616–1622. 152. Neoh, S.H., C. Gordon, A. Potter, and H. Zola. 1986. The purification of mouse monoclonal antibodies from ascitic fluid. J. Immunol. Methods 91:231. 153. Nevins, J.R. 1983. The pathway of eukaryotic mRNA formation. Annu. Rev. Biochem. 52:441–466. 154. Niimura, M., M. Kobayashi, and S. Kojima. 1992, Immunodiagnostic value of an a α-stichocyte-derived antigen isolated by affinity chromatography for trichinosis. Japanese J. Parasitol. 41(4):287–293. 155. Nugent, C.E., T. Gravius, P. Green, et al. 1988. Prenatal diagnosis of cystic fibrosis by chorionic villus sampling using 12 polymorphic deoxyribonucleotide markers. Obstet. Gynecol. 71:213–215. 156. Nygren, H. and M. Stenberg. 1985. Kinetics of antibody-binding to surface-immobilized antigen: Influence of mass transport on the enzyme-linked immunosorbent assay (ELISA). J. Colloid Interface Sci. 107:560–566. 157. Ohyama, Y. and K. Endoh. 1955. Power characteristics of gas-liquid contacting mixers. Chem. Eng. Japan 19:2. 158. O’Farrell, P. 1981. Replacement synthesis method of labeling DNA fragments. Focus 3(3):1–3. 159. Oldshue, J.Y. 1983. Fluid Mixing Technology. New York: McGraw-Hill. 160. Olson, J.C. 1993. Use of synthetic peptides and site-specific antibodies to localize a diphtheria toxin sequence associated with ADP-ribosyltransferase activity. J. Bacteriol. 175(3):898–901. 161. Orfaniotis, A., M. Lalane, N. Doubrovine, C. Fonade, and A. Moser. 1996. Oxygen transfer and scale up in bioreactors using hydro-ejectors for gas-liquid contacting. Bioprocess. Eng. 14:211–218. 162. Pelczar, M.J. and R.D. Reid. 1972. Microbiology. New York: McGraw-Hill.

PH2270_book.fm Page 365 Tuesday, May 16, 2006 9:37 AM

Bibliography and Recommended Reading

365

163. Palmiter, R.D., R.R. Behringer, C.J. Quaife, F. Maxwell, I.H. Maxwell, and R.L. Brinster. 1987. Cell lineage ablation in transgenic mice by cell-specific expression of a toxin gene. Cell 50:435–443. 164. Parker, C. and D. DiBiasio. 1987. Effect of growth rate and expression level on plasmid stability in saccharomyces cerevisiae. Biotech. Bioeng. 29:215–221. 165. Parvatiyar, M.G., R. Govind, and D.F. Bishop. 1996. Treatment of trichloroethylene (TCE) in a membrane biofilter. Biotechnol. Bioeng. 50:57–64. This study presents a hollow-fiber bioreactor where TCE contaminated air streams are treated by a combined aerobic-anaerobic biofilm. Because of oxygen diffusion limitation, part of the biofilm becomes anaerobic and reductive dechlorination of TCE is achieved, metabolites are biodegraded in the aerobic part of the biofilm. A mathematical model is presented. 166. Perlman, D. 1979. Use of antibiotics in cell-culture media, in Methods in Enzymology 58:110, edited by W.B. Jakoby and I.H. Pastan. New York: Academic Press. 167. Piatyszek, M.A., A. Jarmolowski, and J. Augustyniak. 1988. Iodogen-mediated radioiodination of nucleic acids. Anal. Biochem. 172:356–359. 168. Podell, D.N. and G.N. Abraham. 1978. A technique for the removal of pyroglutamatic acid from the amine terminus of proteins using calf liver pyroglutamate amino peptidase. Biochem. Biophy. Res. Comm. 81(1):176–185. 169. Popko, B., C. Puckett, E. Lai, H.D. Shine, C. Readhead, N. Takahashi, S.W. Hunt, R.L. Prober, J.M., G.L. Trainor, R.J. Dam, et al. 1987. A system for rapid sequencing with fluorescent chain-terminating dideoxy-nucleotides. Science 238:336–341. 170. Proudfoot, N.J., A. Gil, T. Maniatis. 1982. The structure of the human zeta-globin gene and a closely linked, nearly identical pseudogene. Cell 31:553–563. 171. Putnam, J.E. 1987. In Commercial Production of Monoclonal Antibodies: A Guide for Scale Up, 119–138, edited by S.S. Seaver. New York: Marcel Dekker. 172. Quesnel, L.B. 1987. Sterilization and sterility, in Basic Biotechnology, 197–215, edited by J. Bulock and B. Kristiansen. London: Academic Press. 173. Renz, M. and C. Kurz. 1984. A colorimetric method for DNA hybridization. Nucl. Acids Res. 12:3435–3444. 174. Reuveny, S. 1985. In Advances in Cell Culture, 4:213–247, edited by K. Maramorosch. New York: Academic Press. 175. Reynolds, O. 1886. On the dynamical theory of incompressible viscous fluids and the determination of the criteria. Phil. Trans. R. Soc. Lond. 177:157. 176. Richards, J.W. 1961. Studies in aeration and agitation. Prog. Ind. Microbiol. 3:143. 177. Rigby, P. W.J., M. Dieckmann, C. Rhodes, et al. 1977. Labeling deoxyribonucleic acid to high specific activity in vitro by nick translation with DNA polymerase I. J. Mol. Biol. 113:237–251. 178. Roberts, R.J. 1981. Restriction and modification enzymes and their recognition sequences. Nucl. Acids Res. 9:75–96. 179. Roberts, P.L. 1985. Comparison of fluorographic methods for detecting radioactivity in polyacrylamide gels or on nitrocellulose filters. Anal. Biochem. 147:521–524. 180. Roca, E., J. Flores, M.J. Núñez, and J.M. Lema. 1996. Ethanolic fermentation by immobilized Saccharomyces cerevisiae in a semipilot pusing packed-bed bioreactor. Enzyme Micro. Tech. 19:132–139. 181. Rosell, C.M. and A.M. Vaidya. 1996. Twin-core packed-bed reactors for organicphase enzymatic esterification with water activity control. Appl. Microbiol. Biotechnol. 44:283–286.

PH2270_book.fm Page 366 Tuesday, May 16, 2006 9:37 AM

366

Producing Biomolecular Substances

182. Rosell, C.M., A.M. Vaidya, and P.J. Halling. 1996. Continuous in situ water activity control for organic phase biocatalysis in a packed bed hollow fiber reactor. Biotechnol. Bioeng. 49:284–289. 183. Roychoudhury, R. and R. Wu. 1980. Terminal transferase-catalyzed addition of nucleotides to the 3′ termini of DNA, in Methods in Enzymology, edited by L. Grossman and K. Moldave, Vol. 65, Part I:43–62. New York: Academic Press. 184. Ruddle, F.H. 1981. A new era in mammalian gene mapping: Somatic cell genetics and recombinant DNA. Nature 294:115–120. 185. Rupp, R.G. 1985. In Large-scale Mammalian Cell Culture, 19–38, edited by W.R. Tolbert and J. Feder. New York: Academic Press. 186. Rushton, J.H. 1951. The use of pilot plant mixing data. Chem. Eng. Prog. 47:467. 187. Sanger, F. and A.R. Coulson. 1975. A rapid method for determining sequences in DNA by primed synthesis with DNA polymerase. J. Mol. Biol. 94:444–448. 188. Sanger, F., S. Nicklen, and R.A. Coulsen. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463–5467. 189. Schneider, Y., X. Veithen, and A. Lavoix. 1990. Hybridoma cell growth and monoclonal antibody secretion upon a continuous cultivation in bioreactor with a serumand protein-free nutritive medium. Biotech. Lett. 12:481–486. 190. Schwartz, L.S., N.B. Jansen, N.W.Y. Ho, and G.T. Tsao. 1988. Plasmid instability kinetics of the Yeast S288C pUCKm8 [cir+] in nonselective and selective media. Biotech. Bioeng. 32:733–740. 191. Scopes, R.K. 1982. Protein Purification. New York: Springer-Verlag. 192. Seller, P., A. J. Jeffreys, V. Silson, et al. 1984. Organization of the human myoglobin gene. Embolism J. 3:439–446. 193. Sidman and L. Hood. 1987. Myelin-deficient mice: Expression of myelin-basic protein and generation of mice with varying levels of myelin. Cell 48:713–721. 194. Sieder, E.N. and G.E. Tate. 1936. Heat transfer and pressure drop of liquids in tubes. Ind. Eng. Chem. 28:1429. 195. Smith, L.M., S. Fung, M.W. Hunkapiller, et al. The synthesis of oligonucleotides containing an aliphatic amino group at the 5’ terminus: Synthesis of fluorescent DNA primers for use in DNA sequence analysis. Nucl. Acids Res. 13:2399–2412. 196. Snyder, L.R. and J.J. Kirkland. 1979. Introduction to Modern Liquid Chromatography, 2nd ed. New York: John Wiley and Sons. 197. Southern, E.M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503–517. 198. Southern, E. 1979. Gel electrophoresis of restriction fragments. Methods Enzymol. 68:152–176. 199. Steinberg, R.A. 1985. The molecules of life. Sci. Am. 253:48–57. 200. Strauss, W. 1975. Industrial Gas Cleaning. 2nd ed. Oxford: Pergamon Press Ltd. 201. Su, Xinzhuan, D.D. Morris, N.A. Crowe, J.N. Moore, K.J. Fischer, and R.A. McGraw. 1992. Equine tumor necrosis factor alpha: Cloning and expression in Escherichia coli. Generation of monoclonal antibodies and development of a sensitive enzymelinked immunosorbent assay. Hybridoma 11(6):715–728. 202. Sutcliffe, J.G., R.J. Milner, T.M. Shinnick, and F.E. Bloom. 1983. Identifying the protein products of brain-specific genes with antibodies to chemically synthesized peptides. Cell 33:671–682. 203. Sutcliffe, J.G., R.J. Milner, J.M. Gottesfeld, and W. Reynolds. 1984. Control of neuronal gene expression. Science 225:1308–1314. 204. Syman, A.R., and R. sh*te. 1980. A highly polymorphic locus in human DNA. Proc. Natl. Acad. Sci. USA 77:6754–6758.

PH2270_book.fm Page 367 Tuesday, May 16, 2006 9:37 AM

Bibliography and Recommended Reading

367

205. Tax, W.J.M., et al. 1984. Fc Receptor for mouse IgG1 on human monocytes: Polymorphism and mold in antibody-induced T cell proliferation. J. Immunol. 133:1185–1189. 206. Tchen, P., R.P. Fuchs, E. Stye, et al. 1984. Chemically modified nucleic acids as immuno-detectable probes in hybridization experiments. Proc. Natl. Acad. Sci. USA 81:3466–3470. 207. Thomas, K.R., and M.R. Capecchi. 1987. Site-directed mutagenesis by gene targeting in mouse embryo-derived stem cells. Cell 51:503–512. 208. Tisseyre, B., J.C. Coquille, and P. Gervais. 1995. Conception and characterization of a continuous plug flow bioreactor. Bioprocess. Eng. 13:113–118. Several aspects of the newly developed bioreactor including KLa, contamination between loops and liquid mixing are described. 209. Treyball, R.E. 1985. Mass Transfer Operations. 3rd ed. New York: McGraw-Hill. 210. Underwood, P.A. and P. Bean. 1985. The influence of methods of production, purification and storage of monoclonal antibodies upon their observed specificities. J. Immunol. Methods 80:189. 211. Van Riet, K. 1979. Review of measuring methods and results in non-viscous gasliquid mass transfer in stirred vessels. Ind. Eng. Chem. Process Des. Dev. 18:357. 212. Viney, J.L., H.M. Prosser, C.R.A. Hewitt, J.R. Lamb, and M.J. Owen. 1992. Generation of monoclonal antibodies against a human T cell receptor a-chain expressed in transgenic mice. Hybridoma 11(6):701–714. 213. von Essen, John A. 1983. Liquid mixing: Scale-up procedures. Presented at InterAmerican Congress and VII Chilean Congress of Chemical Engineering, 6–11 November, Santiago, Chile. 214. von Wedel, R.J. 1987. In Commercial Production of Monoclonal Antibodies: A Guide for Scale Up, 175–195, edited by S.S. Seaver. New York: Marcel Dekker. 215. Wang, S.J. and J.J. Zhong. 1996. A novel centrifugal impeller bioreactor I. Fluid circulation, mixing, and liquid velocity profiles. Biotech. Bioeng. 51:511–519. 216. Wang, S.J. and J.J. Zhong. 1996. A novel centrifugal impeller bioreactor II. Oxygen transfer and power consumption. Biotech. Bioeng. 51:20–527. 217. Watson, J.D. and F.H.C. Crick. 1953. Molecular structure of nucleic acids: A structure for deoxyribose nucleic acid. Nature 171:737–738. 218. Watson, J.D., J. Tooze, and D.T. Kurtz. 1983. Recombinant DNA: A Short Course. New York: W.H. Freeman. 219. Watson, J.D., N.H. Hopkins, J.W. Roberts, J.A. Steitz, and A.M. Weiner. 1987. Molecular Biology of the Gene. 4th ed. New York: Benjamin-Cummings. 220. Wauon, J.D. and F.H.C. Crick. 1953. General implications of the structure of deoxyribonucleic acid. Nature 171:964–967. 221. Wehner, J.F. and R.H. Wilhelm. 1956. Boundary conditions of flow reactors. Chem. Eng. Sci. 6:89–93. 222. Weintraub, H.M. 1990. Antisense RNA and DNA. Sci. Am. 262(1):40–46. 223. Weissman, D., D.J. Parker, T.L. Rothstein, and A. Marshak-Rothstein. 1985. Methods for the production of xenogeneic monoclonal antibodies in murine ascites. J. Immunol. 135:1001. 224. Whiteway, M.S. and A. Ahmed. 1984. Recombinational instability of chimeric plasmid in saccharomyces cerevisiae. Mol. Cell Biol. 4(1):195–198. 225. Wong, Z., V. Wilson, J. Jeffreys, et al. 1986. Cloning a selected fragment from a human DNA “fingerprint”: Isolation of an extremely polymorphic mini-satellite. Nucl. Acids Res. 14:4605–4616.

PH2270_book.fm Page 368 Tuesday, May 16, 2006 9:37 AM

368

Producing Biomolecular Substances

226. Wood, W.G., and A. Gadow. 1983. Immobilization of antibodies and antigens on macro solid phases — A comparison between adsorptive and covalent binding. J. Clin. Chem. Clin. Biochem. 21:789–797. 227. Work, T.S. and E. Work, eds. Laboratory Techniques in Biochemistry and Molecular Biology. Amsterdam: Elsevier/North-Holland. 228. Wyman, A.R. and R. White. 1980. A highly polymorphic locus in human DNA. Proc. Natl. Acad. Sci. USA 77:6754–6758. 229. Yee, D., E. Yankelevich, V. Bystritskii, and T. Wood. A dual-treatment system for the degradation of 2-4 dichlorophenol which utilizes pulsed-electric discharge reactor and bioremediation. Paper 103e presented at the AIChE 1996 Annual Meeting, Chicago, November 10–15. This paper presents preliminary results of a combined pulsed-electric discharge reactor (PED) and a bioreactor. PED cannot usually be applied to water effluents because of inefficient and costly treatment. The use of aerosols is presented for the first time. The overall process showed high treatment efficiency. 230. Yunis, J.J. 1976. High resolution of human chromosomes. Science 191:1268–1270. 231. Zola, H., I.G.R. Beckman, J. Bradley, D.A. Brooks, A. Kupa, P.J. McNamara, I.J. Smart, and M. Thomas. 1980. Human lymphocyte markers defined by antibodies derived from somatic cell hybrids III. A marker defining a subpopulation of lymphocytes which cuts across the normal T-B-null classification. Immunology 40:143.

RECOMMENDED READING TOPICS AIRLIFT BIOREACTORS 1. Chisti, Y., Airlift Bioreactors, Elsevier, London, 1989, p. 355. 2. Chisti, Y., Appl. Mech. Rev., 51, 33–112 (1998). Pneumatically agitated bioreactors in industrial and environmental bioprocessing: Hydrodynamics, hydraulics and transport phenomena. 3. Chisti, Y. and M. Moo-Young, Chem. Eng. Prog., 89(6), 38–45 (1993). Improve the performance of airlift reactors. 4. Chisti, Y. and U.J.Jauregui-Haza, Biochem. Eng. J., 10, 143–153 (2002). Oxygen transfer and mixing in mechanically-agitated airlift bioreactors. 5. Rubio, F.C., J.L Garcia, E. Molina, and Y. Chisti, Chem. Eng. J., 84, 43–55 (2001). Axial inhom*ogeneities in steady-state dissolved oxygen in airlift bioreactors: predictive models. 6. Wenge, F., Y. Chisti, and M. Moo-Young, Chem. Eng. Comm., 155, 19–44 (1996). Split-cylinder airlift reactors: Hydraulics and hydrodynamics of a new mode of operation.

BIOKINETICS, CONTROL, ETC. 1. Allsop, P.J., Y. Chisti, M. Moo-Young, and G.R. Sullivan, Biotechnol. Bioeng., 41, 572–580 (1993). Dynamics of phenol degradation by Pseudomonas putida. 2. Banerjee, U. C., Y. Chisti, and M. Moo-Young, Biotechnol. Tech., 7, 313–316 (1993). Spectrophotometric determination of mycelial biomass. 3. Chisti, Y. and M. Moo-Young, Trans. I. Chem. E., 74A, 575-583 (1996), Bioprocess intensification through bioreactor engineering.

PH2270_book.fm Page 369 Tuesday, May 16, 2006 9:37 AM

Bibliography and Recommended Reading

369

4. Oh, G., M. Moo-Young, and Y. Chisti, Biochem. Eng. J., 1, 211–217 (1998). Automated fed-batch culture of recombinant Saccharomyces cerevisiae based on on-line monitored maximum substrate uptake rate. 5. Wenge, F., Y. Chisti, and M. Moo-Young, Ind. Eng. Chem. Res., 34, 928–935 (1995). A new method for the measurement of solids holdup in gas-liquid-solid three-phase systems. 6. Zhang, Z., M. Moo-Young, and Y. Chisti, Biotechnol. Adv., 14, 401–435 (1996). Plasmid stability in recombinant Saccharomyces cerevisiae.

BIOMASS

AND

SECONDARY METABOLYTES

1. Banerjee, U.C., B. Saxena, and Y. Chisti, Biotechnol. Adv., 10, 577–595 (1992). Biotransformations of rifamycins: Process possibilities. 2. Ejiofor, A.O., Y. Chisti, and M. Moo-Young, J. Ind. Microbiol., 16, 102–109 (1996). Fed-batch production of baker’s yeast using millet (Pennisetum typhoides) flour hydrolysate as the carbon source. 3. Ejiofor, A.O., Y. Chisti, and M. Moo-Young, Enzyme Microb. Technol., 18, 519–525 (1996). Culture of Saccharomyces cerevisiae on hydrolysed waste cassava starch for production of baking-quality yeast. 4. Moo-Young, M., Y. Chisti, and D. Vlach, Biotechnol. Lett., 14, 863–868 (1992). Fermentative conversion of cellulosic substrates to microbial protein by Neurospora sitophila. 5. Sharma, R., Y. Chisti, and U.C. Banerjee, Biotechnol. Adv., 19, 627–662 (2001). Production, purification, characterization, and applications of lipases. 6. Tamer, I.M. and Y. Chisti, Enzyme Microb. Tech., 29, 611–620 (2001). Production and recovery of recombinant protease inhibitor alpha-1-antitrypsin.

BIOPLASTICS 1. Grothe, E., M. Moo-Young, and Y. Chisti, Enzyme Microb. Tech., 25, 132–141 (1999). Fermentation optimization for the production of poly (beta-hydroxybutyric acid) microbial thermoplastic. 2. Tamer, I.M., M. Moo-Young, and Y. Chisti, Bioprocess Eng., 19, 459–468 (1998). Optimization of poly (beta-hydroxybutyric acid) recovery from Alcaligenes latus: Combined mechanical and chemical treatments. 3. Tamer, I.M., M. Moo-Young, and Y. Chisti, Ind. Eng. Chem. Res., 37, 1807–1814 (1998). Disruption of Alcaligenes latus for recovery of poly (beta-hydroxybutyric acid): Comparison of high-pressure hom*ogenization, bead milling, and chemically induced lysis.

BIOPROCESS ENGINEERING 1. Chisti, Y. and M. Moo-Young, Fermentation technology, bioprocessing, scale-up, and manufacture in Biotechnology: The Science and the Business, 2nd edition, Moses, V., R.E. Cape, and D.G. Springham, eds., Harwood Academic Publishers, New York, 1999, pp. 177–222. 2. Chisti, Y., Fermentation (industrial): Basic considerations, in Encyclopedia of Food Microbiology, Robinson, R., C. Batt, and P. Patel, eds., Academic Press, London, 1999, pp. 663–674.

PH2270_book.fm Page 370 Tuesday, May 16, 2006 9:37 AM

370

Producing Biomolecular Substances 3. Chisti, Y. and M. Moo-Young, On bioreactor design in Basic Biotechnology, 2nd edition, Ratledge, C., and B. Kristiansen, eds., Cambridge University Press, Cambridge, 2001, pp. 151–171. 4. Chisti, Y. and M. Moo-Young, On bioreactors, in Encyclopedia of Physical Science and Technology, Vol. 2, Meyers, R. A., ed., Academic Press, San Diego, 2002, pp. 247–271.

ENGINEERING

FOR

MONOSEPTIC OPERATIONS

1. Chisti, Y., Chem. Eng. Prog., 88(1), 55–58 (1992). Build better industrial bioreactors. 2. Chisti, Y., Chem. Eng. Prog., 88(9), 80–85 (1992). Assure bioreactor sterility. 3. Chisti, Y., Modern systems of plant cleaning, in Encyclopedia of Food Microbiology, Robinson, R., C. Batt, and P. Patel, eds., Academic Press, London, 1999, pp. 1806–1815.

GENERAL 1. Chisti, Y. and M. Moo-Young, Fermentation technology, bioprocessing, scale up and manufacture, in Biotechnology: The Science and the Business, second edition, Moses, V., R.E. Cape, and D.G. Springham, eds., Harwood Academic Publishers, New York, 1999, pp. 177–222. 2. Chisti, Y., Fermentation (industrial): Basic considerations in Encyclopedia of Food Microbiology, Robinson, R., C. Batt, and P. Patel, eds., Academic Press, London, 1999, pp. 663–674. 3. Chisti, Y. and M. Moo-Young, On bioreactor design in Basic Biotechnology, 2nd ed., Ratledge, C., and B. Kristiansen, eds., Cambridge University Press, Cambridge, 2001, pp. 151–171. 4. Chisti, Y. and M. Moo-Young, On bioreactors in Encyclopedia of Physical Science and Technology, vol. 2, Meyers, R.A., ed., Academic Press, San Diego, 2002, pp. 247–271.

HEAT TRANSFER 1. Ouyoung, P.K., M.Y. Chisti, and M. Moo-Young, Chem. Eng. Res. Des. 67, 451–456 (1989). Heat transfer in airlift reactors.

MASS TRANSFER

AND

HYDRODYNAMICS

1. Chisti, Y. and M. Moo-Young, Biotechnol. Bioeng., 31, 487–494 (1988). Hydrodynamics and oxygen transfer in pneumatic bioreactor devices. 2. Chisti, Y., B. Halard, and M. Moo-Young, Chem. Eng. Sci., 43, 451–457 (1988). Liquid circulation in airlift reactors. 3. Contreras, A., Y. Chisti, and E. Molina, Chem. Eng. Sci., 53, 41514154 (1998). A reassessment of relationship between riser and downcomer gas holdups in airlift reactors. 4. Chisti, Y., Mass transfer, in Encyclopedia of Bioprocess Technology: Fermentation, Biocatalysis, and Bioseparation, Vol. 3, Flickinger, M.C. and Drew, S.W., eds., Wiley, New York, 1999, pp. 1607–1640.

PH2270_book.fm Page 371 Tuesday, May 16, 2006 9:37 AM

Bibliography and Recommended Reading

371

5. Chisti, Y. and M. Moo-Young, Fermentation technology, bioprocessing, scale-up and manufacture in Biotechnology: The Science and the Business, 2nd ed., Moses, V., R.E Cape, and D.G. Springham, eds., Harwood Academic Publishers, New York, 1999, pp. 177–222.

OTHER BIOREACTORS 1. Chisti, Y. and M. Moo-Young, J. Chem. Technol. Biotechnol., 58, 331–336 (1993). Aeration and mixing in vortex bioreactors. 2. Chisti, Y. and M. Moo-Young, Trans. I. Chem. Eng., 71C, 209214 (1993). Airlift bioreactors with packed beds of immobilized biocatalysts: Theoretical evaluation of the liquid circulation performance. 3. Chisti, Y., Bioprocess. Eng., 9, 191–196 (1993); Process Biochem., 28, 511–517 (1993). Animal cell culture in stirred bioreactors: Observations on scale up. 4. Chisti, Y., Solid substrate fermentations, enzyme production, and food enrichment, in Encyclopedia of Bioprocess Technology: Fermentation, Biocatalysis, and Bioseparation, Vol. 5, Flickinger, M.C. and S.W. Drew, eds., Wiley, New York, 1999, pp. 2446–2462. 5. Chisti, Y. and U.J. Jauregui-Haza, Biochem. Eng. J., 10, 143–153 (2002). Oxygen transfer and mixing in mechanically-agitated airlift bioreactors.

PHOTOBIOREACTORS 1. Mirón, A.S., A.C. Gómez, F.G. Camacho, E.M. Grima, and Y. Chisti, J. Biotechnol., 70, 249–270 (1999). Comparative evaluation of compact photobioreactors for largescale monoculture of microalgae. 2. Grima, E.M., F.G.A. Fernández, F.G. Camacho, and Y. Chisti, J. Biotechnol., 70, 231–247 (1999). Photobioreactors: light regime, mass transfer, and scale up. 3. Molina, E., J. Fernández, F.G. Acién, and Y. Chisti, J. Biotechnol., 92, 113–131 (2001). Tubular photobioreactor design for algal cultures. 4. Acién Fernández, F.G., J.M. Fernández Sevilla, J.A. Sánchez Pérez, E. Molina Grima, and Y. Chisti, Chem. Eng. Sci., 56, 2721–2732 (2001). Airlift-driven external-loop tubular photobioreactors for outdoor production of microalgae: Assessment of design and performance. 5. Molina, E., F.G Acién Fernández, F. García Camacho, F. Camacho Rubio, and Y. Chisti, J. Appl. Phycol., 12, 355–368 (2000). Scale up of tubular photobioreactors.

PROTEINS

AND

ENZYMES

1. Rouf, S.A., M. Moo-Young, and Y. Chisti, Biotechnol. Adv., 14, 239–266 (1996). Tissue-type plasminogen activator: Characteristics, applications, and production technology. 2. Garrido, F., U.C. Banerjee, Y. Chisti, and M. Moo-Young, Bioseparation, 4, 319–328 (1994). Disruption of a recombinant yeast for the release of beta-galactosidase. 3. Zhang, Z., Y. Chisti, and M. Moo-Young, Bioseparation, 5, 329–337 (1995). Isolation of a recombinant intracellular beta-galactosidase by ammonium sulfate fractionation of cell hom*ogenates.

PH2270_book.fm Page 372 Tuesday, May 16, 2006 9:37 AM

372

Producing Biomolecular Substances 4. Chisti, Y., Solid substrate fermentations, enzyme production, and food enrichment, in Encyclopedia of Bioprocess Technology: Fermentation, Biocatalysis, and Bioseparation, Vol. 5, Flickinger, M.C. and S.W. Drew, eds., Wiley, New York, 1999, pp. 2446–2462. 5. Tamer, I. M. and Y. Chisti, Enzyme Microb. Technology, 29, 611–620 (2001). Production and recovery of recombinant protease inhibitor alpha-1-antitrypsin. 6. Sharma, R., Y. Chisti, and U.C. Banerjee, Biotechnol. Adv., 19, 627–662 (2001). Production, purification, characterization, and applications of lipases.

SHEAR EFFECTS 1. Moo-Young, M. and Y. Chisti, Biotechnology, 6(11), 1291–1296 (1988). Considerations for designing bioreactors for shear-sensitive culture. 2. Chisti, Y. and M. Moo-Young, Biotechnol. Bioeng., 34, 1391–1392 (1989). On the calculation of shear rate and apparent viscosity in airlift and bubble column bioreactors. 3. Zhang, Z., Y. Chisti, and M. Moo-Young, J. Biotechnol., 43, 33–40 (1995). Effects of the hydrodynamic environment and shear protectants on survival of erythrocytes in suspension. 4. García Camacho, F., E. Molina Grima, W. Sánchez Mirón, V. González Pascual, and Y. Chisti, Enzyme Microb. Technology, 29, 602–610 (2001). Carboxymethyl cellulose protects algal cells against hydrodynamic stress. 5. Chisti, Y., Crit. Rev. Biotechnol., 21, 67–110 (2001). Hydrodynamic damage to animal cells. 6. Chisti, Y., Trends Biotechnol., 18, 420–432 (2000). Animal cell damage in sparged bioreactors.

PH2270_IDX.fm Page 373 Tuesday, May 23, 2006 10:02 AM

Index A Adaptive controls, 168, 237 Aflatoxins, 101, 124, 126 Alkaline phosphatase, 76 Altered immunogeneity, 129 Amino acids, 104, 114–115 symbols and abbreviations, 333 Amphotericin B, 105, 215 Ampicillin, 105 Amplification by DNA synthesis, 85–87 by ligation, 90–92 reagents, 85–95 manufacturers, 95 by RNA replication, 92–94 by RNA transcription, 87–90 sequence-based, 94 Anthrax, 124 Antibiotics, 104–105, 115, 123, 214–215, 222 as contaminants, 101 peptide, 332 Antibody(ies), 207 affinity, 208–209 affinity-purified, 210–211 immunoassay, 210 monoclonal. See Monoclonal antibody(ies) OKT3, 223 polyclonal, 210 Antigen–antibody binding, 207–208 Antiseptic chemicals, 170 Ascites fluid MAB production in, 213–224 adventitious agents for, 221–223 antibiotics used, 214–215 bioreactor vs., 223–224 host animals for, 214 injecting hosts, 215 porous microcarriers for, 219–221 sacrificing hosts for, 216 suspension vs. anchorage-dependent cells for, 219 Aspergillus nidulans, 80 Attachment factors, 105, 115

Automatic control systems, 167, 236, 237 Avian myeloblastosis virus, 87

B Bacillus anthracis, 124–125 Bacillus stearothermophilus, 172 Bacillus subtilis, 125 Bacteria, 1, 9, 219, 287 as biocatalysts, 121 cloned, 16, 43 critical oxygen concentration, 290 in DNA synthesis, 329 expression, 20, 49 genetically altered, 13 gram-negative, 20, 49, 105, 112, 128, 215 gram-positive, 22, 51, 105, 129, 215 growth characteristics, 49 growth hormone yield, 345 killing, 170, 174 peptide expression, 338, 345 plasmid, 13, 39 recombinant proteins, 128 as serum contaminant, 101 Baculovirus infection, 20, 56, 58, 131 Batch culture, 140, 150, 188, 239, 248, 267, 275 change in cell concentration, 266, 300 growth cycle, 264 accelerated stage, 264 death, 264 decelerated stage, 264 latent stage, 264 static period, 264 performance equation, 139, 300 simulation curve, 191, 242 Bifurcation analysis, 321 Bio-weapons of mass destruction, 124 Biodrugs, 28 Biological license application, 36 Biomolecular foundations, 9 Bioprocess(es), 5–7 biomolecular products, 29 cell-culture process, 6 cGMP standards, 28

373

PH2270_IDX.fm Page 374 Tuesday, May 23, 2006 10:02 AM

374

Producing Biomolecular Substances

complex mixtures, 29 contract manufacturing, 28 detailed specifications, 33 differing features at larger scale, 32 downscaling to scale up, 30 downstream, 33 fermentation procedure, 7 increasing scale, 32 intracellular vs. extracellular, 57 joint manufacturing, 29–30 kinetic models, 28 lab-to-pilot-plant-to-production, 31 large-scale, 31 cultures, 21 optimizing product yield, 27 purification, 7 quality assurance, 8 recovery, 7 scaling-up in-house or in a contract facility, 31 separation, 7 simulating environment at scale, 32–33 virus and foreign-DNA removal, 7–8 Bioprocessing aseptic conditions for, 131, 206 downstream, 33–34 facility operations, 8 reagents used, 8 serum-free media, 102 techniques, 1 Bioproduction, 1, 4–5. See also Bioprocessing Bioreactor(s), 201–323 adaptive controls, 237 aseptic conditions, 206 automatic control systems, 236, 237 batch, cell-growth modeling in, 314–315 bench top, 206 calculating multiple steady states, 317 cell growth thermodynamics, 243–248, 283–285 del factor, 247–248 extracellular product heat release, 284–285 heat release due to growth, 243–245, 283–284 heat release from extracellular products, 245 heating, 246–247 modeling bioreactor sterilization, 245 cell yield, 242–243 complex controls, 237 containment vessels, 205 continuous-culture, 320 dilution rate, 152 maximizing productivity, 152

multiplicity and steady state stability in, 317 continuous dynamics, 315–317 control ranges, 234 derivative controllers, 237 design, 203, 302–304 program, 321–323 dynamics, 232–233 engineering and design, 224–232 basic physical elements, 224–227 complex physical elements, 227–230 heat generation, 230–232 fed batch reactor dynamics, 320 feed-forward controllers, 237 feedback controllers, 236 fermenters vs., 202 growth kinetics, 248–257, 262–282, 285–287 factors affecting growth rate, 253, 290 metabolic quotient and rate expression, 252–253 product formation kinetics and, 253–257, 290–292 specific growth rate value, 250–252 heat generation, 230–232 increased size vs. complexity, 205–206 integral controllers, 236 kinetic modeling, 238–243, 305–309 cell yield, 242–243 stoichiometric coefficient measurements, 243 structured models, 238–242 manufacturers, 323–324 maximizing productivity, 310–311 metabolic quotient and rate expression, 288–290 mixing and agitation, 204–205 modeling eukaryotic cell growth and production, 262–282 cell composition, 277 cell yield, 279–282 factors affecting specific growth rate, 270–271 growth reaction, 277–278 kinetic growth, 262–269, 297–301 kinetic models, 238–243, 271–275, 305–309 maximizing productivity, 276–277 optimal conditions, 275–276 stoichiometric coefficient measurements, 277, 282 monoclonal antibody production, 206–224 as multivariate systems, 232 nonlinear dynamics, 232 operating modes, 202 optimal conditions, 310

PH2270_IDX.fm Page 375 Tuesday, May 23, 2006 10:02 AM

Index oxygen balance, 260–261, 294–295 oxygen transfer in, 257–262, 292–297 factors affecting mass-transfer coefficient, 261–262, 295–297 metabolic oxygen demand and, 257, 292 volumetric oxygen mass-transfer coefficient, 257–260, 292–294 pilot-scale, 206 process control components, 236 functions, 233 system design, 232–233, 236 process measurement and control, 233–235 process monitoring, 235–237 adaptive controls, 237 automatic control systems, 236, 237 complex controls, 237 derivative controllers, 237 feed-forward controllers, 237 feedback controllers, 236 in-line sensors, 235 integral controllers, 236 off-line sensors, 236 on-line sensors, 235 process control, 236 production-scale, 206 recombinant culture kinetics, 311–314 scale-up, 216, 226 size and scale, 203–204 specific growth rate, 287–288 steady state stability, 317 stirred-tank, 204 closed-loop linear stability analysis of, 319–320 comparison, 204 open-loop linear stability analysis of, 318–319 phase plane analysis of, 319 proportional control, with Monod, 317–318 structure and configuration, 202–203 typical characteristics, 203 Biotechnology applications, 1 defined, 1 focus of, 1 historical perspectives, 1, 2–4 Botulinum, 124 Bovine papillomavirus, 22 Bovine serum, 21, 50, 100 Buffers, 105–108, 116 Earle’s balanced salt solution, 107 Hanks’ balanced salt solution, 107 HEPES buffer, 107–108

375 L-15 medium, 107 sodium bicarbonate, 106–107

C Calcitonin, 344 Camel-Pox virus, 126 Cavitation, 66 Cell(s) eukaryotic, 1, 46 cell-culture media, 97–98, 217–218 DNA, 72 expression systems, 21 extracellular expression, 53, 130 genes designer, 18 exons, 17, 45 expression, 44, 80 growth characteristics, 50–51 intracellular expression, 52 introns, 20, 49 modeling growth and production, 262–282 noncoding sequences, 20, 49 organization, 12 protein coding sequence, 20, 49 structure, 71 growth, 27, 277–278 cell composition, 277 conditions, 275–276 factors affecting rate, 270–271 kinetic, 262–269, 297–301 maximizing, 276–277 models, 238–243, 271–275, 305–309 stoichiometric coefficient, 277, 282 yield, 279–282 manufacturers of cell lines, 61 products derived from, 8 growth, 49–51, 134–151 thermodynamics, 243–248, 283–285 del factor, 247–248 extracellular product heat release, 284–285 heat release due to growth, 243–245, 283–284 heat release from extracellular products, 245 heating, 246–247 modeling bioreactor sterilization, 245 prokaryotic, 1, 24, 27, 56, 157 culture medium, 97 enzymes, 129 gene designer, 18, 46

PH2270_IDX.fm Page 376 Tuesday, May 23, 2006 10:02 AM

376 expression, 20, 56 introns, 49 organization, 12 structure, 71 growth, 50, 57 intracellular expression, 53 manufacturers of cell lines, 60–61 variations, 9–10 Cell-culture defined, 201 large-scale, 5, 21, 28 media, 97–119 attachment factors, 105, 115–116 buffers, 105–108, 116–117 Earle’s balanced salt solution, 107 Hanks’ balanced salt solution, 107 HEPES buffer, 107–108 L-15 medium, 107 sodium bicarbonate, 106–107 defined, 98, 102–103 epidermal growth factor, 109 eukaryotic, 97–98, 217–218 fibroblast growth factor, 110 gels, 103 growth factors, 109–110, 117–118 lectins, 110–111, 118 manufacturers, 112–119 prokaryotic, 97 serum, 98–102, 114 bovine, 100 contamination of, 101–102 supplements, 104–105, 114–115 amino acids, 104, 114–115 amphotericin B, 105 ampicillin, 105 antibiotics, 104–105 tylosin solution, 105 toxins, 111, 118 transport factors, 111, 119 vitamins, 111–112, 119 water, 112, 119 process, 6–8 quality assurance, 8 separation, recovery, and purification, 7 virus and foreign-DNA removal, 7–8 Cell lines, 5–6, 19, 47–48 Cell yield, 242–243, 279–282. See also Yield optimizing, 27 Cellular production, 201 cGMP standards, 28 Chemostatic theory, 150, 275 Cleaning, 151, 275. See also Sterilization containment vessel, 205 offline, 175 time, 262, 296, 309

Producing Biomolecular Substances Cleaning-in-place system, 168 threaded fittings, 169 Clone libraries, 16, 44 Cloning, 13–16, 39–41 applications, 15 screening and selection, 15–16 Cloning vectors manufacturers, 59–61 Clostridium botulinum, 125 Clostridium perfringens, 125 Cofactors, 66 Continuous culture systems, 21, 50, 128, 139, 148 basis, 31 benefits, 149 chemostatic theory and, 150, 275 costs, 149, 273, 308 design and analysis, 149, 273, 308 dilution rate, 152 for hybridoma, 276 maximizing productivity, 152, 276–277, 310–311 microorganisms in, 112 requirements, 137, 157, 202 scale-up, 31, 196–199. See also Scale-up size requirements for, 137, 298 Continuously stirred tank fermenter, 134 with cell cycling, 178 continuous culture, 148 economics, 156 feed stream, 149, 274, 309 growth rate, 147, 148 material balance for culture, 145, 155 at maximum productivity, 146, 176 product residence time, 149 productivity, 146, 176, 177 series arrangement, 177 STF vs., 177 time for, flux and throughput vs., 146 Control(s) complex bioreactive, 168, 237 process, 236 system design, 232–233, 236 ranges, 234 system functions, 162 Controller(s) derivative, 168, 237 feed-forward, 168, 237 feedback, 166, 236 integral, 167, 236 Cysteine proteinase, 69

D Dephosphorylation, 66

PH2270_IDX.fm Page 377 Tuesday, May 23, 2006 10:02 AM

Index

377

Designer gene, 18, 46 Dipeptidyl aminopeptidase I, 53 DNA amplification ligation-based, 90–92 PCR, 85–87 Qβ replicase, 92–93 reagents, 85–95 by RNA replication, 92–94 by RNA transcription, 87–90 sequence-based, 94 base-pair complementarity, 10–11 ligases, 76–77, 90–92 phosphorylation and dephosphorylation, 76 polymerases, 77–79 sequence determinations, 13 synthesis, 16–19, 43–47, 325 synthesizers. See DNA synthesizers DNA probe, 16 assay, 42 synthesis, 43 DNA synthesizers, 325–329 advantage of multiple chemistries, 327–328 automated units, 328 coupling chemistries, 326–327 manufacturers, 329 scale up, 329 throughput capability, 328–329 DNase I, 71 Dot blot, 17

E Earle’s balanced salt solution, 107 Ebola virus, 124 Endonuclease, 72, 80 Endopeptidase cleavage, 130 Endotoxin, 102 Endotoxin testing, 112 Enzyme(s), 63–83, 133–134 catalytic activity, 2, 63, 67, 69 methods to maintain, 66–67 cavitation, 66 chemical, 66 microencapsulation, 66 physical, 66 categories, 64 DNA polymerases, 77–79 families and compatible ends, 71–80 hydrolytic, 133 isoschizomers, 71 manufacturers, 80–83 miscellaneous modifying, 80 modifying, 65, 66, 76, 80

nucleases, 71–73 palindromic recognition, 70 phosphodiesterases, 79–80 prokaryotic cell, 129 proteases, 74–76 protein engineering, 67–68 proteolytic, 66 protoplast-forming, 80 restriction, 64, 66 classes, 69–70 specificity, 70 ribonucleases, 73–74 star activity, 70 T4 gene 32, 78 tissue dissociation and cell isolation, 66 transformation in nonaqueous systems, 68–69 Epidermal growth factor, 109 Equilibrium dialysis, 208 Escherichia coli, 39, 128 colony, 21 DNA ligase, 76 DNA polymerase I, 77 exonuclease III, 78 growth, 49 MABs expressed from, 19 as prototypical microorganism, 20 RNA polymerase, 78 Eukaryotic cell(s), 1, 46 cell-culture media, 97–98, 217–218 DNA, 72 expression systems, 21 extracellular expression, 53, 130 genes designer, 18 exons, 17, 45 expression, 44, 80 growth characteristics, 50–51 intracellular expression, 52 introns, 20, 49 modeling growth and production, 262–282 noncoding sequences, 20, 49 organization, 12 protein coding sequence, 20, 49 structure, 71 growth, 27, 50–51, 277–278 cell composition, 277 conditions, 275–276 factors affecting rate, 270–271 kinetic, 262–269, 297–301 maximizing, 276–277 models, 238–243, 271–275, 305–309 stoichiometric coefficient, 277, 282 yield, 279–282

PH2270_IDX.fm Page 378 Tuesday, May 23, 2006 10:02 AM

378

Producing Biomolecular Substances

manufacturers of cell lines, 61 products derived from, 8 Exons, 17, 45 Exonuclease III, 72 Expression. See also Gene(s), expression extracellular, 53–54, 129 intracellular vs., 23, 52 intracellular, 52–53, 129 extracelluar vs., 23, 52 levels, 22, 51–52 systems, 19–25, 48 economic concerns, 24 eukaryotic cells, 21 further posttranslational modification, 24 general considerations, 24 heterologous, 127 intra- vs. extracellular, 23 levels, 22 prokaryotic cells, 20–21 regulatory, 25 vector construction, 20 Extrachromosomal DNA fragment, 128

F Feedstocks, 202, 238 Fermentation. See also Continuous culture systems cell growth and production, 134–151 continuous culture, 21, 50, 128, 139 enzyme, 133–134 historical perspectives, 2, 3, 121–127 introduction to, 121 kinetic modeling, 184–192 optimal conditions, 151–152 for penicillin, 3 for pharmaceuticals, 3 procedure, 7 process, 127–133 recombinant culture kinetics, 152–176 Fermenter(s), 121–200 adaptive bioreactive controls, 168 automatic control systems, 167 bioreactors vs., 202 bubble-column, 179–180 characteristics, 158 cleaning and sterilization, 168 complex bioreactive controls, 168 containment vessel, 160 continuous-culture scale up, 196–199 continuously stirred tank, 134 with cell cycling, 178 continuous culture, 148 economics, 156

feed stream, 149, 274, 309 growth rate, 147, 148 material balance for culture, 145, 155 at maximum productivity, 146, 176 product residence time, 149 productivity, 146, 176, 177 series arrangement, 177 STF vs., 177 time for, flux and throughput vs., 146 control system functions, 162 derivative controllers, 168 design, 176–184 basic physical elements, 181–184 cell kinetics, 176–179 complex physical elements, 184 feed-forward controllers, 168 feedback controllers, 166 increased size and complexity, 161–162 integral controllers, 167 manufacturers, 199–200 mixing and aeration, 160–161 operating modes, 156–157 oxygen transfer resistance, 160–161 packed-bed, 177, 178 power number method for STF geometric scale up, 194–196 process monitoring and control, 162–165 sensor location and function, 166 scale-up, 183, 192–194 size and scale, 158–159 static maintenance, 150 sterilization, 168–169. See also Sterilization stirred tank, 139, 158, 192–196 batch, 177 scale-up, 192–194 bench-top, 159 comparison, 156 continuously, 134 design additions, 180, 181 geometric scale up, 194–196 inoculation, 138 with internal loop, 181 structure and configuration, 157 supplementary equipment, 176 suppliers and products, 193 Fetal bovine serum, 21, 50 viral contaminants, 101 Fibroblast growth factors, 110 Food and Drug Administration, 34–37, 58–59 biological license application, 36 Center for Biologics Evaluation and Research, 28 Center for Drugs and Biologics, 35 history and agency structure, 35–36

PH2270_IDX.fm Page 379 Tuesday, May 23, 2006 10:02 AM

Index

379

Modernization Act (1997), 34 responsibility for therapeutic biologics, 37 Fusion protein, 128

G Gas gangrene, 125 Gelatin-shard-inverted microcarriers, 221 Gene(s), 9 designer, 18, 46 eukaryotic cell designer, 18 exons, 17, 45 expression, 44, 52, 80 introns, 20, 49 noncoding sequences, 20, 49 organization, 12 protein coding sequence, 20, 49 structure, 71 expression, 9, 11, 49. See also Expression applications, 18 codon usage and, 20, 49 complexity, 17, 45 methanol regulation, 21, 50 methylated residues, 80 phosphate-free, 131 repression of, 21, 128, 156, 314 temperature-dependent control, 156, 314 tracking, 341 processed, 12 prokaryotic cell designer, 18, 46 expression, 20, 56 introns, 49 organization, 12 structure, 71 segments, interrupting, 12 Genetic coding, 11–12 Gentamicin, 214 Glycosylation, 23, 54 insect cell, 55 mammalian cell, 54 Growth, 49–51, 134–151 accelerated, 136, 141, 264 bacteria, characteristics of, 49 cycle, 136–137, 264 accelerated stage, 136, 141, 264, 268 death, 137, 145, 264 decelerated stage, 137, 264 exponential growth period, 137, 141, 142, 152, 264 latent stage, 136, 140, 141, 264 static period, 137, 145, 264 decelerated, 137, 264

eukaryotic cell, 27, 277–278 cell composition, 277 conditions, 275–276 factors affecting rate, 270–271 kinetic, 262–269, 297–301 maximizing, 276–277 models, 238–243, 271–275, 305–309 stoichiometric coefficient, 277, 282 yield, 279–282 exponential, 137, 141, 142, 152, 264 factors affecting rate of, 253 fermentation and, 134–151 heat release due to, 243–245 latent, 136, 140, 141, 264 optimal conditions, 275–276 prokaryotic cell, 49–50, 57 rate, 135 factors affecting, 143–145, 290, 304–305 specific, 141 reaction, 238 static, 137, 145, 264 yeast, 280–281 yield, 139, 147, 242 Growth factors, 109–110 Growth hormone, 18, 46, 110 human, 53, 130

H Hanks’ balanced salt solution, 107 Heat, 246–247 generation, 230–232 as means of sterilization, 169 release, 243–245, 283–284 Hemorrhagic conjunctivitis virus, 124, 126 HEPES buffer, 107–108, 117–118 hGH. See Human growth hormone (hGH) Human growth hormone (hGH), 53, 130 Hybridoma, 5, 6, 8, 211–212 clones, 212 continuous culture for, 276 fusion process, 212 high-density cell line growth, 102 interspecies, 209 MABs from, 19, 47, 212– 213 Milstein-Kohler procedure, 206 Hydrogen fluoride, 345 Hydrolytic enzymes, 133

I IgG affinity chromatography, 51

PH2270_IDX.fm Page 380 Tuesday, May 23, 2006 10:02 AM

380

Producing Biomolecular Substances

Immunoassay(s), 209–210 antibody-based, 210 limited reagent, 210 optimizing and validating, 210 reagent excess, 210 Immunogeneity, altered, 129 Immunoglobulins, 207 Introns, 20, 49 Isoschizomers, 71

K Kanamycin, 215

L L-15 medium, 107 Langmuir equation, 209, 210 Lectins, 110–111, 118 Ligase(s), 65, 76–77 Lipase(s), 63 Lipopolysaccharide, 112

M Media supplements, 104–105 amino acids, 104 amphotericin B, 105 ampicillin, 105 antibiotics, 104–105 tylosin solution, 105 Methionine aminopeptidases, 129 Methylase Hpa II, 80 Methylation, 66 Microencapsulation, 66 Microheterogeneity, 58 Milstein-Kohler procedure, 206 Modernization Act (1997), 34 Modifying enzymes, 65, 66, 76, 80 Monoclonal antibody(ies) production, 206–224 in ascites fluid, 213–224. See also Ascites fluid, MAB production in by cell culture, 216 hybridoma creation for, 212–213 methods of, 212 Monod constant, 250 Mung Bean nuclease, 72 Mutagenesis, 76, 206, 325 site-directed, 19, 47, 67, 327 amino acid exchange in, 74 efficiency of, 73

elongation of oligonucleotides in, 77 gap-filling in, 78 Mycoplasma spp., 102, 105, 222 Mycotoxin, 124

N Neomycin, 215 Neurospora crassa, 72, 80 Noncoding sequences, 49 Northern blot, 17, 44 Novobiocin, 215 Nuclease(s), 65, 71–73 manufacturers, 81–82 Nuclease Bal 31, 72 Nuclease P1, 72 Nuclease S1, 73 Nuclease S7, 72–73

O OKT3 antibody, 223 Oxygen accumulation, 260 Oxygen transfer, 160–161, 201 factors affecting mass-transfer coefficient, 261–262, 295–297 metabolic oxygen demand and, 257, 292 volumetric oxygen mass-transfer coefficient, 226, 257–260, 292–294

P Packed-bed fermenter, 177, 178 Parainfluenza 3, 101 PCR. See Polymerase chain reaction (PCR) Penicillin, 214 Peptide(s), 332, 334 biologically active, 331–332 phage-generated libraries, 339–340 solid-phase, 334–335 synthesis cell-free translation systems in, 341–342 cleavable linkers and, 340–341 enzymatic, 341 scale-up, 346 equipment for, 342–343 fragment condensation, 338 recombinant, 338 scale-up, 345 scaling-up, 343–346 screening for, 338–339 sequential, 338

PH2270_IDX.fm Page 381 Tuesday, May 23, 2006 10:02 AM

Index solid-phase scale-up, 345 solution, 336–338 scale-up, 344 split approach, 340 Peptide synthesizers manufacturers, 346 process-scale, 343 solid-phase, 335–336 Peptide technology, 330–332. See also Peptide(s), Peptide synthesizers Phosphodiesterase(s), 65, 79–80 Phospholipase(s), 69 Phosphorylation, 56 Pichia pastoris, 21, 128 Pilot Fermentation Procedure, 7 Plasmid-carrying cell growth-rate kinetics, 152–156 Polymerase(s), 65 DNA and RNA, 77–80 E. coli DNA, 77 exonuclease III, 78 labeling-grade Klenow enzyme, 77 sequencing-grade Klenow enzyme, 77 T4 DNA polymerase, 78 T4 gene 32 enzyme, 78 taq DNA, 77 terminal transferase, 78 Polymerase chain reaction (PCR), 9, 13–15, 85–87, 325 ligation-based amplification vs., 92 problems, 87 Qβ replicase amplification vs., 92, 93 sequenced-based amplification vs., 94 TAS amplification vs., 89, 90 Porous microcarriers, 219–221 Prescription Drug Free Acts, 34 Primer(s), 77, 89, 93 extension, 85, 86, 87 hybridization, 86 labeling, 77 oligonucleotide, 14 polyribonucleotide, 79 RNA sequences, 76 sequences, 76, 87 Prokaryotic cell(s), 1, 24, 27, 56, 157 culture medium, 97 enzymes, 129 expression systems, 20–21, 131 gene designer, 18, 46 expression, 20, 56 introns, 49 organization, 12 structure, 71

381 growth, 49–50, 57 intracellular expression, 53 manufacturers of cell lines, 60–61 Pronase, 75 Protease(s), 65 pronase, 75 proteinase K, 75–76 sequencing-grade, 75–76 Protein-coding sequence, 20, 49 Protein-free media, 218 Proteolytic enzymes, 66 Protoplast-forming enzymes, 80 Purification, 7

Q Qβ replicase amplification system, 92–93 Quality assurance, 8

R Rabies, 122 Recognition sequence of factor Xa, 75 Recombinant culture kinetics, 152–176, 311–314 Recombinant DNA materials and methods, 39–61 applications, 41, 43 cell lines, 47–48 cloning, 39–41 DNA synthesis, 43–47 economic considerations, 56–58 expression systems, 48 glycosylation, 54–55 manufacturers, 59–61 posttranslational modifications, 54–56 regulatory concerns, 58–59 screening and selection, 43 vector construction, 48–49 Recovery, 7 Rennet, 63 Restriction endonuclease(s), 69 Restriction-endonuclease-digested fragment patterns, 45 Restriction enzyme(s), 66, 69 classes, 69–70 manufacturers, 80–81 specificity, 70 Restriction fragment length polymorphisms (RFLP), 17 Reverse transcriptases, 71, 81 RFLPs. See Restriction fragment length polymorphisms (RFLPs) Rhinotracheitis, 101

PH2270_IDX.fm Page 382 Tuesday, May 23, 2006 10:02 AM

382

Producing Biomolecular Substances

Ribonuclease(s), 65, 73–74 DNase-free RNase, 73 manufacturers, 82 RNase A, 73 RNase CL3, 74 RNase H, 74 RNase T, 74 RNase U2, 74 Ricin, 124, 126 RNA blot, 17, 44 polymerases, 77–79 replication, amplification by, 92–94 transcription, amplification by, 87–90 RNase-free DNase I, 72 Rotavirus, 124, 126

S Saccharomyces cerevisiae, 56, 128, 133 advantages for large-scale cultures, 21 growth, 49–50 protoplast formation, 80 recombinant interferon, 53 Scale-up, 33 batch, 192–194 dimensionless numbers method, 193–194 Froude number for, 194 geometric method, 192–193 scale-of-agitation method, 192 bioreactor, 216, 226 for botulinum toxin production, 125 continuous-culture, 196–199 assessing productivity loss, 198 contamination and mutant selection, 198 key biological aspect, 197 laboratory to pilot plant to production, 196–197 maintaining steady state, 197–198 problems, 198–199 theoretical basis, 196 costs, 53, 130 data, 33 defined, 31 design, 33 fermenter, 183 geometric, 194–196 peptide synthesis, 343–346 Scatchard equation, 208, 210 Separation, 7 Serum-free media, 102, 218 Sips equation, 209 Site-directed mutagenesis, 19, 47, 67, 68, 327 amino acid exchange in, 74

efficiency of, 73 elongation of oligonucleotides in, 77 gap-filling in, 78 Slot blot, 17 Snake venom, phosphodiesterase from, 80 Sodium bicarbonate, 106–107 Southern blot, 16, 17, 44, 45 Spodoptera frugiperda, 23, 55 Star activity, 70 Sterilization, 31, 32, 100, 151, 168–175 for bench-scale units, 160, 205 bioreactor, 245–248 continuous, 171–173 Del factor in, 170–171 double-wall steam, 160, 205 mathematical expression, 247 means of, 169–170 antiseptic chemicals, 170 heat, 169–170 ultrafiltration, 170 ultrasound, 170 ultraviolet light, 170 modeling, 245–248 repeated, 164, 235 in situ, 158, 203 steam, 173–175 clean or pure, 174–175 double-wall, 160, 205 plant or utility, 174 system components, 173–174 system design, 175 Sterilization-in-place system, 168 STF. See Stirred tank fermenter(s) Stirred tank fermenter(s), 139, 158, 192–196 batch, 177 scale-up, 192–194 bench-top, 159 comparison, 156 continuously, 134 with cell cycling, 178 continuous culture, 148 economics, 156 feed stream, 149, 274, 309 growth rate, 147, 148 material balance for culture, 145, 155 at maximum productivity, 146, 176 product residence time, 149 productivity, 146, 176, 177 series arrangement, 177 STF vs., 177 time for, flux and throughput vs., 146 design additions, 180, 181 geometric scale up, 194–196 scale-of-agitation method for, 194–196

PH2270_IDX.fm Page 383 Tuesday, May 23, 2006 10:02 AM

Index

383

inoculation, 138 with internal loop, 181 Stoichiometric coefficients, 243, 277, 282 Streptomyces spp., 105 Streptomycin, 214 Substrate saturation constant, 250, 287 Superoxide dismutase, 22, 24, 56, 129, 130

T T4 gene 32 enzyme, 78 T4 polynucleotide kinase, 76 Taq DNA polymerase, 77 TAS. See Transcriptional amplification system (TAS) Teratogens, 101 Terminal transferase, 78 Terminator signal, 49 Throughput, 146, 328–329 Toxins, 111, 118 Transcriptional amplification system (TAS), 87–90 modified protocol, 89 Transfected blastocystic embryonic stem cells, 46 Transgene, 46 Translation initiation, 49 Transport factors, 111, 119 Trichothecene mycotoxin, 126 Trypsin, 63 Tumor necrosis factor, human, 22 Tylosin solution, 105 Tylosinate, 215

U Ultrafiltration, 170 Ultrasound, 170 Ultraviolet light, 170 Uracil-DNA Glycosylase, 73

V Vector construction, 20, 48–49 Vitamins, 3, 111–112, 119

W Water, 112, 119 Watson-Crick DNA model, 10 Western blot, 17, 44 Wheat cover smut, 126

Y Yeast, 49–50, 80, 128, 133 advantages in bioprocessing, 21, 50 as alternative for N-acetylated heterologous proteins, 56, 131 contamination, 222 enzymes, 63 genes, introns in, 49 growth, 280–281 on ethanol, 280 on glucose, 281 interferon expression, 53 manufacturers of expression systems, 60 peptide expression, 338, 345 phosphate-free gene expression, 56 SOD and, 130 Yield, 31, 132, 196, 215, 242–243, 279–282 constant, 190 factor, 151 growth, 139, 147, 242 high-end product, 137 increasing, 133 optimal, 4, 27 quality, 326 with serum-free medium, 218 in solid-phase peptide synthesis, 334 staphylococcal protein A, 22, 51, 129

PH2270_IDX.fm Page 384 Tuesday, May 23, 2006 10:02 AM

Producing Biomolecular Substances with Fermenters, Bioreactors, and Biomolecular Synthesizers - PDF Free Download (2024)

References