Contributors to V o l u m e 70 Article numbers are in parentheses following the names o f contributors. Affiliations lis...
14 downloads
855 Views
44MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Contributors to V o l u m e 70 Article numbers are in parentheses following the names o f contributors. Affiliations listed are current.
FRANK L. ADLER (30), Division of Immu-
BERNARD F. ERLANGER (4), Department of
nology, St. Jude Children's Research Hospital, Memphis, Tennessee 38101 LOUISE T. ADLER (30), Division of Immunology, St. Jude Children's Research Hospital, Memphis, Tennessee 38101 AUGUSTIN BAER (31), Station .fdddrale de Recherches Laitikres, 3097 LiebefeldBern, Switzerland
W1GGO FISCHER-RASMUSSEN (22), Depart-
SARA BAUMINGER (7), Institute of Repro-
WARREN D. GEHLE (27), Litton-Bionetics,
Microbiology, Columbia University Health Sciences Center, New York, New York 10032 ment of Obstetrics and Gynecology, KObenhavns Kommunes, Hvidovre Hospital, University of Copenhagen, DK2650 Copenhagen-Hvidovre, Denmark Kensington, Maryland 20795
ductive Endocrinology, Municipal Governmental Medical Center, Tel Aviv, and Department of Hormone Research, The Weizmann Institute of Science, Rehovot, Israel
C. N. HALES (24), Department of Clinical
Biochemistry, University of Cambridge, Addenbrooke's Hospital, Cambridge CB2 2QR, England
HUGH J. CALLAHAN (2), Department of BiD-
H. J. HANSEN (20), Roche Research Center,
chemistry, Thomas Jefferson University, Philadelphia, Pennsylvania 19107
Hoffmann-La Roche Inc., Nutley, New Jersey 07110
SYLVIA R. CHALKLEY (21), Department of
MA~OR1E R. HEPBURN (16), Department of
Child Health, Westminster Children's Hospital, London SWIP 2NS, England
Pathology, University of Michigan, Ann Arbor, Michigan 48109
M. CHANTLER (5), Wellcome Reagents Limited, Wellcome Research Laboratories, Beckenham, Kent BR3 3BS, England T. CHARD (18), Department of Reproductive Physiology, Joint Unit of Obstetrics and Gynaecology and Reproductive Physiology, St. Bartholomew's Hospital Medical College and The London Hospital Medical College, London ECI, England J. W. COEEEY (20), Department of Pharmacology, Hoffmann-La Roche Inc., Nutley, New Jersey 07110 S. L. COMMEREORD (14), Medical Department, Brookhaven National Laboratory, Upton, New York 11973 FRANK J. DIXON (ll), Department o f l m munopathology, Scripps Clinic and Rd'search Foundation, La Julia, California 92037
B. A. L. HURN (5), Wellcome Reagents
SHIREEN
Limited. Wellcome Research Laboratories, Beckenham, Kent BR3 3BS, England W|LLIAM P. JENCKS (31), Department of
Biochemistry, Brandeis University, Waltham, Massachusetts 02154 ELVIN A. KABAT (1), Departments of Mi-
crobiology, Human Genetics and Development, Neurology and the Cancer Center, Columbia University, New York, New York 10032 JOHN J. LANGONE (13, 25), Laboratory of
lmmunobiology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20205 JENS LARSEN (22), Department of Obstetrics and Gynecology, Rigshospitalet, University of Copenhagen, DK-2100 Copenhagen, Denmark
EVA ENGVALL (28), La Julia Cancer Re-
LAWRENCE LEVINE (31), Department of
search Foundation, La Julia, California 92037
Biochemistry, Brandeis University, Waltham, Massachusetts 02154 ix
X
CONTRIBUTORS
MICHAEL G. MAGE (6), Protein Chemistry
Section, Laboratory of Biochemistry, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 2O205
TO V O L U M E 70 A. RENSHAW (21), Division of Clinical
Chemistry, Clinical Research Centre, Harrow, Middlesex HAl 3UJ, England ROBERT T. RUBIN (23), Division of Biological Psychiatry, Department of PsychiaH. L. J. MAK1N (19), Steroid Laboratory, try, Harbor-UCLA Medical Center, Department of Chemical Pathology, The Torrance, California 90509 London Hospital Medical College, LonE. R. SAUERZOPE (20), Roche Research don El 2AD, England Center, Hoffmann-La Roche Inc., Nutley, New Jersey 07110 PAUL H. MAURER (2), Department of BiDchemistry, Thomas Jefferson University, SEYMOUR I. SCHLAGER (15), Laboratory of Philadelphia, Pennsylvania 19107 Immunobiology, National Cancer Institute, National Institutes of Health, BePATRICIA J. McCONAHEY (I1), Department thesda, Maryland 20205 of Immunopathology, Scripps Clinic and Research Foundation, La Julia, Cali- MORTON B. SIGEL (23), Lutcher Brown fornia 92037 Center for Diabetes and Endocrinology, Scripps Clinic and Research Foundation, J. MERRETT (26), RAST Allergy and Research La Julia, California 92037 Unit, Benenden Chest Hospital, Cranbrook, Kent TNI7 4AX, England KENDALL O. SMITH (27), Department of Microbiology, The University of Texas Health T. G. MERRETT (26), RASTAllergy and ReScience Center, San Antonio, Texas 78284 search Unit, Benenden Chest Hospital, B. DAVID STOLLAR (3), Department of BioCranbrook, Kent TN17 4AX, England chemistry and Pharmacology, Tufts UniM. E. MEYERHOFF (29), Department of versity School of Medicine, Boston, MassaChemistry, University of Michigan, Ann chusetts 02111 Arbor, Michigan 48109 BARBARA B. TOWER (23), Division of BioA. REES MIDGLEY, JR. (16), Department of logical Psychiatry, Department of PsyPathology, University of Michigan, Ann chiatry, Harbor-UCLA Medical Center, Arbor, Michigan 48109 Torrance, California 90509 MARTIN MORRISON (12), Department of BiDD. J. H. TRAEEORD (19), Steroid Laborachemistry, St. Jude Children's Research tory, Department of Chemical Pathology, Hospital, Memphis, Tennessee 38101 The London Hospital Medical College, WlLLIAM D. ODELL (17), Department of London El 2AD, England Medicine, University of Utah College of HELEN VAN VUNAKIS (10), Department of Medicine, Salt Lake City, Utah 84132 Biochemistry, Brandeis University, WalJACQUES OUDIN (9), Institut Pasteur, tham, Massachusetts 02154 28 rue du Docteur Roux, 75015 Paris, W. P. VANDERLAAN (23), Lutcher Brown France Center for Diabetes and Endocrinology, RUSSELL E. POLAND (23), Division of BioScripps Clinic and Research Foundation, logical Psychiatry, Department of PsyLa Julia, California 92037 chiatry, Harbor-UCLA Medical Center, J. P. VANDEVOORDE (20), Roche Research Torrance, California 90509 Center, Hoffmann-La Roche Inc., Nutley, - New Jersey 07110 G. A. RECHNITZ (29), Department of Chemistry, University of Delaware, ME1R WILCHEK (7), Department of BioNewark, Delaware 19711 physics, The Weizmann Institute of Science, Rehovot, Israel MORRIS REICHLIN (8), Veterans Administration Medical Center, Departments of J. S. WOODHEAD (24), Department of MediMedicine and Biochemistry, State Unical Biochemistry, Welsh National School versity of New York at Buffalo School of of Medicine, Heath Park, Cardiff CF4 Medicine, Buffalo, New York 14215 4XN, Wales
Preface Immunochemical procedures provide an important supplement to the battery of available chemical and instrumental methods and can often yield new information not readily obtainable in other ways. Antibodies are extraordinary analytical reagents since they can have specificity for macromolecules (proteins, nucleic acids, and polysaccharides) as well as for small molecules belonging to almost every chemical class. In a field that is moving so rapidly, an exhaustive compilation of immunochemical techniques is neither possible nor practical. Our purpose is to provide the investigator with significant examples and sufficient background information so that he can properly assess and adapt these techniques to his research. This is the first of several volumes to be devoted to the description and application of immunochemical techniques. It deals with the basic principles of antigen-antibody reactions, production of reagent antibodies, as well as with purification and characterization of antibodies and antigens. Antibodies to individual compounds can be used with tracer molecules to develop sensitive, specific, rapid, and reproducible immunoassays. General procedures required to develop suitable radioimmunoassays and solid phase immunoradiometric assays are described in detail. Immunoassays that utilize means of detection other than radioactivity are included. Such assays are becoming increasingly popular because they possess the desirable attributes of good analytical methods yet avoid the generation of radioactive waste. Already in progress is a second volume that will supplement the topics presently covered and include other important techniques. Innovative and practical applications of immunochemical methods have been described in other volumes of this series. We have avoided duplication so far as possible, and have included a cross-reference bibliography for each section (see pp. 481-484) to direct the reader to related papers in other volumes. Subsequent volumes will be involved with the development and application of immunoassays for specific compounds as well as for different classes of compounds. We are grateful to the authors for their contributions and to the staff of Academic Press for invaluable assistance. Carla Langone is to be commended for her competent handling of the correspondence. Timely advice and constructive criticism were given by Nathan O. Kaplan and Sidney P. Colowick. Although this is the seventieth volume in the Methods in Enzymology Series, their enthusiasm, interest, and concern remain undiminished. HELEN VAN VUNAKIS JOHN J. LANGONE
xi
METHODS IN ENZYMOLOGY EDITED BY Sidney P. Colowick and Nathan O. Kaplan VANDERBILT UNIVERSITY
DEPARTMENT OF CHEMISTRY
SCHOOL OF MEDICINE
UNIVERSITY OF CALIFORNIA
NASHVILLE, TENNESSEE
AT SAN DIEGO LA JOLLA, CALIFORNIA
I. II. III. IV. V. VI.
Preparation and Assay of Enzymes Preparation and Assay of Enzymes Preparation and Assay of Substrates Special Techniques for the Enzymologist Preparation and Assay of Enzymes Preparation and Assay of Enzymes (Continued) Preparation and Assay of Substrates Special Techniques VII. Cumulative Subject Index
xiii
METHODS IN E N Z Y M O L O G Y EDITORS-IN-CHIEF S i d n e y P. C o l o w i c k
N a t h a n O. K a p l a n
VOLUME VIII. Complex Carbohydrates
Edited by ELIZABETHF. NEUFELD AND VICTOR GtNSBURG VOLUME IX. Carbohydrate Metabolism
Edited by WILLIS A. WOOD VOLUME X. Oxidation and Phosphorylation
Edited by RONALD W. ESTABROOK AND MAYNARD E. PULLMAN VOLUME XI. Enzyme Structure
Edited by C. H. W. HIRS VOLUME XII. Nucleic Acids (Parts A and B) Edited by LAWRENCEGROSSMAN AND KIVIE MOLDAVE VOLUME XIII. Citric Acid Cycle
Edited by J. M. LOWENSTEIN VOLUME XIV. Lipids
Edited by J. M. LOWENSTEIN VOLUME XV. Steroids and Terpenoids
Edited by RAYMOND B. CLAYTON VOLUME XVI. Fast Reactions
Edited by KENNETH KUSTIN VOLUME XVII. Metabolism of Amino Acids and Amines (Parts A and B)
Edited by HERBERT TABOR AND CELIA WHITE TABOR VOLUME XVIII. Vitamins and Coenzymes (Parts A, B, and C)
Edited by DONALD B. MCCORMICKAND LEMUEL D. WRIGHT XV
xvi
METHODS IN ENZYMOLOGY
VOLUME XIX. Proteolytic Enzymes Edited by GERTRUDE E. PERLMANN AND LASZLO LORAND
VOLUME XX. Nucleic Acids and Protein Synthesis (Part C)
Edited by KIVIE MOLDAVE AND LAWRENCE GROSSMAN VOLUME XXI. Nucleic Acids (Part D)
Edited by LAWRENCE GROSSMAN AND KIVIE MOLDAVE VOLUME XXII. Enzyme Purification and Related Techniques
Edited by WILLIAM B. JAKOBY VOLUME XXIII. Photosynthesis (Part A)
Edited by ANTHONY SAN PIETRO VOLUME XXIV. Photosynthesis and Nitrogen Fixation (Part B)
Edited by ANTHONY SAN PIETRO VOLUME XXV. Enzyme Structure (Part B)
Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME XXVI. Enzyme Structure (Part C)
Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME XXVII. Enzyme Structure (Part D)
Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME XXVIII. Complex Carbohydrates (Part B)
Edited by VICTOR GINSBURG VOLUME XXIX. Nucleic Acids and Protein Synthesis (Part E)
Edited by LAWRENCE GROSSMAN AND KIVIE MOLDAVE VOLUME XXX. Nucleic Acids and Protein Synthesis (Part F)
Edited by KIVIE MOLDAVE AND LAWRENCE GROSSMAN VOLUME XXXI. Biomembranes (Part A)
Edited by SIDNEY FLEISCHER AND LESTER PACKER VOLUME XXXlI. Biomembranes (Part B)
Edited by SIDNEY FLEISCHER AND LESTER PACKER
METHODS IN ENZYMOLOGY
xvii
VOLUME XXXIII. Cumulative Subject Index Volumes I-XXX
Edited by MARTHA G. DENNIS AND EDWARD A. DENNIS VOLUME XXXIV. Affinity Techniques (Enzyme Purification: Part B)
Edited by WILLIAMB. JAKOBY AND MEIR WILCHEK VOLUME XXXV. Lipids (Part B) Edited by JOHN M. LOWENSTEIN VOLUME XXXVI. Hormone Action (Part A: Steroid Hormones)
Edited by BERT W. O'MALLEY AND JOEL G. HARDMAN VOLUME XXXVII. Hormone Action (Part B: Peptide Hormones)
Edited by BERT W. O'MALLEY AND JOEL G. HARDMAN VOLUME XXXVIII. Hormone Action (Part C: Cyclic Nucleotides)
Edited by JOEL G. HARDMAN AND BERT W. O'MALLEY VOLUME XXXIX. Hormone Action (Part D: Isolated Cells, Tissues, and Organ Systems) Edited by JOEL G. HARDMAN AND BERT W. O'MALLEY VOLUME XL. Hormone Action (Part E: Nuclear Structure and Function)
Edited by BERT W. O'MALLEY AND JOEL G. HARDMAN VOLUME XLI. Carbohydrate Metabolism (Part B)
Edited by W. A. WOOD VOLUME XLII. Carbohydrate Metabolism (Part C)
Edited by W. A. WOOD VOLUME XLIII. Antibiotics
Edited by JOHN H. HASH VOLUME XLIV. Immobilized Enzymes
Edited by KLAUS MOSBACH VOLUME XLV. Proteolytic Enzymes (Part B)
Edited by LASZLO LORAND VOLUME XLVI. Affinity Labeling
Edited by WILLIAM B. JAKOBY AND MEre WILCHEK
xviii
METHODS
IN E N Z Y M O L O G Y
VOLUME XLVII. Enzyme Structure (Part E)
Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME XLVIII. Enzyme Structure (Part F)
Edited by C. H. W. HIRS AND SERGE N. TIMASHEFE VOLUME XLIX. Enzyme Structure (Part G)
Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME L. Complex Carbohydrates (Part C)
Edited by VICTOR GINSBURG VOLUME LI. Purine and Pyrimidine Nucleotide Metabolism
Edited by PATRICIA A. HOFEEE AND MARY ELLEN JONES VOLUME LII. Biomembranes (Part C: Biological Oxidations)
Edited by SIDNEY FLEISCHERAND LESTER PACKER VOLUME LIII. Biomembranes (Part D: Biological Oxidations)
Edited by SIDNEY FLEISCHERAND LESTER PACKER VOLUME LIV. Biomembranes (Part E: Biological Oxidations)
Edited by SIDNEY FLEISCHERAND LESTER PACKER VOLUME LV. Biomembranes (Part F: Bioenergetics)
Edited by SIDNEY FLEISCHERAND LESTER PACKER VOLUME LVI. Biomembranes (Part G: Bioenergetics)
Edited by SIDNEY FLEISCHERAND LESTER PACKER VOLUME LVII. Bioluminescence and Chemiluminescence
Edited by MARLENEA. DELUCA VOLUME LVIII. Cell Culture
Edited by WILLIAM B. JAKOBY AND IRA H. PASTAN VOLUME LIX. Nucleic Acids and Protein Synthesis (Part G)
Edited by KIV1E MOLDAVE AND LAWRENCE GROSSMAN VOLUME LX. Nucleic Acids and Protein Synthesis (Part H)
Edited by KIVIE MOLDAVE AND LAWRENCE GROSSMAN
METHODS IN ENZYMOLOGY
VOLUME 61. Enzyme Structure (Part H)
Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME 62. Vitamins and Coenzymes (Part D) Edited by DONALD B. MCCORMICK AND LEMUEL D. WRIGHT VOLUME 63. Enzyme Kinetics and Mechanisms (Part A: Initial Rate and Inhibitor Methods) Edited by DANIEL L. PURICH VOLUME 64. Enzyme Kinetics and Mechanisms (Part B: Isotopic Probes and Complex Enzyme Systems) Edited by DANIEL L. PURICH VOLUME 65. Nucleic Acids (Part I)
Edited by LAWRENCE GROSSMAN AND KIVIE MOLDAVE VOLUME 66. Vitamins and Coenzymes (Part E)
Edited by DONALD B. MCCORMICK AND LEMUEL D. WRIGHT VOLUME 67. Vitamins and Coenzymes (Part F) Edited by DONALD B. MCCORMICK AND LEMUEL D. WRIGHT VOLUME 68. Recombinant DNA
Edited by RAY Wu VOLUME69. Photosynthesis and Nitrogen Fixation (Part C)
Edited by ANTHONY SAN PIETRO VOLUME 70. Immunochemical Techniques (Part A) Edited by HELEN VAN VUNAKIS AND JOHN J. LANGONE VOLUME 71. Lipids (Part C) (in preparation)
Edited by JOHN M. LOWENSTEIN VOLUME 72. Lipids (Part D) (in preparation)
Edited by JOHN M. LOWENSTEIN
xix
[1]
PRINCIPLES
[1]
OF
Basic Principles
ANTIGEN--ANTIBODY
REACTIONS
of Antigen-Antibody
3
Reactions
By ELVlN A. KABAT Definitions
Antigen: An antigen is any substance that, when introduced parenterally into an animal, will induce the formation of antibodies. The antibody formed is generally found in serum or other biological fluids and should react with the antigen used to induce its formation. The term immunogen is often used to apply to substances that will induce a state of cellmediated immunity as well as the formation of antibodies regardless of their specificity. It is possible to increase immunogenicity by introducing groups onto proteins that do not alter their specificity or become part of an antigenic determinant. Antibody: A serum protein belonging to the family called immunoglobulins. There are five classes of immunoglobulins: IgG, IgM, IgA, IgD, and IgE. These are all built of two chains--heavy and light chains (see Figs. 1 and 4). Antibodies may be found in the serum of normal individuals, presumably as a consequence of contact with antigens by the oral, respiratory, enteric, or parenteral routes. Many antibodies result from apparent or inapparent infections with microorganisms. It is generally thought that all immunoglobulin molecules in normal serum are antibodies resulting from contact with unknown antigens. The capacity to form the repertoire of 105 to 107 different antibody specificities is generally considered to be genetically determined in all vertebrates. During embryonic development lymphocyte precursors differentiate, the process ending with each lymphocyte having the capacity to form one kind of antibody combining site. Synthesis and secretion of antibody on contact with antigen is a highly regulated process involving interactions between different types of lymphocytes that may enhance or suppress their maturation into antibody-secreting plasma cells. Hapten: A substance, generally of low molecular weight, that, when injected, does not induce the formation of antibodies; it can react with antibodies induced to it when it is coupled to a protein, polypeptide, or other substance to form an antigen. It is also used to describe an antigenic determinant, that portion of a complete antigen which enters into the combining site of antibodies, the formation of which the antigen containing it has triggered. Lectin: A plant or animal protein having a receptor site specific for a sugar or oligosaccharide unit. METHODS IN ENZYMOLOGY, VOL. 70
Copyright © 1980 by Academic Press, Inc. All fights of reproduction in any form reserved. ISBN 0-12-181970-1
4
PRINCIPLES AND METHODS
[1]
General Considerations It was only when a basic understanding of the nature of antigen-antibody interaction became clear in studies begun over half a century ago by Michael Heidelberger and his school, 1-e that it became possible to apply these insights in a conceptual manner to the development of the highly sensitive array of analytical immunochemical methods now available to molecular and cellular biologists, clinicians, and scientists in other disciplines. Although the principles originally involved quantitative studies of the precipitin and agglutination reactions using the heterogeneous populations of antibodies generally produced in animals by hyperimmunization, 5-1° they have since found direct application to the study of the interactions of lectins 1H3 with polysaccharides and glycoproteins, of immunoglobulins with protein A of staphylococcus 14 and to the monoclonal antibodies present in sera of humans with multiple myeloma or Waldenstr6m's macroglobulinemia ~5 and of BALB/c and NZB mice with plasmacytomas induced with mineral 0iU6-2°; they will soon be applied to
1 M. Heidelberger, Chem. Rev. 24, 323 (1939). 2 M. Heidelberger, Bacteriol. Rev. 3, 49 (1939). 3 M. Heidelberger, "Lectures in Immunochemistry." Academic Press, New York, 1956. 4 E. A. Kabat and M. M. Mayer, "Experimental Immunochemistry," 1st ed. Thomas, Springfield, Illinois, 1948. 5 E. A. Kabat, "Kabat and Mayer's Experimental Immunochemistry," 2nd ed. Thomas, Springfield, Illinois, 1961. e E. A. Kabat, "Structural Concepts in Immunology and Immunochemistry," 2nd ed. Holt, New York, 1976. 7 M. Sela, ed., "The Antigens," Vol. l (1973); Vol. 2 (1974); Vol. 3 (1975); Vol. 4 (1977). Academic Press, New York. 8 C. A. Williams and M. W. Chase, eds., "Methods in Immunology and Immunochemistry," Vol. 1 (1967); Vol. 2 (1968); Vol. 3 (1971); Vol. 4 (1977); and Voi. 5 (1976). Academic Press, New York. a j. Garvey, N. E. Cremer, and D. H. Sussdoff, "Methods in Immunology," 3rd ed. Benjamin, Reading, Massachusetts, 1977. 10 D. M. Weir, "Handbook of Experimental Immunology" Vol. 1, "Immunochemistry," 3rd ed. Blackwell, Oxford, 1978. 11 I. J. Goldstein and C. E. Hayes, Adv. Carbohydr. Chem. Biochem. 35, 127 (1978). 12 E. A. Kabat, J. Supramol. Struct. 8, 79 (1978). 13 M. E. A. Pereira and E. A. Kabat, Crit. Rev. Immunol. 1, 33 (1979). 14 G. Mota, V. Ghetie, and J. Sj6qvist, Immunochemistry 15, 639 (1978). 15 R. J. Slater, S. M. Ward, and H. G. Kunkel, J. Exp. Med. 101, 851 (1955). 18 M. Potter, Adv. Immunol. 25, 141 (1977). 17 E. A. Kabat, Adv. Protein Chem. 32, 1 (1978). 18 M. Leon, N. M. Young, and K. R. Mclntyre, Biochemistry 9, 1023 (1970). 19 j. Cisar, E. A. Kabat, J. Liao, and M. Potter, J. Exp. Med. 139, 159 (1974). zo C. P. J. Glaudemans, Adv. Carbohydr. Chem. 31, 313 (1975).
[1]
PRINCIPLES OF A N T I G E N - A N T I B O D Y
REACTIONS
5
hybridomas (pages 34-35). 2m~ Quantitative precipitin reactions and inhibition by haptens of quantitative precipitin reactions and more sensitive methods, such as competitive binding assays, have become indispensable to the elucidation of the topology of the specific receptor sites on monoclonal antibodies, lectins, and other biologically active proteins. Indeed, in the absence of X-ray crystallographic data they offer perhaps the only approach to the understanding of the specificities of antibodies and lectins. To the extent that X-ray crystallographic s t u d i e s ~3-32, ec 16,17 have been carried out on these substances, they corroborate fully the inferences as to site size and structure made from the immunochemical studies. The earliest analyses of washed specific precipitates of hemoglobin and its antibody were carried out in Peking by H. Wu et al.,33 who showed that both hemoglobin and antibody were contained in the precipitate. A thorough examination of the course of the precipitin reaction was made by Heidelberger and Kendall,34-36 who first studied the reaction of the nitrogen-free specific capsular polysaccharide of the type III pneumococcus with horse type III antipneumococcal sera and subsequently several protein-antiprotein reactions 37,3s,el. 1-6 and established that antigen and anti21 G. K6hler and C. Milstein, Eur. J. lmmunol. 6, 511 (1976). 22 F. Meichers, M. Potter, and N. L. Warner, eds., "Lymphocyte Hybridomas. Second
Workshop on Functional Properties of Tumors ofT and B Lymphocytes." Springer-Verlag, Berlin and New York, 1978. 23 M. Schiffer, R. L. Girling, K. R. Ely, and A. B. Edmundson, Biochemistry 12, 4260 (1973). 2~ A. B. Edmundson, K. R. Ely, R. L. Girling, E. E. Abola, M. Schiffer, F. A. Westholm, M. D. Fausch, and H. F. Deutsch, Biochemistry 13, 3816 (1974). 25 R. J. Poljak, L. M. Amzel, H. P. Avey, B. L. Chen, R. P. Phizackerly, and F. Saul, Proc. Natl. Acad. Sci. U.S.A. 70, 3305 (1973). 2e D. M. Segal, E. A. Padlan, G. H. Cohen, S. Rudikoff, M. Potter, and D. R. Davies, Proc. Natl. Acad. Sci. U.S.A. 71, 4298 (1974). 2r O. Epp, P. Colman, H. Fehlhammer, W. Bode, M. Schiffer, and R. Huber, Eur. J. 'Biochem. 45, 513 (1974). 2s D. R. Davies, E. A. Padlan, and M. Segal, Contemp. Top. Mol. Immunol. 4, 127 (1975). 29 E. A. Padlan, Q. Rev. Biophys. 10, 35 (1977). 30 F. A. Saul, L. M. Amzei, and R. J. Poljak, J. Biol. Chem. 253, 585 (1978). 31 K. W. Hardman and C. F. Ainsworth, Biochemistry 15, 1120 (1976). 32 j. W. Beeker, G. N. Reeke, Jr., B. A. Cunningham, and G. M. Edelman, Nature (London) 259, 406 (1976). 3a H. Wu, L. H. Cheng, and C. P. Li, Proc. Soc. Exp. Biol. Med. 25, 853 (1927). M. Heidelberger and F. E. Kendall, J. Exp. Med. 50, 809 (1929). 35 M. Heidelberger and F. E. Kendall, J. Exp. Med. 55, 555 (1932). 36 M. Heidelberger and F. E. Kendall, J. Exp. Med. 61,563 1935). 37 M. Heidelberger and F. E. Kendall, J. Exp. Med. 62, 697 (1935). 3a E. A. Kabat and M. Heidelberger, J. Exp. Med. 66, 229 (1937).
6
PRINCIPLES AND METHODS
[1]
body combine in multiple proportions. Unlike the usual reactions of simpler compounds that combine in multiple proportions, forming complexes of defined composition, the multivalence of the antigen and the bior multivalence of the antibody or lectin resulted in a smooth curve when increasing quantities of antigen (or macromolecule) were added to a given quantity of antibody (or lectin, etc,). This is the typical quantitative precipitin curve.37, cf. 1-6 It should be remembered that these studies were carried out before anything was known of the valence of antibody and antigen and that the bi- or multivalence of antibody and the multivalence of antigen were key assumptions. Multivalence of antigen is now recognized as a consequence of the occurrence of 1. Repeating units of linear polysaccharides 2. Multiple terminal nonreducing sequences in branched polysaccharides 3. Projecting sequences of amino acids in synthetic polypeptides or of nucleotides in polynucleotides, DNA, or RNA ct 7 4. Various groups of known structure introduced chemically onto polysaccharides, polypeptides, or proteins cf. 5-10 5. Surface patches arising conformationally in various molecules, notably polypeptides39-42 and p r o t e i n s , 43,44 a s helices,/3 sheets, turns, and connecting random coils el"45 Also of importance in understanding antigen-antibody interations are 6. The existence of several different antigenic determinants on the surface of macromolecules such as proteins 43"44 and on complex glycoproteins such as the water-soluble blood group glycoproteins6"46"4r; these often are found in more than a single copy per molecule. Antibodies may be formed to the various antigenic determi39 E. A. Kabat, J. lmmunol. 97, 1 (1966). 4o j. W. Goodman, Immunochemistry 6, 139 (1969). 41 j. W. Goodman, in "The Antigens" (M. Sela, ed.), Vol. 3, p. 127. Academic Press, New York, 1975. 42 M. Sela, Science 166, 1365 (1969). 43 M. J. Crumpton, in "The Antigens" (M. Sela, ed.), Vol. 2, p. 1. Academic Press, New York, 1974. 44 M. Z. Atassi, ed., "Immunochemistry of Proteins," VoE 1 (1977); Vol. 2 1978. Plenum, New York. 45 R. E. Dickerson and I. Geis, "Structure and Action of Proteins." Harper, New York, 1969. 46 E. A. Kabat, in "Chemistry of Carbohydrates in Solution" (H. S. Isbell, ed.), Am. Chem. Soc. Adv. Chem. Ser. 117, 334 (1973). 4r T. Feizi, E. A. Kabat, G. Vicari, B. Anderson, and W. L. Marsh, J. lmmunol. 106, 1578 (1971).
[1]
PRINCIPLES OF ANTIGEN--ANTIBODY REACTIONS
7
nants giving rise to complex populations of antibodies of different specificities (heterogeneity of antibodies). 7. The occurrence on cells, cell membranes, and liposomes of glycolipid, glycoprotein, and lipoproteins, which provide multiple repeats of a given antigenic determinant and lead to aggregation reactions such as agglutination ,6 and of movement in the cell membrane on interaction with antibody with formation of patches and caps of antigen-antibody aggregates on individual cells. 4a It should be remembered that accessibility to the antibody or lectin combining site is an absolute requirement for an antigenic determinant or for the carbohydrate moiety to react with the lectin and that they must therefore be at the surface of the molecule. Denaturation of proteins, uncoiling of helical polypeptides, proteins, and glycoproteins, unwinding of double helices or of synthetic polynucleotides, dissociation of quaternary structure or partial enzymic digestion 4a'5° often expose new antigenic 51,52 determinants. Much useful information is obtainable from such studies. Determinants inaccessible 51-53 in the native structure are often termed hidden antigenic determinants. The assumption of bi- or multivalence of antibodies has been amply verified. 6'~ IgG, I g E ) s and presumably IgD are bivalent; electron micrographs show that IgA may exist as a bivalent molecule or as a tetravalent dimer, and IgM in various species may be a tetramer, 56'5Tpentamer) 5 or hexamer 5s with valences of 8, 10, and 12, respectively, although all of these may not be available for reaction with antigen simultaneously. Bivalence of IgG has been established by equilibrium dialysis, 59,6° fluores-
4s R. B. Taylor, P. H. Duffus, M. C. Raft, and S. de Petris, Nature (London) New Biol. 233, 225 (1971). 49 C. Lapresle, Ann. Inst. Pasteur 89, 654 (1955). 5o M. Raynaud and E. H. Relyfeld, Ann. Inst. Pasteur 97, 636 (1959). 51 C. Lapresle and J. Durieux, Ann. Inst. Pasteur 92, 62 (1957); 94, 38 (1958). 52 C. Lapresle and J. Durieux, Bull. Soc. Chim. Biol. 39, 833 (1957). 5a C. K. Osterland, M. Harboe, and H. G. Kunkel, Vox Sang. 8, 133 (1963). D. Beale and A. Feinstein, Q. Rev. Biophys. 9, 135 (1976). 55 S.-E. Svehag, in "Specific Receptors, Antibodies, Antigens, and Cells" (Third Int. Convocation on Immunol. Buffalo, N.Y., 1972), p. 80. Karger, Basel, 1973. 5~ E. M. Shelton and M. Smith, J. Mol. Biol. 54, 615 (1970). 57 R. T. Action, P. F. Weinheimer, S. J. Hall, W. F. Niedermeyer, E. Shelton, and J. C. Bennett, Proc. Natl. Acad. Sci. U.S.A. 68, 107 (1971). 58 R. M. E. Parkhouse, B. A. Askonas, and R. R. Dourmashkin, Immunology 18, 575 (1970). 5g H. N. Eisen and F. Karush, J. Am. Chem. Soc. 71, 363 (1949). •0 F. Karush, J. Am. Chem. Soc. 79, 3380 (1957).
8
PRINCIPLES AND METHODS
[1]
PAPAIN SPL|TS
F~G. 1. Schematic view of four-chain structure of human IgGx molecule. Numbers on right side are actual residue numbers in protein Eu. eT,eaNumbers of Fab fragment on left side are aligned for maximum homology; light chains are numbered as in Wu and Kabat e9 and Kabat and Wu. TM Heavy chains of Eu have residue 52A and 3 residues 92A, B, C; they lack residues 100A, B, C, D, E, F, G, H, and 35A, B. Thus residue 110 (end of variable region) is 114 in actual sequence. Hypervariable regions or complementarity-determining segments or regions (CDR) are shown by heavier lines. VL and Va: light chain and heavy chain variable region; Cnl, Ca2, and C,3: domains of constant region of heavy chain; CL: constant region of light chain. Hinge region, in which two heavy chains are linked by disulfide bonds, is indicated approximately. Attachment of carbohydrate is at residue 297. Arrows at residues 107 and l l0 denote transition from variable to constant region s. Sites of action of papaln and pepsin and locations of a number of genetic factors are given. Modified from Kabat 7~ in Kabat.17
[1]
PRINCIPLES OF ANTIGEN--ANTIBODY REACTIONS
9
cence quenching, el ultracentrifugation,~ enzymic cleavage with papain ~ and pepsin ~ and by X-ray crystallographice5 and electron microscopic ~ studies. These together with sequencing studies have led to an understanding of the three-dimensional structure of IgG (Figs. 1-3).°7 - n Figure 4 73-75 shows that all classes of immunoglobulins are built with a structure similar to that of human IgG, the multimeric forms of IgA and IgM having additional SH units and a third chain, the J chain, TM which serves to link them.
Sizes and Shapes of Antigenic Determinants A n t i b o d y c o m b i n i n g sites c o m p r i s e o n e o f t h e m o s t u n i q u e r e c o g n i t i o n s y s t e m s k n o w n . T h e t o t a l i t y o f t h e d a t a i n d i c a t e t h a t an a n t i b o d y c o m b i n ing site m a y b e o b t a i n e d t h a t c a n r e c o g n i z e a n y k i n d o f o r g a n i c c o m p o u n d r a n g i n g in size f r o m a l o w e r limit o f a b o u t 4 - 6 / ~ , to an u p p e r limit o f a b o u t 3 4 / ~ , in m o l e c u l a r w e i g h t f r o m p e r h a p s 200 to a b o u t 1000, a n d in a n y s h a p e o r c o n f o r m a t i o n t h a t s u c h an a m o u n t o f m a t t e r c a n a s s u m e . 6,39-41 A n t i b o d y c o m b i n i n g sites as d e t e r m i n e d f o r a n t i p o l y s a c c h a r i d e anti-
8~ S. F. Velick, C. W. Parker, and H. N. Eisen, Proc. Natl. Acad. Sci. U.S.A. 46, 1470 (1960). 62 H. K. Schachman, L. Gropper, S. Hanlon, and F. Putney, Arch. Biochem. Biophys. 90, 175 (1963). R. R. Porter, Biochem. J. 73, 119 (1959). A. Nisonoff, F. C. Wissler, and D. L. Woernley, Biochem. Biophys. Res. Commun. 1, 318 (1959). e5 V. R. Sarma, E. W. Silverton, D. R. Davies, and W. D. Terry, J. Biol. Chem. 246, 3753 (1971). R. C. Valentine and N. M. Green, J. Mol. Biol. 27, 615 (1967). e7 G. M. Edelman, B. A. Cunningham, W. E. Gall, P. D. Gottlieb, U. Rutishauser, and M. J. Waxdai, Proc. Natl. Acad. Sci. U.S.A. 63~ 78 (1969). ea G. M. Edelman, Biochemistry 9, 3197 (1970). eg T. T. Wu and E. A. Kabat, J. Exp. Med. 132, 211 (1970). 70 E. A. Kabat and T. T. Wu, Ann. N.Y. Acad. Sci. 190, 382 (1971). rl E. A. Kabat, in "Specific Receptors, Antibodies, Antigens, and Cells" (Third Int. Convocation Immunol., Buffalo, N.Y., 1972))' p. 4. Karger, Basel, 1973. T~E. M. Silverton, M. A. Navia, and D. R. Davies, Proc. Natl. Acad. Sci. U.S.A. 74, 5140 (1977). 7s j. A. Gaily, in "The Antigens" (M. Sela, ed.), Vol. 1, p. 161. Academic Press, New York, 1973. 74 B. Frangione, in "Immunogenetics and Immunodeficiency." Univ. Park Press, Baltimore, Maryland, 1975. 75 F. W. Putnam, G. Florent, C. Paul, T. Shinoda, and A. Shimizu, Science 182, 287 (1973). r6 M. E. Koshland, Adv. Immunol. 20, 41 (1975).
FIG. 2. Stereoview of the three-dimensional structure of human IgG myeloma protein Dob. The smaller circles represent a-carbon atoms; the larger circles represent carbohydrate hexose units. The Fab arms of the molecule are aligned vertically, and a horizontal twofold axis of symmetry bisects the molecule through the Fc. In this view the light chain is in the foreground of the upper Fab, and the heavy chains in the foreground of the lower Fab compare with Fig. 1. From Silverton et al. TM
FIG. 3. Space-filling view of the Dob IgG molecule. One complete heavy chain is white, and the other is clark gray; the two light chains are lightly shaded. The large black spheres represent the individual hexose units of the complex carbohydrate. In this view the twofold axis of symmetry is vertical. A crevasse is seen between Ca2 of the white heavy chain and the CL domain of the Fab on the left. From Silverton et al. TM
[1]
PRINCIPLES OF A N T I G E N - A N T I B O D Y REACTIONS Human IgG z Mouse Ig2b
Human IgG 1
Human IgG4, IgAt IgAz, A2m(2 )
Human IgG 3
Guinea pig IgG2 Mouse IgGz,
Rabbit IgG
Human IgA2, A2m(I ) Balb/c Mouse IgA
Human IgD
11
Human IgM Human IgA Dimer
,$
I
~
I-I
t
~
FIG. 4. Chain structure and disulfide bonding patterns in immunoglobulins. IgG, IgD, and IgA monomers are from Gaily73; IgM and IgA dimers are based on Frangione74; - - S - - S - bonds of IgM after Putnam et al. 75 From Kabat. a
bodies may be grooves or cavities 77-sz depending upon whether terminal nonreducing ends or internal linear sequences are recognized. Binding of the antigenic determinant is noncovalent and involves hydrophobic and hydrogen bonding, charge interaction, etc., the total binding energy being due to the sum of such interactions and the fit of the antigenic determinant r7 j. Cisar, E. A. Kabat, M. M. Dorner, and J. Liao, J. Exp. Med. 142, 435 (1975). 7a K. Takeo and E. A. Kabat, J. lmmunol. 121, 2305 (1978). 79 A. Wu, E. A. Kabat, and M. G. Wiegert, Carbohydr. Res. 66, 243 (1978). so G. Schepers, Y. Blatt, K. Himmelsbach, and I. Podit, Biochemistry 17, 2239 (1978). al L. G. Bennett and C. P. J. Glaudemans, Carbohydr. Res. 72, 315 (1979). 8, W. Schalch, J. K. Wright, L. S. Rodkey, and D. G. Braun, J. Exp. Med. 149, 923 (1979).
12
PRINCIPLES AND METHODS
[1]
in the antibody combining site. e,83"~ Binding is usually measured as an association constant, Ks, the values varying over a wide range for different antigenic determinants and even for a single antigenic determinant; association constants tend to be of the order of about 104 to 108 M -1 for carbohydrate d e t e r m i n a n t s 77,Ta,8°-a2,84-s7 although a proportion of the antibody in antisera to the group-specific carbohydrate of the streptococcus has been found to have a Ka of 109.82,88,89The Ka value may reach 10l° or more for hydrophobic structures, such as the dinitrophenyl group, 84 for digoxin and fluorescein, and for proteins, such as insulin?° Antigenic determinants of proteins generally have Ka values in the range of 105 to 10s.91-94 For a detailed analysis and additional references, see Karush. s4 To the extent to which they have been compared, enzyme sites and antibody combining sites cover comparable ranges of sizes and shapes. The relative contribution of each sugar to the binding has been measured for the lysozyme site. 95,96 The lysozyme site has been found by X-ray crystallography97 to be a groove accommodating the hexasaccharide of the bacterial cell wall built on alternating N-acetylmuramic acid and Nacetyl-D-glucosamine residues; the myeloma antidextran QUPC52 has also been found to be a groove complementary to an internal chain of six a-(I-->6) linked glucoses, isomaltohexaose. 1s,77 Both these sites are at about the upper limit in size. At the other end of the scale, glycosidases may split a terminal sugar, and antibodies 9s,99,ef-e,39-41 and lectinsef. H-13.100.10~ may react with a single sugar unit plus a portion of the second sugar. s3 F. Karush, Adv. Immunol. 79, 3380 (1962). 84 F. Karush, in "Immunoglobulins" (S. Litman, G. Ward, and R. A. Good, eds.), p. 85. Plenum, New York, 1978. a5 j. W. Kimball, lmmunochemistry 9, 1169 (1972). a6 J.-C. Jaton, H. Huser, W. F. Riesen, J. Schlesinger, and D. Givol, J. Immunol. 116, 1363 (1976). a7 D. G. Strcefkirk and C. P. J. Glaudemans, Biochemistry 16, 3760 (1977). s s j. K. Wright, W. Schalch, L. S. Rodkey, and D. G. Braun, FEBS Lett. 93, 317 1978). s9 W. Schalch, J. K. Wright, L. S. Rodkey, and D. G. Braun, Fur. J. Immunol. 149, 923 (1979). 9o S. Berson and R. S. Yalow, J. Clin. Invest. 38, 1996 (1956). 91 D. H. Sachs, A. N. Schechter, A. Eastlake, and C. B. Anfinsen, Biochemistry 11, 4268 (1972). W. B. Dandliker and S. A. Levison, lmmunochemistry 1, 165 (1968). 93 N. Sakato, H. Fujio, and T. Amano, Biken J. 14, 405 (1971). 94 I. Pecht, E. Maron, R. Arnon, and M. Sela, Fur. J. Immunol. 19, 368 (1971). 95 j. A. Rupley, Proc. R. Soc. London Ser. B 167, 416 (1967). D. M. Chipman, V. Grisaro, and N. Sharon, J. Biol. Chem. 242, 4388 (1967). 97 D. C. Phillips. Sci. Am. 215 (11), 78 (1966). a s y . Arakatsu, G. Ashwell, and E. A. Kabat, J. Immunol. 97, 858 (1966). A. M. Staub and R. Tinelli, Bull, Soc. Chim. Biol. 42, 1637 (1960).
[1]
PRINCIPLES OF A N T I G E N - A N T I B O D Y
REACTIONS
13
T h e Quantitative Precipitin Curve At the time that the quantitative precipitin curve was originally developed, 1-6,34-3s analyses of the washed specific precipitates were carried out by the micro-Kjeldahl procedure and the working range was about 0.10-1.0 mg of total N. Improvements in analytical methods in the intervening half century have reduced the quantities needed per sample to about 1-6/.~g of total N in the precipitate, 1°2 analyses on the washed precipitate being carried out after a Kjeldahl-type digestion followed by the ninhydrin reaction. 1°3 It is doubtful whether smaller amounts can be used, since the "solubility"l°3~ of p01ysaccharide-antibody precipitates at 0 ° is appreciable, values being 0.6-1.0/~g of N per milliliter for horse; 0.71.8/~g of N per milliliter for human, and 3-7/~g of N per milliliter for rabbit antibodies; for protein-antiprotein precipitates, values are 3-10 ftg of N per milliliter. 5 Polysaccharide-lectin precipitates 1°4,1°5 fall into the same range. Thus, in carrying out quantitative precipitin determinations with an upper range of 6 - 8 ftg of N of Ab, lectin, etc., per determination, it is important to work in total volumes of 0.5 ml or less to minimize these effects and to wash the precipitates with small volumes of saline at 0°. Since the reaction mixtures are left at 0 - 4 ° for 5 - 7 days for equilibrium to be reached, and since for most purposes comparative data are needed, results tend to be quite reproducible, but they may be low by an amount corresponding to the "solubility" effect plus small losses in washing due to solubility of the specific precipitate. Quantitative
Precipitin
Curves
Procedure. 5,102,103Determinations on a micro scale are performed in 3ml conical Pyrex centrifuge tubes. A volume of antiserum (containing antibody or myeloma antibody or crude seed extract or purified lectin) previously centrifuged until it no longer deposits sediment and containing about 6 - 8 ~g of antibody N (AbN) or lectin N in a volume of about 50100/.d is added to tubes containing a suitable range of accurately mea-
K. D. Hardman, in "Carbohydrate-Protein Interaction" (I. J. Goldstein, ed.), Am. Chem. Soc. Syrup. Ser. 88, 12 (1979). ~o~ M. Sarkar, J. Liao, E. A. Kabat, T. Tanabe, and G. Ashwell, J. Biol. Chem. 254, 3170 (1979). 102 E. A. Kabat and G. Schiffman, J. lmmunol. 88, 782 (1962). loa G. Schiffman, E. A. Kabat, and W. Thompson, Biochemistry 3, 113 (1964). ~oza ,, Solubility" refers to the effects of carrying out the quantitative precipitin assays with the same amounts of antigen and antibody but varying the total volume with saline. ~04 M. E. Etzler and E. A. Kabat, Biochemistry 9, 869 (1970). ~o5 M. E. A. Pereira, E. A. Kabat, and N. Sharon, Carbohydr. Res. 37, 89 (1974). 1oo
14
PRINCIPLES AND METHODS
[1]
sured quantities of the antigen or substance reacting with lectin. The total volume is adjusted with saline. The contents of each tube are mixed, and the tubes are placed at 37° for 1 hr (or at room temperature with lectins) and then at 4°. The contents of the tubes are mixed twice daily. After 5-7 days they are centrifuged in a refrigerated centrifuge, the supernatants are decanted, and the precipitates are allowed to drain thoroughly with the tubes inverted on a towel and leaning against a rack. The precipitates are washed with 0.5-ml portions of saline (0.9% NaCI solution) at 0°. They are then analyzed for N by the ninhydrin method after a Kjeldahl-type digestion with H2SO4. l°a For work with cold agglutinins47'1°~ all reagents are chilled; the setup is made in an ice bath, and the tubes are kept in ice water for the entire time before washing the precipitates; washing of the precipitates is performed in a cold room. This is also necessary when carrying out quantitative precipitin determinations using horse antisera and cross-reacting polysaccharides s because of their much higher solubility with increasing temperature. Ninhydrin Procedure for Quantitative Precipitin and Quantitative Inhibition Determinations l°a
Reagents 1. Digestion mixture: 1 ml of concentrated H2SO4 diluted to 20 ml 2. H~O2, 30% 3. Standard (NH4)2SO4 solution: 30.0 mg of (NH4)2SO4 in 3.00 ml of distilled water. This stock solution when diluted 1 : 10 for routine use contains 212 ~g of N per milliliter. 4. Sodium acetate buffer, 4 M pH 6.5:136 g of NaAc.3H~O are dissolved in 100 ml of H20 in a hot water bath and allowed to cool; 25 ml of glacial acetic acid are added, and the volume is made to 250 ml with distilled H~O. Adjust to pH 6.5 with NaOH. Store at 4° without preservative. The saturated solution must be warmed to dissolve NaAc before use. 5. KCN, 10 mM 6. Ninhydrin solution: 160 mg of ninhydrin are dissolved in 3 ml of ethylene glycol monomethyl ether plus 1 ml of 4 M acetate buffer, pH 6.5. 7. Prepared just before use by adding 100/~1 of reagent 5 to 4 ml of reagent 6. 1oe M. Heidelberger and A. C. Aisenberg, Proc. Natl. Acad. Sci. U.S.A. 39, 453 (1953).
[1]
PRINCIPLES OF ANTIGEN--ANTIBODY REACTIONS
15
Procedure. Add 25/zl of reagent 1 to each washed specific precipitate in the conical 3-ml centrifuge tubes. (NI-L)2SO4 standards containing 2, 3, and 4 p,g of N are set up at the same time. The samples are digested in a sand bath at 160° for 90 min. Cool, add 15/zl of 30% H20~ to each tube, and redigest the samples in the sand bath at 160°C for an additional 90 minutes. Cool, add 200/~1 of distilled water and 100 tzl of reagent 7 to each tube. Mix gently and place tubes in a water bath at 95° for 20 min. Cool. Add 1.5 ml of 50% ethanol and transfer to a 10-ml volumetric flask; wash tube twice with 1.5 ml of 50% ethanol and bring to the mark with 50% ethanol. Read absorption at 570 nm. The conditions of digestion and assay are chosen so that equal absorption values on a molar basis are given by (NH4)2SO4 and amino acids. Figure 5 and the table give representative data on the precipitation reaction of crystalline hen egg albumin with rabbit anti-egg albumin as studied in 1935 by Heidelberger and Kendall. 3r They provide more insight into the general course of the precipitin reaction than is often seen with the currently used, more micro, procedure given in the preceding paragraph. Not only was the precision substantially greater, solubility effects generally being negligible, but it was standard practice to examine supernatants from each of duplicate determinations for the presence of antigen or antibody. This was generally accomplished by adding antibody or antigen to a portion of each supernatant and examining the tubes for precipitate. Three zones, were recognized: a zone of antibody excess, an equivalence zone in which neither antigen nor antibody was detectable, and a zone of antigen excess. The point on the curve at which free antigen was first seen corresponded to the point of maximum precipitation: with larger excesses of antigen, the amount of antigen-antibody precipitate decreased owing to the formation of soluble complexes, and this was termed the inhibition zone. Tests on supernatants are often carried out by gel diffusion techniques that correlate well with the usual tests and may give improved sensitivity.5,1°7 With the more micro method using 6 - 8 / z g of antibody N per determination and the keeping of tubes in the refrigerator for a week, supernatant tests are often not practical, since the small quantities of precipitate would not be seen. For most purposes one is interested in determining specificity differences by comparing various precipitin curves one with another. With polysaccharide antigens of high molecular weight, and with systems involving IgM antibody, the inhibition zone is reached only with amounts of antigen many times those required for maximum precipitation, whereas with protein antigens, with carbohydrate antigens of lower molecular weight, and with IgG antibodies, inhibition by 1o7 j. Munoz and E. L. Becker, J. lmmunol. 65, 47 (1950).
16
PRINCIPLES AND METHODS
[1]
Ratio AbN: EaN in precipitate
,.=
o 0
< p~ 0 ,.G
:o ~
If
///,, E
o
.-,_
A
a~
~f f '6
z
\ol!
°°"
d
/
m
iV
g~
0 o
--
o
0,/, ~~\. oo
o
~
(6#) peze|!d!0eJd N
o o~=
[1]
PRINCIPLES OF ANTIGEN--ANTIBODY REACTIONS
17
ADDITION OF INCREASING AMOUNTS OF EGG ALBUMIN TO 1.0 ML OF A 1:2 DILUTION OF RABBIT ANTISERUM TO EGG ALBUMIN AT 0 °a'b
EaN added
(~g) 9.1 15.5 25 40 50 65 74 82 90 98 124 135 195 307 490
Antibody Total N by Ratio AbN: EaNpptd Npptd difference EaN in (~g) (p.g) (/~g) precipitate Total Total Total Total Total Total Total Total 87 89 87 [72]~ [48] [4]
156 236 374 526 632 740 794 830 826 820 730 610 414 106 42
147 220 349 486 582 675 720 748 739 731 643 [538] [366]
Antibody Npptd calculated from equation Tests on supernatant (p.g)
16.2 14.2 14.0 12.2 11.6 10.4 9.7 9.1 8.5 8.2 7.4 7.5 [7.6]
137 225 343 499 582 677 714 738 746
Excess Ab Excess Ab Excess Ab Excess Ab Excess Ab Excess Ab No Ab or Ea No Ab, 1/zg of EaN Excess Ea Excess Ea Excess Ea Excess Ea Excess Ea Excess Ea Excess Ea
a Data from M. Heidelberger and F. E. KendallY 0 Ab, antibody; Ea, egg albumin. c Values in brackets are considered to be uncertain.
excess antigen occurs much more rapidly. Further details, representative data, curves, and procedures may be found in Kabat 5 and in Maurer. l°s This review will concentrate on more recent studies using quantitative precipitin data, especially those leading to insights about structure and to the development of more sensitive methods o f assay using immunochemical reagents. With certain horse antitoxins and antiprotein sera '°a-m and in some patients with Hashimoto's thyroiditis 112 who have antibodies to thyroglobulin, one finds a different type o f quantitative precipitation curve, termed a flocculation curve, of"5.6 Precipitation occurs only o v e r a narrow range, and soluble a n t i g e n - a n t i b o d y complexes are formed in the region o f antibody excess as well as of antigen excess. Flocculation curves have 10s p. H. Maurer, in "Methods in Immunology and Immunochemistry" (C. A. Williams and
M. W. Chase, eds.), Vol. 3, p. 1. Academic Press, New York, 1971. los A. M. Pappenheimer, Jr. and E. S. Robinson, J. I m m u n o l . 32, 291 (1937). Ho D. Gitlin,C. S. Davidson, and L. H. Wetterlow, J. Immunol. 63, 291 (1949). m E. H. Relyveld and M. Raynaud, Ann. Inst. P a s t e u r 96, 537 (1959). n2 I. M. Roitt, P. N. Campbell, and D. Doniach, Biochem. J. 69, 248 (1958).
18
PRINCIPLES AND METHODS
[1]
not been used for investigations of antigenic structure and will not be considered. With various kinds of antisera, the course of the precipitin reaction has been described by the equation AbN precipitated =
ax-
(1)
bx 2
where x is the amount of antigen or antigen N added and a and b are constants that differ from one antiserum to another. This equation was found empirically by Heidelberger and Kendall, 34 who subsequently derived it from the law of mass action? e' of. 1-3.H4 Dividing both sides by x gives (AbN)/x in the precipitate
=
a
-
bx
(2)
This is the equation of a straight line in which the ratio of AbN to antigen or antigen N in the precipitate is plotted against x; a is the intercept on the Y axis and - b is the slope (Fig. 5). The equation generally holds throughout the antibody excess region and equivalence zone up to the point of maximum precipitation and was quite useful, especially in comparing various antisera to the same antigen in view of the heterogeneous populations of antibody molecules formed. For systems involving a single antigen and its homologous antibodies, it has been shown that all the added antigen is precipitated throughout the antibody excess and equivalence zones and up to the point of maximum precipitation. 1-e With some homogeneous myeloma proteins and with lectins, one often finds a straight line, but with others the lines are not straight. The need for the equation for homogeneous antibodies from myelomas or hybridomas and for lectins is reduced, since if one uses a given amount of different preparations of the same myeloma antibody or of the same lectin the precipitin curves with the same antigen or glycoprotein are identical. IgA monomeric antibodies often do not precipitate with macromolecular antigens, apparently because of limited flexibility of their hinge, and thus may be effectively monovalent. As such they may under suitable circumstances attach to specific precipitates, and under other conditions they may inhibit precipitation of IgA polymer or of IgG antibodies; these phenomena were described H~,~6 long before the classes of immunoglobulins were recognized. The antibody was termed nonprecipitable antibody and was separated from the precipitating antibody by successive small additions of antigen, a procedure that competitively favored removal of precipitating antibody leaving the nonprecipitating antibody in the superna113 L, Pauling, D. H. Campbell, and D. Pressman, Physiol. Rev. 23, 203, (1943). 114 F. E. Kendall, Ann. N.Y. Acad. Sci. 43, 85 (1942). 115 A. M. Pappenheimer, Jr., J. Exp. Med. 71, 263 (1940). l~e M. Heidelberger, H. P. Treffers, and M. M. Mayer, J. Exp. Med. 71, 271 (1940).
[1]
PRINCIPLES OF ANTIGEN--ANTIBODY REACTIONS
19
tant. 4,5,3s Nonprecipitable antibody was initially hypothesized to be univalent; subsequent studies showed that the nonprecipitable fraction of equine antibody to p-azophenylactoside, which migrated as an IgA, had a valence of two, as determined by equilibrium dialysis against a small hapten, and a sedimentation constant of 7 S; the binding constants of the precipitating and nonprecipitating antibodies were of comparable affinities. 117 It was therefore proposed that the two sites were so distributed that when one site was occupied by a large antigen molecule the other site was sterically unable to react. The 7 S subunit of IgA mouse myeloma MOPC315 did not precipitate with DNP coupled to protein, nor did it agglutinate sheep erythrocytes to which DNP groups had been coupled.llS,119
More recent studies ~2° have shown that the 7 S subunit of MOPC315 would bind only one molecule of DNP-dextran of molecular weight 44,000, containing but one DNP group, but would bind two molecules of DNP coupled to ~-aminocaproic acid. Affinity labeling of varying proportions of DNP sites, measurement of residual sites by equilibrium dialysis, calculation of the theoretical valence for binding to the mw 44,000 univalent DNP dextran, and comparison with the valence observed, satisfied a steric model but not the alternative possibilities of an asymmetric T M or an allosteric model. Similar conclusions had also been drawn ~22for the IgM subunit (IgMs) of a human Waldenstrtm macroglobulin, Lay, which bound only one molecule of IgG. The observed valence of an IgM antibody decreased from 10 to 5 as the molecular weight of the antigen used to measure it increased from 342 to 7100 while the valence of IgG remained unchanged. 12z It is of interest that an IgG1K protein, Dob, TM had a 15-residue deletion in its hinge region, but from the X-ray data this did not seem to affect its flexibility; an explanation for the inability of the 7 S IgA and IgM subunits to bind more than a single macromolecule must await X-ray crystallographic studies on one of these proteins. Estimation of the Total Antibody Content of Antisera For some purposes, one's primary interest may not be in the quantitative precipitin curves as such, but one may wish merely to determine the 117 N. Klinman, J. H, Rockey, and F. Karush, Science 146, 401 (1964). 11s H. N. Eisen, E. S. Simms, and M. Potter, Biochemistry 7, 4126 (1968). 119 M. Potter, Physiol. Rev. 52, 631 (1972). 120 R. Eisenberg and P. Plotz, Biochemistry 17, 480 (1978). m S. I. Chavin and E. C. Franklin, Ann. N.Y. Acad. Sci. 168, 84 (1969). 122 M. J. Stone and ~ . Metzger, J. Biol. Chem. 243, 5977 (1968). 123 S. C. Edberg, P. M. Bronson, and C. J. Van Oss, lmmunochemistry 9, 273 (1972).
20
PRINCIPLES AND METHODS
[1]
antibody content of the antisera. This was originally done by locating the point of m a x i m u m precipitation (Fig. 5) and setting up one or m o r e determinations in this range (for details see K a b a t and MayerS). In practice, h o w e v e r , it is easier to couple the antigen to an i m m u n o a d s o r b e n t , 124"~2~ to add an amount sufficient to r e m o v e all the antibody, mix thoroughly, centrifuge, wash, elute the antibody with acid, and determine the quantity spectrophotometrically. This minimizes solubility effects and permits assays on a much smaller scale; using 100-/zl samples of serum, satisfactory results were obtained with antisera containing about 25 ~g of antib o d y per milliliter (6/~g o f A b N per milliliter); care must be taken to use excess i m m u n o a d s o r b e n t . E s t i m a t i o n of A n t i g e n F r o m Fig. 5 it m a y be seen that with a given antiserum the total N in the precipitate is a function of the quantity o f antigen N added. This relationship for each antiserum serves as a calibration curve, which can be used to determine the quantities of antigen in biological fluids. Thus the antiserum in Fig. 5 could be used to determine quantitatively the a m o u n t of egg albumin in egg white. It is only n e c e s s a r y to prepare and add to the m e a s u r e d volume o f antiserum a suitable dilution of the egg white such that the egg albumin it contains is sufficient to give an amount o f washed specific precipitate N falling in the antibody excess region o f the curve. The total N found is interpolated on the c u r v e to obtain the a m o u n t of egg albumin N in the volume of egg white dilution used. It is evident from examining the curve in Fig. 5 that two points would be found, one in the inhibition zone and one in the antibody excess zone. To be sure that one is getting the correct results, one must be certain that the determination was set up in the region of excess antibody; this was routinely established by tests on the supernatants as described above. On the microscale now used it would be necessary to set up several points to be sure that one was in the zone of antibody excess. Determinations o f antigen by the quantitative precipitin method were used routinely for m a n y years in s o m e institutions for the estimation of IgG in human cerebrospinal fluid, increases in cerebrospinal fluid IgG being found in multiple sclerosis and in neurosyphilis. 126-12s In recent 124A. E. Gurvich, R. B. Kapner, and R. S. Nezlin, Biokhimiya 24, 144 (1959); English translation, p. 129. 125T. J. Gill and C. F. Bernard, lmmunochernistry 6, 567 (1969). 126E. A. Kabat, M. Glusman, and V. Knaub, A m . J. Med. 4, 653 (1948). 12~E. A. Kabat, D. A. Freedman, J. P. Murray and V. Knaub, A m . J. Med. Sci. 219, 55 (1950). 128M. D. Yahr, S. S. Goldensohn, and E. A. Kabat, Ann. N . Y . Acad. Sci. $8, 613 (1954).
[1]
PRINCIPLES OF ANTIGEN--ANTIBODY REACTIONS
21
years more rapid micro procedures, such as competitive binding and fluorometric methods, have essentially replaced quantitative precipitin assays.
Use of Quantitative Precipitin Curves for Structural Insights Figure 6 shows results of quantitative precipitin determinations carried out under these conditions with four antidextran myelomas--three specific for a-(1-~6) linked dextrans W3434, W3129, and QUPC52 and the fourth, UPC102, for a-(l-*3) linked dextrans. 19 The various proportions of a-(1--~6), a-(1-~3)-like, and a-(1--~4)-like linkages in the dextrans are given129,la°; the terms (1-~3)-like and (1--,4)-like are used, since a glucose residue substituted on C-2 and C-4 will not take up periodate and will behave as though it were (1---3) linked and glucoses substituted either on C-2 or on C-4 will behave equivalently, each taking up 1 mol of periodate. When the actual proportions of each linkage are known from methylation T M or other s t u d i e s , 1a°,132,133 these values are often revised. The curves in Fig. 6 for the different myeloma proteins provide certain important insights into the uses of such quantitative data. The analytical precision of the method is clearly seen, in that several of the different dextrans give the same curve. In Fig. 6A for W3434, it is clear that the values for B1299 Fr.S, Bl141, B512, and B1424 all fall on one curve and values for B1498 Fr.S and B1355 Fr.S fall on a second curve. The better reacting groups are those with no or with low a-(1---~3)-like linkages, whereas the poorer groups are high in a-(1---~3)-like linkages. N150N, with a much lower molecular weight than B512 but of the same general structure, reacts much more poorly. The best dextran, B1399, gives somewhat higher results than the first group; this dextran is like the group reacting next best and having high a-(1--~4)-like linkages; these linkages have been shown to be largely a - ( l ~ 2 ) from measurements of optical rotation of the cuprammonium complexes 13zand by the isolation of kojibiose, ~33o-Glc-c~(1---~2)-D-Glc. With W3129, the difference between B1399 and the better reacting group of dextrans is not seen; N150N precipitates about 50% of the total ~9 A. Jeanes, W. C. Haynes, C. A. Wilham, J. C. Rankin, E. H. Melvin, M. J. Austin, J. E. Cluskey, B. E. Fisher, H. M. Tsuchiya, and C. E. Rist, J. Am. Chem. Sco. 76, 5041 (1954). ~30 R. L. Sidebotham, Adv. Carbohydr. Chem. Biochem. 30, 371 (1974). ~3~ F. R. Seymour, M. E. Slodki, R. D. Plattner, and A. Jeanes, Carbohydr. Res. 53, 153 (1977). ~3~ T. A. Scott, N. N. Hellman, and F. R. Senti, J. Am. Chem. Soc. 79, 1178 (1957). ~zz H. Suzuki and E. H. Hehre, Arch. Biochem. Biophys. 104, 305 (1964).
22
[1]
PRINCIPLES AND METHODS LIN KAGES I-hke ~ 3 I_4hk4
LINKAGES ; ~ 6 1"-'3 -hke 1~4 -Ilk e
I~6 O x Ill o
8512orN236 N-15ON Sll41 BI255 G ?42L-R
9G 96 79 86 el
4 4 J8 0 0
0 0 3 14 19
200,dvl W3434
W BI399 65 = Be299-S-3 5 0 Ig BI424 71
(diluted
1/10)
0 0 0
35 50 29
LINKAGES 1--~3 1 ~ 4 1~6 -like -like • e149es 6 2 m 81355-S'4 5 7 Io B742C-SR 61 • BtSOIS 6 5
27 34 21 20
II 9 18 15
Totol Volume.'350,~l
9
a; 6 5 4 3 2 I
o
fx
i
i
, if Or
9 8
I
I
L
J
i
//~'
I
I
i
,~l
50.Avl O U P C 5 2 ( d i l u t e d
1/10)
~.
I
L
I
I
I
T o t o l Volume.'2OO,~vl 0
g
6 5 4 3
I , / / ~ . / ~ _ ~
.
.~Ojul U P C 1 0 2
Tolol
0
~
~O
~-
m # m
Volume.'2OOjvl
15 20 2; 3lO " 40 ~g ANTIGEN ADDED
~C]
60
Fro. 6. Quantitative precipitin curves of mouse myeloma proteins with various dextrans. From Cisar et a l . 19
[1]
PRINCIPLES OF ANTIGEN--ANTIBODY
REACTIONS
23
AbN and the poorer reacting group of dextrans are more effective in precipitating relative to the better group than was seen with W3434. These differences suggest that the two myeloma antidextrans differ somewhat in their combining sites. QUPC52 can readily be seen to be quite different (Fig. 6C) from the other two myeloma antidextrans in that B512 and B1255, which have low proportions of non or-(1---->6)linkages, are very much better in precipitating than are the dextrans with higher proportions of non-a-(1---~6) linkages. QUPC52 differs from W3129 in that it precipitates well TM with a synthetic linear dextran T M whereas W3129 does not. The ability to precipitate with a linear dextran means that the dextran is multivalent and thus that internal sequences of several sugars must constitute an antigenic determinant. 77,Ta,la5 This, plus additional competitive binding data with oligosaccharide inhibitors that will be discussed later, led to the conclusion that the antibody combining site of QUPC52 must be a groove into which internal chains of a-(1--~6) linked glucoses fit. 77 Two other NZB myeloma antidextrans, PC3858 and PC3936, TM also precipitate with the synthetic linear dextran and thus also have groove type sites although their sites appear to be smaller than that of QUPC52. Fluorescence quenching measurements by Bennett and Glaudemans 81 have shown that monovalent Fab fragments of W3129 bind only to the terminal nonreducing ends of dextran molecules of molecular weight 36,000 and give the same Ka per terminal nonreducing end, as does isomaltopentaose. 77 The degree of precipitation with the synthetic linear dextran as compared with the usual native dextrans has been of value in examining antidextrans 77 produced in humans, 'zT"or. ~,6 and in rabbits 98''a6 and has shown that the heterogeneous populations of antidextran may be mixtures of molecules with specificities directed toward internal chains of a-(1---~6) linked residues as well as those with specificities directed toward terminal nonreducing ends of chains. Fractions of such antibodies separated by isoelectric focusing were found to differ in the proportions of their antibodies precipitable by the synthetic linear dextran. 77 Antibodies specific for the terminal nonreducing ends as well as for internal sequences of the group-specific A variant carbohydrate of the streptococcus, a polymer of L-rhamnose with ct-(1---~2) and a-(1---~3) linkages, have also been found in rabbit antisera. 82 Unlike the findings with human antidextran, the binding constants of the antibodies specific for internal sequences were of higher Ka than those directed toward the termi,34 E. R. Ruckel and C. S c h u e r c h , Biopolymers 5, 515 (1967). ,35 W. Richter, Int. Arch. Allergy 46, 438 (1974). ,3s I. M. O u t s c h o o r n , G. Ashwell, F. Gruezo, and E. A. K a b a t , J . lmmunol. 113, 896 (1974). ,37 E. A. Kabat and D. Berg, J. lrnmunol. 70, 514 (1953).
24
PRINCIPLES AND METHODS
[1]
hal nonreducing ends, and this was ascribed to hydrophobic interactions with the CH3 group of C-6 of rhamnose. It is of interest that the terminally specific antidextran ~7'7s and antistreptococcal group A variant molecules s2 had similar Ka for their respective determinants. The existence of linear polysaccharides that are antigenic, such as pneumococcal SIII, S V I I I , sS's6,138-14° S V I a , 141'142 essentially demand the existence of groove-type sites. The fourth myeloma antidextran (Fig. 6D) has an entirely different specificity. The precipitin curves show that it reacts best with three dextrans, B1498, B1355 FR.S, and B1501 FR.S, high in a-(1--*3)-like linkages. A second group, B1399 FR.S., B1255, B742 FR.C, precipitated about half as well, whereas a third group gave very little precipitate. This, plus inhibition data (see later), indicates that the specificity involves a(1-->3) linkages. Another a-(1---~3) specific IgM myeloma, MOPC104E, TM was found not to react at all with dextran B512 with 96% a-(1---~6) linkages, to react best with dextran B1355 FR.S, and to react less well with dextrans with smaller proportions of a-(1---~3) linkages; however, it also reacted with dextrans having high o~-(1---~2) and o~-(1---~4)linkages (Fig. 7). All of these precipitated the same maximum quantity of antibody, unlike the curves in Fig. 6D. Myeloma antibodies to fructosans could be divided into two groups: those specific for/3-(2--->1)-linked fructosans as evidenced by the ability of inulin, a linear/3-(2--> 1) linked polymer, to precipitate, and those specific for/3-(2--->6) linkages as seen from the reaction with ryegrass levan, a linear polymer of/3-(2-->6) linkages. The/3-(2--* 1) specific myeloma proteins did not react with ryegrass levan and the fl-(2--,6) specific myeloma proteins did not react with inulin. '9,a7,79 If one compares Figs. 6 and 7, it is apparent that with the IgA myelomas individual dextrans precipitate different total quantities of antibody N, whereas with the IgM antidextran myelomas all dextrans that can react, regardless of structure, will precipitate all the myeloma antidextran. This difference is ascribable to the presence.in the IgA myelomas of monomers and polymers, the monomers partially inhibiting precipitation as discussed earlier. This distinctive behavior of IgA myelomas holds for antifructosans.~7 With the IgG2a antifructosan myeloma UPC 10, all levans tested, including ryegrass levan, precipitated all the antibody when added ,as R. G. Mage and E. A. Kabat, Biochemistry 2, 1278 (1963). ,3a j. K. N. Jones and M. B. Perry, J. Am. Chem. Soc. 79, 2787 (1957). ~4o A. M. Pappenheimer, Jr., W. P. Reed, and R. Brown, J. Immunol. 150, 1237 (1968). ~41 p. A. Rebers and M. Heidelberger, J. Am. Chem. Soc. 81, 2415 (1959). m M. Heideiberger and P. A. Rebers, J. Bacteriol. 80, 145 (1960).
[1]
PRINCIPLES OF ANTIGEN--ANTIBODY I
I
REACTIONS
25
I
0.4
\ 0.3
~E a
4)-D-GlcNAc-/3-(1--~3)-D-Gal-fl-(l~4)-D-GlcNAc-fl-(l~3)-D-Gal-fl(1---~4)-D-Glc-fl-(1---~1)-ceramide, termed lacto-N-norhexaosylceramide, when incorporated on liposomes, but the structures involved in each specificity have not been further defined. Activity specific for anti-I Step 193 D. Roelcke, Clin. lmmunol, lmmunopathol. 2, 266 (1974). ~ T. Feizi, in "Human Blood Groups" (Fifth Int. Convocation Immunol.) p. 164. Karger, Basel, 1977, 195 R. R. Race and R. Sanger, "Blood Groups in Man," 6th ed. Blackwell, Oxford, 1975. ~9s W. T. J. Morgan, Proc. R. Soc. London Ser. B 151, 308 (1959). ~a7 W. M. Watkins, in "Glycoproteins" (A. Gottschalk, ed.), 2nd ed., p. 830. Elsevier, New York, 1972. 198 K. O. Lloyd, in MTP Int. Rev. Sci. Org. Chem. Ser. 2, Vol. 7. "Carbohydrates" (G. O. Aspinall, ed.), p. 251. Butterworth, London, 1975. ~aa G. Vicari and E. A. Kabat, J. Immunol. 102, 821 (1969). 2 0 o T. Feizi, E. A. Kabat, G. Vicari, B. Anderson, and W. L. Marsh, J. Exp. Med. 133, 39 (1971). 201 K. O. Lloyd and E. A. Kabat, Proc. Natl. Acad. Sci. U.S.A. 61, 1470 (1968). 202 G. Vicari and E. A. Kabat, Biochemistry 9, 3414 (1970). ~03 T. Feizi and E. A. Kabat, J. Exp. Med. 135, 1247 (1972). 2o4 H. Neimann, K. Watanabe, S. Hakomori, R. Childs, and T. Feizi, Biochem, Biophys. Res. Commun. 81, 1286 (1978).
[1]
PRINCIPLES OF ANTIGEN--ANTIBODY REACTIONS
37
q ; ¢.~ :t
,1~ ¢¢
T
i
~02
O
e'~
e.. r3
02 t"
0
.6o=.~ 02
v
~02~
•~
02
"O
02
v-~
:¢~02 02
e'~
~
I
l
, ~ 7 .0~2' oo ~ ~33_ I
-.~ eL 02 ~-.
02
e~
~3. I
e~
/
\ o~ ~
'~
02
02 ,.-
02
02 ,.. t..,
v-4 "F
v.-t v i
"7~ ~q7 O~ 02
O'O
02
¢~
e~
02~
38
PRINCIPLES AND METHODS
o
!
t
~
t
~, ~
.~ .~
[l]
[1]
PRINCIPLES OF ANTIGEN--ANTIBODY REACTIONS
39
has been associated with band 3, the major membrane protein of erythrocytes, z°5 and various other I and i activities have been associated with certain gangliosides of human erythrocytes. 206 The determinant reacting with anti-I Ma (group 1) is one of the most extensively characterized, using not only oligosaccharides isolated by degradation of the precursor OG substance (Fig. 10), which reacted with all anti-I and anti-i sera, but also with oligosaccharides of various possible alternative structures prepared by chemical synthesis in two laboratories.el. 147,148 Figures 1 1 and 12 show the results obtained in two laboratories, one by inhibition of precipitation 147and the other by radioimmunoassay. 148 In each instance all oligosaccharides containing the structure D-Gal-/3(1-~4)-D-GIcNAc-/i-(1---~6)-CHz- were equally active on a molar basis in reacting with anti-I Ma. The oligosaccharide D-Gal-/3-(1-~4)-D-GIcNAc was about one-tenth as active. Changes in linkages of the above structure all reduced the inhibitory activity. As long as the fl-(1-~6)-CH2- portion was present, reduction of the sugar of which it was a part did not affect the activity. Results by radioimmunoassay and by inhibition of precipitation were identical. The evidence that the determinant is no larger than that given above is indicated by the finding that D-Ga/-~- ( 1 ~ 4)-D-GIcHAc- ~- ( I ~ 6)-v- galactitoI v-Ga1-~- ( 1 ~ 4)-D-GleNAc-~- (I~6)-D-Gal D'GaI"~'(I~4)'D'GIcNAc'~'(I"~6)D Gal or v-ga/actltol D_Ga1.~. (i ~ 3)_v. GIcNAe.¢_ ( I j 3 ) "
and v-Gal-/3-(1--~4)-D-GlcNAc-fl-(l-~6)-N-acetyl-D-galactosaminitol all have the same inhibiting activity on a molar basis. Figure 12 shows that a second anti-I serum of group 1, Woj, gave identical results. Recently synthesis of D-Gal-fl-(1-~4)-D-GlcNAc-fl-(1---~ 6)-D-GaINAc-aO-(CH~)sCOOCH3 D-GaI-fl-(I---~4)-D-GIcNAc-fl-(1 x~
6) 4)D-Gal-/i-O-(CHs)sCOOCHa
and
/1 o-Gal-/3-(1--*4)-o-GlcNAc-~8-(1
showed each to be equivalent in activity to the others in Fig. 11 whereas D-GIcNAc-/3-(1-~6)-D-GaINAc-a-O-(CH2)aCOOCHa was inactive) °°a thus, defining the site unequivocally as complementary to D-Gal-/i-(1--~4)D-GIcNAc-/3-(1-o6)-O-CH~-. Radioimmunoassay required about 4~ as z05 R. A. Childs, T. Feizi, M. Fukuda, and S.-I. Hakomori, Biochem. J. 173, 333 (1978). z06 R. A. Childs, S.-I. Hakomori, and M. E. Powell, Biochem. J. 173, 245 (1978). ~06a E. A. Kabat, J. Liao, R. U. Lemieux, and M. H. Burzynska, in preparation.
40
PRINCIPLES
[1]
AND METHODS
100
z
80
y
6O '1I-Z w
40
20
0
• 0.5
,& 1
2
~
I
I
I
3
4
5
6
MICROMOLES INHIBITOR ADDED
FIG. 11. Plots of the inhibition of the precipitation observed on mixing 15/zl of anti-I Ma (group 1) (30 p,l ofa 1:2 dilution) with 14.5/~g of glycoprotein OG (20% from 10%) in a total volume of 400/zl and using the following inhibitors (Kabat e t a1.147): Compound n u m b e r Symbol
1 2 3 4 5 6 7
[] • A • @ ~ O
n-Gal-fl- ( 1 ~ D-Gal-/3- (I ~ D-Gal-/3- (I ~ o-Gal-/3- ( 1 ~ D-Gal-/3- ( 1 ~ D-Gal-/3- ( i ~ D-Gal.~ - ( 1 ~
4)-D- GIeNAe-fl- ( 1 ~ 6)-D-Gal 4)-D- GIeNAc-/3- (I ~ 3)-D-GaI 3)- D-GIcNAc-fl- ( I ~ 6)-D-Gal 3)-o-GIcNAc-fl- ( l ~ 3 ) - n - G a l 4)-D-GlcNAe-/3- ( 1 ~ 6)-D-Gal-fl- 1- O- (CH~)aCOOCHs 4)-D-GIeNAc-/3- (I ~ 3 ) - o - G a l - f l - I- O- (CH~)aCOOCH~ 3)- D- GIeNAc-/3- ( 1 ~ 6) - D-Gal-/3-1- O--(CHz)aCOOCI-Ia
8
•
9
× O +
D-Gal-/3- ( 1 ~ 3)-o-GlcNAc-fl- ( 1 ~ 3)-D-Gal-/3-1- O- (CH2)aCOOCHa OG RLI. 1 D-Gal-/3-(I~4)-D-GlcNAc-/3(I~6)-3, 4-dideoxyhex-3-enitols n-Gal-~- ( 1 ~ 4)- n- GIcNAc D-Gal-~- ( 1 ~ 6 ) - D-GlcNAc
much material as did inhibition o f precipitation 147 •148 ; with the lectin s y s t e m the c o m p e t i t i v e binding a s s a y s required about ~o as m u c h inhibitor. 18z Immunochemical of Proteins
Studies on the Antigenic Determinants
B e c a u s e o f the predominant contribution o f c o n f o r m a t i o n to the tertiary structure o f protein antigens and b e c a u s e protein antigens m a y contain several antigenic determinants, the p r e v i o u s l y described approaches
[1]
PRINCIPLES
OF
ANTIGEN-ANTIBODY
(a) Anti-I Ma1:4000 I00
/
o -20
¢ i I
! / f" o
i I0
41
(b) Anli- I W0j 1:500
~'o
/
O-
REACTIONS
i I00
o o "1
-
I
I
1.000 1 10 01ig05accharideaddedInto01)
t= o
o 1
100
o I
1,000
Fro. 12. Inhibition of binding of anti-] sera Ma and Woj to nSl-labeled blood group l-active glycoprotein by synthetic oligosaccharides. Symbols for synthetic oligosaccharides: D'GlcNAc-fl" (l"~6)D Gal
[]
D-GIcNAc-~3- (1/3) -
D'GaI'/3"(I~4)'D'GIcNAc-~"(I"~6)DGal
v
D-GRI-#3- ( 1 ~ 3) - -GleNAe-/3- (1
From T. Feizi
et
•
/3) "
D-Gal-fl-(I~ 3)-D-GIcNAc-13-(I~ 3)-D-Gal
A
D- Gal-/3- (1 ~ 4) -D- GlcNAc-/3- ( 1 ~ 6)-D- Gal
•
o-Gal-fl- ( 1 ~ 3)-D-GlcNAc-/3- ( 1 ~ 6)-D-GaJ.
O
al. 148
to the location of antigenic determinants must be supplemented by basic chemical procedures. Moreover, protein antigens also generally involve cellular immune responses and contain determinants that act on the thymus-derived lymphocytes and produce delayed-type hypersensitivity. A detailed examination of this vast body of material ef' 207 and the technologycf. 2o8 is outside the scope of this volume, and only a brief outline will be given of the principles applied in localization and identification of antigenic determinants on proteins and polypeptides. Two general approaches are employed. One involves fragmentation of the protein by defined methods so as to obtain fractions containing different antigenic determinantsCC6'7'49-sa; these may be coupled to insoluble adsorbents, to separate the heterogeneous populations of antibodies with a view to obtaining antibodies to the individual determinants. The other 207 M. Z. Atassi and A. B. Stavitsky, eds., "Immunology of Proteins and Peptides I . " Plen u m , N e w Y o r k , 1978. 2o8 I. Lefkovits and B. Pernis, "Research Methods in Immunology." Academic Press, N e w Y o r k , 1978.
42
PRINCIPLES AND METHODS
[1]
involves inhibition assays with oligo- and polypeptides of known structure and increasing size to delineate the determinants in the manner already described for oligosaccharide determinants. Many of the oligopeptides become available as by-products from the determination of the. primary structure of the protein. 43"9~Individual amino acids in these peptides may be modified chemically, and the effect on their capacity to inhibit or to precipitate can be assayed. 2°9 Such studies make it possible to decide whether a given residue in the peptide is or is not reacting with the antibody combining site. The ease with which peptides of any desired structure may be synthesized makes it possible to evaluate the relative contributions of any amino acid to a given antigenic determinant by synthesizing a series of peptides in which one amino acid has been replaced by another. 21° With the heterogeneous populations of antibodies to protein antigens usually found in antisera, antibody to a given antigenic determinant may only comprise 10-15% or even less of the total antibody. 2°9-211,el. 7,91 If inhibition assays were to be carried out with the whole antiserum and the intact antigen as precipitant, the maximum degree of inhibition obtainable with the determinant in question would be only 10-15%, and this could create substantial uncertainty. It is for this reason that whenever fragments containing different populations of antigenic determinants are obtainable, the antiserum should be absorbed to remove antibodies to antigenic determinants other than the one that is being studied; in this way the maximum degree of inhibition and the precision obtainable is increased. Even when an antigenic determinant is defined, antibodies to it produced in individual animals may not be equivalent; for example, with the C-terminal antigenic determinant of myoglobin, in some antisera the Cterminal hexa- and heptapeptides were equal in inhibiting on a molar basis, whereas in others the heptapeptide was more active. T M It may often be difficult to evaluate the contribution of conformation as compared with chain length in defining an antigenic determinant. If the crystallographic structure is available, one may arrive at a decision by seeing which residues of the inhibiting peptide would be at the surface of the molecule and able to react or whether they are internal and are increasing inhibiting power by favoring the conformation present in the intact molecule. Thus Crumpton T M found that in myoglobin a peptide molecule of residues 15-33 was much more active than one with residues 15-29, although both were reacting with the same fraction of the anti2og M. Z. Atassi, Immunochemistry 15, 909 (1978). 210 M. J. Crumpton, Biochem. J. 116, 923 (1970). 21~ M. J. Crumpton and J. M. Wilkinson, Biochem. J. 94, 545 (1965).
[1]
PRINCIPLES OF A N T I G E N - A N T I B O D Y
REACTIONS
43
FIG. 13. The conformation of polypeptide chain of hen egg white lysozyme as deduced from X-ray studies at 2/~ resolution. The loop region is indicated by a circle. From Blake et a / . $12
body; from the three-dimensional structure, residues 30-33 were internal and not able to function as contacting residues, and they were inferred to be contributing to the enhanced binding efficiency by stabilizing the conformation in the intact protein. A very good example is the study of the extent to which residues of the loop polypeptide of lysozyme contributes to specificity. The loop peptide comprises residues 64-80 in which Cys-64 and Cys-80 form a disulfide bond, and thus this structure is highly conformational because opening the disulfide bond destroyed immunological activity. From the X-ray crystallographic study (Fig. 13),212 the loop peptide occurs in an exposed position. Figure 14 shows the sequence of the synthetic loop peptide zla and the capacity of synthetic loop peptides containing residues modified 2~ C. C. F. Blake, D. F. Koenig, G. A. Mair, A. C. T. North, D. C. Phillips, and V. R. Sarma, Nature (London) 206, 757 (1965). 213 R. Arnon, E. Maron, M. Sela, and C. B. Anfinsen, Proc. Natl. Acad. Sci. U.S.A. 68, 1450 (1971).
44
PRINCIPLES AND M E T H O D S
~
7
[1]
0
H2N'-~E ~N/)
Synthetic loop
~' i
HOOC82
80 ~
75
Intactloop,. 100
00¢,
/ Pro~.7//
i 0.1
/
l
I
I
I
1.0
10
100
100(}
P
Syntheticloopderivativesadded(pg) FIG. 14. Effects of replacement by Aia of the indicated residuesin the synthetic loop peptideof lysozymeon ability to inhibit inactivation by antiioop antibodiesof bacteriophageloop conjugate. From Arnon e t al. 213 and Teicher et al. =~4
at different positions to react with antiloop antibodies. 214 In those studies, assay of inhibiting power was based on the ability of the synthetic loop peptides to inhibit inactivation by antiloop antibodies of a conjugate to bacteriophage 215"21e of the loop peptide. 21r A synthetic loop peptide was prepared in which Cys-76 was replaced by Ala. When it was coupled through the COOH of Ala-82 to a poly(DL-Ala)-poly(Lys), and used for 214 E. Teicher, E. Maron, and R. Arnon, Immunochemistry 10, 265 (1973). =15 O. M~ikel~t, Immunology 10, 81 (1966). 2~e j. Haimovitch and M. Sela, J. lmmunol. 97, 338 (1966). 21~ j. Haimovitch, E. Hurwitz, N. Novik, and M. Sela, Biochim. Biophys. Acta 207, 115 (1970).
[1]
PRINCIPLES OF ANTIGEN--ANTIBODY REACTIONS
45
immunization, the antibodies produced reacted with native lysozyme. Also, the loop peptide coupled to Sepharose specifically absorbed from antisera to lysozyme the antiloop antibodies, which could then be eluted. The occurrence of cross-reactions with lysozymes of various species may contribute substantially to the precise delineation of the antigenic determinants. Turkey and bobwhite quail lysozymes differ from that of chicken in that Arg replaces Lys at positions 73 and 68, respectively. Turkey lysozyme reacted identically but quail lysozyme was weaker, thus indicating that Arg-68 makes a greater contribution to the interaction with the antibody combining site. The predominant contribution of the folded native structure is amply illustrated by the findings that complete reduction of disulfide bonds and S-carboxymethylation completely alters the immunochemical reactivity of protein antigens, such as ribonuclease, 218 lysozyme, and other antigens. 2t9,221 However, reduction of but two of the four disulfide bonds in ribonuclease had a negligible effect~2zon the ability to precipitate with antibody to native ribonuclease. Preparation of various derivatives effecting substitution on the protein antigen at known positions and evaluating the effects on reactivity with antisera have also been extensively used, especially by Atassi and coworkers, z°7,2°9 These include modifications or substitutions on tyrosines, tryptophans, methionines, arginines, and free amino or carboxyl groups. An unusual approach to mapping surface antigenic determinants if a three-dimensional structure for the protein is available has been introduced by Atassi 2°7'~°9and termed "surface simulation synthesis." In principle, it consists in measuring distances between the a-carbons of the hypothesized contacting residues of the antigenic determinant on the surface of the molecule and synthesizing a linear peptide with amino acid residues whose side chains would correspond in position to those of the model. Diglycine was used to approximate the - - S - - S bond distances for those determinants with the hypothesized contacting residues from two noncontiguous regions brought into proximity by - - S - - S bonds. Several alternative peptides were also synthesized whose a-carbon distances did not match those of the three-dimensional structure. The proportion of the total lysozyme-antilysozyme precipitate that the peptide was capable of 21s R. K. Brown, R. Delaney, L. Levine, and H. Van Vunakis, J. Biol. Chem. 234, 2043 (1959). 219 j. Gerwing and K. Thompson, Biochemistry 7, 538 (1968). 2zo j. Young and C. Y. Leung, Biochemistry 9, 2755 (1970). ~1 R. Arnon, in "The Antigens" (M. Sela, ed.), Vol. 1, p. 88. Academic Press, New York, 1973. 2~2 H. Neumann, I. Z. Steinberg, J. B. Brown, R. F. Goldberger, and M. Sela, Eur. J. Biochem. 3, 171 (1967).
46
PRINCIPLES
AND METHODS
[1]
inhibiting and the molar ratio of peptide to lysozyme at 50% inhibition were assayed. The most active peptide of minimal size was considered to mimic the surface structure of the antigenic determinant. The lysozyme molecule was determined to have three antigenic sites with residues coming from widely separated portions of the polypeptide chain; the residues proposed as contacting and those synthesized to produce a linear sequence considered as best simulating the active site are seen in Fig. 15. In some instances the peptide synthesized in the "reverse direction"--for example, using the C-terminal amino acid of the hypothesized determinant as the amino terminus--was used as a control. In some instances, the sequence was considered to have directionality whereas in others it did not. One of the uncertainties of the surface simulation method is that all the synthetic peptides proposed are rather short and each contains three highly charged residues. The inferences are based on a limited number of model polypeptides, mostly directional peptides; it would be desirable to make substitutions in the determinants in which the charged groups remain in position, but other presumed contacting residues are changed as well as those in which one charged amino acid is replaced by another. This could provide insight into the specificity of the method. Another question that arises is whether the simulation peptide is indeed reacting with the active site to prevent lysozyme from precipitating; it is conceivable that it might be reacting at another site and causing a conformational change at the site or that it might be reacting around the site to prevent lysozyme from entering rather than interacting by complementarity in the site itself. Changes in the spacing and substitution of other residues in the simulation peptides could increase the weight of the evidence favoring specific vs nonspecific interaction. The antigenic determinants proposed by Atassi 2°7,2°9do not include the loop peptide 64-82 (Fig. 13) to which Maron e t a l . 223 found 8-10 mg of antibody per 100 mg of goat antilysozyme ct.z23a; these antibodies were more restricted in heterogeneity than the total antilysozyme. The problem of antibody heterogeneity may well be complicating the precise delineation of antigenic determinants of proteins. Thus far all efforts to locate the antigenic determinants have been made with heterogeneous antibody populations. The subject must be reexplored with antiprotein antibodies obtained from hybridomas, which would provide monoclonal antibody populations. In this way one cannot have uniform reactivity with only the antigen itself, but must have it also with the vari223 E. Maron, C. Shiozawa, R. Arnon, and M. Sela, Biochemistry 10, 763 (1971). 223a I. M. Ibrihimi, J. Eder, E. M. Prager, A. C. Wilson, and R. Arnon, Mol. lrnmunol. 17, 37 (1980).
[I]
PRINCIPLES OF ANTIGEN--ANTIBODY
REACTIONS
47
Site 1 Constituent
residues:
125
5
Arg
Arg
i
(~C-to-crC, in n m )
0.93
~
_L~
'
O. 58
Distances:
(aC-to-c~C, in n m )
I. 05
'
3.01
Arg--Gly
Gly
13 Lys i
>',= 0.45-DH
'
i*
site:
i
~'~
'
i
The synthetic
14 Arg
i
~
Distances:
7 Glu
'
]
Arg--Gly--Glu--Gly
Gly--Arg
i
Lys
I
Site 2 62 Constituent
residues:
Distances:
(c~C-to-c~C, in rim)
97 Lys
Trp
0.71
i
96 Lys
site:
Distances: (aC-to-c~C, in nm)
89
87
Thr
Asp
=I: 0 . 4 1 ~ - 0 . 5 6 - * / ~ - 0 . 5 1 - ~ 0 . 5 4 I I r I
i.i The synthetic
93 Asn
i
2.73
Phe - -
Gly
Lys - -
,i i
Lys--
i,
Asn - -
Thr--
Asp
2.16
-[
Site 3 Constituent r e s i d u e s : Distances:
(~C-to-~C, in nm)
116 Lys
i
113 Asn - 0.5
~-I~0.4 i
114 Arg
site:
Distances: (c~C-to-c~C, in nrn)
Lys
[I
0.8
~-',~ I
l, The synthetic
34 Phe i
~',~ O. 4--~ I
2.1
Ash
Arg
33 Lys
i
~i i
Gly--Phe--Lys 1.8 p
FiG. 15. Three antigenic sites oflysozyme. The diagram shows the spatially adjacent residues constituting each antigenic site and their numerical positions in the primary structure. The distances separating the consecutive residues and the overall dimension of each site (in its extended form) are given, together with the dimension of each "surface-simulation" synthetic site. The latter assumes an ideal C~-to-C ~ distance of 0.362 nm. From M. Z. Atassi and C.-L. Lee, Biochern. J. 171,429 (1978).
ous constituent polypeptides, simulated peptide stretches, chemically modified antigens, etc. If a sufficient representation of the repertoire of antibody combining sites to the various determinants is obtained, it will be possible not only to localize precisely the individual sequences making up the determinants, but also to ascertain whether individual clones making antibody to a given determinant do or do not see it in different aspects and make antibodies directed against different residues. Various clones may see different residues as immunodominant in the same determinant or may be specific for different lengths of a given determinant, as already
48
PRINCIPLES AND METHODS
[1]
noted for the antisera reacting differently with the C-terminal hexa- and heptapeptide of myoglobin, zH Such studies might help to clarify the recent findings of Hurrell et al.,224 who noted that myoglobins of different species did not cross react even when complete sequence homology existed for the residues comprising a proposed antigenic determinant and suggested the hypothesized determinants may be larger or may be influenced conformationally in their reactivity by other segments of the molecule. Uncertainties that complicate studies on antigenic determinants of proteins are seen in the findings of Benjamini et al. 2zs and of Schechter et al.226 Benjamini et al.225 noted that attachment of a hydrophobic octanoyl group to the N terminus of a tripeptide making up part of a pentapeptide antigenic determinant of tobacco mosaic virus protein made it more active. Schechter et a/. 226 obtained antibodies by immunization to (o-Ala)2Gly-e-aminocaproic acid coupled to protein and found that with some antisera o-Ala4 and o-Ala~ were better than the determinant group of the antigen itself. Heteroclitic antibodies, 227 those that react better with a cross-reacting antigen than with the homologous antigen used for immuninization, are puzzling with respect also to mapping combining sites of protein antigens. Even with respect to carbohydrate determinants, problems in the identification of antigenic determinants remain. Zopfet a/. 22s have demonstrated that antibodies to a tetrasaccharide determinant may be directed toward one side of the molecule whereas others can be of entirely different specificity directed toward the opposite side. A sugar residue substituted on one side can block reactivity with one specificity without influencing reactivity to the antibodies of the other specificity. A n t i g e n - A n t i b o d y Reactions in Gels Since the classical studies of Oudin of"229 and Ouchterlonyof.230 on antigen and antibody interactions in gels and by Grabar and Williamsct. 231 on zz4 j. G. R. Hurreil, J. A. Smith, P. E. Todd, and S. J. Leach, Immunochemistry 14, 283 (1977). 225 E. Benjamini, D. Michaeli, and J. D. Young, Curt. Top. Microbiol. Immunol. ~ , 85 (1972). 2~6 B. Schechter, I. Schechter, and M. Sela, J. Biol. Chem. 245, 1438 (1970). ~27 T. Imanishi and O. M~ikel~i, J. Exp. Med. 1411, 1498 (1974). 228 D. A. Zopf, C.-M. Tsai, and V. Ginsburg, Arch. Biochem. Biophys. 185, 61 (1978). 229 j. Oudin, Methods Med. Res. 5, 335 (1952). ~a00. Ouchterlony, Prog. Allergy, 5, 1 (1958); 6, 3 (1962). 231 p. Grabar and P. Burtin, *'Immunoelectrophoretic Analysis." Elsevier, Amsterdam, 1964.
[2]
PROTEINS AND POLYPEPTIDES AS ANTIGENS
49
immunoelectrophoresis, these have become indispensable tools of the immunologist and immunochemist and have been applied in almost all fields o f biological science. They are invaluable for analyzing complex mixtures o f antigens and antibodies, for establishing that monospecific antisera contain only antibodies to the desired antigen, for detecting antibodies formed to impurities in the antigen used, and for qualitative detection and quantitative estimation of antigens and antibodies. Details are given by Oudin (this volume [9]). Acknowledgments Work of the laboratories is supported by grants from the National Science Foundation PCM 76-81029and a grant to the Cancer Center CA 13696from the National Institutes of Health.
[2] P r o t e i n s a n d P o l y p e p t i d e s
as Antigens
By PAUL H. MAURER and HUGH J. CALLAHAN During the past two decades there has been a tremendous increase in the realization o f the utility of antibodies directed against enzymes as tools in biochemical studies, l'z Antibodies against enzymes can be used (a) to detect and assay quantitatively the concentration of enzymes; (b) to concentrate and purify enzymes from dilute solutions and mixtures; (c) to study the active catalytic sites, multimolecular forms, and conformational structures of enzymes; (d) to localize enzymes in sectioned cells; (e) to study the appearance and modification of enzymes in the course of embryonic and phylogenetic development. Concomitantly in immunology there has been increasing knowledge concerning the many factors that can influence the multifaceted and complex sequence of events of the immune response beginning with the introduction of an antigen into a host (immunogen) to the formation of humoral antibody, a The goal of this chapter is not only to present information about techniques that have become available for producing antibody, but in so doing to make the invesi (B. Cinader, ed.), Ann. N. Y. Acad. Sci. 103, 493-1154 (1963). 2 "Antibodies to Biologically Active Molecules" (B. Cinader, ed.). Pergamon, Oxford, 1967. a See "Essential Concepts," discussed in: L. E. Hood, I. L. Weissman, and W. B. Wood, "Immunology," pp. 1-74. Benjamin-Cummings, Menlo Park, California, 1978.
METHODS IN ENZYMOIX~Y, VOL. 70
Copyright(~)igBOby Academic Press, Inc. All fights of reproduction in any form reserved. ISBN 0-12-1gl970- I
[2]
PROTEINS AND POLYPEPTIDES AS ANTIGENS
49
immunoelectrophoresis, these have become indispensable tools of the immunologist and immunochemist and have been applied in almost all fields o f biological science. They are invaluable for analyzing complex mixtures o f antigens and antibodies, for establishing that monospecific antisera contain only antibodies to the desired antigen, for detecting antibodies formed to impurities in the antigen used, and for qualitative detection and quantitative estimation of antigens and antibodies. Details are given by Oudin (this volume [9]). Acknowledgments Work of the laboratories is supported by grants from the National Science Foundation PCM 76-81029and a grant to the Cancer Center CA 13696from the National Institutes of Health.
[2] P r o t e i n s a n d P o l y p e p t i d e s
as Antigens
By PAUL H. MAURER and HUGH J. CALLAHAN During the past two decades there has been a tremendous increase in the realization o f the utility of antibodies directed against enzymes as tools in biochemical studies, l'z Antibodies against enzymes can be used (a) to detect and assay quantitatively the concentration of enzymes; (b) to concentrate and purify enzymes from dilute solutions and mixtures; (c) to study the active catalytic sites, multimolecular forms, and conformational structures of enzymes; (d) to localize enzymes in sectioned cells; (e) to study the appearance and modification of enzymes in the course of embryonic and phylogenetic development. Concomitantly in immunology there has been increasing knowledge concerning the many factors that can influence the multifaceted and complex sequence of events of the immune response beginning with the introduction of an antigen into a host (immunogen) to the formation of humoral antibody, a The goal of this chapter is not only to present information about techniques that have become available for producing antibody, but in so doing to make the invesi (B. Cinader, ed.), Ann. N. Y. Acad. Sci. 103, 493-1154 (1963). 2 "Antibodies to Biologically Active Molecules" (B. Cinader, ed.). Pergamon, Oxford, 1967. a See "Essential Concepts," discussed in: L. E. Hood, I. L. Weissman, and W. B. Wood, "Immunology," pp. 1-74. Benjamin-Cummings, Menlo Park, California, 1978.
METHODS IN ENZYMOIX~Y, VOL. 70
Copyright(~)igBOby Academic Press, Inc. All fights of reproduction in any form reserved. ISBN 0-12-1gl970- I
50
PRINCIPLES AND METHODS
[2]
tigator aware not only of factors that might enhance antibody formation, but also of factors that might be operative in suppression of the immune response. 4 As a rule biochemists and enzymologists work with limited amounts of purified materials. In addition to the usual biochemical methods, it should be realized that immunochemical techniques exist that allow one (a) to determine whether an enzyme preparation is indeed "pure" and (b) to use antibody against the enzyme to further purify the enzyme preparation via immunoadsorbent techniques) Knowledge of some immunological generalizations may help the investigator before he proceeds with the preparation of antibody. The ability of an animal to elicit an immune response depends upon complex interactions between the specific immunogen being presented to the host, the properties of the antigen, and the physiological state of the specific animal of choice. It is known that not all proteins or polypeptides can be immunogenic, i.e., can elicit an immune response under a standard set of conditions, and that the method by which the antigen is presented can influence the response. There also can be major differences in responses to the same macromolecule from species to species and among strains (or animals) within a specific species. 6 Nevertheless, it is known that by employing the correct "carrier" and conjugation procedure for the "nonimmunogen" there are ways of eliciting a response to almost any macromolecule. The molecular weight and the complexity in structures of the macromolecule influence the nature of the immune response. In general, the greater the molecular weight and the more complex the protein or polypeptide structure, the greater the response that can be expected. From studies with synthetic polymers of amino acids, it has been learned that ordinarily one does not elicit significant responses against homopolymers of amino acids, and that decreased responses are obtained against high molecular weight polymers containing a-D- or y-D-amino acidsY ,8 Determination of the antigenic structures of proteins has posed a chemical challenge of enormous proportions for years. ° Many investigators, therefore, have employed synthetic polymers of amino acids in im4 D. H. Katz, "Lymphocyte Differentiation, Recognition, and Regulation." Academic Press, New York, 1977. I. Parikh and P. Cuatrecasas, in "'Immunochemistry of Proteins" (M. Z. Atassi, ed.), Vol. 2 pp. 1-44. Plenum, New York, 1977. 8 Reviewed in: "Immunogenicity--Physico-Chemicai and Biological Aspects" (F. Borek, ed.). North Holland/American Elsevier, Amsterdam, 1972. 7 p. H. Maurer, Prog. Allergy 8, 1 (1964). s M. Sela, Science 166, 1365 (1969). 9 "'Immunochemistry of Proteins" (M. Z. Atassi, ed.), Voi. 1. Plenum, New York, 1977.
[2]
PROTEINS AND POLYPEPTIDES AS ANTIGENS
51
munochemical studies in the hope that information derived from these systems might be useful in the understanding of the immunochemistry of proteins. Although many data on these synthetic polymer systems have been accumulated in several laboratories, and polymers have contributed to elucidating aspects of the immune mechanism, the information gained from amino acid polymers has not always been helpful in understanding completely the immunochemistry or basis for the immunogenicity of proteins. Knowledge of the antigenic sites of protein antigens may help elucidate further not only the mechanisms of the immune response, but many immunological disorders at the molecular level. Although the last decade has witnessed a great deal of activity investigating the immunochemistry of protein antigens, 5,a so far the antigenic structure of sperm whale myoglobin 1° and of lysozyme have been completed. 1 Factors Influencing the I m m u n e R e s p o n s e
Animal Species The more common animals used for immunization are rabbits, goats, sheep, chickens, horses, guinea pigs, and mice. a2'~3The eventual goals of having antibody against a specific protein or polypeptide may determine the specific species for immunization. Important considerations in choosing a species are the source and availability of the immunogen. As might be expected the larger animals require more antigen for the production of antibody, but when responding can yield more serum than others. Ordinarily one cannot obtain large amounts of serum from repeated bleedings of mice. However, there are techniques available for producing large amounts of ascites fluid rich in antibodies, a4-~9 In addition, recent developments in the technology of the production of "hybridomas" of mye10 M. Z. Atassi, lmmunochemistry 12, 423 (1975). 11 M. Z. Atassi, lmmunochemistry 15, 909 (1978). 12 "Methods in Immunology and Immunochemistry" (C. A. Williams and M. W. Chase, eds.), Vol. 1. Academic Press, New York, 1967. 13 "Handbook of Experimental Immunology," Vol. 3, Application of Immunological Methods (D. M. Weir, ed.), 3rd ed. Blackweil, Oxford, 1978. 14 j. S. Garvey, N. E. Cremer, and D. H. Sussdorf, "'Methods in Immunology," 3rd ed. Benjamin-Cummings, Reading, Massachusetts, 1977. 15 j. Munoz, Pro¢. Soc. Exp. Biol. Med. 95, 757 (1957). le E. C. Herrmann, Jr. and C. Engle, Proc. Soc. Exp. Biol. Med. 98, 257 (1958). lr E. S. Takasingh, L. Spence, and W. G. Downs, Am. J. Trop. Med. Hyg. 15, 219 (1966). 18 A. C. Sartorelli, D. S. Fischer, and W. C. Downs, J. lmmunol. 96, 676 (1966). 19 A. S. Tung, S. Ju, S. Sato, and A. Nisonoff, J. lmmunol. 116, 676 (1976).
52
PRINCIPLES AND METHODS
[2]
loma cells fused with normal antibody-producing mouse spleen cells has afforded the production of larger amounts of antibody both in vivo and in vitro 2°-~4 (see below). If one wants to increase the likelihood of obtaining a good response against most of the antigenic determinants in a protein, a species as phylogenetically removed as possible from the source (species) of the immunizing material should be injected. However, if a goal is to obtain antisera directed only against a few dissimilar structures or peptide sequences in an immunogen, then the same species should be injected. For instance, antibody against rabbit T-globulin or mouse T-globulin can be produced in rabbits and mice, respectively. However, depending on the genetic background of the host, the antibody produced can be directed against the minor "allotypic" structures. 25 Generally, rabbits, sheep, goats, and horses produce much more antibody per milliliter of serum than do guinea pigs and mice. In addition, most of the antibodies obtained following hyperimmunization of rabbits, sheep, and goats precipitate with the homologous antigens, whereas not all of those produced in guinea pigs and mice precipitate easily, but tend to form soluble antigen-antibody complexes. Genetic Factors
That many "genetic" factors govern immune responses to simple and complex synthetic polymers as well as to proteins such as lysozyme has been recently reviewed. 26 Immune responses of inbred strains of guinea pigs, mice, and rats to many immunogens are controlled by immune response genes present in the major histocompatability complex of the respective species. This has accounted for the unique finding of responders and nonresponders to polypeptides and proteinsY For instance, strain 2, but not strain 13, guinea pigs respond to the random copolymers s0 G. Kohler and C. Milstein, Nature (London) 256, 495 (1975). 2~ G. Kohler and C. Milstein, Fur. J. lmmunol. 6, 511 (1976). z2 G. Galfre, S. C. Howe, C. Milstein, G. W. Butcher, and J. C. Howard, Nature (London) 266, 550, (1977). 23 "'Lymphocyte Hybridomas" (F. Meichers, M. Potter, and N. L. Warner, eds.), in Curr. Top. Microbiol. Immunol. 81 (1978). L. A. Herzenberg, L. A. Herzenberg, and C. Miistein, in "Handbook of Experimental Immunology" (D. M. Weir, ed.), 3rd ed., Ch. 25. Vol. 3, Blaekweli, Oxford, 1978. 25 L. A. Herzenberg and L. A. Herzenberg, in "'Handbook of Experimental Immunology'" (D. M. Weir, ed.), 3rd ed., Vol. 3 Ch. 12. Biackweil, Oxford, 1978. 26 "Genetic Control of Immune Responsiveness: Relationship to Disease Susceptibility" (H. O. McDevitt and M. Landy, eds.). Academic Press, New York, 1973. 37 "Immunogenetics and Immunodeficiency" (B. Benacerraf, ed.), Univ. Park Press, Baltimore, Maryland, 1975.
[2]
PROTEINS AND POLYPEPTIDES AS ANTIGENS
53
(Glue°Al#°) n and (Glue°Lys4°)n and the opposite pattern is noted with the polymer (GluS°Tyr5°)~. Similar situations exist with inbred mice, i.e., mice of H-2 haplotypes a, b, d, f, k, s respond to the random polymer (Glua°Al#°)n, but mice of haplotypes p and q do not. In addition a polymer or protein that is immunogenic in one species, need not be immunogenic in another species; i.e., mice do not respond to (Glu6°Lys4°), although guinea pigs and rabbits do. 2s Because of the multigenic control of most immune responses, it is recommended that outbred animals be immunized first. Even with this precaution, it is likely that not all animals will respond similarly. However, the coupling of a poor immunogen either covalently or via electrostatic interactions with an immunogenic protein "carrier" can convert the nonimmunogen to a conjugate that is immunogenic in nonresponders and responders. Responses are then obtained to the complete conjugate, i.e., the carrier as well as the "haptenic" determinants. Some of the limitations to the above procedure are that the conjugation technique may alter the antigenic structure of the determinant and that "antigenic competition" between carrier and coupled macromolecule may occur if the carrier is very immunogenic.
Properties of the Antigen Any consideration of the immunogenicity of a protein must take into account how its physiochemical properties can dictate the outcome of the immune response. We will consider only some of the intrinsic properties of antigens that are important in this respect; other properties will be considered in the section Methods of Immunization. The state of aggregation of a protein has long been recognized as a factor involved in its immunogenic potential. Using bovine y-globulin, Dresser z9 found that mice normally responsive to this protein were nonresponsive to preparations freed of particulate or aggregated material by centrifugation or column chromatography. Others 3° have demonstrated similar results in rabbits with human y-globulin as the immunogen. In both cases the failure to mount an immune response seemed to be due to the induction of a tolerant (paralyzed) state by the deaggregated preparations. Denaturation of protein antigens has been extensively studied over the past 70 years and in general has been shown to decrease immunogenicity relative to that o f the native form. In addition the antigenic specificity of P. Pinchuck and P. H. Maurer, J. Exp. Med. 122, 665 (1965). 2a D. W. Dresser, Immunology 5, 378 (1962). 30 C. Biro and G. Garcia, Immunology 8, 411 (1965).
54.
PRINCIPLES AND METHODS
[2]
many native proteins is lost and new specificities are created. Several reviews 6"31'32 o f this subject are available, and thus an extensive survey o f the literature will not be attempted here, instead a few specific examples will be given that typify the results obtained. Early studies demonstrated that denaturation o f ovalbumin (OA) achieved by any o f several means, e.g., heat, ultraviolet irradiation, sonication, alcohol, urea, or chemical modification resulted in a loss o f reactivity with anti-native OA antibodies. In 1951 Maurer and Heidelberger a3 showed that upon deamination two fractions could be obtained from ovalb u m i n - - o n e lightly deaminated (27-36%), which appeared not to be denatured, and another highly deaminated ( 4 0 - 8 0 ~ ) , which was denatured and insoluble at its isoelectric point. The lightly deaminated preparation reacted completely with anti-native OA serum, and conversely antisera to this preparation reacted completely with native OA. The highly deaminated preparation, however, reacted only weakly with either antiserum. More recently, Jacobsen e t a l . 34 employing a series of chemically modified human serum albumins, have shown that a distinct relationship exists between the antigenicity of these proteins and the Stokes radius values as determined by gel filtration experiments. Using the Stokes radius as an indicator o f unfolding, they have shown that an increase of approximately 0.5 nm results in a 50% loss in activity and an increase of 1.7 nm abolishes activity. Analogous results have been obtained with reduced-carboxymethylated bovine serum albumin, 3~ performic acid-oxidized ribonuclease, 36 reduced-carboxymethylated lysozyme, 37 and several other proteins. The studies described above, as well as numerous others, have led to the conclusion that protein antigens may contain two general classes o f determinant structures, namely, sequential and conformational. Sequential determinants would be those occurring in a linear conformation, as in the unfolded form o f a protein, and conformational determinants would be those that are recognized by their homologous antibodies only when they o c c u r in a particular conformation. The latter class would include determinants formed from amino acids that are located at distant points in the 31 E. A. Kabat, in "Kabat and Mayer's Experimental Immunochemistry,'" 2nd ed. Thomas, Springfield, Illinois, 1961. 32 M. Reichlin, Adv. l m m u n o l , p. 29. Academic Press, New York, 1975. 38 p. H. Maurer and M. Heidelberger, J. A m . Chem. Soc. 73, 2076 (1951). 34 C. Jacobsen, L. Funding, N. P. H. Moller, and J. Steengard, Fur. J. B i o c h e m 30, 392 (1972). 35 E. J. Goetzl and J. H. Peters, J. l m m u n o l . 108, 785 (1972). 86 R. K. Brown, J. Biol. C h e m 237, 1162 (1962). 37j. D. Young and C. Y. Leung, Biochemistry 9, 2755 (1970).
[2]
PROTEINS AND POLYPEPTIDES AS ANTIGENS
55
peptide chain but come into proximity with each other when a particular conformation is achieved. An elegant demonstration of conformational effects has been reported by Arnon and Sela. 38 They were able to isolate a 20-amino acid peptide (Cys°4-Leua3) from peptic digests of egg white lysozyme, which was joined at residues 64 and 80 by a disulfide bridge, thus forming a " l o o p " . This " l o o p " peptide was conjugated to a synthetic branched polypeptide and then used to immunize rabbits. Antibodies immunospecifically isolated from these sera, or from sera of animals immunized with lysozyme, react with the " l o o p " peptide, but only poorly or not at all with the open chain form produced by reduction and carboxymethylation, indicating the requirement for a particular three-dimensional structure in the antigen to bind with antibody. It also seems likely that proteins displaying quaternary structure have unique determinants that can be considered "conformational." Thus Reichlin32 has shown that, in complement fixation tests, antisera against methemoglobin reacted better with oxyhemoglobin than with deoxyhemoglobin. This is most likely due to the known difference in quaternary structure between the two proteins. In addition, the isolated a and /3 chains, which were inactive with the antiserum, could be rendered active by recombination with the appropriate chains from a different species. With myoglobin similar results were obtained, i.e., antisera to myoglobin detected differences between the heme-containing protein and the heinefree protein. It appears at present that most determinants in globular proteins are of the conformational type whereas one finds both types in structural proteins such as collagen, a9 There are numerous properties, other than those mentioned, that are associated with antigens and would affect their ability to induce antibodies or our ability to detect these antibodies. We will consider only four here, since they are more applicable to proteins and polypeptides: accessibility of determinants, complexity of amino acid composition, molecular weight, and effect of ions. It is well documented that groups that function as antigenic determinants are those that are exposed to solvent. This has been elegantly demonstrated by Sela's group s with branched-chain synthetic polypeptides. They found that when tyrosine and glutamic acid residues were attached in a linear fashion to poly(oL-alanine), and the alanine is then linked to poly(L-lysine) as a branch, antibodies were formed against the terminal glutamic acid and tyrosine positions (Fig. 1). However, if alanine was placed in the terminal position and linked to the poly38 R. Arnon and M. Sela, Proc. Natl. Acad. Sci. U . S . A . 62, 163 (1969). 39 M. Crumpton, in "Defence and Recognition" (R. R. Porter, ed.), MTP Int. Rev. Sci. Set. l, Vot. 10. Butterworth, London, 1973.
56
PRINCIPLES A N D M E T H O D S
°ll
[2]
l
Y.
FIG. 1. A multichain copolymer in which L-tyrosine and L-glutamic acid residues are attached to multi-poly(DL-alanyl)-poly(L-lysine). Left: Tyrosine and glutamic acid located in
terminal positions. Right: Tyrosine and glutamic acid positioned internally. Horizontal lines, poly(L-lysine); hatched area, poly(DL-alanine); 0 , L-tyrosine; ©, L-glutamic acid. From M. Sela. s
lysine through the glutamic acid or tyrosine, then most of the antibodies were alanine specific. These data strongly argue that, in order to be antigenic, determinants must be exposed to the environment. This concept is supported by Atassi's studies ~° with myoglobin. Atassi was able to identify five antigenic regions in this protein, and all were located in solventaccessible regions, in this case on the surface of the molecule. It is generally accepted that a relationship exists between the structural complexity of a compound, i.e., the variety of its components, and its ability to induce an immune response. For example, homopolymers of amino acids by themselves are very poor antigens, 4°-42 however, when used in a complex (e.g., with phosphorylated serum albumin), they induce normal levels of antibodies .43 The same effect can be accomplished by the introduction of a second or third different amino acid. 44 Analogous situations exist in some naturally occurring macromolecules. Thus, the low level of antibodies induced by gelatin could be greatly elevated by the introduction of tyrosyl residues. 45"46At the present time, we cannot explain 40 S. B-Efraim, S. Fuchs, and M. Sela, Immunology 12, 573 (1967). 41 p. H. Maurer, Proc. Soc. Exp. Biol. Med. 96, 394 (1957). 42 D. Subrahmanyam and P. H. Maurer, Fed. Proc. 18, 600 (1959). 43 H. Van Vunakis, J. Kaplan, H. Lehrer, and L. Levine, Immunochemistry 3, 393 (1966). 44 p. H. Maurer, Ann. N. Y. Acad. Sci. 103, 549 (1963). 45 M. Sela and R. Arnon, Biochem. J. 75, 91 (1960). 46 M. Sela, B. Schechter, I. Schechter, and F. Borek, Cold Spring Harbor Syrup. Quant. Biol. 32, 537 (1967).
[2]
PROTEINS AND POLYPEPTIDES AS ANTIGENS
57
this phenomenon, although it has been suggested that cooperation between T and B cells in the immune response requires that different specificities exist within the antigen. 47 In considering the relationship of an antigen's molecular size to its capacity to stimulate an immune response, one must examine both humoral and cellular immunity, since their requirements are somewhat different. In general there is a direct relationship between molecular weight and the ability to induce an immune response in high molecular weight compounds. Thus, many polymers, e.g., flagellin, 4s dextran, and pneumococcal polysaccharide, 4a demonstrate increased humoral and cellular responses with increased size, but this is by no means absolute, since several synthetic polypeptide antigens of the same overall composition but widely different molecular weights have induced the same amount of antibody. 5° Attempts to determine the minimum size necessary to evoke a response have been much more definitive. Schlossman e t al. 51 showed, by using a homologous series of a-DNP-oligo(L-lysine) compounds ranging in size from the tetramer to the nonamer, that a chain length of seven units was the smallest size that could induce humoral and cellular immunity in guinea pigs. Smaller oligomers were ineffective. These results were confirmed by Stupp e t al. 52 using ¢-DNP-oligo(L-lysine) compounds, but it was also discovered that the incorporation of oligopeptides into Freund's adjuvant made significant differences in response patterns. Thus, mono-¢DNP-oligo(L-lysine) compounds, containing as few as two lysine residues, when emulsified with mycobacteria in complete Freund's adjuvant, were capable of stimulating anti-DNP antibody production but not delayed hypersensitivity reactions. No antibody was produced when incomplete adjuvant or saline was used. As an explanation it was suggested that the mycobacteria in the adjuvant formed complexes with the positively charged pepfides and thus acted as a carrier. Although it has been known for many years that high concentrations of many salts can inhibit antigen-antibody interaction or dissociate ira-
47 j. Goodman, in " T h e Antigens" (M. Sela, ed.), Voi. 3, pp. 127-183. Academic Press, New York, 1975. 4s M. J. Becker, H. Levin, and M. Sela, Eur. J. lmmunol. 3, 131 (1973). 49 K. Jann and O. Westphal, in "The Antigens" (M. Sela, ed.), Vol. 3, pp. 1-110. Academic Press, New York, 1975. 5o T. J. Gill, III, H. W. Kunz, and D. S. Papermaster, J. Biol. Chem. 242, 3308 (1967). 51 S. F. Schlossman, S. Ben-Efraim, A. Yaron, and H. A. Sober, J. Exp. Med. 123, 1083
(1966). Y. Stupp, W. E. Paul, and B. Benacerraf, Immunology 21, 583 (1971).
58
PRINCIPLES
AND
METHODS
[2]
mune complexes, it was only recently discovered that in some systems physiological concentrations of specific ions are necessary to affect interaction. Using a synthetic polypeptide composed of 60% glutamic acid, 30% alanine, and 10% tyrosine (GAT), Maurer e t a l . 53 found that some animals (particularly sheep) respond to immunization with the production of two distinct populations of antibodies--one that reacts with the antigen only if divalent cations are present, and another having no such requirement. It was shown with several sheep and rabbit antisera that addition of a chelating agent (EDTA) with the antigen prevented precipitation of 10-90% of the antibody. These experiments led to the isolation of the antibodies and subsequently to a physicochemical explanation of the cation's role. Liberti e t al. ~ showed that cations, especially calcium, neither affect the antibody itself nor enhance precipitation of preformed antigenantibody complexes. They did show that calcium affects the antigen by decreasing intrinsic viscosity, increasing sw,20 and changing the optical rotatory dispersion pattern. Bivalent cations probably induce conformational changes in the antigens, perhaps by bridging the carboxyl groups of glutamic acids, leading to the creation of " n e w " antigenic determinants. In addition to synthetic polypeptides, this cation effect has been found with polysaccharide and protein antigens. Approximately 10-20% of rabbit antibodies raised against two pneumococcal polysaccharides (type III and type VIII) had a calcium requirement for interaction with the homologous antigen. 55 Not surprisingly, both polysaccharides contain glucuronic acid residues in their determinant structures. Favre and Vollotton 56 found 7 of 14 rabbits immunized with angiotensin II had antibodies that bound the antigen maximally in the presence of calcium. More recently ~7calcium requiring anti-human serum albumin antibodies from both rabbits and sheep have been isolated and characterized. Methods of Immunization Very detailed and helpful techniques dealing with the preparation of immunogens for immunization and the use of various vehicles and methods for immunization are presented in ~'Methods in Immunology and Immunochemistry, ''12 Volume 1, and in the "Handbook of Experimental Immunology."13
53 p. H. Maurer, L. G. Clark, and P. A. Liberti, J. lmmunol. 105, 567 (1970). P. A. Liberti, H. J. Callahan, and P. H. Maurer, Adv. Exp. Med. Biol. 48, 161 (1974). 55 H. J. Callahan and P. H. Maurer, lmmunol. Commun. 4, 537 0975). 66 L. Favre and M. B. Vallotton, Immunochemistry 10, 43 (1973). 5~ M. E. Frankel, H. J. Callahan, and P. A. Liberti, Fed. Proc. 36, 1286 (1977).
[2]
PROTEINS AND POLYPEPTIDES AS ANTIGENS
59
In dealing with limited amounts of valuable protein or polypeptide one should use techniques that might enhance the immune response. Although responses can be obtained against either solutions of soluble antigens or suspensions of particulate antigens, better responses are elicited with the proper use of "adjuvants." In principle, the adjuvants allow use of much less immunogen for some of the following reasons. Not all of an immunogen administered to a host persists long enough to become an effective stimulator for antibody formation. Adjuvants increase the persistence of antigen in the host and can protect the antigen from degradation by the usual proteolytic enzymes. This can allow more antibody-forming cells to be exposed to the limited amount of antigen. Particulate materials are more immunogenic, and therefore it is advisable, if possible, to aggregate the protein artificially providing the procedure does not change the conformation or biological activity. Aggregates of human y-globulin and bovine serum albumin are immunogenic whereas the "monomeric" form of these proteins can be tolerogenic, sa Macromolecules that are highly charged, and even those that are nonimmunogenic, can be made to react with such carriers as methylated bovine serum albumin or phosphorylated bovine serum albumin. The charge interaction leads to an insoluable aggregate, which can be immunogenic. This technique has been successful for producing antibody against charged macromolecules, such as DNA, 59 polynucleotides,6° polyglutamic acid. 42 However, one has to be aware that the reaction might lead to changes in the structure of the immunogen or to masking or creation of new immunogenic determinants. Processing of an antigen by macrophages is an important aspect of the immune response, and therefore any procedure that enhances the uptake by macrophages before presentation to the lymphocytes augments the response. Adjuvants allow concentration of the antigen onto a particulate carrier so that the amount of antigen administered per unit volume is increased, the immunogen can be localized in specific areas for long periods of time, and local destruction and elimination of antigen is retarded. The commonly recognized adjuvants are remarkable for their diversities. Soluble immunogens can be adsorbed onto the following kinds of inorganic suspensions: alumina cream, aluminum phosphate, and aluminum sulfate. Adsorption onto organic carriers such as blood, charcoal, calcium alginate, or polyacrylamide gels has also enhanced responses. A Maalox
58 W. O. Weigle, "Natural and Acquired Immunologic Unresponsiveness." Cleveland World Publ., Cleveland, Ohio, 1967. 59 O. J. Plescia, W. Braun, and N. C. Palczuk, Proc. Natl. Acad. Sci. U.S.A. 52,279 (1964). 6o "Nucleic Acids in Immunology" (O. J. Plescia and W. Braun, eds.). Springer-Verlag, Berlin and New York, 1968.
60
PRINCIPLES AND METHODS
[2]
[AI(OH)3] suspension has also been used in conjunction with the addition of bacteria, such as Bordetella pertussis. The most popular and successful adjuvants have been the water in oil emulsions developed by Freund. The basic ingredients of light mineral oil (Bayol) and emulsifying agents mixtures such as Arlacel (A or C) are available commercially. The reagents are emulsified with either solutions or suspensions of the immunogen (incomplete Freund's adjuvant). The addition of mycobacteria (Mycobacterium butyricum, M. tuberculosis) in small amounts to the suspension (complete Freund's adjuvant) leads to a further enhancement of the immune response. This has been attributed to the increased local inflammatory response caused by the mycobacteria, el The well mixed and stable emulsion is injected either intraperitoneally, intradermally, or in the footpads of the host. If complete Freund's adjuvant has been used for the first injection, the secondary injections should not contain the mycobacteria, as further immune responses against the mycobacteria can be detrimental to the host and also lead to enhanced inflammatory responses. There are some additional advantages to using the complete adjuvant. The class of antibody formed is sometimes altered, leading to precipitating antibody, and ascites fluid can be formed following intraperitoneal injections in mice. Although it is not always predictable mice can produce large amounts of this fluid, which has concentrations of antibody almost equal to that found in the serum. On the negative side the investigator should be aware that the emulsion might destroy or mask some of the antigenic determinants of labile antigens. As a matter of convenience the immune response is divided into two stages. The primary response, resulting from an initial interaction with antigen, and the secondary response, resulting from subsequent contact with the same antigen. Quite often it is difficult to clearly delineate between the two; for example, an animal may have seen the antigen previously (particularly microbial antigens) or may have been exposed to cross-reacting antigens; however, for the present these factors will be discounted. The primary immune response is characterized by the appearance, within a few days of 19 S (IgM) antibodies. This is followed by a decline in 19 S levels and a rise in 7 S (IgG) antibody levels, which in general is dose dependent; for example, a good immunogen given in very low doses elicits little 7 S antibody, whereas in high doses quite significant amounts are formed. This phenomenon is nicely illustrated in the work of Uhr and Fine~ R. G. White, in "The Immunologically Competent Cell: Its Nature and Origin" (G. E. Wolstenholme and J. Knight, eds.). Churchill, London, 1963.
[2]
PROTEINS AND POLYPEPTIDES
AS A N T I G E N S
61
50 OOSEOF ~X
/,,~
1011
l/ iI
L p
°.1f
109 10 e
0.01
0.001
1
5
I 10
I
15
DAYS
FIG. 2. The primary 19 S ( - - ) and 7 S (---) antibody responses to ~bX174 bacteriophage (&X) in the guinea pig. Representative responses to intravenous injections of 1011, 109, or liP phage particles are shown. From Uhr and Finkelstein. 6z
kelstein. °a Using the bacteriophage 4 × 174, a very potent antigen, they showed that injection of 10s of 10a particles led to good IgM response with no IgG production. However, administration of 1011 phage led to high IgM titers followed several days later by a large increase in IgG (Fig. 2). After about 2 weeks the IgM response had fallen off and the antibody was mostly of the IgG class. There has been some question as to whether soluble protein antigens induce the same pattern of primary response as particulate ones, and it now appears that they do, although the length of time between injection of antigen and appearance of antibody (lag phase) may be delayed. The secondary response results from the readministration of antigen at a later time and is characterized by a rapid increase in antibody levels consisting mainly of IgG, but with some transient IgM. The most striking effect observed is the great increase in total serum antibody levels over that obtained in the primary response e3 (Fig. 3). After 2 - 3 weeks there is a e2 j. W. Uhr and M. S. Finkelstein, Prog. Allergy 10, 37 (1967). J. W. Uhr, M. S. Finkelstein, and J. B. Bauman, J. Exp. Med. 115, 655 (1962).
62
PRINCIPLES AND METHODS
[2]
10o000
1,000
100
k
10-
0
/
- ~,,
10
/
,,
/ / i -
X
i- 19SANTIBODY ~,X 7SANTIBODY
i/ o 0
I I J
.... 5
t .... I0
I,,,,,J,l,,
....
30
15
I .....
35
0.1 I ....
40
DAYS
FIG. 3. Antibody response to ~bX174 in the guinea pig after two intravenous injections. ---, ~bX174 bacteriophage (~bX); - - , 19 S antibody; response; ---, 7 S antibody response.
rapid decline in the IgG level until antibody concentrations reach a plateau, at which they may persist for weeks or months. The dose of antigen used for immunization, in addition to modulating the classes of immunoglobulin formed, can also have a profound effect on the ability of the animal to produce antibody at all. A refractory state, called tolerance, can be induced with most proteins having a low or moderate molecular weight (e.g., serum proteins) provided that the amount of antigen administered is within a given range. This tolerant state may be defined as the inability or diminished ability of an animal to react to a normally immunogenic material that has been induced by previous administration of the same material. It has shown that tolerance can be achieved within two distinct zones of antigen dosage. High zone tolerance is achieved when quantities of antigen much greater than the optimum immunizing dose are presented. Mitchison, 64 for example, induced tolerance to bovine serum albumin in mice by injection of 10 mg three times per week for 10 weeks. This type of tolerance is readily achieved with weaker immunogens (e.g., soluble proteins), but it is difficult to induce with potent immunogens because the quantities needed tend to be toxic or impractical to use. N. A. Mitchison, in "Immunogenicity" (F. Borth, ed.), p. 87. North-Holland Publ., Amsterdam, 1972.
[2]
PROTEINS AND POLYPEPTIDES AS ANTIGENS
63
Low zone tolerance is induced with subimmunogenic amounts of antigen, approximately 1/zg or less in the mouse. Repeated administration of antigen is usually necessary to induce or maintain the tolerant state, or both. There is still no foolproof method for choosing the route of administration of antigen in order to evoke a humoral response. In general, weaker immunogens are used with Freund's adjuvant and given intramuscularly or subcutaneously. With protein antigens the intravenous route is often used in tolerance induction; the intradermal route, minus adjuvant, is employed to induce delayed hypersensitivity. 65 A final consideration in regard to antigen dosage is that of the amounts, affinity, and specificity of the antibodies produced during the immune response. Siskind e t a l . 6~ have shown that in rabbits immunized with DNP-bovine y-globulin, high doses of antigen (50 mg) resulted in a rapid increase in serum antibody levels followed later in the immune response by a decrease and plateauing at low levels. When a low dose (0.5 mg) was used, the response began slowly but increased with time and eventually exceeded by threefold the high-dose level. These authors also showed a progressive increase in antibody affinity with time after immunization. Although all the doses used induced this effect, the increase was much greater with the lower ones. After a consideration of the cellular events involved in antibody production, it was suggested that these results could be interpreted as follows: the large doses of antigen injected might induce tolerance in "high affinity" cells resulting in a decrease in the amount and affinity of antibody produced. An alternative possibility is that large doses of antigen would favor differentiation of cells for antibody production resulting in the rapid appearance of high antibody titers but with a concomitant depletion in proliferating cells, thus limiting the response. Lower concentrations would give a more sustained and eventually greater response: In brief, the current explanation for an affinity increase is that after initial immunization a number of cell types, of both high and low affinity, are stimulated. Later, as the concentration of available antigen decreases owing to metabolism, only those cells that can bind antigen strongly (high affinity) are stimulated. This theory is also used to explain the generally observed increase in cross-reaction of antibodies that occurs with time during the immune response, by noting that, as affinities increase, determinants that would not have bound well with early (low affinity) antibodies now can be bound well enough to be detected. es j. W. Uhr, Physiol. Rev. 46, 359 (1966). G. W. Siskind, P. Dunn, J. G. Walker, J. Exp. Med. 137, 55 (1968).
64
PRINCIPLES AND METHODS
[2]
Typical Protocols for Antibody Production
In Vivo Techniques After consideration of the many factors that influence the immune response and the availability of material, most investigators prefer to incorporate the antigen in Freund's complete or incomplete adjuvant. The preparation for immunization need not be absolutely homogeneous, but the antigen to be used in the assay of the immune response should be highly purified. At times the purity of the immunogen may be important, as one might encounter the phenomenon of antigenic competition, wherein the response to the impurities might mask the responses against the putative purified enzyme. For immunization of large animals, such as rabbits, sheep, or goats, small to moderate amounts of antigen should be used (0.1-1.0 mg per kilogram of body weight). A typical protocol would employ 5 or 6 rabbits injected either in the footpads or intramuscularly into multiple sites with about 1-5 mg in complete Freund's adjuvant in a total volume of 0.250.5 ml. With smaller animals, such as mice and guinea pigs, microgram amounts (1-100/~g in 0.02-0.2 ml) are injected in the footpads subcutaneously or intraperitoneally. The animals are bled weekly for 4 - 6 weeks beginning about 3 weeks after the immunization; the sera are separated and tested qualitatively (see below) for the production of antibody. Rabbits and larger animals can be bled from the ear vein or via jugular vein puncture. With mice, sera can be obtained via retroorbital tappings of blood. Several techniques have been developed for producing ascites fluid in mice. ~5-19 In addition to the original adjuvant technique, a modification used in our laboratory involves injecting the antigen in complete Freund's adjuvant intraperitoneally followed by an injection 3 days later of 0.5 ml of pristane. TM Mice are boosted 7-10 days later and again after another 7-10 days. Another effective method is to administer intraperitoneal injections of Sarcoma 180 cells TM subsequent to several intraperitoneal injections of complete Freund's adjuvant. Distension of the abdomen is indicative of ascites development. The amount of ascites fluid that can be obtained at each tapping varies from 0.25 ml to 20 ml. If the level of antibody is rising significantly, or there is no response at all, booster injections need not be given. It is important to have intervals between injections and to be careful not to reinject if there are high levels of antibody. When the antibody level is rising slowly or has decreased, it is beneficial to reinject the animals with incomplete Freund's adjuvant or the antigen solution intraperitoneally. In the presence of high levels of antibody, anaphylactic shock can ensue in mice or rabbits.
[2]
PROTEINS AND POLYPEPTIDES
AS A N T I G E N S
65
Pooling of sera from different outbred animals is not recommended, as reactions to different determinants might have occurred in different animals. In addition, if the immunizing preparation is not absolutely pure, it is conceivable that some animals might have responded very well to the impurities. Changes do occur in the properties of the antibody produced over a period of time. In general high affinity antibody follows immunization with low doses of antigen, and the best sera in rabbits can be obtained about 3 - 5 months after immunization. Although not always predictable, animals that respond early after immunization usually produce the best antibody, n7 In Vitro Techniques
A unique and revolutionary adaptation of cell hybridization techniques to the construction of myeloma-like cell lines producing monoclonal antibodies with desired reactivities has revolutionized the approach to the production of immunospecific reagents. ~°-~4 Large amounts of specific antibody can be obtained after hybridization of lymphoid cells from an appropriately immunized donor (mouse) with cells from a mouse myeloma that has been adapted to growth in culture. Although the normal in vivo immune response to complex antigens leads to a very heterogeneous population of antibody molecules directed against many determinants, with the hybridoma technique each hybrid clone theoretically produces a single species of antibody specific for a single antigenic determinant. A recent symposium on cell hybridomas ~3and other publications 2°-24 discuss in depth both the technology and the applicability of the procedure. Essentially the protocol (Fig. 4) involves hybridizing spleen cells from a hyperimmunized donor with cells from an in vitro adapted enzyme-deficient myeloma. A neoplastic cell line used for producing fusions is X63AGS, a clonal line of myeloma MOPC21 that has been adapted to growth in vitro, in 8-azaguanine, and lacks the enzyme hypoxanthine guanine phosphoribosyltransferase (HGPRTase) required for rapid growth in tissue culture medium containing hypoxanthine, aminopterin, and thymidine (HAT medium). The fusing agent is polyethylene glycol (PEG) of a specific concentration and molecular weight. After fusion (hybridization) the cells are cultured in HAT medium. Before subculturing, the supernatants of the initial hybrid cells can be assayed for antibody production by a number of sensitive techniques that allow one to determine which clones from a complete spleen are involved in antibody production and therefore are worth further subculturing. Once established the clones can sr G. W. Siskind and B. Benacerraf, Adv. lmmunol. 10, 1 (1969).
66
PRINCIPLES AND METHODS
[2]
P3/X63-Ag 8 tumor cells grown in HAT medium
\
/
107 Tumor cells
108 Spleen cells
/
\
Mixed, fused in PEG
I
Spleen and fused cells cultured in HAT medium
I Fused celts diluted and cloned for growth
I
Cloned cells screened for antibody iroduction
Antibody-producer cells grown in mass culture
Injected in mice for antibody production (serum/ascites fluid)
FIG. 4. Multistep methods f o r eliciting specific antibody-producing ceils (hybridomas).
HAT, hypoxanthine, aminopterin, and thymidine; PEG, polyethylene glycol.
[2]
PROTEINS AND POLYPEPTIDES AS ANTIGENS
67
be kept growing either in culture or in vivo for many months. In the in vivo technique the culture-grown antibody-forming clones are injected into pristane primed mice. Within a few weeks the ascites-containing fluid that appears has monoclonal antibody. When grown in mass culture, antibody concentrations can reach about 50/zg/ml. Immunoadsorbent techniques can then be used to concentrate the specific antibody. In a number of situations where the hybridized cells have been injected in vivo, concentrations of 5-20 mg per milliliter of antibody have been produced. In the specific area of enzymology, success has been achieved in producing clones against human alkaline phosphatase, hen egg lysozyme, and horseradish peroxidase. °a A major advantage of this technique is that one can obtain monospecific antibodies directed against the immunogenic determinants of the enzyme, some of which might be against the active site of the enzyme. T e c h n i q u e s for Assaying for Antibody Initially a qualitative test, then a semiquantitative and, if deemed useful, a quantitative estimation of antibody can be performed.12-14 The table presents the sensitivity of some of the in vitro serological tests that can be used. (The advantages and disadvantages of the different techniques will be discussed by other authors.) When detecting immune responses in different species, one should reckon with the fact that not all antibody systems lead to a precipitation reaction. Rabbit sera containing precipitating COMPARISON OF MINIMAL CONCENTRATION OF ANTIBODY DETECTABLE BY SPECIFIC TEST
Immunological test 1. Fluid precipitation Interfacial (ring) test 2. Gel precipitation Double diffusion (Ouchterlony) Single diffusion (Oudin) 3. Hemagglutination Passive (indirect) 4. Radioimmunoassay
Antibody detectable (/zg N/ml) 20-30 3-15 10-100 0.001-0.03 0.001-less
See references to hybridomas produced against enzymes in "Lymphocyte Hybridomas" (F. Melchers, M. Potter, and N. L. Warner, eds.), Curt. Top. Microbiol. Immunol. 81, 19-22 (1978).
68
PRINCIPLES AND METHODS
[2]
antibody can be screened by both the agar diffusion techniques and/or reactions in liquid medium. It is best first to test the sera by precipitation in gel techniques. Although this is a secondary reaction, based upon complex interactions of antigen-antibody complexes following the initial interaction, a positive reaction (band formation) is indicative of significant concentrations of antibody. The ring or interfacial test involves carefully overlaying a solution of antibody and its dilutions with antigen so that a sharp interface is formed. Diffusion occurs between the two components until an optimal ratio (equivalence) for precipitation of the complexes is established. The gel diffusion tests involve carrying out the reactions between antigen and antibody in a semisolid medium. There are many modifications of this technique, i.e., Preer, Ouchterlony, and Oudin, all of which lead to " b a n d " formation. 14 In addition to detecting antibody qualitatively and estimating the approximate equivalence ratio needed for optimal precipitation between antigen and antibody, the agar diffusion techniques can indicate the number of antigen-antibody systems that might be present, providing that the diffusion effects due to temperature and concentration are controlled. A disadvantage of the reaction in liquid medium resides in the fact that if the concentration of antigen added is too high, soluble antigen-antibody complexes will form in antigen excess and it may appear as though there is no significant amount of antibody present. Therefore varying concentrations of antigen must be added to both undiluted and varying dilutions of serum in a checkerboard fashion. Although there are many in vitro and in vivo methods for estimating antibody in serum or other fluids, only the quantitative precipitin reaction gives a reliable and accurate measure of antibody in absolute weight units. When a proper assessment of the many factors influencing the reaction is made and rigorous procedures of quantitative analysis are employed in analyzing the washed antigen-antibody specific precipitate, quantitive antibody values are obtained, al In addition to the direct reaction of the antigen with antibody, techniques are available for performing "indirect" reactions. Antigens can be coupled chemically or via a "tannic acid" procedure to red blood cells. The antigen-coated erythrocytes in the presence of specific antibody then agglutinate (clump) as do red blood cells in the presence of antibody to the red blood cells (hemagglutination). Rather than measuring the capacity of an antiserum to combine with antigen, all the above-mentioned antibody tests measure the capacity of an antiserum to produce secondary effects, i.e., precipitation and complement fixation, following the primary antigen-antibody interaction.
[2]
PROTEINS AND POLYPEPTIDES AS ANTIGENS
69
Measuring the antigen binding capacity rather than the absolute amount of antibody in a precipitate becomes important when it is realized that some classes of antibodies, and antibodies from some species, do not precipitate well even in the presence of "optimal" proportions of antigen and antibody. A number of reactions do exist that measure the primary binding between antiserum and low molecular weight haptens employing modifications of equilibrium dialysis techniques, s9 However, the application to large macromolecules has been more difficult to develop.The ammonium sulfate test (Farr) was developed with the antigen bovine serum albumin (BSA), which was labeled with lalI, to fill the need for a primary binding test suitable for nondialyzable large macromolecules. ~° As devised it can measure the capacity of antisera to combine with the soluble macromolecular antigens and can detect both precipitating and nonprecipitating antibody. The principle of the reaction depends upon the fact that the antigen (BSA) is soluble in 50% saturated ammonium sulfate, whereas antigen-antibody complexes, which assume the solubility properties of the antibody, are insoluble under the same conditions. One of the serious limitations of this technique is the need for the antigen to be soluble in 50% saturated ammonium sulfate. However, modifications have been developed employing anti-immunoglobulin serum (rather than ammonium sulfate), which precipitates the soluble complexes. 71 The anti-immunoglobulin serum is directed against the immunoglobulins of the specific species being tested. This reagent must be checked for its ability to precipitate all the immunoglobulin in the serum being assayed. The precipitation of the radioactive antigen in the presence of increasing dilutions of serum is a measure of the amount of the antibody; that is, the greater the dilution of antiserum against a specific immunogen that still combines with and precipitates a constant amount of antigen, the greater is the strength of the serum. This double-antibody or radioimmunoprecipitation test has been used to measure a variety of hormonal, microbial, and tumor antigens. Both the ammonium sulfate and anti-immunoglobulin techniques have been used not only to measure the presence or the concentration of antibody, but to detect very small amounts of antigen in fluids or solutions. (Discussions and applications of the various radioimmunoassay procedures referred to here are given elsewhere in this volume.)
C. W. Parker, "Radioimmunoassay of Biologically Active Compound." Prentice-Hall, New York, 1976. 7o R. S. Farr, J. Infect. D/s. 103, 239 (1958). 74 p. Minden, R. S. Farr, and J. Trembath, lmmunochemistry 12, 477 (1975). s0
70
PRINCIPLES
AND METHODS
[3]
Acknowledgments T h e authors' research was supported by grants from the National Institutes o f Health, Institute for Allergy and Infectious Diseases, A0107825; The American C a n c e r Society, IM5H; the National F o u n d a t i o n - - M a r c h o f Dimes, 1-492.
[3] T h e
Experimental Induction to Nucleic Acids
of Antibodies
By B. DAVID STOLLAR
Antibodies to nucleic acids have found many uses in the specific measurement of naturally occurring or modified nucleic acids both in solution and in situ. To obtain the required antibodies, it has been necessary to link nucleic acids or their components to cartier proteins or synthetic polypeptides to form immunizing complexes because injection of purified nucleic acids alone into normal animals does not stimulate significant antibody production. Once the antibodies are formed, they react with the nucleic acid in the absence of carrier. A variety of immunogens have been developed, with either small fragments, such as nucleotides or oligonucleotides conjugated covalently to proteins, or with high molecular weight polynucleotides in physical complexes with protein carriers. With these immunogens, antibodies specific for each of the normal bases of DNA and RNA, or for modified bases or base sequences, have become available as selective reagents. Antibodies that recognize helical shapes have also become available. The applications of anti-nucleic acid antibodies have included localization of specific modified bases in ribosomes 1 or chromosomes2; identification of denatured DNA in replicating DNA in situ3; studies of the denaturation and renaturation of DNA4; identification of double-stranded RNA intermediates of viral replicationS; gene localization by in situ hybrid detection6; isolation of DNA enriched in specific genes7; measurement of ultraviolet i S. M. Politz and D. G. Glitz, Proc. Natl. Acad. Sci. U.S.A 74, 1468 (1977). 2 R. R. Schreck, V. G. Dev., B. F. Erlanger, and O. J. Miller, Chromosoma 62, 337 (1977). 3 W. J. Klein, S. M. Beiser, and B. F. Erlanger, J. Exp. Med. 125, 61 (1967). 4 L. Levine, J. A. G o r d o n , and W. P. J e n c k s , Biochemistry 2, 168 (1963). 5 V. Stollar, T. E. Shenk, and B. D. Stoilar, Virology 47, 122 (1972). s G. Rudkin and B. D. Stollar, Nature (London) 2,65, 472 (1977). 7 W. E. S t u m p h , J. R. W u , and J. Bonner, Biochemistry 17, 5791 (1978).
METHODS IN ENZYMOLOGY, VOL. 70
CopyrightO 19$0by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-181970-1
70
PRINCIPLES
AND METHODS
[3]
Acknowledgments T h e authors' research was supported by grants from the National Institutes o f Health, Institute for Allergy and Infectious Diseases, A0107825; The American C a n c e r Society, IM5H; the National F o u n d a t i o n - - M a r c h o f Dimes, 1-492.
[3] T h e
Experimental Induction to Nucleic Acids
of Antibodies
By B. DAVID STOLLAR
Antibodies to nucleic acids have found many uses in the specific measurement of naturally occurring or modified nucleic acids both in solution and in situ. To obtain the required antibodies, it has been necessary to link nucleic acids or their components to cartier proteins or synthetic polypeptides to form immunizing complexes because injection of purified nucleic acids alone into normal animals does not stimulate significant antibody production. Once the antibodies are formed, they react with the nucleic acid in the absence of carrier. A variety of immunogens have been developed, with either small fragments, such as nucleotides or oligonucleotides conjugated covalently to proteins, or with high molecular weight polynucleotides in physical complexes with protein carriers. With these immunogens, antibodies specific for each of the normal bases of DNA and RNA, or for modified bases or base sequences, have become available as selective reagents. Antibodies that recognize helical shapes have also become available. The applications of anti-nucleic acid antibodies have included localization of specific modified bases in ribosomes 1 or chromosomes2; identification of denatured DNA in replicating DNA in situ3; studies of the denaturation and renaturation of DNA4; identification of double-stranded RNA intermediates of viral replicationS; gene localization by in situ hybrid detection6; isolation of DNA enriched in specific genes7; measurement of ultraviolet i S. M. Politz and D. G. Glitz, Proc. Natl. Acad. Sci. U.S.A 74, 1468 (1977). 2 R. R. Schreck, V. G. Dev., B. F. Erlanger, and O. J. Miller, Chromosoma 62, 337 (1977). 3 W. J. Klein, S. M. Beiser, and B. F. Erlanger, J. Exp. Med. 125, 61 (1967). 4 L. Levine, J. A. G o r d o n , and W. P. J e n c k s , Biochemistry 2, 168 (1963). 5 V. Stollar, T. E. Shenk, and B. D. Stoilar, Virology 47, 122 (1972). s G. Rudkin and B. D. Stollar, Nature (London) 2,65, 472 (1977). 7 W. E. S t u m p h , J. R. W u , and J. Bonner, Biochemistry 17, 5791 (1978).
METHODS IN ENZYMOLOGY, VOL. 70
CopyrightO 19$0by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-181970-1
[3]
ANTIBODIES TO NUCLEIC ACIDS
71
irradiation-induced damage to DNAS; and quantitation of circulating components, such as thymidine a or modified bases of tRNA. 1° Immunogens P r e p a r e d by Covalent Linkage of Nucleic Acid Components to Proteins Principles of I m m u n o g e n Formation Nucleic acid components have been used as classical haptens, with immunogens prepared as covalent hapten-protein conjugates in a number of ways. Purine or pyrimidine bases have been conjugated directly to protein for this purpose, 1' but nucleosides, nucleotides, and oligonucleotides have been used much more extensively. Three procedures for preparing such immunogens will be described in detail in this chapter. For ribonucleosides or ribonucleotides, the most convenient linkage involves periodate oxidation of the furanose ring bearing adjacent hydroxyl groups; this is followed by condensation of the dialdehyde product of oxidation with lysine amino groups of protein. A stable covalent bond is then formed on addition of a reducing agent such as sodium borohydride.12 With deoxyribonucleotides, which do not have adjacent hydroxyl groups, water-soluble carbodiimides have been used to form phosphoramidate conjugates through the 5'-phosphate and protein amino groups.13 With cyclic nucleotides, which do not have free adjacent hydroxyl groups or a free phosphate group, the 2'-hydroxyl group can be succinylated with succinic anhydride and this product can be conjugated to protein amino groups with a carbodiimide reagent. 14 Deoxyribonucleosides also lack both the adjacent hydroxyls and free phosphate group; in this case, the 5'-hydroxyl has been oxidized to a carboxyl group, which is then linked to a synthetic polypeptide carrier with a water-soluble carbodiimide. 15 This procedure has been described in this series (see Vol. 12B [175]). Two other procedures have been used less exs E. Seaman, H. Van Vunakis, and L. Levine, J. Biol. Chem. 247, 5709 (1972). 9 W. L. Hughes, M. Christine, and B. D. Stollar, Anal. Biochem. $$, 468 (1973). ~0 L. Levine, T. P. Waalkes, and L. Stolbach, J. Natl. Cancer Inst. 54, 468 (1975). 1~ V. P. Butler, S. M. Beiser, B. F. Edanger, S. W. Tanenbaum, S. A. Cohen, and A. Bendich, Proc. Natl. Acad. Sci. U.S.A. 48, 1597 (1962). 12 B. F. Erlanger and S. M. Beiser, Proc. Natl. Acad. Sci. U.S.A. 52, 68 (1964). ~3 M. J. Halloran and C. W. Parker, J. l m m u n o l . 96, 373 (1966). 14 A. J. Steiner, D. M. Kipnis, R. Utiger, and C. Parker, Proc. Natl. A c a d . Sci. U . S . A . 64, 367 (1967). 15 M. Sela, H. Ungar-Waron, and Y. Schechter, Proc. Natl. Acad. Sci. U.S.A. 52, 285 (1%4).
72
PRINCIPLES AND METHODS
[3]
tensively. One involves formation of a p-aminobenzoate conjugate of the 3'-hydroxyl of thymidine, followed by its diazotization and reaction of the diazonium salt with tyrosine in the protein. 16 The second, applicable to guanine-containing oligonucleotides, involves the photooxidation of guanine with methylene blue and visible light, and linkage through the oxidized product to amino groups of protein. 17 Principles of Specificity Antibodies induced by nucleosides or nucleotides show marked specificity for the purine or pyrimidine base component. When cross-reactivity does occur with other bases, it may involve distinct populations of antibodies that can be removed by absorption, or it may involve all the antibody reacting with a lower affinity, so that large concentrations of crossreacting nucleoside are required. Specific sera can be obtained and used, at suitable dilutions, as selective reagents for any of the normal bases of RNA or DNA or for a wide range of modified base analogs. TM Antibodies to nucleosides usually react with nucleosides equally or only slightly better than with nucleotides (Fig. 1). They also react well with polymers in which the bases are accessible, such as denatured DNA. When the periodate oxidation technique (with ring breakage and later reduction) is used for immunogen preparation, the resulting antibodies react slightly better with deoxyribonucleosides than with ribonucleosides. The use of nucleotide-protein immunizing conjugates induces antibodies that recognize the base, sugar, and phosphate group~a; they show marked specificity for the base component, but also react much better with the nucleotide than with the nucleoside or free base (Fig. I). Still, there is not a great differentiation between the ribonucleotide and deoxyribonucleotide. When dinucleotide-protein or trinucleotide-protein" conjugates are used as immunogens, there is some specificity for the base sequence used, 2°-22 though mononucleotides or partial sequences show some cross-reactivity. In such cases, the innermost base (closest to the
is j. p. Coat, S. David, and J. C. Fischer, Bull. Soc. Chim. Fr. 2489 (1965). ~7 E. Seaman, L. Levine, and H. Van Vunakis, Biochemistry 5, 1216 (1966). 18 B. D. Stollar, in '*The Antigens" (M. Sela, ed.), Vol 1, p. 1. Academic Press, New York, 1973. 19 M. Z. Humayun and T. W. Jacob, Biochim. Biophys. Acta 331, 41 (1973). s0 S. S. Wallace, B. F. Erlanger, and S. M. Beiser, Biochemistry 10, 679 (1971). zl B. Bonavida, S. Fuchs, M. Sela, P. W. Roddy, and H. Sober, Fur. J. Biochem. 31, 534 (1972). z z R. M. D'~lisa and B. F. Erlanger, Biochemistry 13, 3575 (1974).
[3]
73
ANTIBODIES TO NUCLEIC ACIDS 100
75
5G
c 2.'
"at- i
0.25
4-
,
,'1~
0.5
1
p6Aoles
I
I
I
0.25
2
|
0.5
t
I
1
2
Inhibitor
FIG. 1. Specificity of antinucleoside and antinucleotide antibodies. (A) Inhibition of the precipitation of antiadenosine antibodies and adenosine-serum albumin by adenosine (O), deoxyadenosine (O), adenosine 5'-monophosphate (A), adenine (A), and ¢ytidine, thymidine, or guanosine (+). (B) Inhibition of the precipitation of anti-GMP antibodies and GMPserum albumin by GMP (11), dGMP ([3), and guanosine (x).
protein in the conjugate) may contribute an unexpectedly large part of the specificity.2° Antibodies to mono-, di-, and trinucleotides of the usual nucleic acid bases react with denatured DNA but not with native DNA, in which the bases are not accessible. It has been difficult in many cases to measure their reactions with ribosomal RNA or single-stranded viral RNA, partly because of the extensive secondary and tertiary folding that may mask many of the bases and partly because of the difficulty in removing all ribonuclease from serum. Some reaction with RNA was measurable with an antiadenosine serum after careful efforts were made to remove ribonuclease. 2a Antibodies to anticodon sequences reacted with tRNA, 22 as did antibodies to modified bases of tRNA 24 and antibodies occurring spontaneously in sera of N Z B / N Z W mice or human patients with systemic lupus erythematosus. 2s,26 ~3 B. J. Rosenberg, B. F. Erlanger, and S. M. Beiser, Biochemistry 12, 2191 (1973). 24 R. Salomon, S. Fuchs, A. Aharonov, D. Giveon, and U. Z. Littauer, Biochemistry 14, 4046 (1975). 25 D. P. Eilat, P. Di Natale, A. D. Steinberg, and A. N. Schechter, J. Immunol. 118, 1016 (1977). ~ D. Eilat, A. D. Steinberg, and A. N. Schecter, J. lmmunol. 120, 550 (1978).
74
PRINCIPLES
AND METHODS
[3]
Preparations of I m m u n o g e n s
Ribonucleoside-Protein Conjugates The preparation of immunogens by periodate oxidation of ribonucleosides or nucleotides was introduced in 1964 by Erlanger and Beiser ~2 and has been used extensively since. The original procedure has been modified~° and can be simplified to the following steps. About 0.1 mmol of nucleoside or nucleotide (20-40 mg) is dissolved in a solution of 0.1 M sodium metaperiodate in water. The volume of the periodate solution is chosen so as to give about an equivalent amount of this reagent; thus 1.0 ml of a 0.1 M solution is used for 0.1 mmol of nucleoside. The oxidizing mixture is stirred at room temperature for 20 min and then added, dropwise, to 2 ml of hemocyanin or other protein (5-10 mg/ml in 0.1 M bicarbonate-carbonate buffer, pH 9.5); the pH is readjusted to 9.5, if necessary, with 5% K2CO3. The solution is stirred at room temperature for 1 hr. Then 100 mg of sodium borohydride in 5 ml of water are added, and the mixture is placed at 4° for 1 hr or, if more convenient, overnight. The solution is then dialyzed extensively against 0.1 M NaCI, 30 mM NaHCO3 or a neutral buffered saline solution; some conjugates become insoluble if the pH is allowed to drop to less than 6. In this procedure, the extent of substitution is most profoundly affected by the pH of the reaction mixture containing oxidized nucleoside and protein (Fig. 2). Modifications of the basic technique are required in some instances. Guanosine is not soluble until it is oxidized. To avoid formation of unmanageable clumps during oxidation, the 0.1 mmol of guanosine is first dispersed in 1 ml of distilled water. Periodate solution is then added, and the solution is stirred well during oxidation. Nearly all the oxidized guanosine dissolves, usually producing a viscous solution or gel. This state is maintained until the sodium borohydride addition step. A different problem occurs with 7-methylguanosine. The purine base of this nucleoside is hydrolyzed at basic pH, and very significant degradation occurs in an hour at pH 9.5 at room temperature. The mixture of oxidized nucleoside and protein, therefore, is kept at pH 9.1 and at 0-4 °, at which only very limited hydrolysis of the base occurs in 1 hr. 27 A different reducing agent, tert-butylamine borane (Aldrich Chemical Co.), is used; the reduction is done for only 30-60 min at 4 °, and the product is separated from free nucleoside on a Sephadex G-25 column at 4 °. The nucleotide of the 7-methylguanosine is more stable than the nucleoside, and it also has been used to induce antibody to the 7-methylguanine structure. ~8 27 L. Rainen and B. D. Stoilar, Nucl. Acids Res. 5, 4877 (1978). 2s R. D. Meredith and B. F. Erlanger, Fed. Proc. 37, 1503 (1978).
[3]
ANTIBODIES TO NUCLEIC ACIDS .c_
//
20
o o_
O
"(3
75
10
o
z O I
7
I
8 pH
I
9
I
10
FIG. 2. The dependence of nucleoside-protein conjugation on pH. A mixture of periodate-oxidized adenosine and cytidine was added to bovine y-globulin to give final concentrations of 4 mM nucleoside (0.9 mg/ml) and 6.7/aM protein (1 mg/ml) in 0.2 M Veronal buffer titrated to varying pH. These mixtures were incubated at room temperature for 1.5 hr. Then sodium borohydride was added to a final concentration of 0.4 M (15 mg/ml), and samples were incubated for 2.5 hr at 4°. They were then dialyzed extensively against 0.1 M NaCI and analyzed for protein and nucleoside composition.
In this case, the reducing agent was cyanoborohydride.2s To verify that the conjugate contains intact purine base, a difference spectrum of the hapten-protein conjugate minus protein should be obtained; it should be very close to the spectrum of the intact nucleoside or nucleotide alone. Further, the antibodies should be specific for the intact base in comparison with the degradation productY Modifications are made also in order to use this reaction to conjugate nucleosides to erythrocyte surfaces, allowing use of the coated cells as targets for assays of antibody-forming splenic lymphocytes.29 Nucleoside, 10-20 mg, is oxidized in 1.5 ml of 0.1 M sodium periodate in 0.15 M NaHCO3 for 20 min at room temperature; the reaction is stopped by the addition of 15/zl of ethylene glycol. Sheep erythrocytes are washed twice with 0.15 M NaHCOa, and 0.5 ml of packed cells is then suspended in 2.0 ml of the bicarbonate solution in a 40-ml centrifuge tube. The oxidized nucleoside is added dropwise to the cell suspension and the mixture is kept at room temperature for 15 min. tert-Butylamine borane (Aldrich Chemical Co.), 100 mg in 5 ml of 0.15 M NaHCOa, is added. The suspension is kept at room temperature for 3 min, and the tube is then quickly filled with bicarbonate solution and centrifuged at 1500 rpm for 10 min. 29 B. D. Stollar and Y. Borel, Nature (London) 267, 158 (1977). This procedure is quoted with permission of Macmillan Journals Ltd., publishers of Nature.
76
PRINCIPLES AND M E T H O D S
[3]
The modified cells are washed twice more with the bicarbonate solution and are ready for use as target cells in the hemolytic plaque assay.
Carbodiimide-Linked Nucleotide-Protein Conjugates This procedure was introduced by Halloran and Parker for use with mono- and oligonucleotides. 13 Humayun and Jacob modified it to reduce the amount of insoluble aggregate they obtained with the original method, especially with purine nucleotides. 19 In a reaction following the modified procedure, about 0.25 mmol of solid 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide is added to 0.1 mmol of nucleotide in 0.5 ml of water. The pH is adjusted to 7 with dilute NaOH if necessary, and the solution is incubated at 60° for 10 min. The mixture is then added dropwise to a protein solution of 10-20 mg in 0.2-0.4 ml of 0.15 M NaCl, and this solution is kept in the dark overnight at room temperature. The product is then separated from the free nucleotide and the hydrolyzed carbodiimide reagent by dialysis or gel filtration. For some purposes, it may be important to note that the carbodiimide can modify the protein carboxyl groups, especially at low pH. 3° The conjugate, therefore, is often very positively charged and will form a precipitate with DNA. The carboxyl group modification can be reversed to a large extent by dialysis of the conjugated protein at pH 10, with no loss of nucleotide substitution. Exposure of protein to carbodiimide at high pH (9.5) produces a stable modification of lysine amino groups. 3° It should be noted also that carbodiimide can modify guanine, thymine, and uracil if the pH is above 8. 31
Succinylated Nucleotides Steiner et al. prepared an immunogen by 2'-O-succinylation of cyclic AMP and linkage of the product to protein with a water-soluble carbodiimide. 14 Although amino functions, such as the 6-amino group of adenine, can be succinylated also, the reaction of the 2-hydroxyl group is much more rapid and can be achieved selectively. 3~ This procedure may be of more general use, as with oligonucleotides having a free 3'- or 5'-hydroxyl group. In the use of this procedure with 3',5'-cyclic AMP (cAMP), 33 morphoa0 j. p. Riehm and H. A. Scheraga, Biochemistry 5, 99 (1966). This detailed study o f this modification used p H 4.5. We have found it to occur to some extent at pH 6 as well. a~ N. W. Y. Ho and P. T. Gilham, Biochemistry 6, 3632 (1967). as j. G. Falbriard, T. H. Pasternak, and E. W. Sutherland, Biochim. Biophys. Acta 148, 99 (1967). a3 A. L. Steiner, C. W. Parker, and D. M. Kipnis, J. Biol. Chem. 247~ 1106 (1972). This procedure is quoted with permission o f the author and the Journal of Biological ChemistO'. We have also applied it to noncyclic deoxyribonucleotides.
[3]
ANTIBODIES TO NUCLEIC ACIDS
77
line N,N'-dicyclohexylcarboxamidine (0.76 mmol) is dissolved in 7.5 ml of hot anhydrous pyridine and 0.7 mmol of cAMP (free acid) is added slowly over a period of 30-60 min. After this mixture has cooled, 10 mmol of succinic anhydride are added and the suspension is stirred at room temperature for 18 hr. Unreacted succinic anhydride is hydrolyzed by the addition of 3.75 ml of water, and the reaction mixture is allowed to stand for an additional 2 hr at 4 °. Thin-layer chromatography of the reaction mixture on cellulose with butanol-glacial acetic acid-water (12:3:5 v/v) shows 2'-O-succinyl-cAMP which runs ahead (Re 0.42) of cAMP (Re 0.30). Pyridine is removed by repeated rotary evaporation at 40 ° under reduced pressure, and the residue is dissolved in 2-3 ml of water after the pH is adjusted to 4.5. The succinyl-cAMP is purified by chromatography on a column (1.5 × 44 cm) of Dowex 50 (H ÷ form) with distilled water as the eluent, at a flow rate of 30 ml/hr. Succinic acid appears in the first 50 ml, succinylated cAMP elutes in tubes 30 to 45, and cAMP in tubes 50 to 65. The yield of succinylated cAMP varies from 45 to 60%. The product is conjugated to protein by incubation of 20 mg of protein in 2 ml of water with l0 mg of succinylated cAMP and 10 mg of l-ethyl-3(3-dimethylaminopropyl)carbodiimide, at pH 5.5 for 16 hr at room temperature in the dark. The conjugated protein is separated from free reagents by dialysis.
Measurement of Hapten Substitution The extent of substitution is determined from the ultraviolet (UV) absorbance spectrum and the protein concentration. The latter is determined by a standard chemical assay. The protein contribution to the A~e0 of the conjugate is determined from the known spectrum of unmodified protein for the measured concentration. The difference between the total A260of the conjugate and the A~eodue to protein is contributed by the nucleoside. The molar concentration of nucleoside is calculated from this difference, using published extinction coefficients.34 Immunization A wide range of doses and immunization schedules has been used successfully. A convenient schedule for immunization of rabbits is to inject 200-500 ~g of conjugated protein, emulsified in complete Freund's adjuvant, intradermally and subcutaneously on the first day, and to inject the same dose in incomplete adjuvant on days 14 and 21. Serum can be obtained 5-7 days later. Further bleedings may be done weekly, with intradermal booster injections given if the serum antibody levels fall. After an 34"Handbookof Biochemistry,MolecularBiology"(G. Fasman, ed.), 3rd ed., Section B, Vol I. Chem. Rubber Publ. Co., Cleveland,Ohio, 1976.
78
PRINCIPLES AND METHODS
[3]
intravenous booster injection, the serum antibody level may peak and fall quickly, especially in early courses of immunization. In a recent study, involving immunization with 7-methylguanosine-bovine serum albumin, a schedule of small doses and intradermal and subcutaneous injections induced more antibody (and antibody that was less cross-reactive with guanosine) than did intravenous immunization. ~7 In mice, the response varies in different strains, but in both moderateand high-responding strains, a dose of 200/~g of nucleoside-hemocyanin induces higher antibody production than l0/zg or 50/~g.a5 Injections of antigen in complete Freund's adjuvant are given intraperitoneally on days 1 and 21, and high levels of antibody are present on day 26; a further injection on day 31 gives still higher levels on day 36; further immunization does not usually increase the antibody levels beyond that.
Assays for Antibody Two-dimensional immunodiffusion assays indicate qualitatively whether specificantihaptcn antibodies have been formed; ifthey arc present in thc serum, the precipitation line sccn with conjugated protein will spur ovcr that seen with the carrier protein alone, and precipitation will occur with haptcn linked to an unrelated carrier molecule. Quantitative precipitation tests indicate both h o w much haptcn-spccific antibody is present and how much unconjugatcd protein would bc required to absorb out all the antibodies directed against the protein alone. This absorption can then bc done by incubating equivalence conccntrations of antigen and serum and removing the precipitate.Absorption can also bc done with protein-agarosc affinitycolumns. In addition, the antibody can be purifiedwith hapten-bearing affinitycolumns, from which hapten-spccific antibody can be elutcd with 2 M acetic acid or with excess hapten. Antinucleotidc scra have been tested also for binding of radiolabcled hapten-protein conjugates or labeled D N A and with quantitative complement fixation assays. Once characterized, specific serum or purified antibodies may bc uscd also in immunohistochcmical tests at the level of light or electron microscopy. Single-Stranded Polynucleotides as H a p t e n s Principles of Immunization and Specificity Plescia e t al.36 contributed a major advance in the induction of anti-nucleic acid antibodies when they demonstrated that insoluble complexes of 35y. Borel and B. D. Stollar,Eur. J. Immunol. 9, 166 (1979). ~eO. J. Plescia, W. Braun, and N. C. Palczuk,Proc. Natl. Acad. Sci. U.S.A. 52, 279 (1964).
[3]
ANTIBODIES TO NUCLEIC ACIDS
79
denatured DNA and the positively charged methylated bovine serum albumin (MBSA) could elicit antibodies to the nucleic acid portion of the complex. This procedure has been applied since to synthetic homopolymers, 37 to modified DNA such as UV-irradiated 8 or photooxidized (visible light and dye) DNA, 17 to poly(adenosine-diphosphoribose), as and to a variety of helical nucleic acids as well. Native DNA, native tRNA, and single-stranded viral RNA have not been rendered immunogenic in this way.
Immunization with denatured D N A - M B S A gives rise to antibodies, often mainly of the IgM class, 3a that react with single-stranded DNA but not with native DNA. The largest antigenic determinant for such antibodies is about the size of a pentanucleotide. 4° When DNA with an unusual base, such as the glucosylated hydroxymethylcytosine of T-even phage, is used, the specificity is directed largely to the modified base and much more IgG may be produced. 41"42 Similarly, when UV-irradiated DNA or photooxidized DNA is used, the modified bases of the lesions provide the major specificity determinants, a'l~'4a Immunization with homopolynucleotide-MBSA complexes gives rise to antibodies that are specific for an oligonucleotide segment of the polymer. They probably recognize a number of bases in sequence, perhaps in a stacked array. Poly(I), poly(C), and poly(A) each induce specific antibodies that show little cross-reaction with the other homopolynucleotides, and slight cross-reaction with denatured DNA. 37 Procedures for preparation of the immunizing antigen, injection schedules, and assays for antibody formation are similar to those used for helical polynucleotide-MBSA antigens and are discussed in detail in the following section and in Vol. 12B [174] of this series. Helical Nucleic Acids as Antigens Principles of Specificity Antibodies can distinguish fine structural differences among helical polynucleotides. 44,45In doing so, they probably recognize conformational 37 E. Seaman, H. Van Vunakis, and L. Levine, Biochemistry 4, 1312 (1965). 3s y . Kanai, M. Miwa, T. Matsushima, and T. Sugimura, Biochem. Biophys. Res. Commun. $9, 300 (1974). 39 A. L. Sandberg and B. D. Stoilar, Immunology 11, 547 (1966). 40 A. Wakizaka and E. Okuhara, lmmunochemistry 12, 843 (1975). 4~ E. Seaman, L. Levine, and H, Van Vunakis, Biochemistry 4, 2091 (1965). 42 R. Gruenewald and B. D. Stollar, J. lmmunol. 111, 106 (1973). 43 L. A. Zamchuk, N. A. Braude, and D. M. Goldfarb, Immunochemistry 13, 81 (1976). 44 F. Lacour, E. Nahon-Merlin, and W. Michelson, Curt. Top. Microbiol. lmmunol. 62, 1 (1973). 45 B. D. Stollar, Crit. Rev. Biochem. 3, 45 (1975).
80
PRINCIPLES
AND METHODS
[3]
variations that alter the steric relationship of the backbone phosphate and furanose groups. This allows the development of a particularly useful type of antibody reagent that reacts with a given class of helical structure. For example, one can obtain antibodies that react with double-stranded RNA (dsRNA) of any origin, but do not react with helical native DNA and react only very weakly with R N A - D N A hybrid helices. Similarly, one can induce antibodies that react with R N A - D N A hybrids as a class of structure but not with dsRNA or dsDNA. A third reagent, present in sera of some patients with systemic lupus erythematosus (SLE), but not yet inducible in normal experimental animals, reacts with native dsDNA but only weakly with R N A - D N A hybrids and not at all with dsRNA. Used in suitable assays, these reagents allow the selective measurement of dsRNA, dsDNA, or R N A - D N A hybrid in the presence of large amounts of other nucleic acids. Consistent with the suggestion that specificity is determined by helical conformation of the antigen (and important for the wide application of the antibodies) is the finding that reactions of these antibodies do not depend on specific base sequences or even base composition in the polynucleotide. Antibodies induced by poly(A).poly(U) react very well with poly(I).poly(C) or viral dsRNA of mixed base composition. Quantitative distinctions do occur within this class of structures, 46 and there may be some advantage to using a viral dsRNA as immunogen, for example, to obtain the strongest reactions with natural dsRNA of mixed base composition47; however, antibodies induced by the synthetic forms have been effective reagents for viral dsRNA in several studies. Similarly, antibodies to the hybrid poly(A).poly(dT) react well with poly(I).poly(dC) or hybrids of natural RNA and DNA. The SLE anti-native DNA antibodies react with synthetic poly(dAT) or native DNA of any viral, plant, or animal origin. Narrower specificities have been obtained with antibodies to some unusual helical structures. Poly(dG).poly(dC) induces antibodies specific for the immunogen and unreactive with other deoxyribonucleotide polymers, such as poly(dAT) or native DNA. Double-helical polyribonucleotides with modified furanoses, such as poly(A).poly(2'-O-methylU), induce antibodies that react with a number of polymers bearing 2'-furanose substitutions (such as methyl or ethyl groups on either the purine or pyrimidine-containing strand). 46 Poly(G).poly(C) induced antibodies of narrow specificity in our studies, but Lacour and co-workers obtained antipoly(G).poly(C) that cross-reacted with several forms of viral RNA. 48 Antibodies specific for triple-helical polynucleotides clearly distin4e M. I. J o h n s t o n and B. D. Stoilar, Biochemistry 17, 1959 (1978). 47 R. I. B. Francki and A. O. Jackson, Virology 48, 275 (1972). 48 E. Nahon-Merlin, A. M. Michelson, C. Verger, and F. L a c o u r , J. lmmunol. 107, 222 (1971).
[3]
ANTIBODIES TO NUCLEIC ACIDS
81
guish between three-stranded and two-stranded structures, and among different three-stranded structures. Examples of polymers that are qualitatively distinct are poly(U).poly(A).poly(U), poly(U).poly(dA).poly(U), and poly(U).poly(A).poly(I).49,5° Preparation of I m m u n o g e n s Synthetic polynucleotides are convenient immunogens for obtaining antibodies to helical nucleic acids; naturally occurring helical forms may be preferable if available in adequate supply (100/zg or more). The synthetic forms may be purchased as helical polymers, such as poly(A).poly(U) for dsRNA or poly(A).poly(dT) for hybrids. Alternatively, a wide variety of double- or triple-helical forms can be prepared from mixtures of homopolymers; because they have perfect base pairing homology, most complementary homopolymers anneal readily. Their interaction can be verified by measurement of the UV absorbance spectra of the separate homopolymers and of the annealed mixture; hypochromicity will be observed in the mixture (though not always at 260 nm). To establish whether a double- or triple-stranded structure is produced, one can mix the homopolymers in varying proportions and determine whether the greatest hypochromicity occurs with a 1 • 1 mixture or a 1 : 2 mixture of homopolymers. The assay of stoichiometry by measurements of hypochromicity with varying proportions can help to minimize the presence of an excess of free single-stranded homopolymer and the formation of corresponding antibody. If some antibody to one of the homopolymers is formed, it should be removed from the serum by absorption. To form double-stranded poly(A).poly(U), equimolar amounts of the two homopolymers are mixed at a concentration of 0.6-1.5 p,M nucleotide (about 200-500/zg/ml) in 0.1 M NaC1, 0.01 M phosphate pH 7 at room temperature and allowed to anneal for a few hours. Hypochromicity occurs at 260 nm when a double helix is formed and at both 260 and 280 nm when a triple helix is formed. 51 At room temperature and 0.1 M NaC1, only the stoichiometry determines which helix will form. At high ionic strength (0.7 M NaCl), and at high temperature, even a 1 : 1 mixture will form a triple helix with time; triple-strand formation is also favored by Mg2+.51 Poly(I) and poly(C) form only a double helix, and hypochromicity is greatest at 250 nm. Since poly(I) can form helical structure by itself, this mixture of homopolymers is heated to 100° in a boiling water bath to melt the poly(I) structure, and the mixture is allowed to cool slowly to room 49B. D. Stollarand V. Raso, Nature (London) 250, 231 (1974). 50L. Rainen and B. D. Stollar,Biochemistry 16, 2003 (1977). 51M. Riley, B. Maling,and W. Chambedin,J. Mol. Biol. 20, 359 (1966).
82
PRINCIPLES AND METHODS
[3]
temperature. Mixtures with poly(G) or poly(dG) are also heated to disrupt homopolymer secondary structure. Poly(A).poly(dT) shows hypochromicity at 260 nm; a triple-helix can form with excess poly(dT) but only in solutions of high ionic strength. 5~ Poly(dA) and poly(U) form only a triple helix, with hypochromicity at all wavelengths from 220 to 285 nm. 5' For preparation of the aggregated polynucleotide-MBSA complex, it is convenient to use the polynucleotide at a concentration of about 200500/xg/ml. A stock solution of 10 mg of MBSA per milliliter in water is prepared; the powdered or lyophilized MBSA does not dissolve readily in saline. An amount of MBSA equal to the total weight of the polynucleotide is added to the annealed polymer. A white precipitate should be visible after the MBSA is added. It is an easily handled fine suspension with most polymers, but fibrous strands may be formed with samples of very high molecular weight. Preliminary mechanical homogenization or sonication usually reduces such polymers to a size that will form a manageable suspension with the MBSA. The polynucleotide-MBSA suspension is then emulsified with an equal volume of Freund's adjuvant until a stable thick white emulsion is formed. Complete adjuvant is used for the first immunization, and incomplete adjuvant for subsequent injections. Immunization Schedule A primary dose of 50-200/zg of polynucleotide, in the MBSA complex and emulsified with complete adjuvant, is injected into each rabbit at several intradermal sites along the back and subcutaneously; a total volume of 1 ml per rabbit is convenient. Similar doses, but with incomplete adjuvant, are given intradermally and subcutaneously on days 14 and 21, and the animals are bled 5 - 7 days later. They can be bled at weekly intervals, and additional booster immunizations can be given if antibody levels should fall. During early courses of immunization, they may peak and fall rapidly .39 Assay of Antibodies Sera can be tested by immunodiffusion or, with greater sensitivity and speed, by counterimmunoelectrophoresis (Fig. 3). In this assay, the negatively charged polynucleotide antigens move toward the anode and antibodies move toward the cathode during electrophoresis, and they meet and precipitate in less than an hour. About 7 ml of melted 0.8% agar in 50 mM Tris-HC1, pH 8, is placed on a 5 x 7.5 cm microscope slide. It is important to use agar, not agarose, since the mobility of the antibody depends on endosmotic flow. A trough for 150/zl of antiserum is placed opposite several wells; each well receives 50/.tl of antigen at a concentration of 5-10 ~g/ml. Electrophoresis, in 50 mM Tris-HC1, pH 8, buffer is
[3]
ANTIBODIES TO NUCLEIC ACIDS
83
FIG. 3. Counterimmunoelectrophoresis assay of anti-poly(A).poly(U) antiserum. Serum (150/zl) was placed in the trough, and 50 V.1containing 0.25 p.g of antigen was placed in each well. The antigen preparations were, from left to right, poly(A); poly(A) + poly(U), 85 : 15; poly(A) + poly(U), 67 : 33; poly(A) + poly(U), 50 : 50; poly(A) + poly(U) 33:67; poly(A) + poly(U), 10:90; and poly(U). Electrophoresis was run at 300 V, giving 10 mA per slide, for 45 min. The gel was 0.8% agar in 50 mM Tris-HC1 pH 8. run at 10 m A per slide for 3 0 - 4 5 min. With this assay one can quickly test reactivity with the annealed helical antigen, separate h o m o p o l y m e r s , and potentially cross-reacting p o l y m e r s . F o r m o r e quantitative assay, sera can be tested for binding of radiolabeled antigen; a double-antibody assay, using the y-globulin fraction of an anti-rabbit I g G serum, is preferred. Quantitative precipitation and quantitative m i c r o - c o m p l e m e n t fixation are also valuable assays for these antibodies. I n c r e a s i n g Specificity b y A b s o r p t i o n When antibodies to single-stranded p o l y m e r s or to cross-reacting helices are present, they m a y be r e m o v e d by absorption. A preliminary small-scale quantitative precipitation c u r v e (with 50 or 100 kd o f serum and 2 - 5 0 ~g of antigen per tube) will indicate the amount of p o l y m e r required for equivalence for a given volume of serum. A corresponding mixture is then p r e p a r e d on a larger scale and incubated for 1 hr at 37 ° and overnight at 4 °. The precipitate is r e m o v e d by centrifugation. Since excess free p o l y m e r m a y remain in the supernatant, the a b s o r b e d serum is
84
PRINCIPLES AND METHODS
[3]
passed through a small (2-5 ml) column of DEAE-cellulose equilibrated with 15 mM potassium phosphate, pH 7; antibody passes through directly, whereas polynucleotide binds to the column. Sera may be absorbed also by affinity columns, such as one on which polynucleotide is linked to CNBr-treated agarose. 52 Immunospecific Purification of Antibodies from Specific Precipitates Antibodies specific for helical structure can be purified by a procedure that does not require acid, alkali, or denaturants such as urea or guanidine. 53 Equivalence proportions of polynucleotide and serum (determined from a preliminary quantitative precipitin curve) are mixed and incubated at 37° for 1 hr and at 4° overnight. The precipit~ite, obtained by centrifugation, is washed with cold phosphate-buffered saline (0.14 M NaCl, l0 mM phosphate, pH 7) three times to remove extraneous serum proteins. It is then drained thoroughly, and excess buffer is wiped off the wall of the tube. The precipitate is resuspended in distilled water (in a volume close to that of the starting serum sample) and heated to 45-50 °. This denature s the helix, freeing the antibody. Pancreatic ribonuclease (or a mixture of pancreatic and T1 ribonucleases) is added to digest one or both of the homopolymer strands, and the mixture is incubated at 50° for 1 hr. It usually becomes turbid because the free immunoglobulin is poorly soluble in distilled water. One-tenth volume of 1.5 M NaCl is then added to dissolve the free antibody. Residual precipitate is removed, and the soluble material is applied to a Sephadex G-200 column. This separates purified IgG and IgM antibody populations from each other and from oligonucleotide fragments and residual ribonuclease. Immunospecific Purification with Affinity Adsorbants Guigues and Leng have described the preparation of polynucleotideSepharose columns for this purpose, with linkage of oligo- and polyribonucleotides to the gel through a 6-aminohexanoic acid spacer. ~4 For this linkage, Sepharose-aminohexanoic acid is added to 50 mg of nucleic acid in 50 ml of 0.1 M NaCl. Then 100 mg of a water-soluble carbodiimide is added, and the mixture is incubated at pH 5 for 2-5 hr. The substituted Sepharose is washed sequentially with high-salt solution, 2 M acetic acid, and then neutral buffer. (With purine-containing polydeoxyribonucleotides, the acid step should be avoided at this stage, since such polynucleotides are subject to depurination below pH 5.) Serum or its ~/-globulin 52M. S. Poonian,A. J. Schlabach,and A. Weissbach,Biochemistry 10, 424 (1971). Methods for preparing affinitycolumnswith nucleic acids have been reviewed by H. Potuzak and P. D. G. Dean, FEBS Lett. 88, 161 (1978). 53B. D. Stollarand V. Stollar, Virology 42, 276 (1970). M. Guiguesand M. Leng,Eur. J. Biochem. 69, 615 (1976).
[4]
ANTIGENIC H A P T E N - C A R R I E R CONJUGATES
85
fraction is applied to the column, and the column is washed with phosphate-buffered saline. Additional washing with a solution of 0.125 M borate, pH 8.5, 1.0 M NaCl, 0.1% Tween 20 removes nonspecifically bound protein. 55The antibodies are eluted with 2 M acetic acid in the cold and neutralized immediately. The immunospecifically purified antibodies have been particularly useful in physicochemical studies of the antibody-nucleic acid interaction; for some purposes, the monovalent Fab fragments of the purified antibodies have been used. 54,56 The use of absorbed and purified antibodies also ensures the specificity of immunofluorescent and serological measurements of helical nucleic acids of a given class in the presence of other nucleic acid forms. Acknowledgments Research in the author's laboratory has been supported by grants (currently grant PCM-79-04057) from the National Science Foundation and grant AI14534 from the National Institutes of Health. 5s j. A. Smith, J. G. R. Hurrell and S. J. Leach, Anal. Biochem. 87, 299 (1978). M. Leng, M. Guigues, and D. Genest, Biochemistry 17, 3215 (1978).
[4] T h e P r e p a r a t i o n o f A n t i g e n i c H a p t e n - C a r r i e r Conjugates: A Survey By BERNARD F . ERLANGER
Substances of molecular weight less than 1000 are not ordinarily antigenic. However, antibodies can be raised to small molecules by immunization with conjugates made up of low molecular weight substances (haptens) covalently linked to proteins or synthetic polypeptides. The ability to couple many different structures to macromolecules, the high degree of antigenicity of many of the conjugates, the development of sensitive methods of detecting and quantitating reactions between antibody and hapten, and the perfection of techniques for obtaining highly purified preparations of antihapten antibodies have contributed to the development of many of our modern immunological concepts. Much of our current knowledge of the requirements for immunogenicity, the structure of antigenic determinants, and the nature of antibody--its purification, hetMETHODS IN ENZYMOLOGY,VOL. 70
Copyright© 19~0by AcademicPress, Inc. All rightsof reproductionin any form reserved. ISBN 0-12-181970-1
[4]
ANTIGENIC H A P T E N - C A R R I E R CONJUGATES
85
fraction is applied to the column, and the column is washed with phosphate-buffered saline. Additional washing with a solution of 0.125 M borate, pH 8.5, 1.0 M NaCl, 0.1% Tween 20 removes nonspecifically bound protein. 55The antibodies are eluted with 2 M acetic acid in the cold and neutralized immediately. The immunospecifically purified antibodies have been particularly useful in physicochemical studies of the antibody-nucleic acid interaction; for some purposes, the monovalent Fab fragments of the purified antibodies have been used. 54,56 The use of absorbed and purified antibodies also ensures the specificity of immunofluorescent and serological measurements of helical nucleic acids of a given class in the presence of other nucleic acid forms. Acknowledgments Research in the author's laboratory has been supported by grants (currently grant PCM-79-04057) from the National Science Foundation and grant AI14534 from the National Institutes of Health. 5s j. A. Smith, J. G. R. Hurrell and S. J. Leach, Anal. Biochem. 87, 299 (1978). M. Leng, M. Guigues, and D. Genest, Biochemistry 17, 3215 (1978).
[4] T h e P r e p a r a t i o n o f A n t i g e n i c H a p t e n - C a r r i e r Conjugates: A Survey By BERNARD F . ERLANGER
Substances of molecular weight less than 1000 are not ordinarily antigenic. However, antibodies can be raised to small molecules by immunization with conjugates made up of low molecular weight substances (haptens) covalently linked to proteins or synthetic polypeptides. The ability to couple many different structures to macromolecules, the high degree of antigenicity of many of the conjugates, the development of sensitive methods of detecting and quantitating reactions between antibody and hapten, and the perfection of techniques for obtaining highly purified preparations of antihapten antibodies have contributed to the development of many of our modern immunological concepts. Much of our current knowledge of the requirements for immunogenicity, the structure of antigenic determinants, and the nature of antibody--its purification, hetMETHODS IN ENZYMOLOGY,VOL. 70
Copyright© 19~0by AcademicPress, Inc. All rightsof reproductionin any form reserved. ISBN 0-12-181970-1
86
PRINCIPLES AND METHODS
[4]
erogeneity, valence, size of combining site, and biological p r o p e r t i e s - has resulted from the use of such conjugates. The immunochemistry o f low molecular weight molecules had its beginning in the pioneering work of Landsteiner. In 1917, when Landsteiner set out to prepare what he called "artificial conjugated antigens," it was " t o investigate an almost dogmatic belief . . . that a special chemical constitution, peculiar to proteins, was required for the production of antibodies. ''1 We know better now, of course, but it is to these studies by Landsteiner that we owe much of what appears in this volume. The earliest of his conjugated proteins were prepared by the acylation of the amino groups o f serum albumin with chlorides or anhydrides of butyric, isobutyric, mono-, di-, and trichloroacetic, anisic, and cinnamic acids. This was followed by his better-known studies in which diazonium compounds were allowed to react with histidine, tyrosine, and tryptophan residues o f a protein. With these conjugates, he established that the original specificity of the protein carrier was changed by the newly introduced groups which, by themselves, were not antigenic, and that cross-reactions among sera depended now upon the structural relationships among the acyl or azo groups that were covalently linked to the protein. He also noted that, in most cases, antibody was produced to the protein carrier as well and, to be certain of antibodies to the new determinant group, one had to test the sera with conjugates made with an unrelated or homologous (to the immunized animal) protein. It was his practice to r e m o v e any of the anticarrier antibody by absorption of the serum with the free protein. This is still done, although it is not necessary for radioimmunoassays. Landsteiner also sought to determine the optimal n u m b e r of haptenic groups that gave the best antibody response, and he concluded that too much or too little hapten led to a poor response. With serum albumin as the carrier, 10 haptenic groups seemed to be optimal. The major finding by Landsteiner, however, related to the exquisite specificity of the antisera, as was so beautifully demonstrated by his classical studies with L-, O-, and meso-tartaric acids. Thus, Landsteiner's work established many of the ground rules by which we operate today. Our contributions since his time have been mainly refinement o f techniques and procedures and the expansion o f his ideas. The major exception to this statement, and a crucial one indeed, is the development by Berson and Yalow 2 of the technique o f radioimmunoi K. Landsteiner, "The Specificityof Serological Reactions" Harvard Univ. Press, Cambridge, Massachusetts, 1945. z S. A. Berson and R. S. Yalow, Radioimmunoassay: A status report, in "Immunobiology: Current Knowledge of Basic Concepts in Immunology and Their Clinical Application" (R. A. Good and D. W. Fisher, eds.), pp. 287-293. Sinauer, Stamford, Connecticut, 1971.
[4]
ANTIGENIC HAPTEN--CARRIER CONJUGATES
87
- A M I N O G R O U P S O F LYSINE RESIDUES (59)
×O CH2--CI-12--CH2--CH2--CH--C ~" N'H--
I
I
NH 2
NH
I
a-AMINO G R O U P S (1) /O -- CH--CI---NH
-
I NH 2
PHENOLIC HYDROXYL
G R O U P S O F TYROSINE RESIDUES (19) zO NH
I G R O U P S O F C Y S T E I N E RESIDUES (I)
SULFHYDRYL
×O CI-12--¢H--C"/ NH--
i
I
SH
NH
I I M I D A Z O L E G R O U P S O F HISTIDINE RESIDUES (/79 zO HC
C --CHz---
I
I
N.~c/NH H
CH--C'/---NH
-
l NH I
FIG. 1. Available functionalgroups in bovine serum albumin.
assay. It is this procedure that has led to the dramatic expansion of immunological techniques into the fields of biochemistry and pharmacology. In 1956 our laboratory, in collaboration with Beiser and Lieberman, became interested in preparing steroid-protein conjugates that were to be used to elicit antisteroid antibodies. An examination of the literature at that time showed that the azo coupling techniques of Landsteiner were still dominant. Like him, we chose to use the serum albumins because they were inexpensive and likely to yield soluble conjugates. However, an examination of the amino acid content of bovine serum albumin (BSA) (Fig. 1) convinced us that substitution by such relatively complex haptens as steriods should be attempted by reaction with the more plentiful Eamino groups of the lysine residues rather than by an azo coupling reaction with tyrosine, tryptophan, and imidazole residues. This meant forma-
88
PRINCIPLES AND METHODS
[4]
tion of amide bonds, for which a number of convenient new methods had been developed for the synthesis of peptides.Z A systematic approach was developed in which carboxylic acid groups were introduced into the haptens in various ways so that reaction with the amino groups of the protein carrier could be effected. Rather than deal with the steroid work separately, we will incorporate it into a general survey of the methods of preparing immunogenic haptenprotein conjugates in which the hapten is a pharmacologically interesting compound. The arrangement will be governed by the nature of the reactive functional groups of the hapten. In this way, the information can be applied most easily to new compounds being considered for use as determinant groups. No attempt will be made to present an exhaustive review of the literature. Instead, the various procedures described will be illustrated by specific examples to which the reader can refer for practical aspects of the experimental methods. Choice of Carrier The protein carriers used in various laboratories include globulin fractions, the serum albumins of various species, hemocyanin, ovalbumin, thyroglobulin, and fibrinogen. Hapten-protein conjugates of serum albumin are, in general, more soluble than conjugates of y-globulin or of ovalbumin. Thus, for example, steroid-protein conjugates of bovine, rabbit, and human serum albumin were soluble above pH 5.54"5; similar conjugates made with y-globulin and egg albumin frequently precipitated out of solution during preparation and could not be redissolved. Insoluble conjugates can be used for immunization, but subsequent characterization of the antibody then becomes a more difficult problem. Under certain circumstances, it may be advantageous to have both soluble and insoluble conjugates containing the same determinant group. The latter can be used for the isolation and purification of hapten-specific antibody. 6 For a review of insoluble hapten-carrier conjugates, we refer the reader to Jakoby and Wilchek r and Williams and Chase. 8 a j. p. Greenstein and M. Winitz, "'Chemistry of the Amino Acids," Vol.2. Wiley, New York, 1961. 4 B. F. Erlanger, F. Borek, S. M. Beiser, and S. Lieberman, J. Biol. Chem. 228, 713 (1957). 5 B. F. Erlanger, F. Borek, S. M. Beiser, and S. Lieberman, J. Biol. Chem. 234, 1090 (1959). 6 H. Szafran, S. M. Beiser, and B. F. Erlanger, J. lmmunol. 103, 1157 (1969). 7 W. B. Jakoby and M. Wilchek, eds., This series, Vol. 34. s C. A. Williams and M. W. Chase, eds., "Methods in Immunology and Immunochemistry," p. 335 et seq. Academic Press, New York, 1967.
[4]
ANTIGENIC HAPTEN-CARRIER CONJUGATES
89
W h e t h e r the choice o f c a r d e r significantly influences the antihapten response is a controversial subject. In the author's opinion, no really definitive study has been c a r d e d out. On the other hand, there are some who believe that K L H is a superior carrier2 Thyroglobulin is a choice of others. ~° As noted above, we chose the serum albumins, as have the great majority of the laboratories engaged in developing radioimmunoassays. With respect to the advantage of using a protein rather than a synthetic polypeptide as a carrier, we can refer to Jaffe e t a l . , 11 who showed that an active fragment ofgastrin, its C-terminal tetrapeptide amide, was antigenic when covalently attached to serum protein carriers but not when linked to poly(L-lysine) or to poly(L-glutamic acid). Walker e t al.lZ in 1973 made similar comparisons with steroid conjugates and obtained the same result, i.e., bovine serum albumin was a better carrier than poly(L-lysine). (But also see below. 42-44) Optimal E p i t o p e D e n s i t y Another important question concerns the optimal number o f haptens bound to the carrier protein (i.e., optimal epitope density). N i s w e n d e r and Midgley, 13 using steroid protein conjugates, suggested that at least 20 molecules of hapten should be covalently attached to a BSA carrier. Less than that results in an inferior antigen. Klause and Cross ~4 in studies on (DNP)nBSA conjugates obtained good responses with as few as five D N P groups, with excellent booster responses. Comparable responses were obtained with (DNP)19BSA. On the other hand, (DNP)50BSA and DNPh0BSA elicited an IgM response only; no change (i.e., boost) in titer occurred after 21 days, even with repeated immunization. It has been our experience that the nature o f the hapten exerts an influence, but that good antibody titers can usually be obtained with epitope densities anywhere between 8 and 25. On the other hand, we have not hesitated to immunize with conjugates with fewer haptenic groups (as few as two) if we were unable to prepare " b e t t e r " conjugates (for example, with expensive olig o n u c l e o t i d e - p r o t e i n conjugates). We have n e v e r failed to obtain a response, although on occasion we have had to wait longer for a suitable titer. 9 M. B. Rittenberg and A. A. Amkraut, J. lmmunol. 97, 421 (1966). 10F. Bartos, G. D. Olsen, R. N. Leger, and D. Bartos, Res. Commun. Chem. Pathol. Pharmacol. 16, 131 (1977). It B. M. Jaffe, W. T. Newton, and J. E. McGuigan, Immunochemistry. 7, 715 (1970). 12C. S. Walker, S. J. Clark, and H. H. Wotiz, Steroids 21,259 (1973). 13G. D. Niswender and A. R. Midgley,Jr., in "Immunological Methods in Steroid Determination" (F. G. Peron and B. F. Caldwell, eds.), Ch. 8. Appleton, New York, 1970. 1~G. G. B. Klause and H. M. Cross, Cell. lmmunol. 14, 226 (1974).
90
PRINCIPLES A N D M E T H O D S
[4]
Location of the Linkage Site Before we go on to describe the various chemical reactions used to prepare conjugates, another important consideration must be recognized: the point of attachment on the hapten. Landsteiner established in his early studies that antibody specificity is directed primarily at the part of the hapten molecule farthest removed from the functional group that is linked to the protein carrier. 1 Our experiences have been similar. For example, many steroids share a common ring A structure. In agreement with Landsteiner, it was found that anti-testosterone-3-BSA was better able to distinguish among closely related steroids than was anti-testosterone-17BSA. 15-17 Other similar studies exist in the literature, la'~s Even better specificity has been obtained with conjugates in which the attachment to the steroid is via a spacer joining a protein to a position on the haptenic molecule that is not important for its biological specificity, e.g., the C-6 position of an estrogen, ~a-z~ on the C-6 or C-11 of progesterone. TM An excellent investigation on the effect of the site of conjugation of corticosteroids is that of Nishina e t a l . 22 The antibody raised to these conjugates is thus specific for all the important structural features of the hapten. Preparation of the Conjugates Regardless of the protein carrier used, the same functional groups are available for attachment to the hapten: the carboxyl groups of the C terminal and of the aspartic and glutamic acid residues, the amino groups of the N terminal and the lysine residues, the imidazo and phenolic functions of the histidine and tyrosine residues, respectively, and the sulfhydryl group of cysteine residues (Fig. 1). All have been used for the preparation of immunogenic hapten-protein conjugates. Theoretically, the guanidino group of arginine is also available, but, to our knowledge, it has not been utilized in the preparation of conjugates. The functional groups of the hapten govern the selection of the method to be used to conjugate the hapten to the functional groups of the carder. The procedures described below, therefore, have been classified accordis S. M. Beiser and B. F. Erlanger, Nature (London) 214, 1044 (1967). 16 S. M. Beiser, B. F. Erlanger, F. J. Agate, and S. Lieberman, Science 129, 564 (1959). 17 S. Lieberman, B. F. Erlanger, S. M. Beiser, and F. J. Agate, Recent Prog. Horm. Res. 15, 165 (1959). 18 j. E. Buster and G. E. Abraham, Anal. Lett. 6, 147 (1973). 19 H. R. Lindner, E. Peril, A. Friedlander, and A. Zeitlin, Steroids 19, 357 (1972). zo D. Exley, M. W. Johnson, and P. D. G. Dean, Steroids 19, 605 (1971) 21 S. L. Jeffeoate and J. E. Searle, Steroids 19, 181 (1972). 23 T. Nishina, A. Tsuji, and D. Fukushima, Steroids 24, 861 (1974).
[4-]
ANTIGENIC HAPTEN--CARRIER CONJUGATES
91
ing to the functional group of the hapten utilized for conjugation. In this way, it is hoped that the information may be applied most easily to compounds being considered for use as determinant groups. H a p t e n s with Carboxyl Groups This class of haptens includes those that have a carboxyl group, such as acetylsalicylic acid (aspirin) or the peptides angiotensin and bradykinin. In addition, many haptens, such as some steroids, may have reactive groups to which a carboxyl group can be attached as, for example, by reaction with succinic anhydride (see below). For conjugation to proteins, the same procedures may be used regardless of whether the carboxyl group is present as an inherent part of the hapten or as an added moiety.
Mixed Anhydride Procedure This is a simple, direct procedure 3'2a that does not require the preparation and isolation of an active derivative. The coupling procedure is carfled out directly with the hapten, and the product usually contains 1~-25 hapten groups per molecule of albumin. As an example of this method, the coupling of cortisone-21-hemisuccinate to protein 4 is illustrated (Fig. 2). The haptenic group was converted in situ to an acid anhydride, which could then react in an aqueous-acetone solution with the amino groups of serum albumin. Uridine 5'-carboxylic acid, ~4 testosterone-17-hemisuccinate,4 monosuccinyl ecdysterone,~5 3-O-succinyldigitoxigenin,26 cholic acid) 7 thyroxine, 2a prostaglandins) a synthetic estrogens, 3° clonazepam-3-hemisuccinate, al and reserpine 32 are among the compounds linked in this way.
Carbodiimides This is another direct method that has been used extensively in preparing conjugated antigens. Uridine 5'-carboxylic acid was coupled to a mulza j. R. Vaughan, Jr. and R. L. Osato, J. Am. Chem. Soc. 74, 676 (1952). 24 M. H. Karol and S. W. Tanenbaum, Proc. Natl. Acad. Sci. U.S.A. 57, 713 (1967). 25 M. L. deRiggi, M. H. Him, and M. A. Delaage, Biochem. Biophys. Res. Commun. 66, 1307 (1975). 2e G. C. Oliver, Jr., B. M. Parker, D. L. Brasfield, and C. W. Parker, J. Clin. Invest. 47, 1035 (1968). 27 G. J. Beckett, N. M. Hunter, and W. P.-R. Iain, Clin. Chim. Acta 88, 257 (1978). 2s W. H. Churchill and D. F. Tapley, Nature (London) 202, 29 (1964). 39 B. M. Jaffe, J. W. Smith, W. T. Newton, and C. W. Parker, Science 171, 494 (1971). 30 R. J. Warren and K. Fotherby, J. Endocrinol. 62, 605 (1974). 31 W. R. Dixon, R. L. Young, R. Ning, and A. J. Liebman, J. Pharm. Sci. 66, 235 (1977). 33 A. Levy, K. Kawashima, and S. Spector, Life Sci. 19, 1421 (1976).
92
PRINCIPLES AND METHODS 0
0
H
CHz" 0 "C'(CHz)2"COOH
t
C=O • OH
0///,,,~
[4]
It
CHz .OC. (CH2}2,,CO-O.CO.OR
I
o
II
RO.C.Cl,
C=O OH
pH9-9.5 / PROTEIN.NHz ~ Hz'OC'ICHz)z* CO, NH ) C" 0
"-~'OH /
PROTEIN
+ CO2 -h ROH
FIG. 2. Preparation of cortisone-21-hemisuccinate conjugate by mixed anhydride procedures.
tichain polypeptide, poly(DL-alanyl)-poly(L-lysine) with dicyclohexylcarbodiimide in a 95% dimethylformamide medium as solvent, a3 Coupling reactions can be carried out in aqueous solution by use of the water-soluble carbodiimides, such as 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide • HCI or 1-cyclohexyl-3-[2-morpholinyl-(4)-ethyl]carbodiimide methop-toluenesulfonate,a4,a5 both commercially available reagents. Angiotensin and bradykinin, both small polypeptides (MW approximately 1000), were coupled to proteins in the first utilization of this procedure, a" The authors believed that the reaction was between the N-terminal amino group of the peptide and the protein, but provided no evidence for this. On the other hand, they later used similar techniques to couple angiotensin to polylysine with the water-soluble reagent N-ethylbenzisoxazole,37 a reaction that is possible only if the carboxyl group of angiotensin participates. One case in which it was definitely established that carbodiimides activate the carboxyl group of the peptide relates to the production of antibody to gastrin tetrapeptide. 11 In this case, the amino end of the peptide was blocked by a t e r t - b u t y l o x y c a r b o n y l (t-BOC) group. The t-BOC group was subsequently removed with trifluoroacetic acid. Carbodiimides were also used by Dietrich 3a and by Haber et al. a9 Additional components of a3 M. Sela, H. Ungar-Waron, and Y. Schechter, Proc. Natl. Acad. Sci. U.S.A. 52, 285 (1%4). J, C. Sheehan, P. A. Cruickshank, and G. L. Boshart, J. Org. Chem. 26, 2525 (1%1). a5 j. C. Sheehan and J. S. Hlavka, J. Org. Chem. 21, 439 (1956). ae T. L. Goodfriend, L. Levine, and G. Fasman, Science 143, 1344 (1964). a7 T. L. Goodfriend, G. Fasman, D. Kemp, and L. Levine, lmmunochemistry 3, 223 (1966). as F. M. Dietrich, Immunochemistry 4, 65 (1%7). aa E. Haber, L. B. Page, and G. A. Jacoby, Biochemistry 4, 693 (1965).
[4]
ANTIGENIC H A P T E N - C A R R I E R CONJUGATES
93
biological interest that were coupled to carriers with water-soluble carbodiimides include gastrin, 4° adenosine Y,5'-cyclic phosphate, 41 morphine, 42 lysergic acid diethylamide, 43 and prostaglandin. 44,45 In three cases, 4z-44 the carrier used was polylysine and immunization was done with a complex of the conjugate and succinylated hemocyanin. This improved technique minimized the extent of the immunological response to the carrier portion of the immunogen. Tobramycin, 46 1-/3-E-arabinofuranosylcytosine,47 cocaine metabolites, 4a prednisone-21-hemisuccinate, 4a synthetic narcotic analgesic drugs, 50 DL-methadol-hemisuccinate, 1° ochratoxin A, 51 digitoxigenin, 26 and gentamicin 52 are among additional compounds linked to proteins by means of carbodiimides. Digitoxigenin 26 was linked to a carrier protein, by both the carbodiimide and the mixed anhydride procedures, in the same laboratory. The mixed anhydride yielded a conjugate with 13 haptenic groups compared with 5 groups via the carbodiimide procedure. Specific antibodies to a-melanotropin were obtained with carbodiimide using the peptide with its internal lysine blocked by an ~-methylsulfonylethoxycarbonyl group. Linkage to the amino groups of a protein then occured via the single glutamic acid residue remaining in the peptide, after which the protecting group was removed by 2 N Na2CO3. 53 If the haptenic molecule is not soluble in water, it can be dissolved in water-miscible solvents, such as dimethylformamide, and added to the aqueous protein solution. 45 40 j. D. Young, D. J. Byrnes, D. J. Chisholm, F. B. Griffiths, and L. Lazarus, J. Nucl. Med. 10, 746 (1969). 41 A. L. Steiner, D. M. Kipnis, R. Utiger, and C. Parker, Proc. Natl. Acad. Sci. U.S.A. 64, 367 (1969). 42 H. Van Vunakis, E. Wasserman, and L. Levine, J. Phmmacol. Exp. Ther. 180, 514 (1972). 43 H. Van Vunakis, J. T. Farrow, H. B. Gjika, and L. Levine, Proc. Natl. Acad. Sci. U.S.A. 68, 1483 (1971). 44 L. Levine and H. Van Vunakis, Biochem. Biophys. Res. Commun. 41, 1171 (1970). 45 F. A. Fitzpatrick and G. L. Bundy, Proc. Natl. Acad. Sci. U.S.A. 75, 2689 (1978). A. Broughton, J. E. Strong, L. K. Picketing, and G. P. Brodey, Antimicrob. Agents Chemother. 10, 652 (1976). 47 T. Okabayashi, S. Mihara, D. B. Repke, and J. G. Moffat, Cancer. Res. 37, 619 (1977). 4a B. Kaul, S. J. Millian, and B. J. Davidow, Pharmacol. Exp. Ther. 199, 171 (1976). 49 A. W. Meikle, J. A. Weed, and F. H. J. Tyler, J. Clin. Endocrinol. Metab. 41,717 (1975). 5o H. Van Vunakis, D. S. Freeman, and H. B. Gjika, Res. Commun. Chem. Pathol. Pharmacol. 2, 379 (1975). 51 O. Aalund, K. Brunfeldt, B. Hald, P. Krogh, and K. Poulsen, Acta Pathol. Microbiol. Scand. C 83, 390 (1975). 52 j. E. Lewis, J. C. Nelson, and H. A. Elder, Nature (London) 94, 214 (1972). H. G. Kopp, A. Ebede, P. Vitius, W. Lichensteiger, and R. Schwyzer, Fur. J. Biochem. 75, 417 (1977).
94
PRINCIPLES AND METHODS + 0 NR' 0 NR' O NHR' II II II II H+ II II RC-OH + C.--~RC-O-C ~ RC-O-C INIR' INHR' /HR, O ,, R~-NR'
I C:O
[4]
R"NHz ropid
>
R"NHz very slow
0II H L RC--NR" O+
R' H II __ N HR' N--C
NIl.iR' Fio. 3. Mechanism of carbodiimide-mediated preparation of amides.
The conditions of the reaction are very simple. The carder, an excess of hapten and the reagent are simply stirred together in an aqueous solution for 30 min to several days depending upon the procedure. The reaction is followed by dialysis, and the product is isolated by lyophilization. The reaction mechanism is as shown in Fig. 3. There are two possible pathways, the desired one being catalyzed by H ÷. The protein carrier, however, is most reactive at higher pH, where dissociation of the lysine ammonium groups occurs. A compromise is therefore necessary to provide the most favorable conditions; a pH near 6 is usually chosen. In our experience, the use of water-soluble carbodiimides has not always been successful. On occasion, extensive alteration of the carder has occurred with little if any substitution by haptenic groups. It is possible to be led astray because antibody is produced to the altered protein, and this antibody does not react with the protein in its original state. Nevertheless, it is a generally efficacious method of preparing conjugates. It is of interest that water-soluble carbodiimides have been used to couple nucleotides directly to proteins, presumably by formation of a P-N bond.54-s8
Miscellaneous Carboxyl Methods An aspirin-protein conjugate was prepared by first converting aspirin (acetylsalicylic acid) to the acetylsalicylazide.59 The azide was coupled to rabbit serum globulin in a 1:1 dioxane-water solution maintained alkaline to phenolphthalein by the addition of base. About 25 to 35 haptenic M. J. Halloran and C. W. Parker, J. lmmunol. 96, 373 (1966). 55 M. J. Halloran and C. W. Parker, J. lmrnunol. 96, 379 (1966). se M. Z. Humayun and T. M. Jacob, Biochim. Biophys. Acta 331, 41 (1973). 5T S. A. Khan, M. Z. Humayun, and T. M. Jacob, Nucl. Acids Res. 4, 2997 (1977). 5s S. A. Khan and T. M. Jacob, Nucl. Acids Res. 4, 3007 (1977). 5a G. C. Butler, C. R. Harington, and M. E. Yuill, Biochem. J. 34, 838 (1940).
[4]
ANTIGENIC HAPTEN--CARRIER CONJUGATES
95
o
O RC" OH+HO o
II H
DCC~ R 'O
R'NH2 RC-NR'
o
o
\\
FIG. 4. Preparation of amides using N-hydroxysuccinimide. DCC, dicyclohexyicarbodiimide. groups were conjugated per molecule of globulin. A similar procedure was used for thyroxine.e° Antibodies specific for thyroxine have also been obtained by using, as antigen, tetraiodothyropropionic acid coupled to protein by the mixed anhydride method. 28 The conversion of aspirin to an acid chloride that can react directly with protein was also reported. 61 The insect juvenile hormone DL-10,11-epoxyfarnesoic acid was coupled to protein by a procedure that should find extensive use. 6~,6s We had been unable to effect the reaction with water-soluble carbodiimides, obtaining only unsubstituted, altered protein (see above). The N-hydroxysuccinimide ester was prepared by reaction o f the juvenile h o r m o n e with N-hydroxysuccinimide (commercially available) in the presence of dicyclohexylcarbodiimide (Fig 4). N - H y d r o x y s u c c i n i m i d e esters are used in peptide synthesis. 64 They are quite stable if kept dry but react quickly and in good yield with amino groups to form amide or peptide bonds. Conjugates containing 20 juvenile hormone groups were used to raise specific antibodies in rabbits. Antibodies to e c d y s o n e were made similarly. 6~ Carbonyldiimidazole is another commercially available reagent that has been used to link haptens to proteins by means o f amide bonds, for example in the preparation o f a BSA conjugate of fluoxymesterone. 66 R. F. Clutton, C. R. Harrington, and M. E. Yuill, Biochem. J. 32, 1119 (1938). el L. M. Weiner, M. Rosenblatt, and H. Howes, J. lmmunol. 90, 788 (1963). 62 R. C. Lauer, P. Soloman, K. Nakanishi, and B. F. Erlanger, Fed. Proc. 32, 500 (1972). sa R. C. Lauer, P. H. Soloman, K. Nakanishi, and B. F. Erlanger, Experientia 30, 558 (1974). 84G. W. Anderson, J. E. Zimmerman, and F. M. Callahan, J. Am. Chem. Soc. 86, 1839 (1964). e5 R. C. Lauer, P. H. Soioman, K. Nakanishi, and B.F. Erlanger, Experientia 30, 560 (1974). 66W. A. Colburn, Steroids 25, 43 (1975). so
96
PRINCIPLES AND METHODS
[4]
H a p t e n s with Amino Groups Two classes of haptens with available amino groups must be conside r e d - t h e aromatic amines and the aliphatic amines.
Aromatic Amines Much of Landsteiner's pioneer work 1 was carried out with haptens that were aromatic amines. The compounds were converted to diazonium salts with nitrous acid and allowed to react with proteins at alkaline pH (approximately 9). Reaction occurred primarily with histidine, tyrosine, and tryptophan residues of the protein carrier. For a representative procedure, see Kabat 67 (p. 799 et seq.). An interesting application of this procedure was the preparation of a chloramphenicol-protein conjugate which was used to elicit antibodies specific for chloramphenicol. °s In this case, a prior reduction of the nitro group of chloramphenicol to an amino group was required. As early as 1937, carcinogenic compounds were conjugated to protein carriers by means of their isocyanate derivatives which were prepared from amines. 69 Immune sera were raised, and their properties were studied. 69,7°
Aliphatic Amines Aliphatic amines can be caused to react with proteins by using watersoluble carbodiimides. Examples include bradykinin and angiotensin, 36 tobramycin, 46 gentamicin, 53 adriamycin, 71 5-hydroxytryptamine (serotonin), 72 cortisol-21-amine, 7z and spermidine, r4 Aliphatic amines can be converted to a p-nitrobenzoylamide by reaction with p-nitrobenzoyl chloride. The amide derivative can then be re-, duced to a p-aminobenzoyl derivative which can be coupled to proteins by diazotization, as described above. Among the haptens conjugated this 67 E. A. Kabat, "Kabat and Mayer's Experimental Immunochemistry," 2nd ed. Thomas, Springfield, Illinois, 1961. 68 R. N. Hamburger, Science 152, 203 (1966). ~ H. J. Creech, Cancer Res. 12, 557 (1952). 70 H. J. Creech and W. R. Franks, Am. J. Cancer 30, 555 (1937). 71 y . -H. Chien and L. Levine, lmmunochemistry 12, 291 (1975). 72 B. Peskar and S. Spector, Science 179, 1340 (1973). 73 y. Kobayashi, T. Ogihara, K. Amitani, F. Watanabe, T. Kigushi, I. Ninomiya, and Y. Kumahara, Steroids 32, 137 (1978). 74 F. Bartos, D. Bartos, A. M. Dolney, D. P. Grettie, and R. A. Campbell, Res. Commun. Chem. Pathol. Pharmacol. 19, 295 (1978).
[4]
ANTIGENIC HAPTEN--CARRIER CONJUGATES
97
way are a series of C-terminal peptide sequences of the tobacco mosaic virus protein 75-77 and angiotensin, r8"79 Angiotensin has also been attached by its N-terminal amino group to the amino groups of a carrier by means of the bifunctional reagent m-xylylene diisocyanate,aa Tolylene 2,4-diisocyanate has been used in a similar manner to prepare bradykinin conjugates, s° Haptens containing amino groups have also been covalently linked to amino groups of protein carriers with glutaraldehyde. Among the haptenic groups conjugated in this manner are adrenocorticotropic hormone (ACTH), sl glucagon, 82 and normetanephrine.83 A novel procedure has been used to link nortryptyline to BSA. 84 Its aliphatic secondary amine was converted to a succinamic acid derivative that was caused to react with the protein by using a water-soluble carbodiimide. The antibody was used to assay for various tricyclic antidepressants, including imipramine. In another novel procedure involving nortryptyline, the secondary amine was allowed to react with N-(4-bromobutyl)phthalimide. After removal of the phthalimido group, the resulting primary aliphatic amine was caused to react with carboxyl groups on BSA by using a water-soluble carbodiimide. 85"86 H a p t e n s with Available Hydroxyl Groups This class of haptens includes alcohols, phenols, sugars, polysaccharides, and nucleosides. In most cases, derivatives of this class of compounds must be made in order to introduce functional groups capable of reacting with proteins. Hemisuccinates
A simple procedure, first introduced in our work with steroid-protein conjugates, is the conversion of the alcohol to the half ester of succinic 75 F. A. Anderer, Biochim. Biophys. Acta 71, 246 (1963). 76 F. A. Anderer and H. D. Schlumberger, Biochim. Biophys. Acta 115, 222 (1966). 77 F. A. Anderer and H. D. Schlumberger, Biochim. Biophys. Acta 97, 503 (1965). 78 S. D. Deodhar, J. Exp. Med. 111,419 (1960). 79 S. D. Deodhar, J. Exp. Med. 111, 429 (1960). s o j. Spragg, K. F. Austen, and E. Haber, J. Immunol. 96, 865 (1966). s~ M. Reichlin, J. J. Schnure, and V. K. Vance, Proc. Soc. Exp. Biol. Med. 128, 347 (1968). s2 L. A. Frohman, H. Reichlin, and J. E. Sokal, Endocrinology 87, 1055 (1970). 83 B. A. Peskar, B. M. Peskar, and L. Levine, Eur. J. Biochem. 26, 191 (1972). 84 D. J. Brunswick, B. Needleman, and J. Mendels, Life Sci. 22, 137 (1978). s5 G. W. Aherne, E. M. Piall, and V. Marks, Br. J. Clin. Pharmacol. 3, 561 (1976). s 6 K. P. Maguire, G. D. Burrows, T. R. Norman, and B. A. Scoggins, Clin. Chem. 24, 549 (1978).
98
PRINCIPLES AND METHODS
[4]
acid (i.e., the hemisuccinate). The hemisuccinate has an available carboxyl group that can be made to react by any of the procedures described above. Conversion to the hemisuccinate requires a reaction with succinic anhydride in pyridine. A representative procedure can be found in the papers on steroid-protein conjugates. 4~ Another example is the preparation of the hemisuccinate of cyclic adenosine monophosphate (cAMP). 41 More recently, in an excellent study of the specificity of antiestrogen antibodies, s7 1la-hydroxyhemisuccinates of estrone and estradiol were prepared and linked to BSA by the mixed anhydride technique. Other examples include hemisuccinates of fl-dl-methadol, ~° 3-hydroxyclonazepam,31 ecdysterone, 25 propanolol, 88 Aa-tetrahydrocannabinol,sa and 1-fl-D-arabinofuranosylcytosine.47
Chlorocarbonates Another alternative is the reaction of the determinant group with an equimolar quantity of phosgene to yield the highly reactive chlorocarbonate, which reacts directly with the amino groups of the protein in the presence of bicarbonate. An example is the conversion of testosterone to testosterone- 17-chlorocarbonate.4
Aminophenyl Derivatives Phenols can be converted to active reagents by reaction with diazotized p-aminobenzoic acid. In this way, a carboxyl group is introduced into the molecule. This type of reaction was carried out successfully with 17fl-estradiol. 9° More recently, a similar procedure was used to make conjugates of AT-tetrahydrocannabino189 and reserpine, zz The classical procedure for the coupling of sugars to proteins involves the formation ofp-nitrophenylglycosides, the conversion of the latter by hydrogenation to p-aminophenylglycoside, and then attachment to the protein by diazotization. This method was used by Landsteiner 1 for a number of preparations. A variant of this method, used by Goebel 9L9zand Goebel and Hotchkiss, 93 was conversion to the aminobenzyl ether followed by diazotization. F. C. den Hollander, B. K. van Weemen, and G. F. Woods, Steroids 23, 549 (1974). K. Kawashima, A. Levy, and S. Spector, J. Pharmacol. Exp. Ther. 196, 517 (1976). s a p . T. Tsui, K. A. Kelly, M. M. Ponpipon, and A. H. Sehon, Can. J. Biochem. 52, 252 (1974). 9o N. Weliky and H. H. Weetall, Immunochemistry 2, 293 (1965). 91 W. F. Goebel, J. Exp. Med. 64, 29 (1936). 92 W. F. Goebel, J. Exp. Med. 68, 469 (1938). 93 W. F. Goebel and R. D. Hotchkiss, J. Exp. Med. 66, 191 (1937). s7 ss
[4]
ANTIGENIC HAPTEN--CARRIER CONJUGATES
p
IOZ U
II
OH OH
0
pH8.5-9.2
~'or BHaCN-
\N/
99
0
I ProtoNHz
HO-'CH HC-OH
I
\N/ I
Prot
Prot P= Purine or Pyrimidine R = H ,- PO3H2 or 5'- nucleotide
Prot*NH2= Carrier protein, NHz groups
FIG. 5. Preparation of conjugates using periodate procedure.
Oxidation to Dialdehydes
A relatively simple procedure developed for the preparation of nucleoside- and nucleotide-protein conjugate ~-98 makes use of the reaction of vicinal hydroxyl groups with periodate to yield dialdehydes (Fig. 5). The dialdehydes, without isolation, are caused to react with the amino groups of protein at pH 9.5 in aqueous solution to yield aldimines, which are stabilized by reduction with sodium borohydride. Only the final conjugate is isolated in this procedure, which is simple to run and yields conjugates with as many as 30 determinant groups per molecule of albumin. It should be applicable to all compounds with vicinal hydroxyl groups, such as glycols, glycerol derivatives, and glycosides, and has been used successfully for the preparation of digoxin-protein conjugates 99 and ouabain.1°° Modi-
~" B. F. Erlanger and S. M. Beiser, Proc. Natl. Acad. Sci. U.S.A. 52, 68 (1964). 95 B. F. Erlanger, D. Senitzer, O. J. Miller, and S. M. Beiser, Acta Endocrinol. (Copenhagen) Suppl. 168, 206 (1972). R. M. D'Alisa and B. F. Erlanger, Biochemistry 13, 3575 (1974). 97 R. M. D'Alisa and B. F. Erlanger, J. Immunol. 116, 1629 (1976). S. M. Beiser, S. W. Tanenbaum, and B. F. Erlanger, this series, Vol. 12B, p. 889. V. P. Butler and J. P. Chen, Proc. Natl. Acad. Sci. U.S.A. 57, 71 (1967). 1oo T. W. Smith, J. Clin. Invest. 51, 1583 (1972).
100
PRINCIPLES AND M E T H O D S
[4]
fications have been made to prepare conjugates with alkaline-sensitive nucleosides and nucleotides. 1°"°3 In a recent procedure to detect carcinogen-DNA adducts by radioimmunoassay, 1°4 N-(guanosin-8-yl)acetylaminofluorine was linked to BSA by the periodate method, and the conjugate was used to elicit specific antibodies to this product, which is formed when N-acetoxy-2-acetylaminofluorene reacts with DNA. Oxidation to Carboxyl Oxidation of the 5'-hydroxyl groups of uridine, z4"1°5 pseudouridine, 24 and other nucleosides 33 has made it possible to conjugate these compounds to proteins by methods amenable for the reaction of carboxylic acid derivatives with proteins. Miscellaneous Hydroxyl Methods Another method of seemingly general applicability to carbohydrates was used by Coat et al. 106to conjugate uridine to proteins. The isopropylidine derivative was allowed to react with p-nitrobenzoyl chloride to yield the 5' ester. Removal of the isopropylidine protecting group and hydrogenation of the nitro group made it possible to link the uridine derivative to the protein by a diazotization reaction. Some rather novel chemistry was described in two recent very interesting papers describing a radioimmunoassay procedure for ADP-ribose. 1°7 Adenine-N6-carboxymethylated NAD was prepared and converted to N6-carboxymethylated ADP-ribose by NAD glycohydrolase. N6-Carboxymethylated ADP-ribose was then linked to BSA using watersoluble carbodiimide. The bifunctional reagent sebacoyl dichloride has been used to convert alcohols to acid chlorides, which, at pH 8.5, react readily with proteins. This procedure was used by Bailey and Butler l°s to prepare a cholesterolprotein conjugate. ~o~ L. Rainen and B. D. Stollar, Nucl. Acids Res. 4, 4877 (1978). lo2 R. D. Meredith and E. F. Erlanger, Fed. Proc. 37 (6), 1503 (1978). 1o3 j. Wollack, Fed. Proc. 37 (6), 1632 (1978). ~o4 M. C. Poirer, S. H. Yuspa, I. B. Weinstein, and S. Blobstein, Nature (London) 270, 186 (1977). ~05 M. Sela and H. Ungar-Waron, Fed. Proc. 24, 1438 (1965). ~0~ j. p. Coat, S. David, and J. C. Fischer, Bull. Soc. Chim. Ft. 21, 2489 (1965). ~07 R. Bredehorst, A. M. Ferro, and H. Hilz, Fur. J. Biochem. 82, 105 (1978). los j. M. Bailey and J. Butler, in "'The Reticuloendothelial System and Atherosclerosis" (N. R. DiLuzio and R. Paoletti, eds.), pp. 433-441. Plenum, New York, 1967.
[4]
ANTIGENIC HAPTEN--CARRIER CONJUGATES
101
H a p t e n s with C a r b o n y l G r o u p s K e t o n e s and aldehydes can be used as haptenic determinant groups by converting them to O - ( c a r b o x y m e t h y l ) oximes. This is done by reacting t h e m with O - ( c a r b o x y m e t h y l ) h y d r o x y l a m i n e ( N H 2 O C H 2 C O O H , sold c o m m e r c i a l l y as c a r b o x y l m e t h o x y l a m i n e or a m i n o o x y a c e t i c acid). This serves to introduce a carboxyl group, which is exploited as described above. E x a m p l e s o f this m e t h o d o l o g y can be found in the coupling of testosterone-3-(O-carboxymethyl) oxime, estrone-17-(O-carboxymethyl) oxime, and progesterone-20-(carboxymethyl) oxime to bovine serum albumin with the mixed anhydride technique. 4,5 Prepared in a similar manner were the 3-(O-carboxymethyl) oxime derivative of m e d r o x y p r o g e s terone acetate, l°a the 3 - c a r b o x y m e t h y l oxime of aldosterone-18-21-diacetate. 11° Similarly, O - c a r b o x y m e t h y l derivatives were prepared from the synthetic progestogens norethisterone and norgestrel. H1 5 a - D i h y d r o t e s t o s t e r o n e - l l - ( O - c a r b o x y m e t h y l ) oxime was synthesized H2 in an elegant multistep p r o c e d u r e that included a microbiological reduction and a selective hydrolysis of a dioxime. The final p r o d u c t was p r e p a r e d by reaction of 17fl-hydroxy-5a-androstane-3, 11-dione-11-oxime with sodium chloroacetate to give the O - c a r b o x y m e t h y l oxime. The ketone groups of aldosterone, corticosterone, and cortisol were derivatized with p - h y d r a z i n o b e n z o i c acid. ~13 The resulting carboxylic acid derivatives could be linked to BSA with water-soluble carbodiimide. Aldehydes can be conjugated to proteins directly by Schiff base formation followed by stabilization of the bond by reduction with sodium borohydride. Pyridoxal and pyridoxal p h o s p h a t e are examples of haptens conjugated in this manner. ~4'H5 Other Reactions Penicillenic acid was conjugated to protein by an interesting p r o c e d u r e that included modification o f the protein carrier, u6 Penicillenic acid has a reactive sulfhydryl group capable of forming disulfide bonds with other lo9 M. Hiroi, F. Z. Stanczyk, U. Goebelsmann, P. F. Brenner, M. E. Lumkin, and D. R. Mishele, Jr., Steroids 26, 373 (1975). 110C. A. Bizoleon, J. -F. Riviere, P. Franchimont, A. Faure, and B. Claustrat, Steroids 23, 809 (1974). 111R. J. Warren and K. Fotherby, J. Endocrinol. 62, 605 (1974). 112T. S. Baker and D. Exley, Steroids 29, 429 (1977). 113B. Africa and E. Haber, lmmunochemistry 8, 479 (1971). 114F. Cordoba, C. Gonzalez, and P. Rivera, Biochim. Biophys. Acta 127, 151 (1966). 115R. Ungar-Waron and M. Sela, Biochim. Biophys. Acta 124, 147 (1966). ,e A. L. deWeck and H. N. Eisen, J. Exp. Med. 112, 1227 (1960).
102
PRINCIPLES AND METHODS
[4]
sulfhydryl groups. The carrier proteins (e.g., human y-globulin or bovine y-globulin) were artificially enriched in sulfhydryl groups by reaction with N-acetylhomocysteine thiolactone. 117 The coupling reaction with an excess of penicillenic acid was then carried out in acetate buffer at pH 4 in the presence of H202. Antibody to progesterone was also obtained by immunization with conjugates prepared from thiolated proteins.H.8 Bovine serum albumin was thiolated by reaction with S-acetylmercaptosuccinic anhydride. After deacetylation, coupling was achieved with 6fl-bromoprogesterone. Bis-diazotized benzidine can be used as a bridging reagent between proteins and haptens containing aromatic groups that react with diazonium compounds. A conjugate of thyrotropin-releasing hormone (which contains a reactive histidine residue) was obtained in this way.~9 A novel approach has been to allow serotonin to react with protein via the Mannich reaction? 2° This is a simple reaction that enables one to use formaldehyde as a bridge between the amino groups of a protein and compounds containing one or more reactive hydrogens. The Mannich reaction has also been used to prepare reserpine conjugates, s2 The antibody titers were not as satisfactory as those elicited by conjugates prepared by a pcarboxyazobenzene derivative linked to BSA by the mixed anhydride procedure. Among the low molecular weight haptens that have been used as determinant groups are substances that are reactive enough to be coupled to proteins directly. Dinitrofluorobenzene has been used to prepare antigens for the stimulation of antidinitrophenyl antibodies. These have been very useful in studies of the binding characteristics and structure of immunoglobulins. Antibodies that react with deoxyribonucleic acid (DNA) have been elicited by immunization with the product of the reaction of 6-trichloromethylpurinex21-~3 with BSA. Nucleotide protein conjugates have also been made by using carbodiimide to link the nucleotide to the protein. ~6-~s Antipenicillin antibodies have been produced by immunization 117 R. Benesch and R. E. Benesch, in "Sulfur in Proteins" (R. Benesch, R. E. Benesch, P. D. Boyer, I. M. Kiotz, W. R. Middlebrook, A. G. Szent-Gyrrgyi, and D. R. Schwartz, eds.), pp. 15-24. Academic Press, New York, 1959. 11s C. N. Pang and D. C. Johnson, Steroids 23, 203 (1974). H9 R. M. Bassiri and R. D. Utiger, Endocrinology 90, 722 (1972). 12o N. S. Ranadive and A. H. Sehon, Can. J. Biochem. 48, 1701 (1967). 121 S. Cohen, E. Thom, and A. Bendich, Biochemistry 2, 176 (1963). 1~ S. Cohen, E. Thom, and A. Bendich, J. Org. Chem. 27, 3545 (1962). l~a V. P. Butler, Jr., S. M. Beiser, B. F. Erlanger, S. W. Tanenbaum, S. Cohen, and A. Bendich, Proc. Natl. Acad. Sci. U.S.A. 48, 1597 (1962).
[4]
ANTIGENIC HAPTEN--CARRIER CONJUGATES
103
with penicillin-protein conjugates. The latter were prepared by the reaction o f penicillin with protein under slightly alkaline conditions. 124-12e A t e t r a h y d r o c a n n a b i n o l - B S A conjugate has been made by reaction with 10-iodotetrahydrocannabinol-9-isocyanate. 89 L-Phenylalanine mustard is another example of a reactive hapten,127 C h a r a c t e r i z a t i o n of t h e C o n j u g a t e s Generally, the haptenic group has an absorbance spectrum that can allow one to differentiate it from the protein carder. This is particularly true for azo derivatives, which absorb in the visible range. H o w e v e r , even if there is overlap in the two spectra, reasonably accurate determinations of the number of haptenic molecules per carrier protein can be determined from difference spectra. A more convenient and direct procedure, introduced by Abraham e t a l . , 128 is the incorporation o f some radioactive hapten in the conjugation procedure. A direct estimation of the extent o f substitution can be made by counting undialyzable radioactive material. A procedure introduced by us for s t e r o i d - p r o t e i n conjugates 4 was the estimation of the remaining free amino groups with the dinitrophenylation technique of Sanger.129 e-Dinitrophenyllysine was not isolated but was estimated directly by spectrophotometry after ether extraction of the acid hydrolyzate. A control with unsubstituted c a r d e r was always run concomitantly, and the difference between the two was taken to be the extent o f substitution by hapten. We have also used a procedure of H a b e e b 13° in which trinitrobenzene sulfonic acid is used as the reagent for estimation of free amino groups in the conjugate. Spectrophotometric comparison of theintactprotein conjugate and the original carrier protein is possible; no acid hydrolysis is required. We found this procedure to be convenient and entirely satisfactory, an example being the case of insect juvenile h o r m o n e - p r o t e i n conjugates. 62,63,6~
n4 B. B. Levine, N. Engl. J. Med. 275, 1115 (1966). 125B. B. Levine and Z. Ovary, J. Exp. Med. 114, 875 (1961). ~z6H. Smith, J. M. Dewdney, and A. W. Wheeler, Immunology 21, 527 (1971). 127j. F. Burke, V. H. Mark, A. H. Soloway, and S. Leskowitz, CancerRes. 26, 1893(1966). ~2sG. E. Abraham, R. Swerdloff, D. Tulchinsky, and W. D. Odell, J. Clin. Endocrinol. Metab. 32, 619 (1971). ~29F. Sanger, Biochem. J. 45, 563 (1949). lao A. F. S. A. Habeeb, Anal. Biochem. 14, 328 (1966).
104
PRINCIPLES
AND
METHODS
[5]
General Comments It was not the purpose of this review to present a comprehensive survey of hapten-protein conjugates, but rather to provide sufficient information to guide the researcher in the design of his or her particular experiments. On the other hand, the most practical approaches to the preparation of hapten-protein conjugates were cited. Many of the methods used to prepare immunogenic conjugates have also been used to link drugs to carrier molecules (including antibodies) in order to "target" cytotoxic drugs. Two reviews that are useful in that they describe many of the methods used to make the carder-drug conjugates are those by Trouet T M and by Ghose. 'a2 The information in these reviews should be useful to immunologists as well. ,a, A. T r o u e t , Eur. J. Cancer 14, 105 (1978). ,3z T. G h o s e , J. Natl. Cancer Inst. 61, 657 (1978).
[5] P r o d u c t i o n
of Reagent
Antibodies
By B. A. L. HURN and SHIREEN M. CHANTLER Immunization The explosion of interest in immunoassay procedures during the last two decades has resulted in an enormous volume of literature describing, for the most part, satisfactory results of immunization. During the same period, knowledge of the underlying mechanisms of the immune response has advanced greatly from an original state of almost total ignorance. Perhaps unfortunately, those who have pursued basic understanding have seldom been much concerned with the practical problems of making useful reagent antibodies. As a result, with few exceptions the literature of immunization does no more than describe successful procedures, and the variety of these is legion. In the usual way of things, abortive attempts are seldom mentioned, let alone described, yet anyone with practical experience who has also discussed the matter with colleagues will be well aware that all the successful methods have also, at other times or in other places, singularly failed to give the desired results. Not surprisingly, the failure rate is higher when making antisera for more demanding test systems, such as radioimmunoassay, than for immunoprecipitin methods, for instance. Much of the uncertainty over the outcome of immunization may be ascribed to variations in individual animal response; however, when an METHODS IN ENZYMOLOGY, VOL. 70
Copyrighl © 19~0by Academic Press, Inc. All fights of reproduction in any form reserved. ISBN 0-12-181970-1
104
PRINCIPLES
AND
METHODS
[5]
General Comments It was not the purpose of this review to present a comprehensive survey of hapten-protein conjugates, but rather to provide sufficient information to guide the researcher in the design of his or her particular experiments. On the other hand, the most practical approaches to the preparation of hapten-protein conjugates were cited. Many of the methods used to prepare immunogenic conjugates have also been used to link drugs to carrier molecules (including antibodies) in order to "target" cytotoxic drugs. Two reviews that are useful in that they describe many of the methods used to make the carder-drug conjugates are those by Trouet T M and by Ghose. 'a2 The information in these reviews should be useful to immunologists as well. ,a, A. T r o u e t , Eur. J. Cancer 14, 105 (1978). ,3z T. G h o s e , J. Natl. Cancer Inst. 61, 657 (1978).
[5] P r o d u c t i o n
of Reagent
Antibodies
By B. A. L. HURN and SHIREEN M. CHANTLER Immunization The explosion of interest in immunoassay procedures during the last two decades has resulted in an enormous volume of literature describing, for the most part, satisfactory results of immunization. During the same period, knowledge of the underlying mechanisms of the immune response has advanced greatly from an original state of almost total ignorance. Perhaps unfortunately, those who have pursued basic understanding have seldom been much concerned with the practical problems of making useful reagent antibodies. As a result, with few exceptions the literature of immunization does no more than describe successful procedures, and the variety of these is legion. In the usual way of things, abortive attempts are seldom mentioned, let alone described, yet anyone with practical experience who has also discussed the matter with colleagues will be well aware that all the successful methods have also, at other times or in other places, singularly failed to give the desired results. Not surprisingly, the failure rate is higher when making antisera for more demanding test systems, such as radioimmunoassay, than for immunoprecipitin methods, for instance. Much of the uncertainty over the outcome of immunization may be ascribed to variations in individual animal response; however, when an METHODS IN ENZYMOLOGY, VOL. 70
Copyrighl © 19~0by Academic Press, Inc. All fights of reproduction in any form reserved. ISBN 0-12-181970-1
[5]
PRODUCTION OF REAGENT ANTIBODIES
105
experimental comparison of different procedures is made in such a way as to overcome the effect of individual animal variation, the results may well be inconclusive or irreproducible despite the considerable effort involved.1 Regrettably, then, it must be said that information concerning methods of immunizing laboratory animals is almost entirely anecdotal. The available evidence strongly suggests that there are influences as yet unrecognized that may be as important to success as any of the factors already known. Nevertheless, while acknowledging the significance of art, green fingers, or even plain luck, it is worth considering the known factors briefly so as to provide some evidence in support of the methods of immunization recommended later; they are related to the immunogen, the adjuvant, the choice of animal, the route of injection, and the dosage schedule. The Immunogen
Particulate (cellular) materials, such as heterologous erythrocytes or bacteria, are usually intensely immunogenic, producing a rapid response when administered without adjuvant of any sort. The major problem likely to be encountered is lack of the desired specificity in the resultant antiserum, since the particles have a complex antigenic structure much of which may be shared with other more or less closely related cell types. Short immunization courses are usually adequate but often give rise to a high proportion of IgM antibody, which may be very satisfactory in agglutination techniques but tends to be less stable during storage than IgG. Most antigens of interest to immunoassayists are soluble materials that vary greatly in their immunogenicity dependent on their chemical structure and molecular size. Since soluble substances are readily cleared from the circulation, either by some metabolic pathway or by excretion, through routes that largely bypass lymph nodes, spleen, and other reservoirs of immunopotent cells, they rarely stimulate the production of effective reagent antibodies unless administered with some sort of adjuvant, as described below. Even then, they vary widely in immunogenicity. Proteins and the larger polypeptides of molecular weight greater than about 5000 will readily stimulate a potent immune response. Many may exist in dimer or polymer form, either naturally or as a result of minor denaturation during purification, and this may increase their immunogenicity (major denaturation may be associated with loss of native antigenic characteristics, however, and should be avoided). The smaller the peptide 1 S. Lader, B. A. L. Hum, and G. Court, in "Radioimmunoassay and Related Procedures in Medicine," p. 31. International Atomic Energy Agency, Vienna, 1974.
106
PRINCIPLES AND METHODS
[5]
within the molecular weight range of 5000-1000, the more difficult it seems to be to make avid antisera, although the correlation is much less than perfect. In this size range, closely related (or even identical) peptides are found in all the usual species of laboratory animal, so the element of "foreigness" of the antigen is lost. Many small peptides may lack the clearly defined tertiary structure that is presumably necessary for a substance to be recognized as a unique antigen. Finally, degradation of these substances in the tissues and circulation, by specific enzymes and by nonspecific proteases, may well be so brisk as to prevent effective contact with immunopotent cells. With the exception of some of the larger polysaccharide molecules, no substances other than the proteins and larger polypeptides are effective immunogens in themselves. Nevertheless, antisera of high avidity and specificity can be raised to steroids, glycosides, oligopeptides, and the like if they are first chemically bonded to a large carrier molecule, preferably a protein that is in itself immunogenic in the species under immunization. Current immunological theory suggests that the initial stages of immunization require cooperation between T and B lymphocytes, the T lymphocytes first binding with a recognizably "foreign" substance and then presenting the bound antigen to B lymphocytes bearing suitable receptors. This cooperation is impossible if the antigen is too small to be shared between T and B cells, but a complex of the antigen with a suitable carrier becomes fully effective. Small, nonimmunogenic antigens of this type are known as haptens and, in the form of drugs, steroid hormones and small peptides, have been of great interest to immunoassayists during the last decade. The method of coupling carrier and hapten should be carefully chosen so as to avoid unwanted structural alteration of the latter and so that the linkage does not involve the immunochemically distinctive part of the hapten molecule. Antibodies produced in response to immunization with conjugated haptens generally "recognize" that part of the hapten farthest from the point of linkage, which thus determines their specificity. Highly substituted carriers are usually most effective, and molar ratios of 15-30:1 (hapten :carrier) are desirable, when possible. For best results the carrier should be a protein foreign to the immunized species-thyroglobulin and keyhole limpet hemocyanin are used quite widely, but bovine (or other) serum albumin is fully effective and more easily available. The subject has recently been well reviewed in relation to steroid conjugates by Pratt. 2 The purity of the immunogen is of controversial importance. For synthetic substances, however, no argument exists--the likelihood of closely related substances (such as "error peptides") being present in imJ. J. Pratt, Clin. Chem. 24, 1869 (1978).
[5]
PRODUCTION OF REAGENT ANTIBODIES
107
pure preparations, subsequently leading to the most objectionable variety of nonspecific antibody, means that maximum possible purity is essential. For particulate antigens, especially bacteria, there is also no reason for lack of purity, but the needs in respect to soluble substances extracted from natural sources are somewhat different. There is no doubt that relatively crude preparations are highly immunogenic, often more so than purer materials, so that many workers have thought of the impurities as having some adjuvant-like activity. The probability, however, is that greater purification has led to concomitant subtle chemical changes (such as deamidation) so that the immunogen stimulates antibodies that fail, to a greater or lesser extent, to " s e e " the native antigen. Despite this, a high degree of purification of immunogen must sometimes be sought in order to eliminate certain types of cross-reactivity in antisera. At the other extreme, gross impurity should be avoided, even when the cross-reactions are unimportant, because antigenic competition may then prevent formation of any specific antibody. In practical terms, about 10% purity is the minimum required to make a significant specific antibody response reasonably likely. T h e Adjuvant A wide variety of substances are known to have the property of potentiating the humoral antibody response to injected immunogen. Among them are inorganic adsorbents, such as aluminum hydroxide gel; mineral oils, such as liquid paraffin; and bacterial cell wall components. The diversity of materials having adjuvant properties, which has been the subject of a recent review by Whitehouse,3 makes it difficult to identify a simple mechanism of action. Three major effects are involved, albeit to different degrees for each adjuvant type. First, the release of immunogen from the site of injection is slowed, either by adsorption to solid particles or by incorporation into an oily emulsion. This leads to a "sustained release" from a depot at the injection site, where labile immunogens are also protected from breakdown by tissue enzymes. A secondary benefit is that any direct toxic effects of the immunogen on the recipient will be , minimized. Second, adjuvants have a stimulatory effect on reticuloendothelial cells, attracting a local infiltration of the injection area by mononuclear cells and stimulating phagocytosis by macrophages presumably by presenting soluble immunogen in a particulate or partially aggregated form. Adjuvant-treated macrophages with antigen inoculated into histocompatible recipients give rise to a higher antibody response than transfer 3 M. W. Whitehouse, in "Immunochemistry: An Advanced Textbook" (L. E. Glynn and M. W. Steward, eds.), p. 571. Wiley, New York, 1977.
108
PRINCIPLES AND METHODS
[5]
of macrophages containing antigen alone. 4 Third, it has been shown that adjuvants induce an increased circulation of lymphocytes through lymphoid tissues in the drainage area, 5 macrophages being important for the initiation of these lymphocyte traffic changes, e The increased flow of cells in regional lymph nodes is likely to allow greater contact between antigen and antigen-reactive cells, thus facilitating increased antibody production. The local granulomatous lesions formed at the sites of injection may also serve as foci of antibody production. The most important advance in adjuvant technology arose from the observation of Dienes and Schoenheit 7 that antigen injected into tuberculous granulomata stimulated higher antibody titers than did antigen injected at nontuberculous sites. These findings led Freund and co-workers to develop a series of adjuvants containing mycobacteria, mineral oil, and emulsifier, s which to this day remain the most potent tools available to the aspiring immunologist. The mixture most widely used in the preparation of reagent antibodies contains 9 parts of mineral oil to 1 part of detergent. The detergent (usually Arlacel A) contains a high level of both hydrophilic and lipophilic groups, thus facilitating dispersion of the oily and aqueous (immunogen) phases and allowing the formation of a stable emulsion. The simple oil-detergent mixtures are termed "incomplete" Freund's adjuvant; incorporation of heat-killed Mycobacterium tuberculosis or M. butyricum (0.5 mg/ml) into the oily mixture yieldS "complete" Freund's adjuvant. The latter is the more effective, probably as a result of greater stimulation of the local cellular response, and must now be regarded as an essential aid to the production of reagent antibodies against soluble immunogens. Preparation of Freund's Emulsions. For maximum efficiency, it is necessary to obtain a stable, water-in-oil emulsion. Several ways of pre-, paring such emulsions have been described, but there is no doubt that the simplest and most efficient, at least for the relatively small volumes that most people require, is the double-hub connector method described here. It may occasionally be difficult to persuade the phases to combine as water in oil rather than oil in water or mixed emulsions. Cooling the separate phases before mixing may help, but an infallible way of overcoming the problem is to use 2 - 4 volumes of oily adjuvant to 1 volume of aqueous 4 E. R. U n a n u e , B. A. A s k o n a s , and A. C. Allison, J. Immunol. 103, 71 (1969). 5 p. Frost and E. M. Lance, in " I m m u n o p o t e n t i a t i o n , " C I B A Found. Symp. 18 (New series), p. 29. E x c e r p t a Medica, A m s t e r d a m , 1973. 6 p. Frost and E. M. L a n c e , Immunology 26, 175 (1974). 7 L. Dienes and E. W. Schoenheit, J. lmmunol. 19, 41 (1930). s j. F r e u n d and K. M c D e r m o t t , Proc. Soc. Exp. Biol. Med. 49, 548 (1942).
[5]
PRODUCTION OF REAGENT ANTIBODIES
109
immunogen. Experience has shown these oil-rich emulsions to be at least as effective (probably more effective) than the 1 : 1 ratio usually recommended. A subsidiary advantage is that they flow more easily, so both mixing and injection are less of a chore. If it is essential to use a 1 : 1 ratio (because of volume restrictions when the immunogen solution is very dilute, for instance), the formation of water in oil emulsions can be reliably achieved by adding the aqueous phase in three increments, mixing after each. The necessary apparatus consists of the double-hub connector and two syringes, each large enough to contain the total emulsion volume without overfilling. The largest practicable volume that can be handled by someone with averagely large, reasonably powerful hands is 14-16 ml, using two 20-ml syringes. The best type of syringe for the purpose is an all-glass, center-hub pattern; metal-and-glass types tend to leak at the piston, and the common plastic syringes become very stiff while making the emulsion. Plastic syringes are reasonably satisfactory in the smaller sizes, however, since less force is needed for small volumes. If Freund' s complete adjuvant is required, shake it very thoroughly to resuspend the bacterial cells immediately before use. Pour out sufficient of the adjuvant into a small beaker (to avoid contaminating the remainder) and draw the required volume up into one of the syringes. Attach the double-hub connector, and carefully expel all air until the oil rises up into the farther end of the connector. Draw the aqueous immunogen into the other syringe, remove the needle, and again expel all air until the syringe hub is full of liquid, then connect it to the open end of the connector: any air left in the apparatus will be trapped in the emulsion and, because of its compressibility, will make injections more difficult. At this stage make sure that both syringes are firmly inserted into the connector, but be careful from now on not to place any bending stress on the rather unwieldly apparatus, especially if the syringes have glass hubs. A little oil will almost certainly be squeezed out of the connector (the whole process is somewhat messy) and it may be as well to wipe the apparatus and fingers with a tissue before proceeding. To form the emulsion, begin by squirting the aqueous phase into the oil as vigorously as possible, then continue squirting the total contents toand-fro from one syringe to the other a minimum of 10 times each way (20 times is better, if your thumbs can stand it). To avoid bending stress at the connections, practise deliberate relaxation of the "receiving" hand so that the filling syringe just rests on the palm as the other hand is grasping and pressing on the plunger. Especially as the hands tire it is tempting to let the receiving hand try to help the other, but this inevitably places strain on the syringe hubs. A fracture of a hub (or sudden falling apart of a
110
PRINCIPLES AND METHODS
[5]
carelessly made connection) causes an explosive shower of emulsion to contaminate everything with a radius of several feet (including the operatot's face) and is sufficiently unpleasant to encourage more care thereafter. If the aqueous phase is to be added in several aliquots in order to promote the formation of water-in-oil emulsions at the 1 : 1 ratio, it will be necessary to break one of the connections each time more of the water phase is needed. About five each-way strokes of the syringe will be sufficient for the intermediate mixing, after which the next aliquot of immunogen is taken up into the empty syringe, the connection is made again and, as before, mixing begins by squirting the water into the oily emulsion. Repeated disconnection and reconnection make it all the more difficult to exclude air and prevent messy oil from leaking out: it is better to use a larger oil to water ratio and avoid the problem altogether whenever possible. In the authors' experience, the above method will lead infallibly to the proper type of emulsion, and testing is therefore unnecessary. For those who wish to confirm success, however, the simplest way is to take a beaker half full of water and drop two small, separate drops of emulsion onto the water surface. The first drop always spreads somewhat, but the second will remain a discrete, white globule with no spreading at all if the emulsion is, indeed, water in oil. If the second drop disintegrates into bits and pieces that spread around over the surface of the water, the emulsion was oil in water, at least in part, and should be prepared afresh. Read the above instructions again first, though. After use, plastic syringes should be thrown away, but other apparatus must be washed up. The connector can first be pushed into a piece of rubber tubing connected to a hot tap and flushed through for a few minutes. Syringes should be cleaned with washing-up detergent, then soaked, together with the connector, in a decontaminating detergent (such as Decon 70) for a day or two before rinsing and drying. Residues from emulsions are probably difficult to remove completely, and the syringes should never again be used for any purpose other than preparing such materials. T h e Choice of Animal There are few instances in which categorical evidence has shown one species of common laboratory animal to give consistently better responses than another to any particular immunogen. Some fairly well know exceptions are the superiority of guinea pigs for production of antiinsulin sera (presumably because the endogenous hormone in this species is most unlike the other mammalian insulins) and of horses for preparation of antisera for immunoelectrophoresis. The latter preference is due to the
[5]
PRODUCTION OF REAGENT ANTIBODIES
111
solubility of horse antibody immune precipitates in excess antibody (all immune precipitates are soluble in antigen excess) yielding unusually narrow, clear-cut precipitin arcs. Apart from these examples, however, the literature abounds with indications of the personal preferences of the authors for which the evidence is apocryphal and often contradictory. In most instances the choice of species may reasonably be made on the basis of what is available and the volume of antiserum required--the larger the animal, the bigger the yield. It will be understood that it is usually sensible to immunize a species that is "foreign" to the antigen in question. If homologous immunogens are used, it should be for a valid reason (production of tissue typing sera, for instance, relies on antibodies produced in the same species as the donor). Homologous immunization, when it produces a result at all, will yield antibodies that recognize fine, interindividual differences in the antigen; by contrast, immunization of a foreign species readily yields much more abundant antibody but any reactivity with the structurally minor, idiotypic variants of the antigen is almost always lost in the reactivity against the gross, interspecies difference. A well provided laboratory may have access to guinea pigs, rabbits, sheep or goats, donkeys, and horses. There is little doubt that rabbits should be the first choice for most purposes unless very large amounts of serum are needed. Rabbits are cheap, easy to care for, robust in the face of quite intensive immunization, and easy to bleed. The other species may best be held in reserve in case of a failure with rabbits. Another reserve species that may be available is the chicken--again quite easy to handle, but producing antibodies that behave differently from those of mammalian species 9 and hence best avoided if possible. Whichever species is chosen, it pays to immunize several individuals (which is a good reason for avoiding the larger, more expensive species to begin with). Individual variation in response is often very striking, especially to the more "difficult" immunogens, and groups of at least four or five animals should be started if any difficulty whatsoever is to be expected in preparing satisfactory antisera. Nonproductive animals can be disposed of once it is clear that they will not improve (this may not be for several months with some immunogens) whereas the better responders can be kept under immunization for a year or more and bled repeatedly. Obviously, such a course cannot be followed where the early antibody is desired, for instance in the production of hemolytic serum with minimal hemagglutinating activity for use in complement fixation tests. 9 A. A. Benedict, in "Methods in Immunology and Immunochemistry" (C. A. Williams and M. W. Chase, eds.), Vol. 1, p. 229. Academic Press, New York, 1967.
112
PRINCIPLES AND METHODS
[5]
Immune responsiveness to certain antigens has been shown to be genetically determined. '° The importance of this in the context of antigens of general interest is not known, but it would seem to be desirable to use random-bred animals whenever possible, to give the best chance of a good response in one or more, unless previous experience has already shown that a particular inbred strain responds well to the immunogen in question. Whatever animals be chosen, they should be kept clean, healthy, and well fed if they are to perform well as antibody factories. The subject of animal husbandry is dealt with in a number of works (see, e.g., Short and Woodnott 11and Chase 12) but is, perhaps, of no direct interest to the readers of this chapter. T h e R o u t e of Injection For soluble immunogens, it is generally believed that the efficiency of stimulation of the immune response is related to the site of inoculation. A probable series, in order of increasing effect, is intravenous < intramuscular < subcutaneous < intraperitoneal < intradermal < intraarticular < intranodal. The principal reasons for the differences in efficiency are the speed with which antigen is lost from the site of injection and the likelihood of it passing through the lymph nodes or other centers of immunological activity on the way. These considerations, however, are radically affected by the use of adjuvants, especially oily adjuvants, which may stimulate a brisk local cellular reaction and release antigen over a period of several weeks or even months. Using oily adjuvants, then, the injection site can be chosen principally with a view to minimizing discomfort to the animal. Generally this means intramuscular injections in rabbits and larger animals or subcutaneous injections in guinea pigs; note that water in oil emulsions must never be given intravenously because of the virtual certainty of fatal fat embolism. Subcutaneous or intradermal injection of Freund's emulsions almost invariably leads to ulceration, but provided the sites are well chosen (see below) rabbits and guinea pigs show no sign of distress or loss of condition. Some authors (see Herbert 13) have suggested that Freund's emulsions should not be injected subcutaneously since ulceration may lead to 1o I. Green, W. E. Paul, and B. Benacerraf, Proc. Natl. Acad. Sci. U.S.A. 64, 1095 (1969). 11 D. J. Short and D. P. Woodnott, eds., "The I.A.T. Manual of Laboratory Animal Practice and Techniques," 2nd ed. Crosby Lockwood, London, 1969. ,2 M. W. Chase, in "Methods in Immunology and Immunochemistry" (C. A. Williams and M. W. Chase, eds.), Vol. 1, p. 254. Academic Press, New York, 1967. ,3 W. J. Herbert, In "Handbook of Experimental Immunology" (D. M. Weir, ed.), 2rid ed., App. 2. Blackwell, Oxford, 1973.
[5]
PRODUCTION OF REAGENT ANTIBODIES
1 13
loss of the depot: in the experience of the present authors this has never given rise to difficulty. Occasionally deep abscesses form after intramuscular injection and lead to loss of condition. The abscesses are frequently "sterile" and, in our experience, have usually been related to overzealous attempts to improve on sterile injection techniques (cleaning the skin over the injection area, for instance) rather than to the use of unsterile immunogens. Difficulties in preparing antisera against some of the antigens of interest in radioimmunoassay have led people to try a wide variety of methods of immunization. Most of these variations have been irrational (which does not mean to say they have not worked on occasion), but two deserve special mention. By injection of immunogen (angiotensin I, adsorbed on carbon black and emulsified in Freund's adjuvant) directly into rabbit lymph nodes and spleen, Boyd and Peart 14 obtained improved results that they believed to be due to more direct stimulation of the immune system. A subsequent comparative trial gave rather equivocal results, TM however, and the method was too difficult to be widely used. Injection into the Peyer's patches (lymphoid patches in the intestinal wall, quite easily visible in the rabbit) is technically much simpler but has proved no more successful in the authors' hands. Much simpler than the intranodal method, and now quite widely used, is the method of multiple intradermal inoculation introduced by Vaitukaitis e t a l . 16 The immunogen is introduced at 40 or more sites spread widely over the body surface. Antibody response to this primary immunization is much greater than to a first injection given in the usual way, and no more than one booster injection is usually required. Comparison with the usual intramuscular injection schedule ~ showed no great difference in efficiency, although the multiple intradermal technique (with only one booster) required rather less immunogen and yielded effective antisera in a shorter period of time. T h e Dosage of I m m u n o g e n and Timing of Injections Although an animal may be made "tolerant" to soluble antigens given in too low or too high a dose under certain circumstances, the use of a potent adjuvant makes such an outcome extremely unlikely. Nevertheless, the observation that too high a dose can lead to antiserum of rela14G. W. Boydand W. S. Pearl, Lancet 2, 129 (1968). 1~B. A. L. Hurn and J. Landon,in "'RadioimmunoassayMethods" (K. E. Kirkhamand W. M. Hunter, eds.), p.121. Churchill Livingstone,Edinburgh, 1971. 16j. Vaitukaitis,J. B. Robbins, E. Nieschlag, and G. T. Ross,J. Clin. Endocrinol. 33, 988 (1971).
114
PRINCIPLES AND METHODS
[5]
tively low avidity, 17.18presumably owing to stimulation of lymphocytes bearing low-affinity receptors, may certainly be relevant even when using Freund's adjuvant. Since most sensitive immunoassay techniques of current interest rely on antibody of the highest possible avidity, it is evidently desirable (and economical of immunogen) to use the lowest dose that will be fully effective. This dose is very much smaller than most of the published literature recognizes, and a suitable priming (first) inoculation for rabbits or guinea pigs will generally be of the order of 100 ~g. A range of 50-1000/~g should cover all needs, depending on the purity and immunogenicity of the material in question (but it is sensible to start at the lower end, since an animal showing lack of response after a sufficiently long trial can then be given a larger dose, whereas an animal producing poor antiserum after high dosage is beyond hope of salvage). The dosage required for larger animals does not increase in proportion to body weight: 0.255 mg is satisfactory for sheep and 0.5-10 mg for donkeys. For conjugated haptens, incidentally, these figures refer to total conjugate weight. Booster injections are always needed to obtain antisera of the highest titer and avidity. Practical experience suggests that good results will be obtained using a booster dose about half the size of an effective priming dose, given by the same route (not necessarily at the same site) and using Freund's complete adjuvant on each occasion. It is recognized that these recommendations are somewhat at variance both with immunological theory (which would suggest a progressive increase in dose) and with the advice of other authors to avoid repeated use of Freund's complete adjuvant, especially subcutaneously, because of abscess formation and hypersensitivity reactions. There is some documented evidence in support of the suggested reduction in dose, 1 but the repeated use of complete Freund's adjuvant is a recommendation that stems only from satisfactory, albeit uncontrolled, experience. The repeated booster doses that are usually required for the best arLtiserum should not be given too frequently. It has been shown 19that no further rise in titer results from a second injection given before the response to the first is reaching its peak. At least 4 weeks should pass between injections of Freund's emulsions. After the first booster, or sometimes after the second, antibody response may be quite prolonged and many people believe that a rest of 3-6 months is desirable before the next injection if antiserum of the highest avidity is required; the evidence in favor of this approach is not strong, 1 but in general terms there is little doubt that pa17G. W. Siskindand B. Benacerraf,Adv. lmmunol. 10, 1 (1969). 18E. J. Greeneand J. G. Tew, Cell. Immunol. 26, 1 (1976). 19W. J. Herbert, Immunology 14, 301 (1968).
[5]
PRODUCTION OF REAGENT ANTIBODIES
1 15
tience is desirable when making reagent antibodies. It is not unusual to read descriptions of immunization schedules involving weekly injections of quite large amounts of immunogen in Freund's emulsion; published accounts, not surprisingly, tend to report a successful outcome, but the approach is not to be recommended. Many published immunization procedures terminate with one or more intravenous injections of soluble immunogen given without adjuvant after a course of intramuscular Freund's emulsions. In the authors' experience, this produces a less satisfactory response (about half the final titer of avid antibody) than can be obtained with a final injection of intramuscular emulsion. By contrast with the above, particulate immunogens are normally administered intravenously, frequently (perhaps every other day), in increasing doses and for short periods of time. These materials are usually highly immunogenic, partly because the normal mechanism for their removal brings them into close contact with the immune system and partly because many of them (notably bacterial cells) are antigenically very "foreign" to the immunized animal. Antibody production is rapid, and the early IgM response is excellent for agglutination tests. Initial doses of immunogen are extremely variable, owing to the variable toxicity of the substances concerned (especially bacteria containing endotoxins), and for many of the antigens hypersensitivity reactions to later doses may prove rapidly lethal. Subcutaneous injection, with relatively slow absorption, may ameliorate undesirable acute reactions. Although short immunization courses for particulate antigens are the rule, usually in the belief that antisera will become less specific as immunization proceeds, this is not necessarily the case. Prolonged immunization may result in more stable IgG antibody of higher titer and, because of repeated bleeding over a period of time, in much greater yield.
Practical Immunization Schedules Animals often remain under immunization for many months, even years. You may not be personally responsible for their care during this time, but in your own interests you must ensure that either the individual animals or their cages are properly labeled in a manner compatible with your own records at the time of the first injection so that the individual animals can be identified with certainty thereafter. If the cages alone are labeled, you would also be advised to ensure that the method of animal handling, especially during cage cleaning, is such as to prevent animals being moved accidentally from one cage to another.
116
PRINCIPLES AND METHODS
[5]
Rabbits Four or more healthy, young adult rabbits should be treated with each immunogen. Soluble I m m u n o g e n s
Either the intramuscular or the multiple intradermal route may be recommended. As examples of representative immunogens for which highavidity antisera are required, consider a crude preparation of human chorionic gonadotropin (hCG) and the beta subunit of hCG (fl-hCG). The former, at a characteristic potency of 1500-3000 IU/mg, is about 20% pure whereas the latter is of necessity highly purified and in short supply. Appropriate doses for primary immunization are 1 mg and 100/zg, respectively. Booster doses should be half these amounts. Dissolve the immunogen in isotonic saline (other immunogens may require slight acidity, alkalinity, or other special condition) to a volume of 0.5 ml per rabbit for the primary injection or 0.25 ml for boosters (i.e., the same concentration for both injections). Emulsify the solution with three volumes of Freund's complete adjuvant, using a double-hub connector and two syringes as described above. The total volume of emulsion will then be 2 ml per rabbit for the primary inoculation or 1 ml for a booster. Use the emulsion within an hour of preparation. Intramuscular Schedule. Do not shave the animals or attempt to prepare the skin in any way prior to injection. A fairly stout needle of medium length (21 gauge x 1 inch) is convenient and need not be changed between animals unless it becomes blunted for any reason. Injections are given into thigh and/or upper foreleg muscle, where thickest, and the hair can be parted by gently blowing down on to the selected site immediately before injection. For the primary injection, give 0.5 ml of emulsion intramuscularly into each of the four limbs of each animal. Now go away and think about other things for at least 4 weeks, or 6 weeks if possible. For booster injections, give 0.5 ml of emulsion intramuscularly either into each hind limb or into each fore limb, alternately. Bleeds (20-40 ml) may be taken for testing on two occasions between 7 and 10 days after each booster and similarly every 3-4 weeks thereafter if the antiserum is satisfactory. Further boosters may be given at minimum intervals of 4 weeks (but preferably not within 2 weeks of a bleed) although it may pay to rest the animal for 3-4 months after the second or third booster. Animals that fail to show a reasonable response after two or three boosters should be disposed of. This decision must be related to the level of response expected for the particular immunogen u s e d - - s o m e animals
[5]
PRODUCTION OF REAGENT ANTIBODIES
117
may take several months to respond to "difficult" immunogens, and early responders are not necessarily the best in the end. Multiple Intradermal Method. Shave the hair on the back and on the proximal parts of all four limbs of each rabbit. As a guide to spacing the injections, draw six transverse lines across the shaved area of the back, using a felt-tip marker. The injections should be made with a tuberculin syringe and a fine needle (the syringe holds only enough for one animal but may be loaded repeatedly from the syringe in which the emulsion has been prepared, via the double-hub connector). Make 24 intradermal injections each of 0.05 ml, spaced evenly over the back. Distribute the remainder of the emulsion (about 0.8 ml, or sixteen 0.05 ml injections) over the inner and outer aspects of each upper limb, in the shaved areas. Satisfactory intradermal injections are easily recognized by a characteristic, localized bleb; this is easy to achieve on the back of the animal, where the dermis is quite thick and tough, but very difficult on the limbs, where the skin is much more delicate. Try, but do not be unduly discouraged if you fail. Within a few days of the injections the rabbit will present a horrifying sight, covered as it will be with forty, half-inch ulcers. In the authors' experience, the animals are happily unaware of the aesthetics of the situation and continue to thrive without any specific treatment. Some users of the technique have found otherwise, for no known reason. In the interest of animal welfare, if you find your rabbits are greatly upset by this procedure then please revert to the intramuscular procedure, which can be just as effective. After the multiple injections the animals should be left for at least 10 weeks before boosting. Antibody levels rise to relatively high titers during this time, however, and it is certainly worth taking a large bleed for testing after 8-10 weeks. All booster injections are given by the intramuscular route, and the method of treatment from the tenth week onward is thus exactly the same as for the previous schedule.
Particulate Immunogens These ~intigens are commonly administered by frequent, intravenous injection without adjuvant. Results are obtained quickly, the antisera often containing a high proportion of lgM immunoglobulin. There is a risk both of direct toxicity early in immunization and of severe hypersensitivity reactions as a result of later injections. With some at least of the antigens in question, equally satisfactory results can be obtained by intramuscular injection of Freund's emulsions--immunization is slower but less risky.
118
PRINCIPLES AND METHODS
[5]
Production of antisera to Escherichia coli, for use as specific typing reagents, furnishes an example of a typical intravenous schedule. Good bacteriological technique and the selection of an appropriate colonial form of the organism is essential to the specificity and reactivity of the antiserum (this is obviously analogous to the purification of a soluble immunogen). Living organisms are required for expression of the important K antigens in this species, but live coli will kill a high proportion of unprotected animals and the early injections are therefore made with heat-killed suspensions. Antisera to most other microbial species can be prepared against killed suspensions throughout. The following schedule should be followed (all suspensions being prepared to an opacity of Brown's tube 4, and all injections given intravenously). Day 1:0.25 ml of killed suspension Day 3:1.0 ml of killed suspension Day 5:3.0 ml of killed suspension Day 9:0.5 ml of freshly prepared living suspension Day 12:1.0 ml of freshly prepared living suspension Day 16:3.0 ml of freshly prepared living suspension Day 22: test bleed for titer Either Day 23 Bleed out if titer is satisfactory Or Continue weekly injections as for day 16 with test bleeds 5-7 days later, until satisfactory titers are obtained. Guinea Pigs Each animal will yield only 3-5 ml of serum by cardiac puncture or 15-25 ml when bled out. For this reason guinea pigs are best reserved for use when only small quantities of antiserum are required (particularly in radioimmunoassay and similar immunoassays) or when other animals are known not to respond well to the immunogen in question (insulin is such a substance, and, in our experience, parathyroid hormone is another). Groups of up to 10 guinea pigs may conveniently be kept in a single large cage, individuals being identified by natural markings or applied pigments (the latter need to be renewed rather frequently). Soluble immunogens should be administered as Freund's emulsions, injected subcutaneously into the abdominal wall just on either side of the midline. The injection sites will usually ulcerate after a week or so, but the animals are apparently free from discomfort, thrive, and make good antibodies. Prepare the inoculum by emulsifying 1 volume of aqueous immunogen
[5]
PRODUCTION OF REAGENT ANTIBODIES
1 19
in 2 - 3 volumes of Freund's complete adjuvant in the usual way, to give a total volume of 0.5 ml per animal. Injections should be given at intervals of not less than 4 weeks although longer rests later in the course of immunization may be desirable. Because of the low yield of serum and the risk of killing the animals when bleeding by cardiac puncture, it is less practicable to bleed guinea pigs repeatedly than it is to bleed rabbits. Since guinea pigs are cheaper to buy and look after, it is probably best to immunize a relatively large number for a comparatively long period of time, then bleed them out and select the best antisera from the result. Our experience has suggested that at least four injections are desirable if this strategy is employed, and six injections may often be better. The decision depends on the purpose for which the antiserum is required and, in particular, whether the highest possible avidity is needed. Sheep The immunization of sheep offers the possibility of obtaining relatively large amounts of antiserum, not only because each individual bleed is larger (150-300 ml of serum, depending on the size of the animal), but also because the animals may be maintained and bled repeatedly for longer than rabbits. This can be a major advantage when antisera are to be prepared for relatively undemanding, insensitive test systems such as immunoprecipitation, when larger volumes of reagent are required but variations in quality over the course of time are unlikely to cause difficulty. The higher cost of buying and keeping a sheep makes it less attractive when the use of a "difficult" immunogen makes it necessary to immunize a large number of animals. Circumstances alter cases, of course, and an Australian laboratory might have a different view of the relative economy of sheep and rabbits. Immunization of a sheep should proceed according to a schedule similar to that described for a rabbit. Intramuscular injections (as usual, always prepared with Freund's complete adjuvant) should be given with a 1½-2-inch needle deeply into the haunch or shoulder (preferably into all four "corners" for the first injection). As has been mentioned before, dosage is not proportional to size and for a relatively good immunogen such as human IgG an initial injection of 0.2-1 rag, followed by booster doses of half that size, should be sufficient. After the first two or three monthly injections, subsequent boosts should be given at longer intervals depending on the quality of antiserum. Bleeds may be collected on a regular schedule throughout the period of immunization, the best yields being obtained if three bleeds (of 300-600 ml, depending on the size and experience of the animal) are taken over a period of 8 - l 0 days followed by
120
PRINCIPLES AND METHODS
[5]
about 3 weeks rest before the next triple bleed. A healthy animal may remain productive for some considerable time, at least a year or two, but antibody levels will eventually decay and fail to respond to a further booster injection, at which time the animal should be disposed of. Collection and Storage of I m m u n e Serum Animals immunized with Freund's emulsions should be bled 7-10 days after booster injections. If the blood is taken from a vein rather than by cardiac puncture, two or three bleeds can be taken on successive days, but the animal should then be rested for 3 - 4 weeks before further bleeding or before boosting again if the original antiserum was not of satisfactory quality. After intravenous injection antibody levels rise and then fall more rapidly and bleeds should be collected 5-7 days after the last dose. It is often helpful to fast the animals overnight to minimize lipemia, but do not deprive them of drinking water. Blood should be collected in clean, dry, glass bottles and allowed to clot at room temperature or at 37° until the clot has retracted; it may help to "ring" the clot with a glass rod to promote separation. The sample should then be centrifuged and the serum be separated without undue delay in order to avoid unnecessary hemolysis, which looks unaesthetic although it has no obvious deleterious effect on the antibody. When handling large quantities of blood it may be easier to separate serum from the clot by letting it drain through a stainless steel mesh cone supported in a filter funnel--this can even be left to drain overnight in the cold room if the maximum possible yield is required, but in any case a final centrifugation will be required to remove residual red cells. After separation from the clot, antiserum may be stored without significant deterioration for long periods of time under a variety of conditions.Z° A counsel of perfection for reference or otherwise most precious reagents would be to filter sterilize, fill out in appropriate, accurately measured, small amounts (diluting in a suitable carrier medium if necessary), and then freeze-dry prior to storage at 4 ° or below. Experience has shown, however, that IgG antibodies are remarkably robust and that liquid antiserum (even without sterilization) can be kept for many months at 4 ° with 0.1% sodium azide added as an antibacterial agent. Storage at about - 20° in the ordinary laboratory freezer cabinet is, in theory, likely to cause protein denaturation due to the proximity of this temperature to the eutectic of sodium chloride (the complex mixture comprising serum will not be 20K. E. Kirkhamand W. M. Hunter,eds., in "RadioimmunoassayMethods,"pp. 189-193. Churchill Livingstone,Edinburgh, 1971.
[5]
PRODUCTION OF REAGENT ANTIBODIES
121
completely frozen at - 20°) but, again in practice, the freezer has proved most convenient and harmless to antibody protein provided that repeated freezing and thawing is avoided. Storage at lower temperature, preferably not in unreliable mechanical refrigerators, is very satisfactory when available. Gas-phase liquid nitrogen is the ideal low-temperature storage medium, being more reliable and convenient than mechanical or CO~ cabinets, not involving the special restrictions on storage vessels imposed by immersion storage in liquid nitrogen yet virtually guaranteeing lifetime stability of precious antisera (the investigator's lifetime, that is to say). Storage of IgM antibodies is far more of a problem and gives very variable results. Most antisera containing IgM can be handled exactly as described above, with only gradual deterioration that would be inapparent in the relatively undemanding test systems in which this class of antibody is generally used. Some, on the other hand, prove much less stable. On occasions, this instability is associated with bacterial growth (which seldom causes much loss of IgG antibody activity although it is embarrassing and should be avoided if possible). For this reason it is strongly recommended that IgM antisera should have 0.1% sodium azide added, be sterilized by filtration at the earliest possible opportunity (before bacterial growth and release of enzymes can occur) and be handled in a cleanly fashion thereafter. Even when collected after overnight fasting of the animal, defatted (see below), sterilized and with a bacteriostat added, serum stored at 4° will gradually become turbid and show a deposit, principally of denatured lipoprotein. This does not lead to any loss of antibody activity although it is easily mistaken for bacterial contamination and causes anxiety for that reason. The only practical disadvantage is seen when the antiserum is used in capillary precipitin reactions, when the turbidity can obscure the result unless the antiserum is first clarified by filtration. Further T r e a t m e n t of Antisera D e f a t t i n g A n t i s e r u m 21 Antisera to be used in capillary precipitin tests must be crystal clear so that the faint ring o f precipitation can be easily seen. Untreated sera become turbid on storage, due to precipitation of denatured lipoprotein; such precipitates can be r e m o v e d by membrane filtration prior to use, but it is usually better to reduce the severity o f the problem by extracting the bulk o f the lipoprotein at the time the serum is first prepared. zl A. S. McFarlane, Nature (London) 149, 439 (1942).
122
PRINCIPLES AND METHODS
[5]
Materials Diethyl ether, solvent grade Solid CO2-methylated spirit freezing bath Procedure 1. Place the serum in a beaker and add 3 ml of ether for every 10 ml of serum. 2. Place the beaker in the freezing bath. 3. Stir the serum-ether mixture quite briskly with a glass rod until it has frozen solid. The two liquids are completely miscible in these proportions at the freezing point. 4. Allow the frozen mixture to stand in the freezing bath for another 10 min, then remove the beaker and stand it in tepid water until the frozen plug loosens. 5. As soon as possible, tip the still frozen plug into a glass filter funnel (without filter) leading into a cylindrical separating funnel. Make sure the stopcock on the latter is free running and well lubricated with a silicone grease. 6. Allow the frozen material to thaw and run into the separating funnel at room temperature, then remove the filter funnel and close the separating funnel with a rubber stopper covered in metal foil. 7. Allow the separating funnel to stand undisturbed, at 4° if possible, overnight. 8. The next day the serum-ether emulsion will have separated into a lower layer of clear serum shading gradually into an opalescent zone of residual emulsion that has a sharp interface with the uppermost, opaque, fatty layer. Collect the serum by running off the bottom and intermediate layers. 9. Remove the bulk of the residual ether by boiling offunder reduced pressure, ideally with the aid of a rotary evaporator. 10. Add preservative and sterilize the serum by filtration prior to storage. NOTE: Due care should be taken to avoid the risk of fire or explosion when handling ether.
Absorption of Nonspecific Antisera The production of potent antiserum almost always results in a reagent with some degree of reactivity against nonspecific antigens, either because of impurities in the immunogen used or because there are "shared determinants" present in both specific and nonspecific antigens. Whatever the cause of the unwanted reactivity, it is usually necessary to re-
[5]
PRODUCTION OF REAGENT ANTIBODIES
123
move it by absorbing the antiserum with an appropriate antigen. Although absorption may be carried out with solutions of the antigen (the immune precipitate being removed afterward by centrifugation or filtration), excess antigen or soluble complexes of antibody and antigen will inevitably remain in the absorbed antiserum. A more satisfactory method, therefore, is the use of a solid phase immunoadsorbent prepared from the appropriate antigen, which can be added in excess and easily recovered for later re-use if required. Such adsorbents may be made from antigen alone by the use of cross-linking reagents (such as glutaraldehyde for protein antigens) or can be more complex reagents prepared by chemical coupling of the antigen to a solid support such as Sepharose. The latter technique is covered in a subsequent section on the use of IgG-Sepharose immunoadsorbent, but the present example describes the preparation of a glutaraldehyde polymer of F(ab')2 suitable for removal of light-chain crossreactivity from class-specific anti-immunoglobulin sera. It should be noted that the pH optimum for efficient polymerization by this method varies considerably depending on the protein to be treated; if a polymer of whole serum is required, for instance, a pH of 4.4 will be optimal.
Preparation of F(ab')2 lmmunoadsorbent by Glutaraldehyde Polymerization 22 Materials F(ab')2 prepared by pepsin digestion of IgG-z3 Phosphate buffer, 0.1 M, pH 7.0 Glutaraldehyde, 2% in saline Glycine-HCl buffer, 0.1 M, pH 2.5 Tris-HCl, 0.1 M, pH 8.0 Phosphate-buffered saline, 10mM, pH 7.5 (PBS) Procedure 1. Dialyze 100 mg of F(ab')2 preparation (20-50 mg/ml) against phosphate buffer at 4° overnight. 2. Place the F(ab'h solution on magnetic stirrer and add 0.4 ml of glutaraldehyde solution dropwise from a Pasteur pipette. 3. Allow the gel that forms to remain at room temperature for 3 hr and then place at 4 ° overnight. 4. Homogenize the gel in phosphate buffer then centrifuge hard in a bench centrifuge and discard the supernatant. 5. Repeat step 4 using glycine-HCl buffer. ~ S. Avramcas and T. Ternynck, lmmunochemistry 6, 53 (1969). 23 L. H. Madsen and L. S. Rodkey, J. Immunol. Methods 9, 355 (1976).
124
PRINCIPLES AND METHODS
[5]
6. Repeat step 4 using Tris-HCl buffer. 7. Repeat step 4 using PBS. 8. Wash polymer in PBS until the washings have negligible absorption at 280 nm. Tube Absorption Procedure 1. Mix 2 volumes of serum with 1 volume of packed polymer and stir on a magnetic stirrer at 37° for 1 hr. 2. Centrifuge at 4500 rpm for 5 min. 3. Transfer supernatant to another tube and recentrifuge. 4. Remove the supernatant and test it for specificity. NOTE: The polymer can be "regenerated" for further use by washing extensively with PBS followed by incubation with 3 M sodium thiocyanate, pH 6.6, for 30 min at room temperature to elute adsorbed protein. Wash the polymer finally with PBS and store at 4° in PBS containing 0.1% sodium azide. Preparation of Immunoglobulin Fractions from Whole Serum
Precipitation with Rivanol and Ammonium Sulfate Materials Rivanol (2-ethoxy 6,9-diaminoacridine lactate) Activated charcoal Saturated ammonium sulfate solution Isotonic saline Procedure 1. Adjust antiserum to pH 8.5 by careful addition of 0.1 N NaOH. 2. For each 10 ml of antiserum add 35 ml of 0.4% Rivanol solution dropwise from a separating funnel. Stir the serum gently on a magnetic stirrer throughout. 3. Decant the supernatant (containing the immunoglobulins) into universal bottles and centrifuge in a bench centrifuge to remove remaining sediment. 4. Decant the supernatant into a conical flask and add activated charcoal (1 - 1.5 g per 100 ml) to decolorize the solution. Agitate gently for approximately 10 min. 5. Remove charcoal from the protein solution by filtering through a double layer of moistened filter paper (Whatman No. 42) in a Biichner funnel. Transfer filtrate to a beaker. 6. Add an equal volume of saturated ammonium sulfate solution dropwise from a separating funnel, stirring gently on a magnetic stirrer throughout.
[5]
PRODUCTION OF REAGENT ANTIBODIES
125
7. When all the ammonium sulfate solution has been added, place the beaker at 4 ° for at least 6 hr to allow the immunoglobulin precipitate to flocculate. 8. Centrifuge at about 4000 g for 20 min, preferably in a refrigerated centrifuge, and discard the supernatant. 9. Dissolve the precipitate in a volume of saline approximately equivalent to half the volume of original antiserum. 10. Place the immunoglobulin solution in Visking tubing and dialyze extensively against several changes of saline to remove sulfate ions. (Alternatively, remove sulfate by chromatography on a Sephadex G-25 column.) 11. Check for residual sulfate ions by adding a few drops of the immunoglobulin solution to a tube containing a small volume of barium chloride solution. Any cloudiness indicates the presence of sulfate ions and the need for further dialysis. 12. Measure the volume of immunoglobulin solution and calculate the protein concentration by measuring the absorbance of a 1 : 25 dilution at a wavelength of 280 nm using a cuvette of 1 cm path length. concentration = (OD~80 × 25)/1.34 mg/ml (The factor 1.34 can be used for the immunoglobulins of most animal species).
Precipitation with Caprylic
A c i d 24
Materials Acetate buffer, 60 mM, pH 4.0 Caprylic acid Isotonic saline Procedure 1. Add 2 volumes of acetate buffer to the antiserum in a beaker. Check and adjust the pH of the mixture to 4.8. 2. For each 10 ml of starting antiserum add 0.74 ml of caprylic acid dropwise. Stir the mixture continuously on a magnetic stirrer at room temperature. 3. Continue stirring for 30 min. 4. Centrifuge at 4000 g to remove the precipitate (or filter on a Biichner funnel). 5. Retain the supernatant (containing the immunoglobulin) and dialyze extensively against saline at 4 °. N. Steinbuch and R. Audran, Arch. Biochem. Biophys. 134, 279 (1969).
126
PRINCIPLES AND METHODS
[5]
6. Measure the volume of the immunoglobulin solution and calculate the protein content as described above. NOTE. Since the final volume of immunoglobulin solution is approximately three times the volume of starting antiserum, concentration is usually necessary. This may be achieved by ammonium sulfate precipitation as described above, pressure ultrafiltration, or dialysis against hypertonic polyethylene glycol. If the latter procedure is used, the immunoglobulin preparation should subsequently be redialyzed against saline to remove any polyethylene glycol that has diffused into the dialysis bag, thereby contributing to the absorbance at 280 nm. Ion Exchange Chromatography Immunoglobulins, in particular IgG, may be separated from whole serum by ion exchange chromatography. The technique relies upon differences in the net charge of serum proteins: at low ionic strength and neutral pH, IgG carries a neutral or slight net positive charge and will not be adsorbed to diethylaminoethyl (DEAE) cellulose, unlike all other serum proteins. Although the principle of the method remains the same for the serum proteins of different species, the exact conditions of pH and ionic strength required for good separation of IgG will vary. A method for preparation of rabbit IgG by ion exchange chromatography using a batchwise procedure is outlined below. Materials Diethylaminoethyl (DEAE) microgranular preswollen cellulose (Whatman DE-52) Phosphate buffer, 5 mM pH 6.5 Procedure. The batchwise procedure of Stanworth 25 is used. 1. Equilibrate approximately 5 g of DEAE-cellulose with several changes of phosphate buffer. 2. Dialyze 20 ml of serum against phosphate buffer at 4° overnight. 3. Place the cellulose slurry in suitable containers such as universal bottles or large test tubes and centrifuge to sediment the particles. Check the pH of the supernatant buffer against that of the starting buffer to ensure that equilibration is complete. Discard the supernatant. 4. Add dialyzed antiserum to the packed cellulose and mix by gentle rotation for 1 hr at room temperature. 5. Centrifuge gently to sediment the cellulose, then carefully transfer 2s D. R. Stanworth, Nature (London) 185, 156 (1960).
[5]
PRODUCTION OF REAGENT ANTIBODIES
127
the supernatant (containing immunoglobulin) to a clean container. Discard the cellulose. 6. Recentrifuge to remove any remaining cellulose, and decant the supernatant immunoglobulin solution. 7. Calculate the protein content as described previously. If this preparation contains serum proteins other than immunoglobulin the process may be repeated using a fresh aliquot of equilibrated DEAEcellulose. Preparation of Immunospecific (Affinity-Purified) Antibody For some purposes it is necessary to use specific antibody rather than whole antiserum or a crude immunoglobulin fraction. Immunospecific antibody can be prepared by passing antiserum or a globulin fraction through an immunoadsorbent column containing antigen chemically coupled to an inert solid phase. Specific antibody combines with the immobilized antigen and can be eluted subsequently with "chaotropic" ions (such as thiocyanate) or low pH buffers. A method for preparation of human IgG immunoadsorbent and the elution of specific anti-IgG antibodies is given here. Use o f lgG-Sepharose Immunoadsorbent Prepared by Periodate Oxidation 2~,27 Materials Sepharose CL4B Sodium metaperiodate Ethanediol Isotonic saline Carbonate-bicarbonate buffer, 0.1 M, pH 9.5 Phosphate-buffered saline 10 AM, pH 7.5 (PBS) Sodium borohydride Sephadex G-50, suspended in PBS Sodium thiocyanate, 3 M, adjusted to pH 6.6 Procedure ACTIVATION OF SEPHAROSE
1. Suck dry some of the Sepharose CL4B slurry. Weigh out 20 g of the gel and wash it with saline in a Biichner funnel containing two Whatman No. 54 filter papers. 26C. J. Sanderson and D. V. Wilson,Immunology 20, 1061 (1971). 27T. J. G. Raybouldand S. M. Chantler,J. lmmunol. Methods 27, 309 (1979).
128
PRINCIPLES AND METHODS
[5]
2. Make up 40 ml of 10% sodium metaperiodate solution in distilled water. 3. Suck the Sepharose dry on the Biichner funnel and transfer the pad to the periodate solution. Mix or stir gently for 2-4 hr at room temperature. 4. Transfer the Sepharose slurry to a Biichner funnel containing two Whatman No. 54 papers. Wash quickly with saline to remove periodate. 5. Pour on 40 ml of 10% aqueous ethanediol, allowing the liquid to run through the gel very slowly to ensure thorough washing. 6. Wash the activated Sepharose finally with sodium carbonate-bicarbonate buffer and suck dry. C O U P L I N G OF ANTIGEN .
2. 3. 4. 5.
.
Prepare 100 ml of IgG solution at a concentration of 1.0 mg/ml in sodium carbonate-bicarbonate buffer. Add the activated Sepharose to the IgG solution and mix or stir gently for 18 hrs at room temperature (or 4° if preferred). Transfer the slurry to a Btichner funnel containing two Whatman No. 54 papers. Suck dry and wash with PBS. Prepare 20 ml of a 5 mg/ml aqueous sodium borohydride solution. Transfer the Sepharose pad to the borohydride solution and mix or stir gently for 2 hrs at room temperature. (CARE: Borohydride reduction is accompanied by evolution of hydrogen and should be carried out in a loosely stoppered vessel in a well ventilated area). Transfer the gel to a Biichner funnel containing two Whatman No. 54 papers and suck dry. Wash extensively with PBS, and finally resuspend in PBS to desired concentration. The gel is now ready for use.
PREPARATION OF IMMUNOADSORBENT COLUMN
1. Clamp a column (approximately 1.5 cm x 40 cm) to a stand, and with the outlet closed run a small volume of PBS into the column. 2. Pour the Sephadex G-50 suspension into the column and allow to settle until approximately 1 cm of column length is filled. Open the outlet to allow a flow of PBS, which facilitates column packing. Add more Sephadex slurry to give a packed volume of one-third of the column length, with a reasonable depth of PBS above the packed Sephadex. Close the outlet. 3. Pour the IgG-Sepharose slurry into the column carefully so as not to disturb the surface of the Sephadex and allow it to settle in a separate layer on top of the Sephadex. 4. Cut a circle of Whatman No. 54 filter paper the same size as the internal diameter of the column and allow to float onto the settled
[5]
PRODUCTION OF REAGENT ANTIBODIES
129
surface of the IgG-Sepharose. This prevents disturbance of the surface of the column during subsequent sample and buffer applications. 5. Open the column outlet to allow excess buffer to run through and wash the column contents by passage of PBS until the absorbance of the effluent at 280 nm is equivalent to that of washing buffer. APPLICATION OF ANTISERUM OR GLOBULIN SAMPLE 1. Dialyze 1 ml of the serum or globulin solution against PBS overnight at 4° . 2. Open the column outlet to allow the head of buffer to pass into the gel. Close the outlet. 3. Apply the dialyzed sample to the top of the column, taking care to avoid disturbance of the Sepharose. 4. Open the outlet and allow the sample to run into the Sepharose column, closing the outlet when all the liquid has been absorbed. 5. Run PBS onto the top of the column and allow to flow through slowly by opening the outlet slightly. Ensure that a head of PBS is always present to avoid drying out. 6. Unadsorbed serum proteins will pass through the column and can be detected by a suitable monitor. When all the protein has emerged allow the remaining head of PBS to pass into the column and then close the outlet. E L U T I O N OF BOUND ANTIBODY
1. Gently apply 5 ml of sodium thiocyanate solution to the column and allow to run into the gel by opening the outlet. 2. As soon as the thiocyanate solution has entered the gel, close the outlet, apply PBS, reopen the outlet and allow PBS to flow continuously through the column as previously. Eluted antibody contained in the thiocyanate solution will pass through the Sepharose and into the lower, Sephadex portion of the column. The molecular sieving properties of the Sephadex will serve to separate antibody rapidly from the thiocyanate and reduce the risk of denaturation. 3. Collect fractions containing the antibody, pool, and concentrate to approximately 5 mg/ml. 4. Measure the volume and protein content. Store frozen or freezedried in suitable size aliquots. NOTE. This method of purification will select all antibodies reacting with the antigen on the immunoadsorbent, including any that may crossreact with other antigens by virtue of shared determinants. If such antibodies are likely to be present (as, for instance, will be the case in antisera raised against whole IgG), they should be removed by straightforward ab-
130
PRINCIPLES AND METHODS
[5]
sorption as described in a previous section: Absorption of Nonspecific Antisera. Absorption may be carried out before or after preparation of the immunospecific antibody, but the former is to be preferred for logistic reasons. Preparation of Fluorochrome and E n z y m e - L a b e l e d Antibodies
Selection of Antisera for Conjugation Satisfactory conjugates can be prepared only from potent antisera of the required immunological specificity; it is important therefore that the purest available antigen be employed as immunogen and that multiple injections be given to ensure the production of antisera in which the ratio of antibody globulin to nonantibody globulin is high. Ideally, antiserum should be selected on the basis of tests of potency and specificity carried out prior to labeling. This preliminary evaluation may conveniently be performed by titration in conventional gel diffusion and assessment of specificity in immunoelectrophoresis. If the lack of a suitable soluble antigen makes such tests impossible, indirect immunofluorescent or immunoenzyme tests should be done utilizing a range of dilutions of pre- and postimmunization sera as the intermediary layer followed by the appropriate labeled anti-species immunoglobulin. Test samples, which may be histological preparations or cell films should be appropriately prepared (some prior knowledge of the system is almost essential) and should represent both "positive" (antigen containing) and "negative" (non-antigen containing) materials. Antisera exhibiting the highest level of activity and specificity should be selected for labeling. Labeling should be carried out on immunoglobulin preparations derived from the selected antisera, so as to maximize the proportion of specific antibody to total protein and hence reduce non-specific activity in the final reagent. Immunoglobulin can be prepared by any of the methods described above, but only in the most demanding systems will it be necessary to prepare immunospecific antibody rather than a crude immunoglobulin fraction.
Fluorescein Labeling of Antibody Globulins Materials Fluorescein isothiocyanate, isomer I (FITC) Carbonate-bicarbonate buffer, 0.1 M, pH 9.0 Immunoglobulin preparation (10 mg/ml in saline) Phosphate-buffered saline, pH 7.5 (PBS) Sephadex G-50 medium
[5]
PRODUCTION OF REAGENT ANTIBODIES
131
Procedure 1. Prepare a solution of FITC in carbonate-bicarbonate buffer to give a solution containing 1 mg of dye per milliliter. 2. Place a measured volume of the immunoglobulin solution in a small beaker and cool to 4°. Place on magnetic stirrer. 3. Add one-tenth volume of carbonate-bicarbonate buffer. 4. Add one-tenth volume of FITC solution dropwise while stirring the immunoglobulin solution at 4° (approximately 1 mg of dye per 100 mg of protein). 5. Check pH after addition of FITC and if necessary adjust to pH 9.0 with 0.1 N NaOH. 6. Cover reaction vessel and stir gently at 4° overnight. (Alternatively the reaction can be carried out at room temperature for 1 - 2 hr if the volume to be labeled is less than 20 ml.) Removal of unreacted free FITC is preferably performed by dialysis followed by gel filtration chromatography on Sephadex G-50 (medium). 7. Dialyze conjugate against several changes of phosphate-buffered saline (PBS). 8. Prepare Sephadex G-50 column equilibrated with PBS such that the packed volume is at least six times the volume of conjugate to be applied. Allow a disc of filter paper, cut to fit the dimensions of the column, to float onto the top of the column. This facilitates the even application of conjugate. 9. Allow the PBS to run through the column until no buffer remains above the top of the column. 10. Stop the flow of buffer and apply the conjugate. 11. Allow the conjugate to flow into the column by opening the tap. When all the conjugate has passed into the column elute with PBS. 12. Collect the first colored peak to emerge (this contains the labeled immunoglobulins) and concentrate to the original conjugate volume. 13. Conjugates can be stored at 4° or in aliquots at - 20° after the addition of a preservative such as 0.1% sodium azide. Repeated freezing and thawing is to be avoided.
Peroxidase Labeling of Antibody Globulins Although a variety of methods can be used for coupling enzymes to antibody, 2s the conjugation procedures most commonly used with horse2s S. Avrameas, T. Ternynck, and J. L. Guesdon, Scand. J. Imrnunol. 8, Suppl. 7, 7 (1978).
132
PRINCIPLES AND METHODS
[5]
radish peroxidase (HRP) are the two-stage glutaraldehyde 29 and periodate oxidation 3° methods. In the former procedure peroxidase is first mixed with an excess of the dialdehyde glutaraldehyde, which reacts with free amino groups of the enzyme via only one of its active aldehyde groups. After gel filtration chromatography to remove excess glutaraldehyde, the activated enzyme is mixed with the immunoglobulin preparation to allow the free aldehyde group to combine with an amino group of the immunoglobulin. Conjugates prepared in this way have been shown to contain a homogeneous derivative 29~1 with a molecular weight of 90,000, but the coupling efficiency is poor at around 25% and 5% for antibody and enzyme, respectively. 32 The low efficiency in this system appears to be due to the relative paucity of reactive amino groups in HRP. In contrast the periodate oxidation method of conjugation 3°'33 is not dependent on the presence of reactive amino groups but relies upon the generation of active aldehyde groups after periodate oxidation of the carbohydrate moiety of peroxidase. These aldehyde groups combine with the amino groups of added immunoglobulin to form Schiff bases, which are subsequently stabilized by reduction with sodium borohydride. Conjugates prepared by this procedure contain high molecular weight derivatives, 3°-32but the coupling efficiency is increased to approximately 60% for both antibody and enzyme )4 Recent studies using a modification of the method described by Kato et al. 35 have shown that peroxidase can be satisfactorily coupled to antibody by coupling via sulfhydryl groups introduced into both the immunoglobulin and enzyme structures. 36 Conjugates prepared in this way contain active derivatives that are heterogeneous in relation to molecular weight but retain good enzyme and antibody activity. 37 Glutaraldehyde Conjugation M e t h o d 2s Materials
Horseradish peroxidase RZ 3.0 Stock solution of glutaraldehyde, 25% in water 29 S. Avrameas and T. Ternynck, lmmunochemistry 8, 1175 (1971). 3o p. K. Nakane and A. Kawaoi, J. Histochem. Cytochem. 22, 1084 (1974). z~ M. Mannick and W. Downey, J. Imrnunol. Methods 3, 233 (1973). 32 D. M. Boorsma and J. G. Streefkerk, J. Histochem. Cytochem. 24, 481 (1976). 33 M. B. Wilson and P. K. Nakane, in "Immunofluorescence and Related Staining Techniques" (W. Knapp, K. Holubar and G. Wick, eds.), p. 215. Elsevier/North-Holland, Amsterdam, 1978. D. M. Boorsma, J. G. Streefkerk, and N. Kors,J. Histochem. Cytochem. 24, 1017 (1976). 35 K. Kato, Y. Hamaguchi, H. Fukui, and E. Ishikawa, J. Biochem. 78, 423 (1975). ~6 p. D. Weston, J. A. Devries, and R. Wrigglesworth,Biochim. Biophys. Acta 612, 40 (1980). 37 S. M. Chantler and L. S. Cooper, unpublished observations, 1978.
[5]
PRODUCTION OF REAGENT ANTIBODIES
133
Phosphate buffer, 0.1 M, pH 6.8 Sephadex G-25 Isotonic saline Immunoglobulin preparation, 5 mg/ml in saline Carbonate-bicarbonate buffer, 0.5 M, pH 9.5 Lysine solution, 1.0 M pH 7 Phosphate-buffered saline, pH 7.5 (PBS) Saturated ammonium sulfate Glycerol Procedure 1. Dissolve 10 mg of peroxidase in 0.2 ml of a freshly prepared 1 : 25 dilution of the stock glutaraldehyde solution in phosphate buffer and allow to stand at room temperature for 18 hr. 2. Pass through Sephadex G-25 column equilibrated with saline to remove excess glutaraldehyde. 3. Collect the brown fractions, which contain the activated peroxidase, pool, and concentrate to 1 ml. 4. Add 1 ml of immunoglobulin solution (previously dialyzed against saline) to the peroxidase solution. 5. Add 0.2 ml of carbonate-bicarbonate buffer and leave for 24 hr at 4 °. 6. Add 0.1 ml of lysine solution and leave the mixture at 4 ° for 2 hr. 7. Dialyze against several changes of PBS at 4°. If desired remove free enzyme by precipitation with saturated ammonium sulfate as described in steps 8-10. 8. Add an equal volume of saturated ammonium sulfate to the conjugate and allow to stand at 4° for 30 min. 9. Centrifuge for 20 min at 4000 g and discard supernatant. 10. Dissolve precipitate in approximately 1 ml of saline and dialyze extensively against several changes of PBS. (Alternatively, sulfate ions may be removed by gel filtration chromatography on Sephadex G-50.) 11. Preserve by adding an equal volume of glycerol, and store at 4 °. Periodate Oxidation Conjugation Method Two procedures have been described by Nakane and co-workers. In the first of these 3° free amino groups on the peroxidase are blocked by fluorodinitrobenzene (FDNB) treatment prior to the production of active aldehyde groups by periodate oxidation. A recent modification of this method, described here, omits FDNB blocking and recommends periodate oxidation of the enzyme at low pH prior to coupling with immunoglobulin .33
134
PRINCIPLES AND METHODS
[5]
Materials Horseradish peroxidase RZ 3.0 (HRP) Sodium metaperiodate (freshly prepared), 0.1 M Acetate buffer, 1 mM, pH 4.4 Carbonate-bicarbonate buffer, 10 mM, pH 9.5 Immunoglobulin preparation Carbonate-bicarbonate buffer, 0.2 M, pH 9.5 Sodium borohydride Sephacryl S-200 Phosphate-buffered saline pH 7.5 (PBS) Procedure 1. Dissolve 4 mg of HRP in 1 ml of distilled water. 2. Add 0.2 ml of freshly prepared periodate to the enzyme solution and stir for 20 min at room temperature. 3. Dialyze against acetate buffer overnight at 4° . 4. Prepare globulin solution containing 8 mg of protein in 1 ml of 10 mM carbonate-bicarbonate buffer. 5. Adjust activated HRP solution to approximately pH 9 by addition of 20/zl of 0.2 M carbonate-bicarbonate buffer. 6. Immediately add the globulin preparation to the HRP-aldehyde and stir for 2 hr at room temperature. 7. Add 0.1 ml of freshly prepared sodium borohydride solution containing 4 mg/ml and leave at 4 ° for 2 hr. 8. Separate unreacted enzyme from the mixture by chromatography on a column of Sephacryl S-200 equilibrated with PBS or by salt precipitation with ammonium sulfate as described above. 9. If purification of conjugates is performed by gel chromatography, the appropriate fractions should be pooled and concentrated prior to storage at - 20°. Addition of albumin (10 mg per milliliter of conjugate) or an equal volume of glycerol prior to freezing in small aliquots is recommended. Repeated freezing and thawing should be avoided.
Evaluation of Conjugates A variety of tests should be used to determine the efficiency of conjugation and the suitability of the conjugate in use. The extent of the testing performed, particularly with respect to specificity, will vary with the intended use of the reagent. Efficacy of labeling can be determined very simply by measuring the absorbance of the conjugate both at the 280 nm protein peak and at the maximum absorbance wavelength of the label used. For immunohistologi-
[5]
PRODUCTION OF REAGENT ANTIBODIES
135
cal studies the ratio of OD495 to OD2s0 for fluorescein-labeled reagents should lie between 0.6 and 0.9; and the ratio of OD4o3 to OD~s0 for peroxidase-labeled conjugates, between 0.3 and 0.6. This test, however, fails to show whether biological activity is present in the conjugate. This should be determined initially by using the conjugate as the antibody in appropriate gel diffusion or immunoelectrophoresis tests (if a suitable soluble antigen preparation is available), followed by testing in the immunofluorescent or immunoenzyme system in which it is to be used. Performance testing by titration (direct method) or chessboard titration (indirect method) is e s s e n t i a l in order to select the optimal working dilution of the reagent and to assess its specificity under working conditions. Tests of immunological specificity carded out by other methods (e.g., gel diffusion) are irrelevant and may even give misleading results because of the widely varying sensitivity shown by different test systems. 3s
Antibody Production by Lymphocyte Hybridomas 3sa Conventional immunization by injection of antigen into an animal stimulates the production of a heterogeneous population of antibodies that differ in respect of both their affinity and their specificity. Although the immunization procedure or prior treatment of the recipient may be manipulated to favor the production of antibodies of predominantly high or low affinity, the specificity of the antibody response is less amenable to control and antibodies directed against each of several antigenic determinants present in the immunogen will usually be present. The extent of this heterogeneity of response will differ not only among members of different species but also in individual animals of the same species despite the use of identical immunogen preparations and immunization schedules. These biological factors influence both the ease and reproducibility with which antisera of the desired immunological specificity can be prepared. The application of cell fusion techniques for in vitro production of antibodies of defined specificity offers a significant potential alternative to conventional methods of reagent antibody production. In 1973, Cotton et al. a9 successfully fused cells of two plasmacytoma lines to produce hybrid cells capable of synthesizing both myeloma proteins. Subsequently, hybrid cells derived by fusion of a murine myeloma with spleen cells from appropriately immunized donors were shown to seas S. M. Chantler and M. Haire, Immunology 23, 7 (1972). aaa The authors wish to acknowledge the helpful criticism of Dr. Jane Hewitt during the preparation of this section. 39 R. G. H. Cotton, D. S. Secherand, and C. Milstein, Eur. J. Immunol. 3, 135 (1973).
136
PRINCIPLES AND METHODS
[5]
crete antibodies against the immunogen used. 4° These hybrid cells (hybridomas) could be grown in tissue culture, producing antibodies of defined specificity in vitro; alternatively, antibody secretion could be obtained in vivo by inoculation of the hybridoma cells subcutaneously or intraperitoneally into syngeneic recipients. This approach thus offered the possibility of the production of monoclonal antibody of defined specificity by selective cloning procedures, avoiding the need for highly purified immunogen or elaborate antibody purification procedures. Although it is now well established that the fusion of mouse myeloma cells and antibody-secreting splenic lymphocytes is an effective means of producing homogeneous antibody of defined specificity, a number of technical variables remain. The same basic principles are applicable to many systems, but fairly extensive preliminary investigation is required to define the optimal conditions, particularly in relation to the choice of immunization schedule and donor species used. Investigators interested in detailed methodology should refer to the recent proceedings of a workshop on lymphocyte hybridomas. 41 Choice o f Fusion Partners
The myeloma line selected should exhibit good growth characteristics in vitro, a high fusion frequency (one hybrid per 105 to 106 normal cells)
and should be sensitive to the selective medium HAT. If the cell line is lacking in either of the enzymes hypoxanthine guanine phosphoribosyltransferase (HGPRT) or thymidine kinase (TK), growth in this selective medium (which contains hypoxanthine, aminopterin, and thymidine) will be impossible. 4z Only after hybridization with a normal cell containing the enzymes can DNA synthesis and growth occur; thus hybrid cells alone survive in the selective medium. A limited number of myelomas exhibiting these features are available, and the one most commonly used is P3-X63Ag8 of B A L B / c origin. Hybrids obtained by fusion of an antibody-secreting normal cell and a myeloma cell such as the above produce specific antibody together with the myeloma protein and the products of mixed genetic combinations. This heterogeneous immunoglobulin production may not always pose a problem, but in applications where greater purity is necessary the difficulty can be avoided by using a nonsecreting myeloma line that produces 4oG. Krhler and C. Milstein, Nature (London) 256, 495 (1975). 41 F. Melchers, M. Potter, and N. L. Warner, eds., "'Lymphokine Hybridomas,'"in Curr. Top. Microbiol. lmrnunol. 81. (1978). 42j. W. Littlefield, Science 145, 709 (1964).
[5]
PRODUCTION OF REAGENT ANTIBODIES
137
no immunoglobulin of its own but still supports the synthesis of spleen cell-derived immunoglobulins. 43,44 The phylogenetic relationship between the cells utilized in hybridization studies determines the functional success of the hybrids produced. Murine myeloma lines have been successfully fused to both syngeneic and allogeneic mouse spleen cells 45,46 and to rat spleen cells, 4r but fusion with human lymphocytes and with cells of rabbit or frog origin has been less successful. 4s Recently Galfre and his colleagues 49 have described a rat myeloma line that has been successfully fused to rat spleen cells, but as yet no suitable human myeloma lines are available. The ontological derivation of potential fusion partners is also important. 5° It appears that optimal results are dependent upon fusion with cells of the B lymphocyte series at an appropriate stage of differentiation. Although the exact characteristics of the cell have not been identified, an activated B lymphocyte at an early stage of differentiation appears to be preferable. It follows therefore that selection of fusion partners of compatible phylogeny and ontogeny together with preselection of suitably differentiated B lymphocytes will increase the success rate of obtaining functional hybridomas. In practice, splenic cells from immunized mice have been most extensively used in experimental work because of the availability of suitable murine myelomas. Although the myeloma line (P3-X63Ag8) commonly used in fusion studies is derived from the B A L B / c mouse strain, it is not essential to use this inbred strain as a source of donor cells. Instead, it is preferable to use a strain that provides the best response to the immunogen in question; however, if it is intended finally to inoculate the hybrid clones into animals in order to produce antibodies in vivo, then clearly the recipient animal must be histocompatible. This can be achieved by using, as recipients, F1 hybrids of B A L B / c and the strain selected for initial immunization.
43 M. Schulman, C. D. Wilde, and G. Krhler, Nature (London) 276, 269 (1978). 44 G. Krhler, S. C. Howe, and C. Milstein, Fur. J. lmmunol. 6, 292 (1976). 45 G. Krhler and C. Miistein, Eur. J. Immunol. 6, 511 (1976). 46 G. Krhler, T. Pearson, and C. Milstein, Somatic Cell Genet. 3, 303 (1977). 47 G. Galfre, S. C. Howe, C. Milstein, G. W. Butcher, and J. C. Howard, Nature (London) 266, 550 (1977). 48 G. Krhler and M. J. Schulman, in "Lymphocyte Hybridomas" (F. Melchers, M. Potter and N. L. Warner, eds.), Curr. Top. Microbiol. Immunol. 81, 143 (1978). 49 G. Galfre, C. Milstein, and B. Wright, Nature (London) 277, 131 (1979). 50 p. Coffino, B. Knowles, S. Nathenson, and M. D. Scharff, Nature (London), New Biol. 231, 87 (1971).
138
PRINCIPLES AND METHODS
[5]
Immunization Procedure In most somatic cell hybridization studies the potential spleen cell donor is immunized in order to increase the proportion of cells producing specific antibody. This enrichment of functionally active cells has been shown to increase the percentage of hybridomas exhibiting the desired specific antibody activity. The type of immunization schedule adopted will depend upon the physical nature of the antigen and its immunogenicity, so that the variables, such as use of adjuvant, route of injection, and the timing of injections, will differ in different studies. Immunization commonly involves an initial subcutaneous injection of immunogen followed by a booster intravenous injection. The animals are tested 2 - 5 days after the boost, and a good responder is given a second intravenous injection, spleen cells being harvested 2 - 5 days later.
Preparation of Spleen Cells Separation of nucleated cells from red blood cells present in the spleen cell suspension is rarely performed. Spleen cells are washed twice in serum-free medium, the yield from one spleen being approximately 1 x 108 nucleated cells. A ratio of 10 spleen cells : 1 myeloma cell is used for fusion. If it is possible to enrich the proportion of plaque-forming cells in the spleen suspension--for instance, by rosetting with antigen-labeled red blood cells followed by centrifugation in Ficoll-Isopaque--the ratio used for fusion may be reduced to 1 : 1. Such an enrichment procedure not only decreases the number of cells that need to be distributed into individual culture wells after fusion, but also increases the percentage of hybridomas that secrete antibody of the desired specificity, thereby reducing the number of tests performed in the selection of appropriate hybrid clones at a later stage.
Cell Fusion In early studies fusion was promoted by the use of Sendai virus, but more recently polyethylene glycol (PEG) of molecular weight 1000-6000 has been preferred. The sediment of spleen and myeloma cells is gently resuspended in the small volume of washing medium remaining after centrifugation, and approximately 2 ml of 50% PEG solution diluted in the serum-free medium is added. After incubation at 37° for 1 min, the cell mixture is diluted slowly with medium, approximately 5 ml being added over a period of 5 min. The suspension is then centrifuged and resuspended in the selective HAT medium (containing serum) to a final density of approximately 106 cells per milliliter. This procedure yields approximately 100 ml of suspension from one spleen.
[5]
PRODUCTION OF REAGENT ANTIBODIES
139
Growth of Hybrid Cells The fused cells, suspended in HAT medium, are seeded into individual tissue culture wells, putting approximately 10n cells into each well. Unfused myeloma cells cannot grow in this selective medium, and normal spleen cells are incapable of prolonged growth, so only the hybrid cells survive. These "microcultures" are examined periodically, and those showing growth visible over approximately 30% of the base of the individual wells are tested for specific antibody activity, this stage being reached in successful wells between 7 and 20 days after seeding. The percentage of wells showing growth will depend on the number of cells originally introduced: approximately 90% of wells exhibit growth when l0 s cells are placed in each culture well. The majority of the wells will contain multiple clones derived from different parent hybrid cells, the products of many of which are irrelevant to the particular study. The proportion of wells containing functional hybrids of the desired specificity will vary considerably, but approximately 5% of those showing growth may contain appropriate hybrids.
Evaluation of Activity of Hybrid Products The supernatants obtained from individual culture wells exhibiting growth must be tested to determine whether any hybrids present in that culture are secreting antibody of the required specificity. Since the level of immunoglobulin secretion is low (approximately 10-50 tzg/ml) and the number of wells to be tested may be relatively large, it is essential that highly sensitive and specific assays that are readily performed on small volumes of supernatants be used for screening. Radioimmunoassays are most widely used, but hemagglutination, hemagglutination inhibition, and (in cases where localization of activity is relevant) immunofluorescence and immunoenzyme procedures have been applied.
Cloning of Active Hybrids As previously mentioned, culture wells containing antibody of the appropriate specificity may contain a heterogeneous population of hybrid cells secreting a variety of products. Individual hybrid cells can be separated only by additional cloning procedures, either by growth in soft agar or by using the limiting dilution method. Cloning by the soft agar method is carried out in petri dishes 3 cm in diameter that contain a layer of normal spleen "feeder" cells (10e per plate) in 5% agar, over which is then layered a dilution of the hybrid cells (obtained from positive wells) suspended in a medium containing 20% fetal calf serum in 2.5% agar. A range of different dilutions of the hybrid
140
PRINCIPLES AND METHODS
[5]
cells may be treated in this way. After incubation, individual clones of cells are detectable within 1-2 weeks. These discrete colonies are then transferred to microculture wells and their products are again tested for activity of the required specificity. Cultures exhibiting appropriate activity are immediately recloned at least twice in order to select stable functional cell lines, which are stored by freezing in vials or transferred to larger culture vessels. The limiting dilution method of cloning involves culturing serially diluted suspensions of hybrid cells together with normal spleen cells, each dilution being set up in 6-12 wells. The average number of hybrid cells dispensed in each series lies between 240 and 0.1 cells per well. At high cell levels growth is observed in most wells, but statistical considerations suggest that if only one-third of the wells seeded at a particular cell dilution show growth, then it is highly probable that the cells growing within each of the individual wells are derived from a single parent cell. These wells are then tested for antibody activity, and the cellular contents are recloned to establish functional stability in the same way as those derived by soft agar cloning procedures. Antibody Production
Once stable hybrid clones secreting antibody of defined specificity have been isolated, methods of obtaining maximal amounts of antibody become important. These may involve in vitro culture or in vivo growth in a suit~tble recipient. The hybridomas may be maintained in continuous culture in vitro for several months at a cell density within the range of 104 to 4 × 106 cells per milliliter. Under these conditions an antibody yield of 10-100/~g/ml can be obtained, but in most cases loss of functional activity eventually occurs. The reason for such functional instability is not clear, but it is likely to be due to loss of chromosomes during a period of time in culture. For this reason in vitro antibody production is more satisfactorily performed in limited rather than continuous culture, selected stable clones being stored by freezing at an early stage in their life cycle so that a new vial of cells can be thawed when required to initiate a fresh culture. As an alternative, antibody can be produced in vivo. Many cultured hybridomas have been successfully transplanted to genetically compatible recipients, 45,49,51and hybridomas derived from cells of nonmurine origin have been successfully transplanted to athymic nude mice. 52 In vivo antibody production is achieved by inoculating the cloned, hybrid cells 51 T. Pearson, G. Gaifre, A. Ziegler, and C. Milstein, Eur. J. lmmunol. 7, 684 (1977). H. Koprowski, Z. Steplewski, D. Herlyn, and M. Herlyn, Proc. Natl. Acad. Sci. U.S.A. 75, 3405 (1978).
[5]
PRODUCTION OF REAGENT ANTIBODIES
141
subcutaneously or into the peritoneal cavity. If the latter route is used, mineral oil is given several days prior to inoculation in order to encourage the production of ascitic fluid. The level of antibody obtained by in vivo culture is reported to be 100- to 1000-fold greater than for in vitro culture .39,45,49,51 In assessing the efficacy of in vitro versus in vivo production of reagent antibody, consideration must be given both to the relative concentrations of antibody and to the volumes obtainable. Subcutaneous inoculation of cells in a mouse yields approximately 1 ml of serum 2 weeks later; the yield of ascitic fluid harvested 7-14 days after intraperitoneal injection is between 5 and 15 ml. As the concentration of antibody produced in this way is at least 100-fold greater than in tissue culture, 10 ml of ascitic fluid from one mouse would be equivalent to at least 1 liter of tissue culture fluid. In this context, the recent description of hybridomas produced by fusion of a rat myeloma line with rat spleen cells is likely to be of considerable practical significance because of the larger volume of serum obtainable following inoculation of hybridomas in these rodents? 9 Nonhybridoma Techniques
The production of a thriving, functional hybridoma is dependent on a close phylogenetic relationship between the two parent cell lines; the only species to have provided suitable cell lines so far are mice and rats. An alternative approach to the production of nonrodent antibodies in cell culture is provided by the transformation of B lymphocytes on exposure to Epstein-Barr virus (EBV). Adult human peripheral blood cells exposed to EBV have been shown to release polyclonal secretory immunoglobulin. 53 Cultures of human peripheral blood lymphocytes exposed to antigen (sheep red blood cells) and EBV produce specific antibody. 54 Preselection of human peripheral blood lymphocytes exhibiting surface binding of tetanus toxoid and the hapten NNP (4-hydroxy-3,5-dinitrophenacetic acid) followed by viral transformation has been shown to yield cells capable of antibody production in vitro, ss.5e Although the cultures have been shown to be active for some months, methods of increasing both the yield of antibody (at present only some 10 ng per milliliter of culture fluid) and longterm stability have yet to be devised. 57 Attempts to establish stable spesa A. Rosen, P. Gergely, M. Jondal, and G. Klein, Nature (London) 267, 52 (1977). 54 A. L. Luzzatti, H. Hengartner, and M. H. Schreier, Nature (London) 269, 419 (1977). ss V. R. Zurawski, E. Haber, and P. H. Black, Science 199, 1439 (1978). M. Steinitz, G. Klein, S. Koskimies, and O. Makel, Nature (London) 269, 420 (1977). s7 V. R. Zurawski, S. E. Spedden, P. H. Black, and E. Haber, in "Lymphocyte Hybridomas" (F. Melchers, M. Potter, and N. L. Warner, eds.), Curr. Top, Microbiol. Immunol. 81, 152 (1978).
142
[6]
PRINCIPLES AND METHODS
cific antibody-secreting cell lines by somatic hydridization with the murine myeloma P3-X63Ag8 have been unsuccessful. 58
Summary The successful fusion of normal and neoplastic lymphocytes has laid the foundation for the production of a variety of antibody specificities of practical relevance in research and diagnosis. The technical problems associated with this approach should not be underestimated, but one cannot fail to recognize the enormous range of applications that lie ahead once these problems have been overcome. H. Hengartner, A. L. Luzzatti, and M. Schreir, in "Lymphocyte Hybridomas" (F. Melchers, M. Potter, and N. L. Warner, eds.), Curr. Top. Microbiol. Immunol. 81, 92 (1978).
[6] P r e p a r a t i o n o f F a b F r a g m e n t s f r o m I g G s of Different Animal Species By MICHAEL G. MAGE
The light and heavy polypeptide chains of the IgG molecule are folded into a series of globular regions called domains 1 (Fig. 1). The portion of the polypeptide chain between the CT1 and CT2 domains of the heavy chain, known as the "hinge region, ''2 is relatively accessible to proteolytic enzymes. When whole IgG molecules are incubated with the proteolytic enzyme papain, in the presence of low concentrations of sulfhydryl compounds, one or more peptide bonds in the hinge region are split,3 leading to the release of the Fab and Fc fragments (Fig. 1). The Fab fragments of IgG antibodies thus consist of the light chain, and the Vx and CT 1 domains 1of the heavy chain. Fab fragments are univalent, in that each fragment contains a single antibody combining site, composed of parts of the variable regions (VL and Vn) of the light and heavy chains. Because of their univalency, Fab fragments can be used to advantage in procedures where it is desirable to bind antigen to antibody in solution without cross-linking or precipitation or to bind to antigen on cell surfaces without producing "patching" or "capping. T M G. 2 D. 3 S. 4 F.
M. Edelman and W. E. Gall, Annu. Rev. Biochem. 38, 415 (1969). S. Smyth and S. Utsumi, Nature (London) 216, 332 (1967). Zappacosta, A. Nisonoff, and W. J. Mandy, J. lmmunol. 100, 1268 (1968). Loor, L. Forni, and B. Pernis, Eur. J. Immunol. 2, 203 (1972).
METHODS IN ENZYMOLOGY, VOL. 70
ISBN 0-12-181970-1
142
[6]
PRINCIPLES AND METHODS
cific antibody-secreting cell lines by somatic hydridization with the murine myeloma P3-X63Ag8 have been unsuccessful. 58
Summary The successful fusion of normal and neoplastic lymphocytes has laid the foundation for the production of a variety of antibody specificities of practical relevance in research and diagnosis. The technical problems associated with this approach should not be underestimated, but one cannot fail to recognize the enormous range of applications that lie ahead once these problems have been overcome. H. Hengartner, A. L. Luzzatti, and M. Schreir, in "Lymphocyte Hybridomas" (F. Melchers, M. Potter, and N. L. Warner, eds.), Curr. Top. Microbiol. Immunol. 81, 92 (1978).
[6] P r e p a r a t i o n o f F a b F r a g m e n t s f r o m I g G s of Different Animal Species By MICHAEL G. MAGE
The light and heavy polypeptide chains of the IgG molecule are folded into a series of globular regions called domains 1 (Fig. 1). The portion of the polypeptide chain between the CT1 and CT2 domains of the heavy chain, known as the "hinge region, ''2 is relatively accessible to proteolytic enzymes. When whole IgG molecules are incubated with the proteolytic enzyme papain, in the presence of low concentrations of sulfhydryl compounds, one or more peptide bonds in the hinge region are split,3 leading to the release of the Fab and Fc fragments (Fig. 1). The Fab fragments of IgG antibodies thus consist of the light chain, and the Vx and CT 1 domains 1of the heavy chain. Fab fragments are univalent, in that each fragment contains a single antibody combining site, composed of parts of the variable regions (VL and Vn) of the light and heavy chains. Because of their univalency, Fab fragments can be used to advantage in procedures where it is desirable to bind antigen to antibody in solution without cross-linking or precipitation or to bind to antigen on cell surfaces without producing "patching" or "capping. T M G. 2 D. 3 S. 4 F.
M. Edelman and W. E. Gall, Annu. Rev. Biochem. 38, 415 (1969). S. Smyth and S. Utsumi, Nature (London) 216, 332 (1967). Zappacosta, A. Nisonoff, and W. J. Mandy, J. lmmunol. 100, 1268 (1968). Loor, L. Forni, and B. Pernis, Eur. J. Immunol. 2, 203 (1972).
METHODS IN ENZYMOLOGY, VOL. 70
ISBN 0-12-181970-1
[6]
PREPARATION OF FAB FRAGMENTS
"Hinge" "
~ ~ ~ 1
/ Regi°n~
\CL \ J -- ~
//
""; //
VL
Papain
)
/
/~_.Splits,
143
Fab
// ~-S-S-
lII
C~y2\\! ../
//
FIG. 1. A schematicdiagram of an IgG molecule, showingthe relationship of the Fab and Fc fragments to the intact ]gO molecule, and the cleavage by papain of the heavy chain between the C71andC72domains [modified from W. E. Gall and P. D'Eustachio,Biochemistry 11, 4621 (1972)]. Because they lack the Fc region (the C3,2 and C73 domains of the heavy chain1), Fab fragments are also useful where antigen binding is desired in the absence of effector functions, such as complement fixation, or where whole IgG molecules, particularly if complexed to antigen or if aggregated, could bind via the Fc portion of the heavy chain to cellular receptors for Fc. ~ This is of particular importance in studies using fluorescent antibody to surface antigens of cells that also have Fc receptors. Fab fragments have also been used therapeutically for the specific binding and excretion of small toxic molecules, 6 taking advantage of the Fab fragment's smaller molecular size, rapid clearance from the circulation, and lesser immunogenicity than whole IgG molecules. Fab fragments are usually prepared from whole IgG molecules by digestion with papain, as originally described by PorterF There are distinct subclasses of IgG, that in some animal species (e.g., mouse,a guinea pig, 9 sheep1°), in addition to having antigenically distinct Fc portions, differ with respect to electrophoretic mobility and binding to ion-exchange H. B. Dickler, Adv. lrnmunol. 24, t67 (1976). 8 V. P. Butler, Jr., D. H. Schmidt, T. W. Smith, E. Haber, B. D. Raynor, and P. Demartini, J. Clin. Invest. 59, 345 (1977). 7 R. R. Porter, Biochem. J. 73, 119 (1959). 8 M. Potter, Methods Cancer Res. 2, 105 (1967). B. Benacerraf, Z. Ovary, K. J. Bloch, and E. C. Franklin, J. Exp. Med. 117, 937 (1963). 10 E. T. Harrison and M. G. Mage, Biochim. Biophys. Acta 147, 52 (1967).
144
PRINCIPLES AND METHODS
[6]
media. The more acidic subclass, usually called IgG1, binds to DEAE-cellulose under conditions of low ionic strength, and can be eluted with a gradient of increasing salt concentration. TM However, it may be contaminated by small amounts of other serum proteins. The more basic class, IgG2, does not bind to DEAE-cellulose under these conditions, or elutes at the start of the gradient. Subsequent purification of the Fab fragments by ion-exchange chromatography is facilitated if the starting material is not a mixture of two IgG subclasses of different electrophoretic mobility. After purification of an immune IgG, one should check both subclasses for antibody activity, as antibody activity is not infrequently found to be predominantly in one subclass of IgG. ~1 The following procedure, modified from Keckwick TM and from Sober and Peterson, 13 can be used to prepare IgG from immune serum. To 100 ml of immune serum are added 75 ml of "36%" (36 g of Na2SO4, anhydrous, plus 100 g of H20) Na2SO4 solution slowly, with stirring. (Phenol, to a final concentration of 0.25%, can be added to Na2SO4 solution, as a preservative.) After standing for 1 hr at room temperature, the suspension is centrifuged for 20 min at 8000 rpm. The precipitate is redissolved in 20 ml of 0.15 M NaC1 and reprecipitated with 15 ml of the Na2SO4 solution. After 1 hr, the suspension is recentrifuged for 20 min at 8000 rpm. The pellet is redissolved and dialyzed against phosphate buffer, 10 mm phosphate, pH 7.6. Following dialysis, the retentate is passed through a column of DEAE cellulose (200 ml bed volume) equilibrated with the same buffer. The effluent consists of the more basic molecules of IgG (IgG2). Following emergence of the unbound IgG, an exponential gradient to 1 M NaCI can be used to elute the more acidic IgG molecules (IgG0, which emerge at the start of the gradient. The following procedure, modified from Porter, 1° can be used for proteolysis of IgG by papain. One milligram of crystalline papain or mercuripapain (Worthington Biochemical Corp., Freehold, New Jersey) is added to a solution of 100 mg of IgG in 10 ml of phosphate-buffered (10 mM phosphate, pH 7.3) 0.15 M NaCI, with 1 mM EDTA and 25 mM mercaptoethanol. The mixture is incubated for 1 hr at 37°. Further proteolysis is ended, and sulfhydryl groups are alkylated by adding iodoacetamide to a final concentration of 30 mM and incubating for an additional 15 min at 37°. 11 W. K. Ashe, M. Mage, R. Mage, and A. L. Notkins, J. Immunol. 101, 500 (1968). 12 R. A. Kekwick, Biochern. J. 34, 1248 (1940). la H. A. Sober and E. A. Peterson, Fed. Proc. 17, 1116 (1958).
[6]
PREPARATION OF FAB FRAGMENTS
145
FIG. 2. Immunoelectrophoresis of undigested goat IgG (upper well), and of Fab and Fc from papain digestion of goat IgG (lower well), showing the electrophoretic difference and the reaction of nonidentity between the Fab and the Fc fragments. The slot contains antibody to goat IgG.
After digestion, especially if preparing fragments from IgG of a previously uncharacterized animal species, the extent of fragmentation can be conveniently monitored by immunoelectrophoresis of the digest, TM where the electrophoretically separated fragments are visualized by precipitation with antiserum to the whole IgG. Since Fab and Fc are derived from different portions of the IgG molecule and share no antigenic determinants, there will be two arcs of precipitation that cross each other in a reaction of nonidentity, whereas whole IgG gives a single arc of precipitation (Fig. 2). 14 Purification of the Fab fragments is most conveniently done by ion-exchange chromatography. Choice of ion-exchange media depends on the animal species and subclass of IgG. In the rabbit, the Fc fragments are more basic than the Fab fragments, 7 and the separation is done with CMcellulose, the Fc fragment being most tightly bound and eluting last. In most other species (goat, 14 sheep, l°,H mouse, s human, ~5 horse, TM guinea 14 A. L. Notkins, M. Mage, W. K. Ashe, and S. Mahar, J. lmmunol. 100, 314 (1968). 15 B. Frangione and E. C. Franklin, J. Exp. reed. 12,4, 715 (1966). le j. H. Rockey, J. Exp. Med. 125, 249 (1967).
146
PRINCIPLES AND METHODS
[6]
FIG. 3. Immunoelectrophoresis of separated Fab (upper three wells) and Fc (lower three wells), following DEAE chromatography of a papain digest of goat IgG. The slots contain antibody to goat IgG. pig, 9,1r r a t , TM c o w , 19,2° c h i c k e n 21) t h e F c f r a g m e n t is m o r e a c i d i c t h a n t h e F a b f r a g m e n t , a n d D E A E - c e l l u l o s e c a n b e used.14 Z o n e e l e c t r o p h o r e s i s in s t a r c h gel, 22,2a P e v i k o n , 24 o r a g a r 25 h a s a l s o b e e n u s e d to p u r i f y h o r s e 16,~4 a n d h u m a n 26 F a b f r a g m e n t s . S e p a r a t i o n o f F a b b y i o n - e x c h a n g e c h r o m a tography can be done as follows: T h e d i s g e s t is d i a l y z e d a g a i n s t t h e s t a r t i n g b u f f e r f o r t h e s u b s e q u e n t c h r o m a t o g r a p h y ( a c e t a t e buffer, 10 m M a c e t a t e , p H 5.5 f o r C M - c e l l u l o s e , o r 10 m M p h o s p h a t e b u f f e r , p H 7.6, f o r D E A E - c e l l u l o s e ) . 17 R. C. Q. Leslie, M. D. Melamed, and S. Cohen, Biochem. J. 121, 829 (1971). ~8 V. Nussenzweig and R. A. Binaghi, Int. Arch. Allergy Appl. Imrnunol. 27, 355 (1965). 19 N. E. Kuchinskaya, A. Ya. Kulberg, and V. S. Tsvetskova, Biokhimia 30, 1065 (1965). 30 S. I. Wie, J. Immunol. 121, 98 (1978). 31 G. Dreesman and A. A. Benedict, J. Imrnunol. 98, 855 (1965). 33 H. G. Kunkel and R. J. Slater, Proc. Soc. Exp Biol. Med. SO, 42 (1952). 3a O. Smithies, Biochem. J. 71, 585 (1959). 3, H. J. Miiller-Eberhard, Scand. J. Clin. Lab. Invest. 12, 33 (1960). 25 D. V. Stefani arid A. Ya. Kulberg, Vopr. Med. Khim. 10, 279 (1964). 3s H. G. Kunkel, F. G. Joslin, G. M. Penn, and J. B. Natvig, J. Exp. Med. 132, 508 (1970).
[6]
PREPARATION OF FAB FRAGMENTS
147
FIG. 4. Rabbit Fab (well 1) and Fc (well 2) fragments, separated by chromatography on CM cellulose, showing reaction of nonidentity between Fab and Fc. Wells 3, 4, and 5 have antibody to rabbit IgG.
After dialysis, the retentate is placed on the appropriate ion-exchange column and unbound material (Fab) is eluted with starting buffer before starting the gradient. A gradient with limit buffer of 1 M sodium acetate, pH 5.5, is used for CM-cellulose. For DEAE-cellulose the limit buffer can be 1 M phosphate, pH 7.6. On applying the gradient, the first material to elute is additional Fab, followed by Fc. Chromatography can be done at room temperature or at 4°. Figure 3 shows the immunoelectrophoresis of goat Fab and Fc fragments separated by chromatography on DEAE-cellulose. Figure 4 shows the reaction of nonidentity of rabbit Fab and Fc fragments, separated by chromatography on CM-cellulose. After separation of the fragments, when preparing Fab fragments from a new animal species, the Fab fragment can be identified by antigen-binding activity, for example by inhibition by Fab of the precipitation reaction between antigen and whole antibody molecules (Fig. 5). T M For commonly used animal species, Fab fragments from nonimmune IgG can be detected
148
PRINCIPLES AND METHODS
[6]
FIG. 5. Precipitation of antigen (mouse lgG1) (well 1) by goat antibody to the same antigen (wells 2, 5) is inhibited by the Fab fragment of antibody to the same antigen (well 7), but not by Fab from another antibody (to mouse lymphoma EL4) (well 3). Fc from antibody to mouse IgGl (well 6) and rabbit IgG (well 4) likewise fail to inhibit the precipitation.
by precipitation in gel with anti-light chain antibodies, which are commercially available. Fab fragments purified by ion-exchange chromatography may still contain undigested or partly digested IgG molecules that still have some or all of the Fc portion of the IgG molecule attachedY "z8These contaminants can be detected by double diffusion in gel against antibody to the Fc fragment or to whole IgG. When the Fab is destined for use in a procedure ~r j. W. Goodman, Biochemistry 4, 2350 (1965). 38 T. E. Michaelson and J. B. Natvig, Scand. J. lmmunol. 1, 255 (1972).
[6]
PREPARATION OF FAB FRAGMENTS
149
FIG. 6. After chromatography on DEAE cellulose, a preparation of Fab from goat IgG (welt 2) still had material reacting with anti-Fc (well 1). Following passage through a column of protein A-Sepharose, the purified Fab (well 3) no longer had material reacting with the anti-Fc (well 1) but still reacted with anti-Fab (well 4).
where the absence of Fc is important, the Fc-containing fragments can be removed by passing the partially purified Fab fraction through an affinity column containing insolubilized anti-Fc antibodies. 29 Fragments containing Fc bind to the column, and the Fab fragments emerge unbound in the effluent. A procedure for preparing such an anti-Fc affinity column, quoted with permission za and modified, follows. One milliliter of horse anti-Fc serum (Behring-Werke Inst., Frankfurt, GFR) is added to 25 ml of 0.5 M NaCI-0.1 M NaHCOa buffer, pH 29 F. DeLaFarge and P. Valdiguie, J. Chromatogr. 123, 247 (1976).
150
PRINCIPLES AND METHODS
[6]
8.3, and the mixture is stirred (for 2 hr at room temperature) with 15 g of CNBr-activated Sepharose 4B (Pharmacia): This couples the antiserum to the CNBr activated Sepharose, and the activated Sepharose (with bound anti-Fc serum) is placed in a column (21 × 1.6 cm i.d.). Unbound material is washed from the column with the coupling buffer, and any remaining groups are allowed to react with 1 M ethanolamine at pH 8 for 1-2 hr. Three washing cycles are then used to remove noncovalently absorbed protein, each cycle consisting of a wash at pH 4 (0.1 M acetate buffer containing 1 M NaCl), and at pH 8 (0.1 M acetate buffer containing 1 M NaCl). When the IgG is of a subclass whose Fc binds to S t a p h y l o c o c c u s protein A, 3° an affinity column of insolubilized protein A can be used to remove Fc-containing fragments31 as follows.
aureus
Five grams of protein A-Sepharose (Pharmacia Fine Chemicals, AB, Uppsala, Sweden) are rehydrated according to the manufacturer's directions, packed into a column, and washed with l0 mM phosphate-buffered 0.15 M NaC1, pH 7.4. Goat Fab (23 mg in 13 ml) containing some molecules with Fc determinants still present (Fig. 6) was passed through the column. The effluent contained Fab but no longer reacted with anti-Fc (Fig. 6). When the Fc-containing fragments are sufficiently larger in size than Fab fragments, they can be removed by gel filtration on Sephadex G-150. 6"a2Fab fragments have been directly purified by binding to insolubilized antigen, followed by specific elution with hapten, e or with acetic or propionic acid. 33 They have also been purified by binding to an affinity column of insolubilized anti-Fab, followed by elution with 0.1 M glycineHC1 buffer, pH 2.8. 29 It should be noted, however, that such affinity purification by specific binding of the Fab fragment cannot be expected to remove undigested IgG molecules or partly digested fragments containing both Fab and Fc.
30H. Spiegelberg,Adv. ImmunoL 19, 259 (1974). al j. W. Goding,J. Immunol. Methods 20, 241 (1978). ~ H. G. Van Eyck, Biochim. Biophys. Acta 127, 241 (1966). 33j. L. Spratt and S. B. Jones, Life Sci. 18, 1013(1976).
[7]
C A R B O D I I M I D E S FOR I M M U N I Z I N G C O N J U G A T E S
151
[7] The Use of Carbodiimides in the Preparation of Immunizing Conjugates By S A R A B A U M I N G E R a n d M E I R W I L C H E K
Introduction and Principle Carbodiimides comprise a group of compounds whose general formula is R - - N ~ - - - C ~ N m R ', where R and R' are aliphatic, such as diethylcarbodiimide (C~H5--N~C~---N--CzHs), or aromatic, such as diphenylcarbodiimide
Sheehan and Hess ~ introduced the use of carbodiimides for the synthesis of peptide bonds. The preparative procedure is simple and easy to perform, and therefore it became the most important coupling method. The reaction may be represented as a dehydration and expressed as shown in Eq. (1).
R,
R2
RNHCHCOOH + H2NCHCOORs+R4N=C=NR 4
RNHCHCONHCHCOOR 3
+ R4NHCONHR 4
(1) The mechanism of the reaction is not yet fully understood. It is postulated that an intermediate is formed that can react either with an amine to give the desired peptide or to rearrange to an acyl urea [Eq. (2)].
J. C. Shcehan and J. P. Hess, J. Am. Chem. Soc. 77, 1067 (1955).
METHODS IN ENZYMOLOGY, VOL. 70
Copyright © 1960by Academic Press, Inc. All dightsof reproduction in any form reserved. ISBN 0-12-181970-1
152
[7]
PRINCIPLES AND METHODS RI I
RNHCHCONR 4 I
CO I
NHR4 (Acyl Ureo)
RI
NR 4 II RNHCHCO0-C I
(2)
I
NHR 4
Ri
R2
I
I
RNHCHCONHCHCOOR 5
+ R4NHCONHR 4
The acyl ureas are the main side products at elevated temperatures, and in order to shift the reaction toward peptide bond formation, temperatures around 0° should be used. Presence of an amino group during the reaction, also reduces the formation of acyl urea. The most commonly used carbodiimide for peptide synthesis performed in organic solvents is
urea, which is the product of the reaction, is very insoluble and precipitates in most solvents. Therefore it is easy to remove it by filtration. On the other hand, when the reaction is performed in aqueous solutions, the carbodiimides used are usually water soluble. The most useful water-soluble carbodiimides are 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (I) and 1-cyclohexyl-3-[2-morpholinyl-(4)-ethyl]carbodiimide(II). /CH~ CH3CH2 N = C = N CH2 CH2 CH2 N H(~)CI(-) CH3
(~)
CH3
@
N=C= NCH2CHPN
0
(3)
(rr)
The urea formed during the reaction is water soluble and can be extracted with water if the peptide synthesis is performed in organic solvents. It can be removed by dialysis or by gel filtration when used to couple haptens to high molecular weight carriers. In addition to the use of carbodiimides for the direct formation of peptide bonds, they can also be applied for the preparation of active esters, such as hydroxysuccinimide
[7]
CARBODIIMIDES FOR I M M U N I Z I N G CONJUGATES
153
orp-nitrophenyl esters. These esters can then be used for the formation of amide bonds according to Eq. 4.
0,,\ RCOOH+HO- N ~
0,,,, DCC • RCOON~
+ H2N-P
~RCONHP
(4) In the field of immunology, the main use of carbodiimides has been in the conjugation of weakly immunogenic or nonimmunogenic compounds to larger carder proteins or to synthetic antigens, thus enhancing their immunogenicity. Among the immunogens synthesized are complexes containing low molecular weight peptides, such as protein fragments (e.g., loop peptide of lysozyme 2) or hormones (e.g., ACTH, a bradykinin, 4 and gonadotropin-releasing hormoneS), steroidal hormones (e.g., estrogens, progestins, and androgense), prostaglandins, 7,8 cyclic nucleotides (e.g., adenosine 3',5'-cyclic phosphatea), and plant hormones (e.g., genistein 1° and gibberellic acidH). The conjugation of two compounds by the carbodiimide method requires the presence of an amino and a carboxyl group. In most cases the amino groups involved in the reaction are lysyl residues of the protein carrier 3-e or lysyl and alanyl residues of synthetic polypeptide carriers. 24° The carboxyl groups are, in most cases, contributed by the hapten. These either are originally present in the hapten or may be introduced into the molecule using a variety of chemical procedures. The introduction of such reactive groups has been performed either when the native molecule lacks such groups, a or when the native functional groups are also responsible for biological activity of the compound, and coupling through one of these groups may lead to their masking as antigenic determinants. 6,12The latter case relates in particular to steroid hormones, where methods have been 2 R. Arnon and M. Sela, Proc. Natl. Acad. Sci. U.S.A. 62, 163 (1969). 3 j. McGuire, R. McGili, S. Leeman, and T. L. Goodfriend, J. Clin. Invest. 44, 1672 (1965). 4 T. L. Goodfriend, L. Levine, and G. D. Fasman, Science 144, 1344 (1964). 5 y . Koch, M. Wilchek, M. Fridkin, P. Chobsieng, U, Zor, and H. R. Lindner, Biochem. Biophys. Res. Comrnun. 55, 616 (1973). e S. Bauminger, F. Kohen, and H. R. Lindner, J. Steroid Biochem. 5, 739 (1974). r L. Levine and H. Van Vunakis, Biochem. Biophys. Res. Commun. 41, 1171 (1970). s S. Bauminger, U. Zor, and H. R. Lindner, Prostaglandins 4, 313 (1973). 9 A. L. Steiner, D. M. Kipnis, R. Utiger, and C. Parker, Proc. Natl. Acad. Sci. U.S.A. 64, 367 (1969). 10 S. Bauminger, H. R. Lindner, E. Perel, and R. Arnon, J. Endocrinol. 44, 567 (1969). 11 S. Fuchs and Y. Fuchs, Biochim. Biophys. Acta 192, 528 (1969). l= S. L. Jeffcoate and J. E. Searle, Steroids 19, 181 (1972).
154
PRINCIPLES AND METHODS
[7]
devised for the insertion of chemical handles at different positions into the steroid molecule.6,12 Insertion of carboxyl groups to peptide hormones for similar reasons has been reported. 5,13 Among the mary methods for introducing carboxyl groups to molecules are succinylation,9 preparation of azo derivatives, 5 formation of chloroformate, hemisuccinate, or O-carboxymethyl oxime, TM and formation of thioether alkanoic acid derivatives.15'16 The compounds containing the native or inserted carboxyl groups are then attached to the macromolecular carrier using carbodiimide as the coupling reagent. Experimental
Methods for Introduction of Carboxyl Groups A carboxyl group can be introduced in almost any compound by various methods. In cases where a free thiol is present, the carboxyl group can be introduced easily by reaction with bromo- or iodoacetic acid. a7 This reaction is very mild and is usually performed at pH around 8-9. When the hapten contains a hydroxyl group, carboxylic acid may be introduced by one of the following methods: (a) carboxymethylation of the hydroxyl group with bromo- or iodoacetic acid; 17 (b) esterification with dicarboxylic acid anhydrides, such as succinic anhydride, to yield hemisuccinates, 14 which are unstable above pH 9; (c) reaction with phosgene, which results in the formation of chlorocarbonates. TM Haptens containing amino groups can be coupled directly to carriers containing carboxyl groups, or the amino group can be converted to a carboxylic acid by reaction with succinic anhydride, a'ls When a keto or an aldehyde group is present in the hapten, it can be converted to a carboxyl via reaction with O-(carboxymethyl)hydroxylamine15 or with hydrazides. Haptens containing double bonds can be made to react directly with mercaptoacetic or mercaptopropionic acid, TM or a two-step reaction may be performed: bromination followed by reaction with mercaptocarboxylic acid. a When a phenol or an imidazole is present in the hapten, the carboxylic 1~ y . Koch, T. Baram, and M. Fridkin, FEBS Lett. 63, 295 (1976). 14 B. F. Erlanger, F. Borek, S. M. Beiser, and S. Lieberman, J. Biol. Chem. 234, 1090 (1959). 15 H. R. Lindner, E. Perel, A. Friedlander, and A. Zeitlin, Steroids 19, 357 (1972). le A. Weinstein, H. R. Lindner, A. Friedlander, and S. Bauminger, Steroids 20, 789 (1972). 17 F. R. N. Gurd, This series, Vol. 11, p. 532. is I. M. Klotz, This series, Vol. II, p. 576. 19 R. Wagner and H. G. Gassen, Biochem. Biophys. Res. Commun. 65, 519 (1975).
[7]
CARBODIIMIDES FOR I M M U N I Z I N G CONJUGATES
155
acid is most easily introduced via a reaction with diazonium salts, such as diazophenylacetic acid. 5 Haptens containing guanido groups may be made to react with p-carboxyphenylglyoxal to yield carboxyl groups. Typical examples for introduction of carboxylic group are described below. ll~-Hydroxyprogesterone Hernisuccinate. This compound is prepared essentially by the method of Buzby et al. 2° l la-Hydroxyprogesterone dissolved in dry pyridine is refluxed for 20 hr with a fivefold excess of succinic anhydride. The solution is then poured into iced water and extracted with 1 ml of ether. After washing with water, the hemisuccinate is extracted from the ether with bicarbonate solution and precipitated with 10% hydrochloric acid. The compound is usually recrystallized from methylene chloride-hexane. Testosterone-3-(O-carboxymethyl) Oxime. This compound is prepared according to the general method of Erlanger et al.~4 Testosterone dissolved in ethanol is refluxed for 3 hr with a solution of O-(carboxymethyl)hydroxylamine in 2 N KOH. The ethanol is then removed by evaporation, water is added, and the mixture is washed with ethyl acetate. The aqueous solution is acidified to pH 2 with hydrochloric acid. The precipitate formed is filtered, washed with water, and recrystallized from ethyl acetate-petroleum ether. Gonadotropin-Releasing Hormone-Azophenylacetic Acid (LH-RH Azophenylacetic Acid). 5 L H - R H has blocked N and C terminals, and therefore a carboxyl group has to be introduced in order to make it available for conjugation to protein. This is performed by attaching p-diazophenylacetic acid to synthetic LH-RH, which results in the formation of an azo derivative. p-Aminophenylacetic acid in cold 2 N HCI is diazotized by the addition of nitrite in cold (4°) water. After 8 min at 4 °, the solution is brought to pH 8.5 with a cold solution of sodium bicarbonate and immediately made to react with a solution of LH-RH in 60% aqueous dimethylformamide (DMF) containing 20% NaHCOa. The reaction mixture turns orangebrown within a few minutes, and the reaction is allowed to proceed for 12 hr at 4 °. The mixture is acidified to pH 2 with 2 N HCI and washed with ether, The aqueous phase is adjusted with 0.5 N NaHCOa to neutral pH and used as such after lyophilization or further purification by chromatography. Since LH-RH contains one histidine and one tyrosine, the product is a mixture of azohistidyl and azotyrosyl derivatives. 20
G. C. Buzby, Jr., D. Hartley, G. A. Hughes, H. Smith, B. W. Gadsby, and A. B. Jansen, J. Med. Chem. 10, 199 (1967).
156
PRINCIPLES AND METHODS
[7]
Synthesis of Compounds The coupling of carboxyl-containing compounds to amino groups of the polypeptide carrier is usually performed with an excess of carboxyl groups and an equivalent amount of the water-soluble carbodiimide around pH 5 at room temperature. In some cases where the hapten is not water soluble, it is usually dissolved in DMF or dioxane. In such cases it is also possible to prepare the hydroxysuccinimide ester using dicyclohexylcarbodiimide and couple this ester directly to the carrier. The experimental procedure will be demonstrated by two examples. Genistein-poly(DL-alanine) (Genistein-p DLAla--pLys). lo Genistein2-carboxylic acid, prepared according to Baker and Robinson, 21 is attached to pDLAla--pLys (MW 175,000)22 through the o~-amino groups of alanine as follows: 0.2 g of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide-HCl (Ott Chemical Co.) is dissolved in 4 ml of H,O and added to a mixture of 1 g of pDLAla--pLys in 10 ml of H20 and 0.4 g of genistein-2carboxylic acid in 10 ml of DMF. After 20 hr at room temperature, the solution is dialyzed against 50% aqueous DMF for 24 hr, then against H20 for 48 hr and finally lyophilized. Spectrophotometric analysis indicated an average of 22 residues of genistein coupled per molecule of polymer.
Prostaglandin E~--Bovine Serum Albumin (PGE~-BSA). 8 PGE~ (100 mg, Upjohn Company, Kalamazoo) and 1.5 × 10-8 Ci of [3H]PGE~ (New England Nuclear Corp., 100 Ci/mmol) are dissolved in 1 ml of DMF. Then 60 mg of dicyclohexylcarbodiimide and 70 mg of N-hydroxysuccinimide are added. The reaction mixture is stirred at room temperature for 30 min. The precipitated dicyclohexylurea is removed by centrifugation, and the supernatant is added to a solution of 250 mg of BSA in 10 ml of 0.1 N sodium bicarbonate. The mixture is stirred at 4° for 2 hr, dialyzed against 50 mM sodium phosphate buffer, pH 8.0, and stored at - 2 0 ° . Measurement of the radioactivity in the conjugate indicated that an average of 15 residues of PGE~ were bound to each molecule of BSA.
Characterization of Conjugates Before characterizing the conjugate, it is essential to strip the complex of any molecule not covalently bound to the carrier. This is usually done by exhaustive dialysis TM or by gel filtration. 5 After these processes, the number of hapten molecules linked to each molecule of the macromolecular carrier can be determined by several ways. 21 W. Baker and R. Robinson, J. Chem. Soc. London 1926, 2713 (1926). 22 M. Sela, E. Katchalski, and M. Gehatia, J. A m . Chem. Soc. 78, 746 (1956).
[7]
CARBODIIMIDES FOR I M M U N I Z I N G CONJUGATES
157
1. The easiest way is to introduce a radioactive tracer. The number of residues attached to each molecule of the carrier can then be determined by comparing the specific activity of the hapten and the conjugate.~5 2. If the bound compound absorbs light at a different region from the cartier, the analysis can be done spectrophotometrically. 14 3. If the attached groups are not labeled and do not contain chromophoric groups, they can be quantitated by dinitrophenylation or deamination of the unoccupied lysines of the carrier. 23 The amino acid analysis of the conjugate before and after deamination or dinitrophenylation enables determination of the number of side chains per carder molecule. Usually only a fraction of the lysine residues present in the carder molecule is accessible for coupling, e.g., about 30-35 out of 59 lysine residues of BSA. In general, for a compound to yield a good antiserum, 15-30 residues of hapten molecules have to be attached to each molecule of BSA.
Immunization The methods of immunization used by various investigators are numerous, and therefore we shall describe only the method we have used for the raising of antibodies to steroid hormones. This method may be applied also to other immunogens. The antigen (1 mg per milliliter of saline) is emulsified with an equal volume of complete Freund's adjuvant and of a vaccine against Bordetella pertussis (Pertussis Vaccine fluid, USP, Eli Lilly & Co., Indianapolis, Indiana; 1.6 units per animal) and injected into multiple intradermal sites of the rabbits. Booster injections are given at monthly intervals, and the rabbits are bled 10-14 days after each booster injection. The same schedule, using only 100-200/~g of antigen, may be applied to rats.
Specificity of the Antibodies Produced The specificity of the antibodies produced to the hapten is affected mainly by the site of attachment of the hapten to the protein carrier; recently it has also been reported that different species of animals may produce antibodies differing in specificity when challenged with the same antigen (prostaglandin-protein conjugate). 24 The influence that the site of attachment of the hapten to the peptide carrier has on the specificity of the antibodies produced was studied in great detail, mainly in the field of antibodies to steroid hormones. When 23 C. B. Anfinsen, M. Sela, and J. P. Cooke, J. Biol. Chem. 237, 1825 (1962). 24 S. Bauminger, J. lmmunol. Methods 13, 253 (1976).
158
[7]
PRINCIPLES AND METHODS SPECIFICITY OF ANTISERA RAISED IN RABBITS WITH DIFFERENT CONJUGATES OF PROGESTERONE Cross-reaction (%) antisera to Compound
P-20- BSA a
P-7 - BSA b
P- 11- B SAc
Progesterone 11-Deoxycorticosterone Testosterone 17-Hydroxyprogesterone 20a-Hydroxy-4-pregnen-3-one 20fl-Hydroxy-4-pregnen-3-one Estradioi-17fl
100 96 95 98 34 96 95% before use. Antisera. These were kindly provided by many colleagues H and were purchased from commercial sources (rabbit antitestosterone serum, Searle Diagnostics, High Wycombe, Bucks., U. K.). 6 B. E. P. Murphy, J. Crn. Endocrinol. Metab. 27, 973 (1976). 7 M. A. Binoux and W. D. Odell, J. Clin. Endocrinol. Metab. 36, 303 (1973). s S. L. Jeffcoate and J. E. Searle, Steroids 19, 181 (1972). 9 G. C. Forrest, Ann. Clin. Bioehem. 14, 1 (1977). ~0 Immo P h a s e - - a commercial immunoassay kit for plasma cortisol (Corning Medical, Halstead, Essex, CO9 2DX United Kingdom). 11 D. J. H. Trafford, P R. Ward, A. Y. Foo, and H. L. J. Makin, Steroids 27, 405 (1976). lz C. J. O. R. Morris and P. Morris, eds. "Separation Methods in Biochemistry," 2nd ed., p. 179. Pitman, London, 1976. 13 A. F. Hofman, in "Biochemical Separation Techniques" (A. T. James and L. T. Morris, eds.), p. 283. Van Nostrand-Reinhold, Princeton, New Jersey, 1964.
[19]
HYDROXYAPATITE Radioactive
293
IN R A D I O I M M U N O A S S A Y
counts
(cpm)xlO 3 15 Antibody -bound
14
fraction
13
12
11
10
9
8
7
6
5
4 ~ Free.
l
t
3
2
1
0
I
I
I
I
I
I
I
10
20
30
40
50
60
70
Fraction
number
FlG. 1. Separation of free and antibody-bound [3H]testosterone on small columns (0.4 cm i.d. x 8 cm) of hydroxyapatite. Jail]Testosterone was incubated with antibody at 37 ° for 1 hr and applied to the column in 5 mM phosphate buffer. Radioactivity was eluted from the column by slowly increasing the concentration of eluting phosphate buffer. The concentration (C, in moles per liter) o f the phosphate eluent was related to time (t, in minutes) by the equation: C × 0.5 - 0.445e -t~4°. The flow rate was 0.1 ml/min, and 0.2 ml fractions were collected.
294
RADIOIMMUNOASSAYS
AND
IMMUNORADIOMETRIC
ASSAYS
[19]
I m m u n o a s s a y . This assay was carried out as described by Trafford et al. 11 unless otherwise specified. Antiserum [200/xl appropriately di-
luted with 0.1% bovine serum albumin (BSA) in phosphate buffer, pH 7.0, so as to give the maximum binding at zero together with the maximum displacement of labeled tracer over the range of hapten concentration being examined] was added to a dried extract of plasma, or standard steroid solution, to which 3H-labeled steroid had been added. After incubation at 37 ° for 1 hr, 50/.d of hydroxyapatite suspension were added, and tubes were gently shaken and centrifuged. Supernatant (100/.d) was removed for liquid scintillation counting, at least 1000 counts being collected for each pot. Efficiencies were approximately 25% with a background count rate of 40 cpm. It has been found possible to carry out the
%Unbound 100
1: 200 diln. present throughoutincubation)
.~..f.O(HA
8O e/
o..~-'~" I"
60
o 1:600 diln. (Normal, HA added after incubation)
f~
f
.,//
'°,7 20
1:600diln. (HA presentthroughout incubation)
= = y O "Oj o j -
0
I
I
40
i
I
80
i
I
120
I
I
160
I
i
200
Testosterone (pg) F I G . 2. Standard curves for the immunoassay of testosterone either adding hydroxyapa-
tite (HA) after incubation or with hydroxyapatite present during incubation.
[19]
HYDROXYAPATITE IN RADIOIMMUNOASSAY
295
EFFECT OF HYDROXYAPATITE ON VARIOUS LOW MOLECULAR WEIGHT COMPOUNDS
Labeled compounds [l~q]Triiodothyronine [l~q]Thyroxine [aH]Guanosine 3',Y-cyclic phosphate 3-O-Succinyldigitoxigenin-L-tyrosine, IZq-labeled [3H]Estradiol-17/3 [3H]Testosterone [all] 17a-Hydroxyprogesterone
Percentage of added radioactivity recovered in supernatant after addition of hydroxyapatite 91.8 -+ 1.7 a,o 99.7 -+ 3.8a 98.8 -+ 2.4a 105.3 -+ 3.8a 97.2 - 9.7c 97.2 _+ 5.2c 96.3 -+ 11.6c
a In 50 mM Tris-HCl buffer, pH 7.5. 0 Value +- standard deviation. c In 5 mM phosphate buffer, pH 7.0. immunoassay with hydroxyapatite present throughout the incubation with the antiserum, although the concentration of antiserum used has to be increased. At the end of the incubation, the tubes are simply centrifuged and the supernatant is sampled as before. Figure 2 shows standard curves for testosterone immunoassay obtained by adding hydroxyapatite after the incubation and also by having hydroxyapatite present throughout the incubation. E v a l u a t i o n of H y d r o x y a p a t i t e as a n A d s o r b e n t of A n t i b o d y - B o u n d Steroid. 11 Investigations have shown that the use o f hydroxyapatite appears to be a very satisfactory procedure for the separation o f free and antibodybound steroids and other low molecular weight compounds (see the table). The adsorption of antibody-bound steroid appears to be independent of time in contact with hydroxyapatite (up to 1 hr), temperature (4-37°), and amount of hydroxyapatite added. Buffers of p H > 7.5 appear to prevent adsorption to varying degrees, and phosphate buffers elute antibody-bound steroid from hydroxyapatite (see Fig. 3). H o w e v e r , TrisHC1 buffers up to molarities of 0.1 M have no appreciable effect on the adsorption o f antibody-bound steroid. No significant effects on the crossreactivity o f the various antisera used has been noted. E f f e c t of P r o t e i n Because of the relative lack of specificity of many steroid antisera, it has been the custom to extract the hydrophobic steroids from plasma or
296
[19]
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS %Unbound
80
60
40
20 10', 0 0"001
I 0"01
I 0-1
I 0-2
Concentration of phosphate ( m o l e s / I ) FIG. 3. Effect of increasing concentrations of phosphate buffer, pH 7.0, on the adsorption of antibody-bound steroid. [all]Testosterone was incubated with antibody without added unlabeled testosterone. Hydroxyapatite suspensions in phosphate buffers of different molarities were added; tubes were shaken and centrifuged, and the supernatants were counted. Phosphate buffer strengths recorded are the final concentrations after addition of the hydroxyapatite suspension.
urine prior to immunoassay. The presence of plasma protein has therefore been largely irrelevant. With the increasing specificity of modem antisera, it has become more common to dispense with the organic extraction step, and many steroids are now being assayed in unextracted plasma. ~4 Studies on the effect of protein on hydroxyapatite adsorption of antibody-bound steroid have been inconclusive. Hydroxyapatite adsorbs protein, and the presence of increasing concentrations of added BSA or gelatin (up to 2% protein) in the incubation medium merely requires the addition of extra hydroxyapatite in order to achieve the same adsorption of antibody-bound steroid. The addition of plasma or serum is more com14 E. A. S. AI-Dujaili and C. R. W. Edwards, J. Clin. Endocrinol. Metab. 46, 105 (1978).
[19]
297
HYDROXYAPATITE lN RADIOIMMUNOASSAY
plex, however, and there appears to be some component present in plasma that prevents the adsorption of antibody-bound steroid by hydroxyapatite. This effect was first noted during attempts to utilize hydroxyapatite in an immunoassay for insulin that is assayed directly in plasma: good standard curves were obtained in 1% human serum albumin but after the addition of plasma samples, hydroxyapatite was ineffective. Hydroxyapatite, when added to solutions of BSA in sufficient amounts, reduced the extinction of the protein solution at 280 nm to background, indicating total adsorption of protein. A similar experiment using diluted plasma gave no reduction in the extinction at 280 nm. It has been found, therefore, that hydroxyapatite cannot be used in the presence of significant concentrations of plasma or serum. Figure 4 shows the effect
•Unbound
1°°f , ,,e Undiluted 80 i
•
I :,,lO,oOOO
/ ~
~ • 1 in 10
•
I in 1,000
4°If/~~ ot
I
ill
100
I
200
Testosterone (pg)
FIG. 4. The effect of charcoal-stripped plasma of various dilutions on the standard immunoassay curve for testosterone, l m m u n o a s s a y s were carded out as described in the text in plasma diluted with 0.1% bovine serum albumin.
298
R A D I O I M M U N O A S S A Y S A N D I M M U N O R A D I O M E T R I C ASSAYS
[19]
of various dilutions of charcoal-stripped plasma on the immunoassay of testosterone. Plasma has been treated in various ways in attempts to identify and remove the interfering component. Dialysis, heat treatment (80° for 20 min), and oxidation with periodate have had no effect. Precipitation of plasma proteins with trichloroacetic acid or perchlorate, and subsequent dialysis, produced a dialyzate that had no appreciable effect on hydroxyapatite. This approach is impractical, however, because it cannot easily be applied to plasma immediately prior to assay. So far, we have not managed to find a simple solution to this problem that will enable hydroxyapatite to be used to separate free and antibody-bound steroid in the presence of significant concentrations of nonextracted plasma. Comparison of a Radioimmunoassay for Plasma Testosterone Using Hydroxyapatite with Other I m m u n o a s s a y s Methods The use of hydroxyapatite as a means of separating free and antibodybound steroid in a routine immunoassay for plasma testosterone n was evaluated by comparing the results obtained using hydroxyapatite with the mean values obtained on the same plasma samples estimated by a number of other laboratories taking part in the U. K. Supraregional Assay Service Quality Control scheme (by courtesy of Dr. M. Wheeler). Four female and six male plasma samples were provided in the Quality Control scheme, and each was analyzed a number of times on different occasions. A total of 56 assays were compared. The correlatin coefficient was 0.9650 and the equation of the regression line was: y(SAS) = 0.84(HA) + 0.28. The correlation coefficient was not significantly different from 1.00, and the intercept was not significantly different from zero. The slope was, however, highly significantly different from 1.00 (p < 0.001). Summary Hydroxapatite has proved to be very useful for the separation of free and antibody-bound steroid in the immunoassay of various steroids. Successful immunoassays have been set up, using hydroxyapatite, for the estimation of plasma testosterone, estradiol-17fl, 17a-hydroxyprogesterone, and progesterone. No steroid so far examined has been found to be bound to hydroxyapatite. Some other low molecular weight compounds, for which immunoassays are available, have also been found not to be bound to hydroxyapatite (see the table), and it may therefore be possible to use hydroxyapatite in immunoassays for compounds other than steroids.
[20]
ZIRCONYL P H O S P H A T E GEL IN RAD1OIMMUNOASSAY
299
[20] U s e o f Z i r c o n y l P h o s p h a t e G e l for t h e S e p a r a t i o n of Antigens from Antigen-Antibody Complexes in Radioimmunoassays: Assays for Carcinoembryonic Antigen, ot-Fetoprotein, and Insulin
By J. W.
COFFEY,
J. P. V A N D E V O O R D E , E. R. and H. J. HANSEN
SAUERZOPF,
During the last two decades, radioimmunoassay has become one of the most widely used techniques for the quantitation in both clinical and experimental situations of proteins, hormones, drugs, and many other physiologically important molecules that are present at very low levels in biological fluids. As a result, many methods have been developed for the separation of free antigen from the antigen-antibody complexes in incubation mixtures (see Felber 1 for review). Zirconyl phosphate (Z-gel) is a convenient and inexpensive reagent for the separation of free antigen from antigen-antibody complexes in certain radioimmunoassays. Examples of radioimmunoassays in which Z-gel has been used successfully, such as assays for carcinoembryonic antigen2 (CEA), a-fetoprotein (AFP), and insulin,3 are described in this report. Stragand and Hagemann4 have also used Z-gel in their assay for cyclic AMP. Webb and NowowiejskP have described the use of Z-gel in a radioimmunoassay for prostaglandins. Methods
Preparation of Zirconyl Phosphate Gel Zirconyl chloride.8H20 (Matheson-Coleman and Bell), 100 g, is dissolved at room temperature in 15 liters of 0.1 M HCI; 200 ml of 14.6 M HsPO4 are added slowly as the solution is stirred vigorously. The resultant precipitate is allowed to settle for 2 days, and then the supernatant is removed by aspiration. The settled gel is brought to a final volume of 15 liters with distilled H20, resuspended by stirring, allowed to settle as 1 j. p. Felber, Adv. Clin. Chem. 20, 129 (1978). z H. J. Hansen, L. Hainsselin, D. Donohue, R. Davis, O. N. Miller, and J. P. Vandevoorde, Methods Cancer Res. 11, 355 (1975). 3 j. W. Coffey, C. F. Nagy, R. Lenusky, and H. J. Hansen, Biochem. Med. 9, 54 (1974). 4 j. j. Stragand and R. F. Hagemann, Experientia 31, 1375 (1975). 5 D. R. Webb and I. Nowowiejski, Cell. lmmunol. 41, 72 (1978). METHODS IN ENZYMOLOGY, VOL. 70
Copyright © 1980by Academic Press, Inc. All rightsof reproduction in any form reserved. ISBN 0-12-181970-1
300
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
[20]
above; then the supernatant is aspirated as before. Washing of the gel with distilled H20 at room temperature is repeated five times; the gel is then brought to a volume of 15 liters with distilled H20. Concentrated acetic acid (17.6 M) is added with stirring to a final concentration of 0.1 M. The gel is allowed to settle for 2 days in this solution of 0.1 M acetic acid; then the supernatant is removed completely by aspiration. The pH of the resultant slurry of Z-gel is adjusted to 6.25 by the addition of concentrated NH4OH (15.1 M); the slurry is stored at room temperature after the addition of 0.01% (w/v) thimerosal as a preservative. The percentage of solids should be 37-45% when centifuged at 1000 g for 10 min in a hematocrit tube.
Preparation of Antigen, 125I-Labeled Antigen and Antisera Since the purpose of this report is to illustrate the usefulness of Z-gel as a reagent for the separation of antigens from antigen-antibody complexes, the methods 6,7 used for the purification of CEA and AFP will not be described. Bovine insulin was purchased from Schwarz-Mann, and anti-bovine insulin serum was purchased from Miles Laboratories. Antisera to CEA and AFP were prepared in goats and rabbits, respectively. Standard solutions of nonradioactive CEA (125 ng/ml) and AFP (180 ng/ml) are prepared in 50 mM borate buffer (pH 8.4) containing 10% (v/v) group A human plasma for CEA or 0.25% human albumin for AFP and 0.01% thimerosal. The bovine insulin standard (100/~IU/ml) is prepared in 40 mM phosphate buffer, pH 7.4, containing 0.15 M NaCI, 0.5% bovine serum albumin (BSA), and 0.01% thimerosal (phosphate buffer 1).
Iodination of CEA and AFP Proteins were iodinated by the method of Hunter and Greenwood. s Briefly, CEA and AFP are dissolved in 50 mM borate buffer (pH 8.4) at a concentration of 1 mg/ml. Fifty microliters of the protein solution are added to a vented V vial containing 5 mCi of carrier-free Na125I (pH 8-10) from Amersham Searle (IMS-30). Twenty microliters of freshly prepared chloramine-T (5 mg/ml in borate buffer) are added to the vial with a Ham-" ilton syringe, and the contents are mixed immediately. After 1 rain at room temperature, 20 t~l of Na~S205 (10 mg/ml in borate buffer) are added to stop the iodination reaction. The reaction mixture is combined with e j. Krupey, P. Gold, and S. O. Freedman, J. Exp. Med. 128, 387 (1968). 7 N. G. Anderson, D. W. Holladay, J. W. Caton, E. L. Candler, P. J. Dierlan, J. W. Eveieigh, F. L. Ball, J. W. Holleman, J. P. Breillatt, and H. J. Coggin, Jr., Cancer Res. 34, 2066, (1974). 8 W. M. Hunter and F. C. Greenwood, Nature (London) 194, 495 (1964).
[20]
ZIRCONYL PHOSPHATE GEL IN RADIOIMMUNOASSAY
301
1.0 ml of 5% human albumin in the case of AFP, or of group A human plasma in the case of CEA. The nSI-labeled proteins are separated from free 1251 by filtration through a Sephadex G-100 column (2.5 × 90 cm) equilibrated with 0.1 M Tris-HCl (pH 7.0) containing 0.15 M NaCI and 0.02% (w/v) NAN3. The column at 4° is eluted at a flow rate of 0.5 ml/min, and 5-ml fractions are collected into tubes containing 0.5 ml of 5% human albumin (AFP) or 0.5 ml of group A human plasma (CEA). The amount of 1251 in a 25-/~1 aliquot of each fraction is determined in a gamma scintillation spectrometer. The tubes containing the '25I-labeled protein are pooled, and the pool is diluted in 50 mM borate buffer (pH 8.4) supplemented with 0.25% human albumin (AFP) or with 10% group A plasma (CEA) and 0.01% thimerosal to yield a working solution of 125I-labeled CEA, which contains 200,000 dpm per 25 /~1, and of [125I]AFP, which contains 120,000-160,000 dpm per 50t~l.
Iodination of Insulin Twenty-five micrograms of insulin in 25 ~1 of 40 mM phosphate buffer (pH 7.4) are added to a vial containing 5 mCi of Na125I, and the iodination reaction is started by the addition of 10 t~l of chloramine-T (10 mg/ml in 40 mM phosphate buffer, pH 7.4). The reaction is allowed to proceed at room temperature for 2 min and then is stopped by the addition of 21~ tzl of Na~S205 (10 mg/ml in phosphate buffer). The contents of the iodination vial are combined with 0.2 ml of 1% BSA, and the p25I]insulin is separated from free 1251on a 2.5 × 20 cm column of Sephadex G-25 (4°) equilibrated with 0.1 M Tris-HCl (pH 7.4). The column is eluted at a flow rate of 0.5 ml/min, and 1.0-ml fractions are collected into tubes containing 1 ml of 1% BSA. The fractions containing [125I]insulin are pooled, the pool is diluted with phosphate buffer 2 (40 mM phosphate buffer, pH 7.4, 0.5% BSA, and 0.01% thimerosal) to yield a working solution containing approximately 90,000 dpm per 100/zl.
Titration Curve To Determine Appropriate Dilutions of Antisera for Radioimmunoassays Carcinoembryonic Antigen (CEA). Twelve disposable glass test tubes (1.5 × 15 cm, Packard Instrument Co.) are labeled appropriately, and 5 ml of EDTA buffer (3.7 mM EDTA, pH 6.5, 0.017% sodium azide, and 0.002% BSA) are added to each tube. Two-hundred microliters of 50 mM borate buffer, pH 8.4, containing 10% group A human plasma and 0.01% thimerosal are added to the two blank tubes, and 200/zl of five increasing dilutions (appropriate dilutions must be determined experimentally for each antiserum) in the same buffer of anti-CEA serum are added in dupli-
302
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
[20]
cate to the other 10 tubes. Then 25 ~1 of 125I-labeled CEA are added to all tubes. The contents of the tubes are mixed, and then the tubes are incubated at 45 ° for 30 min in a H20 bath. At the end of the incubation period, 2.5 ml of the Z-gel slurry are added rapidly to each tube, and the tubes are mixed on a Vortex mixer. The Z-gel is collected by centrifugation (1000 g for 6 min) at room temperature, and the supernatant is decanted. The pellet of Z-gel is washed once by resuspension with a Vortex mixer in 5 ml of 0.1 M ammonium acetate (NH4Ac), pH 6.25, followed by recentrifugation. The supernatant is decanted again, and the amount of 1251 associated with the pellet of Z-gel is determined in a gamma scintillation spectrometer. Under these conditions, approximately 70% of the ~25I in the preparation of 125I-labeled CEA binds to Z-gel in the presence of excess antibody, and less than 10% binds nonspecifically to Z-gel in the absence of antibody. A dilution of antiserum that will bind approximately 60% of the antibody-precipitable azsI-labeled CEA is used in the radioimmunoassay. a-Fetoprotein (AFP). Titration curves with anti-AFP serum and [I~5I]AFP are prepared using the procedure described for CEA with the following changes. The amount of EDTA buffer per tube is reduced to 2.5 ml, and in addition the EDTA buffer is made 1% (v/v) with respect to normal goat serum. Fifty microliters of p25I]AFP are added to each tube, and the tubes are incubated at 37° for 90 min. Approximately 70% of the 1251 in the [IzSI]AFP preparation binds to the Z-gel under these conditions in the presence of excess antibody; however, approximately 30% of the 1251binds nonspecifically to Z-gel in the absence of antibody. A dilution of antiserum that will bind approximately 60% of the [125I]AFP in the incubation mixture is used in the radioimmunoassay. Insulin. Twelve glass test tubes are labeled, and 0.2 ml of phosphate buffer 1 is added to each tube. The two blank tubes receive 0.1 ml of phosphate buffer 2, and the remaining I0 tubes receive in duplicate 0.1 ml of five increasing dilutions of anti-insulin serum in phosphate buffer 2. A constant amount (0.1 ml) of [125I]insulin in phosphate buffer 1 is added to each tube, and the tubes are incubated at 37° for 45 min. The reaction is stopped by the addition to each tube of 5 ml of the Z-gel slurry followed by 10 ml of 0.1 M NI-I4Ac (pH 6.25). The contents of the tubes are mixed by inversion, and the Z-gel is collected by centrifugation (1000g for 5 min) at room temperature. The supernatant is decanted, and the pellet is washed by resuspension in 10 ml of 0.1 M NH4Ac followed by recentrifugation. The supernatant is decanted, and the amount of ~2zI associated with the pellet of Z-gel is measured. Approximately 80% of the 1251in the preparation of [~5I]insulin binds to Z-gel in the presence of excess antibody, but only 10% binds in the absence of antibody. A dilution of the
[20]
ZIRCONYL PHOSPHATE GEL IN RADIOIMMUNOASSAY
303
antiserum that will bind to the Z-gel albproximately 50% of the [x2sI]insulin is used for the preparation of standard inhibition curves.
Preparation of Standard Inhibition Curves Carcinoembryonic Antigen (Fig. IA). Five mililiters of EDTA buffer is added to each of 10 glass test tubes and then 10, 25, 50, and 100/~1 (1.25, 3.125, 6.25, and 12.5 ng) of the CEA standard are added in duplicates to 8 of the tubes. Two tubes receive no nonradioactive CEA. Twenty-five microliters of the appropriate dilution (see titration curve) of anti-CEA serum are added to each tube; the tube contents are mixed and then incubated in a H20 bath for 30 min at 45 °. After this initial incubation period, the tubes are removed from the bath, and 25 tzl of ~25I-labeled CEA are added to each tube. The tubes are returned to the 45 ° bath and are incubated for an additional 30 min. The tubes are removed from the bath, Zgel is added to stop the reaction, and the tubes are processed as described for the titration curve. ~-Fetoprotein (Fig. 1B). Standard inhibition curves for AFP are prepared essentially as described for CEA with some modifications. The
A
40
501"
B
20,
C
1
i g
P
I
N ¢
o
(2_
I
x x
u
~
2C I 0
I
I
4
Corclnoembryonlc
I 8
I
I 12
Antigen rig/tube
4 -
0
I
I
4
I
I
8
I
I
12
I
1
16
Q- Fetoproteln ng/tube
I
I
0
i 4
I
I 8
ImLulln /~Iu/tube
FIG. 1. Standard inhibition curves for the measurement of carcinoembryonic antigen (A), ct-fetoprotein (B), and insulin (C) by radioimmunoassays utilizing Z-gel. The total amounts of ~25I in the tubes used for the preparation of the inhibition curve were 90,000 cpm, 79,000 cpm, and 45,000 cpm for carcinoembryonic antigen (CEA), ,x-fetoprotein (AFP), and insulin, respectively, with a gamma scintillation counter efficiency of 45%.
I
304
RADIO1MMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS'
[20]
amount of EDTA buffer is reduced to 2.5 ml per tube, and 1% normal goat serum is added; 10, 25, 50, and 100/zl (1.8, 4.5, 9.0, and 18.0 ng) of the AFP standard are added in duplicate to 8 tubes. Two tubes with no nonradioactive AFP are included in the assay. Fifty microliters of the appropriate dilution of anti-AFP are added to each tube, the contents of the tubes are mixed, and then the tubes are incubated at 37° for 90 min. The tubes are removed from the bath, and 50/xl of [125I]AFP are added to each. After a second incubation period of 90 min at 37°, 1 ml of Z-gel is added to each tube; then the tubes are processed as described (see titration curve). Insulin (Fig. 1C). Standard inhibition curves are prepared using 10 tubes, each of which contains 0.1 ml of buffer 2 and 0.1 ml of the appropriate dilution of anti-insulin serum. Four pairs of tubes receive 0.1 ml of the insulin standard diluted so as to contain 1.0, 2.5, 5.0, or 10.0/zlU of insulin. The tubes are incubated at 37° for 2 hr, and then 0.1 ml of [~25I]insulin is added to each. After an additional incubation period of 45 min at 37°, 5 ml of Z-gel are added and the tubes are processed as described (see titration curve). Comments This radioimmunoassay for CEA has been widely used clinically to quantitate the level of CEA in perchloric acid extracts of plasma. 9 The sensitivity of this assay is approximately 0.5 ng of CEA per milliliter of plasma. '° Multiple analyses on the same sample have established standard deviations of 0.76 to 1.03 for plasma samples with CEA titers from 0 to 5.0 ng/ml, of 1.12 to 2.31 for plasma samples with titers from 5.1 to 10 ng/ml, and of 2.29 to 5.47 for samples with higher titers. 1° a-Fetoprotein can be quantitated in serum samples of 25/A using the Z-gel,procedure; however, the accuracy and reproducibility of the procedure have not been studied extensively. The sensitivity and reproducibility of the radioimmunoassy with Z-gel for insulin are comparable to those for other published procedures. 3 Z-gel can be used theoretically to separate a variety of antigens from their antibody complexes provided certain conditions can be satisfied; i.e., if a pH can be found at which the antigen and its antibody complex carry opposite net charge. Z-gel is a negatively charged gel at pH 6.25; it will bind positively charged molecules, e.g., proteins at a pH below their isoelectric pH. I n the three radioimmunoassays described in this report, Z-gel at a pH of 6.25 is used to adsorb the positively charged antibody g H. J. Hansen,J. J. Snyder,E. Miller,J. P. Vandevoorde,O. N. Miller, L. R. Hines, and J. J. Burns, Human Pathol. 5, 139 (1974). 10T. M. Chu and G. Reynoso,Clin. Chem. 18, 918 (1972).
[21]
MICROFILTRATION IN RADIOIMMUNOASSAY
305
molecules (p! > 7) with their associated antigens, while leaving the negatively charged free antigens (CEA, AFP, and insulin) in solution. Experience has shown that NH4 ÷ is the preferred cation as Na ÷ and K ÷ decrease the ability of Z-gel to adsorb antigen-antibody complexes. H It should be possible to use Z-gel as a simple, convenient, and inexpensive reagent for the separation of antigen from antigen-antibody complexes in any radioimmunoassay for which the investigator can establish experimental conditions that meet the charge requirements of Z-gel, provided that nonspecific adsorption to Z-gel is not a problem and that Z-gel does not dissociate the antigen-antibody complex, as was noted in the case of angiotensin. H 11 j. p. Vandevoorde and H. J. Hansen, unpublished data.
[21] M i c r o f i l t r a t i o n a s a M e a n s o f S e p a r a t i n g Free Antigen from Antigen-Antibody Complexes in I m m u n o a s s a y B y SYLVIA R . CHALKLEY a n d A . R E N S H A W
In immunoassay techniques, efficient separation of free antigen from the antigen-antibody complexes is essential for reproducible results and maximum sensitivity. Microfiltration under suitable conditions gives efficient separation and ease of handling large or small numbers of samples. The technique of microfiltration is more adaptable to mechanization than centrifugation. Various methods of microfiltration are described here that use either disposable microfilters, 1'2 discrete filter disks, 3-5 or continuous filtration through glass fiber paper tape. ~ i S. R. Chalkley and A. Renshaw, Clin. Chim. Acta 62, 377 (1975). 2 C. J. Sanderson, G. H. Taylor, and R. Geary, J. Immunol. Methods 4, 17 (1974). a G. A. Taylor, C. J. Sanderson, R. Geary, and J. C. Meakin, Lab. Pract., August, 521 (1975). 4 j. O'Brien, S. Knight, N. A. Quick, E. H. Moore, and A. S. Platt, J. lmmunol. Methods 27, 219 (1979). s K. D. Bagshawe, Lab. Pract., September, 573 (1975). A. Pollard and C. B. Waldron, Technicon Symp. 1, 49, (1966).
METHODS IN ENZYMOLOGY, VOL. 70
Copyright © 1980 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-181970-1
[21]
MICROFILTRATION IN RADIOIMMUNOASSAY
305
molecules (p! > 7) with their associated antigens, while leaving the negatively charged free antigens (CEA, AFP, and insulin) in solution. Experience has shown that NH4 ÷ is the preferred cation as Na ÷ and K ÷ decrease the ability of Z-gel to adsorb antigen-antibody complexes. H It should be possible to use Z-gel as a simple, convenient, and inexpensive reagent for the separation of antigen from antigen-antibody complexes in any radioimmunoassay for which the investigator can establish experimental conditions that meet the charge requirements of Z-gel, provided that nonspecific adsorption to Z-gel is not a problem and that Z-gel does not dissociate the antigen-antibody complex, as was noted in the case of angiotensin. H 11 j. p. Vandevoorde and H. J. Hansen, unpublished data.
[21] M i c r o f i l t r a t i o n a s a M e a n s o f S e p a r a t i n g Free Antigen from Antigen-Antibody Complexes in I m m u n o a s s a y B y SYLVIA R . CHALKLEY a n d A . R E N S H A W
In immunoassay techniques, efficient separation of free antigen from the antigen-antibody complexes is essential for reproducible results and maximum sensitivity. Microfiltration under suitable conditions gives efficient separation and ease of handling large or small numbers of samples. The technique of microfiltration is more adaptable to mechanization than centrifugation. Various methods of microfiltration are described here that use either disposable microfilters, 1'2 discrete filter disks, 3-5 or continuous filtration through glass fiber paper tape. ~ i S. R. Chalkley and A. Renshaw, Clin. Chim. Acta 62, 377 (1975). 2 C. J. Sanderson, G. H. Taylor, and R. Geary, J. Immunol. Methods 4, 17 (1974). a G. A. Taylor, C. J. Sanderson, R. Geary, and J. C. Meakin, Lab. Pract., August, 521 (1975). 4 j. O'Brien, S. Knight, N. A. Quick, E. H. Moore, and A. S. Platt, J. lmmunol. Methods 27, 219 (1979). s K. D. Bagshawe, Lab. Pract., September, 573 (1975). A. Pollard and C. B. Waldron, Technicon Symp. 1, 49, (1966).
METHODS IN ENZYMOLOGY, VOL. 70
Copyright © 1980 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-181970-1
306
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
[21]
Materials and Methods
Filter Paper Glass fiber paper, Whatman grade GF/B (W. & R. Balston Ltd., Maidstone, Kent, U. K.) is the most suitable for use in immunoassay. This has an optimum flow rate of 10 ml/min through a disk 1 cm in diameter, and the pore size is such that all particles over 0.5/xm 3 are retained. The filter should be prewashed immediately before use with a protein solution, such as aqueous 0.25% albumin or gelatin in order to minimize protein adsorption. 10--
9
8
7
~6,.m .,m
~-5Z
ou N4-
5-
2
1
I
0
I
I
10 20 WASHING VOLUME (rnl)
FIG. 1. Effect of washing the CRC thimble, i--U, (2ZNa). From Chalkley and Renshaw. 82
i
30
Precipitate (1251); o - - o , supernatant
[21]
MICROFILTRATION
IN
RADIOIMMUNOASSAY
307
This filter paper is fragile when wet and must be supported. Glass filter paper has a high diffusion ratio giving an even flow throughout the filter matrix under negative pressure. An adequate flow rate is achieved with one hole 0.8 mm in diameter for an area 6 mm ~ and a negative pressure of 15 psi (83 kN/m2). Under these conditions the precipitate is spread evenly over the surface of the filter. When glass fiber paper tape is used, this can be strengthened by the application of a rubber solution along each edge, 6 or with an under tape of strong filter paper having coarse filtration properties. After drying, the tape must be sealed on the upper surface with transparent cellulose tape. The volume of the wash used on the filter does not appear to be critical above a minimum volume, which is 0.4 ml for a disk 1 cm in diameter (Fig. 1). If this volume is used to wash the container in which the reaction has taken place, the amount of precipitate on the filter increases, presumably owing to recovery of residual precipitate. The amount of supernatant retained on the filter decreases sharply and becomes constant at about 0.1% of the original volume, after 0.4 ml of wash has passed through the filter. Increasing the volume above 0.4 ml does not remove more supernatant, nor does it remove precipitate from the filter.
CRC Thimbles The disposable microfilters, "CRC thimbles" (W. & R. Balston Ltd., Maidstone, Kent, U. K.) shown in Fig. 2 consist of a glass fiber filter disk 1 cm in diameter with an effective filter area of 33 mm 2, sealed between two polypropylene cylinders. The outer cyclinder has a base perforated
O
2
I
I 1 cm
FIG.2. Explodedviewof CRC thimble. 1, Innercylinder;2, glassfiberfilterdisc; 3, outer cylinder with perforated base. From Chalkleyand Renshaw.~
308
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
[21]
by seven holes, which supports the glass fiber, and a groove just below the upper rim into which a corresponding projection on the inner cylinder locks. The CRC thimbles are adaptable for use in basic manual apparatus or may be incorporated into automatic equipment. Apparatus A manual microfiltration apparatus (Fig. 3) has been made by the present authors for use with a CRC thimble. The apparatus is constructed of stainless steel with neoprene rubber seals. A hollow base holds a test tube for collection of the filtrate, if required, and has a connection for a vacuum pump. The top, which fits onto the base with an O ring, has on its upper surface a well to hold the CRC thimble and a support for the removable filter cover. The cover has a probe that is inserted into the solution to be filtered. When negative pressure is applied to the vacuum connection and the CRC thimble is held in place by the filter cover, the airtight seals enable solution to pass through the probe. A prototype apparatus (Fig. 4) developed by Chalkley and Renshaw, 1 utilizing the CRC thimble will automatically filter the contents of 100 tubes in pairs. The basis of the apparatus is a box that contains tubing and a sequence programmer and holds in place electrical and mechanical com-
2
|
I
6
FIG. 3. Microfiltration apparatus. 1, Stainless steel base; 2, Stainless steel top; 3, neoprene rubber seal; 4, vacuum connection; 5, CRC thimble; 6, test tube for collection of filtrate; 7, airtight filter cover with probe; 8, airtight seating.
[21]
MICROFILTRATION IN RAD1OIMMUNOASSAY
309
FIG. 4. Automated filtration apparatus. I, Processing turntable; 2, upper sample tube turntable; 3, lower turntable holding collecting tubes for CRC thimbles; 4, vertical cylinders for loading CRC thimbles.
310
RADIOIMMUNOASSAYSAND IMMUNORADIOMETRIC ASSAYS [21]
ponents; three turntables and five delta pumps transfer liquids from the sample tubes onto the microfilters, and the others add liquid directly to the tubes or to the microfilters. The apparatus will (a) load paired microfilters onto the processing turntable; (b) prewash them with protein solution; (c) transfer the contents of the paired sample tubes on the first circular tray onto the microfilters; (d) wash both the tubes and the microfilters; (e) drop the microfilters into the tubes in the second circular tray. When the fiftieth pair of microfilters have been dropped into their tubes, the apparatus switches itself off. In cell culture techniques, the need to harvest cells from the culture medium has resulted in the development of harvesters that will filter a large number of small-volume samples. Some of these are suitable for immunoassay work. Various harvesters that can filter twelve samples at a time have been developed by Sanderson and his colleagues. 2"3 The original model transferred the contents of a microplate onto CRC thimbles, and the later versions transfer the contents of either tubes or microplates onto glass fiber filter disks. Ganged peristaltic pumps were used in both models. The microplate, or row of 12 tubes, is carried on a sliding platform, which locates positively by a ball and spring, allowing each row of wells or tubes to fit under an intake manifold. The vacuum manifold in the original version holds CRC thimbles in recesses in the top. Each recess connects with the cavity of the vacuum manifold. The manifold slides into position under the outlets from the peristaltic pump, and in this position the cavity connects to the vacuum line. In later models, the vacuum manifold consists of two separate Perspex blocks2 The lower block has a row of 12 sharpedged platforms on its upper surface, each of which is perforated by 8 holes joining the cavity connected to the vacuum line. The upper block has 12 holes, which are shaped to give an exact fit over the platforms on the lower block. For use, a strip of glass fiber filter paper is laid over the lower block and the top block is pressed down onto it, so that 12 disks are cut and clamped into place between the two blocks. A simplified version of this is manufactured by Ilacon Limited, Tonbridge, Kent, U. K. Two other filtration modules, which both formed part of automated radioimmunoassay systems, have been developed. 5,6 One of these, 6 was based on conventional continuous flow techniques. At the filtration stage, a continuously moving strip of glass fiber filter paper was strengthened and wetted, the reaction mixture was filtered, and the precipitate was washed by two streams of buffer. The strip was dried and overlaid with cellulose adhesive tape before being counted as it passed between two end-window radioactivity detectors and wound onto a take-up spool. In the second system, 5 glass fiber filter disks are mounted at intervals over perforated segments of a flexible plastic carrier tape. The contents of five
[21]
50
MICROFILTRATION IN RADIOIMMUNOASSAY
~ ~ | I II
311
MINIMUM RETENTION FOR GFB FILTER
i I I I I I I I I
._1 In,-
20 neW m
\
h-
0
.I
0.31
I
il f 0.53
I
I
0.75
I
0.97
I
I
1.19
I
I
1.41
I
I
1.63
I
I
L85
I
I
2.07
I
I
2.29
I
I
2.51
PARTICLE SIZE (~rn 3) FXG. 5. Particle size distribution after 6 hr final incubation, o - - o , Standard; n - - l l , human serum. GFB, Whatman grade GF/B glass fiber paper. From Chalkley and Renshaw. 62
reaction tubes at a time are transferred onto these disks, which have previously been wetted with protein solution. Washing solution is pumped into the reaction tubes and transferred onto the filters, which are also washed directly. A pressure plate bears on the lower surface of the filters and provides negative pressure for the transfer. When the filtration is complete, the pressure plate is lowered, and the filter tape and reaction tubes then move on five places. The filter tape is dried, overlaid with adhesive cellulose tape, and wound onto a take-up spool. When the full cycle is completed, the tape is rewound onto a supply spool, ready to be counted in a separate radioactivity detector module.
Preparation of Samples Microfiltration is ideal for use with immunoassay methods that use pre-precipitated antibody, 7 solid phase techniques where the antibody is r K. D. Bagshawe, C. E. Wilde, and A. H. Orr, Lancet 1, 1118 (1966).
312
RADIOIMMUNOASSAYS AND 1MMUNORADIOMETRIC ASSAYS
[21]
bound to Sephadex or cellulose, 8 or chelation of the first antigen-antibody complex by staphylococcal protein A instead of a second antibody. 9 In these methods the solid phase particles are all larger than the minimum retention size of the glass fiber filter. Although the coated iron particles produced by Technicon Corporation (Ardsley, New York) are intended to be separated from the mixture by magnetic methods, 1° these could be separated as easily by microfiltration. In double-antibody methods where the precipitating antibody is the last reagent to be added, the final incubation must be at least 17 hr, 1 unless accelerators have been added. H Tests have suggested, x that precipitates form at different rates in the presence or the absence of serum. This is shown by the particle size distribution found in reaction mixtures containing standard solution or human serum after an incubation period of 6 hr (Fig. 5) or 17 hr (Fig. 6). In a comparison of 94 3OrMINIMUM RETENTION FOR GFB FILTER
2o
5~-
lO
o
I 0,51
I 0.53
I 0,75
I 0.97
| L19
i 1.41
I 1.65
i 1.85
i 2.07
i 2.29
| 2.51
PARTICLE SIZE (u.m 3)
FIG. 6. Particle size distribution after 17 hr final incubation, o - - o , human serum. From Chalkley and Renshaw. 82
Standard; i - - m ,
8 L. E. M. Miles and C. N. Hales, in "'Protein and Polypeptide Hormones" (M. Margoulies, ed.), Proceedings of the International Symposium, Liege, p. 61. Excerpta Med. Found., Amsterdam, 1968. 9 M. E. Soergal, F. L. Schaffer, J. C. Sawyer, and C. M. Prato, Arch. Virol. 57, 271 (1978). 1o L. S. Hersh and S. Yaverbaum, Clin. Chim. Acta 63, 69 (1972). 11 B. Morris, personal communication, 1979.
[21]
MICROFILTRATION
IN
313
RADIOIMMUNOASSAY
eP' • °qa
o w cr n,i F-
3
L~_
2
"2
"1C~ I-/0
0
-1
•
eoen
•
-2
-5 -3
•
f
I
-2
-1
0
I
I
I
I
I
I
i
1
2
3
4
5
6
7
,l
LOG TSH (H.U/rnt) CENTRIFUGED F i G . 7. C o r r e l a t i o n o f T S H v a l u e s after 17 hr final i n c u b a t i o n . F r o m C h a i k l e y and R e n shaw.8 2
samples, 1 which were assayed in duplicate and then separated by either microfiltration or centrifugation after a final incubation of 17 hr (Fig. 7), the correlation was 0.91 (p < 0.001) and the equation of the regression line, y = 1.01 × - 0 . 3 4 does not differ significantly from the line of identity. Immunoassay methods involving adsorption of antigen by charcoal particles were found to be unsuitable for use with microfiltration techniques. The fine particles quickly block the glass fiber, with a sharp drop in the flow rate, so that variable amounts of the supernatant are retained with the precipitate.
Measurement of Radioactivity The CRC thimble and glass fiber filter disks are suitable for use in radiolabeled techniques. With y-ray counting, the microfilters are placed in containers (for example, disposable cellulose acetate tubes) and sealed
314
RADIOIMMUNOASSAYS AND IMMUNORADIOMETR1C
ASSAYS
[21]
before being measured in a well-type counter. Where fl-emitting isotopes have been used, liquid scintillation counting is most appropriate, although the glass fiber disks could be counted in a planchet counter. For liquid scintillation, the microfilters should be placed in screw-top counting vials and dried before the addition of the chosen scintillant. The polypropylene CRC thimble becomes translucent in toluene-based scintillation fluid, and tests have shown that the counts acquired are similar whether counted with the thimble intact or with the filter removed and counted separately .2 The orientation of the thimble in the vial does not significantly affect the counts. The filter tape is most conveniently counted with two end-window detectors, one on either side of the tape. Although it is possible to cut the tape into suitable lengths to be counted in a well-type counter, this is tedious and the position of each sample must be marked exactly. Conclusion Microfiltration using glass fiber filter disks is an efficient alternative to centrifugation for the separation of free antigen from antigen-antibody complexes. When small numbers of samples are involved in manual methods or for mechanized aparatus, microfiltration has advantages over centrifugation. For immunoassay techniques, cellulose acetate filters offer no advantage over glass fiber filter paper, and the slow flow rate and the handling problems of cellulose acetate when compared with glass fiber make it less suitable for use when large numbers of samples are involved. The retention characteristics of GF/B filter paper are adequate for most immunoassay techniques. However, in receptor assays, TM the particle sizes are close to the minimum retention size of the GF/B filter paper, and cellulose acetate filters are being used for filtration of these samples.
~2 p. C u a t r e c a s a s , Proc. Natl. Acad. Sci. U. S. A. 69, 318 (1972).
[22]
ANTIGEN SEPARATION BY GEL CENTRIFUGATION
315
[22] G e l C e n t r i f u g a t i o n : S e p a r a t i o n o f F r e e L a b e l e d Antigen from Antibody-Bound Labeled Antigen B y JENS LARSEN a n d WIGGO FISCHER-RASMUSSEN
Separation of protein-bound steroid from free steroid by gel filtration on stationary columns with Sephadex G-25 was elaborated by Murphy and Pattee 1 for estimation of cortisol in plasma. Giese et al. 2 developed the quicker method of gel centrifugation for separation of prelabeled antigen from antibody-bound labeled antigen in a radioimmunoassay (RIA) for angiotensin I. Before using this method of separation for the successful estimation of estrogens in serum, 3,4 we found it necessary to evaluate the completeness and the precision of the gel centrifugation. Material and Methods Solvents and R e a g e n t s
Spectrograde diethyl ether, benzene, and methanol, and analytical grade absolute ethanol were obtained from Merck AG, (Darmstadt), as were crystalline estradiol-17/3, estrone, and estriol. Other materials were human albumin fraction V (Sigma Chemical Company), Merthiolate (British Drug House), Sephadex LH-20 and Sephadex G-25 fine (AB Pharmacia, Sweden), Insta-Gel (Packard Instrument Company, U.S.A.), counting vials of polyethylene (Scintitec Laboratorieudstyr A/S, Denmark). [2,4,6,7-3H]Estrone (115 Ci/mmol), [2,4,6,7-~H]estradiol-17fl (100 Ci/mmol), and [2,4,6,7-3H]estriol (112 Ci/mmol) were purchased from New England Nuclear Corporation (Boston, Massachusetts). Human [125I]7-globulin (64/zCi/ml) was obtained from the Department of Nuclear Medicine, Rigshospitalet, Copenhagen. Antibodies. The antibody A-1 against estrone and estradiol-17/3 was raised in a sheep against estradiol-17fl coupled at C-17 to bovine serum albumin. At 50% displacement of tritiated estradiol-17/3, the cross-reac1 B. E. P. Murphy and C. J. Pattee, J. Clin. Endocrinol. 24, 919 (1964). J. Giese, M. J~rgensen, M. D. Nielsen, J. O. Lund, and O. Munck, Scand. J. Clin. Lab Invest. 26, 355 (1970). 3 j. Larsen and W. Fischer-Rasmussen, Ugeskr. Laeg. 138, 1082 (1976). 4 j. Larsen, W. Fischer-Rasmussen, P. Henger, and B. Ovlisen, Ugeskr. Laeg. 138, 1086 (1976).
METHODS IN ENZYMOLOGY, VOL. 70
Copyright © 1960by Academic Press, Inc. All rightsof reproductionin any form reserved. ISBN 0-12-18197(L1
316
RADIOIMMUNOASSAYS AND IMMUNORAD1OMETRIC ASSAYS
[22]
tivity of the antibody with estrone was 52% and with estriol 4%. By means of a Scatchard 5 plot the affinity constant for estradiol-17/3 was calculated to be 1.76 x 10TM 1/M and for estrone to be 1.00 x 10TM 1/M. The antibody A-2 specific for estradiol-17/3 was raised in a rabbit against estradiol-17/3 coupled at C-6 to human serum albumin. The cross reaction with estrone was 6%, and with estriol 0.1%. By means of a Scatchard plot the affinity constant was calculated to 5.70 × l09 1/M with respect to estradiol-17B. The antibody A-3, highly specific for estriol, was obtained from a sheep by immunization with estriol coupled with human serum albumin at C-6. The cross-reaction against estrone was less than 0.01%, and against estradiol-17/3, 0.1%. Preparation o f Materials
Gelatin buffer was prepared from glass-distilled water containing phosphate buffer (10 mM, pH 7.0), sodium chloride (0.14 M), Merthiolate (0.01%), and gelatin (0.1%). Albumin buffer was prepared by addition of human albumin (0.1%) to gelatin buffer. Preparation of Sephadex G-25 fine minicolumns was performed in the following way: The barrels of 2-ml polyethylene syringes (50 x 9 mm) were furnished with a circular filter of Whatman No. 41 paper. After swelling for at least 3 hr in surplus of gelatin buffer Sephadex G-25 fine was packed into the syringes leaving the upper 4 mm of the barrels empty. Rubber caps were placed on the nozzles of the syringes, which were stored at 2° and used within 8 days. Method in Detail
Estrone, estradiol-17/3, and estriol have been estimated in serum by means of RIA including a step of separation with Sephadex G-25 fine minicolumns. After diethyl ether extraction from 1-ml serum samples and chromatography on Sephadex LH-20 columns, estrone and estradiol-17/3 were estimated after incubation by the method of Hotchkiss et al. ~ with antibody A-1. In order to obtain quicker methods without chromatography, the more specific antibodies A-2 and A-3 have been used to estimate estradiol-17/3 and estriol, respectively. These two antibodies were incubated directly with the residues after extraction of serum with dichioromethane. G. Scatchard, Ann. N. Y. Acad. Sci. 51, 660 (1949). 6 j. Hotchkiss, L. E. Atkinson, and E. Knobil, Endocrinology 119, 177 (1971).
[22]
A N T I G E N SEPARATION BY GEL C E N T R I F U G A T I O N
317
After incubation for 1 hr at 2°, the separation of antibody-bound antigen from free antigen was performed by Sephadex G-25 fine minicolumns pretreated in the following way. The rubber caps of the cold Sephadex G-25 fine columns were removed, and the columns were coated by application of 500/zl of ice-cold albumin buffer on the top surface of the gel particles and subsequent spinning of 320 g for 5 min at 2 °. Aliquots of 600/zl of the cold incubates were transferred to the albumin-coated columns and subsequently 200 ~l of ice-cold albumin buffer were applied in order to wash down residue of the samples on the inside wall of the syringes into the gel. The columns were kept cool for 5 min before centrifugation at 2° at 320 g for another 5 min. During centrifugation the free steroid was retained in the gel particles and the fraction bound to antiserum was spun out directly into the counting vials in which the minicolumns were suspended by their wings. To each counting vial 10 ml of Insta-Gel were added, and the counting was performed in a Nuclear Chicago model Isocap/300 liquid scintillation system to a level of at least 104 counts. Results
Completeness of Separation 1. As much as possible of the antibody with the bound fraction of the labeled estrogen should pass through the Sephadex G-25 fine minicolumns. As an indicator of this passage, human [125I~-globulin in gelatin buffer was transferred to the Sephadex G-25 fine columns and treated as incubates. Two series, each with eight samples, were used with 3 ng and 150 ng of y-globulin, respectively. These amounts of y-globulin were chosen as the amount of antibody used in the assay in one vial was estimated to be 15 ng. The total counting activity was determined by transfer of identical amounts of [l~5I]y-globulin to four control vials for each of the two series. After gel centrifugation the counting activity was measured of both the column eluates and the used Sephadex G-25 fine minicolumns, which were placed directly in y-counting vials and counted in a Packard TriCarb scintillation spectrometer Model 3003. From the results in Table I it is seen that more than 97% of the radioiodinated y-globulin passed through the Sephadex G-25 minicolumns. 2. As little as possible of the free fraction of estrogen should pass through the Sephadex G-25 fine minicolumns. This was estimated by incubation of tritiated estrogen without antibody followed by centrifugation through the Sephadex G-25 fine minicolumns. As seen in Table II, 1% or less of the tritiated estrogen applied will pass through the minicolumns.
318
R A D I O I M M U N O A S S A Y S A N D I M M U N O R A D I O M E T R I C ASSAYS
[22]
TABLE I PASSAGE OF HUMAN [~25I]y-GLOBULIN THROUGH THE SEPHADEX G-25 FINE MINICOLUMNS [12~I]y-Globulin in control vials (mean cpm a - SD, n=4)
[~25I]y-Globulin in column effluent (mean cpm - SD, n=8)
Recovery in column effluent (mean % ± SD, n=8)
Rest activity in columns after use (mean % ± SD, n =8)
431,972 ± 1204 (approx. 150 ng) 11,495 - 106 (approx. 3 ng)
421,254 -- 2386
97.5 ± 0.6
1.3 - 0.3
11,197 --- 109
97.4 - 0.9
1.6 ± 0.2
cpm = counting activity as counts per minute.
T A B L E II PASSAGE OF TRITIATED ESTROGEN THROUGH THE SEPHADEX G-25 FINE MINICOLUMNS AFTER INCUBATION WITHOUT ANTIBODY A m o u n t of tritiated estrogen applied to minicolumns (mean cpm a _+ SD, n = 4)
Tritiated estrogen in column effluent (mean cpm ± SD, n = 8)
RIA including Sephadex LH-20 chromatography Estrone, 2447 ___ 42 Estradiol-17fl, 1878 - 32
16.2 - 1.8 (0.7 - 0.1%) 18.8 ± 6.5 (1.0 --- 0.4%)
RIA without Sephadex LH-20 chromatography Estradiol-17/3, 8292 - 175 Estriol, 10,338 ± 193
84.8 --_ 8.5 (1.0 --- 0.1%) 23.3 +-- 2.9 (0.2 -+ 0.03%)
a cpm = counting activity as counts per minute.
3. During the gel centrifugation of incubates with antibody it was of major importance that the fraction of tritiated estrogen bound to antibody should not be contaminated by the free fraction, which will later pass through the columns by continued elution (Fig. 1). This was investigated by repeated centrifugation of estrogen samples from standard curves at 400 and 1000 pg levels after application of 250/zl of albumin buffer. Table III shows that approximately 0.5% of the amount of tritiated estrogen in the sample was eluted. However, after repeated application of 500/zl of albumin buffer and gel centrifugation somewhat more tritiated estrogen appeared in the effluent. 4. After application of the incubates onto the minicolumns, the equilibrium obtained during incubation is changed by the stripping effect to
[22]
ANTIGEN SEPARATION BY GEL CENTRIFUGATION
319
1000. 750 500 250
ml ALBUMIN BUFFER FIG. I. Counting activity (cpm, ordinate) of minicolumn effluent and volume (milliliters, abscissa) of albumin buffer applied onto the minicolumn. Centrifugation was repeated after every application. The first peak represents antibody-bound tritiated estrogen: the second peak represents e]ution of tritiated free estrogen retained in the minicolumn.
TABLE III RESULTS OF TWO REPEATED SEPHADEX G-25 FINE CENTRIFUGATIONS WITH ALBUMIN BUFFER AFTER CENTRIFUGATION OF THE SAMPLES Albumin buffer Sample
250/zl
500 p.1
12.8 - 3.6 0.36
23.5 ± 3.9 0.67
14.0 --_ 2.2 0.46
23.3 -+ 4.9 0.76
52.2 - 4.1 0.63
286.0 -+ 20.9 3.45
51.1 ± 6.0 0.45
130.5 ± 9.0 1.16
RIA including Sephadex LH-20 chromatography Estrone samples (n = 4), 3511 cpm a, 400 pg Mean cpm ± SD Percentage Estradiol-17/] samples (n = 4), 3056 cpm, 400 pg Mean cpm - SD Percentage
RIA without Sephadex LH-20 chromatography Estradiol-17/3 samples (n = 4), 8292 cpm, 1000 pg Mean cpm - SD Percentage Estriol samples (n = 4), 11,255 cpm, 1000 pg Mean cpm - SD Percentage a cpm, counting activity as counts per minute. t h e a d v a n t a g e o f t h e f r e e f r a c t i o n o f t r i t i a t e d e s t r o g e n . T h i s is r e t a i n e d in t h e m i n i c o l u m n a n d t h e a n t i b o d y - b o u n d f r a c t i o n in t h e c o l u m n e f f l u e n t is reduced. Tables IV and V show that after contact of the incubates with the minicolumns for 5 min the stripping effect of the minicolumns during the next 5-min period will be less pronounced than after further contact.
320
RADIOIMMUNOASSAYS
AND
IMMUNORADIOMETRIC
[22]
ASSAYS
TABLE IV INFLUENCE OF THE DURATION OF CONTACT OF THE INCUBATES WITH THE M1NICOLUMNS ON THE AMOUNT OF ANTIBODY-BOUND TRITIATED ESTRONE AND ESTRADIOL-17fl PASSING THROUGH THE SEPHADEX G - 2 5 FINE MIN1COLUMNS a Estrone Time of incubate on microcolumn at 2 ° (rain) 0 5 10 15 20 30 60 120
E s t r a d i o l - 1713
0 pg, n = 4 (cpm b - SD)
100 p g , n = 4 ( c p m -+ S D )
1319.4 1169.5 1134.1 1044.5 1012.2 933.9 867.8 804.4
499.1 448.1 430.2 398.6 371.2 343.3 299.4 304.0
-+ -+ ---+ -+ -+
22.7 22.9 24.2 34.8 20.1 24.3 22.8 7.9
-+ --+ --+-
23.7 15.0 12.6 11.4 20.6 10.7 10.5 11.4
0 pg, n = 4 (cpm - SD) 1081.3 1005.3 982.1 948.9 872.0
-+ _+ _+ ---
100 p g , n = 4 ( c p m -+ S D )
28.5 23.1 18.2 16.0 25.8
8 0 2 . 2 --- 2 3 . 6 7 8 4 . 5 --- 2 4 . 8
571.6 492.5 487.1 482.9 478.1 418.7 404.1 399.8
-+ -+ -+ -+ -+ ---+
15.8 17.3 10.5 9.3 20.4 22.4 10.5 20.8
a A n t i b o d y A - 1 w a s i n c u b a t e d w i t h u n t r i t i a t e d e s t r o n e a n d e s t r a d i o l - 1 7 / 3 in a m o u n t s o f 0 p g a n d 100 p g a n d w i t h t h e s a m e t r i t i a t e d e s t r o g e n , r e s p e c t i v e l y . b cpm = counting activity as counts per minute.
TABLE V INFLUENCE OF THE DURATION OF CONTACT OF THE INCUBATES WITH THE MINICOLUMNS ON THE AMOUNT OF ANTIBODY-BOUND TRITIATED ESTRADIOL-17/3 AND ESTRIOL PASSING THROUGH THE SEPHADEX G - 2 5 FINE MINICOLUMNS a E s t r a d i o l - 17/3 Time of incubate on microcolumn a t 2 ° (rain) 0 5 10 15 20 30 60 120
0 pg, n = 4 ( c p m -+ SD) 2471.8 2400.2 2367.7 2257.6 2165.6 2107.3 2011.0 1960.1
-+ -+ + + -+ + -+ -+
33.1 45.1 48.1 14.3 43.3 53.7 32.8 33.9
Estriol
500 p g , n = 4 ( c p m --+ S D ) 418.2 397.3 391.3 385.7 362.9 347.5 328.3 327.6
+ + + + -+ -+ + -+
13.0 45.1 11.2 12.7 11.4 8.4 14.7 17.7
0 pg, n = 4 (cpm + SD) 1838.0 1764.4 1711.6 1595.5 1446.2 1237.4 1175.4 1123.2
+ + + -+ -+ -+ -+ -+
19.1 8.4 11.2 31.3 47.6 49.3 21.4 46.9
500 pg, n = 4 ( c p m --- S D ) 1131.6 1010.3 989.5 927.1 852.3 812.3 754.5 700.9
+ + + + -+ -+ -+ -+
31.4 34.7 14.9 33.3 33.9 16.0 29.1 32.5
a A n t i b o d y A - 2 w a s i n c u b a t e d w i t h u n t r i t i a t e d e s t r a d i o l - 1 7 f l a n d a n t i b o d y A - 3 w a s inc u b a t e d w i t h u n t r i t i a t e d e s t r i o l b o t h in a m o u n t s o f 0 a n d 5 0 0 p g a n d w i t h t h e s a m e tritiated estrogen, respectively. b Cpm = counting activity as counts per minute.
[22]
A N T I G E N SEPARATION BY GEL C E N T R I F U G A T I O N
321
TABLE VI PRECISION OF SEPHADEX G-25 FINE CENTRIFUGATION ESTIMATED BY APPLICATION TO 20 MINICOLUMNS OF SAMPLES FROM ONE MAJOR INCUBATION VOLUME OF EACH ESTROGEN, RESPECTIVELY
Estrogen (pg/vial)
Measured value Coefficientof (mean +- SD) variation(%)
RIA including Sephadex LH-20 chromatography Estrone, 250 Estradiol-17/3, 250
246.9 - 6.3 239.6 - 6.4
2.5 2.7
240.4 ± 7.2 945.6 - 51.5
3.0 5.4
RIA without Sephadex LH-20 chromatography Estradiol-17/3, 250 Estriol, 1000
Precision Major incubation volumes were prepared for each of the estrogen assays, and from each incubate 600-/~1 samples were applied onto 20 minicolumns that were centrifuged together. Table VI reports the values that were measured with coefficients of variation between 2.5% and 5.4%.
Practicability One hundred Sephadex G-25 fine minicolumns are prepared during half an hour. By means of the Sephadex G-25 fine gel centrifugation procedure, about 100 separations are performed in 1 hr. Comments The reliability of the separation step in RIA of estrogens by means of gel centrifugation using Sephadex G-25 fine minicolumns is demonstrated by the following observations. 1. More than 97% of [l~sI~/-globulin will pass the minicolumns. This is supposed to apply also to the antibody-bound fraction of tritiated estrogen, 2. Only 1% or less of the free fraction of tritiated estrogen will escape retention in the minicolumns, 3. The counting activity is low in the column efluent, appearing by repeated application of albumin buffer and centrifugation immediately after the passage of the antibody-bound fraction of tritiated
322
RAD1OIMMUNOASSAYS
AND
IMMUNORADIOMETRIC
ASSAYS
[23]
estrogen and before the elution of the unbound fraction retained in the minicolumns, 4. The stripping effect of the minicolumns is present only to a low degree and kept constant by a fixed period of 5 min from application of incubate before centrifugation. Abraham 7 reported that using an antibody with an affinity constant of more than 109 1/M the stripping effect should be negligible in a RIA including charcoal as a separating agent. Lindberg et al. 8 were not able to use dextran-coated charcoal in their estriol RIA because of an initial binding of only 3%, but changing of the separation step to include precipitation with ammonium sulfate gave an initial binding of 81%. It is regarded as an advantage to use only one type of separation technique for different assays in the laboratory, and the versatility of the Sephadex G-25 fine minicolumn method is demonstrated as the affinity constant of the estriol-antibody A-3 was only 5.10 x 108 1/M. The precision of separation by the minicolumns was reduced with a low affinity constant but still reasonable with a coefficient of variation of 5.4% in the estriol assay. 7 G. E. Abraham, Acta Endocrinol. (Copenhagen) Suppl. 183 (1974). 8 B. S. Lindberg, P. Lindberg, K. Martinsson, and E. D. B. Johansson, Acta Obstet. Gynecol. Scand. Suppl. 32 (1974).
[23] T h e T a l c - R e s i n - T r i c h l o r o a c e t i c Acid Test for Screening Radioiodinated Polypeptide Hormones By BARBARAB. TOWER, MORTON B. SIGEL, RUSSELL E. POLAND,
W. P. VANDERLAAN, and ROBERT T. RUBIN A rapid, reliable physicochemical method for predicting the immunologic performance of ~q-labeled hormones in radioimmunoassays (RIA) has been developed: the talc-resin-TCA test. 1 The search for an accurate physicochemical indicator of the quality of a labeled hormone has been undertaken for three decades,2-8 but most techniques have had limB. B. Tower, M. B. Sigel, R. T. Rubin, R. E. Poland, and W. P. VanderLaan, Life Sci. 23, 2183 (1978). S. A. Berson, R. S. Yalow, S. S. Schreiber, and J. Post, J. Clin. Invest. 32, 746 (1953). S. A. Berson, R. S. Yalow, A. Bauman, M. A. Rothschild, and K. Newerly, J. Clin. Invest. 35, 170 (1956). 4 p. Sonksen and S. Refetoff, in "Radioimmunoassay Methods" (K. E. Kirkham and W. M. Hunter, eds.), p. 89. Churchill-Livingstone, Edinburgh, 1971. METHODS IN ENZYMOLOGY, VOL. 70
Copyright © 1980 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-181970-1
322
RAD1OIMMUNOASSAYS
AND
IMMUNORADIOMETRIC
ASSAYS
[23]
estrogen and before the elution of the unbound fraction retained in the minicolumns, 4. The stripping effect of the minicolumns is present only to a low degree and kept constant by a fixed period of 5 min from application of incubate before centrifugation. Abraham 7 reported that using an antibody with an affinity constant of more than 109 1/M the stripping effect should be negligible in a RIA including charcoal as a separating agent. Lindberg et al. 8 were not able to use dextran-coated charcoal in their estriol RIA because of an initial binding of only 3%, but changing of the separation step to include precipitation with ammonium sulfate gave an initial binding of 81%. It is regarded as an advantage to use only one type of separation technique for different assays in the laboratory, and the versatility of the Sephadex G-25 fine minicolumn method is demonstrated as the affinity constant of the estriol-antibody A-3 was only 5.10 x 108 1/M. The precision of separation by the minicolumns was reduced with a low affinity constant but still reasonable with a coefficient of variation of 5.4% in the estriol assay. 7 G. E. Abraham, Acta Endocrinol. (Copenhagen) Suppl. 183 (1974). 8 B. S. Lindberg, P. Lindberg, K. Martinsson, and E. D. B. Johansson, Acta Obstet. Gynecol. Scand. Suppl. 32 (1974).
[23] T h e T a l c - R e s i n - T r i c h l o r o a c e t i c Acid Test for Screening Radioiodinated Polypeptide Hormones By BARBARAB. TOWER, MORTON B. SIGEL, RUSSELL E. POLAND,
W. P. VANDERLAAN, and ROBERT T. RUBIN A rapid, reliable physicochemical method for predicting the immunologic performance of ~q-labeled hormones in radioimmunoassays (RIA) has been developed: the talc-resin-TCA test. 1 The search for an accurate physicochemical indicator of the quality of a labeled hormone has been undertaken for three decades,2-8 but most techniques have had limB. B. Tower, M. B. Sigel, R. T. Rubin, R. E. Poland, and W. P. VanderLaan, Life Sci. 23, 2183 (1978). S. A. Berson, R. S. Yalow, S. S. Schreiber, and J. Post, J. Clin. Invest. 32, 746 (1953). S. A. Berson, R. S. Yalow, A. Bauman, M. A. Rothschild, and K. Newerly, J. Clin. Invest. 35, 170 (1956). 4 p. Sonksen and S. Refetoff, in "Radioimmunoassay Methods" (K. E. Kirkham and W. M. Hunter, eds.), p. 89. Churchill-Livingstone, Edinburgh, 1971. METHODS IN ENZYMOLOGY, VOL. 70
Copyright © 1980 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-181970-1
[23]
TALC--RESIN--TCA TEST FOR HORMONES
323
ited success in yielding consistently reliable results. 5'e'a'x° With the present t a l c - r e s i n - T C A test, however, we have successfully anticipated the RIA behavior of x25I-labeled human (h) and rat (r) prolactin (PRL), growth hormone (GH), thyroid-stimulating hormone (TSH), and luteinizing hormone (LH) prepared from more than 100 iodinations. Direct immunoreactive assessment of a25I-labeled hormones by the setting up of standard curves or serum standards may be costly and time consuming. Similarly, the excess antibody test may be wasteful of limited reagents and may provide only incomplete information. 5 The t a l c - r e s i n TCA test, on the other hand, provides an immediate, accurate answer as to whether one should commit valuable samples and technical time to an RIA, using a given labeled hormone preparation. The t a l c - r e s i n - T C A test is based on three different physicochemical properties of iodohormones: adsorption to talc, exclusion from resin, and precipitation by trichloroacetic acid (TCA). The test patterns are distinct for monomeric iodohormone, aggregated iodohormone, and 125I-labeled salts (or free 1251). Thus, a usable iodohormone may be identified quickly and accurately by its meeting the criteria of greater than 90% talc adsorption, less than 25% bound to resin, and greater than 90% TCA precipitation. In addition, the talc and TCA percentages should agree within 3%. Radioiodinated human and rat PRL, TSH, GH, and L H have been prepared by different iodination procedures. All purified monomeric iodohormones producing sensitive and reliable RIA results showed the characteristic talcresin-TCA test results of >90% talc adsorption, 90% TCA precipitation. T h e T a l c - R e s i n - T C A Test
Reagents: Talc tablets, 100 mg, Ormont Drug & Chemical Co., Englewood, New Jersey
s W. M. Hunter, in "Radioimmunoassay Methods" (K. E. Kirkham and W. M. Hunter, eds.), p. 1. Churchill-Livingstone, Edinburgh, 1971. 8 R. S. Yalow and S. A. Berson, Trans. N. Y. Acad. Sci. 28, 1033 (1966). 7 H. Pinto, B. L. Wajchenberg, O. Z. Higa, I. Torres de Toledo e Souza, R. S. Werner, and R. R. Pieroni, Clin. Chim. Acta 60, 125 (1975). 8 H. Pinto, A. C. Lerario, I. Torres de Toledo e Souza, B. L. Wajchenberg, E. Mattar, and R. R. Pieroni, Clin. Chim. Acta 76, 25 (1977). 9 S. A. Berson and R. S. Yalow, in "Radioisotope Techniques in the Study of Protein Metabolism," p. 29. International Atomic Energy Agency Technical Reports Series No. 45, Vienna, 1965. 1o F. C. Greenwood, in "'Principles of Competitive Protein Binding" (W. D. Odell and W. H. Daughaday, eds.), p. 288. Lippincott, Philadelphia, Pennsylvania 1971.
324
R A D I O I M M U N O A S S A Y S A N D I M M U N O R A D I O M E T R I C ASSAYS
[23]
Bio-Rad anion exchange resin, Dowex AG 1-xl0, 200-400 mesh, chloride form, was used as received. However, this resin has been discontinued by the manufacturer, but we have obtained comparable results by using the Bio-Rad AG l-x8 resin, 200-400 mesh, chloride form, as a replacement (Bio-Rad catalog No. 1401451). A plastic 1.0 ml syringe may be used to measure the resin. Trim off tip of syringe at the zero graduation. Retract the plunger to measure 0.25 ml, and tamp syringe into resin. Dispense resin into test tube by pushing clown syringe plunger (0.25 ml should equal 150 mg resin). Trichloroacetic acid (TCA), 10%, reagent grade RIA assay buffey with 1% bovine serum albumin (BSA) added (1% BSA) lZSI-labeled hormone diluted in 1% BSA to l04 cpm or the counts normally added to each assay tube Procedure
1. Prepare three 12 × 75 mm test tubes, each containing 1 ml of diluted 125I-labeled hormone. 2. To the first tube, add 1 talc tablet. 3. To the second, add 150 mg (0.25 ml) of resin. 4. To the third, add 1 ml of 10% TCA. 5. Vortex each tube; walls of the talc and resin tubes may be washed down with an additional 1 ml of 1% BSA. Centrifuge 10-20 min at 2000-2500 rpm. 7. Carefully decant supernatants into clean 12 × 75 mm tubes. 8. Count the supernatant (super.) and precipitate (ppt) for each of the three tests. . Calculate the percentage of precipitate for each test as .
% ppt =
ppt × 100 ppt + super.
General Considerations
As mentioned, optimal RIA results are obtained from monomeric iodohormones showing greater than 90% adsorbed to talc, less than 25% bound to resin, and greater than 90% precipitated by TCA, with the talc and TCA results within 3% agreement. Any change in these physicochemical parameters indicates a change in the structure of the iodohormone. Since a relatively minor change in the structure of a polypeptide hormone may lead to drastically altered immunoreactivity, it follows that rigidly adhering to the optimum t a l c - r e s i n - T C A test guidelines to screen
[23]
TALC--RESIN--TCA TEST FOR HORMONES
325
the iodohormones before use will eliminate a major cause of unreliable assays. Talc will adsorb the iodohormone, and it has been used in RIA to separate bound from free hormone 1' and to purify iodohormones. TM Hormonal aggregates will easily dissociate from talc, ~3 and thus more than 90% of the labeled hormone should remain adsorbed to talc. Less than 90% talc adsorption indicates the formation of either free iodide or basic aggregates. Since acidic aggregates also may adsorb to talc, the talc and TCA percentages should agree within 3%. The Dowex AG 1-xl0 ion exchange resin used in our test has the capacity to remove both acidic aggregates and free [~25I]iodide salts. This resin has been used to monitor the iodination reaction 14-17 and to remove free iodide formed during storage of the purified labeled material/s TCA precipitation of protein has been the standard test used to monitor the quality of radiolabeled material for in vivo studies 2,3 and for use in RIA. TM The nonprecipitable material can be either free iodide or acid-soluble peptide fragments. In either case, material showing less than 90% TCA precipitability will have a markedly decreased immunoreactivity. We again stress that for optimal RIA results the TCA and talc percentages should agree within 3%.
Results
The glucose oxidase-lactoperoxidase (GO-LPO) iodination method of Tower et al. 2° was used to produce [~25I]hPRL. After purification on a column of Sephadex G-100, the major peaks of iodinated material were screened for use in RIA by the t a l c - r e s i n - T C A test (Fig. 1). Peak III material obtained from the Sephadex G-100 gel filtration purification was the 11 G. Rosselin, R. Assan, R. S. Yalow, and S. A. Berson, Nature (London) 212, 355 (1966). 1~ R. S. Yalow and S. A. Berson, Nature (London) 212, 357 (1966). 13 p. Cuatrecasas and M. D. Hollenberg, Biochem. Biophys. Res. Commun. 62, 31 (1975). 14 A. S. McFarlane, in "Radioisotope Techniques in the Study of Protein Metabolism," p. 1. International Atomic Energy Agency Technical Reports Series No. 45, Vienna, 1965. 1~ T. Chard, M. J. Kitau, and J. Landon, J. Endocrinol. 46, 269 (1970). le T. Goodfriend, in "Protein and Polypeptide Hormones" (M. Margoulies, ed.), p. 624. Excerpta Med. Found., Amsterdam, 1969. 17 R. J. Lefkowitz, J. Roth, W. Pricer, and I. Pastan, Proc. Natl. Acad. Sci. U. S. A. 65, 745 (1970). is M. A. Lesniak, J. Roth, P. Gordon, and J. R. Gavin, Nature (London), New Biol. 241, 20 (1973). 19 A. Von Zur Muhlen, in "Radioimmunoassay Methods" (K. E. Kirkham and W. M. Hunter, eds.), p. 83. Churchill-Livingstone, Edinburgh, 1971. s0 B. B. Tower, B. R. Clark, and R. T. Rubin, Life Sci. 21, 959 (1977).
PERCENT OF ADDED COUNTS PRECIPITATED Talc Resin TCA "X'optimol >90---% 90% Peak [ 74 Peak ]] 44 47 95 *Peak l"rT 98 22 34 Peak ~. 35 94
"~
5 --
]~ A
4 -
/ I I
/ /
x Q. 2
20
50
40
50 60 70 FRACTION ( 2 m l )
80
90
FIG. 1. Gel filtration profile of ~=SI-labeled human prolactin, p=5I]hPRL, (VLS No. 4) iodinated by the glucose oxidase-lactoperoxidase ( G O - L P O ) method and purified on a 2 x 50 cm Sephadex G-100 column. Elution solvent was 50 mM phosphate buffer, pH 7. l, containing 0.15 M NaCl, 0.1% sodium azide, and 0.1% crystalline bovine serum albumin. After chromatography, fractions of the major peaks were checked with the talc-resin-trichIoroacetic acid (TCA) test. Peak III, the material suitable for use in radioimmunoassay, showed >90% talc adsorption, 90% TCA precipitation. Peaks I and II are aggregated hormone, and peak IV is dissociated l=SI-labeled salts and peptid¢ fragments (Tower et al.1).
PERCENTOF ADDED COUNTS PRECIPITATED Talc Resin TCA ~oplimol >9o~ 90% 94
Tn'
5-
95
4x ~3a. 2-
20
I
30
I
40
I
I
I
50 60 70 FRACTION ( I ml)
I
80
I
90
FIG. 2. Gel filtration profile o f ~25I-labeled human prolactin, [I=S[]hPRL ( V L S N o . 3) iodinated by the chloramine-T method and purified on a 1.5 x 30 cm Sephadex G-100 column, precoated with b o v i n e serum albumin. E l u t i o n solvent was 10 m M phosphate buffer, p H 7.5, containing 0.15 M N a C I and 1 : 10,000 M e r t h i o l a t e . The labeled h o r m o n e was purified just p r i o r to use. The t a i c - r e s i n - t r i c h l o r o a c e t i c acid ( T C A ) test results showed that peak ] I [ material was suitable f o r use in radioimmunoassay ( T o w e r et al.t).
[23]
TALC--RESIN--TCA TEST FOR HORMONES
327
TALC-RESlN-TRICHLOROACETIC ACID (TCA) TEST RESULTS FOR 125-I-LABELED HORMONES PREPARED BY THE GLUCOSE OXIDASE--LACTOPEROXIDASE(GO-LPO) IODINAT1ON METHOD Peak III 125I-hormone preparation (NIH batch no.) a hGH (HS1394) hTSH (AFP 1) hPRL (VLS 4) hPRL (AFP 1) hLH (LER960) rGH (I-2) rGH (1-3) rTSH (1-4) rPRL (1-2) rPRL (1-3) rLH (1-4)
Optimal:
Talc
Resin
>90%
90%
97 95 98 96 93 96 97 98 90 93 96
3 20 22 19 11 13 15 9 22 18 13
96 94 95 96 91 97 98 98 93 94 93
a Peak III material from 2 x 50 cm G-100 gel filtration column, as in Fig. 1. Hormone preparations were obtained from the National Pituitary Agency, through NIAMDD or Dr. A. F. Parlow.
usable monomeric iodohormone which produced reliable radioimmunoassays. The chloramine-T iodination method of Hunter and Greenwood 21 also was used to produce [125I]hPRL, as shown in Fig. 2. Again, peak III of the G-100 column purification yielded talc-resin-TCA results within the defined limits of usable material, which was verified by reliable RIA results. 22 Iodo-hPRL prepared by either iodination method thus will yield reliable RIA results if the talc-resin-TCA test shows >90% adsorbed to talc, 90% precipitated by TCA (agreeing with talc -+3%). The G O - L P O iodination and G-100 gel filtration purification of human and rat PRL, TSH, LH and GH all yielded results similar to those shown in Fig. 1. The peak III iodohormone that gave reliable assay results always fell within the aforementioned limits when screened by the talc-resin-TCA test (see the table). Discussion
Storage Labeled Hormone. Two-milliliter fractions of the peak III material prepared by the G O - L P O iodination method (Fig. 1) should be stored at 21 W. M. Hunter and F. C. Greenwood, Nature (London) 194, 495 (1962). ~ Y. N. Sinha, F. W. Selby, U. J. Lewis, and W. P. VanderLaan, J. Clin. Endocrinol. Metab. 36, 509 (1973).
328
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS %
n7
PERCENT OF ADDED COUNTS PRECIPITATED Talc Resin TCA ~optimal >90~/. < 2 5 % >90%
54
[23]
Peak I Peak ~
58 43
56 47
92 75
~PeakTFr Peok'rV
90 33
22 95
93 (2
o_ t.) 2 1-
20
30
40
50
60
70
80
90
FRACTION (2ml)
FIG. 3. Elution pattern of nSl-labeled rat prolactin, [12~I]rPRL (AFP 1-2), prepared by the glucose oxidase-lactoperoxidase iodination procedure and purified as described in Fig. I for human PRL. The unlabeled hormone was stored as a dry powder, desiccated, and weighed out just prior to iodination; this resulted in a maximum yield of peak III material. The peak III monomeric hormone is suitable for use in radioimmunoassay (cf. Fig. 4) (Tower et al.~).
- 20 ° with 50 mg of resin added to each tube. Extra BSA may be added for protection; 0.1 ml of 0,5% BSA in phosphate-buffered saline, for example, may be added to each tube. The radiolysis of aqueous solutions produces radicals that attack the iodohormones, and BSA is the traditional scavenger added to counteract this process. 6 A note of caution about the BSA used with iodohormones: some lots of BSA contain substances, believed to be enzymes, that will degrade the labeled preparation33 This degradation will be apparent by an increased deiodination rate, an increase in aggregate formation, or a combination of both. Careful appraisal of each lot of crystalline BSA will eliminate the "bad" lots. Unlabeled Hormone. Rat prolactin (AFP 1-2) was labeled by the G O LPO method and purified by gel filtration on Sephadex G-100 (Fig. 3). This material resulted in a high yield of peak III material. The rPRL was stored as received from Dr. A. F. Parlow, through the NIAMDD rat hormone distribution program. The rPRL was weighed out and diluted just before iodination. Contrast this storage procedure with the results shown in Fig. 4. Here the same rPRL 1-2 preparation was diluted in PO4 buffer, aliquoted, and frozen before iodination. After 1 month of storage, an ali=a A. E. Freedlender, in "Protein and Polypeptide Hormones '' (M. Margoulies, ed.), p. 608. Excerpta Med. Found., Amsterdam, 1969.
[23]
TALC--RESIN--TCA TEST FOR HORMONES
5-
x
PERCENT OF ADDED COUNTS PRECIPITATED Talc Resin TCA %ptimal >90~, 90% Peak I 60 53 92 Peak ~T 63 55 91
I 1"1" ~ A /
%
/
I
~
4-
329
Peok 177"
82
60
88
t
3-
13_ 2I-
20
i
i
50
40
50
60
i
i
~
70
80
90
FRACTION (2ml}
FIG. 4. Elution pattern of l=5I-labeled rat prolactin, [125I]rPRL (AFP I-2), prepared by the glucose oxidase-lactoperoxidase iodination method and purified as described in Fig. 1 for human PRL. Before this material was labeled, it had been diluted in phosphate buffer to 100/~g/ml, aliquoted into separate reaction vials, and stored at - 2 & for I month. During storage, the unlabeled hormone apparently converted to aggregate, as the talc-resin-(TCA) test indicated that no usable monomeric hormone remained available for iodination (cf. Fig. 3) (Tower et al.1).
quot was defrosted and iodinated by the G O - L P O method. Most of peak III had converted to the unusable, aggregated peak I or peak II material, and the t a l c - r e s i n - T C A test results showed that no material obtained from this iodination could be used for RIA. This was confirmed with an antibody binding check on the peak III material; less than 10% of the expected immunoreactivity remained. Therefore, in order to obtain a maximum yield from the valuable pituitary hormone preparations, they must be stored as a dry, lyophilized powder, in a desiccator, at 4°. The preparation should be weighed out just before use and diluted immediately before iodination with cold phosphate buffer. The phosphate buffer must not contain azide or any other enzyme inhibitors, as they will completely inhibit the G O - L P O iodination reaction. The shelf life of any unlabeled hormone preparation is not known. Interbatch differences in the quality of unlabeled human PRL preparations have been noted. ~a Some materials will produce reliable RIAs for years, whereas others will show a rapid decline in immunoreactivity and an increase in aggregate formation. The careful monitoring of the behavior of the labeled hormone by the t a l c - r e s i n - T C A test will detect subtle changes in the hormone preparations. ~4 V. Fang, J. Armstrong, and I. G. Worsley, Clin. Chem. 24, 941 (1978).
330
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC
ASSAYS
[23]
125Ifor Iodination Interbatch differences in the quality of the radioiodine purchased from commercial suppliers have been apparent, zS-2r The use o f " b a d " iodine can result in several different problems, including a low incorporation rate of iodine and an increased formation of aggregated iodohormone either during the iodination reaction or upon storage. We have requested Union Carbide Corp. (Tuxedo Park, New York) to supply us with Na~25I at a concentration of 100 cubic feet per minute draw, in which the iodination reaction is performed. Pasteur pipettes for transfer of iodination reaction mixture to Sephadex G-50M column. Plastic-backed paper to line the floor of the hood. Several layers should be used and removed, one layer at a time, to be discarded in the radioactive waste if spillage occurs during the procedure. Protocol sheets, prepared before the procedure and available for reference during the iodination. Procedure. Add, in order, to the reaction tube: 0.3-0.5 mCi of 1251; 40/zl of phosphate buffer; 10/~1 of LPO, 100/zg/ml in phosphate buffer; 40/.d (4-7/zg) of hormone; 10-20/zl of GO, 10/~g/ml or 100/Lg/ml in phosphate buffer; and 20 ~1 of 1% fl-o-(+)glucose in H~O. Wait 5-30 min, then stop the reaction by adding phosphate buffer containing 0.1% azide. Transfer the iodination reaction to a Sephadex G50 column and wash the reaction vial with 5% BSA-phosphate buffer, transfering the washes to the column. Start the column flow after all the washes and transfers have been completed. Collect 0.4-ml fractions in each of 20 tubes. Count each fraction for 0.1 min. Save the first (protein) peak, and dis-
332
RADIO|MMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
[23]
card the second (1251) peak into the radioactive waste. The first peak must be purified on a 2 x 50 cm Sephadex G-100 column, or by other purification methods of your choice. Clean up the hood and monitor the area for contamination promptly upon completion of the iodination. We have been impressed with the favorable radiation safety aspects of the GO-LPO iodination technique. Since we have been using this technique, personnel monitoring for radiation contamination both by bioassay (thyroid counts) and by dosimeter readings have not been above background (NAB). Discussion. Iodohormones prepared by the G O - L P O iodination method are the most stable labeled preparations we have used. Standard curves obtained from a purified hPRL preparation (VLS No. 4) on the day of iodination and after 3 months of storage at - 20° are illustrated in Fig. 5. When stored properly, the stable G O - L P O iodohormones should give reliable, reproducible RIA results without further purification. However, chloramine-T-labeled hormones must be rechromatographed after storage, just prior to use. Regular monitoring of the quality of stored iodohormones before use in an RIA is easily accomplished with the talcresin-TCA screening test.
Exceptions The classical parameters of >90% talc adsorption and TCA precipitation, with ---3% agreement between the talc and TCA values, and 85%), and specific activity (100 _+ 10 Ci/mmol) of the product. This reagent labels e-amino groups of lysine residues. Protein A has 52 lysines, and introduction of > 1 atom of 1251per PA molecule does not appear to affect adversely the functional activity. 7 Unless otherwise indicated, [lZSI]PA prepared by this method was used in the experiments described in this report. Chloramine- T Method ~s'z2 Reagents
Buffer: sodium phosphate 0.15 M, pH 7.5 NalZSI (New England Nuclear or Amersham/Searle), 1 mCi PA, 50/zg (50 p,l of 1 mg/ml buffer) Chloramine-T, 4/.~g (5/zl of 0.8 mg/ml buffer) Sodium metabisulfite, 5/zg (5/zl of 1 mg/ml buffer) Ovalbumin, 5%, 0.25 ml in 0.15 M phosphate, pH 7.1 Procedure. To the Na125I add the PA, followed by chloramine-T. Shake the reaction mixture for 1 min at 23 °, then add sodium metabisulfite followed by ovalbumin. [a25I]PA is isolated by chromatography on a column (1.5 × 30 cm) of Sephadex G-25. Although the authors 1~ suggest that the product be stored at - 2 0 °, precipitation and inactivation can occur under these conditions. Storage at 4° in 0.04% sodium azide is recommended, r Bolton-Hunter Reagent n,zz Reagents
Buffer: sodium phosphate, 50 mM, pH 8.0 125I-labeled Bolton-Hunter reagent, 0.2 mCi in benzene-0.2% dimethylformamide (Amersham/Searle, > 1400 Ci/.mmol) PA, 50/~g (100/~1 of 0.5 mg/ml buffer) Glycine, 100/zg (100/zl of 1 mg/ml buffer) Procedure. Evaporate the solution of Bolton-Hunter reagent to dryness under a gentle stream of dry nitrogen or air. Add the PA and shake the mixture occasionally during 15 min at 23 °. Add the glycine and incubate the mixture for an additional 30 min. [125I]PA is isolated by chromatography on a column (1.5 x 30 cm) of Sephadex G-25 wet-packed and eluted with VBS-gel (Veronal-buffered isotonic saline containing 0.15 mM Ca 2÷, 1 mM Mg ~+, and 0.1% gelatin, pH 7.2). Fractions of 12 ml are collected, and the pooled product is stored at 4 ° in the presence of 0.04% sodium azide. The specific activity is reproducibly 100 ___ 10 Ci/mmol, and the product is functionally stable for at least 8 weeks.
[25]
125I-LABELED PROTEIN A
359
When 2/~g of PA were iodinated under similar conditions, the specific activity ranged between 5000 and 10,000 C i / m m o l ) T M Tritiation Procedures
General 3H-Labeled PA has been prepared by two different methods. 2°,21 These reagents have not been widely tested, but reportedly are functionally stable. They have been used to detect antibody bound to antigen on the cell surface 2° or immobilized to a solid support. 2' Since [3H]PA may be a useful alternative to [12sI]PA, the syntheses are included here.
[3H]Acetyl-PA Z° Reagents Buffer: sodium phosphate, 0.3 M pH 7.2 25 mCi [3H]acetic anhydride (Amersham; 10 txmol, 2.5 Ci/mmol) PA, 5 mg in 0.45 ml of buffer Procedure. Add the PA solution to [3H]acetic anhydride and allow the mixture to incubate at room temperature for 2 hr. Collect 3H-labeled acetyl-PA by chromatography on a column (1.5 × 25 cm) of Sephadex G-25 fine by elution with 10 mM sodium phosphate, pH 7.2. Collect 0.5 ml fractions at a flow rate of 10 ml/hr. The specific activity is approximately 25 Ci/mmol. This procedure should primarily label ~-amino groups of lysine residues. [3H]PA by Reductive Methylation 21 Reagents Buffer: borate, 0.20 M pH 9.0 Formaldehyde: 1.8 mg (62 tzmol); 0.5 ml of 3.7% aqueous solution PA, 10 mg in 1.0 ml of buffer NaB3H4 (specific activity 60 Ci/mmol): 15/zmol as 0.3 ml of 3 Ci/ml solution in 0.1 M sodium hydroxide. Use immediately after preparation. Procedure. Add formaldehyde solution to PA followed by NaB3H4. After the mixture has stood at 4° for 30 min, [zH]PA is isolated by chromatography on Sephadex G-25M that was prewashed with 15% BSA and PBS, pH 7.4. The authors 21 mix the product with 10% fetal calf serum or 2.5% BSA and store aliquots in liquid nitrogen. Thawed samples report24L. Levine, I. Alam, and J. J. Langone,Prostagl. Med. 2, 177 (1979).
360
RADIOIMMUNOASSAYSAND IMMUNORADIOMETRIC ASSAYS [25]
edly are stable for over a month at 4 °. E-amino groups of lysine residues are methylated by this procedure. Quantitation of Fluid-Phase IgG Using ['25I]PA 7 General Considerations
Two properties of [x25I]PA make it a useful analytical reagent: specificity and the ability to react with antibody without inhibiting antigen-antibody binding. Although PA reacts mainly with IgG, the specificity is not absolute either in terms of Ig class, subclass, or species. 4 However, for practical purposes, the reaction with IgG is the important one, since IgG generally is the principal class of antibody produced against antigens and haptens in hyperimmunized animals. Principle
The principle of the method is illustrated in Eq. (1). RalgGb - *PA RalgGb + IgGf+ *PA ~ IgGf- *PA
(1)
Fluid-phase IgG (IgGf) and rabbit IgG bound covalently to polyacrylamide beads (RaIgGb) compete for a limited amount of [125I]PA (*PA). Binding of [x2sI]PA is determined in the absence of standard IgGf. The degree of competition, calculated as percentage of inhibition of maximum binding, is plotted as a function of IgGf added. This standard curve is used to determine the concentration of IgG in a test sample based on the observed percentage of inhibition. All rabbit IgG reportedly binds to PA, ze so rabbit IgG is a useful substrate for comparing the relative reactivity of IgG from different species. Procedure Optimal A m o u n t s o f Reagents. The functional purity of [125I]PA and the optimal amounts of beads and tracer to use for routine assay of fluidphase (PA reactive) IgG are determined from the binding curves shown in Fig. 1. Increasing amounts of beads (0.1 ml; 5-200/~g of beads corresponding to 15-600 ng of rabbit IgG) are incubated for 60 min at 30° with [125I]PA (0.1 ml). The beads are washed with two 3-ml portions of buffer by centrifugation at 1500 g (4°) for 5 min or by filtration on polycarbonate filters, and the radioactivity in the bead pellets is determined. In this ex-
~5I. Alam, J. J. Langone,and L. Levine,Prostagl. Med. 2, 167 (1979). 26I. Lind, I. Live, and B. Mansa,Acta Pathol. Microbiol. Scand. Sect. B 80, 702 (1970).
[25]
lzsI-LABELED PROTEIN A i
24 22 20
361
,
r
cpm Added: 28,300_.....t][
1 /ag Beads= 3 ng IgG ~
~
j f
x 12
14,200
E 6
~
7,300___________--
4 2 "
' ; 40 2's
5'0
160
26o
Rabbit IgG BeadsAdded (pg) FIG. 1. Binding of l=5I-labeled protein A (PA) to immobilized rabbit IgG. Increasing amounts of beads (0.1 ml; corresponding to 15-600 ng of IgG) were incubated for I hr at 30° with either 28,1)01) (0-----0), 14,2111)(O---O), or 7301) cpm (&--&) of [~sI]PA. The beads were washed with two 3-ml portions of buffer, and radioactivity in the bead pellets was determined.
ample the functional purity is approximately 85%, based on maximum uptake of radiolabel. For routine work it is convenient to use an amount of beads that will bind approximately 10,000-15,000 cpm. Assay oflgG. The ability of IgG from different species to inhibit the binding of [~25I]PA to immobilized rabbit IgG was tested with concentrations of beads and tracer established from Fig. 1. A representative protocol is shown in Table I. In addition to determination of maximum binding (mixture 7) and inhibition by different amounts of IgG (mixtures 1-6), control samples include binding of [~25I]PA in the presence of the highest concentration of test sample and in buffer alone, both with no beads present (mixtures 8 and 9). These controls normally are 250-300 cpm out of 40,000 cpm added. Typical inhibition curves are shown in Fig. 2 for rabbit, swine, mouse, and rat IgG. The amount of IgG required to inhibit binding by 50% can be used to compare the relative specificity of [~25I]PA. These results are summarized in Table II for the 12 species tested. Reactivity of IgG ranges over a factor greater than 103 and agrees with the available data on PA specificity .4 This procedure, in which dilutions of test sample are analyzed along with standard IgG, has been used to determine the concentration of IgG in normal human, rabbit, and guinea pig serum. Levels shown in Table III agree well with available reported values.
ir~--
~l
I~ o
~
I~
~rj
~
~J
=
~
z
z L~
E
~ !
~
-=
[25]
lzsI-LABELED PROTEIN A I
]
363 I
100 L9 z Q z 80
RABB,T
_/
MOOSE
en
60 ii
o z
_o 4O t~
z -
20
10
100
1,000 NANOGRAMS ]gG ADDED
10,000
100,000
FXG. 2. Inhibition of nSl-labcled protein A (PA) binding to 20/~g of rabbit IgG Immunobeads by differing amounts of rabbit ( H ) , swine (O--©), mouse (A--&), and rat (A--A) IgG. Out of 37,000 cpm of [~25I]PA added, approximately 10,800 cpm were bound. From Langone. 27
TABLE II INHIBITION OF BINDING OF lsSI-LABELED PROTEIN A TO IMMOBILIZED RABBIT IGG BY IGG FROM DIFFERENT SPECIESa
Species
IgG required to inhibit by 50% (ng)
Rabbit Human Guinea pig Pig Dog Cow Mouse Horse Sheep Goat Rat Chicken
60 60 60 135 290 3,000 4,500 5,000 40,0O0 > 100,00& > 100,000c > 100,00(F
a Reproduced from Langone27 with permission. b Inhibition at this level: 45%. c Inhibition at this level: < 1 5 % .
364
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS [25] TABLE III LEVELS OF I G G IN NORMAL SERA
Species Sample
Human
1
10.5 --- 0.5 9.0 ± 0.5 7.3 ± 0.3 8.9 ± 0.8 10.3 ±- 0.5
2 3 4 5
Rabbit 3.8 5.2 4.3 4.7 5.3
-+ 0.1 ± 0.1 ± 0.1 ± 0.3 ± 0.2
Guinea pig (strain 2)a 3.3 2.0 5.9 5.9 5.9
--- 0.3 ± 0.1 ± 0.1 ± 0.4 ± 0.1
a Langone et al. r Immunoassay
o f F l u i d - P h a s e A n t i g e n s a n d H a p t e n s 2r
Principle
125I-labeled protein A as a t r a c e r for I g G has b e e n e x t e n d e d to a general i m m u n o a s s a y m e t h o d f o r fluid-phase antigen and hapten. T h e steps i n v o l v e d are s u m m a r i z e d in Eqs. (2) and (3). AB + (Lb + Lf) ~ (Ab - Lb) + (Ab - Lf)
(2)
Ab - Lb + *PA ~ Ab - Lb -- *PA
(3)
I m m o b i l i z e d (Lb) and free (Lf) ligand c o m p e t e for I g G a n t i b o d y binding sites [Eq. (2)]. T h e a m o u n t o f Lf p r e s e n t will determine the a m o u n t o f a n t i b o d y b o u n d to Lb, After the b e a d s are w a s h e d , (excess) [125I]PA is a d d e d and a s e c o n d i n c u b a t i o n carried out. T h e a m o u n t o f [12sI]PA b o u n d is a quantitative m e a s u r e o f a n t i b o d y , and indirectly o f Lf. A s t a n d a r d inhibition c u r v e is o b t a i n e d u n d e r optimal a s s a y conditions using k n o w n a m o u n t s o f h o m o l o g o u s Lf and used to relate o b s e r v e d inhibition to conc e n t r a t i o n o f Lf in test samples. Immobilized
L i g a n d s 27
L i g a n d s are c o u p l e d by amide b o n d s to p o l y a c r y l a m i d e b e a d s that h a v e free c a r b o x y l ( I m m u n o b e a d s ) o r a m i n o g r o u p s (Affi-Gel 701; BioRad). Reagents
Buffer: s o d i u m p h o s p h a t e , 3 m M , p H 6.35 Ligand: 0 . 1 - 1 . 0 mg o f protein o r 0 . 1 - 0 . 3 m g o f h a p t e n dissolved in 0.5 ml o f buffer 27 j. j. Langone, J. Immunol. Methods 24, 269 (1978).
[25]
1251-LABELED PROTEIN A
365
Beads: 100 mg of Immunobeads or 500 mg of Affi-Gel 701 washed and suspended in 9.0 ml of buffer 1-Ethyl-3-(3-dimethylaminopropyi)carbodiimide (EDAC), 6 mg (0.3 ml of 20 mg/ml buffer) Procedure. Mix the ligand and bead suspension at 4 °, then add EDAC. Allow the mixture to rock at 4° for 4-25 hr. In the cold, wash the beads with three 20-ml portions of coupling buffer, three 20-ml portions of 5 M guanidine hydrochloride, pH 7.2, then several times with phosphate-buffered saline, pH 7.2. After the beads have stood in this last buffer at 4 ° for 2 hr, they are washed twice with 20 ml of VBS-gel and resuspended in 25 ml of this buffer containing 0.04% sodium azide and stored at 4 °. Generally, beads are stable for several months and can be prepared with reproducible activity.
General Assay Procedure Optimal Bead and Antibody Concentrations. Working concentrations are determined from curves similar to those shown in Fig. 1. Reagents Beads, serially diluted (0.1 ml) Antibody--either diluted whole serum or IgG fraction (0.1 ml) [125I]PA, 40,000-50,000 cpm (0.1 ml) Procedure. Add antibody to increasing amounts of beads and incubate the mixture at 30° for 60 min. Wash with two 3-ml portions of buffer. Add [125I]PA, carry out a similar incubation and washing procedure, and determine the number of counts per minute bound. Controls include a set of tubes containing increments of beads with no antibody (0.1 ml buffer), antibody alone plus buffer (0.1 ml), and buffer alone (0.2 ml). Typical binding curves shown in Fig. 3 were obtained by incubating increasing amounts of immobilized human chorionic gonadotropin (HCG) (2.5-100 ~g of beads) and rabbit antiserum diluted either z~, 5r~o, 2~o, or 6r-~. Note that binding of [125I]PA to the beads in the absence of antibody is insignificant. Excess [125I]PA. At the concentrations of antibody and immobilized ligand used, sufficient [125I]PA must be added to saturate the receptor sites on the bound IgG. A set of tubes containing replicate samples of antibodycoated beads are prepared by the procedure described above. Increasing amounts of [~2H]pA (0.1 ml) are added to the bead pellets, and the mixtures are incubated at 30 ° for 60 min. The beads are washed, and bound radioactivity is determined. Typical binding curves are shown in Fig. 4 for four sets of anti-HCG (25-~) coated-HCG beads that were treated with [~25I]PA ranging between 2900 and 100,000 cpm. When 40/~g of beads were used, excess [~25I]PA was not present, even when 100,000 cpm was
366
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
28
I
I
[25]
I
ANTI-HCG 1/250 24
,.;'- 20 _
~,
O
1/750
_
II1
1/2,250 8
4 J-]p" ~ 0 IIr~ 1 2
1/6.750 e~ 5
t5 ~,, NO ANTIBODY 10 20 RELATIVE CONCENTRATION OF IMMOBILIZED HCG
~d] 40
FIG. 3. Binding of l=5I-labeled protein A (PA) (0.1 ml, 38,000 cpm added) to differing amounts of human choriomic gonadotropin (HCG) beads that were treated with rabbit antiHCG diluted either rk~ (O---Q), r ~ (©--©), r ~ (A--A), or ~ (&--&). Binding of [12sI]PA to beads in the absence of antibody (I-q--l-q) was also determined. The beads were incubated with antibody for 60 min at 30°, washed, and then incubated with [nsI]PA for 60 min at 30° before the number of counts per minute bound was determined. A relative concentration of 1 = 2.5/~g of HCG beads. From Langone. 27
added. However, with 13/zg of beads, approximately 35,000 cpm was sufficient to saturate the Fc antibody binding sites. As less beads were used, less [125I]PA was required to reach a saturating dose as the maximum binding decreased. Standard Inhibition Curves: Sensitivity and Specificity. The ability of homologous ligand to inhibit antibody binding is measured as inhibition of [125I]PA binding. Table IV shows a sample protocol including the necessary controls. In the initial reaction [Eq. (2)] beads (0.1 ml) and antibody (0.1 ml) are incubated at 30° for 60 min in the presence of different amounts of inhibitor (0.1 ml) or in buffer to determine maximum binding (tubes 1-9). The beads are washed twice, then incubated again with [x25I]PA (40,000 cpm, 0.1 ml) [Eq. (3)]. Radioactivity bound to the beads is determined, and inhibition curves are plotted. Typical results shown in Fig. 5A were obtained using a constant amount of HCG beads (13 p.g) and
[25]
]25I-LABELED PROTEIN A 50
i
40; z2;
i
367 i
!
i
~
---, 20 X
18
a Z 0
16
r~ < r~ Z
.~=~ °~
[25]
lgSI-LABELED PROTEIN A 100
I
A 80
I
ANTI-HCG • 11250 o 1/750
369
I
JA~/"
50
40 Z
z
~u
o z
20
100
B
k-
~:
T
0
80
I
I
l
I
I
i
..]
HCG BEADS (/~g) •
13
50
20
1.0
10
I
I
100
1,000
-3 NANOGRAMS HCG ADDED (= 1U X 10 )
FIG. 5. Inhibition of the binding or rabbit anti-HCG to human chorionic gonadotropin (HCG) beads by differing amounts of HCG as measured by inhibition of [12M]PA binding. (A) Effect of varying antibody concentration on assay sensitivity. Inhibition curves were obtained using 13/~g of HCG beads and anti-HCG diluted ~ ( H ; 14,200 cpm bound), r ~ (O---O; 8000 cpm bound), or r,~n (L~--/x; 4250 cpm bound). (B) Effect of varying bead concentration on assay sensitivity. Inhibition curves obtained using anti-HCG diluted ~-~ and either 40/~g ((3---0; 23,500 cpm bound), 13/~g ( H ; 14,200 cpm bound), or 4.3 p.g (/x--/x; 6000 cpm bound) of HCG beads. In each case, 38,000 cpm of [1=sI]PA were added. From Langone.27
dilution) serum from a species with PA-reactive IgG (Table II) is analyzed. To avoid this problem, samples are transferred to new tubes before counting. Alternatively, IgG can be removed by absorption of the sample with PA-Sepharose. zs ~s j. j. Langone, M. D. P. Boyle, and T. Borsos, Anal. Biochem. 93, 207 (1979).
370
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
[25]
2. Nonspecific sticking of serum lipid to beads when concentrated serum (< ~o dilution) or culture fluid is analyzed. High binding of radiolabel will result. This problem also can be solved by preabsorbing the sample as indicated above. When a preabsorption is carried out, control samples should be included to account for possible absorption of the target molecule. Absorption Procedure ~s Reagents PA-Sepharose stock suspension: Suspend 1.5 g PA-Sepharose (Pharmacia) in 10 ml of VBS-gel. Allow the mixture to rock at room temperature for 30 min, then wash the beads with two 10-ml portions of buffer. Collect the beads by centrifugation or filtration, and resuspend them in 5.3 ml of 0.15 M saline. One milliliter of suspension is equivalent to 2 mg of PA. Serum, 1 ml, centrifuged at 3000 g for 5 min Procedure. Centrifuge 1 ml of PA-Sepharose suspension and discard the supernatant liquid. Add serum, vortex, and allow the mixture to incubate at room temperature for 60 min. Collect the serum by centrifugation (1500 g for 5 min), being sure to remove all the beads. This procedure reproducibly absorbs > 99% of the PA-reactive IgG from human or rabbit serum containing up to 12 mg of IgG/ml. 28 Examples HCG in Urine. ~7 Optimal assay conditions for HCG were established by the procedure described above. Levels of HCG in the urine of women who were in the second or third trimester of pregnancy are shown in Table V. The concentrations are given as HCG equivalents because luteinizing hormone, which also is produced during pregnancy, crossreacted in the anti-HCG immune s y s t e m Y No immunoreactivity was detected in urine from females who were not pregnant, nor in the urine from male subjects. Immunoglobulin Levels in Serum. 27.~s Similar immunoassays were developed for human IgM and IgE. Levels of these Igs were determined in the sera of normal individuals and are shown in Table VI. The IgE analyses were performed on samples that were absorbed with PA-Sepharose as described above to remove components responsible for nonspecific (lipids) and specific (IgG) interference in the assay. The IgE levels were comparable to values obtained for the same samples by double-antibody RIA. 28 The concentration of IgG in these sera (Table VI) was determined by the procedure described above [Eq. (1)].
[25]
125I-LABELED PROTEIN A
371
TABLE V LEVELS OF HUMAN CHORIONIC GONADOTROPINS (HCG) EQUIVALENTS IN URINE a
Concentration of HCG equivalents (/~g/ml = IU/ml) c
Subject b D.S. C.B. E.M. C.L. D.L. Da. L. M.B. J.L.
6.8 ± 0.8 15.8 ± 0.5 13.5 +-- 0.0 -+
~
+
[26]
O 0 ----)
OO
SP IGE ALLG ANTI= HI/LO • +DISC" - - - b COMPLEX + I G E , k - ~ . SCORE
:>'+
)
+
O0 O0
)
FIG. 1. The radioaUergosorbent test (RAST). See text for details.
acterized allergens, since the heterogeneity of allergen extracts--even relatively simple pollens7--will raise doubts as to whether or not biologically relevant allergens have been labeled. By comparison, the RAST has the very important practical advantage that only one radiolabeled substance (anti-IgE) is required for the measurement of specific IgE antibody levels to a variety of allergens. R A S T Reagents Although the technology for estimating IgE levels was available in the 1950s, 2 IgE was not reported until 1966, s and sufficient IgE reagents for the technique to be used by clinical allergists awaited the discovery of an IgE myeloma protein) About one myeloma patient with very high circulating levels of IgE has since been discovered each year (currently, 15 cases have been reported), and the purified protein has been used for radiolabeling and to raise Fc-specific anti-IgE. A high degree of immunochemical expertise is required in the preparation of RAST reagents, and so their commercial availability (Pharmacia Diagnostics AB, Uppsala, Sweden) is to be welcomed. Solid Phases
Cyanogen bromide (CNBr)-activated insoluble matrices, such as microcrystalline cellulose particles, 1° or paper disks, 1~ can be chemically M. D. Topping, J. Brostoff, W. D. Brighton, J. Danks, and M. Minnis, Clin. Allergy 8, 33 (1978). s K. Ishizaka, T. Ishizaka, and M. H. Hornbrook. J. Immunol. 97, 65 (1966). 9 S. G. O. Johansson and H. Bennich, Immunology 13, 81 (1967). ~0 L. Wide, in "Immunoassay of Gonadotrophins" (E. Diczfalusy, ed.), Acta Endocrinol. 63, Suppl. 142, 207 (1969). 11 M. Ceska and U. Lundkvist, Immunochemistry 9, 1021 (1972).
[26]
RAST AND MIXED-ALLERGENRAST
379
linked to amino groups of allergens under mild conditions. An alternative matrix, which requires no preactivation, is polymaleic anhydride. TM
CNBr-Activated MicrocrystaiHne Cellulose Particles Reagents Microcrystalline cellulose (Sigma Chemical Co.) Cyanogen bromide Sodium hydroxide, 1 M Sodium bicarbonate, 5 raM, pH 7-7.2 Distilled water 25%, 50%, and 75% acetone/distilled water (v/v) and acetone Suspend cellulose particles (10 g) in distilled water (100 ml), and dissolve CNBr (10 g) in distilled water (100 ml). Mix the two solutions at 25 ° at pH 4-6. Commence the activation by adding 1 M NaOH (40 ml) dropwise, with constant stirring, and maintain the temperature at 25° and the pH between 10 and 10.5. Transfer the suspension to a Biichner funnel fitted with a sintered glass support (G4) and wash first with cold 5 mM NaHCO3 (3 liters) and then cold distilled water (1 liter), stirring all the time to ensure adequate washing. The particles may then be lyophilized or shrunk by washing successively with 500 ml each of 25%, 50%~, and 75% acetone in water, and finally in 1 liter of acetone. Store at 4 °.
CNBr-Activated Paper Disks Reagents A hard grade of filter paper (e.g., Whatman 2 MM, or Munktells OOH), punched into disks 5 mm in diameter Cyanogen bromide Sodium hydroxide, 1 M Sodium bicarbonate, 5 mM Distilled water 25%, 50%, and 75% acetone/distilled water (v/v) and acetone Suspend 10 g of the paper disks (about 5000 disks) in distilled water (100 ml) for 3 min, then mix with a fresh CNBr solution (10 g per 300 ml of distilled water). Commence activation by dropwise addition of 1 M N a O H (50 ml) and maintain a pH of 10.5 and the temperature at 25 °. Aspirate the liquid and wash the disks 10 times in 500 ml of 5 mM NaHCOa, allowing 2 min between each wash. Similarly, wash disks successively with 500 ml of 25%, 50%, and 75% acetone in water, and then with one liter of acetone. Dry the disks by allowing the remaining acetone to evaporate at room temperature; store at 4 ° . ~ J. Merrett and T. G. Merrett, Clin. Allergy 8, 235 (1978).
380
RADIO1MMUNOASSAYS AND IMMUNORADIOMETRIC
ASSAYS
[26]
Polymaleic Anhydride (PMA) Particles Reagents Polymaleic anhydride (British Drug House Ltd.) These particles require no activation, but can be made super-fine by stirring a 10% suspension in 0.1 M NaHCOa, pH 8.5-9, with a high-speed mixer (e.g., Silverson Laboratory Mixer-Emulsifier).
Solid-Phase Allergens Commercially available allergens are compared by the RAST-inhibition technique (see RAST Reagents: l~5I-Labeled Anti-IgE), and the most potent should be selected for insolubilization. Most pollen extracts are satisfactory, but animal danders or epithelia and, particularly, molds have to be carefully selected. Once selected, the allergen should be coupled to the solid-phase support in the ratio of 1 : 100 (w/w), e.g., 10 mg of allergen per gram of solid-phase particles or disks. When the weight of allergen is not certain, assume that 1 ml of skin-prick test solution contains 1 mg, but be prepared to optimize amounts by determining the least allergen concentration that produces the highest RAST results against a panel of allergic sera.
Reagents Allergen to be coupled, dialyzed against 0.1 M NaHCO3 CNBr-activated cellulose particles CNBr-activated paper disks PMA particles NaHCO3, 0.5 M, NaHCO3, 0.1 M Tris, 0.1 M, pH 8.5 Sodium acetate, 0.1 M, pH 4 RAST assay buffer: 50 mM phosphate buffer, pH 7.4, containing 1% Tween 20, 0.3% bovine serum albumin, and 0.05% thiomersal Suspend 1 g of solid phase in 10 ml of the dialyzed allergen extract and vertically rotate for 60 hrs at room temperature. Wash the solid phase with 10 ml of 0.5 M NaHCO3, mix for 3 hr with 1 M Tris, pH 8.5, and then wash twice with 10 ml of RAST assay buffer with continuous mixing for each wash. Finally, store in 100 ml of assay buffer at 4°; this is sufficient for 2000 assay tubes.
12~I-Labeled Anti-IgE It has been suggested that the radiolabeled IgG fraction of anti-human IgE can be used in the RAST, TM but it is our experience that anti-IgE must 1~ j. K. G. Sarsfield and G. Gowland, Clin. Exp. Immunol. 13, 619 (1978).
[26]
RAST AND MIXED-ALLERGEN RAST
381
be immunosorbent purified if falsely positive results are to be avoided. Therefore we describe immunosorbent purification of anti-IgE.
Purification o f lgE Reagents Myeloma IgE serum; conceivably, high titer serum in excess of 100,000 U/ml could be used as an IgE source Phosphate buffer, 10 mM, pH 7.5 Phosphate buffer, 30 mM, pH 7.5 DEAE-cellulose 52 (Whatman Biochemical Co Ltd.) Sephadex G-200. The method is similar to that first described by Nezlin et al. 14 Dialyze myeloma IgE serum against 10 mM phosphate buffer, pH 7.5, and apply 10 mg to a column of DEAE-cellulose 52 equilibrated in the same buffer. Discard the protein removed in this buffer, and collect the bulk of IgE in 30 mM phosphate pH 7.5 buffer wash. Filter this IgE protein through a column (110 × 3 cm) of Ultragel AcA 34 in 50 mM Tris-HC1 buffer with 0.28 M NaC1, pH 8.0. Test the column fraction by RIA, and pool the located pure IgE. Preparing CNBr-activated Sepharose 4B-IgE Immunosorbent Column Reagents Purified IgE Coupling buffer: 0.1 M NaHCO3, pH 8.5, containing 0.5 M NaCI CNBr-activated Sepharose 4B (Pharmacia AB) HC1, 1 mM Tris buffer, 1 M, pH 8.5 Acetate, 0.1 M, pH 4, containing 0.5 M NaC1 Dialyse 1 mg of IgE against the coupling buffer. Reconstitute, and wash 300 mg of freeze-dried CNBr-activated Sepharose 4B on a sintered glass filter (G3) with 70 ml of 1 mM HCI. Suspend the Sepharose in 10 ml of IgE solution, and mix by slow vertical rotation for 60 hr at room temperature. Wash away excess protein with the coupling buffer, and block any remaining active imidocarbonate groups with 1 M Tris buffer, pH 8.5, for 2 hr at room temperature. Finally, wash away excess Tris and adsorbed protein with the coupling buffer, followed by 0.1 M acetate at pH 4 containing 0.5 M NaCI, followed again by coupling buffer. Store at 4° until required. ~4 R. S. Nezlin, Y. A. Zagyansky, A. I. Kaivarainen, and D. V. Stefani, Immunochemistry 10, 681 (1973).
382
R A D I O I M M U N O A S S A Y S AND I M M U N O R A D I O M E T R I C
ASSAYS
[26]
Affinity chromatography--Preparation of lmmunosorbent The immunosorbent-purification of rabbit anti-IgE described is similar to that recorded in detail by Bennich and Johansson, ~ except for the substitution of a commercial anti-IgE produced in rabbit.
Reagents CNBr-activated Sepharose-IgE Rabbit anti-IgE (Behringwerke-Hoechst) NaHCOa, 0.1 M, pH 8.5, containing 0.5 M NaC1 Acetic acid, 0.1 M The procedure should be performed at 4 °. Overnight, stir 20 ml of rabbit antiserum specific for human IgE with an equal volume of SepharoseIgE. Pour into a chromatography column and elute unbound protein with bicarbonate-buffered saline. Desorb the y-globulin fraction of anti-IgE serum with 0.1 M acetic acid; it is now suitable for labeling with 1251.
Labeling lmmunosorbent-Pure Anti-IgE with 1251 This is essentially the chloramine-T procedure first described by Greenwood, Hunter and Glover, ~ and the aim is to label 1 mol of anti-IgE with 1 atom of 1251.
Reagents Immunosorbent-pure anti-IgE Chloramine-T, 2 mg per milliliter of 50 mM phosphate, pH 7.4 12~I, 2 mCi per 10/zl of 50 mM phosphate, pH 7.4 Sodium metabisulfite, 4 mg per milliliter of 50 mM phosphate, pH 7.4 RAST assay buffer: 50 mM phosphate buffer, pH 7.4, with 1% Tween 20, 0.3% bovine serum albumin and 0.05% thiomersal Sephadex G-200 equilibrated in albumin buffer Perform the reaction in a well-ventilated fume-chamber in an ice-water bath. Use a micropipette to add 2 mCi of 125I (10/~1) to 10/.d of anti-IgE (10/.~g) in a 55 × 12 mm polystyrene tube and then in quick succession, with rota-mixing, add 10/zl of chloramine-T (20/zg), 10 txl of sodium metabisulfite (40/zg), and 1 ml of albumin buffer. Separate the radiolabeled anti-IgE from denatured products and free iodide by filtration through Sephadex G-200 with albumin buffer. Store at 4 °. 15 H. Bennich and S. G. O. Johansson, Adv. l m m u n o l . 13, 1 (1971). ~e F. C. Greenwood, W. M. Hunter, and J. S. Glover, Biochem. J., 89, 114 (1963).
[26]
RAST AND MIXED-ALLERGEN RAST
383
R A S T Procedure Paper Disks Reagents Allergen coupled to paper disks Test and reference sera Immunosorbent-pure [lzSI]anti-IgE RAST assay buffer: 50 mM phosphate, pH 7.4, with 1% Tween 20, 0.3% bovine serum albumin, and 0.05% thiomersal Saline, 0.9% Polystyrene tubes (55 × 12) mm Dilute 50/zl of test (or reference) serum to 150/zl with incubation buffer; incubate overnight with the appropriate allergen-paper disk. Stop the reaction by washing the disk three times with 2 ml of 0.9% saline, and then "develop" the disk by adding 17 nCi of [125I]anti-IgE in 100/~1 of incubation buffer; again incubate overnight at room temperature. Finally, remove the unbound radioactivity by again thoroughly washing the disk with normal saline. Measure the bound radioactivity in a gamma scintillation spectrometer. The amount of specific IgE in the test sample is compared with a set of four reference sera and assigned a RAST grade based on the commercially developed scoring system (Pharmacia, Phadebas RAST): zero is negative; 1 is equivocal; 2, 3, and 4 are increasingly positive. The test time can be reduced by half, at the expense of using three times as much labeled anti-IgE, using undiluted test serum (50/.d), and reducing the first incubation period from overnight to 3 hr. Cellulose or PMA Particles The RAST can be performed using allergens linked to particles instead of to paper disks. Reagents Allergen particles Test and reference sera Immunosorbent-pure [x25I]anti-IgE RAST assay buffer: 50 mM phosphate, pH 7.4, with 1% Tween 20, 0.3% bovine serum albumin, and 0.05% thiomersal Decanting aid: 1% Tween 20 in 0.15 M NaCI NaCI, 0.15 M Allow the test serum (100/~1) to react with a suspension of allergen particles (50 ~1) on a horizontal shaker for 3 hr at room temperature. Stop
384
RADIOIMMUNOASSAYS AND I M M U N O R A D I O M E T R I C ASSAYS I*
~
I~ = i~+~._ I~ ~
Io-
OO
~ l, i k . ~ _ + l ~ l l ~
TEST+Si:~.IGE ALLG.
[26]
+ ALLG.DiSC
.a.~.+OO ~ ~ ~ OO OO
I~
+ ANTIIGEA"
FIG. 2. Reverse RAST. See text for details,
the primary reaction by adding 2 ml of decanting aid, centrifuge the assay tubes at 2000 g for 1 min, and decant the supernatant solution. Mix 17 nCi of xzsI-labeled anti-IgE (100/zl) with the particles, and allow to stand overnight at room temperature. Finally, thoroughly wash the particles three times with 2 ml of 0.9% saline and measure the bound radioactivity in a gamma scintillation spectrometer. Multiple suction devices (Garner Engineering Co. Ltd., Sherborne, England) are very useful in speeding the washing and decanting steps. R A S T Inhibition Technique
The RAST principle can be used for the assay of allergens by either direct or indirect techniques, ar The indirect method has been commonly used and depends on the ability of the test allergen to inhibit the RAST result of a known positive reference serum. (Fig. 2). Reagents (In addition to those necessary for the RAST Technique, see p. 383. Reference serum pool, from patients with grade 4 RAST scores to the allergen being tested. Patients should not have received specific immunotherapy. Test allergen Prepare serial threefold dilutions of the test allergen in RAST assay buffer; the number of dilutions depends on the strength of the allergenic extract. To aliquots (100/xl) of the serial dilutions add 50/~1 of reference serum and incubate the mixture overnight at room temperature. Add the solid phase allergen (disk, or 50/xl of particles) and perform the RAST in the usual manner. x7 L. Yman, G. Ponterius, and R. Brandt, Dev. Biol. Standard. 29, 151 (1975).
[26]
RAST AND MIXED-ALLERGEN RAST
385
Mixed-Allergen H A S T Allergen extracts coupled to disks or particles usually contain several allergens, each of which may have one or more allergenic determinants. Consequently, a specific IgE reference serum probably has a spectrum of IgE ~intibodies directed to the extract. The spectrum will vary among patients, and for this reason an allergen-independent reference system was developed based on a direct, sandwich RIA for total IgE. TM This assay is similar to the RAST, except that anti-IgE instead of allergen is coupled to either disks or particles. It is simpler to standardize allergen particles for the mixed-allergen RAST, la and so we have limited this part of the text to cellulose particles.
Total IgE Reference Curve Reagents Rabbit anti-IgE serum (Behringwerke-Hoechst) CNBr-activated microcrystalline cellulose particles (see RAST Reagents: Solid Phases: CNBr Activated Microcrystalline Cellulose Particles RAST assay buffer: 50 mM phosphate, pH 7.4, with 1% Tween 20, 0.3% bovine serum albumin, and 0.05% thiomersal Sodium Sulfate, 18% NaHCO3, 0.1 M Decanting aid: 1% Tween 20 in 0.15 M NaCI IgE reference serum dilutions: 0.5, 2, 5, 10, 20, 50, and 100 U/ml in 50% horse serum/RAST assay buffer Immunosorbent-pure [125I]anti-IgE Precipitate the T-globulin fraction from 1 ml of anti-IgE serum by mixing 1 ml of serum with 0.18 g of anhydrous sodium sulfate and 1 ml of 18% sodium sulfate. Allow to stand overnight at room temperature. Wash the precipitate twice with 2 ml of 18% sodium sulfate, reconstitute in 1 ml of 0.1 M NaHCO3, and covalently link to 100 mg of CNBr-activated microcrystalline cellulose particles (see RAST Reagents: Solid Phase Allergens). Incubate 50/~1 of anti-IgE particles for 3 hr at room temperature with dilutions of an IgE reference serum previously calibrated against an international standard (e.g., WHO 69/204). Stop the reaction by adding 2 ml of is U. Lundkvist, in "Advances in Diagnosis of Allergy: RAST" (R. Evans, III, ed.), p. 85. Symposia Specialists, Miami, Florida, 1975. la J. Merrett and T. G. Merrett, Clin. Allergy 8, 235, (1978).
386
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC
ASSAYS
[26]
decanting aid, centrifuge the assay tubes at 2000 g for 1 min, and decant the supernatant solution; add 17 nCi of immunosorbent-pure [12~I]anti-IgE in 100/.d of RAST assay buffer, mix, and allow to stand overnight at room temperature. Finally, wash the particles three times with 0.9% saline (2 ml) in the usual way, and measure the radioactivity associated with the particles.
Estimating the Potency of Allergen Particles Individual particle RASTs must be related to the total IgE reference curve. Reagents [In addition to those necessary for the IgE reference curve (see Mixed Allergen RAST) and the cellulose particle RAST (see RAST Procedure)]. Allergen-specific IgE reference pools: for inclusion in these pools, a serum must have a RAST score of 4 to selected allergen, but no more than 2 to other common allergens Low IgE serum pool: composed of sera from nonatopic patients with total IgE levels below 20 U/ml The object is to determine the amount of allergen particles that, when made to react with a ¼ dilution of the appropriate reference serum pool, produces a count rate in the particle RAST equal to that produced by 20 U/ml in the IgE reference curve. Perform total IgE and particle RAST assays at the same time; it is usually sufficient to use doubling dilutions of allergen particles ranging from x (i.e., undiluted) to ~ in a volume of 50 t~l; the more potent the allergen, the fewer particles will be required. Having determined the optimal concentration of allergen particles, first check that the nonspecific binding is minimal; i.e., when 50 Izl of the titrated amount of particles are reacting with 100 ~1 of the low IgE serum pool, the count rate should be less than that produced by 0.5 U/ml in the IgE reference curve. If the nonspecific binding is high, then a more pure allergenic extract should be used to prepare the allergen particles. Finally, dilute the reference serum pool with serum from the low IgE pool and check that the RAST curve is approximately parallel to the IgE reference curve, and that the particles are unsaturated at a reference serum pool dilution of one-third. Generally, the serum dilutions equivalent to RAST scores of 4, 3, 2, and 1 produce radioactive counts equivalent to 20, 5, 1.2, and 0.6 U/ml when read off the IgE reference curve (Fig. 3).
Mixed-Particle RAST The mixed-particle RAST is performed identically to the particle RAST, except that the 50 tzl suspension will contain more than one vari-
[26]
RAST AND MIXED-ALLERGEN RAST
387
/ 4000
TOTAL IGE
/
GRASS POLLEN-SPECIFICIGE
3000 •
,t/4)
D--
OC W O. u') I.Z 0
2000
I000
i iiiiii
i
I
1
~ iiiill
t0 UNITS IGE PER ML
i
i illlll
1
I00
I
I
1/100 Ii10 SERUMDILUTIONS
I
I
FIG. 3. Estimating the potency of allergen particles. IGE, immunoglobulin E; SP, serum pool. See text for details.
ety of allergen particle. For example, instead of 0.5 mg of grass pollen, the 50 ~l might contain in addition 0.5 mg ofDermatophagoides pteronyssinus and 1 mg of cat epithelial particles--that is to say, 2 mg of particles instead of 0.5 mg. It is likely that the nonspecific binding will increase slightly with the use of more particles, and so a "nonatopic serum" blank should always be run with each assay. This is sometimes more of a problem with PMA than with cellulose particles. Returning to the immunological definition of the atopic person as being one who readily makes IgE antibodies to common environmental allergens, ~ we should like to emphasize the great potential of a mixed-particle RAST screen in any preliminary allergy investigation. For example, in the United Kingdom three common allergens are grass pollen, D. pteronyssinus, and cat epithelium, and we believe that a mixed-particle RAST using these three allergens will detect at least 90% of the atopic population. In other climates, of course, different allergens may be more important, but the principle remains the same: the mixed-allergen RAST technique can be a very efficient screening test for atopic allergy.
388
RADIOIMMUNOASSAYS
[27] S e m i a u t o m a t i o n Magnetic
AND
IMMUNORADIOMETRIC
of Immunoassays Transfer Devices
ASSAYS
[27]
by Use of
B y K E N D A L L O . S M I T H a n d WARREN D . G E H L E
This chapter is designed to review some of the approaches now used for automating immunoassays, specifically the radioimmunoassay (RIA) and the enzyme-linked immunosorbent assay (ELISA), and to describe the use of magnetic devices for processing solid phase elements in these assays. The RIA and ELISA fields, though relatively new, are broad, and newly reported technique variations are accumulating very rapidly. Since the first reports of Berson and Yalow,l'2 radioimmunoassay methods have been used as powerful analytical and quantitative tools. The value of the RIA in biology and medicine is suggested by the fact that Dr. Yalow was awarded a Nobel prize for her pioneering work in this area. The ELISA technique, originally described by Engvall and Perlmann 3 in 1972 for quantitation of antibodies and antigens, has likewise been widely exploited, particularly in the area of rapid serodiagnosis. Several excellent reviews of RIA and ELISA techniques have appeared in the last few years that describe in detail the chemical and physical principles involved, the myriad of technical variations that have been devised, some of the sources for components required for setting up the techniques, and the tremendous variety of applications in which RIA and ELISA have been used. 4-a It would be helpful to the reader to consider from the outset that the RIA and ELISA techniques are similar in many respects, differing principally in the fact that the label in one case is an isotope (e.g., 125I), while in the other case it is an enzyme (e.g., horseradish peroxidase). As will be indicated later, results with one system are often comparable or identical with results with the other, assuming that the immune reagents used are 1 S. A. Berson and R. S. Yalow, J. Clin. Invest. 38, 1996 (1959). R. S. Yalow and S. A. Berson, J. Clin. Invest. 39, 1157 (1960). 3 E. Engvall and P. Perlmann, J. lmmunol. 109, 129 (1972). 4 A. J. Moss, G. V. Dalrymple, and C. M. Boyd, "Practical Radioimmunoassay.'" Mosby, St. Louis, Missouri, 1976. 5 D. S. Skelley, L. P. Brown, and P. K. Besch, Clin. Chem. 19, 146 (1973). e j. p. Felber, Adv. Clin. Chem. 20, 129 (1978). 7 F. W. Spierto and W. Shaw, Crit. Rev. Clin. Lab. Sci. 7, 365 (1977). s j. L. Sever and D. L. Madden, J. Infect. Dis. Suppl. 136, 257 (1977). 9 A. Voller, A. Barlett, D. Bidwell, M. Clark, and A. Adams, J. Gen. Virol. 33, 165 (1976). METHODS IN ENZYMOLOGY, VOL. 70
Copyright © 1980 by Academic Press, Inc. All rights of reproduction in any form r~served. ISBN Ikl2-181970-1
[27]
MAGNETIC DEVICES FOR USE IN IMMUNOASSAYS
389
the same except for the tag. Mechanical methods that were first used for RIA work or for ELISA work can frequently be used easily and effectively with the other technique. The quantitation of radiation (counts per minute) or the intensity of color development at the end of the two procedures constitutes the main difference in their mechanics.
A Case for Using a Solid Phase Leiva 1° contrasted and compared the many separatory techniques currently employed in various RIA systems. He observed that the choice of the separation technique to use in any assay is of prime importance because that choice will largely govern the kind of equipment required, the maximum feasible batch size in an assay run, the manipulations required for extractions and/or washings, the reproducibility of the test, and the sensitivity of the test. Separatory techniques now employed include antibody precipitation (second antibody precipitation of antigen-antibody complexes), nonspecific adsorption of immune complexes (e.g., with charcoal, talc, dextran), selective precipitation of immune complexes by salting-out or partitioning (e.g., with ammonium sulfate, organic solvents, or polyethylene glycol), ion exchange techniques, molecular sieving and
solid phase systems. A strong case can be made for using solid phase methods for separating bound from unbound reagents on the basis of greater technical simplicity. A solid phase can be separated from various reagents by simply aspirating, decanting, or mechanically moving by magnetic force. Perhaps the most attractive point is that washing immune complexes attached to a solid phase is extremely efficient. The simplicity, reproducibility, and sensitivity of most existing solid phase assays are such as to make conversion from more cumbersome separatory methods to solid phase methodology very tempting. Nonspecific binding to the solid phase can be greatly reduced by use of purified labeled reagents and by blockage of nonspecific binding sites with immunologically "indifferent" proteins (such as bovine serum albumin). Centrifugation, a time-consuming and awkward procedure required in some separatory techniques, is not required in most solid phase systems. Very tiny amounts of antigen or antibody are required for coating solid phase units, so thousands of units can be coated with the appropriate reagent, dried, and stored indefinitely in a highly stable state. For example, we have coated plastic beads with herpes simplex virus antigen, soaked 10 B. Leiva, Am. J. Med. Technol. 43(1), 41 (1977).
390
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
[27]
them in 1% bovine serum albumin, dried them, and found no change in antigenic potency after storage for 6 months at - 70 °, - 24 °, or + 5°. Solid Phase Materials The primary requirement for a solid phase material is either that it be capable of nonspecifically adsorbing useful amounts of the desired reagent (usually a protein) or that it be possible to chemically link the solid phase to the reagent. Since the solid phase in an immunoassay is usually used only once and discarded, it is also desirable that the material be relatively inexpensive and obtainable in quantity. Ready availability of large, uniform lots of solid phase material (e.g., plastic) is important for test reproducibility. 1°-12 Uniformity in the total surface area on which the immune reactions take place is important for precision; thus the shape of the solid phase is important. Solid phases that have been used for RIA or ELISA include the following ones: Plastic test tubes 13--16 Plastic microtiter plates 17-20 Plastic disks or beads 21-26 Plastic filters or disks 27,2s Glass fiber filters 29 n p . Atanasiu, S. Avrameas, J. Beale, P. S. Gardner, M. Grandien, K. Mclntosh, D. M. McLean, A. Schuurs, O. Sobeslavsky, and A. Voller, Bull. WHO 56, 241 (1978). lz B. S. Chessum and J. R. Denmark, Lancet 1, 161 (1978). la K. J. Catt and G. W. Tregear, Science 158, 1570 (1967). 14 S. E. Salmon, G. Mackey, and H. H. Fudenberg, J. lmmunol. 103, 129 (1969). 15 H. Daugharty, D. T. Wartield, and M. L. Davis, Appl. Microbiol. 23(2), 360 (1972). le M. Ceska, F. Grossmfiller, and F. Effenberger, Infect. Imrnun. 19, 347 (1978). 17 j. D. Rosenthal, K. Hayashi, and A. L. Notkins, J. Immunol. 109, 171 (1972). is R. H. Prucell, D. C. Wong, Y. Moritsugu, J. L. Dienstag, J. A. Routenberg, and J. D. Boggs, J. Immunol. 116, 349 (1976). 19 A. R. Kalica, R. H. Purcell, M. M. Sereno, R. G. Wyatt, H. W. Kim, R. M. Chanock, and A. Z. Kapikian, J. lmmunol. 118, 1275 (1977). z0 E. Reiss, H. Hutchinson, L. Pine, D. W. Ziegler, and L. Kaufman, J. Clin. Microbiol. 6, 598 (1977). 21 K. J. Catt, H. D. Niall, and G. W. Tregear, Aust. J. Exp. Biol. Med. Sci. 45, 703 (1967). 22 K. O. Smith, W. D. Gehle, and A. W. McCracken, J. Immunol. Methods 5, 337 (1974). 23 K. O. K. Kalimo, R. J. Marttila, K. Granfors, and M. K. Viljanen, Infect. Immun. 15, 883 (1977). 2~ B. R. Ziola, M.-T. Matikainen, and A. Salmi, J. lmmunol. Methods 17, 309 (1977). 25 O. H. Meurman, M. K. Viljanen, and K. Granfors, J. Clin. Microbiol. 5, 257 (1977). 2e j. j. Langone, M. D. P. Boyle, and T. Borsos, J. Imrnunol. Methods 18, 281 (1977). 27 p. Davis, A. S. Russell, and J. S. Percy, Am. J. Clin. Pathol. 67, 374 (1977). 38 S. P. Halbert and M. Anken, J. Infect. Dis. 136, Suppl., $318 (1977). 39 E. D. Sevier and R. A. Reisfeld, Immunochemistry 13, 35 (1976).
[27]
MAGNETIC DEVICES FOR USE IN IMMUNOASSAYS
391
Porous glass 30 Paper disks or filters 31.32 Infected cells adherent to culture vessels 22,33-36 Sepharose (agarose) beads 37,38 Bentonite (clay) 39 Polyacrylamide gel 4o Staphylococcal protein A 41,42 Magnetically attractable cellulose 43 Magnetically attractable polyacrylamide-agarose beads 44 Magnetically attractable ferric oxide particles 45-47 Magnetically attractable plastic coated beads 4s A finely divided solid phase (employing small particles) has the advantage of (a) high efficiency in surface:volume ratio; (b) capability for being dispersed throughout the reaction mixture, allowing immune reactions to proceed to completion very quickly. However, single-piece solid phases have the distinct advantage of being easier to manipulate and process. Moreover, many immunoassays employing single-piece solid phases are sufficiently sensitive that it is not necessary to allow the immune reactions to go to completion in order to obtain good dose-response quantitation. Dilution of the unknown sample and the labeled reagenta, as well as incubation temperature, can be adjusted upward or downward to give sufficiently sensitive immunoassay results within the time frames required for most laboratory testing. a0 H. H. Weetall, Biochem. J. 117, 257 (1970). 31 R. Pauwels, H. Bazin, B. Platteau, and M. van der Straeten, J. lmmunol. Methods 18, 133 (1977). 32 H. Watanabe and I. H. Holmes, J. Clin. Microbiol. 6, 319 (1977). 33 K. Hayashi, D. Lodmeli, J. Rosenthal, and A. L. Notkins, J. lmmunol. 110, 316 (1973)." 34 N. H. Levitt, H. V. Miller, and G. A. Eddy, J. Clin. Microbiol. 4, 382 (1976). 35 B. Forghani, N. J. Schmidt, and E. H. Lennette, J. Clin. Microbiol. 4, 470 (1976). 30 M. A. Jankowski, E. E. Petersen, and J. F. B6cker, Acta Virol. 21, 405 (1977). 37 y . S. Choi, Immunochemistry 14, 53 (1977). 38 A. A. A. Ismail, P. M. West, and D. J. Goldie, Clin. Chem. 24, 571 (1978). 39 W. C. Cheng and D. W. Taimage, J. Immunol. 103, 1385 (1969). 40 G. D61ken and G. Klein, J. Natl. Cancer Inst. 58, 1239 (1977). 41 M. E. Soergel, F. L. Schaffer, J. C. Sawyer, and C. M. Prato, Arch. Virol. 57(3), 271 (1978). 42 p. B. Jahrling, R. A. Hesse, and J. F. Metzger, J. Clin. Microbiol. 8, 54 (1978). 43 M. J. Anderson, Med. Lab. Sci. 35(2), 173 (1978). 44 j. L. Guesdon and S. Avrameas, Immunochemistry 14, 443 (1977). 45 L. S. Hersh and S. Yaverbaum, Clin. Chim. Acta 63, 69 (1975). 4s L. Nye, G. C. Forrest, H. Greenwood, J. S. Gardner, R. Jay, J. R. Roberts, and J. Landon. Clin. Chim. Acta 69, 387 (1976). 47 N. Gochman and L. J. Bowie, Anal. Chem. 49, 1183A (1977). 4s K. O. Smith and W. D. Gehle, J. Infect. Dis. 136, Suppl., $329 (1977).
392
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC
ASSAYS
[27]
Automated Processing It is important that mechanical processing of solid phase units during an immunoassay be efficient and highly uniform. In fact, the mechanical facility with which the solid phase can be manipulated during the assay procedure is sufficiently important that it probably should be a primary determining factor in the choice of the assay system. M a n u a l manipulation of each solid phase unit, particularly during the multiple washing steps (e.g., handling units one at a time with forceps), severely restricts the number of tests that can be accomplished within an acceptable time frame. Operator boredom and consequent fatigue are significant negative factors in highly repetitious operations such as immunoassays. Moreover, reproducibility is adversely affected when time-consuming manual manipulations are required, particularly in assays involving large numbers of specimens and short incubation periods (the first and last units can hardly be equivalent in treatment if each unit must be handled separately by a technician). Major manual steps involve pipetting and diluting specimens, dispensing reagents, incubation, multiple rinsings, gamma counting or spectrophotometry, and data-reduction interpretation. Efficient use of technician time is difficult if these steps are all done manually. For these reasons there has been considerable interest in automated systems for RIA and ELISA; consequently, several more or less automated systems have been devised. The review by Gochman and Bowie 47 compared the following commercially available (or soon to become available) systems: (a) the Union Carbide "Centria" (a solid phase is not used); (b) the Micromedics "Concept 4," which uses antibody-coated tubes as a solid phase; (c) the Becton-Dickinson "Aria II," which is designed around a flow-through, • reusable solid phase chamber; (d) the Technicon automated RIA system, which employs antibody-coated ferric oxide particles as a solid phase. 4a Sevier and Reisfeld 29 described a semiautomated RIA processor that utilizes a solid phase double-antibody procedure and glass fiber filters. Brooker e t al. 49 devised an automated system called "Gamma-flow" which uses column separation of reactants instead of a solid phase adsorbent. Ismail e t al. 3~ designed a fully automated, continuous flow system employing agarose (Sepharose) beads as the solid phase. The Gilford Corporation has developed an automated system for accomplishing ELISAs; it uses plastic cuvettes as the solid phase, automatic spectrophotometry, and computerized reduction of data. Cordis Laboratories, Inc. has marketed an ELISA system employing plastic-coated disks (composition is proprietary information), which are processed with mechanical devices 49 G. Breoker, W. L. Terasaki, and M. G. Price, Science 194, 270 (1976).
[27]
MAGNETICDEVICES FOR USE IN IMMUNOASSAYS
393
that they provide. Organon Diagnostics has devised a semiautomated solid phase ELISA for detecting viral antigens in clinical specimens. 5° Leinikki and Passila, 51 in cooperation with Finnipipette-Labsystems, developed a semiautomated ELISA for viral antibodies employing disposable polystyrene cuvettes with optically clear bottoms. The capacity of most of these automated systems for handling specimens is probably in the range of I0-40 test samples per hour. 38,47Claims for greater capacity have been made, but such claims seem to ignore the substantial time required for specimen dilution, instrumental reading of the results, and data analysis. Galen, ~2 speaking from the position of a hospital pathologist, expressed the view that "industry efforts directed toward total automation of RIA procedures have been numbingly earnest in their intentions and crushingly dull in their execution." A Case for Choosing Semiautomation Judicious use of labor saving devices obviously increases work efficiency by decreasing the boredom and personal fatigue that result from excessively repetitious manual operations. Moreover, mechanical devices obviously perform some tasks more precisely and economically than technicians can. Fully automated systems, however, can be prohibitively expensive as capital investments, trouble-prone, expensive to maintain, and highly restrictive in how they can be used (e.g., they may be inflexible in the kinds of tests they can perform). The manufacturers of automated equipment may deliberately design components of their system so that they will be the sole source of these components (e.g., proprietary reagents and odd-shaped tubes) and this can be a big disadvantage. Between the extremes of n o automation and c o m p l e t e automation there is an area of semiautomation that most laboratories may wish to consider because of cost and flexibility considerations. The marketplace is becoming crowded with devices designed to accomplish the following major steps in solid phase immunoassays: sample pipetting, diluting, dispensing diluted samples into appropriate tubes or wells, dispensing reagents, adding the solid phases to appropriate tubes or wells (or the reagents to the solid phase tube or well), processing the solid phases through various washes and specific reagents and, finally, measurement of the resulting color (in the ELISA) or radioactivity (in the RIA). Automated pipetting-diluting stations are commercially available in 5oQ. R. Miranda, G. D. Bailey,A. S. Fraser, and H. I. Tenoso,J. Infect. Dis. 136, Suppl., $304 (1977). s~ F. Leinikki and S. P/isillfi,J. Infect. Dis. Suppl., 136, 294 (1977). 5~R. S. Galen, Human Pathol. 8(4), 359 (1977).
394
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
[27]
configurations that accommodate various numbers and sizes of test tubes or wells, including 96-well Microtiter plates. Either serial dilutions or single, screening dilutions of a sample can be accomplished with some of these devices. Gamma counter manufacturers are marketing pipetter-diluters and computer data processors as ancillary equipment for immunoassay work. Most gamma counters are automated in that they provide automatic sequencing of 300 or more samples per load and a printout of counts per minute on tapes. Reduction and processing of this kind of data can be accomplished with relatively simple, programmable desk top computers by methods such as those described by Grotjan and Steinberger sa and Naus et al.S4 A spectrophotometric system has recently been offered commercially that allows measurement of substrate color development in each of the 96 wells of a Microtiter plate within 60 sec and performs a number of dataprocessing functions, s4a Both RIA and ELISA techniques have been steadily miniaturized; the Microtiter plate level of mechanical miniaturizing may be a reasonable compromise in size for test accuracy and precision, and still provide good economy of reagents (volumes of 100/zl or less are required for tests). The Gilford Processor/Reader 54b uses 300/~1 volumes and allows optional manual control or automation in the washing, reagent-dispensing, and data collection steps of the ELISA. Flexibility in immunoassay design is important because the state of the RIA and ELISA art is changing so rapidly that new improvements must be taken into account continually. Accepting rigidity in assay design options by choosing complete, "hands-off' automation may, therefore, be unwise at this particular time. Technical flexibility is highly desirable, but is possible only if each of the major steps in the immunoassay can be independently modified to fit specific needs of the laboratory. Thus, a strong case can be made at the present time for semiautomation, which sufficiently minimizes technical drudgery, yet gives the needed flexibility for optimizing the immunoassay through modification of any one or a combination of technical steps. Most of the labor-saving components are now commercially available for assembling semiautomated systems for either the RIA or ELISA, tailored exactly to the needs of a particular laboratory. H. E. Grotjan and E. Steinberger, Comput. Biol. Med. 7, 159 (1977). A. J. Naus, P. S. Kuppens, and A. Borst, Clin. Chem. 23, 1624 (1977). u~ Titertek Multiskan Plate Reader, Flow Laboratories, 1710 Chapman Avenue, Rockville, Maryland 20852. ~b Gilford Instrument Laboratories, Inc,, Oberlin, Ohio 44074.
[27]
MAGNETIC DEVICES FOR USE IN IMMUNOASSAYS
395
An Example of Semiautomation: Magnetic Transfer Devices for Processing One of the primary purposes of this chapter is to present a detailed description of magnetic transfer devices for simultaneously processing solid phase units (beads) in either RIA or ELISA systems. It is understood that automatic pipetting devices, automatic instrumentation for reading v-radiation or color development, and programmable calculators could, at the discretion of the investigator, be used in conjunction with the magnetic processing devices described below. We have already described the design and use of 24-place and 60-place devices for use in solid phase RIA or ELISA techniques. 48 Since that time, we have designed modifications that have increased the numbers of specimens handled simultaneously to 400 and have miniaturized the me-
FIG. 1. Crosman BB shot " b e a d s " plated with "bright nickel" and "black nickel" (for color coding purposes), then coated with polycarbonate by immersing in 5% solution of methylene chloride, scattering on slick paper, and allowing to dry. Various antigens or antibodies are adsorbed onto this plastic surface; the beads are soaked in a 1% solution of bovine serum albumin in phosphate-buffered saline, drained, air dried, and stored below 0° until needed.
FIG. 2. Device " A " consists of 96 soft iron rods embedded in a I-inch Lucite plastic sheet; the rods protrude about ~ inch on one side and are flush with the reverse side of the plastic sheet (outer dimensions are 123 m m × 80 mm).
FIG. 3. Device "B'" is a powerful permanent ceramic magnet (obtained from Texas Magnetics Corp., Garland, Texas); the outer dimensions are the same as those of device " A . "
[27]
MAGNETIC DEVICES FOR USE IN IMMUNOASSAYS
397
FIG. 4. Test tube rack containing small (6 mm x 50 mm) test tubes (also see Fig. 11); foreground; U-bottom Microtiter plates (Linbro S-MRC 96).
chanical system so as to allow use of smaller volumes of reagents (70/zl, in Microtiter plates). For several reasons the Microtiter configuration may be ideal for immunoassay work (much ancillary equipment has already been designed and marketed for Microtiter work). Therefore, the following description is aimed toward accommodating the Microtiter design. The apparatus required for performing an RIA or ELISA with this particular magnetic processing system includes: plastic-coated steel beads (see legend of Fig. 1 for details of preparation); a 96-place device ( " A " ) consisting of short iron rods embedded in a Lucite plastic sheet (Fig. 2); a permanent magnet ( " B " ) of similar dimensions (Fig. 3); two or three disposable-type Microtiter plates (depending upon whether the system is RIA or ELISA); and a test tube rack that holds small test tubes in the same spatial configuration as the Microtiter plate wells (Fig. 4). Each piece of apparatus is designed to facilitate the simultaneous processing of from 1 to 96 solid phase units (plastic-coated steel beads) through the vail-
398
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
PLASTIC
[27]
PLASTIC CLASSSPECIFIC
COATED VIRAL HUMANANTI-HUMAN,,~ COATED ANTI-HUMAN HUMAN ANTI-HUMANjq::,~. BEAD ANTIGEN ANTIBODYANTIBODYt~I~ BEAD GLOBULIN GLOBULIN GLOBULIN~ r
(~= IzsI LABEL
BOVINESERUMALBUMIN
SERUMALBUMIN
FIG. 5. (A) Diagram showing sequence of the "indirect" or labeled second antibody reaction (left to right): plastic-coated bead is coated with viral antigen, soaked in bovine serum albumin, dried, stored, withdrawn, placed in a 1 : 100 dilution of human serum, then placed in enzyme-labeled antihuman globulin. The final step (see text) is development of a color reaction by dropping the bead-antigen-antibody complex into wells containing 2,2'-azinodi(3-ethylbenzthiazoline) sulfonate. (ABTS). (B) Diagram showing sequences for quantitating class-specific human globulin (left to right): plastic-coated bead is coated with purified antihuman globulin, soaked in bovine serum albumin, dried, stored, withdrawn, placed in serial dilutions of serum (also tears, saliva, etc.), then placed in ~"I-labeled antihuman globulin. The final step is reading counts per minute in a gamma counter. These results are plotted on semilog graph paper and compared with similar results with human globulin of known concentration (standard curve).
ous reaction mixtures and rinsing baths to the point in the procedure at which test tubes containing the beads are placed either in a gamma counter (for RIA) or, alternatively, the beads are dropped into the wells of a Microtiter plate containing an appropriate enzyme substrate (for ELISA). The processing sequence can be varied with complete freedom in order to accomplish different purposes in an assay (i.e., for either RIA or ELISA; for either antigen or antibody detection; by either the "direct" or the "indirect" techniques). As an example, the following sequence of steps is appropriate for ELISA detection of human antibody to a specific viral antigen, by the "indirect" (labeled second antibody) technique (Fig. 5A shows diagrammatically how this can be accomplished). Step 1. Antigen-coated beads are placed, with plastic tipped forceps, onto the magnetic terminals of the 96-place transfer device (as in Fig. 6); all beads are dropped simultaneously into wells of a Microtiter plate containing either serial serum dilutions or a screening dilution of many differ-
[27]
MAGNETIC DEVICES FOR USE IN IMMUNOASSAYS
399
FIG. 6. Placing of antigen-coated beads onto iron terminals of device " A , " preparatory to dropping the beads (simultaneously) into a Microtiter plate whose wells contain diluted human sera. Dropping of beads into the wells is accomplished by simply removing the magnet (device " B " ) from the iron terminals of device " A . "
ent human sera (1:50 - 1 : 100 dilution is frequently a good choice of a screening dilution). Dropping of the beads is accomplished by simply removing device " B " (the magnet) from the top of device " A " (the soft iron terminals). Step 2. Beads are incubated for 1 hr at room temperature, then removed simultaneously by placing device " A " over the plate and placing the magnet ( " B " ) over the top of device " A " (see Fig. 7); beads then " j u m p " by magnetic attraction from the bottom of the microtiter wells to the iron terminals of device " A " and are held there very firmly (the magnet is powerful) (see Fig. 8). Step 3. Beads are washed sequentially in two separate water baths by dipping them six times in and out of each bath; washing is more efficient if beads are positioned upward (Fig. 9). Step 4. By removing magnet " B " from device " A , " beads are simultaneously dropped into a second Microtiter plate, each of whose wells contain 70/zl of horseradish peroxidase linked (labeled) antihuman globulin.
400
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
[27]
FIG. 7. Placing of magnet (device " B " ) over iron terminals of device " A , " so as to cause beads at bottom of Microtiter plate wells to jump to the terminals and be held there securely (beads on right have jumped; those on left have not).
Step 5. Incubation and bead removal (step 2) is repeated. Step 6. Washing (step 3) is repeated. Step 7. Step 4 is repeated (dropping of beads into wells), except that beads are simultaneously dropped into a third Microtiter plate whose wells contain a solution of the HRP substrate 2,2'-azino-di(3-ethylbenzthiazoline) sulfonate (ABTS). Step 8. Beads are incubated for 5-20 min at room temperature, or until the positive serum controls turn dark blue-green but before the negative serum controls develop a detectable color (Fig. 10). Color development in the substrate wells can be quantitated by spectrophotometry or, less precisely, by eye. If spectrophotometry is to be done directly in the Microtiter plate wells, plates with fiat-bottomed wells are used and the beads are removed so as to allow an unobstructed beam of light to pass through the colored substrate in each well (as with the Flow Laboratories plate reader). The standard curve principle can be used, as for RIA quantitation) 5 If the human eye is used for reading color 55 W. R. Patterson and K. O. Smith, J. Clin. Microbiol. 2, 130 (1975).
[27]
MAGNETIC DEVICES FOR USE IN IMMUNOASSAYS
401
FIG. 8. Devices " A " and " B " are in place, with all beads adhering securely to the iron terminals.
reactions, visual reference needs to be made to positive and negative controls in order to evaluate each unknown. The standard curve principle can be used effectively to obtain reasonably good quantitation by visually comparing the color intensity of an unknown at a single dilution with the nearest-match color intensity in a serially diluted known positive sample (standard). Reproducibility of this method (visually matching unknowns with dilutions of a standard) is usually within one twofold dilution, which is adequate for many purposes. The main problem with this visual method of quantitation is that a very strong positive unknown has such a dark color that it is difficult to compare with similar dark colors in a standard sample. Therefore, an unknown sample must be diluted to a low level to determine whether it is positive and to a higher level for adequate visual quantitation (the eye seems to " s e e " only so much color intensity, and no more). As already indicated, the steps used for magnetically processing plastic-coated beads can be varied easily to accomplish different tasks. A further example is given below, which illustrates RIA quantitation of an "antigen" in human sera or saliva; the "antigen," in this particular case, is either IgG, IgA, or IgM (see Fig. 5B for a diagrammatic explanation of
402
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
[271
FIG. 9. Rinsing of beads by immersion of device " A " into a water bath. The inverted, diagonal position is preferable. Alternate total removal and total submersion gives efficient washing. Alternatively, this washing step may be accomplished by serial transfer of beads from one Microtiter plate to another (the wells simply containing water), or by holding the beads under a moving stream of water.
this assay; the solid phase bead is coated with either anti-IgG, anti-IgA, or anti-IgM; antiglobulins are prepared in goats or rabbits and are heavychain specific). Step 1. Antibody-coated beads are placed onto the magnetic terminals of the transfer device (see Fig. 6) then dropped simultaneously into wells of a Microtiter plate (Fig. 7) containing appropriately diluted specimens (serum or saliva). Step 2. Beads are incubated for 1 hr at room temperature, then removed simultaneously by magnetic force (see Fig. 8). Step 3. Beads are washed sequentially in two separate wash baths (see Fig. 9). Step 4. Beads are simultaneously dropped into a second Microtiter plate whose wells contain l~H-labeled antihuman globulin (prepared against the Fc portion). Step 5. Incubation and transfer (step 2) is repeated.
[27]
MAGNETIC DEVICES FOR USE IN IMMUNOASSAYS
403
FIG. 10. Appearance of a completed ELISA for toxoplasma antibodies. Row 1 contained a serially diluted, positive control, human serum (1 : 100-12,800). Row 2 contained a similarly diluted negative control human serum (no color developed). All remaining wells contained 1 : 100 dilutions of " u n k n o w n " human sera being tested for toxoplasma antibodies.
Step 6. Washing (step 3) is repeated. Step 7. Beads are simultaneously dropped into small test tubes (see Fig. 11), then placed in a gamma counter. Step 8. Beads (in the tubes) are read in the gamma counter. Dose-response curves are plotted for the unknown specimens (dilution factor vs counts per minute) and for serially diluted standards (containing known amounts of IgG, IgA, and IgM). From these curves, one can calculate the concentration of immunoglobulin in each of the unknown specimens. Advantages in using a magnetic processing system of this kind for immunoassays are summarized as follows: 1. The equipment can be used either for RIA or ELISA equally well; the sensitivity of the two systems is about the same4S; also see Section VIII.
404
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
[27]
FIG. 11. Beads that have been completely processed in the radioimmunoassay procedure and are about to be dropped into small test tubes, then placed in a gamma counter for reading (the last step in the assay).
2. Any number of beads (up to 96) can be processed as quickly and easily as a single bead; antigen-sensitized beads and expensive reagents are not wasted because only the number of beads required for a particular test run are used (when the Microtiter well is sensitized and used as the solid phase, an entire plate of 96 sensitized wells must be used if one test is done). 3. Beads are dropped into reaction mixtures simultaneously and are later removed and washed simultaneously, so that large number of antigen-antibody reactions are instantly started and stopped. Thus, highly reproducible results are obtainable when large numbers of assays are performed in a short time period (when unknown samples are added to 96 sensitized wells one at a time, the first and last samples cannot be incubated identically). 4. In the ELISA technique, beads can be removed simultaneously from the substrate-containing wells, thereby instantly stopping further color development due to specific enzyme-substrate turnover. In techniques where the wall of the Microtiter plate well is the solid phase, simultaneously stopping development of the color reactions in all wells is difficult or impossible. 5. Steel BB shot are uniform in diameter (see Table I), inexpensive,
[27]
M A G N E T I C D E V I C E S F O R USE IN I M M U N O A S S A Y S
405
TABLE I UNIFORMITY OF BB SHOT FROM THREE DIFFERENT SOURCES
Manufacturer o f BB shot Crosman P o w e r Line Daisy
Mean Diameter (inches) a
Standard deviation o f diameter (%)
Standard deviation o f calculated surface area ~ (%)
0.1723 0.1741 0.1702
0.21 0.27 2.59
0.59 0.78 7.7
a Eleven different shot, from each o f three different commercial sources, were m e a s u r e d with a m i c r o m e t e r caliper. E a c h shot was m e a s u r e d in three different directions, and the m e a n diameter m e a s u r e m e n t was calculated. T h e m e a n diameter o f 11 different shot was then calculated and is s h o w n in the chart. b Surface area = ~'r a.
and commercially available in very large lots. They can be coated with almost any plastic (including polycarbonate), are magnetically attractable, and thus provide an almost ideal solid phase surface upon which immune reactions can take place. The small size of the beads allows scaling down reagent volumes to very small proportions and, therefore, permits substantial conservation of precious reagent materials. If larger wells and volumes are used, several color-coded beads can be incubated in the same well at the same time, each bead coated with a different antigen or antibody, and all beads processed simultaneously) e Color coding of beads can help avoid errors in usage, several different antigens can be used for simultaneous use in one 96-well plate, and there is complete flexibility in choice of combinations. 6. Reproducibility of immunoassay results can be adversely affected by variations in the chemical formulation of different lots of plastics used as the solid phase (see section on Solid Phase Materials). Therefore, use of a thin coating of one lot of plastic over a material like metal for millions of tests could be important in standardizing immunoassay techniques. Extremely small amounts of polycarbonate (or other plastics) are required for putting a thin coating on steel beads. Therefore, one small commercial lot of plastic (e.g., one 212 pound 4 x 8 foot sheet of polycarbonate) would be sufficient to prepare beads for over 100 million tests. Plastic solid phases employing Microtiter plates, test tubes, and polystyrene beads require relatively large quantities of plastic for fabrication and, cons6 W. D. Gehle, K. O. Smith, D. A. Fuccillio, A. Perry and A. P. A n d r e s e , J. Immunol. Methods 26, 381 (1979).
406
RADIOIMMUNOASSAYS AND IMMUNORADIOMETRIC ASSAYS
[27]
sequently, the lots of plastic change frequently. Such changes in the formulation of plastic lots might cause devastating variations in immunoassay results. Reagents
Antibodies The quality of immune reagents are of at least equal importance to the choice of the mechanics for solid phase immunoassays. The entire globulin or y-globulin fraction from hyperimmune animals can be labeled and used successfullyg; however, the specific antibody playing a useful part in the immunoassay is a very small fraction of the total labeled proteins. Therefore, immune affinity chromatography is highly recommended for purifying the appropriate specific antibody fraction prior to labeling. Several good methods are available for this purpose (Cuatrecasas and Anfinsen57; Pharmacia Fine Chemicals Monograph, 1974). The main benefit in labeling only the desired, specific antibody fraction is that the background ("noise") in the assay is diminished because labeled nonspecific proteins inevitably adsorb to some extent to solid phase surfaces, in spite of prior "blocking" of the solid phase surface with nonspecific proteins such as BSA. Moreover, there is a large savings in the quantitiy of the ~25Ior enzyme label required for labeling, since only specific antibody is labeled. Labeling antibody with 1251can be accomplished in several different ways (see the review by Skelley et al.5); among the commonly used methods are (a) various modifications of Hunter and Greenwood's 58 chloramine-T oxidation method, including that of Bolton and Hunter; 59 (b) the use of catalytic amounts of lactoperoxidase for iodination, s° It is important that unbound ~25Ibe thoroughly removed from the reaction mixture, after labeling, by chromatography with an anion exchange resin, ~5 or by dialysis through a cellophane membrane. Otherwise, unlabeled ~5I may adsorb nonspecifically to the solid phase and increase background noise in the assay. Labeling antibody with an enzyme can be accomplished rapidly and efficiently by the gluteraldehyde linkage method. 9"6~ Other widely used methods for labeling antibodies with enzymes include those of Nakane and Kawaoi e2 and Engvall and Perlmann. 3 Volleff has described in techni5~ p. Cuatrecasas and C. B. Anfinsen, Annu. Rev. Biochem. 40, 259 (1971). W. M. Hunter and F. C. Greenwood, Nature (London) 194, 495 (1962). 5g A. E. Bolton and W. M. Hunter, Biochem. J. 133, 529 (1973). ~o j. j. Marchalonis, Biochem J. 113, 299 (1969). 61 S. Avrameas, lmmunochemistry 6, 43 (1969). ~2 p. K. Nakane and A. Kawaoi, J. Histochem. Cytochem. 22, 1084 (1974).
[27]
MAGNETIC DEVICES FOR USE IN IMMUNOASSAYS
407
PLASTIC COATED VIRAL HUMAN SIMIANANTIBEAD ANTIGEN ANTBODY ENZYME
ANTI-PRIMATE ~ELANT/BODY BOVINSERUM E ALBUMIN FIG. 12. Diagram showing sequence for detecting viral antibody by the peroxidase-antiperoxidase (PAP) method of Sternberger et al. ~ This method uses a "bridging" antibody (anti-primate antibody) between the human and simian antibody. Artificial (chemical) linkage of enzyme with antibody is replaced by an antigen-antibody reaction.
cal detail a "double antibody sandwich" method for detection of a n t i g e n and a simple method for detecting a n t i b o d y , using an "indirect" ELISA in Microtiter plates. This general approach to doing ELISAs (equivalent to methods we and others have described for the RIA) is now in general use (the inner wall of the Microtiter plate well can be used as the solid phase). Alternatively, the "indirect" ELISA method of Sternberger et al. ~3 or the modification of this approach by Butler et al. ~ can be used; this involves the production of soluble immune complexes of antienzyme globulin and enzyme (see diagram in Fig. 12). The Sternberger-Butler method, called by Butler the "amplified E L I S A , " is interesting because affinity chromatographed, purified antiglobulins are n o t required for enzyme labeling; artificial linkage by chemical (covalent) bonding is completely circumvented. The Organon Corporation holds patents on processes for covalent linkage of enzymes to antibodies and use of these reagents in the ELISA (enforcibility of these patents remains to be established in a court case, now pending, against one of the many commercial sources of ELISA reagents). No patents now exist on the Sternberger-Butler approach to producing labeled antibody, i.e., use of an enzyme-antibody reagent resulting from coupling by an i m m u n e reaction.
sa L. A. Sternberger, P. H. Hardy, Jr., J. J. Cuculis, anti H. G. Meyer, J. Histochem. Cytochem. 18, 315 (1970). J. E. Butler, P. L. McGivern, and P. Swanson, J. Immunol. M e t h o d s 20, 365 (1978).
408
R A D I O I M M U N O A S S A Y S A N D I M M U N O R A D I O M E T R I C ASSAYS
[27]
Protein A Protein A, a structural component of certain strains of staphylococci, has been used successfully in the place o f the second antibody in radioimmunoassay techniques, ~6 as well as the solid phase adsorbent of antibody (mentioned in the section Solid Phase Materials). This protein can be heavily labeled with ~25Iand is, therefore, a reasonable alternative for the use of labeled antiglobulins in immunoassays.
Antigens Immunoassays can be accomplished in systems where an unknown antigen (rather than antibody) competes with labeled antigen for specific antibody adsorbed onto the surface of a solid phase (such as the lower inner wall of a plastic tube). This method can be used effectively for measuring serum antibody levels (Clinical Assays Division of Travenol Laboratories, 620 Memorial Drive, Cambridge, Massachusetts, 02139); the more highly purified the labeled antigen, the less background noise there is likely to be in the assay system.
Enzyme-Substrate Combinations Judicious choice of an enzyme-substrate combination is important if optimal ELISA results are to be obtained. For example, the relative merits of peroxidase, alkaline phosphate, and glucose oxidase in the ELISA have been studied, 44 and substantial differences in both dose-response kinetics and assay reproducibility have been demonstrated. Solubility of the colored substrate derivative and the manner in which the color reaction is measured may also influence the investigator's choice of an enzyme-substrate combination. Some investigators may read ELISA results by eye rather than by spectrophotometer, therefore an easily visualized color reaction should be chosen. Optimal lighting conditions should be arranged for reading the results, and colors apt to cause visual perception problems in partially color-blind people should be avoided. Two combinations that are now widely accepted for ELISA work are (a) alkaline phosphatase-p-nitrophenyl phosphate (PNPP) substrate; (b) horseradish peroxidase-ABTS substrate. The former gives a bright yellow reaction product; the latter gives a bluish green color; both colors can be visualized easily or quantitated spectrophotometrically. Differences between shades of very dark colors are hard to distinguish by eye, therefore lighter shades are preferable if visual quantitation is desired.
[27]
MAGNETIC DEVICES FOR USE IN IMMUNOASSAYS
409
A Case for Selecting an E n z y m e R a t h e r T h a n a Radioisotope for Labeling Most RIA methods arc highly sensitive and reproducible, but there are several distinct disadvantages associated with isotopic labeling. 28 Probabl~, the greatest of these is the short half-life of the most frequently used isotopes (only 8 days for 131I and 60 days for 1251), which drastically reduces the useful shelf life of labeled reagents. Since it is impractical to make large lots of such reagents, reagents must be labeled frequently and proper standardization requires careful quality control on each lot of labeled reagents. Other problems are the licensing requirement for use of radioisotopes, special training requirements for personnel, health risks, disposal of radioactive waste, and the requirement for relatively expensive equipment (gamma counters) for readout. In contrast, use of an enzyme label has several major advantages: labeled reagents are stable on storage (shelf life of lyophilized reagents is probably infinite), disposal of labeled waste is not a problem, no special licensing is required for handling enzymes, and spectrophotometric readout equipment required for ELISA techniques is usually less expensive than gamma counters. Moreover, many workers prefer to read ELISA end points by eye, so no readout equipment is required. RIA and ELISA techniques are probably about equal in sensitivity, 4s although ELISA methods theoretically could be more sensitive. In experiments we have done comparing the relative sensitivity of the two methods by serial dilution-end point assays, 29 of 30 sera gave RIA and ELISA end points within twofold of each other when titrated for herpes simplex type 1 antibody. Similar comparisons of titrations for human cytomegalovirus (CMV) and toxoplasma antibody showed good correlation between the two methods; 26 of 27 and 27 of 29 sera showed titers within twofold of each other (CMV and toxoplasma assays, respectively) with the RIA. In all of these comparisons, ELISA results were read by eye, not with a spectrophotometer. There is a large advantage associated with the ELISA for screening specimens visually for the rare positive or negative (for example, assaying for hepatitis B antigen in human sera). A visual glance at a panel of 96 wells will establish the identity of one or two positive specimens within a few seconds. Close to 2 hr are required to cycle 96 specimens through a gamma counter to find these positive specimens, assuming 1 min counting time and time for mechanical movement of tubes. Therefore, counting tubes in a gamma counter can actually require as much time as the rest of the RIA procedure. ELISA systems employing visual readings seem ideally suited to situations involving an occasional or rare outcome, whether the result be positive or negative.
410
R A D I O I M M U N O A S S A Y S A N D I M M U N O R A D I O M E T R I C ASSAYS
[27]
Application of the Solid Phase (Bead) Magnetic Processing System; Some Examples Until recently most of our own serologic work has utilized isotopic labeling and the RIA, but we are convinced that ELISA methodology will gradually replace the RIA for much immunoassay work. The technical transition from RIA to ELISA is easy, especially when steel beads are used as solid phase units. Differences in the sequential procedure merely involve use of enzyme-labeled reagents, placing the bead (with the enzyme-labeled immune complex) into a suitable substrate solution (instead of the gamma counter), and estimating color development in the substrate solution rather than measuring gamma radiation. Examples of how we have used the magnetic processing system in immunoassays follows.
Correlation Studies between RIA and Other Serologic Methods Sera were obtained from several different sources in order to compare the RIA with the serologic assays commonly used in the past in clinical laboratories for detecting antibody to members of the TORCH (toxoplasma, rubella, cytomegalovirus, and herpesvirus) group of agents. Ninety-eight sera were obtained under code from source A. Most of these sera had been previously tested by older methods for antibody to toxoplasma, rubella virus, cytomegalovirus (CMV), and herpes simplex virus type 1 (HSV-1). One set of sera was chosen to represent a cross section of the normal population; that is, the proportion of positive and negative sera in this sample was similar to the proportion likely to be seen in the clinical laboratory setting. Because these sera represented a cross section of the normal population, there were relatively few sera with antibody to toxoplasma and only a few without antibody to rubella virus. To ensure that the RIA could detect antibody to toxoplasma, 28 additional toxoplasma-positive sera were obtained from source B. To ensure that the RIA did not give false positive reactions in the rubella test, 20 additional rubella-negative sera were obtained from source C. Besides this large number of sera, 43 sera from source D and 27 sera from source E were obtained under code. Most of these sera had been previously tested by older methods for antibody to either CMV, HSV-1, or both. Not all sera were used from each group for all tests because of limited quantities available. Toxoplasma. The commonly used serologic tests for antibody to toxoplasma are the fluorescent antibody (FA) test and the indirect hemagglutination (IHA) test. Table II shows a comparison of the RIA with these two tests. The most important point to be seen is the agreement between the RIA and the other tests in distinguishing positive and negative sera.
[27]
MAGNETIC DEVICES FOR USE IN IMMUNOASSAYS
411
TABLE II m COMPARISON OF TOXOPLASMA SEROLOGIC ASSAYS WITH THE MAGNETIC
RADIOIMMUNOASSAY (RIA) Correlation with the RIA Source of sera
Type of test used ~
No. of negative sera
No. of positive sera
Negative (%)
Positive (%)
A B
IHA FA
82 --
16 28
~- (98%) --
~ (100%) ~ (96%)
a IHA, indirect hemagglutination; FA, fluorescent antibody.
This correlation was excellent, approaching an overall agreement of approximately 98%. The 28 toxoplasma-positive sera that had been titrated for antibody by FA were titrated by RIA, and the titers of the two tests were correlated, using Spearman's rank-order correlation test. The correlation coefficient was 0.68, and probability that this value differs from zero (no correlation) by chance alone was '- 60
=
451
cyclic
;
cyclic
~
N 40
20
i
10-8
h
10-7
i
i
10-6 10-5 nucleotide c o n c . , M
i
10-4
=
10-3
FIG. 6. Calibration curves obtained for c A M P and cGMP using a urease-cAMP conjugate and cAMP antibody. Data obtained in 0.1 M Tris-HCI-EDTA, pH 7.5, using 1.0 ml of
nucleotide standard, 30/zl of 1:10 rabbit anti-cAMP antibody, 30 ttl of 1:10 uretase- cAMP conjugate, and 300 ttl of second antibody suspension.
cAMP to get equal inhibition, indicating the high selectivity of the antibody. However, the relative insensitivity (> 10-7 M) toward cAMP presented a problem if one wanted eventually to use such a system in physiological samples where cAMP levels are quite low (10 -a to 10-e M). Van Weemen and Schuurs 28 demonstrated that for EIA-hapten systems, the nature of the hapten linked to the enzyme can have a profound influence on sensitivity. They found that if the hapten of interest is linked to the enzyme in the same manner as the hapten was linked to the immunizing protein, a relatively insensitive assay may result because antibody production is elicited to the bridging group as well as the rest of the haptenic structure. Evidence of this type of antibody specificity in cAMP antibodies has been shown by RIA methods. Immunization with the 0 2'monosuccinyl derivative gives rise to antibody production with strong recognition of the ester linkage at the 0 2' position of the ribose ring. Initial acetylation of cAMP samples (at the 0 2' position) has brought forth increased sensitivity in the RIA method 31,~2 as a result of stronger affinity between antibody and the acetylated free cAMP. To increase sensitivity al j. E. Harper and G. Brooker, J. Cyclic Nucleotide Res. 1,207 (1975). 32 M. L. Goldberg, Clin. Chem. 23, 576 (1977).
452
1MMUNOASSAYS
100
[29]
AMP, G M P
80 ~_
~
cyclic
•-> 60 ~ cyclic
"X,,.
40
20
i
10 -9
i
i
10-8
10-7
i
10-6 n u c l e o t i d e conc., M
i
10-5
10-4
FIG. 7. Calibration curves obtained from inhibition response to cAMP, cGMP, AMP, and GMP, using a urease-cGMP conjugate and cAMP antibody. Data were obtained as for Fig. 6.
for EIA systems, Van Weemen and Schuurs showed that one could alter the site of hapten attachment to the enzyme, change the nature of the briding group (i.e., succinyl to glutaryl), or use a hapten structurally similar to that to be measured. This last approach was pursued here in an attempt to improve the sensitivity of the cAMP assay. Figure 7 shows a typical calibration curve for cAMP when using a u r e a s e - c G M P conjugate with cAMP antibody in the electrode-based system. Comparison with Fig. 6 illustrates the dramatic improvement in sensitivity obtained. Inhibition of label binding begins to occur at less than 10-9 M cAMP. The u r e a s e - c G M P conjugate is prepared with the O vmonosuccinyl derivative of cGMP and differs from the initial ureasecAMP conjugate and immunogen only by a guanine instead of adenosine moiety in the haptenic structure. This substitution effectively decreases the binding constant between the label and the cAMP antibody, thus allowing free cAMP to inhibit at lower concentrations. Direct comparison of the two conjugates is possible because final activity and degree of nucleotide conjugation were very similar for both. Figure 7 also shows the selectivity of this assay system over cGMP, GMP, and AMP. As expected, in switching to a u r e a s e - c G M P conjugate, selectivity over cGMP itself is reduced, but it still takes 20 times more cGMP than cAMP to produce the same amount of inhibition. This again
[29]
UREASE C O N J U G A T E S IN E N Z Y M E I M M U N O A S S A Y S
453
can be explained by the relative affinity of cAMP antibody for free cGMP vs cGMP conjugated to the enzyme. The succinyl group present in the enzyme conjugate causes greater affinity for enzyme-linked cGMP than for free cGMP. The system is highly selective over the corresponding noncyclic nucleotides AMP and GMP. This result agrees well with previous RIA systems, indicating that the cAMP antibodies have strong recognition of the cyclic phosphate ring. 19 Along this same line, a urease-cIMP conjugate was prepared and tested in the EIA system for cAMP. Figure 8 shows the resulting inhibition curves obtained for both cAMP and cGMP. Sensitivity for cAMP is not as good as when using the urease-cGMP conjugate and selectivity over cGMP also appears to be somewhere between that obtained with the other conjugates (approximately 80 times). These results seem appropriate in view of the fact that cIMP is more similar in structure to cAMP than to cGMP. It is important to note here that cIMP was not tested for inhibition because it has not yet been found to exist in physiological fluids at detectable levels. Cyclic GMP itself is present in serum and urine, but at levels 10 times lower than cAMP. 32 Furthermore, the antiserum to cAMP used throughout this work was taken from a single rabbit and there was no attempt to produce higher quality antisera in other rabbits, which, if obtainable, could lead to even more sensitive and selective assays.
• •
100
cyclic
•> 60
~
cyclic
4C
2C
i
10 -9
i
i
i
10-6
10-7
10-6
nucleotide
i
10-5
i
10-4
conc. , M
FIG. 8. Calibration curves obtained from inhibition response to c A M P and c G M P using a urcasc-clMP conjugate and c A M P antibody. Data were obtained as f o r Figs. 6 and 7.
454
IMMUNOASSAYS
[29]
REPRODUCIBILITY OF cAMP CALIBRATION CURVES
Percentage of activity 'z cAMP conc. (M)
Day 1
Day 2
Day 3
Day 4
Mean percentage of activity
2.5 × 10-s 2.5 × 10-7 2.5 × 10-e
66.1 34.8 8.7
67.7 33.1 8.5
64.0 30.1 10.0
69.0 33.0 8.0
66.7 32.8 8.8
Rates of 100% activity tubes were 11.8, 11.5, 10.8, and 11.5 mV/min for days 1-4; average: 11.5; relative standard deviation: 3.7%
Analytical precision of an EIA method is ultimately limited by how reproducibly one can assay enzyme concentration via activity determinations. Previous potentiometric activity determinations have resulted in excellent relative standard deviations of 10% or less. za'25"27In the preliminary part of this work, analysis of urease by our potentiometric system yielded similar precision, with maximum standard deviations at low urease levels (i.e., < 10-a M). Typical reproducibility of subsequent EIA calibration curves from day to day is summarized in the table. These data represent the relative percentage of activity values for three different cAMP concentrations when using a u r e a s e - c G M P conjugate (as in Fig. 6) over a 4-day period. It can be seen that excellent reproducibility is observed, indicating the good precision of the activity assay as well as the time stability of the reagents involved. In fact, all urease conjugates prepared in this work can be stored for at least 4 months without significant loss of immuno or enzymic activity. Moreover, working buffer conditions have been chosen so that possible inhibitors of urease that may be present in real samples (e.g., heavy metals) would not be expected to create a problem with the assays (EDTA preserves full activity of urease under physiological conditions). 17 We have developed in this work the techniques necessary to utilize urease as a label for EIA of both protein type antigens and low molecular weight haptens through the use of an ammonia electrode to measure bound enzyme. Furthermore, the first EIA procedure for cAMP has been demonstrated using urease-cyclic nucleotide conjugates. In view of the great interest in measuring cAMP in physiological samples, the EIA system described here should provide the basis for developing an attractive alternative to traditional cAMP assay procedures.
[30]
PASSIVE HEMAGGLUTINATION [30] Passive Hemagglutination and Hemolysis Estimation of Antigens and Antibodies
By
FRANK
L. ADLER
and
LOUISE
455 for
T. ADLER
Serological techniques provide specific and sensitive means for the detection and measurement of antibodies and antigenic substances. Among the many procedures that are available, some are more suitable for measuring antibody, others are better adapted for the assay of antigens. Passive hemagglutination (HA) and hemagglutination inhibition (HI) are here described and discussed as highly versatile techniques that require no specialized or expensive equipment and provide semiquantitative answers rapidly. A few remarks concerning the specificity and sensitivity of serological tests seem appropriate to assist the reader in the judicious application and interpretation of the procedures to be described. It is well to remember that the reaction between antibodies and antigenic substances is mediated by the specific binding of antibody to exposed antigenic determinants on the antigen molecule. The specificity of the antibody is for such determinants as, in native proteins, are thought to have the dimensions of tetrapeptides or, in substituted proteins, such as dinitrophenylated ovalbumin, may be as small as the substituent. Some unanticipated reactions of antibodies, superficially suggestive of lacking specificity, can often be traced to the sharing of antigenic determinants by the "cross-reactive" antigens. Antisera usually contain a mixture of antibodies differing from each other with regard to specificity as well as biological and other properties. Immunization with a highly purified antigen does not, in itself, assure the production of highly specific antibodies, and it is not unusual that a trace contaminant may evoke a disproportional amount of antibody. Since tests for the specificity of antisera are often based on precipitin procedures, such as Ouchterlony tests or immunoelectrophoresis, it is important to keep in mind that HA tests are 100-1000 times more sensitive in the detection of antibodies than are precipitin tests. Thus it is obvious that specificity must be ascertained by tests of appropriate sensitivity. Finally, in view of the advent of monoclonal antibodies produced in vitro or in vivo by cloned hybrids of myeloma and antibody producing cells, one should keep in mind that, while the specificity of such antibodies will be restricted to a single determinant, they will still react with diverse antigens that possess this determinant.
METHODS IN ENZYMOLOGY, VOL. 70
Copyright © 1980 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-181970-I
456 Passive
IMMUNOASSAYS
[30]
Hemagglutinafion
Principles In passive hemagglutination (HA) erythrocytes serve as the inert carriers of suitably affixed extraneous antigens, and the clumping Of such indicator cells by antibody specific for the coating antigen is a sensitive test for antibody. Under appropriate conditions the test measures antibody in concentrations of as little as 1-10 ng/ml. Since reasonably potent immune sera will contain at least 1 mg of antibody per milliliter it is apparent that their HA titer, defined as the highest dilution that will bring about agglutination, is often greater than 1:100,000. Since dilutions of the antisera to be tested are usually made in twofold steps, one deals with a semiquantitative test, and because antibodies sharing the same specificity but belonging to different classes, such as IgM or IgG, may differ in their HA activities, the test is more accurately viewed as one for activity rather than amount of antibody. Nevertheless, when applied to antisera obtained after intensive immunization and thus containing mostly IgG antibodies, the HA test results generally correlate reasonably well with those of binding assays for antibody. Soluble antigens can be attached to erythrocytes either through adsorption or by covalent linkage, often through the use of bifunctional agents. Whatever the procedure used, it is evident that firm attachment is required since antigen leakage from the erythrocyte would result in the binding and diversion of antibody and result in the specific inhibition of HA. This, indeed, is the principle of the hemagglutination-inhibition (HI) test to be described later. Some antigens adhere to red cells with sufficient tenacity and in adequate amounts to obviate the need for anything more elaborate than incubation of the cells with the antigen. Among them are many polysaccharides of microbial or protozoan origins. 1 Other antigens, such as most proteins, require treatments that either increase their ability to bind the antigen or their susceptibility to agglutination by antibodies against the coating antigen. The prototype for such procedures is the treatment of red cells with tannic acid, 2 which is to be described in detail below. Some antigenic determinants, such as dinitrophenyl groups, can either be attached to red cells directly3 or can be introduced into an antigenically irrelevant protein, which is then affixed to the erythrocytes.4 A typical method for 1 E. Neter, Bacteriol. Rev. 20, 166 (1956). 2 S. V. Boyden, J. Exp. Med. 93, 107 (1951). 3 M. B. Rittenberg and K. L. Pratt, Proc. Soc. Exp. Biol. Med. 132, 575 (1969). 4 F. L. Adler and C.-T. Liu, J. Immunol. 106, 1684 (1971).
[30]
PASSIVE HEMAGGLUTINATION
457
the covalent linkage of protein antigens to red cells is that employing bisdiazobenzidine, which is to be described in detail below. Procedures that have been successfully employed by various workers also include the use of chromic chloride, 5 water-soluble carbodiimide, e difluorodinitrobenzene, 7 toluene 2,4-diisocyanate, 8 glutaraldehyde, 9 and others. The choice is often determined by preferences of the investigator, but always requires selection of a procedure that will not block or destroy essential determinants and will yield cells of acceptable stability and sensitivity.
Materials and Reagents The HA test can be done in tubes (10 × 75) or in plastic hemagglutination trays. The latter are available either as permanent (Lucite) or disposable units. Those made of rigid plastic and with cone-shaped ( " V " plates) wells are preferred. When trays are used one can supplement them with other components of the Microtiter series, which include diluting devices to replace conventional pipettes and calibrated droppers that deliver 25or 50-/.d amounts. Alternatively, dilutions can be prepared in test tubes and aliquots can then be transferred into the wells of the trays with disposable Pasteur type pipettes of standard bore size that deliver approximately 30-/zl drops. When trays are used, static electricity can lead to difficulties; it should be reduced by wiping the trays with a moist tov~el or by use of an t~-particle emitting Staticmaster. An assortment of graduated conical (15 ml) and plain round-bottom tubes (12-15 ml), serological pipettes, flasks, and a bench-type centrifuge with a swinging bucket head complete the equipment needs. Antisera to be used in HA must be free of bacterial contamination and other foreign matter. It is also important that harvesting of the sera from the blood clot be postponed until the clotting is complete, since residual fibrin or fibrinogen can seriously interfere in HA. The use of preservatives, such as 0.01% sodium azide, is recommended, as is storage of aliquots of the sera at - 20°. Prior to this, sera should be heated at 56° for 30 min, and antibodies reactive with the erythrocytes to be used should be removed by three successive absorptions with washed, packed red cells. Generally 3 × 1 volume of packed cells for 9 volumes of serum with 30 min incubation at room temperature should suffice. Diluted antisera should not be stored because deterioration of HA activity may occur. 5 j. W. Goding, J. Immunol. Methods 10, 61 (1976). e E. S. Golub, R. K. Mishell, W. O. Weigle, and R. W. Dutton,J. lmmunol. 100, 133 (1968). r N. R. Ling, Immunology 4, 49 (1961). 8 D. E. Mahan and R. L. Copeland, Jr., J. lmmunol. Methods 19, 217 (1978). S. Lemieux, S. Avrameas, and A. E. Bussard, Immunochemistry 11, 261 (1974).
458
IMMUNOASSAYS
[30]
Dilutions of the antisera are generally made in 2-fold steps. If the initial dilution to be tested is less than 1:1000 it is advisable to use separate pipettes for each step because antibody from the more concentrated serum carried on the outside of the pipette can introduce serious errors. An initial 1:1000 dilution is most readily made in three successive steps of 10-fold dilutions. For the procedures to be described, phosphate-buffered saline (PBS) consisting of equal volumes of 0.15 M NaC1 and 0.15 M phosphate buffer, pH 7.2-7.3, is satisfactory as the basic diluent or as washing and suspending medium for the cells. It is used without additives in the washing of the erythrocytes and in preparing the original 5% suspension. For antiserum dilutions and for washing and suspending the coated cells the PBS is enriched by the addition of 1.0 ml of heated (56° for 30 min) and absorbed (red cells) normal rabbit serum to 100 ml of PBS. Fetal calf serum can also be used provided that it does not react with the antibody or the coating antigen. The serum addition is required for the maintenance of optimal suspension stability of the cells, which is a critical factor in the test because insufficient stability causes agglutination in controls and excessive stability reduces the sensitivity of the procedure. None of the various substitutes that have been tried have proved to be equal to normal serum. It is sometimes convenient to add some dye, such as Evans' blue, to give a slight tint to the diluent. This helps in differentiating empty from filled wells in the tray, especially when the trays sit on a nonglossy bright surface. Usually sheep red cells (SRBC) are used; human type O cells are equally suitable. Pooled SRBC are better than cells from single animals; they are available commercially in the form of whole pooled blood in modified citrate solution containing about 25% red cells. Such blood samples are stable at 4° for at least 1 month; they should be discarded at the first sign of spontaneous lysis or lysis during washing. Coated cells, prepared by either of the two procedures to be described, are stable for 1 week at 4° .
Procedure Preparation and Coating of Tanned SRBC. Citrated sheep blood is centrifuged in conical graduated tubes for 15 rain at 750 g. The supernatant is removed by suction, and the sedimented cells are thoroughly resuspended in the original volume of PBS. In this and all subsequent steps resuspension is best accomplished by first adding 1 ml of suspending fluid, suspending cells in this small volume with the aid of a Pasteur pipette with attached 2-ml bulb, and then adding the remainder of the suspending fluid.
[30]
PASSIVE HEMAGGLUTINATION
459
The cells are washed at least three times; the supernatant should be free of protein, which is detectable by adding a drop of nitric acid to a drop of supernatant. After the last wash the cells are suspended in PBS to yield a 5% (v/v) suspension. A freshly prepared solution of tannic acid, diluted 1:20,000 (w/v) in PBS is mixed with an equal volume of 5% SRBC. After 15 min at 37° the cells are harvested in round-bottom tubes by centrifugation for 8 min at 750 g. The supernatant is decanted by partial inversion of the tubes and by blotting the tube lips with filter paper. Properly tanned cells are sufficiently sticky to permit this procedure. After the tanned cells have been made into a 5% suspension in PBS, they are coated by mixing 1 volume of cell suspension with 4 volumes of antigen solution (in PBS) and incubating the mixture 15 min at 37° with occasional agitation. The cells are then harvested by centrifugation as described above, washed three times with PBS containing 1% normal serum, and then made into a 5% suspension in this medium. Coated cells stored at 4° should be washed once just before use on subsequent days. The concentration of antigen optimal for coating is the minimal concentration that will yield cells that provide the highest sensitivity in the HA test. It will vary with the nature and degree of aggregation of the antigen and, to some extent, with the avidity of the antisera to be tested. '° For most proteins the optimum will be in the range of 10-100 p,g/ml. Since the amount of antigen removed from the coating solution is minute, it is possible to reutilize coating antigen several times for the coating of successive batches of tanned cells provided dilution is avoided through addition of coating antigen to appropriate amounts of packed tanned cells. The coated cells should be kept at 4° until they are used in the test. Bisdiazobenzidine Procedure. To prepare bisdiazobenzidine (BDB) one dissolves 0.23 g of benzidine (a carcinogen) in 45 ml of 0.2 N HCI. The solution is chilled to 0° in an i c e - w a t e r - s a l t bath, and 0.175 g of NaNO~ in 5 ml of chilled water is added rapidly. The reaction is allowed to proceed for 30 min at 0°, with occasional mixing. A suitable preparation should, upon 15-fold dilution of a sample in PBS, turn brown and turbid in 90 sec at room temperature. If this occurs too fast, addition of a trace of 1% NaNO~ will correct it; if too slow it can be adjusted by addition of a trace of benzidine. The BDB solution is then distributed in 2-ml amounts into plastic tubes, which are then capped and rapidly frozen in a Dry Ice-ethanol bath. The reagent is stored at - 20° or - 70° and remains stable for up to 6 months. When red cells are to be coated, a 5% suspension is prepared exactly ,o G. Wolberg, C.-T. Liu, and F. L. Adler, J. Immunol. 103, 879 (1969).
460
IMMUNOASSAYS
[30]
as described for tanning. To 24 ml of a protein solution in PBS, add 12 ml of the cell suspension and 6.0 ml of BDB thawed and diluted 15-fold in PBS just before use. Rapid and thorough mixing of the reagents is essential. After 15 min at room temperature the cells are harvested, washed, and resuspended to a 5% suspension exactly as described for tanned cells. Here also PBS containing 1% heated normal serum is used. The spontaneous lysis of cells prepared by the BDB procedure, noted by earlier investigators, can readily be prevented either through the use of lesser amounts of BDB or through an increase in the concentration of the coating protein. To attain the desired protein concentration of 18 mg in the 24 ml of coating solution it is often desirable to use some protein that will not participate in the antigen-antibody reaction as a "filler." For example, in studies involving y-globulin as coating antigen we have used 12 mg of y-globulin (antigen) and 6 mg of bovine serum albumin in 24 ml. Coated cells remain in usable condition for at least 5-7 days at 4° . The useful life-span of either tanned or BDB cells can be extended indefinitely by mild fixation with formaldehyde, glutaraldehyde, or other agents.11 Freezing or lyophilization of such preparations is possible. Since these procedures entail some disadvantages and may be of more restricted utility, they will not be described here. Tube Assay. It is convenient to use 10 x 75 mm disposable tubes and 96-place racks to support these tubes. Serial 2-fold dilutions of the sera to be tested should be prepared in PBS containing 1% normal serum, leaving 0.5 ml amounts in each tube. As previously noted, separate pipettes should be used until a dilution of about 1:1,000 is reached. One then adds 0.05 ml amounts of a 2.5% suspension of the coated red cells, mixes the contents of the tubes thoroughly by vigorous shaking of the rack, and incubates for 2 hr at room temperature. One basic control to be included in the assay is a tube containing diluent and test cells (negative control); another is a tube containing antiserum in the highest concentration and cells coated with an irrelevant antigen (negative control). Another desirable control is discussed later as part of the description of the HI test. It is also advisable to select an antiserum to be used as the standard that will indicate variations in the sensitivity of different batches of coated cells (positive control). Hemagglutination patterns are read from below. Absence of agglutination is indicated by the formation of a compact button or a sediment resembling a doughnut, with smooth and regular edges. Agglutination, in contrast, is indicated by the presence of a mat of cells that covers the bottom. If agglutination is very strong this mat develops folds and contracts, leading to a highly irregular shape. 11T. Suzuki, S. Tanaka, and Y. Kawanishi,lrnmunochernistry 11, 391 (1974).
[30]
PASSIVE HEMAGGLUTINATION
461
Plate Assay. Dilutions of the antiserum can be prepared in the wells of the plates if the required spiral or loop diluting devices are available; otherwise they can be prepared in tubes as described above and 0.05-0.06 ml amounts can be transferred to wells using calibrated dropper pipettes (0.05 ml), automatic pipetting devices, or 2 drops from disposable Pasteur pipettes (0.06 ml). One then adds to each well a matching volume of a 0.5% suspension of coated erythrocytes in PBS containing 1% normal serum. The contents of the wells are mixed by vigorous rotating of the tray while it rests on the laboratory bench. The results are read after 1,52 hr of incubation at room temperature. In the V-shaped cups nonagglutinated cells slide to the bottom (tip of the cone) and, as seen from above, appear as a compact button. Agglutinated cells adhere to each other and to the sides of the wells, forming a matlike pattern. Controls and Interpretation It should be stressed again that HA is a semiquantitative test for antibody that, applied to antisera that contain largely IgG antibodies, yields results in relatively good agreement with precipitin assays and binding tests but provides greater sensitivity than the former and greater simplicity than the latter. Essential controls have been described in the preceding paragraphs. One must keep in mind that antigenic impurities in the coating antigen and the presence of corresponding antibodies in antisera may lead to confusion. Since effective coating is generally highly dependent on relatively high concentrations of antigen in the coating solution, we find it advisable to employ the minimal concentration of coating antigen that will assure acceptable, though not necessarily maximal, sensitivity. The use of "filler" protein to meet the requirements of the BDB procedure has been discussed. R e v e r s e Passive Hemagglutination This procedure employs erythrocytes coated with a preparation of purified antibody. It is suitable for the detection of trace amounts of antigen and its semiquantitative measurement. For example, SRBC coated with anti-y-globulin antibody can be used to detect 3,-globulin in concentrations of 1-10 ng/ml. In this reaction the antigen in solution is the agent that links the antibody-coated cells into aggregates. Since most protein antigens are multivalent with respect to their antigenic determinants they are more effective ligands than divalent IgG antibodies. A major shortcoming of the method is the inhibition of agglutination in the presence of excess antigen (prozone) which may occur when cross-linkage is prevented by the excess of antigen binding to all available antibody combining sites on the test cells.
462
IMMUNOASSAYS
[30]
Specific antibody is generally prepared by its absorption to and elution from solid phase immunoadsorbants. Ideally it should be free not only of nonantibody serum proteins but also of antibodies against contaminating antigens. Using the BDB procedure previously described one coats SRBC with a mixture of purified antibody and "filler" protein, such as bovine serum albumin, in a proportion such as 6 mg of antibody plus 12 mg of albumin per 24 ml of coating solution. The optimal amount of antibody must be determined for each preparation. The test itself is done exactly in the same manner as HA. Because the end point in this procedure is not always reached in a single dilution step from complete agglutination to its total absence, it is advisable to include a standard with known concentrations of antigen. It is not difficult to estimate end points by referring to the standard. Hemagglutination Inhibition
Principles The hemagglutination-inhibition test detects and measures antigen with extremely high sensitivity. It is based on the principle of competition for a finite and limiting amount of antibody by antigen in two forms: that coating the indicator red cells and that present in a solution that is to be tested for its antigen content. The sensitivity of the test for protein antigens of molecular weights 50,000 to 200,000 is generally about 0.1 to 1.0 /zg/ml. For smaller molecules, such as morphine (M = 285) it is about 100 to 1000 times more sensitive. The reason for this difference stems from the fact that 2 mol of antigen are required to saturate 1 mol of divalent antibody; thus the relative inhibitory efficiency of antigens (haptens) are inversely proportional to their molecular weights.
Procedure A preliminary HA test is required to determine the HA titer of the antiserum that is to be used. It is advisable to use in the HI test a dilution of the antiserum that is 4-8 times less than that corresponding to its HA titer. An amount of this dilution sufficient for a day's work is prepared in PBS containing 1% normal rabbit serum. It is kept cold until use, and it is not recommended to store it for use on subsequent days. If the tube assay is to be employed, 0.25 ml amounts of serially diluted antigen, made in the same diluent, and 0.25 ml amounts of the selected antiserum dilution are mixed and incubated for 30 min at room temperature. The test cells (0.05 ml of 2.5% suspension) are added, the tubes are shaken virogously, and the agglutination patterns are read after 1.5-2 hr
[30]
PASSIVE HEMAGGLUTINATION
463
of incubation at room temperature. In the plate assay 0.025 ml or 0.03 ml amounts of antigen dilutions and of antiserum, respectively, are used and 0.05 ml of 0.5% coated cells are used. A series of dilutions of antigen of known concentration serve as control and standard and are to be included in each day's titrations. The end point is the highest dilution of the antigen solution that causes complete inhibition of HA. The minimal amount of antigen required for total inhibiton is indicated by the standard. Controls should be included to rule out agglutination of the coated cells by the antigen solution (omit the antiserum), and to rule out inhibition of agglutination by nonimmunological activities of the test solution, such as proteolysis. For this purpose test cells coated with an unrelated antigen and the corresponding antiserum can be used. Optimal conditions for HI call for the use of (a) highly purified antigen in the coating of the red cells; (b) antiserum that in the dilution in which it is to be used is free of effective (agglutinating) amounts of antibodies against residual antigenic contaminants of the coating solution; and (c) dilution of the antiserum to near its agglutination end point to assure both maximal specificity and sensitivity. With these precautions it is generally possible to test for the presence and amount of a given antigen in crude preparations, and thus to have a convenient assay for a given antigen in biological fluids or to monitor progress toward purification of an antigen. Passive Hemolysis The passive hemolytic (PH) test is an extension of HA in which an additional reagent, complement, causes lysis of test cells that is mediated by antibody against the coating antigen. In some instances this modification leads to further enhancement of sensitivity, but this advantage is balanced by a number of complicating factors. Dominant among these is the introduction of complement as an additional variable. Choice of the hemolytic test also imposes limits on the method for linking antigen to the red cells, since some procedures enhance and others diminish the susceptibility of erythrocytes to lysis. Although PH is not widely used in assays for serum antibodies it provides a most useful method for the enumeration and identification of lymphoid cells that produce and secrete antibodies against a given antigen. In a further extension, as passive reverse hemolysis, it is a test applicable to enumeration, identification, and isolation of cells that secrete a given antigen. The basic principle common to the several forms of the test calls for mixing a suspension of cells to be tested with a suitable excess of coated red cells (targets), to incubate the mixture under conditions that prevent movement of the cells, and to score circular clear (hemolytic)
464
IMMUNOASSAYS
[30]
plaques surrounding individual test cells. The complement is added in some procedures to the initial incubation mixture; in others it is added after an initial incubation period. The plaque assay, originally described for cells secreting antibodies against erythrocytes, lz has been modified by numerous workers to allow observations on cells that secrete antibodies specific for antigens that can be effectively attached to red cells. One method and one example only will be described here, namely, an assay for spleen cells making antibody against the trinitrophenyl determinant. Alternate procedures are well described elsewhere. 13
Materials and Reagents Trinitrophenylated SRBC are prepared as originally described? In brief, 20 mg of trinitrobenzene sulfonate are dissolved in 7.0 ml of 0.28 M cacodylate buffer, pH 6.9; 2.0 ml of washed 50% SRBC in complement buffer (5,5-diethylbarbituric acid, 0.575 g; sodium 5,5-diethylbarbiturate, 0.375 g; CaCI~, 0.017 g; MgClz, 0.048 g; NaCI, 8.5 g; distilled water to 1 liter) are added while the mixture is gently stirred. After 10 min at room temperature one adds 6.0 ml of cold complement buffer, harvests the cells by centrifugation at 750 g for 10 min, suspends them in 12 ml of complement buffer containing 7.5 mg of glycylglycine, and washes them three additional times in complement buffer. A 10% suspension is then made in the same buffer. It should be remembered that the reaction and product are sensitive to light. The assay can be conveniently done in glass or plastic petri plates (60 mm) containing about 4 ml of 0.8% agarose in complement buffer. Guinea pig complement is readily available commercially and generally does not require absorption. Lymphoid cells are prepared by teasing spleens or lymph nodes in cold tissue culture medium buffered with Hepes. If the "indirect" plaque assay is to be done, an antiserum specific for IgG of the antibody-secreting cells is also required.
Procedure The freshly teased and uniformly suspended lymphoid cells are adjusted to some convenient starting concentration, such as 107 viable cells per milliliter. A series of dilutions are made, and 0.1 ml amounts are transfered into 10 × 75 mm tubes containing 0.8 ml of 0.8% agarose at 12 N. K. Jerne, C. Henry, A. A. Nordin, H. Fuji, A. M. C. Koros, and I. Lefkovits, Transplant Rev. 18, 130 (1974). in W. J. Herbert, in "Handbook of Experimental Immunology" (D. M. Weir, ed.), 2nd ed., Vol. 1, Ch. 20, Blackwell, Oxford, 1973.
[30]
PASSIVE HEMAGGLUTINATION
465
45 °. Aliquots of 0.1 ml of 10% coated erythrocytes are quickly added, and the contents of the tubes are poured over the bottom layer of agarose in the prepared plates. Incubation for 1 hr at 37° in a tissue culture incubator is followed, in the "direct" assay, by a second incubation for 1 hr at 37°, in air, after the addition of 1 ml of guinea pig complement diluted 1:10 in complement buffer. This procedure develops plaques surrounding cells that make hemolytically efficient antibodies, such as IgM. For more comprehensive count of antibody-secreting cells, an incubation for 1 hr at 37° in air is interpolated between the two incubations mentioned. In this additional step 1 ml of a suitable dilution of an anti-IgG serum is placed on the top layer of agarose; it is removed prior to the addition of complement. This assay yields an estimate of total antibody-secreting cells. Plaques are counted under low magnification under a dissecting microscope, and results are expressed as the number or proportion of secreting cells per total cells applied. An essential control is one for the specificity of plaque formation. In the example cited this consists of plates in which trinitrophenylated serum albumin (100 ~g/ml) is incorporated into the top layer agarose as a specific inhibitor. Plaques forming in these plates are presumed to be antiSRBC, and correction is made by subtracting their number. Another essential control omits complement. This description of the plaque assay is lacking in some details but it will suffice to introduce the reverse hemolytic plaque assay, which probably will be more relevant to the readers' interests. Reverse Hemolytic Plaque Assay This procedure is of great potential utility for the identification and enumeration of cells that secrete an antigenic product. Patterned after the assay just described, it employs similar techniques with two major differences: The coating for the indicator red cells consists of purified antibody, as in the reverse passive hemagglutination previously described~ and an antiserum against the secretion product is used as the developing agent. 14,15 In an application of the method to the detection of murine hepatocytes that secrete albumin, the authors 14 employed SRBC coated with purified antibody against mouse serum albumin and plated such cells together with varying numbers of teased liver cells exactly as described for the hemolytic plaque assay. After the initial incubation, anti-mouse serum albumin x4G. A. Molinaro,E. Maron, W. C. Eby, and S. Dray,Eur. J. lmmunol. 5, 771 (1975). 15W. C. Eby, C. A. Chong, S. Dray, and G. A. Molinaro,J. Immunol. 115, 1700(1975).
466
IMMUNOASSAYS
[30]
antiserum in a suitable dilution was added for 1 hr at 37° in air. This reagent was decanted, and incubation at 37° in air was continued for another hour in the presence of 1.0 ml of guinea pig complement 1:10. The resulting plaques are counted, and the results are expressed in terms of number of secreting cells per number of viable cells. Rosette Test for Cell M e m b r a n e Antigens In contrast to the method just described, the rosette test is applicable to antigens that are membrane-bound or integral components of the cell membrane. It has been used to detect membrane-bound Ig on lymphocytes and will be described in terms of this model, but has unquestionably wider applicability. The principle of the test is based on the specific binding of antibody-coated erythrocytes to the corresponding antigens on the cell membrane as observed by the specific adherence of such indicator cells. In general, adherence of at least three red cells is the minimal criterion for a rosette. Cells that actively secrete the antigen in question may not form rosettes and, in fact, may inhibit rosetting through specific binding of the antibody on the red cells by the secreted antigen. It is believed that rosetting matches or exceeds fluorescent antibody techniques in sensitivity for the detection of membrane antigens. The antibody-coated red cells are prepared as previously described. It is particularly important for this procedure that the indicator red cells are absolutely free of clumps. The sensitivity of coated cells can be assayed by reverse passive hemagglutination if, as in the model under consideration, the antigen is available in soluble form. The cells under study are washed in suitable tissue culture medium or other buffered solution and suspended at a concentration of 107 per milliliter in the same diluent to which serum has been added (usually 1% fetal calf serum). A small volume (50-100/zl) of the cell suspension is placed in a 10 x 75 mm disposable tube. The addition of an equal volume of 1% coated red cells results in a mixture that contains about 25 red cells per lymphocyte. Linkage of antibody on the red cells to the corresponding antigen determinant on the surface of the lymphocyte results in the formation of a "rosette" or lymphocyte surrounded by red cells. The mixing of cells and incubation for at least 1 hr are done in an ice bath. The tubes are then centrifuged very briefly (1 min at 1000 g), and a drop of dye is added to tint the lymphocytes (e.g., crystal violet or brilliant cresyl blue). The mixture is then aspirated four or five times with a Pasteur pipette and examined in a hemacytometer chamber at about 400 x. A cell is scored as a rosette if it is surrounded by three or more adherent erythrocytes, and usually 300 cells are counted.
[31]
MICRO COMPLEMENT FIXATION
467
[31] Quantitative Micro Complement Fixation: Serologic Properties of Pig Liver Carboxylesterase
By
LAWRENCE LEVINE, AUGUSTIN BAER, and WILLIAM P. JENCKS
Changes in quaternary structure of macromolecules are sometimes accompanied by changes in their serologic properties. Some of these altered serologic activities reflect spatial rearrangement of the antigenic determinants whose conformations depend on the structural integrity of other parts of the molecule, or changes in density of such antigenic determinants. 1,2 Quantitative micro complement fixation 3 is a most sensitive serologic method for detection of conformational changes in macromolecules. Denaturation of D N A and polyribonucleotides, 4-s altered conformation of lactic dehydrogenases, ° hemoglobin, 1°-12 myoglobin, 1° collagen, 13 lysozyme, 14 aspartate transcarbamylase, 3,15 pepsinogen and pepsin, ~e-~a carboxypeptidase, 2° and S-100 brain protein zl,zz have been detected and 1 M. Reichlin, M. Hay, and L. Levine, lmmunochemistry 1, 21 (1964). z M. R. Bethell, R. von Fellenberg, M. E. Jones, and L. Levine, Biochemistry 7, 4315 (1%8). 3 E. Wasserman and L. Levine, J. Immunol. 87, 290 (1961). 4 L. Levine, E. Wasserman, and W. T. Murakami, lmmunochemistry 3, 41 (1966). 5 W. T. Murakami, H. Van Vunakis, L. Grossman, and L. Levine, Virology 14, 190 (1%1). 6 D. Stollar, L. Levine, H. I. Lehrer, and H. Van Vunakis, Proc. Natl. Acad. Sci. U.S.A. 48, 874 (1%2). r L. Levine, Fed. Proc. 21, 711 (1%2). a B. D. Stollar, in "The Antigens" (M. Sela, ed.), Vol. 1, pp. 1-85. Academic Press, New York, 1973. 9 R. D. Cahn, N. O. Kaplan, L. Levine, and E. Zwilling, Science 136, 962 (1962). 10 M. Reichlin, M. Hay, and L. Levine, Biochemistry 2, 971 (1%3). 11 M. Reichlin, Adv. lmmunol. 20, 71 (1975). 12 M. Reichlin, M. Hay, and L. Levine, Immunochemistry 2, 337 (1965). lap. F. Davison, L. Levine, M. P. Drake, A. A. Rubin, and S. Bump, J. Exp. Med. 126, 331 (1%7). 14 R. von Fellenberg and L. Levine, Immunochemistry 4, 363 (1967). 15 R. von Fellenberg, M. R. BetheU, M. E. Jones, and L. Levine, Biochemistry 7, 4322 (1968). 16 H. Van Vunakis, H. I. Lehrer, W. Allison, and L. Levine, J. Gen. Physiol. 46, 589(1%3). lr j. Gerstein, H. Van Vunakis, and L. Levine, Biochemistry 2, 971 (1%3). is j. Gerstein, L. Levine, and H. Van Vunakis, Immunochemistry 1, 3 (1964). 19 T. G. Merrett, L. Levine, and H. Van Vunakis, Immunochemistry 8, 201 (1971). 2o H. I. Lehrer and H. Van Vunakis, Immunochemistry 2, 255 (1965). 21 D. Kessler, L. Levine, and G. D. Fasman, Biochemistry 7, 758 (1968). 22 p. S. Dannies and L. Levine, J. Biol. Chem. 246, 6276 (1971).
METHODS IN ENZYMOLOGY,VOL. 70
Copyright© 1960by AcademicPress, Inc. All rightsof reproductionin any formreserved. ISBN 0-12-181970-1
468
IMMUNOASSAYS
[31]
I00
80
60
2O
0.01
0.05
0.1
0.2
/u.g Carboxylesterose FIG. 1. Complement (C) fixing activity of native and dissociated molecules mixed in varying proportions: 100% dissociated ([~); 80% dissociated, 20% native (A); 60% dissociated, 40% native (©); 40% dissociated, 60% native (11); 20% dissociated, 80% native (A); 100% native (@). [From L. Levine, A. Baer, and W. P. Jencks, Arch. Biochem. Biophys., in press (1980), with permission.]
quantified by micro complement fixation. In addition, the extreme sensitivity of the method has been exploited to measure molecular changes induced by evolutionary processes. ~-z7 The principle and the procedures for performing the micro complement fixation test have been presented in detail previously, z8"20Here, we record the use of micro complement fixation to measure the rates of dissociation and the equilibrium constants for dissociation of pig liver carboxylesterase (EC 3.1.1.1) as a function of pH and salt concentration. The pig liver carboxylesterase preparation and some of its physical properties have been described by Barker and Jencks. 3° Antibodies to the carboxylesterase preparation were obtained by immunization of rabbits via the toe pads and muscles with 2 mg of enzyme emulsified in complete 23 A. C. Wilson, N. O. Kaplan, L. Levine, A. Pesce, M. Reichlin, and W. S. Allison, Fed. Proc. 23, 1258 (1964). 24 A. H. Tashjian, Jr., L. Levine, and A. E. Wilhelmi, Endocrinology 77, 563 (1965). 25 L. Nonno, H. Herschman, and L. Levine, Arch. Biochem. Biophys. 136, 361 (1969). E. M. Prager and A. C. Wilson, J. Biol. Chem. 246, 7010 (1971). 27 N. Arnheim, in "The Antigens" (M. Sela, ed.), Vol. 1, p. 377. Academic Press, New York, 1973. 2s L. Levine and H. Van Vunakis, this series, Vol. 35, p. 928. 29 L. Levine, in "Handbook of Experimental Immunology" (D. M. Weir, ed.), p.22.1. Blackwell, Oxford, 1973. 30 D. L. Barker and W. P. Jencks, Biochemistry 8, 3879 (1969).
[31]
MICRO COMPLEMENT FIXATION
469
90 8O
g
70
-.,7.
u_
60
¢.)
E
5O
.E K O
40 3O 20 0
I 20
I 40
I 60
I
I
I
I
I
I
I00
80
60
40
20
0
%
%
I 80
I I00
Notive
Dissocioted
FiG. 2. Maximum complement (C) fixing activity of native and dissociated ~holecules mixed in varying proportions. [From L. Levine, A. Baer, and W. P. Jencks, Arch. Biochem. Biophys., in press (1980), with permission.]
• •
8
x~x=
I00
=
o
75
"2_
N gl ILl
0
o
6
50 .-
O9
5
Z
25 ~.
0 3
I 2
I 3
I 4
I 5
I 6
I 7
I 8
pH FIG. 3. Complement-fixing activity (×) after incubation of esterase (0.1 mg/ml) for 40 min at indicated pH. The sedimentation coefficient (s=o) as a function of pH for 2 mg/ml esterase. [From L. Levine, A. Baer, and W. P. Jencks, Arch. Biochem. Biophys., in press (1980), with permission.]
470
IMMUNOASSAYS
[31]
Freund's adjuvant. At monthly intervals, the rabbits were boosted intramuscularly with 1 mg of enzyme, again emulsified in complete Freund's adjuvant. Immune sera were collected 1 week after each booster injection. The antiserum used in the immunochemical application described here was collected 1 week after the sixth boost. Estimation of Native Carboxylesterase Molecules Varying quantities of carboxylesterase were incubated with a constant amount of rabbit anti-carboxylesterase (1.0 ml of a 1:7000 dilution) and a constant level of complement to give a complete complement fixation curve and a maximum of 80% complement fixed with 0.05/zg of carboxylesterase. Pig liver carboxylesterase (2 mg/ml) undergoes a decrease in sedimentation rate over a relatively narrow pH range (pH 4-5) that reflects dissociation of the whole molecule into subunits. The serologic activities of these whole and dissociated molecules are changed. Whereas 80% of the complement is fixed with the whole molecules, only about 20% is fixed with the dissociated molecules. Maximum complement fixation is still observed with about 0.05/zg of the dissociated protein. A decrease in complement fixing properties accompanies dissociation. In order to relate dissociation and decreased complement fixing activity, dissociated molecules were mixed with the whole molecules in varying proportions, and complement-fixing activities of the admixtures were determined (Fig. 1). The whole molecules reacted more effectively than mixtures of the dissociated and native molecules. Moreover, the decrement of maximum complement fixation with the several mixtures of whole and dissociated molecules reflected their percentage, in weight. This decrement probably results from inhibition by dissociated molecules of the complement fixed with whole molecule s; i.e., the antigenic determinants of the dissociated molecules are recognized by the antibodies, but lattice formation requisite for complement fixation zl is affected by the changed density of these antigenic determinants. With all these mixtures, maximum complement fixation is still obtained with about 0.05/zg of esterase. Maximum complement fixation, as a function of the percentage of whole and dissociated molecules, is shown in Fig. 2. Such a calibration curve was used to estimate the number of native molecules in unknown carboxylesterase solutions. The sedimentation coefficients and the percentage of native molecules, as measured serologically, remaining in carboxylesterase preparations incubated for 30 min at 20° at various pH values are shown in Fig. 3. 31 A. G. Osier and B. M. Hill, J. Immunol. 75, 137 (1955).
IxlO - 9
I x l 0 -10
Ixl0 -u
o/
/
I xl0 -i2 N ~E Cr Y
o I x l 0 -13
I xlO-14
I x l 0 -15 0
I xlO -16
4
0
0
0
I
I
i
I
I
I
I
5
6
7
8
9
10
II
pH FIG. 4. Equilibrium constants, as measured by complement (C) fixation, f o r dissociation o f pig liver carboxylcstcrasc as a function o f pH. [From L. Levinc, A. Baer, and W. P.
Jcncks, Arch. Biochem. Biophys., in press (1980), with permission.]
472
IMMUNOASSAYS
[31]
There is a sharp decrease in sedimentation properties between pH 4 and 5. There is also a sharp decrease in the content of native molecules over this pH range. In this experiment and in all subsequent experiments on dissociation of carboxylesterase, further dissociation or association during the course of serologic analysis was stopped, or at least minimized, by dilution of the reaction mixtures into ice-chilled Tris buffer (10 mM Tris pH 7.4, 0.14 M NaCI containing 0.1% gelatin) to give a carboxylesterase concentration of 0.1 ~g/ml (in some cases to 0.05 ~g/ml), and immediate addition of the diluted enzyme solution into the antiserum and complement for the complement fixation assay. The combination of dilution to pH 7.4 at low temperature and the interaction of the antibodies with the carboxylesterase, which is very rapid, effectively prevents continuous dissociation. While this is difficult to prove unequivocally, the following considerations attest to its validity. Antibodies to pig liver carboxylesterase quantitatively precipitate the enzyme in the regions of the equivalence and excess antibody zones of the precipitin curve. All the catalytic activity of the carboxylesterase is recovered in these immune precipitates, suggesting that the enzyme is catalytically active in the associated state. In our complement fixation procedure, in which the concentrations of antibodies and carboxylesterase are around 10-l° M , the reaction of antibody with the associated molecule would be sufficiently rapid to inhibit continuous dissociation of the carboxylesterase. It should be recalled that complement fixation analyses are performed at pH 7.2-7.4 where the associated state is most stable, and at 2-4 °, where the rate of dissociation is also relatively slow. Determination of Equilibrium Constants Studies by Junge and Krisch 32 and by AuneY reported in 1973, of the subunits of pig liver carboxylesterase suggest that the whole enzyme is composed of three subunits of about 60,000 molecular weight. Thus, our values for the equilibrium constants for dissociation were calculated according to the equation K~q = [T]3/[W] where T and W refer to the molar concentrations of third and whole molecules. In experiments designed to estimate equilibrium constants, varying quantities of carboxylesterase were incubated at specified conditions for 24 and 48 hr. At the time of measurement, the carboxylesterase was di32 W. Junge and K. Krisch, Mol. Cell. Biochem. 1, 41 (1973). aa K. Aune, Arch. Biochem. Biophys. 156, 115 (1973).
[31]
MICRO COMPLEMENT
FIXATION
473
TABLE I PERCENTAGE OF COMPLEMENT FIXATION AND EQUILIBRIUM CONSTANTS FOR DISSOCIATION 24 AND 48 HR AFTER INCUBATION IN 10 m/~ TRIS, p H 7.2, AT 35 °a
% Complement fixation
% Whole molecules
Keq (10-1° M 2) W ~ 3T
Carboxylesterase (/,~g/ml)
24 Hr ~
48 Hr ~
24 H r b
48 Hr e
24 Hr
48 Hr
1.0 2.0 3.0
43 60 59
49 60 63
33 54 53
35 49 53
8.4 6.9 18.2
7.6 13.5 18.0
a From L. Levine, A. Baer, and W. P. Jencks, Arch. Biochem. Biophys., in press (1980), with permission. b 76% maximumcomplement fixation with untreated carboxylesterase in these analyses. c 79% maximumcomplement fixation with untreated carboxylesterase in these analyses.
luted into ice-chilled Tris buffer (pH 7.4, 10 m M Tris, 0.14 M NaC1 containing 0.1% gelatin) to contain 0.15, 0.1, or 0.05/zg of esterase per milliliter and assayed for serologic activity. E v e n at a p H around neutrality and at low ionic strength, equilibrium has been reached by 24 hr, so that in most of our experiments only one time o f incubation (24 hr) was used. To illustrate the data generated in these complement fixation experiments and to measure equilibria after 24 and 48 hr of incubation, an experiment was performed to estimate equilibrium constants in 10 m M Tris, pH 7.2, at 35 °. At 24 and 48 hr, aliquots were diluted to contain 0.1/zg/ml, and complete complement fixation curves, similar to those shown in Fig. 2, were obtained. At the same time, the complement fixation o f a carboxylesterase solution not previously incubated was obtained in order to normalize day-to-day variation in the technique. The percentage of native molecules in the carboxylesterase preparations was calculated from their normalized maximum complement fixation values and the calibration curve shown in Fig. 2. The equilibrium constants at 24 hr (1.1 x 10-15 M 2) and at 48 hours (1.3 × 10-15 M z) are in good agreement (Table I). The equilibrium constants for dissociation as a function of pH are shown in Fig. 4. R a t e of Dissociation of Pig Liver C a r b o x y l e s t e r a s e As was expected from the experiments in which complement fixation was measured with varying mixtures of whole and dissociated esterase (Fig. 1), a decrease in complement fixation is observed with increased dissociation. Moreover, the point at which maximum complement fixation is obtained, around 0.05/~g, did not change. Such a series o f complement
474
IMMUNOASSAYS
[31]
I00
80
/"~\ //
\\
~ ' \ \ o
u
40
2O
0.01
0.05
0.1
0.2
# g Corboxylesterose
FIG. 5. Complement-fixing activity of carboxylesterase after incubation for varying periods of time in 10 mM sodium acetate buffer (pH 4.5). No incubation (1); 5 min (A); 10 min (11); 20 min (O); 40 min (A); 60 min (D); 120 min (x). [From L. Levine, A. Baer, and W. P. Jencks, Arch. Biochem. Biophys., in press (1980), with permission.]
I00 8O 6O
=
40
Q
1~
20
I0
I
0
~l
I
I
20
40
I
I
60 80 Time (Minutes)
I
I00
I
I
120
140
FIG. 6. Dissociation of carboxylesterase at pH 4.0 (©), pH 4.8 (O), and pH 5.5 (&) as measured by complement (C) fixation. [From L. Levine, A. Baer, and W. P. Jencks, Arch. Biochem. Biophys., in press (1950), with permission.]
[31]
MICRO COMPLEMENT FIXATION
475
1.0
0.5 0.3
0.1
T" .S
0.05
0.03 T
/ 0.01
O.OO5 O.003
0.001 4.0
I
I
I
I
I
I
I
5.0
6.0
7.0
8.0
9.0
I0.0
I1.0
pH
FIG. 7. Effect o f p H on the rates of dissociation of carboxyesterase as measured by complement (C) fixation. [From L. Levine, A. Baer, and W. P. Jencks,Arch. Biochem. Biophys., in press (1980), with permission.]
fixation curves was obtained when carboxylesterase (10/~g/ml) was incubated at pH 4.5 for various periods of time (Fig. 5). Therefore, a procedure for measuring the rate of dissociation by removing 0.05/.~g from the reaction mixture and adding it directly to the complement fixation system was used. This level of esterase gives maximum complement fixation and reflects the percentage of whole molecules remaining. The percentage of complement fixation and the percentage of native molecules calculated from that value during dissociation of carboxylesterase at pH 4.0, 4.8, and 5.5 are shown in Table II. The rates of dissociation are shown in Fig. 6. The rates of dissociation as a function of pH are shown in Fig, 7.
0
,% Z