■Jli-il^:::^^S^^: Jmmuno-Catalysis Immuno-Catalysis And Related Fields of Bacteriology and Biochemistry (Second Edition, Revised and Enlarged) By M. G. SEVAG, Ph.D. Associate Professor Bepartment of Bacteriology School of ^Medicine, Vniversity of Pennsylvania Philadelphia, Pennsylvania With a Preface By STUART MUDD, M.A., M.D. Professor of Bacteriology School of Medicine, Vniversity of Pennsylvania Philadelphia, Pennsylvania CHARLES C THOMAS • PUBLISHER Springfield • Mnois • U.S.A. CHARLES C THOMAS • PUBLISHER Bannerstone House 301-327 East Lawrence Avenue, Springfield, Illinois Published simultaneously in the British Commonwealth of Nations hy BLACKWELL SCIENTIFIC PUBLICATIONS, LTD., OXFORD, ENGLAND Published simultaneously in Canada hy THE RYERSON PRESS, TORONTO This monograph is protected by copyright. No part of it may be reproduced in any manner without written permission from the publisher. Copyright 1951, hy CHARLES C THOMAS • PUBLISHER First Edition, 1945 Second Edition, 1951 Printed in the United States of America Pref ace CULTIVATION of the Special fields of science at first tends to express the attributes of the particular subjects of investigation, the immediate practical objectives and the technical instrumentalities which prove serviceable in each field. Specialization of investiga- tional method and equipment, of vocabulary and of the body of accumulating fact and conclusion may become extreme, resulting in an unfortunate degree of isolation of the field and its workers. With growing insight, however, the special lore of the particular field is gradually seen to represent special aspects of the same truths which may be perceived also from other fields, to exemplify laws of more general validity. Isolationism in science is then a phase of narrow understanding and immaturity. Integration enriches the special fields in technique and content, and broadens the horizons of the workers. The studies of Pasteur on the alcoholic, and tartaric, butyric and acetic acid fermentations brought about by particular micro-organisms were certainly among the principal original sources of scientific insight into the specific causation of infectious diseases. Interrelationships between the phenomena of immunity and of biocatalysis, also, have been perceived by men of discernment working in both fields, Ehr- lich, Morgenroth, Zinsser, Landsteiner, Avery, Marrack, Todd and Northrop, to mention but a few. Ehrlich, indeed, borrowed his famous lock and key simile expressing the specificity of antibody with respect to antigen from Emil Fischer's studies on enzyme specificity. As a matter of practice, however, physiological chemists working with enzymes have only rarely concerned themselves with infectious disease and immunity, and bacteriologists and immunologists have paid too little heed to the many developments in enzymology paralleling devel- opments in their own fields. The fullness of the integration possible between the fields of enzyme chemistry, immuno-chemistry and the mechanisms of infectious disease, has, indeed, in the writer's belief, VI PREFACE been indicated for the first time in this volume by M. G. Sevag, Immuno-Catalysis. In Sevag's treatment of immuno-catalysis we discern that enzyme, substrate, and specifically inhibitive reaction products, have their respective counterparts in antigen, the antibody precursors, and spe- cific antibodies. This is no mere analogy, for antigens in truth do determine the specificity of newly forming antibodies by a catalytic mechanism; many protein enzymes have been proved to be antigenic; the chemical configurations upon which specificity depends in enzyme and immune reactions are often analogous or identical. Specialists either in enzymology or immunology will find the first three sections of the book a reservoir of experimental fact, considered with insight and woven together in a remarkable synthesis. The enzymes of plants which are pharmacologically active, the enzymes of snake venoms and the enzymes of pathogenic bacteria are considered in detail with reference to their chemical activities in vitro, their pharmacodynamic actions in vivo, and with respect to specific immunity against them. Consideration of this array of data, which is little known to most students of immunity and of infectious disease, leads the reader to the conclusion that many diverse symptoms and lesions of infectious dis- ease may find their explanation in terms of the action of biocatalysts of the parasite which can alter vital substrates of the host. Much of the pathological symptomatology of bacterial and viral disease may thus ultimately come to be understood as special manifestations of enzyme action. Chemical and pharmacological studies with the lecithinases, proteases and nucleases of snake venoms and of those associated with pathogenic Clostridia, and studies of hyaluronidase and fibrinolysin afford striking pertinent cases. Success in research and success in teaching are in no small part dependent upon discernment of the interrelationships of things; the deeper the insight into phenomena which seem superficially unrelated, the more clearly they may often be perceived to rest upon more funda- mental, general laws of matter and energy. Teachers and investigators dealing with biochemistry, with bacteriology and immunity and with the mechanisms of infectious disease will, I believe, find in this in- formative book an expanded horizon and a challenge to further investigations. PREFACE VU In the five years between the appearance of the first and the present edition of Immuno-Catalysis the intimate relationships between biocatalysis and innumerable phenomena in the medical sciences have become even more clearly apparent. The subject is maturing. The author's treatment of the subject of Immuno-Catalysis has like- wise matured. It is our sincere belief that the present volume will prove stimulating and helpful to the many biochemically trained investigators and teachers who are reinvigorating the medical sciences. Stuart Mudd, M.D. Vhiladel'jphia AUTHOR'S COMMENTS AND ACKNOWLEDGMENTS THE excellent reception accorded to the first edi- tion of this book was highly gratifying and stimulating. This necessitated the preparation of a second edition, work on which started in early 1946. The contents of the book have been enlarged by including many new subjects of biochemical, enzy- mological and immunological interest, including a chapter on the physiology and biochemistry of ana- phylaxis. The author takes this occasion to express his in- debtedness to many friends, readers and reviewers here and abroad whose comments have been a con- stant source of stimulation. - During the preparation of the manuscript for this edition I have enjoyed the privilege of carrying on research under grants to the University of Pennsyl- vania from the Josiah Macy, Jr., Foundation and from the United States Public Health Service. I wish to express my appreciation to Dr. Otto Rosenthal of the Department of Research Surgery, and to Dr. Seymour S. Kety of the Department of Pharmacology of the Graduate School of Medicine, and to Dr. John R. Preer of the Department of Zoology, for valuable aid received in the preparation of certain sections. Special thanks are due to Dr. John Flick and Professor Stuart Mudd for their con- stant stimulating interest and their painstaking read- ing of the entire manuscript. Their most valuable suggestions are incorporated into the text. M. G. Sevag Contents Page Preface v Author's Comments and Acknowledgments ix PART I ANTIGENS AS BIOCATALYSTS Introduction 3 A. The Formation and Properties of Antibodies 7 1. The Chemical Nature of Antibodies 9 B. The Role of Catalysis in Chemical Reactions and Its Bearing on the Formation of Antibodies 10 1. Catalysis of Organic Reactions 13 a. Mutarotation of Glucose by Acid and Base Catalysis . . 16 b. Acid Catalyzed Enolization 18 c. Enolization in Dilute Acid 19 d. Acid-Base Catalysis of Condensation Reactions ... 19 e. Acid Catalysis of Condensation Reactions 20 f. Hydrolysis of Esters 21 g. Acid Catalyzed Esterification and Hydrolysis .... 22 2. Acid Catalysis in Non-Aqueous Solvents (Aprotic) ... 23 3. Characteristics Which Are Common to Inorganic Catalysts, Enzymes and Antigens 25 a. Disproportionality between the Amount of Inorganic Cat- alysts and the Amounts of Substrates Catalyzed ... 29 b. Disproportionality between the Amounts of Enzymes and the Amounts of Substrates Catalyzed 29 c. Disproportionality between the Amount of the Antigen Used and the Amount of Antibody Produced .... 30 d. Absence of Inorganic Catalysts, Enzymes and Antigens in the Catalyzed Reaction Products 35 e. Enzymes as Antigens 38 4. Do Catalysts (Antigens) Make a New Reaction Possible"? 44 5. Does Antibody Synthesis Involve New Processes Which Did Not Already Exist in the Animal System? 45 3d j666; Xll CONTENTS PART II MECHANISM OF ANTIBODY FORMATION Page The Factors Controlling the Production and Persistence of Anti- body 49 1 . The Relation of the Specificities of Host Enzymes to the Anti- genicity of Substances Foreign to the Species of the Host 49 2. An Inquiry into the Nature of Factors Controlling the Balance of Antigen and Antibody During Immunization .... 61 Theories on the Site of Antibody Formation 67 1. The Lymphocytic Theory 67 a. The Hormonal Control of Anamnestic Response and Lym- phocytes 70 b. Findings Contrary to the Theory of the Hormonal Control of the Release and Fabrication of Antibody 72 2. The Reticulo-Endothelial Theory 72 a. The Theory of Sabin 72 b. Plasma Cells as Antibody Producers 73 Theories on the Mechanism of Antibody Formation .... 77 1. An Introductory Comment on the Antigenic Specificity of Proteins 78 2. The Role of Determinant Groups on the Antigenic Specificity of Conjugated Proteins 84 3. Mechanism of Antibody Formation 86 a. The Theory of Breinl and Haurowitz 86 b. The Theory of Mudd 86 c. The Theory of Alexander 87 d. The Theory of Pauling 87 4. The Concept of Catalysis as Implied by the Above Theories on the Mechanism of Antibody Formation 89 5. Theory of Burnet et al. of Antibody Formation and Adaptive Enzyme Process 90 a. The Premises of the Theory of Burnet et al 90 b. Consideration of the Adaptive Enzyme Concept with Re- spect to Antibody Production 95 c. Adaptive Enzymes— A Consideration of Facts and Assump- tions 96 d. Consideration of the Primary Role of Antigen in Antibody Production as Postulated by Burnet, et al 100 e. Consideration of Antigens as Toxic Agents Causing the Mutation of Globulin Synthesizing Enzyme System . 102 6. On Jordan's Autocatalytic Theory of Antibody Formation 104 CONTENTS XIU Page The Concepts of "Proteinogen" and Enzyme-Precursors Con- sidered in the Light of the Specificities of Antigens and Anti- bodies and Their Digestion Products 106 a. Consideration of Enzyme-Precursors 107 b. Derivatives of Chymotrypsin and Their Properties . . . 110 c. Virus-Host Relationship 113 d. Structural Specificity of Polypeptides of Low Molecular Weight 114 e. Does Antigen Function as Coenzyme in the Production of Antibody? 116 f. Specificity of Cleavage Products Derived from Antibody . 118 Anti-Antibodies 122 a. Presence of Serologically Reactive Basic Amino Groups in Antibody Globulins 123 b. The Directive Influence of Optically Active Catalysts in Producing Optically Active Substances 125 c. Experiments Dealing Directly with the Question of Anti- Antibody Formation 131 d. Non-Identity of the Combining Sites for Antigen and Anti- body in an Antitoxin Molecule 134 e. Further Experimental Data Concerning the Presence of a Common Group in Different Antibodies from a Species of Animal 135 f. A Comment on the Use of Antigen- Antibody Complex for the Production of Anti- Antibodies 137 g. Acquisition of New Antigenic Specificity by Antitoxins , 138 h. Crystalline Diphtheria Antitoxin Antigenically Distinct from Normal Serum Components 139 PART III ANTIBODY AS A SPECIFIC ENZYME INHIBITOR A. Nature of the Analogy between Immune and Enzyme Reactions 144 1. Serological Specificity of Stereoisomeric Conjugated-Protein Antigens 147 2. Specificity of Enzymes Which Catalyze Stereoisomeric Sub- strates 151 3. Comparison of the Stereoisomeric Specificities of Immune and Enzyme Reactions 153 4. Immune and Enzyme Reactions— A Comparison of Reaction Mechanisms 156 5. Zinsser's View on the Formation and the Role of Antitoxins 159 XIV CONTENTS Page B. The Formation of Specific Inhibitors in Enzyme Reactions 160 1. Pepsin Inhibitor 161 2. Tr)'psin Inhibitor of Pancreatic Extracts 162 3. Tr)'psin Inhibitor of Blood Serum 163 4. Inhibition of Carbohydrases by the Reaction Products . 163 5-13. Similarly Formed Inhibitors 165-173 PART IV ANTI-ENZYME IMMUNITY A. Analysis of Certain "Controversial" Aspects of Anti-Enzyme Im- munity 175 1. Analysis of Bayliss' Objections Against the Existence of Anti- Enz^Tnes 176 a. The Nature of the Trypsin Inhibitor Present in Sera in Relation to Anti-Enzyme Antibody 178 b. The Nature of the Tr)'psin Inhibitor in Egg White 179 c. The Resistance of "Living" Protein to the Action of Pro- teolytic Enz)Tnes 180 d. The Inhibition of Enzymes by Enzymes and Viruses 183 2. The Question of Non-Specific Adsorption of Toxins or En- zjones on Coexisting Protein-Anti-Protein Precipitates 185 a. Adsorption of Toxins on Red Blood Cells and Charcoal without Loss of Activity 188 b. Failure of Protein-Anti-Protein Precipitates to Adsorb Toxins 188 c. Failure of Protein-Anti-Protein Precipitates to Adsorb Non- Specific Colored Proteins 189 d. Failure of Protein- Anti-Protein Precipitates to Adsorb Non- Specific Proteins 190 e. Failure of Agglutinated Bacteria to Adsorb Non-Specific Proteins 191 3. Effect of pH of Optimal Activity on the Nature and Extent of the Antigen-Antibody or Enzyme-Anti-Enzyme Combinations 192 B. A Critical Consideration of the Antigen- Antibody Reactions in Relation to the Inhibition of Enzymes by Specific Antibodies . 195 1 . An Analysis of the Reactions of the Active Groups of Proteins in Relation to Their Biological Specificities 195 a. Studies on the Reactive Groups of Proteins 196 b. Reaction of Formaldehyde with Proteins and Amino Acids 196 c. Acylation of Proteins 198 d. lodination of Proteins 200 e. Reactions wdth Sulfhydryl and Disulfide Groups of Proteins 201 CONTENTS XV Page 2. Identity of the Nature of the Inhibition of Enzymes by Homol- ogous Antibodies with Antigen-Antibody Reactions . 204 C. Antibody Against Carbohydrases 213 1. Antibody Against Emulsin 214 a. Inhibition of the Amylase Activity of Emulsin Preparation by Anti-Emulsin Serum 215 2. Specificities of Amylases 218 a. Antibody Against Malt Amylase 220 3. Invertase and Its Properties 223 a. Antibody Against Invertase 224 4. Enz)Tnatic Synthesis of Serologically and Physiologically Ac- tive Polysaccharides 225 5. Antibody Against Bacterial Polysaccharidases 230 6. Antibody Against Crystalline Lysozyme 230 D. Hyaluronidase or the Permeability Factor 231 1. Discovery of the Permeability Factor 231 2. Enz)Tnic Nature of the Permeability Factor and the Nature of the Substrate 233 3. Mechanism of the Action of Hyaluronidase 237 a. Quantitative Methods of Measuring Hyaluronidase Acti\aty 239 4. Distribution and Physiological Significance of Hyaluronidase or the Permeability Factor 241 a. Relation of Hyaluronidase Production to Virulence of Bac- teria 241 b. Relation of Capsular Polysaccharide and Hyaluronidase to the Virulence of Streptococcus and Pneumococcus . 244 c. Hyaluronidase in Infected Tissue Extracts 245 d. Effect of Hyaluronidase on Fertilization 246 5. Non-Specific and Specific Inhibitions of Hyaluronidase 246 a. Non-Specific Inhibition of Hyaluronidase 246 b. Neutralization of Hyaluronidase or the Permeability Factor by Specific Immune Serum 247 E. Antibody Against Proteolytic Enzymes 249 1. Antibody Against a "Tr\^sin" Preparation 249 a. Antitryptic Property of Guinea Pig Serum Immunized vdth "Trypsin" 251 b. The Toxic and Lethal Action of the Trypsin Preparation 252 c. Neutralization of the Toxic Effects of Trypsin with Im- mune Serum 253 d. Antibody Against Papain 253 e. Antibody Against the Proteolytic Activity' of Snake Venom 255 f. Proteolytic Acti\aty of the Toxin of CI. Welchii . 256 2. Antibody Against Bacterial Proteinases 257 XVI CONTENTS Page a. Immunity Against Proteinases of Various Bacteria 258 b. Antibody Against Streptococcal Proteinase 259 c. Differences of the Proteinases of CI. Sporogenes and CI. Histolyticum and the Specificity of the Anti-Proteinases . 260 d. The Abihty and Inabihty of Normal Serum "Trypsin-Inhib- itor" and Raw Egg White to Inhibit the Proteinases of Pathogenic and Non-Pathogenic Bacteria 262 F. Fibrinolysis 263 1 . Serum Proteinase as Fibrinolytic Enzyme and the Role of Bac- terial Fibrinolytic Factor 263 a. Comment on Nomenclature 263 b. Digestion Products of Fibrin and Fibrinogen .... 264 c. Distribution of Bacterial Fibrinolytic Factor 266 d. Fibrinolytic Factor and Bacterial Invasiveness .... 269 e. Antibody to Fibrinolytic Factor 269 2. Postulates on the Role of the Bacterial Fibrinolytic Factor . 271 a. Comments on the Possible Nature of the Role of Bacterial Fibrinolytic Factor 275 b. A Suggestion as to the Possible Lipolytic Role of Strepto- coccal Factor 278 G. Mechanism of Fibrin and Milk Clot Formation 278 1 . Enzymes Responsible for the Formation of Plasma Clot . . 278 a. Comment on the Use of the Term Coagulation .... 278 b. Factors Responsible for the Formation of Plasma Clot . 279 c. Enzyme Nature of Thrombin 280 d. Conversion of Fibrinogen into Fibrin 280 2. Chemical Reactions Involving -S-S- Linkages in the Conver- sion of Fibrinogen to Fibrin 282 3. Antibody Against Thrombin. Neutralization of Thrombin of Rabbit Serum by Anti-Rabbit Guinea Pig Immune Serum 285 4. Antibody Against the Plasma Clotting Enzymes of Snake Venom 289 a. Enzymic Nature of Blood Clotting Activity of Snake Venoms 289 b. The Clotting of Fibrinogen by Snake Venoms .... 289 c. The Activation of Prothrombin to Thrombin by Snake Venoms 290 d. Anti-Thrombin, Anti-Proteolytic, and Anri-Necrotic Acrivi- ties of Antivenin 291 5. Staphylococcal and Other Bacterial Factors in Fibrin Clot Formation 295 a. Distribution of Plasma Clotting Factor among Bacteria 295 b. Anti-Clotting Factor 295 CONTENTS XVll Page c. The Role of Bacterial Clotting Factor in Phagocytosis and Infection 296 d. Comment on the Nature of the Staphylococcal Clotting Factor 298 e. "Activation" of the Staphylococcal Clotting Factor . 299 f. Properties of Clotting Factor and Activator Substance and Their Possible Role in Fibrin Clot Formation .... 300 6. Antibody Against Rennins 303 a. Certain Properties of Rennin 303 b. Mechanism of Milk Clotting 304 c. Milk and Plasma Clots Compared 304 d. Antibody Against Animal Rennin 304 e. Antibody Against Plant Rennin 306 H. Enzymatic and Pharmacological Activities of Hemolytic Sub- stances 307 1. Hemolytic Lysolecithin Derived by the Action of Snake Venom Lecithinase 307 a. Discovery of Hemolytic Lecithinase and Its Neutralization by Antivenin 308 b. Physiological Consequences of the Action of Lysolecithin 312 2. Ricin 314 a. Chemical Nature of Ricin 314 b. Inhibition of Lecithinase and Hemolytic Activities of Ricin by Anti-Ricin Immune Serum 315 3. Bacterial Hemolysins 316 a. Hemolytic, Dermonecrotic and Lethal Activities of Staphy- lococcal Toxin 317 b. The Relation of the Lipase Activity of Staphylococci to Hemolysis 318 c. Hemolytic Toxin of Clostridium Septicum 319 4. Pneumococcal Hemolysin 319 a. Number of Pneumococcal Hemolysin Molecules Required to Hemolyze One Red Blood Cell 323 b. A Study of the Antipneumolysin Titer of the Sera of Pneu- monia Patients 324 5. Streptococcal Hemolysin 324 a. Streptolysin O 326 b. Probable Enzymic Nature of Bacterial Hemolysins 328 I. Antibody Against Nucleases, Urease and Penicillinase .' 329 1. Antibody Against Ribonucleases 329 a. Liberation of Adenyl Compounds from Perfused Organs Treated with Cobra Venom 329 XVUl CONTENTS Page b. Antibody Against Nucleic Acid-Hydrolyzing Enzymes of Snake Venom 330 c. Inhibition of Ribonuclease by Homologous Anti-Serum 330 2. Neutralization of the Toxic Action of Crystalline Urease by Its Homologous Antibody 331 3. Antibody Against Penicillinase 332 J. Immunity Against the Enzymatic Activities of Bacterial Toxins 335 1. Bacterial Toxins 335 a. Type A Toxin oi Clostridium hotulinum 335 b. Biological Action of Botulinus Toxin 336 c. Crystalline Tetanal Toxin 336 d. Enzymatic Activities of the Toxins of Vibrio Comma . 337 2. Types of the Toxins of Clostridium- Oedematiens .... 338 3. Lecithinase of Clostridium. Welchii 339 a. Classification and Properties of the Toxins of Clostridium Welchii 339 b. Preparation of Toxins 342 c. The Lecithinase Activity of the Toxin 342 d. Characteristics of Lecithinase 345 e. Quantitative Methods Used for the Measurement of Leci- thinase Activity 345 f. Phenomenon of Opalescence in Serum and Lecitho-Vitellin Produced by the Action of the Toxin of CI. Welchii . 347 g. Production of Opalescence in Serum and Lecitho-Vitellin as a Measure of the Potency of the Toxin 349 h. Lecithinase Activity as a Measure of Biological Effects of Toxin 350 4. Immunity to the Pharmacological Actions of Toxins 352 a. The Inhibition by CI. Welchii Toxin of the Oxidation of Succinate by Tissue 352 b. Neutralization of the Inhibition of Tissue Respiration by Toxin as a Measure of the Antitoxin Concentration of Im- mune Serum 354 c. Pharmacological Activity of the Toxins of CI. Welchii 355 5. Action of Antitoxin on Enzymes Causing Histological Changes in Gas-Gangrene 357 PART V ANTIBODIES AGAINST RESPIRATORY ENZYMES A. Respiratory Enzymes 363 1 . Dehydrogenases, Flavoproteins, Cocarboxylase Containing En- zymes, and Heme Containing Enzymes 363 CONTENTS XIX Page 2. Enzymes Involved in the Fermentation and Oxidation of Glu- cose (Phosphorylation of Hexoses and Oxidation-Reduction Reactions of Hexose-Breakdown Products) 365 3. Tricarboxylic Acid Cycle 372 4. Synthesis of Amino Acids from a-Keto Acids 372 a. Oxidative Deamination of Amino Acids 372 b. Synthesis of Amino Acids by Transamination .... 375 B. Problems Dealing with the Production of Antibody Against the Respiratory Enzymes 377 1. Reasons for the Failure to Produce Antibody Against Yellow Enzyme 380 2. Inhibition of Pyruvic Acid Reductase by Specific Immune Serum 381 3. Inhibition of Yeast Hexokinase by Homologous Anti-Serum 382 4. Inhibition of Yeast d-Glyceraldehyde-3-phosphate Dehydro- genase by Specific Anti-Serum 385 5. Reports on the Inhibition of Bacterial Respiration by Specific Immune Serum 388 6. Phosphatase and Antiphosphatase 390 C. Metallo-Proteins as Oxidation Catalysts and Antigens . . . 393 D. Antibody Against Luciferase-Oxidative Enzyme of Luminescence 396 PART VI PHYSIOLOGY AND BIOCHEMISTRY OF SHOCK 1. A Description of Anaphylactic. Shock 399 a. Symptoms in Guinea Pigs 401 b. Symptoms in the Dog 401 c. Symptoms in the Rabbit 401 2. Peptone Shock 402 3. Shocks Caused by Proteolytic Enzymes 403 a. Trypsin 403 b. Papain and Ficin 404 c. Chymotrypsin 404 d. The Question of the Proteolytic Liberation of Histamine as Cause of Trypsin Shock 405 4. A Summary of Some of the Physiological Changes Accompany- ing Shock 407 5. Histamine Shock 407 a. Effects of Histamine 407 b. Distribution of Histamine 408 c. Origin of Histamine 409 XX CONTENTS Page d. The Function of Antihistamine Substances and Their Phar- macological Action 411 e. Histamine and Antihistaminic Substances 412 f. A Comparison of the Syndromes of Histamine and Ana- phylactic Shock 414 6. Metabolic Changes in Shock 414 7. Acetylcholine in Anaphylactic Shock 415 a. Destruction and Synthesis of Acetylcholine 416 b. "Muscarinic" and "Nicotinic" Actions of Acetylcholine 416 c. Antagonists to Acetylcholine 417 d. The Role of Acetylcholine on the Effector Organs . .417 e. Effects of Administered Acetylcholine in Man . .418 f. In Vitro Action of Acetylcholine on Muscle Contraction 419 g. Role of Acetylcholine in Anaphylactic Shock . .421 h. Liberation of Acetylcholine in Anaphylactic Shock . 422 i. Experiments Cited to Support the Acetylcholine Theory of Anaphylaxis 424 j. Summary of Evidence in Favor of Acetylcholine Theory . 426 8. Metabolic Changes in Anaphylactic Shock 427 a. Liberation of Potassium 427 b. Potassium-Calcium Antagonism 428 c. Acidosis in Anaphylactic, Histamine, and Peptone Shock 429 d. Metabolic Factors in Anoxia 430 9. Theories on the Liberation and the Role of Proteinases in Anaphylactic Shock 432 10. The Question of Proteolysis and Release of Histamine in Shock Produced by Proteolytic Enzymes and in Anaphylaxis 435 References 445 Author Index 509 Subject Index 521 Jmmuno-Catalysis Introduction IT IS A, universal phenomenon of nature that every hving form must struggle to perpetuate its existence. In this pursuit its primary energy is spent in the search for food. Its chemical activities determine the extent the food can be made use of. From the standpoint of a free living bacterium, it is immaterial whether it derives its food from dead matter, or within a suitable host; the host is simply another medium for its propagation. The disease, death, and immunity, in case of recovery, caused by its multiplication in a host, are incidental events. Only extremely parasitic living forms are dependent on a host. The focus of our interest must of necessity be the nature and the intensity of the chemical activities of micro-organisms, for these constitute their biologically indispensable characteristics and determine the nature and the fatality of the diseases they cause. Micro-organisms of varying degrees of virulence, such as Group A /^-hemolytic streftococci, Stre^ptococcus viridans, fneumococcus, Sal- monella aerlrycke, C. difhtheriae, Shigella dysenteriae, CI. tetani, CI. hotulinum, CI. welchii, Stafhylococcus aureus, B. fwteus, members of the colon group, etc., produce their exotoxins and endotoxins in vitro under special or routine laboratory growth conditions. It is immaterial how these toxins are produced; whether they are products of bacterial secretion, autolysis, or a result of the activities of certain specific bac- terial enzymes on the medium, or are extracted from the intact cell by chemical manipulation, has no bearing on the basic aspect of the question. As these poisons can be elaborated by the bacteria outside of, or in the body of hosts, they are the products of their specific physio- logical activities in the course of their normal growth and death. Many of the pathogenic bacteria produce acute exotoxic diseases. However, in chronic infections, the disease agents either do not elabo- rate such fatal poisons, or they are rendered less harmful by the host. Whatever may be the answer, at the end a serious condition may develop to a greater or lesser degree as the result of the physiological 3 4 IMMUNO-CATALYSIS activities of the disease agents. In this connection it may be noted that infections by Sfirochaeta 'pallida, M. tuberculosis, the leprosy bacillus, Actinomycetes and malaria parasites etc., are diseases which habitually run a very slow course and do not show evidence of any fatal toxin which can overwhelm the host before the defenses can be mobilized. Cutaneous and mucous eruptions, inflammatory discharges, the swell- ing of the regional nodes, the malaise, the fever and sensation of chilli- ness, numerous disturbances in the normal physiology of the infected host may likewise be looked upon as the consequences of the inter- action between specific bacterial components and the environment offered by the hosts. These considerations bring us to the examination of the possible physiological role of the various constituents of bacteria. Hence the morphological and cultural characteristics, and biochemical activities of bacteria must be studied from the point of view of their relation to infectious diseases, as well as from the view of the biology of the bacterium per se. Examining the facts from the standpoint of the biology of bacteria, we see that the enzyme activities of bacteria are indispensable for their growth and reproduction in vitro, as well as in vivo. Invasiveness and fatalness of a bacterial infection are conditioned generally not only by the degree of the resistance offered by the host but, more specifically, by the degree and nature of the enzyme activities of invasive bacteria. The antigenic property of a bacterium, on the other hand, is a conse- quential property; it is conditioned by its propagation in a host for a requisite length of time. This property of bacterial substances would never have come to light if the bacterium had followed an entirely free-living, non-parasitic existence. These facts indicate no doubt that with the exception of highly parasitic forms micro-organisms are not dependent on a host, and, more specifically, not on the antigenic properties of their constituents for existence and propagation. On the contrary, the antigenic property is not of utility for the bacterium, but is detrimental to its prolonged propagation within a host. For example, certain infections are followed by effective immu- nity which is instrumental in wiping out the bacteria from the body of a host. On the other, in chronic infections the ineffectiveness of the antigenic property of certain disease agents assures them of a pro- longed existence in a host. INTRODUCTION Despite these facts, the property which makes certain bacterial con- stituents antigenic must play a definite role in the biology of a bac- terium. Of these constituents, protein is the most essential one. It is not an exaggeration if we make the statement that life could not have appeared without protein substances. The proteins of the most primi- tive forms of life which have appeared as single cells in the evolution- ary scale of living forms must have possessed attributes responsible for the antigenic property long before the appearance of animal hosts which alone have served as the means of exposing it. Certain bacterial protein substances, as well as those of animals and plants, would appear to be endowed with catalytic powers which are essential for the physiological functions of the living forms from which they are derived. Likewise, catalytic power, as discussed in this treatise, appears to stimulate, or is responsible for the formation of specific anti- bodies in animals to proteins and also to certain non-protein compo- nents. From the standpoint of the biology of bacteria these facts would seem to represent a contradiction; nevertheless it is a fact that chemical forces of bacteria (or their constituents) which accelerate their in vitro activities, as well as their invasive, infectious and growth activities in a host, are also instrumental in setting up in the host highly specific antagonistic immune mechanisms which block or neutralize these very activities of bacteria. The highly specific nature of these relationships, characteristic of enzyme reactions, are strong indications that the mechanism of the above mentioned bacterial activities are of catalytic nature. The action of enzymes usually results in the formation of reaction products which specifically inhibit the action of these enzymes. Simi- larly, as discussed in this treatise, antigens as enzymes, or enzymes as antigens, produce antibodies, as finished reaction products, the only function of which, as far as our present knowledge goes, is to neutralize the specific biological activities of antigens. Practically all proteins foreign to the species possess a special sort of biocatalytic power, namely that of stimulating and directing specif- ically the synthesis of antibody. This familiar property of antigenicity must be considered a special kind of biocatalytic power, and it is pos- sessed by enzyme proteins, in common with other proteins. Since we are at present in the dark concerning the specific functions of the latter proteins in the living cells, it is premature to consider them devoid of 6 IMMUNO-CATALYSIS special properties characteristic of the known enzymes which catalyze anabolic and catabolic reactions. The answer to the question of whether or not antigenicity and enzyme-activity are two distinguishable or identical biocatalytic properties of proteins can come from the results of studies designed with this question in mind. The results of certain studies to be discussed in the text would seem to show that under appropriate conditions specific reactivity as enzyme and as antigen may, and frequently do in fact, reside in the same specific chemical configurations of the protein molecule. This view has brought us to consideration of the well-known analogy that exists between enzyme and immune reactions. Ehrlich, finding that toxin-antitoxin, ricin- antiricin reactions are highly specific, was the first to call attention to the similarity between immune reactions and those of enzymes ob- served by Emil Fischer. Many investigators, more particularly Land- steiner, have referred to this analogy on numerous occasions. The parallelism between the specificities of these two reactions* has been extended to d- and 1-specificity by Landsteiner and Van der Scheer, and to a- and iS-configurational specificity by Avery and Goebel and their associates. However, the question as to why the specificities of these two classes of reaction approach and nearly coincide with each other has remained until the present still very much of a riddle. The present author has critically analyzed the known facts concerning this question, and the belief has grown in him that his interpretation of the basic facts regarding the above mentioned analogy and the related questions present a useful contribution to the sciences of immunology and biochemistry. This belief has prompted him to prepare the present treatise, which he humbly offers to the reader under a new title, "Immuno-Catalysis." Part I Antigens as Biocatalysts IN THE introductory part of this study the concept formulated as a working hypothesis attributed to antigens (bacteria, toxins and proteins) properties similar to those of biocatalysts. Among these properties the possible catalytic role of antigens in the formation of antibodies will be discussed more fully. A clear understanding of this aspect of antigens may yield additional information about immune processes, as well as the possible role the bacterial antigens might play in the biology of bacteria. The most outstanding property of antigens is, by definition, the stimulation of the formation of antibodies. Although the reactions between antigens and their specific antibodies have been extensively studied, the role of antigens in the formation of antibodies has not yet been clearly defined. Certain factors and conditions contributing to the production of antibodies are likewise well knowm, but these scattered facts have not been evaluated and integrated. The present treatise is an attempt in this direction. To determine the conditions which control the formation of a new chemical entity, we must first establish what class of known sub- stances this new entity belongs to. This is achieved by a detailed study of its chemical and physical properties in reference to known substances. With respect to antibody, information concerning the site of its formation, as well as its relationship to a knov^oi class of com- pounds wdll assist us further in determining the type of reactions, or chemical influences which might be responsible for its production. A review of the literature on antigens and antibodies from this view- point may therefore yield valuable information. A. THE FORMATION AND PROPERTIES OF ANTIBODIES With the exception of iso-antibodies (natural hemagglutinins) the tissues and the blood of a normal animal theoretically are believed to 7 8 IMMUNO-CATALYSIS be free of specific antibodies capable of reacting with bacteria, and bacterial, plant and other animal proteins. The injection of these ma- terials into an animal creates changes resulting in the formation of antibodies. With the appearance of antibodies the serum of such an animal acquires the property of entering into specific chemical com- bination (agglutination, precipitation) with the bacteria or the protein used for immunization. Each species of protein thus stimulates the formation of an antibody reactive sfecifically with itself. Several investigators have stated that in parallel with the appear- ance of antibodies there is an increase in the serum globulin, not all of which could be accounted for by the amount of antibody present. The findings of a few investigators will be cited here. It is found (Liu, Chow and Lee, 1937) that at most two-thirds, usually under one-half, of the increase of globulin in rabbit's serum, or immunization with pneumococci is accounted for by antibody. In contrast, the examina- tion (Bj0rneboe, 1939) of 56 anti-pneumococcal rabbit sera of different types for total and specific nitrogen led to the conclusion that the increase in serum protein during immunization was due to antibody- protein. The production of antibody-protein was therefore stated to be an extra production of proteins. Immunization of rabbits with different combinations of proteins— crystalline ovalbumin, Limulus hemocyanin, Vivifarus (snail) hemocyanin, crystalline edestin and Bence-Jones protein— was followed with respect to the increase in serum globulin (Boyd and Bernard, 1937). The sera were fractionated with 13.5 and 17.4 per cent sodium sulfate. The results showed that the greatest increase takes place in the 13.5 per cent sodium sulfate fraction, rising in one rabbit by 1275 per cent. The increase in the 17.4 per cent fraction, though still relatively enormous, was in general not so large as that of the 13.5 per cent fraction. The measurements of antibody showed that the great increase of globulin was not all, or even chiefly attributable to the specific antibody produced. As a rule the antibody increase was much less than that of the globulin, although the two were generally parallel. There was no antibody in the albumin fraction. These and numerous other investigations have thus shown that in response to the stimulus of a foreign protein, the animal system experiences an increase in antibody protein and at times in "non- specific" proteins (see also van der Scheer, et ah, 1942). ANTIGENS AS BIOCATALYSTS 1. The Chemical Nature of Antibodies Chemical tests have shown (Welsh and Chapman, 1908; Dean and Webb, 1926) that the precipitates resulting from antigen-antibody reactions contain considerably more protein than the antigen could account for. A series of quantitative analytical studies (Hartley, 1926; Felton and Bailey, 1926; Heidelberger and Kendall, 1929, 1933, 1935, 1936) have demonstrated that the precipitate obtained with protein- free pneumococcal carbohydrates and immune sera consists largely of serum globulin; 2.5 mg. of pneumococcal polysaccharide Type II precipitated 37 mg. of protein. The serum proteins isolated from these precipitates were found to be 98 per cent antibody. The properties of diphtheria toxin-antitoxin floccules and the pre- cipitate in other types of antigen-antibody precipitation reactions were similar to those of serum globulin (Marrack and Smith, 1930). The particulate antigens, such as bacteria and red blood corpuscles, when fully combined wdth antibody, moved in an electric field as though they were pure globulin (Shibley, 1926; Eagle, 1930; Mc- Cutcheon, Mudd, Strumia, and Lucke, 1930). The effect of proteolytic enzymes, such as pepsin, trypsin and papain, on antibodies has constituted the subject of numerous in- vestigations. These studies have shown that antibodies are destroyed rapidly by pepsin, less rapidly by trypsin. Types I and II pneumo- coccal antisera, and purified Types I, II, and III pneumococcal anti- bodies were slowly destroyed by trypsin (Felton and Kauffmann, 1927). The precipitate formed by Type I pneumococcal polysaccharide and antiserum was incubated with pepsin (pH 4.8). The destruction of precipitin ran parallel to the increase of amino nitrogen (Chow, Lee and Wu, 1937). Rosenheim (1935) found that the O-agglutinins in all serum samples from three horses immunized with B. typhosus were rapidly destroyed by pepsin (pH 4.6-4.8), trypsin (pH 8.6) and activated papain. The H agglutinins in serum samples obtained from each of three horses after the first immunizing course were rapidly de- stroyed by proteolytic enzymes. Those in samples obtained after several immunizing courses were not appreciably destroyed under identical conditions. The H agglutinins which were apparently resist- ant to pepsin and trypsin were not resistant to activated papain. The 10 IMMUNO-CATALYSIS globulin fractions of the serum obtained after the first and subse- quent immunizing courses were hydrolyzed to approximately the same extent by proteolytic enzymes. A certain degree of resistance to pep- sin by diphtheria antitoxin serum is also reported (Pappenheimer and Robinson, 1937). However, since the peptic digestions in these cases were carried out at a pH unfavorable for obtaining maximum pro- teolytic activity, it should not be interpreted that these immune bodies are not of protein nature. It can only mean a relative, rather than an absolute resistance to enzymes in comparison with normal serum globulins under these conditions. Conclusions. Chemical analysis of purified antibody preparations and of normal globulins have failed to indicate any differences in the percentage of amino acids and total nitrogen. Antibodies produced in rabbits to various antigens have so far been found to be indistinguishable by ultracentrifugal and electrophoretic studies from normal rabbit serum globulin. The antibodies produced in the rabbit seem to have the same isoelectric point as the corresponding y-globulin derived from the normal rabbit. How is it then that antibodies, chemically and physico-chemically identical with globulins, are serologically distinct entities? This dif- ference has given rise to various hypotheses, which will be discussed later. It suffices to state at this point that this diflFerence is believed to be due to certain variations in the configuration of normal globulin during its synthesis resulting in the formation of antibodies. B. THE ROLE OF CATALYSIS IN CHEMICAL REACTIONS AND ITS BEARING ON THE FORMATION OF ANTIBODIES In analyzing the nature of various chemical influences which con- tribute to antibody formation in response to antigenic stimuli we cannot at present formulate the mechanism of the chemical reactions involved, but we can at least determine whether an antigen acts as a reactant forming a part of the final reaction product, or as a catalyst directing the synthesis of antibody globulin. Changes in organic and inorganic substances, and changes or varia- tions in the physiological processes of living forms are brought about by chemical influences. No matter how the mechanism of these changes or reactions are formulated they take place according to one ANTIGENS AS BIOCATALYSTS 11 of the known type reactions. All chemical reactions take place ac- cording to a stoichiometrical relationship; that is, there exist numer- ical relations between elements or compounds in combining to form a new compound, or in tautomerizing to a new isomer, or in dissociat- ing into elements or compounds. They do so according to definite proportions by volume and weight. Various influences, such as heat, light, or catalysts do not affect this relationship. The influence of a catalyst does not change the stoichiometrical relationships. The catalyst does not alter the free energy of the reaction; it does not enter into any irreversibly stable union with either the reactants or the reaction products. It simply speeds up the reaction to attain the equilibrium in a shorter period of time. In many cases the reactions, in the absence of a catalyst, proceed at such a slow rate that one cannot detect their presence, but vdth the introduction of a suitable catalyst the reaction becomes obvious. The structures of organic substances express mechanical and chemi- cal meaning. All atoms or groups of atoms at special positions in the space occupied by the structure of a molecule determine the chemical and physical behavior of the whole molecule, and its groups as well. Any change in the spatial relationships of the functional groups as a whole causes a corresponding change in properties. A given empirical formula of an organic compound will often represent several isomeric compounds of the same molecular weight and elementary analysis, such as, for example, the cis and trans isomers, fumaric and maleic nature. HOOC— C— H H— C— COOH II II H—C— COOH H— C— COOH Fumaric acid Maleic acid which have different physical and chemical properties. There are large numbers of isomers which may differ from each other only by a single physical property, namely, by the ability to rotate the plane of polarized light in a different direction, d- and 1-fructose, d- and 1-lactic acid, d- and 1-alanine, etc., represent the class of a large number of optically active isomers. What interests us here is that some of these isomers are interconvertible under the influence of accelerating agents of catalytic nature. In general, chemical reactions take place either (a) by the applica- 1 2 IMMUNO-CATALYSIS tion of heat energy; or, (b) through chemical affinities; or, (c) the apphcation of force to non-spontaneous processes. This force may be electromotive, as in electrolytic processes; or, (d) by the accelerating effect of organic or inorganic catalysts. Two important facts are to be observed. In the first place, we note that the final products either contain the parts of the reactants, or are derived from a single substance, or a single complex molecule is formed out of two or more reactants. In contrast, the enzyme, or organic and inorganic catalysts do not enter into any irreversible union with any of the reaction products. The Synthesis of Sul'phuric Acid hy the Catalytic Contact Process. In this process the interaction of sulphur dioxide and oxygen is the principal reaction. The interaction between the two main reactants can be hastened by the catalytic surfaces of porcelain, ferric oxide, and, more especially, by finely divided platinum spread in a fine grey powder on the fibers of asbestos to obtain maximum active surface. Under controlled temperature conditions the above gases are blown over the catalytic system forming sulphur trioxide. The platinum re- mains unchanged and functions continuously so long as arsenious oxide and other impurities do not poison the active surface of the catalyst. The surface of the finely divided platinum is characterized by strong adsorbing powers. It holds the gaseous molecules with relative firmness, and it is believed that there the molecules undergo activation prior to a chemical union. Once this has taken place the active surface is set free to function again. The Synthesis of Sulphuric Acid in a Homogeneous Catalytic System— Chamber Process.— Water vapor, sulphur dioxide, nitrous anhydride (N2O3), and oxygen participate in the synthesis of sul- phuric acid in this process. When these gases are thoroughly mixed in large leaden chambers it is assumed probably that the greater part of the acid is formed by the following two successive actions (Rideal and Taylor, 1926): O— NO 1 . H.,0 -f 2SO^ -f N0O3 + O, -> 2SO^ Nitrosylsulphuric acid OH ANTIGENS AS BIOCATALYSTS 13 O— NO OH / / 2. 2SO2 +Ho0^2S0. +N0O3 \ " \ " OH OH The gaseous mixture is brown; with the formation of nitrosylsulphuric acid the brown color disappears. With its decomposition into sulphuric acid and N2O3 the color reappears. These two equations represent distinct consecutive actions and not partial equations of an interaction, for one can observe the deposition of nitrosylsulphuric acid crystals when a glass flask is used and the supply of water is deficient. The N0O3 liberated in the second equation is immediately available to take part in the successive cycles a large number of times. In this process NoOa functions as a catalyst and the nitrosylsulphuric acid is the inter- mediate labile complex between a catalyst and the reacting substances. It corresponds to the enzyme substrate complexes generally believed to form in enzyme reactions. 1. Catalysis of Organic Reactions Some of the outstanding general catalyses of organic reactions in homogeneous systems are: (a) general acid and base catalysis; (b) catalysis in aqueous concentrated acid systems; and, (c) acid catalysis in non-aqueous systems. Before dealing with these catalytic processes, it will be helpful to review the modern concepts of acids and bases as well as the role of the solvent in processes involving acids and bases. To extend the considerations of acids and bases beyond treatment of aqueous systems, Bronsted, Lowry, and others have defined acids as substances which yield protons (H+) and bases as substances which accept protons. The relative ease with which the protons are given up (or accepted) is the measure of the acid (or base) strength. The solvent itself may act as an acid or base, and, indeed, the acidity or basicity of the solvent is an important factor in the ionization processes of solutes, as shown by the following generalized reaction scheme: HA + S ^ H+.S 4- A- acid solvent solvated base (base) proton (acid) 14 IMMUNO-CATALYSIS In the solvolytic processes, the basicity of the solvent is closely related to the acidity displayed by the acid. A system of conjugate acid-base pairs is put in equilibrium, A~ being the conjugate base of the acid HA, H+.S the conjugate acid of the solvent S. A striking example of the efficacy of this theory is found in the consideration of aqueous hydrochloric acid solutions. The substance HCl is not itself an electrolyte. In the pure state its bond is strongly covalent and it is a poor conductor of electricity. The theory states that its acidity and electrolytic behavior in w^ater solution is due to a frotolytic reaction with the solvent: HCl -{- H2O ^ H3O+ + Cl- the hydro- nium ion CH+.H2O) The hydrochloric acid, even up to rather high concentrations, ap- pears to be completely ionized because of this process. This is simply due to the fact that the chloride ion is a weak base compared to the H2O molecule, so that the equilibrium is shifted rather completely to the right. In the case of a typical weak acid, such as acetic acid, the ionization is far less complete CH3COOH + HoO ^ H3O+ + CH3COO- for the reason that the conjugate base, the acetate ion, is a rather strong base. Nevertheless, in a more basic solvent, such as liquid ammonia, this same acid is effectively completely ionized, acting as a strong acid : CH3COOH + NH3 ^ NH4+ + CH3COO- It is obvious that the behavior of a substance as an acid or base depends greatly upon the solvent medium. Many substances considered to be strong acids because of their behavior in water may well be weak acids in another solvent or may even be bases. For example, hydro- chloric acid is only slightly ionized in benzene while nitric acid and water act as strong bases with sulphuric acid as solvent, HNO3 4- H2SO4 ^ H2NO3+ + HSO4- H2O + H2SO4 ^ H3O+ + HSO4- This broader view of the acid-base relationships has an important advantage in understanding the catalysis of many reactions. Many ANTIGENS AS BIOCATALYSTS 15 substances other than hydrogen or hydroxy! ions are expected to act as acids or bases from this viewpoint and this reahzation has clarified the picture of a great number of complex reactions. Often the even broader concept of acids and bases proposed by G. N. Lewis is called upon. According to this theory, substances which accept a share in an electron pair are acids, and the donor substances are bases. This picture not only includes as a special case the substances H:Ci: +H:d:H^[^'9'"l'^+[:Cl:l treated by the Bronsted-Lowry concept, but also extends to neutraliza- tions involving no proton transfer at all : CI H CI H C1:'b + :N:H >► C1:B:N:H ci H ci H acid base These viewpoints of the nature of acids, bases, and solvents will be incorporated into the discussion of examples of processes where either specific acids or bases act as catalysts or where any acid or base would be effective to an extent dependent upon its strength. In the latter case, i.e., general acid-base catalysis, the symbols HA and A~ will serve for acids and bases in general. Furthermore, the course of some of the reactions will employ the concept of quantum-mechanical resonance to justify the existence of intermediate substances. Where this resonance stabilizes the inter- mediate, it will be shown in customary style by a double-headed arrow (<^) connecting the several canonical resonating structures which serve as our designation of the resonance hybrid.* In representing the electron pairs in certain of the diagrams, dashes are used. Thus H— O— H instead of H:0:H The hydrogen ion will often be written as a base proton, H"^, although it is realized that the proton is solvated. This procedure is followed largely as a matter of convenience, but it should be pointed out that it is justified on other counts. First, the exact extent of *The reader is referred to such works as L. Pauling's The Nature of the Chemical Bond (1944) for discussion of the fundamental basis of the resonance phenomenon. 16 IMMUNO-CATALYSIS solvation is not known and it is really equally misleading to write a definite structure, such as H3O+ for the hydrogen ion. Secondly, it is not general practice to show the solvation of other ions, although nearly all ions are solvated because of the ion-dipole or ion-induced dipole forces present in their solutions. For these reasons the solvation will be shown only when it serves the purpose of clarifying the interaction of the solvent with the acids and bases present. General Acid and Base Catalysis. The role of acid or base as a catalyst in reactions involving esterification, hydrolysis, condensations, and carbonyl addition reactions is to accentuate the difference in electron density between the two reacting centers, as indicated in the following examples. a. Mutarotation of Glucose by Acid and Base Catalysis. It is known that when glucose is dissolved in water its specific rotation is + 110° which gradually sinks to +52.5° on standing. This change, known as mutarotation, is the result of the transformation of a-glucose HO-C HO-C ^C-OH O C— CHoOH /\H H OH a-d-Glucose C=0 HO-C OH Hxl HO-C C— CHoOH \ / H OH d-Glucose HO\,/H C HO-C HO-C C-CHoOH /Cx H OH ^-d-Glucose ANTIGENS AS BIOCATALYSTS 17 (+110°) into /3-glucose (+19°), and vice versa, until an equilibrium (at +52.5°) is established, i.e. a-glucose ^ ^S-glucose. In a-glucose the hydroxyl groups at Ci and C2 are in cis-, and in /?-glucose in trans- positions. The mechanism of the interconversion of d-glucose, a-d- glucose, and ^S-d-glucose has been extensively studied. The mutarotation of glucose involves the reversible change of one stereoisomeric form of glucose into another which differs only in the spatial arrangement of the groups attached to the carbon marked with *. This reaction requires the breaking and the reformation of one of its bonds. It has been shown that it is not the carbon to hydrogen link which has been severed, for if the mutarotation takes place in D2O this hydrogen is not exchanged by deuterium. It has therefore been assumed that the ether oxygen to carbon link is involved in this con- version. According to the general acid-hase catalysis the reaction proceeds in the following manner. H O— H HO-O 10— H HO— G C— CH2OH \/^H /^ H OH + + A- i H O \ H. HO— C HO— C 10— H + HA \/ ^H H OH C— CH2OH In the acid catalysis there is mobile and reversible addition of a H+ to the ether oxygen which is followed by a comparatively slow reaction with a base (A") producing the symmetrical carbonyl form. While in 18 IMMUNO-CATALYSIS the base-catalyzed reaction the hydrogen of the hydroxy! group attached to the carbon (*) Hnked to the ether-oxygen is first reversibly removed by the catalyst base. Then an acid reacts at a much slower rate with the anion formed to produce the symmetrical form. + HA H Ol HO— C lOI HO— C C— CH2OH H OH H 01 \' «x/ - HO— C 10— H + A "xl I HO— C C-CH2OH \/^H H^%H As can be seen, both an acid and a base are needed to catalyze the mutarotation. In this reaction there is an addition as well as the removal of a hydrogen ion. In either case, the conversion of the carbonyl form of glucose to the ring form may take place in either configuration of the carbonyl carbon, and this is responsible for the change to an equilibrium mixture of a- and /3-glucose, with a change of rotation to that of the mixture (Hammett, 1940, p. 337). b. Acid Catalyzed Enolization. Enolization in Concentrated Aqueous Acid. As an example of enolization in concentrated aqueous acids again the primary reactions involve the formation of the conjugate acid of the ketone which in turn loses its H+ to a molecule of water as a base. Hammett (1940, p. 276), believes that in strong acid this molecule of water forms part of the complex in the transition state of the enolization of acetophenone. ANTIGENS AS BIOCATALYSTS 19 H CeHs— C C . I I O— H H c. Enolization in Dilute Acid. H \ H I CHa R— C = 0— H < > R— C— O— H 1^ j LIBRAF jL^K^ MASS. 7i^ — ' I R— C— OH + HA II CHa The enol may change to the ketone and since it has a symmetrical ion as an intermediate this must lead to the racemization of the ketone in case the a-carbon is asymmetric. In base accelerated reactions of such ketones, satisfactory explana- tion of the course of the reaction involves the reversible formation of a very reactive carbanion as follows: R— C = + A~ . > I CH3 R— C = -<— > R— C— 01 I li CH2 CH2 + HA d. Acid-Base Catalysis of Condensation Reactions. Base Cataly- sis of Condensation Reactions. CH3— C— CH3 + CHsCHjO" — > CH3CH2OH + CH3— C— GHz" II II o o Acetone Ethoxyl ion Ethyl Alcohol Anion of acetone O II CH2— C— CIT3 CH3C— CH2~ + CH3— C+— O— CH2CH3 — > CH3— C— O— CH2CH3 U II I O 01 ~ iQl- Ethyl acetate 20 IMMUNO-CATALYSIS In the above reaction, ethoxyl ion functions as basic catalyst and removes a hydrogen from acetone making it electron rich (anion of acetone). This ion attacks the carbonyl carbon of the ester forming a new carbon to carbon bond. CH2C — CH3 II O CH3— C— O— CH2CH3 vini — IQI CH3— G— CH2C— CH3 + CH3— CHzO" II II o o Acetyl acetone Ethoxyl ion The pair of electrons on the oxygen of the carbonyl carbon will dis- place the ethoxyl group regenerating a new ethoxyl ion which can function as catalyst. e. Acid Catalysis of Condensation Reactions. In the acid catalysis of the above condensation reactions, the same products are obtained, as shown in the following equations. GH3 — C— OCH2CH3" II o H+ CH3 — C — OCH2CH3 11 lOH -> CH3C— OCH2CH3 I lOH CH3 C ^ 0x12"^ -^ CH3 — G — CH2 II 10— H CH3— C+— OCH2CH3 + H2C— C— CH3 I u IQH +QH (^OCHsCHa CH3 — C — CH2 — C — GHs I") I!) IQ— H |0— H — )► CH3C— CH2— C— GHs + GH3GH2OH + H+ II II o o In the acid catalysis the H+ will attach itself to one of the lone pairs of the carbonyl oxygen of the ester thus making it positively charged, which in turn draws out one pair of the double bond electrons of the carbon and oxygen bond thus rendering the carbon electron deficient and vulnerable to attack by the electron-rich methylene group of the acetone molecule. ANTIGENS AS BIOCATALYSTS 21 However, in the case of unsymmetrical reactants different reaction products are obtained by acid and base catalysis shown below (Ham- mett and Getder, 1943). H H CeHa— C = + CH3— C— CH2CH3 ^^^^ CeHs— C = CH— C— CH2CH3 II II o o Benzaldehyde Methyl ethyl ketone Ethyl styryl ketone No solvent, reagents saturated with HCl gas CH3C — C — CH3 II II O CH I CeHs Methyl ^-methyl styryl ketone f. Hydrolysis of Esters. Carboxyliq ester hydrolysis reactions are slow reversible reactions catalyzed by both hydroxyl and hydrogen ions. The alkaline hydrolysis of an ester is a reaction between the ester molecules and hydroxyl ions, while in the acid catalyzed reaction a molecule of water may form a part of the transition state complex. Base Catalyzed Hydrolysis. In this hydrolysis the carbon-oxygen bond that breaks is between the carbonyl carbon and oxygen and not the alkyl carhon and oxygen. This is based on the following observations (Hammett, 1940, p. 354). (a) Using amyl acetate in water containing labeled oxygen (O^^), Polanyi and Szabo (1934) showed that oxygen of the alcohol formed by the alkaline hydrolysis of the ester is the original ester oxygen and did not come from the water. (b) Alkaline hydrolysis of an ester containing an optically active alkyl group does not yield an inverted carbinol as should be expected if the bond between the oxygen and the asymmetric carbon of the alkyl group had been ruptured. (c) No rearrangement takes place during the hydrolysis if the alcohol part of the ester consists of an a-, ^S-unsaturated alcohol, thus eliminating the carbonium type reaction. 22 IMMUNO-CATALYSIS The hydrolysis of the esters derived from tertiary butyl alcohol forms an exception to the above mechanism in that the alkyl-oxygen linkage rather than the acyl-oxygen linkage is ruptured, as shown below. CHs CH3 : I HOH I R— C— O C— CH3 > R— C— OH + HO— C— CH, II : I II I O CHs O CHs (Cohen and Schneider, 1941) The hydrolysis can be represented as follows: 10— H R— C— C— R, + 1 10— H I I > • R— C-^i :^ lOi *- ^Q— Ri -Ri + |iQ-H I R— C=0 + 10— Ri HOH^ ROH + OH~ 10— H Alkaline hydrolysis of an ester is practically irreversible on account of the exhaustion of the anion due to the reaction : R— C— OH + NaOH > R— C— O" + Na+ + H2O II 11 O O As an exception to this mechanism, in esters of sulphuric and sul- phonic acids it is the alkyl oxygen linkage that is severed. In fact, these esters are used as alkylating agents. g. Acid Catalyzed Esterification and Hydrolysis. Acid catalyzed esterification and hydrolysis are reversible and may go towards comple- tion either to the right or left depending upon the excess of water or alcohol. In hydrolysis, the alkyl-oxygen link is not the one that is severed (Roberts and Urey, 1938, 1939), since it is shown by labeled oxygen studies that in esterification the ether oxygen of the ester comes from alcohol as shown below (Hammett, 1940, pp. 356-357). CeHs— C— Oi«H+CH30»«H ^CgHsC— O^s— CHs+HzQi^ II II O" Oi« ANTIGENS AS BIOCATALYSTS 23 Esterification Detailed presentation of the esterification is given below: R— C— OH + HA -7 y II O R— C— O— H I lO— H lOH H I I R— C— O— R lO— H + + A~ + R— O— H < > lO— H I +A~ . > R— C— O— R + HA I lO— H Hydrolysis R— C— O— R + HA -7 > II O R— C— O— R I lO— H R— C— O— R I lO— H + +IO— H:2=> I H H I lO— H I R— C O— R I lO— H + A" lO— H H I I R— C O— R I lO— H + lO— H H I I R-c g— R I lO— H + + A~ >-R— C = + HA + ROH I lO— H 2. Acid Catalysis in Non- Aqueous Solvents (Aprotic) Acid catalysis in non-aqueous solvents such as chlorobenzene have been investigated. The rearrangement of N-bromacetanilide, the inver- sion of 1-menthone and the racemization of C6H5COCH-(CH3)C6H5 and C6H5COCH(CH3)CH2CH3 are some of the examples. The re- actions are not kinetically of integral order but are fractional. With respect to the substrate the order is less than one. This can be attributed to hydrogen bonding type of complex formation between substrate and acid if the complex is chemically unreactive. Direct evidence has been found (Bell and Caldin, 1938) for the rapid and reversible forma- tion of complexes of menthone with various acids resulting in the depression of optical rotation. 24 IMMUNO-CATALYSIS For example, in 0.036m trichloracetic acid the [a]D of menthone was 21.1°, and in 0.762m acid the [a]D was 11.3°. This depression of optical activity was dependent on acid strength; for there was hardly any change in weak acids such as acetic acid, indicating that there was very little complex formation. The inversion of menthone may schematically be represented as follows: H I CH3 — G — CHs I *C— H H2C H2G c=o CH, *C— H I CH3 + ha; H 1 H 1 1 CH3 — C — CH3 1 CH3 — G — GH3 C— H /\ + H2G G = OH C— H H2G G— OH H2G GH2 \/ C H 1 ~s H2G GH2 \/ C— H 1 GH3 CH3 +A- H I GH3 — G — GHs H2G H2G \ G— OH + HA GH2 C— H I CH3 The examples presented here suffice to give a general idea about quite a large class of chemical changes which occur generally accord- ing to the mechanisms underlying the examples cited. From the standpoint of the subject matter the most fundamental aspects of these changes are: in the first place, when a chemical change takes place, a new substance is formed; secondly, the change is stoichiometrical, for every a-d-glucose molecule which undergoes mutarotation a molecule of /?-d-glucose is formed; thirdly, these changes are accelerated by catalytic agents, and none of the final iso- ANTIGENS AS BIOCATALYSTS 25 meric forms, or two or more isomeric forms at a state of equilibrium contain the catalyst as part of the molecule. Let us turn now to the production of antibody in the light of what we know about chemical changes. As discussed in the preceding pages, if the production of a new substance is brought about by the union of two reacting substances, then the reaction product or products should contain parts of the reactants according to stoichiometrical relationship. If on the other hand, such a reaction is accelerated by the presence of a substance which does not enter into the union irre- versibly with the reaction products, the role of this substance is one of catalytic acceleration. The absence of the catalyst or its parts in the final products of a catalytically accelerated reaction is one of the fundamental criteria of catalysis. If it can be demonstrated that the antibody formed in response to an antigenic substance does not contain the antigen molecule or its parts, the formation of the antibody can be assumed to have taken place by the antigen acting as catalyst. If, on the other hand, a chemical union has taken place between the antigen and the tissue substances forming the antibody complex, then this complex must be produced in accordance with stoichiometrical principles. The answer to this question must come from analytical and quantitative studies. 3. Characteristics Which Are Common to Inorganic Catalysts, Enzymes and Antigens Before we present the facts concerning the quantitative relationship between antigen and antibody produced, let us discuss what a catalyst is. For later discussions it is also necessary at this point to specify the difference that exists between an inorganic catalyst on the one hand, and the enzymes or biocatalysts on the other. The inorganic catalysts, no doubt, exist in nature and are produced under natural conditions. However, modern chemistry has produced quite a large number of inorganic catalysts in the laboratory as the need may have demanded, or they are accidentally discovered. These are, therefore, artificial catalysts. All the enzymes or the biocatalysts without exception, on the contrary, are produced by living systems. An enzyme is a complex organic substance produced by a living cell and utilized by the cell 26 IMMUNO-CATALYSIS during the activity of its life-cycle. It would therefore seem quite reason- able to assume that the special physiology of each type of cell is controlled by the specificity of the cellular enzymes. It is to be noted, however, that the catalytic activity of an isolated enzyme is generally independent of the living processes of the cell which produce it, although the activity of such an enzyme as part of the cellular system may be different, or a great deal more active, than when it acts as a single isolated chemical entity. Crystalline Enzymes. Pasteur and his followers regarded the cata- lytic activity of biological systems as an inseparable part of the phenomenon of life, and outside of experimental science. Liebig challenged Pasteur's point of view and advocated the existence of enzymes outside of the cell. Biichner in 1897 settled the argument finally by demonstrating that the fermentation of sugar could be caused by yeast extracts free from living cells; since then, not only many cell-free enzyme preparations have been obtained, but a score of enzymes has been obtained in crystalline form, which fact estab- lishes, beyond doubt, the chemical individuality of enzymes. Crystal- line urease (Sumner, 1926) was the first enzyme obtained in crystalline form. Later many crystalline enzymes and their precursors have been obtained and their purity studied by critical methods. They are: fefsin (Northrop, 1930); amylase (Caldwell, Booher and Sherman, 1931); yellow enzyme (Warburg and Christian, 1932; The- orell, 1934); Chymotrypsin (Kunitz and Northrop, 1935); carhoxy- -peptidase (Anson, 1935); ficin (Walti, 1937); papain (Balls, Line- weaver and Thompson, 1937); lysozyme (Abraham and Robinson, 1937); catalase (Sumner and Bounce, 1937); alcohol dehydrogenase (Negelein and Wulff, 1937); tyrosinase (Dalton and Nelson, 1938); lecithinase (Slotta and Fraenkel-Conrat, 1938); d-rihonuclease (Kun- itz, 1939; muscle d-glyceraldehyde-3-phosphate dehydrogenase (War- burg and Christian, 1939); yeast d-glyceraldehyde dehydrogenase (Krebs and Najjar, 1948); muscle lactic dehydrogenase or pyruvic acid reductase (Straub, 1940); peroxidase (Theorell, 1940); fumarase (Laki and Laki, 1941); Rennin (Hankinson, 1942); phosphorylase (Green, Cori and Cori, 1942); phosphate transporting enzyme, 1,3-diphos- phoglyceric acid + adenosine diphosphate ^ 3-phosphoglyceric acid + adenosine triphosphate (Schelling, 1942); myosin (Szent-Gyorgyi, 1943); serum mucoprotein with high cholinesterase activity (Bader, ANTIGENS AS BIOCATALYSTS 27 Schultz and Stacey, 1944); yeast hexokinase (Kunitz and McDonald, 1946); lif oxidase (Theorell, Holman and Akeson, 1947); bacterial a-amylase (Meyer, Fuld and Bernfeld, 1947); pancreatic desoxyriho- nuclease (Kunitz, 1948). The fundamental characteristic of all catalyzed processes is that they are reactions which, in the thermodynamic sense,* are classed as spontaneously occurring processes. That is, they are reactions which oc- cur with diminution of free energy. Any added substance which can accelerate such a reaction is known as a catalyst. A catalyst accelerates slowly progressing reactions and enables them to attain the equilibrium condition in a very much shorter time. The equilibrium point is not changed but the time necessary to attain this point is shortened. To illustrate this point we may use the following example. We know that iron can combine with oxygen to form various iron oxides. *The themiodjTiamic criterion for a reaction to occur spontaneously is that it do so with a loss in free energy of the system, change in free energy being the difference of the free energies of a system in its initial and final states. Hence, any system, left to itself, will change in such a way as to approach a point of equilibrium where the change in free energy will equal zero. Thus, for example, in a system consisting of two bodies at different energy levels there will be transfer of energy from one to the other at different rates. The body at the higher energy level, transferring energy at a greater rate, tends to raise the energy level of the second body until a condition is reached where both bodies are at the same energy level. No energy is thereby lost, but the capacity for spontaneous change has vanished, and the system is said to have less free energy. Tree energy is a function which indicates the direction in which chemical or other processes take place, and is of great theoretical value in fixing the conditions of equilibrium. The fundamental conception is that of a reversible cycle of operations. The condition of reversibility is that the state of the system at any time does not differ sensibly from equilibrium, for then the slightest variation in the con- ditions will determine the occurrence of the process in the one direction or the other. Kinetic theory postulates that chemical reactions take place only when molecules collide. However from chemical kinetics it has been demonstrated that only those collisions between reactants are effective in which the joint energy contributed by the molecules is equal to or greater than a certain minimum energy value termed energy of activation. A catalyst does not affect the point of equilibrium but accelerates the rate at which the equilibrium state is attained. Thermodynamically this may be expressed by saying that the change of free energy involved in a chemical reaction is the same whether a catalyst is present or not. Kinetically, this means that the catalyst accelerates the rate of the reverse reaction to the same extent as that of the forward reaction, so that the equilibrium constant, equal to the ratio of these rates, is the same for the catalyzed and uncatalyzed reactions. The function of the catalyst is to bring about the desired reaction with a smaller energy of activation. A lower energy of activation gives a more rapid reaction because more molecules have the necessary amount of energy to react. The high energy requirement is avoided by some by-pass. Usually the by-passing consists in forming an unstable activated cofwplex with less energy consumption and then decomposing this intermediate compound in such a way as to regenerate the catalyst. In this way the catalyst is used over and over again. 28 IMMUNO-CATALYSIS But under completely dry conditions a noticeable combination will not be observed. They will, however, combine on ignition forming oxide of iron. The applied heat accelerates the combination till sufficient heat of reaction is produced to make it self-sustaining. On the other hand, when we bring oxygen and iron together in the presence of moisture at ordinary temperature the reaction starts automatically, the moisture acting as catalyst. The role of moisture here was to reduce the energy of activation necessary for the attainment of the equilibrium of the re- action between oxygen and iron. As previously mentioned, there is no stable union between a cata- lyst and the reaction products or reactants although it may form in- termediate, reversible and labile complexes with these substances; but the lives of such complexes are very short; the combination is fugitive. It is stated that the mean life of an active H202-catalase com- plex (catalase contains trivalent iron) is 1.2X10"'^ second. This mean life is unique in being derived from purely chemical data, and does not involve any hypothesis as to collision. Likewise, the mean life of the excited oxyhaemoglobin complex (contains bivalent iron) is 4.0XlO-« second (Haldane, 1931). The fact that a minimum amount of a catalyst is capable of trans- forming a large quantity of substrate is understandable for the above mentioned reason that after the breakdown of the catalyst-substrate complex the catalyst can complete another cycle of acceleration, and continuously thereon. It is true that the unchangeability of an ideal catalyst during a reaction is one of its essential characteristics and responsible for its continued activity, but it does not need to be so, for when a catalyst takes part in a reaction and establishes the equilib- rium it is possible that a certain amount of catalyst is tied by the reaction products as specific inhibitors, or used up by the system so that its redelivery from the theoretical equation: (A and B reactants, A+C^AC; AC+B^AB+C C catalyst, AB reaction products, AC catalyst-reactant complex) is not complete. Also the catalyst might chemically combine by a side-reaction and thus be removed from the field of action. The catalyst might be destroyed by one of the reaction products, such as the destruction of the oxidative enzyme system of xanthine oxidase or pneumococcus by the hydrogen peroxide formed as one of the reaction products. Addi- ANTIGENS AS BIOCATALYSTS 29 tion of a little catalase at the outset to the reaction system decomposes hydrogen peroxide as soon as it forms, and thus the enzyme can continue functioning for many hours without weakening. a. Disproportionality between the Amount of Inorganic Catalysts and the Amounts of Substrates Catalyzed. Following the above discussion regarding the general and qualitative aspect of these proper- ties of catalysts let us now present a few quantitative data. The com- bination of hydrogen and oxygen at ordinary temperature could be brought about by 2.5 ml. of a colloidal solution of platinum containing as litde as 0.17 milligram of platinum, and at the outset the rate of combination was 1.8 ml. of gas per minute. After a period of time dur- ing which 10 liters of gas had undergone combination it was found that the activity of the colloidal solution was still unimpaired (Rideal and Taylor, 1926). The spontaneous slow decomposition of H2O2 can be accelerated by colloidal platinum in a dilution of 1 Mol (194 g. of platinum) in 70,000,000 liters. The presence of 0.000,000,000,000 IN. CUSO4 solution is sufficient to produce a perceptible acceleration of the rate of oxidation of an aqueous solution of sodium sulphite (Titoff, 1903). The oxidation of aniline hydrochloride, in the prepara- tion of aniline black, is carried out in the cold by a solution of potas- sium or sodium chlorate with the aid of metal catalysts, the most active of which is vanadium 'pentoxide, V2O5 of which one part is sufficient for 270,000 parts of aniline and the corresponding amount of chlorate (Sabatier, 1922). b. Disproportionality between the Amounts of Enzymes and the Amounts of Substrates Catalyzed. Calculations (Haldane, 1931) show that one molecule of liver catalase (or one trivalent atom of catalase iron) decomposes (2H202->2H20-l-02) 5.42X10^ molecules of H2O2 per second, at 0°C and 10~^ M substrate concentration. Under the same conditions one molecule of plant catalase decomposes 1.7X10^ molecules of H0O2 per second. One atom of peroxidase iron decomposes 10^ molecules of H2O2 in one second (Kuhn, Hand and Florkin, 1931). The iron of the cytochrome oxidase (Warburg) mani- fests an activity of about 10^ oxygen molecules per second (Warburg and Kubowitz, 1929); one molecule of saccharase hydrolyzes 7.0X10^ molecules of sucrose per second (Moelwyn-Hughes, 1933); one part of pure solid carbonic anhydrase (HsCOa-^COo+HoO) in 7,000,000 parts of solution is sufficient to double the rate of CO2 evolution and 30 IMMUNO-CATALYSIS during the first 1 5 seconds 1 g. of enzyme is responsible for the produc- tion of 825 g. of CO2, at a rate of 1 .24 moles CO2 per sec. per g. of enzyme (Roughton, 1934); urease crystals possess an activity of a little more than 100,000 units per gram, or 100,000 mg. of ammonia nitrogen is produced in five minutes when acting on urea at 20°C. at pH 7.0 (Sumner, 1932); solutions made of crystalline trypsin or pepsin containing less than 1/1,000,000 (0.000,005 M) of a gram of protein nitrogen per ml. have an accurately measurable effect on the digestion of casein, while solutions of pepsin containing less than 1/10,000,000 of a gram of protein nitrogen have a very powerful effect on the clotting of milk (Northrop, 1932). One g. of crystalline pepsin dissolves 50,000 g. of boiled egg white in two hours, clots 100,000 liters of milk and liquefies 2000 liters of gelatin during the same period of time. One g. of purified rennin (Tauber and Kleiner) clots 4,500,000 g. of milk in ten minutes. c. Disproportionality between the Amount of the Antigen Used and the Amount of Antibody Produced. The measurement of the amount of antibody produced in response to a given amount of anti- genic substance encounters considerable technical difficulties, some of which are almost insurmountable. These difficulties are: (1) although the circulating antibodies could be approximately estimated by pre- cipitin and agglutination reactions, we have as yet no way of determin- ing the amount of the antibody fixed in the tissues; (2) we cannot determine the actual amount of antigen responsible for the antibody produced; it is not unlikely that only a fraction of the injected substance is instrumental in the production of antibody; (3) when whole organ- ises are used we have no idea of the number of antigenic molecules in an injected quantity of bacteria. Despite these difficulties the prev- alent opinion among immunologists and bacteriologists is that there is a striking disproportion between the quantity of antigen and the total amount of the resulting antibodies. As early as 1893 it was demonstrated (Roux and Vaillard, 1893) that continuous bleeding of horses actively immunized against tetanus toxin did not diminish the antitoxin content of the regenerated blood. In similar studies (Salomonsen and Madsen, 1898) on diphtheriae toxin used to immunize horses, the above observation was confirmed. Much more interesting was the observation that when the antibody diminished, or nearly completely disappeared, the administration of ANTIGENS AS BIOCATALYSTS 31 non-specific substances, such as pilocarpin stimulated the restoration of the antibody formation. One unit of tetanus toxin produced (Knorr, 1898) 100,000 neutralizing antitoxin units in horses. A man, surviving a typhoid infection, w^ould still contain in his blood agglutinins for months and years, in spite of the fact that a certain fraction of the antibody would be eliminated (Friedberger, 1902) through urine, or other routes. It was shown (Friedberger and Dorner, 1905) that bacteriolysins are produced in rabbits by 1/1000 of a loopful of cholera vibrios killed at 60°C., and antibodies in rabbits by injecting a total of 0.5-1 .0 mgm. (300,000-900,000) goat red blood cells. In the light of some of the above facts concerning the disproportion- ality between the very small amount of antigen used and the many fold quantity of antibody formed, the Biichnerian hypothesis, which con- siders the formation of antibody as a consequence of chemical union between antigen and the antibody, loses its significance (Miiller, 1917). Newer and more accurately obtained quantitative data will be presented below concerning this question. Seibert (1925) found waters spontaneously became contaminated after standing at least four days under non-sterile conditions. Bacterio- logical tests revealed the presence of chromogenic and non-chromo- genic, motile and non-motile bacteria. Four liters of water on concentra- tion was found to contain 0.036 mg. of nitrogen (after subtracting the blank). She calculated that 1 ml. of fever-producing water would have contained 0.000,000,005 g. of protein. Ninhydrin tests on one liter concentrates of highly potent water were negative, as were also the tests on several bacterial filtrates. Immunological Studies. A rabbit was injected with 1/50 ml. of water. After an elapse of 12-14 days it was injected with 10 ml. of the same water; within one hour and ten minutes it developed typical shock symptoms, e.g. scratched its nose, fell to one side paralyzed, collapsed, expelled bloody urine, gasped for air and died with violent convulsions. Autopsy revealed an enlarged right heart and congested liver. Several other experiments gave her similar results. Immunization of the rabbits with traces of solid material present in 20 ml. of non- sterile distilled water with a protein content of 0.000,062 to 0.000,12 g. produced an agglutinating serum which was active in dilutions of 1 : 8000 against bacteria found in such distilled water. Branham and Humphreys (1927) found that the serum of animals 32 IMMUNO-CATALYSIS immunized with sterile filtrates obtained from B. enteritidis cultures incubated in a synthetic medium for six to 14 days contained ag- glutinins, precipitins and complement fixing antibodies. They could not demonstrate the presence of protein in these filtrates by the biuret, ninhydrin and Molisch tests; Millon's test for tyrosine and Ehrlich's and vanillin tests for tryptophane were negative. Ten liters of toxic filtrate were concentrated in vacuo at 40°C. to a syrupy fluid of 200 ml. The protein tests still were negative, although the residue was still toxic for rabbits. After dialysis of the residue against distilled water the dialysate was concentrated in vacuo at 30° to 40°C. to a volume of about 5 ml. This concentrate showed by ninhydrin, vanillin, and diazo tests (the last for histidine) very faint, but definitely positive reactions. By comparing the sensitiveness of these tests to the positive reactions shown by the concentrate they calculated the protein content of the original solution at least as 0.000,000,1 g. per ml. of filtrate. Rabbits on immunization with 30 to 40 ml. filtrates contain- ing a total of 0.000,003 to 0.000,004 g. of protein elicited definite antibody formation. The authors in the light of their work suggest that such infinitesimal amounts should be borne in mind in interpreting results obtained with apparently protein-free materials. Topley (1930) evaluated as nearly as possible the relation between the amount of antigen injected into rabbits and the resulting agglu- tinin titre. The antigenic material consisted of a saline suspension of Bact. 'paratyphosum B. in the type phase, and the antibody studied was the corresponding H or flagellar agglutinin. Bacterial suspensions containing 0.25 per cent formalin were killed by heating at 55°C. for one hour. A series of rabbits were injected intravenously with 10^, 10^ and 10^ bacilli per kilo body weight (k.b.w.); specimens of sera were collected at intervals of three to seven days during the first few weeks and at longer periods thereafter. The whole period of observa- tions varied from 50 to 300 days or more. In estimating the response to inoculation two values were noted: (a) the highest titre attained, and (b) the mean titre during the first 50 days after inoculation. The cal- culations were made from the graph drawn from the actual observa- tions. The tabulated results show that with a single injection into each of three rabbits of a dose of 10^ bacilli per k.b.w. they attained an immune serum with a highest agglutinating titre of 2550 to 4480 and ANTIGENS AS BIOCATALYSTS 33 a mean titre during the first 50 days from 880-1630 for the three rabbits; and with a single dose of 10^ bacilli per k.b.w. three rabbits showed a highest agglutinating titre of 130 to 510 and a mean titre of 37 to 280 during the first 50 days for a second series of rabbits. Thus Topley finds that a dose of 10^ bacilli per k.b.w. is in the neighborhood of the threshold dose for a detectable response. To determine the actual amount of antigen contained in a dose of this order a dense bacterial suspension was dried and brought to con- stant weight over calcium chloride; and the dry weight of 10*^ bacilli was found to weigh 0.55 mg. A dose of 10^ bacilli thus corresponded to about 0.000,005 mg. of solid material. When allowance had been made for the possible presence of non-bacterial material brought in with bacteria by washing out the agar in preparing the suspensions, and for the flagellar material which could not have contained very much antigen, it would seem probable that the dose of active sub- stance administered with 10^ bacilli was of the order of 0.000,000,5- 0.000,005 mg. After such a consideration Topley used a conservative figure of 50 ml. of serum contained in the blood of a rabbit of 2000 g. and evaluated its agglutinating titre as 1 : 500 or more. He thus pre- sented a quantitative study to emphasize the well known and striking disproportionality between the amount of antigen injected and the amount of antibody produced. Fiooker and Boyd (1931) using the data obtained by Topley carried out calculations to determine the amount of antibody produced by a single antigen molecule. They assumed that the above mentioned weight of material— as evaluated by Topley— present in 10^ bacteria as the minimal effective dose of active antigenic substance is of the order of 5X10~^° gram per k.b.w. of rabbit having an agglutinating titre of serum 1 : 500. Considering that the fundamental specific anti- genic unit has a molecular weight of 20,000— a very conservative figure— they arrived at the following conclusions: One gram of material will contain, after Avogadro, 5X10"^ (6.06X10-^)=3X10^^ molecules, and the minimal eff^ective dose of antigen (5X10~^*^ g.) per k.b.w., contains therefore approximately 1.5X10^" molecules. One ml. of serum from rabbit 11 responsive to 8.3X10^° molecules in Topley's experiments contained 5000 or more agglutinin units. A "unit" was estimated to affect the agglutination of 2X10^ bacteria 34 IMMUNO-CATALYSIS or a much larger number of flagella. Fifty ml. of serum (estimated con- servatively by Topley) obtainable from the 2760 g. rabbit would yield 2.5X10^ units of agglutinin, resulting from the stimulus of S.BXlO^'* antigen molecules, and capable of flocking 5X10^^ bacteria. In other w^ords, one molecule of antigenic substance gives rise to an amount of circulating agglutinin capable of flocking 600 bacteria. As they state, this evaluation does not take into account the large reservoir of anti- body in tissue fluids and cells which by repeated daily bleeding would contribute in no small measure to a magnification of this figure. Studies by Heidelberger and Kendall (1930), and Heidelberger, Kendall and Soo Hoo (1933) likewise demonstrated the striking dis- proportionality between an artificial antigen and the amount of anti- body produced. The antigenic substance was colored R-salt-azo- benzidine-azo-crystalline egg albumin. The injections of the antigen were carried out either in solution, or adsorbed on collodion or alum particles. Fifty ml. of a suspension of antigen adsorbed on collodion, the amount used for 16 injections, contained 0.55 mg. of dye and 140 mg. of collodion. In all but one series the injections were given intra- venously, four daily injections each week for four weeks; many animals were given an additional course of two, three or four weeks. Quantitative data showed that as much as 0.73 to 0.94 mg. of cir- culating antibody per ml. of serum may be formed in response to in- jections of antigen totaling 0.35 to 0.55 mg., or a total response for the rabbit of over 100 mg. of circulating antibody for every milligram of antigen injected. Since the ratio of antigen to antibody in the pre- cipitate has been found to average 1 : 7 at the equilibrium point, the authors stated that at least 12 times as much antibody as necessary to combine with the amount of antigen used is formed. This evaluation, of course, does not take into account the antibodies present in the tissues as well. Morgan (1937) extracted an antigenic material from B. dysenteriae (Shiga) by diethyleneglycol. This material contained a polysac- charide, gave positive biuret and negative ninhydrin reactions. On immunizing a horse with a total of 16.2 mg. of material the production of 1.64 mg. of antibody protein per ml. of blood serum was attained. Morgan assumes a blood serum volume of 25 liters obtainable from the immunized horse, or a total of 41 g. of antibody protein in re- sponse to 0.0162 g. of antigenic material, a ratio of (41/0.0162) ANTIGENS AS BIOCATALYSTS 35 2500:1. His findings show also that the ratio of mg. of antibody pro- duced per mg. of antigen injected, for rabbits that have been im- munized was considerably smaller than the one obtained for the horse. In the five sera examined they ranged in the order of 550:1, 550:1, 200:1, 400:1, 1000:1, respectively. Pappenheimer (1940) immunized a horse against egg albumin. Dur- ing the entire course of injections the horse received a total of 1 .9 g. of egg albumin. Assuming that the animal contained 25 liters of serum at the time of the final bleeding the total circulating anti-egg albumin antibody was estimated to be in the neighborhood of 50 g. A horse was immunized vdth a total of 60 mg. (20,000 Lf . units) of diphtheria toxoid by Dr. W. B. Rawlings of the National Drug Com- pany over a period of one month. At the end of this period the horse contained 2150 units of antitoxin per ml. of serum, equivalent to 25 mg. of antitoxin per ml. which was estimated (Pappenheimer, 1940) to correspond to more than 600 g. of total circulating antitoxin or more than 10,000 times the weight of antigen injected. d. Absence of Inorganic Catalysts, Enzymes and Antigens in the Catalyzed Reaction Products. In the preceding pages several quan- titative studies were cited demonstrating that the amount of a catalyst or an enzyme in ratio to the amount of the products of reaction they accelerate is incomparably small. Likewise, experimental data were cited demonstrating that the ratio of the amount of antibody produced to the amount of antigen used is strikingly disproportionate. These quantitative relationships can be interpreted to signify that the antigen could not have entered into stoichiometrical chemical union with y-globulin to form the final antibody complex. Miiller (1917), Heidel- berger and Kendall (1930), Topley (1930), Hooker and Boyd (1931) have expressed this view in opposition to the Biichnerian hypothesis of antibody formation. Hooker and Boyd after having calculated that one molecule of active antigenic substance gives rise to an amount of agglutinin capable of flocking 600 bacteria, and that the surface relationship between one globulin molecule and 600 bacteria is 1:25,000,000, stated that "A theory of antibody formation involving catalysis would seem more promising." If our assumption that antibody produced by catalytic acceleration of antigen is true, then we must be able to demonstrate chemically that antigen is actually absent in the antibody molecule. For when 36 IMMUNO-CATALYSIS colloidal platinum catalyzes the combination of oxygen and hydrogen with the formation of water, copper sulphate catalyzes the rate of oxidation of an aqueous solution of sodium sulphite, and urease the transformation of urea into ammonium carbonate, etc. no traces of catalysts are found in the molecules of the reaction products. Doerr and Friedli (1925) were perhaps the first to undertake the task of showing the absence of antigenic substances in the antibody molecule by chemical analysis of the latter. By immunizing rabbits with atoxyl-containing azo-proteins, highly active anti-atoxyl specific immune sera were obtained. Analysis of these specific sera with a highly sensitive chemical method failed to show the presence of ar- senic. On the basis of their findings they concluded that antibody cannot be a metabolite originating from the substance of antigen. Following the above mentioned work Berger and Erlenmeyer (1931) diazotized the sodium salt of p-aminophenylarsenic acid (atoxyl) and coupled it with normal horse serum. The purified antigen, which contained 0.000,449 g. arsenic per ml. of solution, was used for the immunization of rabbits, administering four intravenous injections of 2.0, 4.0, or 6.0 ml. of antigen solution. The animals were bled 13 days after the last injection. The immune sera thus obtained re- acted with the antigen solution of 1 : 3200-6400 dilution. Using 30 ml. of immune serum for the chemical detection of arsenic they did not find any trace of arsenic in two rabbit sera. The serum of a third rabbit showed a faint trace which was definitely weaker than a positive control containing 10~^ g. (O.Oly) arsenic. A rabbit weighing two kilos was calculated to contain 90 ml. of serum so that the arsenic content of the total serum in none of the three rabbits could have been 0.000,03 mg. (0.03y). For the detection of the arsenic in serum they used a micro method based on the Marsh test whereby an amount of arsenic as little as O.Oly (10~^ g.) could be demonstrated. In order to show whether ar- senic was present in the serum they analyzed the sera of rabbits im- mediately— 314 to 20 minutes— after injecting them with atoxyl. Their positive findings thus showed that arsenic, if present, in serum can be determined under the experimental conditions. The evidence, there- fore, appears to show conclusively that arsenic in antigen is not incorporated in the resulting antibody. Hooker and Boyd (1932) immunized rabbits with casein-diazo-ar- ANTIGENS AS BIOCATALYSTS 37 sanilic acid which had a calculated molecular weight of 114,000 and contained 78 diazo-arsanilic acid groups combined with one casein molecule. A calculation of the serological results showed that the ratio by weight of antigen to antibody was 1 : 18, or adopting a liberal figure, was at least 1 : 10. Analyzing their quantitative data, they stated that if the most probable calculation is made, on the basis of the Biichnerian hypothe- sis, 15 ml. of immune serum should contain IS.By of arsenic. In actual experiments this volume of serum showed no trace of arsenic, though trial runs showed that the method would detect 1 .Oy of arsenic. In a study previously cited, Heidelberger, Kendall, and Soo Hoo (1933) immunized rabbits against a red azo dye, R-salt-azo-benzidine- azo-crystalline egg albumin. If the red colored whole antigen or its red prosthetic group had been incorporated into the antibody ac- cording to the Biichnerian hypothesis, the immune sera, or the isolated antibody should be colored. The immune sera from the rabbits showed no trace of color. Wollman and Bardach (1935) showed the falseness of the Biich- nerian hypothesis by a highly sensitive anaphylactic test. The fact that by means of an anaphylactic test the presence of a protein in a mixture of several proteins can be detected was used to determine whether or not a protein antigen, assumed by the theory to have been incorporated in the homologous antibody molecule would cause ana- phylactic shock in guinea pigs sensitized against the same antigen. The guinea pigs sensitized against egg albumin and horse serum and injected with homologous protein suffered shock and death, but the sera of rabbits immunized against these antigens produced no symp- toms of anaphylaxis in another group of guinea pigs sensitized respec- tively against egg albumin and horse serum. Haurowitz, et al. (1942) found that homologous antibodies pro- duced in rabbits to iodo-proteins, bromo-protein, caseinogen (phospho- protein) and arsanil-azoprotein did not contain any of the determinant group of the antigen or a serologically related group. In the absence of any other plausible explanation which could be offered at present, the most logical possibility therefore is that anti- body production is directed by antigen acting as a catalyst at the site of the antibody formation. This point of view appears to meet the criteria of the well known catalytic processes employed in all branches 38 IMMUNO-CATALYSIS of chemistry. Before we proceed to the elaboration and discussion of other aspects of this problem, we must also discuss the question of whether or not enzymes are antigenic. For the production of a specific antibody in response to an antigen acting as a directive specific catalyst and the antigenicity of an enzyme of which the apparent function is to catalyze biological processes specifically, are intimately related. e. Enzymes as Antigens. It is a generally accepted fact that the antigenic property of a substance, or more specifically, the production of antibody against an antigen is usually dependent on the protein molecule. It is also an accepted fact that the enzymic activity of a substance of biological origin is dependent on the presence of protein in the enzyme molecule. Even in those enzymes the specificity of which is associated with a prosthetic group, this group must necessar- ily be conjugated with a specific protein molecule or a "Kolloidal- trager" of protein nature. In its absence the prosthetic group is enzymically inactive. Likewise, when an antigen is conjugated with a non-protein group, such as atoxyl, the antigenicity is conditioned by the protein molecule, even though the prosthetic group may be the factor determinant of serological specificity. From the standpoint of chemical composition, both antigen and the enzyme proteins are of the same nature. Generally speaking, their amino-acid content, physical properties, molecular size and their behavior to various chemical treatments are of the same nature, or order of magnitude. There is no theoretical basis, therefore, to exclude the enzyme proteins from the family of antigenic proteins. The antigenicity of an enzyme protein can be determined by the well known serological methods: (a) by antigen-antibody reaction in vitro causing the formation of a precipitate, (b) by anaphylactic tests, and (c) by inhibition of the enzymic activity as a consequence of the combination between the enzyme and its specific antibody. Undoubtedly, other methods supplementary to one or the other method can also be employed. Of the methods the first and second will be discussed here briefly, and the third will be reserved for a later dis- cussion when we come to the correlative treatment of the biological significance of antibodies against antigens (enzymes). In characterizing the antigenicity of an enzyme only by the pre- cipitation method, we must be aware of the fact that every enzyme preparation cannot be accepted as free of other protein impurities. ANTIGENS AS BIOCATALYSTS 39 If a preparation contains an enzyme protein and an inactive protein impurity the immune serum produced against such a mixture will react in the precipitation tests with one or both of them; it will hence be difficult to state whether the precipitinogen was the enzyme or the contaminating protein. We will therefore include only immunological studies carried out using the more highly purified crystalline prep- arations. Antibody against Crystalline Urease. Antiurease was the first anti- body produced against a crystalline enzyme, and this is probably the first indubitable proof of the existence of an immune anti-enzyme. Sumner and Kirk (1931; Kirk and Sumner, 1931, 1932, 1934; Sumner, 1937; Howell, 1932) obtained antiurease which behaved like antitox- ins prepared against bacterial toxins in precipitation and neutralization experiments. Antiurease serum inhibited the catalysis of urea into ammonium carbonate by urease, and protected animals against the toxic and fatal effects of ammonia resulting from the dissociation of ammonium carbonate. Crystalline urease prepared from jack bean meal of high urease content had an activity of about 135 units per milligram. The crystals under the microscope were seen to be practically uncontaminated by any other material. When 1 ml. of crystalline urease solution containing 100 units was injected into the ear vein of rabbits they experienced convulsions in a few minutes and death wdthin an hour. When two rabbits were in- jected intraperitoneally or subcutaneously with 0.5 ml. (50 units) of solution death followed within 48 hours. To prevent the death of rabbits 2.5 to 5 units of urease were injected at the start. The injections were given every eight days for 30 days and then every two or three days for 30 days. The last injections contained 600 units of urease. The serum of rabbits showed the presence of precipitin, though of low titer. However, such a serum was capable of neutralizing the urease, and the urease-antiurease combination was used for immunization against urease. Animals therefore were usually started with a suspension of urease-antiurease in large doses (containing 100 to 1000 urease units, per animal). This indicated that a very effective protective antibody had been produced during the first immunization experiments. During the course of this study it was found that the recrystallized urease of highest purity gave better results than the once-crystallized urease. 40 IMMUNO-CATALYSIS The immune sera of the rabbits contained from 30 to 40 antiurease units per ml. and the amount of antiurease in the blood serum of an immune hen varied from five to 24 neutralizing units per ml. The rabbit immune serum precipitated urease in dilutions up to 600,000. Urease which had been denatured by contact with 0.05 N hydrochloric acid for a few seconds and which was then brought to neutrality gave no precipitate with antiurease whatsoever. Urease completely inacti- vated by formalin was non-toxic and produced no antiurease when injected into rabbits. It was shown that antiurease protected animals from poisoning with urease. When two rabbits were given 90 neutralization units of antiurease each and three hours later 90 units of urease were given there were no symptoms of poisoning. A rabbit given 90 units of anti- urease just before being given an injection of 80 units of urease was likewise unaffected. On the contrary, two rabbits which were given injections of 90 units and one rabbit given an injection containing 80 units of urease all died within five hours. All injections were intraperi- toneal except the third rabbit which received the antiurease by ear vein. In order to find out whether rabbits near death from urease poison- ing could be saved by injecting antiurease, six two-kilogram rabbits were injected intraperitoneally with 65 to 70 units of urease. When the rabbits had become totally paralyzed ( 1 V^ to two hours afterwards) each rabbit was given 80 units of antiurease by injection into the ear vein. Four of the rabbits showed immediate improvement and became normal within one hour. The other two died. Similar results were obtained with guinea pigs. Antibody against Trypsin, Chymotrypsin and Chymotrypsinogen. The methods of preparation and chemical properties of some of the crystalline enzymes as found by the Rockefeller group at Princeton are compiled in a book by Northrop, et al. (1948). Since it was dem- onstrated that chymotrypsin differs in its properties from trypsin. Ten Broeck (1934) undertook the demonstration of an immunological difference between the two enzymes. Trypsin and chymotrypsin are the enzymes principally responsible for the proteinase activity of pan- creatic juice. Neither alone digests protein very far, but the two to- gether cause hydrolysis to proceed to the polypeptide stage. Trypsin decreases the clotting time of normal or hemophilic blood, but under ANTIGENS AS BIOCATALYSTS 41 ordinary conditions it does not clot milk. Chymotropsin, on the other hand, clots milk but not blood. It probably represents the pancreatic rennin of Vernon. Trypsin appears to be identical in its specificity with Waldschmidt-Leitz's "Proteinase." Sensitization of Guinea Pigs. The antibody production by the above enzymes and pro-enzymes was tested by anaphylactic test. This test, rather than the less sensitive precipitin reaction, was used in the major portion of the tests for differentiation for the reason that the enzymes may be very closely related. Since the sensitization of an animal is tantamount to immunization and anaphylactic shock depends upon an antigen-antibody reaction, the anaphylactic reactions have the same significance for the question under consideration as the precipitin reaction in vitro (Landsteiner, 1936). Female guinea pigs weighing about 125 g. were given subcutaneous injections of 0.5 ml. of either 0.5 or 1 per cent solution of the enzymes, purified by five crystal- lizations. Injections of chymotrypsinogen produced no visible effect, but the animals receiving the trypsin and chymotrypsin showed necrotic areas at the site of inoculation, and several of them died. Chymotrypsin seemed to be more toxic than trypsin. Between 15 and 20 days after the first injection the uteri of these guinea pigs were tested by means of the Dale technique (Dale, 1931). Two baths were used and the two horns of the uterus were tested, one after the other. The capacity of each bath was 75 ml. and 0.75 mg. of the en- zyme solution was added for each test (1 : 1000 dilution of the enzyme in the bath). The uteri of guinea pigs receiving pig trypsin injections did not react against beef trypsin and chymotrypsin in 1 : 1000 con- centrations; on the other hand, pig trypsin showed definite positive reactions with both left and right horns of the guinea pig uterus sensi- tized with pig trypsin. In three groups of similar tests the results showed that (1) guinea pigs receiving beef trypsin were sensitized against itself, but not against chymotrypsin and chymotrypsinogen; (2) guinea pigs receiving chymotrypsinogen were sensitized against itself but not against chymotrypsin and beef trypsin; and (3) guinea pigs receiving chymotrypsin were sensitized against itself, but not against chymotrypsinogen. Ten Broeck makes the following state- ment as a summary of his studies: "Not all of the animals were sensi- tized, and in some cases there were cross-reactions, particularly between the chymotrypsin and chymotrypsinogen. The results were, however, 42 IMMUNO-CATALYSIS sufficiently clear cut to show that all four of these enzymes and their precursors can be differentiated by this reaction." Antibody against Pefsin and Vefsinogen. Seastone and Herriott (1937) carried out experiments to distinguish by serological methods the pepsins from several different animal species as well as to compare the serological behavior of pepsin and its precursor, pepsinogen. Northrop had previously reported (1930) that crystalline swine pepsin protein gave rise to pepsin precipitating antibodies. Seastone and Herriott were aware of the fact commented on by other investigators, that pepsin is inactivated above pH 6; as a more alkaline condition is approached the enzyme is converted into a typical denatured protein. At normal body temperature and at blood pH 7.6 it is therefore most likely that active pepsin in the body fluids is inactive, and denatured pepsin is responsible for antibodies developed following the injection of active pepsin. The limitations imposed by its denaturation at pH 7.6 must be accepted. Immunization. Rabbits weighing about 2 kg. were given, at weekly intervals, three intraperitoneal injections of 5.0 ml. of a 1 per cent solution of swine pepsin at pH 5.0. This material had been twice crystallized and dialyzed. Two weeks after the last injection, the serum was collected. Precipitin reactions were done by the ring test, and readings were made after \Vi hours at room temperature. Of four rabbit sera, two showed no pepsin precipitins; one precipitated pepsin solution (pH 7.6) at a concentration of 1:1000 and 1:100,000. Anti- sera against swine serum were also prepared, the strongest reacting wdth normal swine sera in dilutions (on the basis of dry weight) of 1:100,000. Pepsinogen gave rise to precipitating antibodies more readily than pepsin. At pH 7.6 it is a stable native protein. Of the four rabbits immunized with pepsinogen, two had a titer of 1 : 100,000 and two 1:1,000,000. Their findings further showed that alkali (pH 7.6) denatured pep- sins from swine, cattle, and guinea pigs precipitated in swine pepsin antiserum; pepsin from the rabbit, chicken, and shark treated simi- larly did not precipitate in swine pepsin antiserum. Pepsin antisera reacted with both pepsin and pepsinogen but did not react with the serum proteins from the homologous species. Anti-sera made with serum proteins did not react with the homologous pepsin or pepsino- gen. Pepsinogen anti-sera reacted with pepsinogen, but not with ANTIGENS AS BIOCATALYSTS 43 twice-crystallized pepsin, nor with the serum proteins from the homol- ogous species. These findings demonstrate beyond doubt that pepsin protein is specifically antigenic. It also appears that pepsin retains the serologi- cal specificity of pepsinogen to an appreciable degree. On the other hand, pepsinogen is serologically an independent entity. Its activation into pepsin, associated with chemical changes (Herriott) partially retains this specificity. Antibody against Catalase. After the isolation of beef liver catalase by Sumner and Dounce (1937) and horse liver catalase by Bounce and Frampton (1939) in crystalline form, Tria (1939), and Campbell and Fourt (1939) reported the immunization of rabbits with crystalline beef catalase. Tria immunized rabbits by injecting them with 12.5 mg. of enzyme every three days for three weeks in the first set of ex- periments. In a second set of experiments he started with 1.25 mg. of enzyme and gave gradually increasing doses. The immune sera had high anti-catalase activity. The sera reacted in 1:10 optimal dilution with a solution of catalase containing 0.1 to 1 mg. Determining the anti-catalase activity of the serum quantitatively he obtained 4 anti- catalase units in 1 ml. serum. A 50-fold purified anti-catalase isolated from the catalase-anti-catalase precipitate after dialysis had 2200 anti-catalase units per g. of antibody. Anaphylactic experiments by Tria also showed the presence of anti- body in the serum of guinea pigs actively immunized against beef and horse liver catalase. Anaphylactic shock resulted in the death of the immune animals within a few minutes. The results of anaphylactic experiments aiming at a serological differentiation of the species specificity of beef, lamb, and horse liver catalases were inconclusive. The precipitation tests by Campbell and Fourt likewise demonstrated the formation of antibody against beef liver catalase. The results obtained from experiments involving the use of dog and horse liver catalases as antigens likewise did not yield con- clusive information regarding the question of the species specificity of catalases. Summary and Conclusions. The formation of antibodies in the animal system in response to antigenic stimuli has been discussed from various viewpoints. The stoichiometrical relationship underlies all known chemical changes— simple combination, decomposition, double 44 IMMUNO-CATALYSIS decomposition, tautomerism and mutarotation. A catalyst does not change stoichiometrical relationships of reactions, does not enter into any irreversible stable union either with the reactants or with the reaction products; its function is therefore one of repetition and con- tinuity. For this reason it can transform a disproportionately large amount of substrates into reaction products. Numerous examples have been cited to emphasize this property of inorganic and organic catalysts (enzymes). Antigens likewise have been shown to produce dispropor- tionally large amounts of antibodies. Catalysts and enzymes in no case have been shown to be part of the reaction products; similarly, highly sensitive qualitative and quantitative analytical tests have not been able to demonstrate the presence of antigens or their parts ("marked" antigens) in the antibody molecules. It thus appears that antigens do not function as reactants in the stoichiometrical sense in the formation or synthesis of antibody molecules. The experimental data which have been presented would therefore seem to show that the role of antigens in stimulating the formation of specific antibodies is one of catalysis, which fact brings the antigens within the class of biocatalysts. If antigens are believed to exercise a catalytic role in the production of antibodies, enzymes likewise have been shown to demonstrate the property of stimulating the production of specific antibodies. This property of enzyme proteins has been a subject of controversy for many years, although recently it has been accepted as an established fact. Thus antigens and enzymes possess two properties in common- catalytic activity and antibody production. These two properties are interwoven in the production of antibodies. 4. Do Catalysts (Antigens) Make a New Reaction Possible? The two properties of antigens and catalysts discussed above meet two of the basic criteria of the concept of catalysis. These properties are: (a) infinitesimal amounts of inorganic catalysts, enzymes and antigens catalyze the interaction of disproportionately large amount of reactants, and (b) neither antigens nor enzymes (and inorganic catalysts) form a part of the reaction products. There is also another criterion of the concept of catalysis which must be satisfied by enzymes and catalysts as well as antigens. A catalyst can only accelerate (not cause) a thermodynamically possible reaction. In other words, a catalyst ANTIGENS AS BIOCATALYSTS 45 does not create a new reaction; it does not make the impossible pos- sible; it simply accelerates a possible reaction. It has been stated that "a catalyst not only accelerates a reaction, but makes a reaction possible" (Willstatter, 1927; Schade, 1923; see also Mittasch, 1935, 1938). This statement should not be taken too literally. This implies, for example, that hydrolysis of proteins is not demonstrable in the absence of proteolytic enzymes, but in the presence of traces of these agents marked hydrolysis occurs. The effect is dramatic. At the surface it may appear that such an effect is equivalent to the creation of new substances by, apparently, non-existent reactions. These processes are, however, thermodynamically possible. The fact that a catalyst encourages and accelerates such tendencies which already theoretically exist is not equivalent to the creation of a new reaction, or making the impossible possible. 5. Does Antibody Synthesis Involve New Processes Which Did Not Already Exist in the Animal System? If we assume that the basic chemical processes responsible for the synthesis of antibody are different from those of the serum globulins we must also assume and demonstrate that there are basic chemical differences between the antibody and serurri globulins. If this should prove to be true, then it could also be assumed that the antigen has produced in the animal system new processes for the synthesis of anti- bodies. This would necessarily create a discrepancy between the con- cept of catalysis and that of antigen exercising the role of a catalyst. If on the other hand, the available experimental data show that the chemical characteristics of both the antibody and serum globulins are practically indistinguishable then it is reasonable to accept the thesis that the synthetic processes involved for both the immune and normal globulins are essentially the same. The directive influences, however, may bring about certain configurational changes in the globulin molecules to account for the serological specificity of antibody globu- lins. In other words, the change from normal globulin to antibody globulin involves a change of the "configurational pattern" and not of the basic structures. Thus the above mentioned discrepancy between the concept of catalysis and that of antigen acting as a catalyst would not exist. This point of view will receive further support if we take 46 IMMUNO-CATALYSIS into consideration tKe occurrence of iso-antibodies as precursors or prototypes of all the antibodies. Iso- Antibodies. From the known facts it can be concluded that the unique influence of antigens apparently is the shaping of certain parts or groups of normal globulins in conformity with the specific parts of antigens. Even this influence does not seem to be a new process of animal cells. The presence of iso-antibodies in the animal systems shows that antibody synthesis is a genetically established process. The formation of additional specific antibodies in response to the injected bacterial and other foreign proteins appears, therefore, to be an extension of the number and art of the synthesis of antibody globulins already being manufactured; in other words, a change of cer- tain details in the genetically determined general scheme of antibody and serum globulin synthesis. Landsteiner (1900, 1901) described and established the occurrence of iso-antibodies and their corresponding agglutinogens in human blood. He divided the blood from normal individuals which he ex- amined, into three groups, namely: A, B, and C, on the basis of the agglutinative reactions. The sera of group A agglutinated the red corpuscles of group B, but not of group C; the sera of group B agglutinated the red corpuscles of A but not of group C; the sera of group C agglutinated the corpuscles of both A and B. The sera of the fourth group described by Decastello and Sturli (1902) failed to agglutinate the corpuscles of the above cited three groups, but its corpuscles were agglutinated by the sera of all of them. The blood of zoologically related species shows immunological relationship; thus the bloods of chimpanzee and man, mouse and rat, etc. are related. The blood cells of pigeon, rabbit and man are agglutinated by goat's serum; and the blood cells of each species will absorb out from goat's serum its own specific agglutinin but not the agglutinin which reacts with the blood cells of the other species. The four classical blood groups are inherited according to Mendelian principles, and all experimental evidence to date points towards the chromosomes as the bearers of the hereditary factors, or "genes." What interests us here most directly is the fact that the synthesis of these antibody globulins is a genetically determined process. The basic antigenic unit in natural and artificial conjugated pro- teins is the natural protein itself. The prosthetic groups coupled with ANTIGENS AS BIOCATALYSTS 47 the protein do not initiate the formation of, but contribute to the architectural pattern (specificity) of the antibodies. It is immaterial whether a protein is used in its natural form or after having been coupled with a chemical group non-existent in nature. The coupled prosthetic groups which merely influence the specificity of antibodies are comparable to the type or group specific carbohydrates of various bacterial antigens. The prosthetic groups of artificial or natural con- jugated antigens in many respects are similar to the conjugated en- zyme proteins. In conjugated protein antigens the prosthetic groups alone are non-antigenic; similarly the prosthetic groups of the con- jugated protein enzymes are catalytically inactive by themselves. The heme group in hemoglobin, catalase and peroxidase is comparatively inactive without a combination with their specific native proteins. Various co-enzymes must likewise be in conjugation with specific proteins to manifest their activity. Part U Mechanism of Antibody Formation A. THE FACTORS CONTROLLING THE PRODUCTION AND PERSISTENCE OF ANTIBODY A PRESENTATION of a chain of reactions leading to the appearance of immune bodies in response to an antigenic stimulus in the complex in vivo environment is obviously impossible because of present lack of precise information. It is likewise difficult to ascertain the nature of the factors which influence the rate of formation, rise and decline in the amount of, and eventual disappearance of antibody from humoral systems. An approximation to these questions may, perhaps, be achieved by an analysis of the available results which may have bearing on these questions. The following discussion is an attempt in this direction. 1. The Relation of the Specificities of Host Enzymes to the Antigenicity of Substances Foreign to the Species of the Host In connection with the problems pertaining to immune response to an antigenic stimulus, there are two questions into which we need to inquire. They are: a. The absence of an immune response to an antigenic substance derived from one individual in another belonging to the same species;* and *There are reported certain results which might appear to be not in full agreement with the view here discussed. For example, Hektoen and Schulhof (1925) reported that a rabbit thyroglohulin produced a precipitin in a rabbit. Also overlapping precip- itin reactions were obtained among thyroglobulins of beef, dog, horse and human; human and rat; horse, human and sheep; beef, sheep and human, and swine, human, sheep and dog. On the other hand, Stokinger and Heidelberger (1937) demonstrated definite species differences among various thyroglobulins in addition to organ speci- ficity, corresponding in general to the biological relationship of the animals from 49 50 IMMUNO-CATALYSIS b. The immediate immune response to an antigenic substance when introduced into an individual belonging to a different species. A reasonable understanding of the difference in immune response under the above two conditions would necessitate that we become acquainted with in vivo metabolism of protein and other non-protein antigenic substances. It is an established fact that antibody protein is a modified globulin. Immune globulin which has been obtained from which the thyroglobulins were derived. Thyroglobulin is a conjugated protein con- taining thyroxine and a protein. Since the thyroxine group, common to the thyro- globuUns of various species, would be expected not to function as a common deter- minant haptenic group, the overlapping precipitin reactions described by Hektoen and Schulhof could be due to the denaturation and thereby loss of the specificity of thjTToglobulins they isolated. Thyroglobulin is known to circulate in normal blood and does not function as an antigen under these circumstances. Its reported antigenic- ity in homologous species might very well be due to its denaturation incurred during its preparation. Kato (1924) reported that the serum of rabbits immunized with a heterologous fihrinogen may react with it and also with that of other species, but not with rabbit fibrinogen; and the serum of rabbits immunized with rabbit fibrinogen reacts with other fibrinogens but not with the immunizing rabbit fibrinogen. These findings woiald indicate that the rabbit fibrinogen exercises species specific and non-specific serological activity as well, and that fibrinogens from various species possess a common denomi- nator or, as antigens, they represent mixtures of denatured and native fibrino- gens. Hektoen and Welker (1927) obtained overlapping reactions among the fi- brinogens from chicken, duck, goose, guinea-hen, pigeon and turkey. They studied the fibrinogens of the common mammals and concluded that they have antigenic elements that are more or less common, but that bird fibrinogen does not belong immunologically to this group; however, it is not wholly distinct and different from manrnialian fibrinogen. Fibrinogen is unique among the serum proteins. It is in- soluble in salt-free water but is soluble in dilute salt solutions. It is the most readily precipitable of all the common blood proteins by concentrated salt solutions e.g., half saturation with sodium chloride or 20 per cent ammonium sulfate. It is also unique among the blood proteins in that it is readily converted into insoluble fibrin by the action of thrombin. It is an asymmetrical molecule with dimensions of 33x900A. Though the coagulation temperature of fibrinogen is 55°C in neutral solution, the heat coagulation of fibrinogen (and fibrin) during a period of one hour, according to Robbins (1945), did not aEFect the antigenic nucleus. In view of these properties, one may wonder whether or not the laboratory prepara- tions of fibrinogen are removed from the native form. In this respect it may be of some interest to reinvestigate the problem with a view to the role of — S— S— or — SH groups in fibrinogen in relation to their possibly masked immunological speci- ficities as has been demonstrated with keratins and ocular lens proteins discussed below, also for the possible reasons that, according to Baumberger's suggestion (1941), -S-S- ^ -SH relationship might play a role in the conversion of fibrinogen into fibrin, and that reducing agents, e.g., cysteine, glutathione, sodium bisulfite (ChargafF, 1945) inhibit fibrin formation (see page 282). For a long time keratins were considered immunologically indistinguishable. Pillemer, Ecker and Wells (1939), Pillemer, Ecker and Martiensen (1939) showed, however, that species specificity is an individual characteristic of keratins and that the observed specificity is dependent on the redox state of the sulfhydryl groupings in the protein molecule. In a similar study, Ecker and Pillemer (1940), contrary to MECHANISM OF ANTIBODY FORMATION 51 the antigen-antibody complex and purified by pbysico-chemical means, has been shown to possess the serological species-specificity of normal globulin and the acquired additional property of reacting specifically with the homologous antigen. It would be of interest, therefore, to consider the metabolism of normal and immune globulins, and other proteins as well. Experimenting with young and adult rabbits, Cannon (1945), Cannon, et at (1943), and Wissler, et al (1946) reported the long accepted indistinguishable specificity of the ■proteins of the ocular lens, demonstrated the species specificity of the proteins of lens from chicken and fish, though this characteristic was not so evident in the closely related species swine and sheep. According to Lewis (1934) casein is iso-antigenic. Sera of two goats, one immunized with goat casein, and the other with cow casein, reacted equally well with both caseins. The chief objection to these results seems to be the method of the prepara- tion of casein which involved precipitation with normal acids and mechanical stirring at the rate of 2000 to 3000 revolutions per minute. The system was exposed to O.IN acid (final concentration) for six to eight hours. The precipitated casein was washed five times vidth water, twice with 95 per cent alcohol and three times with ether. It was then dried by spreading. Under these conditions the species specificity of native casein might easily be destroyed as a result of denaturation, and behave like a non-specific antigen. For it has been showoi that, in contrast to native rabbit serum protein, denatured rabbit serum would produce antibody in the rabbit (Landsteiner, 1945). One may also call to attention the possible non-specific antigenicity of the pros- thetic group, serine-phosphate or glutamyl-serine phosphate (Levene and HUl, 1933; Schmidt, 1933, 1934), common to all mammalian caseins. The phosphate in casein is ortho-phosphoric acid esterified with the hydroxyl group of serine. It is a relatively strong acid and its titration curve shows a sharp inflection between pH 6.0 and 7.5, due to, possibly, secondary ionization of the phosphoric acid group. The antigenicity of phenyl phosphoric acid is an established serological fact. Another possibility which may be worth mentioning is the fact that casein is a pro- tein synthesized by the female mammal by an organ which is entirely inactive and undeveloped in the male. For these reasons one may be tempted to consider it as a protein foreign to any mammalian system. Bruynoghe (1935) reported that the egg- white from a hen, precipitated with ammonium sulfate and dialyzed to remove the salt, produced specific antibody in the same hen and in a rooster. He assumed that egg-albumin is a protein totally different from the components of serum and has not participated at all in the cellular metabolism of the animal. In this manner, he con- cluded, egg-albumin might exercise iso-antigenic activity. However, one must keep in mind the fact that egg-albumin is very susceptible to denaturation during its preparation. Ferritin is a metal protein of 500,000 molecular weight and contains from 17 per cent to 23 per cent Fe. Granick (1943) found that apoferritin derived from crystalline ferritin, with the exception of a slight reaction between horse apoferritin antibody and dog apoferritin, is immunologically species specific, and non-organ specific. Protein crystals of ferritin obtained from one organ reacted with antiserum to ferritin crystals obtained from another organ. Bonnichsen (1947) found that liver and blood crystalline catalases were serologi- cally identical; the values for nitrogen distribution, histidine, arginine, lysine, tyro- sine, cystine, glutamic and aspartic acids were found to agree writhin the limits of error for both catalases. 52 IMM UNO-CATALYSIS that "Adult rabbits made hypoproteinemic by a low protein diet or by low-protein diet supplemented by plasmapheresis exhibited a lessened capacity to produce agglutinins as compared with animals of similar age but supplied with an adequate diet," and "Protein repletion of protein-depleted rats by the feeding of high-quality protein, or a hydrolyzate of a high quality protein led to a markedly increased out- put of hemolysin, evident as early as after two days of repletion and pronounced within seven days." The agglutinative titers revealed that the average agglutinin-output of the well-fed rabbits was about five times that of the protein-depleted ones. The serum protein levels of the well-fed group rose to an average of 6.62 grams per 100 ml. whereas those of the low-protein group of rabbits declined to an average level of 4.76. These results show that factors controlling protein metabolism are directly related to the production of antibody. According to the studies of Whipple and his collaborators there exists a dynamic equilibrium between the proteins of blood and those of tissues. That is, there is a "give and take" between body and plasma proteins. When plasma protein is depleted, replacement is possible by the proteins of the organs (Madden and Whipple, 1940). The results of the studies of Schoenheimer and his collaborators (1942) with isotopic amino acids have shown that the proteins are in dynamic equilibrium with their constituent fragments. When an isotopic amino acid was added to the diet of a host, the concentration of marked nitrogen in the serum proteins increased immediately, but diminished steadily after the addition of isotopic amino acid to the diet was dis- continued. Chemical reactions had thus occurred among the units of the individual serum proteins, resulting in the uptake of dietary nitrogen in the first period and its removal in the second. The above findings may be presented in the following manner: Serum proteins ^ amino acids ^ tissue proteins This dynamic equilibrium of protein metabolism, involving pro- teolysis and protein synthesis is mediated by species specific proteinases. From what we know of the species-specificity of proteins and the speci- ficity of enzymes, it would be reasonable to assume that the enzymes of the individuals belonging to the same species metabolize each other's proteins, given parenterally, in an identical manner. Under these con- MECHANISM OF ANTIBODY FORMATION 53 ditions, they can exercise no antigenic stimuli. The readiness with which these proteins are dispensed with is most Hkely due to an absence of structural difference among the proteins of the individuals of a given species. In contrast, the parenteral injection of a protein, derived from a species different from the species of the recipient host, being struc- turally different will resist greatly the action of the host enzymes. Due to this resistance, the life of the whole protein, or its structurally specific part, will be prolonged. Transferred by phagocytes to a center of protein metabolism, not only will it resist complete degradation but would seem also to influence the course of the synthesis and certain details in the pattern of the normal globulin, yielding immune anti- body globulin. In this manner, the foreign protein, or antigenic unit, exercises the role of specific catalytic modifier, or superimposes a new catalytic role of its own on the enzymes which synthesize globulins. This role continues so long as the antigenic unit remains intact. Any process whereby the life of the antigen in the host is prolonged might result in a degree of antibody response. The difference between a species specific substance and that which is foreign to the host rests on the distinctive differences in the specifici- ties of host enzymes (or the genes which are assumed to be responsible for the origin of enzymes). The ready digestibility and utilization of a species protein by the species specific host enzyme system is understand- able, for the specialized enzyme system is capable not only of digesting such a protein when parenterally introduced but also is capable of synthesizing it in an identical pattern. This may be presented in the following manner: species enzyme Species protein v amino acids On the other hand, a host can either eliminate a foreign protein (or substance) by excretion, or digestion by virtue of the inherent abilities of proteolytic enzymes to split peptide linkages common to all proteins. Because of the basic structural (or architectural) difference of the molecular species, there is no doubt that the elimination of the foreign protein when parenterally introduced, will proceed at a very different rate than the digestion of the species specific protein. In con- trast to this ability of the host enzymes to eliminate a foreign substance, they totally lack the ability to resynthesize or regenerate the foreign- 54 IMMUNO-CATALYSIS protein from its hydrolytic products. This would mean that the host enzyme would catalyze the reactions of a foreign protein only in a forward direction: host host enzyme enzyme Foreign protein >- amino acids ^ host protein, etc. This relationship is particularly worthy of consideration when we are dealing with artificial conjugated antigens. These differences in the rates of the metabolism of species specific and foreign proteins, or the failure to metabolize a foreign protein by a host could no doubt account for the longer life of a foreign substance in a host as will be discussed below. This relationship is of fundamental significance in the production of antibodies to foreign substances. A foreign protein may possess component units uncommon to the host species. Many microorganisms possess unusual proteins or poly- peptides composed of single or different unnatural amino acids, and non-protein uncommon substances as well. Ivanovics and Bruckner (1937, 1940) found that the specifically precipitable capsular sub- stance of B. anthracis is a polypeptide-like substance of a molecular weight of about 6000 made up solely of 40 to 50 d-glutamic acid residues. According to Hanby and Rydon (1946) the capsular sub- stance is made up solely from d-glutamic acid residues. The molecular weight of the native material is greater than 50,000 and is thus of the same order of size as many proteins. Structurally, the capsular sub- stance is a long chain molecule made up of a-peptide chains of 50 to 100 d-glutamic acid residues joined together by y-peptide chains of d-glutamic acid residues. Since d-glutamic acid is of unnatural optical configuration the proteolytic enzymes of the animal system would either fail to digest or effect the digestion of this polypeptide at a very slow rate. The resistance of this polypeptide to proteolysis would not only confer on it a haptenic role but may also serve as an armor for the microorganism against the host enzymes, prolonging the antigenic activity of the cellular components and virulence as well. The occur- rence of polypeptides which are indistinguishable from that of the anthrax bacillus have been likewise found in the organisms of the mesentericus and subtilis groups. In Saccharomyces cerevisiae a poly- peptide composed of 10 to 12 glutamic acid residues was found to be MECHANISM OF ANTIBODY FORMATION 55 linked at the terminal part of the chain to one molecule of p-amino- benzoic acid through a carboxyl group, as reported by Ratner, et at. (1944). It is also to be noted that gramicidin and tyrothricin obtained from B. hrevis are polypeptide containing natural 1- and unnatural d-amino acids. These polypeptides are resistant to the action of crude trypsin, pepsin, papain and papaya latex at several pH values (Hotch- kiss, 1944). This resistance is attributed to the d-amino acid contents of these polypeptides which exercise antibacterial action and toxicity for animal cells and tissues. The consideration of the above cited observations may help us to visualize the structural differences of antigenic substances derived from animal, plant and bacterial sources. These differences, no doubt, play a significant role in resisting the host enzymes and thereby pro- longing the life of antigenic substances and the immune response they produce in a host. In this connection, the following observations are of interest. In a study on the behavior of antibody protein toward dietary nitrogen in active and passive immunization, the following experiment has been carried out (Heidelberger, et al; Schoenheimer, et al. 1942). A rabbit actively immunized against Type III pneumococcus was given a single large injection of Type I antipneumococcal rabbit serum. The administration of isotopic glycine by addition to the stock diet was started two hours before injection and continued for 48 hours. Daily estimations of the amount of circulating antibody (passively admin- istered) and of the isotopic N^^ concentration of antibody and residual proteins were made. Since, in this passive immunization, the antibody introduced could only be hydrolyzed but not be regenerated, its daily estimation gave an idea as to how long it could persist in the rabbit act- ing as a species specific host. At hour the amount of passively intro- duced Type I antibody corresponded to 1 .09 mg. of total antibody nitro- gen/ml. of serum. After 22V^ hours it was 0.88 and after 48 hours 0.49 mg. of total antibody nitrogen/ml. of serum. Thus 55 per cent of the passively introduced antibody was eliminated within 48 hours. Atten- tion was drawn to the fact that the passively introduced antibody enters into metabolic reactions which lead to its disappearance, but not to the regenerative uptake of nitrogen, as in the case of actively produced antibody. The failure of passively introduced antibody to take up heavy nitrogen, they stated, can less reasonably be ascribed to a generalized "foreign protein" character than to the absence of some 56 IMMUNO-CATALYSIS specific function of the tissue cells of the host, meaning the specific function which is operative in active immunity. The rate at which the passively introduced anti-pneumococcal rabbit antibody disappears in the body of the rabbit shows that 37. 6 per cent was eliminated during 22 Vi hours, and 55 per cent during 48 hours. In contrast, the foreign proteins (or polysaccharides) administered to a host for immunization or for therapeutic reasons persist for a decidedly longer period of time. As will be discussed below, production of an antibody can occur within four to 10 hours after the injection of an antigenic material. It is therefore reasonable to assume that the length of the period, during which a species specific protein can persist in a host of the same species, appears to be long enough to exercise an antigenic stimulus. The absence of such a stimulus would emphasize the importance of the difference of molecular structure of the foreign antigenic substances from those of the host as the most critical differ- ence in the stimulation of an immune response. In contradiction to the observation of Heidelberger, et al., Kooyman and Campbell (1948) reported that antibody globulin introduced passively into rabbits can enter into the dynamic equilibrium state of the body without loss of antibody properties. Experiments were carried out by injecting C^^ labeled dl-leucine into a rabbit which was actively immunized against p-azophenylarsonic acid-ovalbumin. Ten days after the last injection, an injection of 10 ml. of concentrated rabbit pneu- mococcus antiserum (Type I) was given intravenously, followed by a similar injection next day. On the same day, an injection of 30 ml. of a one per cent leucine solution was given intraperitoneally. The leucine contained C^^ in the carboxyl group. Similar injections of leucine were given on the three following days, 1.15 g. of leucine being injected in all. Samples of serum were taken on 5, 9, and 16 days after. The anti- bodies were precipitated with corresponding antigens, washed with saline three times, once with distilled water, twice with alcohol and ether, and analyzed for C^^ after drying. These materials were reported to contain labeled carbon. If interpreted to indicate that passively introduced antibody globulin is digested to the stage of amino acids and peptides and the C^* con- taining antibody globulin is resynthesized in the absence of the specific antigen, these findings would be contrary to the facts underlying the concept of the mechanism of antibody formation, namely, that antibody MECHANISM OF ANTIBODY FORMATION 57 synthesis occurs under the specific stimulus of an antigen. No excep- tion to this rule has as yet been found. The presence of C^^ in the assayed antigen-antibody precipitates could, perhaps, be best explained in the following manner: (a) Antibody globulin molecule can ex- change amino acid residues in a dynamic environment without under- going appreciable digestion or losing parts from the basic antibody molecular unit; (b) normal serum components and the minimum antibody molecular units (e.g., diphtheria antitoxin, etc.) can enter into reversible combinations, as shown below: Normal serum component + antibody unit ^ Antibody Complex (containing C^"*) and (c) the extreme precaution taken to remove possible C^* irnpuri- ties from the antigen-antibody precipitates was perhaps inadequate and drastic dialysis at various [H + ] might have been necessary to remove the C^'* impurities from the reactants to start with. With this in mind, the following observations concerning the length of time an antigenic substance can reside in a host are presented. According to Uhlenhuth and Weidanz (1909) 5 ml. of horse serum in- jected intravenously into a rabbit was found in the blood 1 5 days after the injection. With the gradual increase of precipitin there was a cor- responding diminution of the horse serum. Longcope and Mackenzie (1920) reported that 5 ml. of horse serum per kilo body weight of rabbit injected intravenously was detectable after three weeks. These latter investigators studied also the presence of horse serum at given intervals after its injection into human beings for therapy of pneu- monia. The results of experiments with 14 individuals, who received from 100 to 500 ml. of antipneumococcal horse serum intravenously yielded results falling into two groups. In the first group of eight cases, the persistence of horse serum ranged from 18 to 39 days. The injection of horse serum was followed by severe serum disease, lasting from 11 to 28 days, and as a rule the precipitins appeared first or their concentration increased markedly towards the end of the serum disease. On the other hand, the pre- cipitin reaction for horse serum diminished rapidly towards the termina- tion of the serum disease and disappeared shortly thereafter. In the second group of 5 cases, the precipitinogen in the serum persisted from 49 to 69 days. The precipitin formation was either of extremely short 58 IMMUNO-CATALYSIS duration or entirely absent. These patients either had very mild serum disease lasting from one to five days or had none at all. One individual who had received 630 ml. oF horse serum showed the persistence of horse serum in the circulation for 75 days. He also showed precipitins and had severe serum disease lasting 12 days. These observations were made, in the main, by the use of macroscopic precipitin tests, and there- fore are rough estimations. By the use of refined quantitative micro- technique one may be able to detect circulating antigens during a longer period of time. Such estimations, however, would fail to inform us about the amount and the duration of antigens bound in the tissues which may continue to exercise antigenic stimuli during a far greater period of time than are indicated by the tests for their presence in blood and urine. Herdegen, Halbert and Mudd (1947) developed an in vivo tech- nique whereby they demonstrated that 0.002 mg. of Shigella para- dysenteriae Type III antigen was capable of inducing an agglutinin response. This method was used to determine the fate of this antigen in mineral oil emulsion stabilized with lanolin derivatives at the site of subcutaneous inoculation in mice. It was found that the antigen per- sisted at the site of injection for at least 24 weeks when 200 i^g. of antigen were injected. This was demonstrated by covering the injected vaccine from period to period; 200 /tg. of antigen per 0.3 ml. of original volume of vaccine was recovered after a period of two weeks. This de- creased to 20 fj-g. after 12 weeks, but 0.2 //.g. was still present at 24 weeks. The material recovered from the injection site was still anti- genic. The deterioration of antigen at the site of inoculation was paralleled by a fall in the agglutinin titers of the mice receiving the oil vaccine. In connection with the above observations, a reference to the find- ings of Freund and Bonato (1946) is of interest. Using water-in-oil emulsion stabilized by Falba (a mixture of oxycholesterine and choles- terine derived from lanolin) of killed typhoid bacilli as vaccine, they reported the presence of antibodies in the sera of animals three years and one month after one injection and three years and five months after two simultaneous subcutaneous injections. The persistence of pneumococcal polysaccharide in the circulation for a long period of time has been reported by various investigators. Dochez and Avery (1917) carried out a systematic study of the pres- MECHANISM OF ANTIBODY FORMATION 59 ence of soluble specific substance in the blood and urine of experi- mental animals and of patients suffering from lobar pneumonia. A positive reaction for soluble specific substances (Types I, II, III) in the blood was found as early as 12 hours after the initial chill, and was demonstrable in one instance five weeks after defervescence. A positive precipitin test with urine against anti-pneumococcal serum was, in certain cases of Type I pneumonia, demonstrable as late as 42 days after the infection; with Type II pneumonia, after 58 days, and with Type III, 30 days after the infection. Quigley (1918) studied 82 cases of Types I, II, and III, pneumococcus lobar pneumonia and obtained positive urine tests for soluble specific substance in 81 per cent of the cases as late as the 21st day during convalescence. Similar results were obtained by Blake (1918). Pepper (1934) following the excretion of polysaccharide in urine by several cases of lobar pneumonia, reported that two Type I cases of empyema excreted large amounts of S sub- stance late in the disease, one of them until the 41st day and the other until the 27th day. Urinary S substance did not appear until agglu- tinins for Type I pneumococci appeared in the blood. In immunization experiments, Avery and Goebel (1933) found that acetyl polysaccharide, corresponding to the above cited soluble specific substance, persisted in the circulation of the treated rabbits for con- siderable period of time, was slowly excreted by the kidney, and ap- peared in the urine in its naturally acetylated form, though, for reasons as yet not understood, acetyl polysaccharide does not induce any im- mune response in the rahhit. However, the fact remains that the type specific antigenic component of pneumococcal vaccine, acetyl poly- saccharide, is excreted without being attacked by the host enzymes. On the other hand, it is a well-established fact that this same substance, or its slightly modified forms, produce immunity in mice and men (Schiemann and Casper, 1927; Saito and Ulrich, 1928; Wadsworth and Brown, 1931; Zozaya and Clark, 1932, 1933; Sevag, 1934; Felton, 1935; Heidelberger et al., 1946). It is further to be noted that the polysaccharides which persist in the circulation without producing immune response were shown to be antigenic determinants by coupling with serum globulin (Avery and Goebel, 1931). Type III polysaccharide coupled with serum globulin immunized rabbits against infection by virulent Type III pneumococcus, and the immune serum contained type specific antibodies which precipitated Type III 60 IMMUNO-CATALYSIS polysaccharide, agglutinated Type III pneumococci, and specifically protected mice against Type III infection. These findings show that these polysaccharides are, apparently, incapable of forming a polysac- charide-protein antigenic complex in vivo with rabbit proteins, in a manner, perhaps, comparable to those possibly formed when pure polysaccharides are injected into mice and men. An interesting new observation on the effect of a large dose of pneumococcal carbohydrate on antibody production is reported by Felton, et al. (1947). A relatively large dose of antigenic polysaccharide of pneumococci type-specifically "paralyzes" the immunological mech- anism of mice during the life span or for 1 5 months. This paralysis was reported to be due to the presence of polysaccharide. From observations of a large number of "paralyzed" mice, the pneumococcus polysac- charide was found present in the following tissues in order of decreas- ing precipitinogen concentration: liver, spleen, kidney, skin, bone marrow; irregularly present in muscle (especially Type III), lung, in- testines, and urine, and absent in heart and blood. In three instances, the polysaccharide was isolated, partially purified, and tested for im- munizing activity. Type I, 5 gamma protected against 50,000 lethal doses. Type II, 0.5 gamma against 500,000 lethal doses, and Type III, 5 gamma against 500 lethal doses. The amount of polysaccharide in the tissues gradually decreased with increasing interval following injection; but after 1 5 months, it was still present in the liver of animals paralyzed against Type I and II, and, in one experiment. Type III. In addition, mice injected with a paralyzing dose of pneumococcus whole-cell vaccine, showed a similar distribution of polysaccharide in the tissues. These observations may indicate, as the investigators sug- gested, that the large dose of polysaccharide "paralyzes" the antibody forming mechanism of the animal, or that antibody formed is neutral- ized by combining with polysaccharide and is immediately eliminated because of the continuous presence of a large excess of antigenic polysaccharide. The immunological paralysis is a typical specific blocking, for anti- bodies to immunizing doses of heterologous antigens are synthesized. The enzyme systems which synthesize other immune globulins must therefore be in an intact state. Only the response to the polysaccharide antigen is lacking. Since the polysaccharide antigen is found in nearly all the cells of immunologically paralyzed animals, failure to demon- MECHANISM OF ANTIBODY FORMATION 61 strate the presence of homologous antibody was interpreted (Fehon, 1949) to be due to so firm an attachment of the antigen to cell sub- stances that it is no longer free to act as an antigen. If we accept the idea that antigen functions as a catalyst, the above interpretation of "immunological paralysis" would be in agreement with the fact that a catalyst which irreversibly combines with a component of a reaction system would be incapable of catalyzing the specific process. The fact that these antigenic polysaccharides persist in the circula- tion for long periods and are excreted through the kidney unchanged fails to indicate a relationship between these results and an in vitro observation by Dubos and MacLeod (1938) that pneumococci are very rapidly rendered non-antigenic (in regard to type-specific antibody) by leucocyte extracts. 2. An Inquiry into the Nature of Factors Controlling the Balance of Antigen and Antibody During Immuni- zation For an understanding of the theories advanced concerning the mech- anism of the formation of antibodies it may be of interest to consider the possible factors which influence the rate of antibody production, its decline, and also the length of the period during which an antigen may remain in the host system. The amount of antibody formed in response to antigenic stimuli will be dependent on the rate at which antibody is removed, or dissociates from the site of its synthesis into the humoral systems. With the accumulation of antibody in the humoral system and tissues an equilibrium condition will be established.* The attain- ment of this final state of equilibrium may represent the peak of anti- body concentration. The maintenance of the peak of antibody concen- tration would be dependent on several factors. The rate at which *The question of how soon the above reaction equilibria can be set up in vivo may best be answered by a consideration of the following observations. Two hours after subcutaneous injection, egg-white was demonstrated in the urine and blood by the precipitin reaction (Ascoli, 1902; Obermayer and Pick, 1902). According to Dochez and Avery (1917), if rabbits are infected intraperitoneally with pneumococcus a substance specifically precipitable with antipneumococcus serum can be demonstrated in their blood stream, freed from bacteria by filtration, from within two to six hours following the time of infection. One may find other similar observations in the literature. These facts show how rapidly the antigenic material finds its way into the circulation introduced through various routes. 62 IMMUNO-CATALYSIS antibody is formed and the rate at which it is metabolized, antibody ^ amino acids ^ normal proteins, will condition the maintenance or decline of the peak of humoral concentration of antibody. The rate of antibody formation will be governed by three principal factors: (a) The ability of antigen to resist the host's catabolic activities. (b) The rate at which antibody accumulates and thereby shifts the equilibrium from left to right in the following complex reaction mechanism : cell-antigen-antibody ^ cell-antigen -j- antibody. Because of the strong specific affinity between antigen and antibody, antigen- antibody complex formation, in the presence of an accumulated amount of antibody, may predominate, blocking the activity of antigen (see p. 137) to direct the formation of new antibody."^ However, this con- dition cannot last long for the obvious reason that the amount of anti- body will soon diminish by participating in the protein metabolism of the host, or by some special mechanism for eliminating antibody from the antibody forming cells, and, therefore, the reaction will resume its forward trend once again. (c) The magnitude of the dissociation constants derivable from the reaction equilibria. The rates of the backward and forward reactions in the establishment of the states of equilibria are governed also by the magnitude of the dissociation constants of various reactions. The following may be cited to illustrate the possible reaction equilibria: CO Antigen-cell ^ cell -|- antigen (2) Cell-antigen-antibody ^ cell -|- antigen -f- antibody (3) Antigen-cell-antibody ^ antigen -(- cell -|- antibody (4) Antigen-antibody ^ antigen -|- antibody From the standpoint of continued antibody formation, the most critical of the above relationships would appear to be a very small degree of dissociation of the cell-antigen complex,! supplemented with *In the presence of an excess of antigen Csevere infection) small amounts of anti- body formed could not be demonstrated because of antigen-antibody complex forma- tion and elimination. tin connection with the above relationships it may be pointed out that the inter- action between antigens, antibodies, and the susceptible tissue cells could be expected to exercise marked effects on the above reaction equilibria. It has been pointed out by Friedmann (1947), for example, that the avidity of tetanus toxin for the tissue cells will result in slower rate and degree of reaction between the toxin and the antitoxin. The combination between the tetanus toxin and tissue was stated to undergo a gradual increase in firmness and that, consequently, it became increasingly difficult for the antitoxin to dislodge the toxin from the tissue. The resistance of toxin to antitoxin under conditions as they prevail in the natural disease is explained by the high avidity MECHANISM OF ANTIBODY FORMATION 63 a greater degree of resistance of the antigenic molecule to the catabolic actions of the host enzymes. The interrelationship of the various component parts of the complex reaction mechanism of antibody formation may, perhaps, be presented in the following scheme : Cell -\- antigen ^ cell-antigen complex cell-antigen Globulin factors ^= — — i cell-antigen-antibody complex cell-antigen -\- antibody i i cell -|- component amino acids units of antigen normal proteins In the synthesis of antibody globulin, antigen functions as a catalyst when present in trace amounts. When present in large amounts it also functions as a reactant in a stoichiometrical reaction with antibody and thereby exercising a determinant role in regulating the amount of antibody procurable in the host. The above formulations would seem to be amply supported by the following observations. According to Hamburger (1902), a precipitin reaction for egg-white appeared in the serum of rabbits two hours after subcutaneous injec- tion, reached a maximum after 24 hours, persisted until the third day and had disappeared on the fourth day. Ramon (1928) re- ported that following the immunization of horses with the filtrates of formolized broth cultures of pest bacillus there was produced flocculat- ing antibody within from six to eight hours after the injection. The serum obtained 36 hours after the injection contained a four-fold greater amount of antibody. Similar results were obtained in experi- ments with six other horses. The above cited observations may be looked upon as exceptional or unusual cases; however, they indicate the speed with which the immune mechanism may be set in operation following the introduction of antigens into a host. of the toxin for nerve tissues (see further Friedmann, 1947; Wilson and Miles, 1946). That anti-virus antibodies enter into combinations with tissue cells has been shown by Sabin (1935) and Andrews (1928). 64 IMMUNO-CATALYSIS In an extensive investigation on the antigen-antibody balance in lobar pneumonia, Blake (1918) studied: (a) daily blood cultures throughout the course of the disease; (b) daily determinations of the concentration of soluble specific substance in the blood and urine, and (c) daily determinations of agglutinins and precipitins in the blood serum. All patients who failed to excrete soluble antigen in the urine during the course of the disease developed precipitins in the blood at or about the time of crisis, v\'hich may indicate that the development of precipitin in the body keeps pace with the elaboration of the antigen and that the two serve to neutralize each other until finally the develop- ment of precipitin exceeds the formation of soluble antigens, and free precipitin appears for the first time in the blood. Furthermore, the concentration of precipitin, after rising rapidly during the period shortly after crisis, fell rather abruptly coincident with the appearance of soluble antigen in the urine. Apparently, according to Blake, this is associated with the liberation of a considerable amount of antigen by the resolution of the pneumonic consolidation. At least it was coinci- dent with it. The fact that in two cases during this period small amounts of precipitin in the blood and of soluble antigen in the urine were simultaneously present, was explained as due either to antigen and precipitin simultaneously present in approximately equal amounts in the body and not in all instances completely bound, or that the kidney possessed the power of separating the two, excreting the soluble antigen and retaining the precipitin. The fact also that recovery was invariably accompanied by the appearance of agglutinins was striking, and suggested to him that here again a struggle between living antigen (i.e., pneumococcus) and its corresponding antibody was taking place. On the other hand, in those cases, usually severely infected, in which soluble antigen was excreted throughout the course of the disease, and in which precipitin did not appear in a free state in the blood, it appeared to Blake that the formation of precipitin never equalled the elaboration of soluble antigen, which was constantly in excess and readily excreted in the urine. He attributed his inability to detect the presence of soluble antigen in the blood in these cases to probable insufficient refinement of his method of testing. As will be dis- cussed below, it may have been due to the constant neutralization and removal of the soluble antigen by the antibody present in the blood stream and excretion through the kidneys. It was also observed that, in MECHANISM OF ANTIBODY FORMATION 65 fatal cases, there was an increasing septicemia, steady rise in the amount of soluble antigen excreted in the urine, and complete failure to develop either precipitins or agglutinins in the blood. Apparently in fatal cases the body was unable to combat the progress of the infec- tion by the production of sufficient antibodies to neutralize and over- balance either the living antigenic cell or its soluble products. The coincidence of antigen and antibody in the blood of an im- munized animal has been variously reported (for a brief review see Opie, 1923). Using whole serum or egg-white, Opie found that while an antigen can persist in a normal animal for a long time, it is more readily removed when introduced into an immune animal. The pres- ence of antigen coincident with antibody shown in precipitin tests was associated with the use of complex antigenic mixtures such as whole serum. Apparently, differences in the relative amounts of antibodies formed to various antigens in such a mixture were responsible for these observations. When a repeatedly crystallized preparation of egg albumin was used as antigen its presence in the blood of a rabbit immunized (titer 1:100,000) against this substance could not be demonstrated. It caused temporarily the complete disappearance of precipitin, and though it may have appeared in the blood, in no in- stance was antigen and its precipitin simultaneously demonstrable. Apparently, the introduction of an antigen into a well immunized animal brings about a combination between antigen and antibody whereby the latter is temporarily removed from circulation. Its sub- sequent reappearance may either be due to the elimination of antigen or to the synthesis of new amounts of antibody. These observations corroborate oft reported findings that with the progress of immunization the demonstration of antigen in the circula- tion becomes increasingly difficult. Wolfe and Dilks (1946) obtained immune sera from seven chickens, four immunized against goat and three against horse sera. Each chicken was bled 61 or 62 days after the last injection and immediately reinjected with the antigen previ- ously used. Large decreases in antibody concentration were noted at two to five hours after the reinjection which was the earliest interval tried. For example, a chicken which had produced an antigoat serum precipitating at 1 : 102,400 dilution of antigen showed no precipitin in the serum obtained by bleeding four hours after the reinjection of the homologous antigen, and showed a titer of 200 after 22 hours. A 66 IMMUNO-CATALYSIS chicken which had produced antihorse serum precipitating with 1 : 102,400 dilution of antigen, showed no precipitin when bled five hours after the reinjection of homologous horse serum antigen, but the titer had risen to 409,600 when bled six days after the reinjection of antigen. The other chickens showed similar behavior in every respect. Using quantitative chemical technique and horse serum and crystal- line egg albumin as antigens, Culbertson (1935) in a similar study made the following observations : After its injection into the circulation the horse serum as antigen persisted much longer in the blood of normal rabbits than of immune rabbits. In the normal animals, injected for the first time, complete removal of the antigen was never effected before the appearance of the precipitins in the blood. All of the crystalline egg albumin introduced combines with circulat- ing precipitin, which is immediately available for union with foreign protein. When the antigen is given in slight excess, a small residue of circulating precipitin generally remains in the blood. When large excess of antigen is given, antigen persists free in the circulation for about as long as an injection of similar size persists in the blood of a normal animal. The fixed tissue antibody is available in much less amount, and functions only when some of the antigen escapes union with the circulating precipitin and reaches the fixed tissues. The above relationship between the circulating precipitin and homologous antigen is not altered by the injection of numerous non- specific substances into the circulation. Hektoen and Welker (1935) also observed that the reduction or complete disappearance of antibody from the blood after introduction of the specific antigen, is sharply specific and that in the rabbit im- munized against many antigens the injection of one of the antigens as the rule resulted in the disappearance from the blood of the precipitin for that antigen only. Lewis (1912) reported that diphtheria horse antitoxin, if injected into rabbits which had been previously injected with horse serum, was considerably less effective in neutralizing diph- theria toxin subsequently injected than in normal rabbits. Romer and Viereck (1914) reported that, in guinea pigs sensitized against horse serum, antitoxin injected into the blood disappeared more quickly than it did in normal animals. According to Hooker and Follensby (1931) MECHANISM OF ANTIBODY FORMATION 67 in human subjects markedly hypersensitive to horse serum, scarlet fever antitoxin when administered locally, loses its neutralizing effect for specific toxin more quickly than it does in non-sensitive persons. The above cited observations shov^ that in the presence of an insufficient amount of antibody production the antigen persists in the host and continues to exercise an effect until completely destroyed by the host enzymes. When an adequate amount of antibody is produced it may completely block the activity of an antigen. In an ordinary immunization experiment these conditions might correspond to the stage v^'hen the peak of antibody titer is reached, which then is followed by an abrupt decline of antibody production for the above stated reasons. B. THEORIES ON THE SITE OF ANTIBODY FORMATION 1. The Lymphocytic Theory For over 40 years statements bearing on the lymphocytes as the site of antibody formation have been made by various investigators. Pfeiffer and Mark (1898) reported that anti-cholera antibodies ap- peared in the spleen, bone marrow, and lymph nodes before any in- crease could be detected in the blood. Anti-rabbit red cell hemolysins were reported to be present in serum, lymph from the thoracic duct, neck lymph, lymph from the limbs, salivary glands, etc. (Hughes and Carlson, 1908). Intravenous injection of antisera was followed by con- centrations of antibodies in thoracic duct lymph greater than in cervical lymph (Becht and Luckhardt, 1916). Takahashi (1933) reported finding in the lymph of the rabbit anti-human red cell ag- glutinins flowing from lymph nodes, and believed that agglutinins were manufactured there. (For further detailed discussion see Drinker and Yoffey, 1941; Dougherty and White, 1947.) McMaster and Hudock (1935) summarizing numerous reasons, concluded that agglutinins are formed in the lymph nodes; and McMaster and Kidd (1937) observed the antiviral principle in the regional lymph nodes of rabbits which were injected with vaccinia in the ears. Intensive studies on this question have since been carried out by Ehrich and Harris at the University of Pennsylvania, Dougherty, White and Chase at Yale University and Burnet and his co-workers in 68 IMMUNO-CATALYSIS Australia. Ehrich and Harris (1942) injected typhoid bacilli or washed sheep erythrocytes subcutaneously in the hind legs of rabbits and determined the antibody titers in the afferent lymph, in the sub- stance of the lymph node, in the efferent lymph, and in the serum. In a later study, Harris et al. (1945) separated the lymphocytes from the lymph plasma, prepared a cell extract and compared its con- tent of antibody with that of the surrounding fluid. Ehrich and Harris (1945), after a review of the observations pertaining to the reticulo- endothelial theory, summarized their results in the following manner. "When antigens were injected into the pad of the hind foot of the rabbit, antibodies first appeared two to four days after the injection in the lymph draining from the popliteal lymph node (the only node regional to the site of injection). They reached their highest titer after six days. In all experiments it was found that the antibody titer was higher in the efferent lymph; in some cases the concentration was 100 times that found in the lymph of afferent vessel. The produc- tion of antibody was preceded and accompanied by a rise in the output of lymphocytes in the efferent lymph which ranged from 15,000 to 20,000 per c.mm. to 60,000 to 80,000 per c.mm. or more. At the same time hyperplasia of the lymphatic tissue within the node occurred resulting in some experiments in a weight increase of the node from 0.2 g. to 1.0 g. or more. During antibody formation in the popliteal lymph node of rabbits, the lymphocytes in the efferent lymph vessels contain antibodies in a much higher concentration than the surround- ing lymph. The ratio of titers amounted to from eight to 16 in many instances." Commenting on the lack of phagocytic activity of lympho- cytes in relation to the antibody formation, and interpreting differently the results of Sabin (see page 72), they reason that the lymphocyte goes into action only after the raw material, i.e., bacteria or other formed antigens, have been properly prepared by the action of micro- or macrophages. In conclusion, they are of the opinion that the poly- morphonuclear leucocytes and the macrophage as well as the lympho- cyte may be instrumental in the antibody production, or through the cooperation of all these elements the immune bodies may be produced. In a series of studies, Dougherty, Chase and White (1944), and White & Dougherty (1946), immunized mice against washed sheep erythrocytes. They found that the antibody titers in the extracts from three times washed lymphoid cells from immunized mice were ap- MECHANISM OF ANTIBODY FORMATION 69 proximately eight-fold higher than those in the corresponding sera on the basis of nitrogen contents. The absence of antibody titer in the final washings of the lymphoid cells showed that the antibody titer in the extracts was derived from the cells and not from adherent lymph. Salivary gland or muscle tissue, containing reticulum cells, macro- phages, or fibroblasts, from the same immunized mice which had yielded antibody-containing lymphoid cells, showed no extractable ag- glutinins or hemolysins. Also lymphoid cell extracts from non- immunized mice were negative when tested for antibodies. Though these results seem to offer support to the theory that lymph nodes or lymphocytes might be the sites of antibody formation, these investiga- tors stated that sites of antibody production and concentration other than lymph nodes may exist, e.g., bone marrow, spleen, liver, and other organs containing high proportions of reticulo-endothelial cells. These, however, were not examined. On the basis of the above results, though overwhelmingly favorable for the lymphatic system as the site for production of antibody, they were not inclined to conclude that lymphocytes necessarily are concerned with antibody formation. Another study in this direction was reported by Kass (1945). Start- ing with a premise of questionable validity that if antibody is synthe- sized in lymphocytes, normal serum gamma-globulin should also be present, he prepared rabbit antisera to highly purified (electro- phoretically 98 to 100 per cent pure) human gamma-globulin. These sera reacted specifically with extracts from human mesenteric lymph nodes obtained within one hour after death of a patient. Extracts of washed slices of human liver failed to react with the antiserum. This finding was interpreted to show the presence of gamma-globulin in lymphocytes and thereby the synthesis of antibody gamma-globulin within the same cell was assumed. A few years previous to the publications of Ehrich and Harris, Harris, et al., and Kass (cited above), Burnet, et al. (1938, 1941) reported the results of comparative studies pertaining to this question. Burnet and Lush (1938), immunizing mice with a virulent strain of influenza virus, found that antibody production occurred in the lymph node. Studying the spread of herpes virus and the formation of antibody in rabbits, they found that two out of five rabbits showed significant amounts of antibody in the lymph node at six and seven days after injection, while the results with another rabbit were weakly positive. 70 IMMUNO-CATALYSIS At this period there was no significant antibody in the serum, ahhough by ten days considerable amounts were detectable. Similar results were obtained in experiments with dysentery agglutinogen and phage. However, when staphylococcal toxoid was similarly studied, as the most convenient representative of these antigens, no clear evidence of antibody production in the local lymph node could be detected. Since the primary antitoxic response was of such small extent, Burnet, et at. investigated the secondary response (specific anamnestic response) to the staphylococcal toxoid to detect antibody production in the lymph nodes. The toxoid was first injected intravenously; nineteen days later it was given subcutaneously in the right foot. A sharp rise in antitoxic titer of the serum had commenced on the 22nd day, and, believing that the secondary production of antitoxin was at its height, the rabbits were killed on the 24th day. The lymph nodes were re- moved, cleared of fat, and after weighing, they were emulsified (by grinding with quartz powder). After centrifuging the emulsion the filtered supernatant fluid was titrated for its antibody content. Ex- pressing the values per gram of tissue they found that there was a con- siderable enlargement of the lymph node on the inoculated side, but there was only a barely significant difference in the antitoxic titers. Four days after the subcutaneous injections a low titer of antitoxin in the serum of rabbits was obtained, and there was no antitoxin in the left node, and only the smallest detectable trace in the right. In view of these and similar results they stated: "The general conclusion seems justified that antibody is produced by those phagocytic cells of the reticulo-endothelial system which ingest the antigen; which group is actually concerned in any particular instance will be determined by the site of infection of inoculation, and the subsequent spread of the antigen in the body." a. The Hormonal Control of Anamnestic Response and Lympho- cytes. It has frequently been observed that the amount of antibody in the circulatory system increases following the injection of a non-specific protein, or other non-antigenic substances. A reference has already been made (p. 33) to an observation that the administration of a non-specific substance, such as the plant alkaloid pilocarpin, stimulates the restoration of antibody production. This reaction has been known as the anamnestic reaction. This, as evidence for the increased produc- tion of antibody, has, however, not generally been accepted (Land- MECHANISM OF ANTIBODY FORMATION 71 steiner, 1945). In a series of studies, White and associates (White and Dougherty, 1944; Dougherty, White and Chase, 1944, 1947; Dough- erty, Chase and White, 1945; Chase, White and Dougherty, 1946; White and Dougherty, 1946) recently reported that the increase in the antibody content of the blood stream subsequent to the administra- tion of anamnestic agents is not a new antibody formation but the re- lease of antibody concentrated in the lymphocytes. The injection of pituitary adrenotrophic hormone, or adrenal cortical extract caused changes in the lymphocyte tissue structure, decrease in the lympho- cytes in the lymphatic tissue, increase in serum protein concentration and antibody titers. A chronic adrenotrophic hormone treatment diminished lymphoid tissues, produced a consistent lymphopenia and increased the serum protein levels. These changes did not occur in adrenalectomized mice injected with adrenotrophic hormone, though the lymphocytes of these same mice contained appreciable quantities of antibody. These investigators pointed out that this hormonal control of an important lymphocyte function integrates the role of lymphocyte and adrenal cortex in maintaining certain normal protective mech- anisms of the organism. Data show that one of the augmented serum protein fractions is that containing antibodies. These relationships are said to be normally under pituitary adrenotrophic hormone influence and may be altered by a variety of stimuli aff^ecting pituitary-adrenal cortical activity, e.g., non-specific protein injections, hemorrhage, toxic chemicals, etc. Some of these have been demonstrated to produce pituitary-adrenal cortical activation resulting in lymphoid tissue dissolution, lymphopenia, and an increase in total serum proteins. Of these stimuli benzene and arsenite were reported to be very effective in producing the above men- tioned alterations in normal mice but were totally ineffective in adrenalectomized or hypophysectomized mice. (See also Eisen et al, 1947.) The anamnestic response occurring within three to nine hours in rabbits and mice following a single injection of either aqueous adrenal cortical extract, adrenal cortical steroid in oil, or adrenotrophic hormone showed both lymphocyte dissolution and globulin contribution to the serum. In this manner, it was shown that antibodies are an integral part of the lymphocytes, though not considered as evidence that lymphocytes are necessarily concerned with antibody formation. 72 IMMUNO-CATALYSIS b. Findings Contrary to the Theory of the Hormonal Control of the Release and Fabrication of Antibody. In contradiction to the findings discussed above, Eisen et al. (1947) reported that adrenal cortical activity is not essential for the fabrication or release of anti- bodies and gamma-globulin. These investigators found identical concentrations of serum antibodies and gamma-globulin in adrenalecto- mized rats repeatedly injected during immunization with adrenal cortical extract and similar animals not receiving this extract. Stoerk, et al. (1947) studied the effect of adrenal cortical hormones on the turnover of the serum proteins in adrenalectomized rats and likewise found that the administration of adrenal cortical lipo-extract did not exert any effect. 2. The Reticulo-Endothelial Theory a. The Theory of Sabin. In a study to determine the true site and the mechanism of antibody formation, Sabin followed the course and the fate of R-salt-azo-benzidine-azo-egg albumin in the animal, and traced its distribution and disappearance in various body cells. The dye-protein in the form of purplish red-alum-precipitated particles was introduced into rabbits through intraperitoneal, intravenous, intradermal, and subcutaneous routes. After the intravenous injec- tions the dye-protein was located in the Kupffer cells of the liver, in the macrophages of the milk spots of the omentum and in the cor- responding cells of the peritoneal walls, as well as in the endothelium lining the lymphatic sinuses of the retrosternal nodes and in the free macrophages of sinuses and follicles of these nodes. After intradermal and subcutaneous injections it was in local macrophages and in the regional lymph nodes. The dye protein was found in the vacuoles of digestion of the phagocytic cell and altered first by the removal of the dye; thereafter the solid particles of protein disappeared and it was assumed to have been rendered soluble and passed into the cyto- plasm. The vacuoles are considered the cellular organs of digestion, and the cytoplasm the zone of synthesis. Coincident with the time when the dye-protein is no longer visible within these cells, and when there are antibodies in the serum, there is observed a marked ac- celeration (over that of the normal rate) of the shedding of the surface films of the macrophages without damage to them. With the shedding MECHANISM OF ANTIBODY FORMATION 73 of parts of the surface films of these cells, both normal globulin and antibody globulin are carried out into the blood plasma. Thus it is stated that these mononuclear cells, functioning first as macrophages ("big eater"), render the antigen into suitable soluble form within the vacuole, and then as clasmatocytes ("shedding of exoplasm") in- corporate it into the cytoplasm, and there in some way the increase in the synthesis of normal globulin and the modification of some of it into antibody globulin take place. Coupling highly indiffusible blue dye T-1824, Evans blue, with various serum proteins and egg albumin, Kruse and McMaster (1949) prepared intensely blue dye-azo-proteins. A similar antigen consisting of extremely diffusible dye, echt-saure-blau, and of bovine y-globulin was also prepared. On intraperitoneal or intravenous injections, the azoproteins ap- peared in the Kupffer cells of the liver and sinus and reticular cells in lymph nodes, especially the great mesenteric node. These cells were particularly active in the removal of the blue antigens from the blood, but many other reticulo-endothelial cells were found to be active to a lesser degree. The storage of the antigenic material was observed in the cytoplasm only; it was not observed within nuclei, nor was it observed within the brain cells. Blue color was seen in the reticulo- endothelial cells of mice as long as 3Vi months after injecting echt- saure-blau azoglobulin. The seizure of the blue azoprotein by reticulo- endothelial cells almost everywhere in the body was interpreted to indicate that antigenic stimuls to antibody formation can be brought to bear from practically all parts of the body upon those tissues or cells that are capable of antibody formation. b. Plasma Cells as Antibody Producers. The hypothesis that plasma cells might be associated with process of immunization has been advanced by many authors on the basis of pathological-anatomical observations. In sternal bone-marrow punctures, a proliferation of plasma cells in lesions associated with hyperglobulinemia (infections, agranulocytosis, serum sickness and other morbid conditions) has been considered as affording clinical evidence in support of the hypothesis that plasma cells function as antibody producers. Plasma. Cells QPlasmocyte, Phagocytes^. The plasma cell is described to be a cell rich in protoplasm, with eccentrically placed nucleus, relatively small, round or oval, with five to eight bands of chromatin 74 IMMUNO-CATALYSIS extending from the center like the spokes of a wheel; around the nucleus is a lighter zone, whilst the abundant protoplasm otherwise is dark, basophile. It has been claimed that there is a smooth transition between the lymphocytes and plasma cells. Others hold that plasma cells originate from cells belonging to the reticuloendothelial system (Bing, 1940). Markoff (1937) stated that plasma cells of sternal bone marrow are identical with plasma-cellular reticuloendothelial cells. According to him, reticulum cells of sternal marrow functionally can be divided into two groups: one, plasma-cell-like; the others, phagocytic and lymphoid forms. Kolouch (1938) reported that there is an increase in the plasma cell content of the bone marrow running parallel with a rise in the anti- body titer against StrC'ptococcus viridans. According to Bing (1940) and Bing and Plum (1937) various diseases are accompanied by hyperglobulinemia, an increase of plasma cells and reticulo-endothelial cells in or outside of the bone marrow, which makes it probable that the formation of globulin takes place in these cells. Bj0rneboe and Gormsen (1941) immunized rabbits with various types of pneumococci killed with formalin. Intravenous injections were made every other day for a month, each injection containing 500 millions of pneumococci. Six of the immune sera contained 26 to 75 per cent more serum protein than the normal sera, or the globulin concentrations of immune rabbits were 6.1 to 12.9 mg./ml. of serum in comparison to 3.6 mg./ml. of normal serum. Histological examinations revealed a marked increase in plasma cells in spleen, in the capillaries of liver, plasma cell infiltration in the kidney and fatty tissues, slight plasma cell increase in the bone marrow. There was no increase in the reticulo-endothelial cells in these organs. They concluded that increase in globulin concentration of serum parallels the increase in plasma cells, or plasma cells are responsible for the synthesis of globulin. In an extensive later study, Bj0rneboe and Gormsen (1943), em- ploying intensive intravenous immunization technique with several dif- ferent antigens (polyvalent pneumococcal vaccine, polyvalent Salmo- nella vaccine and a mixture of proteins, etc.) simultaneouslv found the antibody concentration amounting to about 70 per cent of the total protein concentration of immune sera. Parallel with the rise of anti- body globulin they observed an increase in plasma cells and young plasma cells with scanty protoplasm and without the characteristic MECHANISM OF ANTIBODY FORMATION 75 "spoke" arrangement of the chromatin of the nucleus, or merely a suggestion of this feature. Many of these cells correspond to the concept of "plasma cellular reticulum cells" which are taken to be the progenitors of plasma cells. Based on such careful analysis of various factors, they concluded that there is constant direct proportionality between the concentration of antibody formed and the degree of plasma cell proliferation in the organism which has been immunized. Weighing various facts, one against the other, their reasonable ex- planation is that plasma cell proliferation essentially originates from elements of the reticulo-endothelial system, and as this system con- stitutes quite a considerable part of the spleen, the splenic weight is naturally increased by the proliferation in the reticulo-endothelial system. In other organs, for instance, the liver, the relative weight of the reticulo-endothelial system is too slight to influence noticeably the weight of the organ during the process of immunization. Studying the cellular reaction in the rabbit spleen during the secondary response, after intravenous reinjections of antigens, espe- cially living bacteria, Fagraeus (1948) observed a very strong plasma cellular reaction which was confined almost exclusively to the red pulp. The antigen (^Salmonella typhi) was found in the periphery of the lymph follicles and in the red pulp, while but little antigen could be distinguished in the follicles. The great development of the plasma cells in the red pulp of the spleen after the injection of the antigen was thus preceded by the concentration of bacteria there. Successive stages of the above-mentioned intense plasma cellular reaction, associated with antibody formation, was accompanied by the finding of large reticulum cells or transitional cells followed 1 to 2 days later by im- mature plasma cells and, a few days later, an increase in the number of plasma cells. The plasma cells were usually not detected in the lymph follicles. In tissue culture studies, suggestive evidence of the production of antibody by spleen tissue was obtained. The antibody thus produced was from 20 to 200-fold in excess of the amount of antibody he could obtain in tissue extracts made from pieces of tissue from the parent spleen. Excised pieces of red pulp and lymph follicles, abundant, respectively, in plasma cells and lymphocytes, were sepa- rately studied for their capacities to form antibody. The antibody production by red pulp cultures was considerably greater than that by lymph follicle cultures. The production of antibody by the former 76 IMMUNO-CATALYSIS system was considered responsible for a reasonable portion of tbe total amount of circulating antibodies. Histological studies showed that tissues containing transitional cells were comparatively poor anti- body-formers. The appearance of more mature cells was associated with an increase in the antibody content of the culture fluid. The decrease in the number of transitional cells was associated with an increase in antibody content of the serum, and the latter distinctly coincided with the number of immature cells showing growth. The transition of the immature plasma cells into mature cells was observed to lead to a decline in this capacity. According to Caspersson (1947), all protein synthesis needs the presence of nucleic acids; quantitatively, the most important nucleic acids in the chromosome are of desoxyribose* nature. The nucleus itself is a cell organelle organized especially for being the main center of the cell for the formation of proteins. In myeloma relatively large amounts of plasma cells appear in the blood. They show the typical picture of intense protein production (Thorell and Wising, 1944; Thorell and Wilton, 1945). Bing, Fagraeus and Thorell (1945) re- ported that a great amount of ribose nucleotides is found in the cytoplasm of the immature plasma cells. The nucleotide content to- gether with the well developed nucleolar apparatus was considered to indicate an intensive production of protein in the cell. They consider the appearance of plasma cells as a sign that a conversion and an *In a recent study Ehrich et al. (1949) observed that there is a parallehsm between the increase in desoxyribosenucleic acid and the increase in the weight of lymph nodes involving active cell multiplication. The peak of pentosenucleic acid increase was found to occur between the 4th and 6th day after vaccine injection when anti- body formation was at its maximum. A histological study showed that on the 5th and 6th day mature plasma cells were the prevailing cells. Most of the pentosenucleic acid was contained in the plasma cells. A comparison of the rates of the proliferation of lymphocytes and plasma cells with the pentosenucleic acid and antibody formation showed that the plasma cells and not the lymphocytes are responsible for antibody formation. A somewhat different point of view regarding the seat of the antibody synthesis is maintained by Harris and Harris (1949). After injecting antigenic and non- antigenic materials into the foot-pads of rabbits, the draining (pophteal) lymph nodes were daily studied histologically, chemically, and serologically. Non-antigenic materials showed no change. In response to antigenic materials the concentration of pentosenucleic acid rose to more than twice its normal value by the 2nd to 5 th day followed by a decline. The peak of this change occurred at or slightly before the appearance of the maximal concentration of antibodies in the same node. The seat of these changes are believed to be the large, young lymphocytes, a view which differs from that which considers the plasma cells as the seats of antibody synthesis. MECHANISM OF ANTIBODY FORMATION 77 intensification of the protein production has occurred in reticulo- endothehal cells, incited by the great antigen doses injected. From a functional point of view, this development culminates in the formation of the immature plasma cells. Maturing of the latter into plasma cells marks the transition to a less active state, and, therefore, the mature plasma cell represents the final link in a chain of development, a cell which has already passed the stage of its great functional intensity. Fagraeus (1948) concluded that "reticulo-endothelial elements under the conditions described, produce antibodies, thereby develofing into a type of cell xvith the morphological characteristics of the plasma cell." She found no evidence which would directly favor the participation of the lymphocytes in the formation of antibodies.* C. THEORIES ON THE MECHANISM OF ANTIBODY FORMATION In Part I, experimental data were cited to indicate that the role of antigen in the production of antibody might be one of catalysis. It was further stated that naturally occurring iso-antibodies, agglutinins against bacteria, and red blood cells, precipitins against animal and plant proteins and toxins appeared to be serum globulins. These would emphasize not only the fact that the normal serum and antibody globulins are closely related but also the fact that the site of their forma- tion might be identical. At the present time, we are handicapped by the lack of precise *Fagraeus (1947) found that the thymus, where the chief production of lympho- cytes takes place (Andreasen and Ottesen, 1945), has an insignificant antigen- phagocyting capacity, and lacks entirely the capacity to form antibodies in vitro. Harris, Rhoads and Stokes (1948) also studied the function of thymus in relation to the formation or retention of antibodies. They likewise found that the thymus did not play a demonstrable role in the formation of antibodies by the young rabbit, or in the retention of antibodies derived by the fetus from the maternal circulation. The formation of antibodies by spleen was also comparatively studied. They found slight evidence of formation of antibodies in the spleen following subcutaneous injections, but high antibody- titers were demonstrated in the spleen extracts following intravenous injection of antigens. This difference was explained by assuming that "the subcutaneously injected rabbits may well have formed the bulk of their anti- body to the Shigella in the lymph nodes draining the site of injection, the spleen being stimulated only by the fraction of soluble antigen-specific material which reached it via the blood stream. In the intravenously injected animals, on the other hand, the spleen could have received a major portion of the injected antigen and would thus have played a relatively greater role in the total antibody response of the animal." 78 IMMUNO-CATALYSIS knowledge as to the finer structure of proteins to formulate a com- prehensive mechanism of antibody formation for correlating its chemical and physical properties with that of the specificity of sero- logical reactions. In this connection it might not be irrelevant to re- view some of the properties of proteins pertinent to our question. 1. An Introductory Comment on the Antigenic Specificity of Proteins We know that the proteins differ in a general way in their physico- chemical properties— solubility, iso-electric point, electric charge, molecular weight and shape, etc. We do not know very much about the spatial configuration, but we know that one protein might diflfer from others in the arrangement and content of amino acids. A protein might be simple or conjugated with nucleic acid, heme, coenzyme, carbohy- drate, lipoid, etc., but none of this information has as yet clarified the subject. It has further been known that all proteins are con- structed of a large number of a-amino acids which are combined through peptide bonds, thus forming a long chain containing many CO-NH groups or peptide bonds. The proteins contain (in the ab- sence of c-amino acid) practically no free amino or a-carboxyl groups. Consisting solely of 1-amino acids with the exception of glycine, which has no asymmetric carbon atom and thus no optical activity, the proteins have uniform optical configuration. Specificity of proteases is singularly limited to the hydrolysis or synthesis of molecules of 1-configuration. These enzymes have no effect on the d-amino acids. For example, glycine (glycocoll or a-amino-acetic acid, NH2CH2- COOH) and alanine (a-amino-propionic acid, CH3CHNH0COOH) can form four different dipeptides: (1) glycylglycine, (2) alanyla- lanine, (3) glycylalanine, (4) alanylglycine. It is significant that peptides 3 and 4, although giving the same amino acids on hydrolysis, are not the same. This gives some clue to the possibilities of com- binations of the amino acids in proteins. It can be seen that when the number of amino acids available for such head-tail amide formation is increased, the number of combinations due to the order in which they are linked increases very rapidly. Emil Fischer (1923) starting with glycylglycine synthesized an octadecapeptide with a molecular weight of 1213. He calculated that out of 30 amino acids, 18 of which MECHANISM OF ANTIBODY FORMATION 79 are dissimilar, 1.28X10"'^, and out of 20 naturally known amino acids 2.4X10^^ protein isomers could be formed. It must be under- stood that such a polypeptide, as well as any protein, exists, or can exist in numerous tautomeric forms, each of which will have properties different from the other. In proteins the amino acids are bound together by peptide bonds through a-amino and a-carboxyl groups and the rest of the amino acid is essentially free. The non-peptide forming parts of the amino acids constitute side groupings in a protein molecule. The amino acids are listed below (Table I); according to their side groupings they con- tribute to the structure of the protein molecule. As will be seen, they are acidic, basic, polar and hydrophobic. These characteristics are related to free carboxyl, amino, amide, guanidine, phenolic, aliphatic hydroxyl, indole imino, imidazole NH, methyl sulfide, disulfide groups, and phenyl and aliphatic hydrophobic hydrocarbon radicals. Studies with enzymes, hormones, viruses (simple), and antigens deal with specific chemical reactions with certain of these side groupings in relation to their reversible and irreversible inactivation. Despite ex- tensive investigation the nature of enzymatic and antigenic specificity of the proteins remains the challenging unsolved problem. It is true, for example, that the oxidation of SH groups, or the reduction of -S-S- groups destroys the activity of certain proteins, but this fact alone does not explain the specific activity of various enzymes to which the same reactive SH or -S-S- groups are attached. There is no doubt that these specificities are governed by the whole protein molecule or parts thereof larger than the SH or similar groups. The integrity of every constituent part of the protein molecule functioning as virus, enzyme, hormone or antigen is essential for full biological activity. A scission in the molecule usually causes complete loss of activity. In this respect, the ability of an antigen to initiate the production of a specific antibody is comparable to those of enzymes and hormones. Splitting of a protein of minimal molecular unit size causes loss of the ability to stimulate antibody formation. This property of antigen being catalytic in character differs from its ability to combine with homologous antibody. In the latter case the reaction is stoichiometrical, is independent of molecular size and, therefore, of the integrity of the whole antigenic molecule. As is well known, the split products, as haptens, enter into stoichiometrical combination with 80 IMMUNO-CATALYSIS Table I AMINO ACIDS WHICH YIELD BASIC, ACIDIC, POLAR AND HYDROPHOBIC SIDE GROUPS IN THE PROTEIN MOLECULES Name of amino acid Structural formula Amino Acids With Basic Side Groups NH2 lC + )-Lysine(a-e-diamina-caproic HOOC-C-(CH.)4-NH. acid} I H NH2 H l(+)-ArginineC8-guanido-a-amino- hOOC-C-(CH.)3-N-C = NH valeric acid J 1 1 H NH2 NH2 I ir A u- .-J- ro ■ -A } ■ HOGG— C— CH2— C===CH IQ— J-Histidine C/3-imidazol-a-amino- , 1 1 propionic acid) . n ^ Nh GH Amino Acids With Acidic Side Groups NH2 lC + )-Glutamic acid (a-amino-glu- HOOG-C-GH2-CH2-GOOH taric acid) 1 H NH2 1 C-)-Aspartic acid (a-amino-succinic HOOG-C-GH2-GOOH acid) I H NH2 I(-)-Tyrosine (^-p-hyroxy-phenyl ^^qq^— c_C— < )— OH o-amino-propionic acid) 1 1 H H NH2 1(— )-Cysteine C/3-thiol-a-amino-pro- hOOG— C— GH — S— H pionic acid) 1 H MECHANISM OF ANTIBODY FORMATION 81 1 (— )Cystine [di-C^-thiol-a-amino- propionic acid)] Amino Acids With Polar Side Groups NH2 I HOOC— C— CH2 S H NH2 I HOOC— C CH2 S I H NH2 1 C-)-Mcthionine (y-methylthiol- HOOC-C-CH2-CH2-S-CH3 a-amino-n-butyric acidj 1 H 1(— )-Serine (jg-hydroxy-a-amino- propionic acid) NH2 HOOC— C— CH2— OH I H NH2 H 1 (— )-Threonine (a-amino-jg-hydroxy- ttqqp n n OH n-butyric acid) 1 I H CH3 1(— )-Hydroxy-proline (4-hydroxy- pyrrolidine-2-carboxylic acid) CH2 CH— OH HOOC— CH CH2 \/ N H 1(— )-Tryptopbane (/3-3-indole- a-amino-propionic acid) NH2 HOOC— C— CH2— C- H HG /\ N H 82 IMMUNO-CATALYSIS Amino Acids With Hydrophobic Side Groups NH2 Glycine (glycocoll, amino-acetic hOOC— C H acid) H NH, l(-j-)-Alanine (o-amino-propionic HOOC C— CH acid) I H l(-|-)-Valine (a-amino-isovaleric acid) NH2 H HOOC— C C— CH3 I I H CH3 NHo H l(+)-Isoleucine (^-methyl-a-amino- jjoOC— C C— CH2CH3 valeric acid) | | H CH3 NH2 H lC-)-Leucine (a-amino-isocaproic jjOOC— C— CH2— C— CH3 acid) I I H CH3 NH2 Norleucine (a-amino-caproic hOOC— C-CH2CH2CH2CH3 acid) I H 1(— )-Phenylalanine (/3-phenyl- a-amino-propionic acid) NH2 I HOOC— C— CH2— < I H CH2 CH2 1 (— )-Proline (pyrrolidine-2-car- boxylic acid) HOOC— C CH, N I H MECHANISM OF ANTIBODY FORMATION 83 antibody. The specific combining capacities of antigens reside, there- fore, in the whole as well as in its component small molecular weight non-antigenic units. These units carry specific configuration and chemically reactive groups. On the basis of these established facts one may conclude that while the specific combining capacities of antigens, or their component units, are dependent on the amino acid composition and order of their linkage yielding specific configuration to each protein molecule, the antibody producing catalytic activity depends on the integrity of the whole molecule. The interdependence between catalytic activity of antigen and its minimum irreducible molecular unit size is a common and basic property of enzymes, hormones and viruses also. No doubt, the speci- ficities of the latter three are similarly dependent on the amino acid composition and order of their linkage yielding specific configuration to each species of protein molecule. These facts may perhaps more fully be appreciated if we consider the fact that the specific activities of the same heme group in various closely related enzymes, cytochromes, cytochrome oxidase, cytochrome peroxidase and catalase, are directed by specific protein carriers. Similarly, there exist distinct differences in antigenic specificity of closely related classes of proteins. One can cite as an example dif- ferences in this respect of a class of a-, (3- and y-globulins from the same species of animal. Kendall (1937, 1938) showed that these globulins of human serum are antigenically specific. The antibodies against these globulins were found to be specific for each of these globulins; that is, the a-globulin antiserum reacted with the a-globulin to give a precipi- tate, but did not react with fS- or y-globulin, with albumin, or with any other component of human serum. Thus, this highly refined im- munological differentiation of the closely related globulins of human serum, which differ from each other in amino acid composition (Table II), electrical mobility, molecular weight, viscosity and iso- electric point, makes the correlation between the structural pattern and the biological specificity of the proteins reasonably plausible. 84 IM MUNO-C ATALYSIS Table II Amino Acid Composition of Human Plasma Proteins g per 100 g. Protein Constituent Albumin y )8 a Fibrinogen Total N 15.95 16.03 15.24 16.9 Total S 1.96 1.02 1.32 1.26 Free a-amino N 0.18 0.11 Amide N 0.88 1.11 Glycine 1.6 4.2 5.6 3.1 5.6 Alanine Valine 7.7 9.7 7.0 5.2 4.4 Leucine 11.0 9.3 7.9 14.2 7.1 Isoleucine 1.7 2.7 5.0 1.7 4.8 Proline 5.1 8.1 7.1 4.7 5.7 Phenylalanine 7.8 4.6 4.7 4.6 4.2 Cysteine 0.7 0.7 1-i f-l 0.4 Half-Cystine 5.6 2.4 2.3 Methionine 13 1.1 L7 1.4 2.5 Tryptophan (0.2) 2.9 2.0 1.9 3.3 Arginine 6.2 4.8 6.8 7.7 7.9 Histidine 3.5 2.5 2.8 2.8 2.8 Lysine 12.3 8.1 6.6 8.9 8.3 Aspartic Acid 10.4 8.8 9.8 9.0 13.6 Glutamic Acid 17.4 11.8 14.5 21.6 14.3 Serine 3.7 11.4 7.1 5.0 9.2 Threonine 5.0 8.4 6.1 4.9 6.6 Tyrosine 4.7 6.8 6.0 4.5 5.8 Totals 105.9 108.3 104.2 102.7 108.8 John T. Edsall (1947): Advances in Protein Chemistry, Volume III. Edited by M. L. Anson and John T. Edsall. New York, Academic Press, 1947. p. 464. 2. The Role of Determinant Groups on the Antigenic Specificity of Conjugated Proteins Although all attempts with natural proteins have failed to obtain an answer to why one protein is antigenically different from the other, MECHANISM OF ANTIBODY FORMATION 85 Studies in another direction by Landsteiner and his followers have shed considerable light on the role of the "determinant groups" in antigens on the specificity of antibodies. There is though no evidence that the natural simple proteins contain characteristic "determinant groups" to account for the specificity of homologous antibodies; however such groups introduced into the artificial proteins have been shown, by numerous examples, to exercise the power of determining the specificity of antibodies against them. Of the considerable number of such examples, which will be taken up in a later chapter on the similarity of the specificities of enzymes and antigens, we will consider at this point only two of the most out- standing examples. It is to be noted that these examples have more or less served as a basis for the theories on the mechanism of antibody formation. Landsteiner and van der Scheer (1928, 1929) demon- strated that d- and l-^^ara-amino-tartranilic acids coupled with proteins produce antibodies specifically reactive. Avery and Goebel (1929) similarly showed that p-amino-phenol-/3-glucoside and p-amino-phenol- /?-galactoside coupled with a protein produce respectively specific anti- bodies. In these examples we have the differences between the "de- terminant groups" reduced to the simplest structural forms known to chemistry. In tartranilic acid antigens the apparent difference is one of optical or space isomerism. In p-amino-phenol-/?-hexoses likewise the difference is one of space isomerism. These differences are associated also with other chemical properties, such as differences in the reactivity to specific enzymes, solubilities, and also physical and chemical dif- ferences of their respective derivatives. How are we then to interpret the specificity of antibodies produced against them? Does it signify that antibody against d-antigen possesses 1-configuration and vice versa? Or are we to assume that antibody against one is the space isomer of the antibody against the other? Or else must we assume that chemical and physical differences resulting from or associated with the difference of optical or space isomerism are responsible for the dif- ferences in the specificities of antibodies against each and every one of them? A satisfactory answer to these questions has an important bearing on the formulation of the concepts of the structure of anti- bodies in relation to the nature and the position of the serologically reactive specific groups. At present our knowledge regarding these questions is not of sufficient scope to formulate the answer satisfactorily; 86 IMMUNO-CATALYSIS we must therefore confine our ideas for the time being to the realm of possibilities. With these limitations in mind, the theories advanced concerning the mechanism of antibody formation are discussed below. 3. Mechanism of Antibody Formation (a) Breinl and Haurowitz (1930), (b) Mudd (1932), (c) Alex- ander (1931), (d) Pauling (1940) and others have offered hypotheses on the possible mechanism of antibody formation. a. The Theory of Breinl and Haurowitz. These authors assume that the cells which synthesize serum globulin possess species specific surfaces containing certain groups with residual valences which attract or repulse amino acids participating in the formation of globulin. An antigen combining with these surfaces produces new groups of differ- ent residual valences. The orientation of amino acids in the formation of antibody globulin are thus governed by surface valences of the new cell-antigen complex. Possible evidence concerning the presence of surface valences are cited from the observations of Landsteiner that only those antigens containing groups with weak residual valences, such as -COOH, -NHo, -HSO3, etc. show antibody determining property whereas aliphatic (paraffin) chains, or plain aromatic rings, which lack such residual valences, are devoid of antibody forming property. b. The Theory of Mudd. Mudd bases his concept of the mechanism of antibody formation on the specific relationship between antigen and antibody. The striking instances cited by him are the d- and 1- tartranilic acid azo-proteins of Landsteiner and van der Scheer (1929) and the p-amino-phenol-i8-glucoside and p-amino-phenol-^-galactoside azoproteins of Avery and Goebel (1929). Injected into rabbits they give rise to antibodies which combine electively with the particular stereoisomer injected. Synthesis by the linking of amino acids to the peptide is supposed to occur in an orienting environment, namely the antigen-protoplasm interface. The chemical groupings linked to the molecule undergoing synthesis at the antigen surface are believed to be adapted spatially and in their chemical affinities to the antigen surface at the region in which linkage occurs. The synthesized anti- body molecule should therefore possess to some degree a stereochemical MECHANISM OF ANTIBODY FORMATION 87 correspondence with the antigen for the reason that each structural unit has been selected and oriented to fit the local configuration and affinities of the antigen surface. Thus, an antibody "specific" for the antigen, i.e., possessing specific stereochemical correspondence with it, is supposed to have been formed. c. The Theory of Alexander. According to Hektoen the minimum sensitizing dose of egg albumin is approximately 0.000,05 mg. It has also been frequently shown that if an animal be bled, the temporary drop in the titer of antibodies in the remaining blood is retrieved and even surpassed. In view of this great disproportionality between the antigen used and the antibody produced, Alexander points to a sim- ilarity between these facts and the role of enzymes or catalysts. He believes that the antigen forms within the (antibody forming) cells themselves, new specific catalysts which are able to direct the forma- tion of antibodies. In this connection three possibilities are considered : (a) modification of a gene, (b) modification of a non-genic catalyst, and (c) fixation of the antigen particle by a non-catalyst cytoplasmic particle in such a manner that the combination functions as a specific catalyst. Variations in the duration of immunity are assumed to cor- respond to variations in the persistence of the antigen-catalyst complex, while inability to establish immunity would indicate the non-formation of such a complex. As will be discussed on pages 104-8 of this treatise we do not sub- scribe to the idea of Alexander that an antigen produces a modification of a gene. Our view concerning this question basically differs from that of Alexander. d. The Theory of Pauling. In disagreement vdth Breinl and Hauro- witz, and Mudd, Pauling does not believe that the effect of an antigen in determining the structure of an antibody molecule is the ordering of the amino-acid residues in the polypeptide chains in a way different from that in the normal globulin. Hypothetically he assumes that all antibody molecules contain the same 'polype-ptide chains as normal globulin, and differ from, normal globulin only in the configuration of the chain; that is, in the way that the chain is coiled in the molecule. It is further stated that "the number of configurations accessible to the polypeptide chain is so great as to provide an explanation for the ability of an animal to form antibodies with considerable specificity for an apparently unlimited number of different antigens, without the 88 IMMUNO-CATALYSIS necessity of invoking also a variation in the amino-acid composition or amino-acid order." A globulin molecule is pictured as consisting of a single polypeptide chain, containing several hundred amino-acid residues. The arrange- ment of the amino-acid residues in the central part is much more stable than any other, whereas the two end parts of the chain are of such a nature that there exist for them many configurations with nearly the same energy. "The atoms and groups which form the sur- face of the antigen will attract certain complementary parts of the globulin chain (a negatively-charged group, for example, attracting a positively-charged group), and repel other parts. As a result of these interactions the configurations of the chain ends which are stable in the presence of the antigen will be such that there is attraction between the coiled globulin chain ends and the antigen, due to their com- plementarity in structure. The configuration assumed by the chain end may be any one of a large number, depending upon which part of the surface of the antigen happens to exert its influence on the chain end and how large a region of the surface happens to be covered by it." The middle part of the antibody molecule thus produced would be like that of a normal globulin molecule. Hence antibodies should have antigenic activity, with essentially complete cross reactions with normal globulin. The two ends, on the other hand, would have con- figurations more or less complementary to parts of the surface of the antigen. According to this picture the active end regions of the anti- body molecule would not have effective antigenic power.* This is explained by assuming that the configurations of the end regions would be different from molecule to molecule, and that an antibody complementary to one antibody end would as a rule not combine with another. *The concept of Pauling appears to be in contradiction to the results provided by the studies of Northrop (1942). He reported that diphtheria antitoxin (mol. wt. 184,000) on digestion with proteolytic enzymes was obtainable in crystalline form. This substance was 90 per cent precipitable with diphtheria toxin and had a molec- ular weight of 90,000. The immune sera prepared against the crystalline substance were active against antitoxin but inactive against the normal serum proteins. The fact that the crystalline substance represents one-half of the original antitoxin molecule, fails to support the hypothesis of Pauling that the antibody molecule is made up of a stable midpiece carrying at each end polypeptide chains susceptible to configurational changes. In view of Northrop's results it would seem difficult to label certain parts of the antibody molecule as corresponding to mid or end pieces. MECHANISM OF ANTIBODY FORMATION 89 4. The Concept of Catalysis as Implied by the Above Theories on the Mechanism of Antibody Formation The above discussed theories on the mechanism of antibody forma- tion imply concepts which support the view presented in this treatise that the antigens exercise the role of catalysts in the formation of specific antibodies. The following statements concerning this question are taken from the above discussed theories: "cells which synthesize serum globulin possess species specific surfaces," and "antigens com- bining with these surfaces produce . . . new antigen-cell complexes directing the synthesis of antibody globulins" (Breinl and Haurowitz); "antigen-protoplasm interface" serving as an orienting environment in the synthesis of antibodies (Mudd); "the antigen molecule, after its desertion by the newly-formed antibody molecule, may serve as the pattern for another" (Pauling), which is a function of all catalysts; "Macrophages render the antigen into suitable soluble form within the vacuole and then as clasmatocytes incorporate it into the cytoplasm, there in some way increasing the synthesis of normal globulin and the modification of it into antibody globulin . . ." (Sabin); "antigen forms within the (antibody forming) cells themselves, new specific catalysts which are able to direct the formation of antibodies" (Alexander), are concepts which support the thesis that the mechanism of immunization is similar to the mechanism of surface catalysis in heterogeneous systems. The similarity between the experimental facts serving as a basis for the above views, and those to be encountered quite extensively in chemical processes, where catalysis plays a determinant role, is quite clear. We know that normal and immune serum globulins are synthesized probably within reticulo-endothelial cells (Sabin). To account for the specific reactivity of the antibody globulin it is necessary to assume the formation of a new cell-antigen catalytic surface. Evidently the cell as such is acting as an anchor or a catalytic support for antigen to exercise its directive influence in modifying the synthesis of normal globulin into antibody globulin. These modifications of surfaces within cells may be pictured as analogous to the mechanism of heterogeneous catalytic surfaces. They remind us of the well-known "mixed" or "supported" catalysts, which represent mixtures of two or more sub- 90 IMMUNO-CATALYSIS Stances, capable of producing a greater specific catalytic effect than can be accounted for if they were to act independently. 5. The Theory of Burnet, et ah of Antibody Formation and Adaptive Enzyme Process A valuable contribution to immunology is the monograph, Pro- duction of Antibodies, by Burnet, et al. (1941; Burnet and Fenner, 1950). Among other topics, the chapter on the "Theoretical Aspects of Antibody Production" concerns us here most. Because of the war conditions this publication was unknown to us, hence no reference to this theory will be found in the 1st edition of Immuno-catalysis which went to press in the summer of 1943. Burnet, et al. advance a theory of antibody production which, in essence, is that antibodies are produced in a manner analogous to the "production" of so-called "adaptive enzymes." The theory postulates that antigen sfecifically ynodifies the 'proteinase producing a new enzyme with the ability to synthesize an antibody specifically reactive with the antigen. The basic tenets of this concept, as will be discussed below, conflict with the experimental basis of the existing theories dealing with the interrelationship of genes and enzymes, the changes a cell may undergo as a result of mutations, and the chemical activities of the cells as the seats of enzyme action, etc. Before we undertake an analysis of these factors it is necessary that the salient points which Burnet, et al. offer in support of their theory be presented. a. The Premises of the Theory of Burnet, et al. ( 1 ) "Antibody is composed of globulin molecules which are pro- duced by and liberated from, those cells of the reticulo- endothelial system, which ingest the antigenic m^olecules or particles. Particulate antigens introduced into the tissues are largely dealt with by cells of the lymph nodes, while, if they (the antigens) reach the blood stream, cells of the spleen, liver and bone marrow are chiefly concerned. There is still some doubt as to where bacterial antitoxins are produced. From Buttle's experiments (1934) the cells of the bone mar- row may represent the main source of antitoxin. (2) "A second or subsequent contact with the same antigen MECHANISM OF ANTIBODY FORMATION 91 'provokes a more active production of antibody. This is seen more clearly with toxoids than with particulate antigens, but when looked for can be observed with all types of antigens. The latent period is shortened, the antibody titer rises more rapidly and to a higher titer, and the rate of subsequent fall is slower, (3) "Antibody in the circulation is being constantly removed at a rate which is approximately proportional to its concentration. This is based largely on the data from passive immunity ex- periments with sera of the same species, but adequate reasons have been given for assuming that it holds also for actively produced antibody. (4) "Antibody production following an antigenic stim,ulus rises to a peak and then diminishes, but continues at a dim,inishing rate often for long periods. The classical example is the persistence of demonstrable yellow fever antibody more than fifty years after the last contact with the virus. (5) 'Antibody production can continue long after the antigen responsible has disappeared from, the body. This conclusion is perhaps debatable, but reasons have been given earlier for adopting it. The alternative, to suppose that the cell retains the antigen unmodified but undetectable by any experi- mental method, is against all biological analogies, creates grave difficulties in interpreting qualitative changes in anti- body with repeated immunization, and could only be adopted if no other interpretation could account for the facts. (6) 'Antibody production is a function not only of the cell originally stimulated, but of its descendants. The cells of the reticulo-endothelial system are well known to vary enor- mously in number and in response to physiological and pathological stimuli. The inoculation of antigenic and non- antigenic foreign material is one of the most potent methods of inducing their proliferation. Although direct proof is im- possible, it is a reasonable assumption from this lability in numbers that the life of any of these cells is a short one, and that when antibody production goes on for months or years other cells than those initially stimulated must be respon- sible." Rejecting the persistence of thorium dioxide in the 92 IMMUNO-CATALYSIS reticuloendothelial cells of liver and spleen for years as evidence for the persistence of the same individual cells, Burnet concluded : "we can see no escape from the conclusion that the antibody-producing mechanism can be transmitted to descendant cells by some hereditary process. (7) "The tjfe of antibody 'produced varies (a) according to the species used, (h) with the age of the animal, and (c) according to the nature and frequency of the antigenic stimuli. The change in character of the antibody following repeated re- inoculation is the difference of most theoretical importance. It would indicate that an antibody-producing mechanism once established can be further modified by new antigenic con- tact. Of the criteria offered by Burnet, et al. in support of their theory, criteria one to four can be explained by the observations considered in the preceding discussions in relation to the factors controlling the im- mune response to an antigen, and those related to the antigen-antibody balance during the process of immunization. The most challenging of these criteria are those condensed in paragraphs five and six. It is here assumed that antibody production continues long after the antigen has disappeared from the body. It is, therefore, concluded that antibody production is a characteristic which is acquired not only by the cell proteinases but also transmitted to the descendant cell proteinases. In other words, it is an acquired characteristic inheritable by generations of cells during the lifetime of the host. In support of these assertions these authors cite, in particular, the lifetime persistence of immunity against measles and yellow fever virus. In the case of yellow fever, based on an observation by Sawyer, they stated that "there is direct evidence that a single infection may induce the formation of antibody which can be detected in the serum 75 years later." Landsteiner (1945), in reference to the nature of long-lasting immunity, and certain claims that antigens leave impressions upon the antibody-producing cells which last after the disappearance of the antigen, made the following statement: "The fact that immunity can last for many years would be a decisive proof, but in the most striking case of virus infections (smallpox, measles, yellow fever), the perma- nence of active virus cannot be excluded." According to Rous (1946), MECHANISM OF ANTIBODY FORMATION 93 the viruses may lie latent permanently in plants and animals yet cause no discernible harm, e.g., the virus of King Edward potato plants, which cause them no perceptible trouble, though capable of killing plants of other varieties, or the virus from human "fever blisters" which causes encephalitis in an inoculated rabbit. The "latent" viruses stimu- late in the animal host the formation of specific protective antibodies. According to Dixon (1945), a lasting immunity to viruses demon- strable by complement fixation, agglutination, precipitation and neutralization following one attack in an animal, may depend on the kind of tissue infected and is probably due to a long-term sojourn or persistence throughout the life of the host. Lasting immunity may not be obtained to influenza or to the common cold virus because the superficial cells lining the respiratory tract are being thrown off at intervals to be replaced by deeper cells and thus do not provide a permanent abode for these viruses. It may further be stated that all studies so far carried out, and intended to determine the relation of the duration of antigen to antibody production in a host, show that the disappearance of antigens (some tagged with radioactive atoms) from host tissue results in the cessation of antibody production (Ehrich, etal 1945; Herdegen, et al 1947; Libby, 1947). According to Taliaferro (1932, 1938) when a trypanocidal crisis is permanently effective, it is because the few parasites which escape destruction cannot produce a relapse because they are prevented from reproducing by anti-reproduction (ablastin) antibody. According to Packchanian (1934) when the animal has recovered from reinfections, it is quite possible that all the trypanosomes (Trypanosoma hrucet) have been destroyed, at least in the blood. If a few latent ones do exist, they are so attenuated that they cannot produce any detectable in- fection in subinoculated test animals. Those reinoculated forest deer mice are considered relatively immune to further reinoculation with Tr. hrucei. Experimental data indicated that even in certain individuals of forest deer mouse (P. L. novehoracensis^ , prolonged and indefinite latency was associated with a few surviving trypanosomes. Thus, 100 days after the initial inoculation, the animal was etherized and almost its entire heart blood, 0.5 ml., was injected into a rat. The latter animal was observed for five months and failed to show symptoms of nagana. The heart, lungs, liver, spleen, kidneys, brain, muscles, and part of the bone marrow and glands of the same mouse were ground into pulp in 94 IMMUNO-CATALYSIS an equal volume of tyrode solution and introduced into the peritoneal cavity of rabbit by means of regular surgical technic. This rabbit later developed symptoms of nagana and trypanosomes were demonstrated in its blood. Two drops of ear blood from this animal were inoculated into a rat which contracted the disease and died after 8 days. Fifty- three days after the original inoculation, this rabbit was found dead. This particular experiment shows that at least a few trypanosomes were present in the tissues or organs of the forest deer mouse, an animal which had been relatively immune on the basis of negative microscopic examination of its blood, as well as on the basis of the noninfectivity of its blood to rats. In another set of studies Packchanian and Tom (1943) reported the presence of agglutinins for Leytosfira icterohaemorrhagiae in the circulating blood of patients for periods from one year to at least 20 years and seven months after recovery from Weil's disease. It is inter- esting that in all cases during the periods indicated there has been a de- cline in the agglutinin titer. However, the sera were never free from ag- glutinins of significant titer. Following up these findings Packchanian and Sonnier (1948) found that tests for agglutinins in the serum samples of 18 rats (R. norvegicus') were negative and no Leptos'pira were found in the kidneys of these rats. However, the serum samples of a wild rat, whose kidneys were positive for motile Le'ptos'pira, gave agglutination reaction with Type I Leptospira icterohaemorrhagiae in dilutions up to 1 : 30,000 and the serum sample from an infected mouse, whose kidneys also were positive for Leptospira, gave an agglutination reaction in dilutions up to 1 : 10,000. These investigators have shown that less virulent strains of Lepto- spira may fail to produce death in guinea pigs, but may produce anti- bodies, such as agglutinins and lysins as a result of subclinical infection. Burnet, et at. interpret the above claimed long-lasting immunity in the following manner. The fractions of the serum globulin which are concerned in immunological reactions are synthesized by cellular pro- teinase units. These liberate partial replicas of themselves. These proteinases, by virtue of their enzymatic function, come into contact with any foreign antigens taken into the cell, are lastingly modified by this contact. The modification of the proteinase unit which is produced by contact with the antigen is not to be regarded as resulting from a synthesis of a new unit in spatial contact with antigen, but rather as a MECHANISM OF ANTIBODY FORMATION 95 process analogous to the production of adaptive enzymes by bacteria. The stress is laid on the assumption that contact with a sugar molecule not normally fermented can impress on a bacterial enzyme a new specific power of acting on this new substrate. A similar speculation related to the "adaptive enzyme" concept is also advanced by Emerson (1945). He assumed that a disaccharide such as maltose can make a transmittable print on a gene which then conveys this to a protein in a manner complementary to that of the gene yielding a specific enzyme. The gene then constructed on this template should continue to produce the enzyme, even in the absence of maltose. This is another version of the "adaptive enzyme" concept which will be discussed below. b. Consideration of the Adaptive Enzyme Concept with Respect to Antibody Production. The basic weakness in the theory proposed by Burnet, et al. would seem to lie in the fact that it is patterned after a concept which in itself is based on inadequate experimentation and interpretation of the factors controlling the assumed production of, or the increment in, the so-called "adaptive enzymes." As discussed previously (Sevag, 1946), the analogy between the claimed mechanism of adaptive enzyme formation and that of antibody production does not appear to be valid. Burnet, et al. postulate that the antigen molecule modifies a cellular proteinase which then is capable of synthesizing an antibody molecule. That is, the proteinase adapts itself not to metabolize the antigen but to synthesize antibody, a basic difference from the assumed acquired ability of adaptive enzymes to metabolize the specific substrate. Another noteworthy difference between anti- body producing "adaptive proteinase" and "adaptive enzymes" is the duration of the former for the liftetime of the host and the inheritabil- ity of this character from parent to daughter host cells. In contrast, the lifetime of an adaptive enzyme is postulated to be limited. That is, it is formed when substrate is present and fails to function or form when the substrate in question is removed. "Adaptive enzyme" and "adaptive antibody" production concepts would have been more com- parable if it could have been demonstrated that once the antibody producing enzyme system is born in a host, it could adaptively function to produce more antibody, provided the same antibody is introduced from time to time into the system which has ceased to produce measur- able antibody. The assumed permanency of the antibody synthesizing 96 IMMUNO-CATALYSIS modified proteinase requires that this be a fact demonstrable in all cases. The proponents of the "adaptive enzyme" concept assume the existence of a freenzyme (Monod, 1942, 1944, 1945, 1947; Lwoff, 1946) with which a certain substrate combines and liberates the "adaptive enzyme" (discussed further below). On the basis of a com- plete analogy between Burnet's theory and the above theory of adaptive enzymes, it is to be noted that there does not seem to exist such a preenzyme, to adapt itself to synthesize antibody in contact with newly introduced antibody. The results of experiments have shown that the reinjection of antibody into a previously immunized host from which the antibody is derived, does not stimulate or initiate the production of more of the same antibody. c. Adaptive Enzymes — A Consideration of Facts and Assump- tions. In a previous communication (Sevag, 1946), enzymatic, chemi- cal and theoretical reasons were given to reject the concept that the "production of adaptive enzymes" involves the synthesis of a new enzyme protein. Let us summarize the reasons for the position we maintain. First: The pertinent data show that the so-called "adaptive enzymes" were present as integral parts of the cells prior to coming into contact with the substrates in question. In the presence of large amount of substrates, the necessary factors seem to be optimal for the maintenance or the demonstration of a measurable activity of the enzyme. The substitution of other structurally related substrates temporarily reduces the activity of the "adaptive enzyme" to a minimum, but does not abolish it. The reversal of the reduced activity to optimal activity with the return of the particular substrate to the metabolic environment has been claimed to be the production of "adaptive enzymes." The restora- tion of the activity to metabolize a substrate is also accomplished when the cells are grown either in the absence of that particular substrate, or also in the absence of a different substrate whose metabolism exercises a deleterious effect on the particular enzyme (Sevag and Swart, 1947). But under no conditions is the particular enzyme completely elimi- nated. Second: "Adaptive enzymes" are formed only with those substrates which are configurationally related to a substrate which is more actively metabolized than others. This indicates a "master-key enzyme" for a group and not a specific enzyme for each substrate. It also means MECHANISM OF ANTIBODY FORMATION 97 differences in the rates of the metabohsm of various structurally related substrates. There is no evidence as to the presence of individual enzymes for each of the class of substrates which possess the same enzyme specific active group. Thus, according to the formulation of Weidennagen (1940), "there is no special key for each lock but a master key for a group of locks." Glucose, mannose, fructose, etc.; arginine, creatine, etc. possessing, respectively, the same configuration, would be metabolized by the same enzyme. Several dipeptides com- posed of different amino acids are hydrolyzed by Anson's crystalline carboxypeptidase, etc. In this connection the concept of Monod (1943, 1944, 1945, 1947) is of interest. According to him, all the adaptive and constitutive carbohydrate-attacking enzymes in bacteria may depend on a common mechanism of synthesis, or, more precisely, on a common precursor. The mutual-exclusion effect results from a competitive interaction of the substrates for the 'preenzymes or the common precursor. This concept likewise agrees with our view that under the influence of a substrate the synthesis of a new specific enzyme protein does not occur. On the other hand, Spiegelman, et al. (1947) postulate that enzyme adaptation involves protein modification requiring energy-yielding metabolic reactions. Such assumptions cannot be accepted as valid un- til the postulated pre-protein and the enzyme derived therefrom are characterized, and the energy relationships of the chemical reactions leading from the former to the latter are established. The principal justification for the claim of the formation of "adaptive enzymes" seems therefore to lie in the fact that the enzyme activity for a certain substrate (for example, galactose)* disappears or is greatly diminished *On the basis of the results of recent studies we are now able to gain an insight into the mechanism of the fermentation of galactose by yeast. Various steps involved in galactose fermentation are described as follows: Galactose -f ATP >- galactose- 1 -phosphate + ADP I glucose— 1 —phosphate >- Galactose— 1 —phosphate y glucose— 6— phosphate II III According to Wilkinson (1949) the first step (reaction I) in galactose fermenta- tion by Dutch Top yeast is the transference of the terminal phosphate group from adenosine triphosphate (ATP) to galactose to give galactose- 1 -phosphate. (There was no evidence for the accumulation of galactose- 1 -phosphate during the pre- adaptive period.) The enzyme catalyzing this reaction is activated by Mg++ and by cysteine, and is named galactokinase. Since galactose is fermented via fructose- 1,6- diphosphate, the question of how the glactose-1 -phosphate is converted to glucose- 1-phosphate (reaction II) arises. According to Caputto et al. (1950), the reaction II 98 IMMUNO-CATALYSIS when it is replaced by a more actively metabolizing substrate (for ex- ample, glucose). The increase in galactose fermenting activity with the elimination of the fermentation of glucose is explained by assuming that the common preenzyme is adapted now for one substrate and then for another. This does not appear to offer sufficient grounds for a theory of the production of a new enzyme protein from a common protein when it is in contact with each member of a group of related sub- strates. The application of the enzyme precursor idea to such problems seems to exaggerate the significance of such changes. In an actively metabolizing environment with or without outside nitrogen source, an enzyme can be reversibly changed to an inactive form. The restoration of activity when in contact with substrates or in their absence in no way justifies the assumptions that the inactive precursor form originated first during the cytoplasmic synthesis of the protein. The fact that the active enzyme can assume an inactive or weakly active state when a substrate (galactose) is eliminated, or when its metabolism is followed by the metabolism of another substrate, shows clearly that the enzyme which was once active, could assume an inactive form, and therefore justifies the belief that the active form of the enzyme is synthesized first. There seems therefore no experimental basis for hypothesizing the existence of enzyme pre- cursors as being synthesized prior to the active enzyme (see further a succeeding section). Third: Cells do not manifest the quality of producing a new, "adaptive enzyme" for a substrate which does not possess an enzyme- specific reactive group common to a substrate which is normally readily metabolized by a given cell. A cell cannot be forced to produce an enzyme to metabolize a substance for which there is no genetic pro- vision in its species specific physiology and economy. Staphylococcus aureus, for example, lacks the ability to metabolize formate in a manner similar to that exercised by E. coli. Staphylococcus lacks the genetic ability to produce the formic hydrogenylase. Hemophilus influenzae has not been shown to adapt itself to synthesize the porphyrin molecule involves a Walden inversion of C-4 and is catalyzed by an enzyme which they call "galactowaldenase." The coenzyme of this reaction has been isolated and character- ized as uridine-difhos'phate-glucose (UDPG). This coenzyme has been found in animal tissues and in yeast not adapted to galactose, a-1 ,6-Glucosedi'phosphate, dis- covered and isolated by Cardini et al. (1949), has been found to function as co- enzyme for phosphoglucomutase (reaction III). MECHANISM OF ANTIBODY FORMATION 99 and coenzyme I no matter how long it may be in contact witK them. Pneumococci do not contain catalase, and are incapable of synthesizing an enzyme no matter how long they remain in contact with catalase or with non-toxic concentrations of hydrogen peroxide as substrate for catalase. Staphylococcus aureus is incapable of synthesizing nico- tinamide and thiamine despite the fact that they are capable of in- corporating them into their metabolic system. The assumption of Spiegelman (1946) that there appear endog- enous sources of protein, which are available for transformation into enzyme protein does not therefore appear to be a valid and experi- mentally tenable conception. His assumption also that when a cell is forced to form a new enzyme it may draw upon existent enzymes as a source of protein has been shown to be an impossibility in those bac- terial cells where the required particular enzyme (so-called "adaptable" enzyme protein) is not in existence to start with. The above cited facts as examples of generally occurring events contradict these assump- tions that an organism can synthesize a new enzyme under pressure. Under any circumstances, if an adaptive enzyme formation is a reality, it is essential that its characteristics be determined not by an increment in a common type of a metabolic effect, but by unequivocal classical methods of chemistry and immunology. The application of the following two critical tests seems to be essential: (1) That there be an absolute qualitative and quantitative differ- ence in the amounts of "galactose apoenzyme" (Spiegelman, Reiner, and Morgan, 1947) in adapted and non-adapted yeasts; and (2) That the assumed presence of this adapted galactose apo- enzyme can serologically be demonstrated in the galactose fermenting adapted and not at all in the non-adapted yeast cells. The latter method of analysis should particularly be useful in view of the enormous number of experimental facts to show that a- and P- configurations, d- and 1-activity, -COOH and -CONHo, etc. closely related groups present in antigens can be finely differentiated. If the observed increment in the fermentation of galactose is due to a de novo synthesis of a species of an enzyme protein, it should be possible to demonstrate this synthesis in the above mannet. The presence of even a few molecules of this enzyme protein in a non-adapted cell means a genetic provision to synthesize this enzyme. A measurable increment 100 IMMUNO-CATALYSIS in this respect in an "adopted" cell is merely a quantitative change and, therefore, does not represent theoretical significence. d. Consideration of the Primary Role of Antigen in Antibody Production as Postulated by Burnet, et al. Burnet, et al. postulate that antigens taken into the antibody forming cells lastingly modify cellular proteinase units which then synthesize antibodies specific to antigens. The claimed lasting character and the inheritability from parent to daughter cells of this modification implies, in the light of the present concept of the role of genes in the synthesis of specific enzymes, that antigens bring about an inheritable variation in the genes. This varia- tion is not a degradative mutation but a building-up process since the specific enzyme system not only is still capable of synthesizing the total quota of normal but also the new species specifically related antibody globulins. This assumption conflicts with the well established genetic characteristics of species. Known facts instruct us that a cell may undergo the following three types of genetic changes: (1) Chromosomal changes and gene mutations. Such directed changes have not been demonstrated in antibody formation against foreign proteins. (2) A building-up process which requires the incorporation and inheritance of genetic factors acquired by a cross-fertilization (Lindegren and Lindegren, 1945), or, in an asexual (?) cell (e.g., bacteria), by the incorporation into the recipient reces- sive cell of cytoplasmic material or of a catalyst of genetic nature from another strain of higher order but of the same species (Avery, et al. 1944). Recombinations might also re- sult in degenerative changes depending on the nature of recombining factors. (3) Degenerative mutations by which a cell undergoes hereditary losses as the consequence of the action of toxic agents, as observed in the development of resistance to drugs (Sevag, 1946). The hereditary cytoplasmic changes in the enzymatic constitution of yeast (Ephrussi, et al., 1949; Slonimski and Ephrussi, 1949) are clearly degenerative changes. The changes in the antigenic types in 'Paramecium aurelia (Sonneborn, 1949) may also be interpreted as degenerative changes. In this case, every potentiation of an antigen re- sults from or is associated with the suppression of another. MECHANISM OF ANTIBODY FORMATION 101 According to the first process, only species specifically related strains of cells cross-fertilize to produce a new strain. No case is known which shows genetic cross-linkings between cells not related species specifi- cally. It would therefore seem improbable that thousands of proteins and thousands of artificial antigens species specifically foreign to the host undergoing immunization are capable of genetically cross-linking with the genes of the host to bring about a building-up process in the antibody producing cells as implied by the theory of Burnet, et al. If the life-long immunity to measles and yellow fever viruses persists long after these agents are eliminated (assuming that this is an absolutely proven experimental fact), one may, perhaps, indulge in speculating that these viruses function as "quasi-genetic factors" re- lated species specifically to the globulin synthesizing cells of the host by being derived from them so as to offer a support for the implica- tions of Burnet's theory. However, there are as yet no traces of experi- mental data that smallpox, measles and yellow fever viruses bear species specific relationship to the host cells involved in antibody globulin synthesis. On the other hand, the generally supported idea emphasizes the fact that the observed life-long immunity to these viruses may be due to their multiplication in the host as variants deprived of their properties to cause observable pathological symptoms, as for example, the cow-pox virus, probably arising as a mutation of human smallpox, modified rabies virus produced by brain-to-brain passage in rabbits, and yellow fever virus strain 17D which was evolved by hundreds of passages in tissue culture in media containing em- bryonic tissue from which neural tissue had been removed. The result- ing virus has lost its neurotropic character and has been widely used for human immunization (Theiler and Smith, 1936). Reference can be made also to six mutant strains of tobacco mosaic virus each producing characteristic leaf symptoms, and possessing different serological properties and amino acid composition (Knight, Stanley, 1941; Ross, 1941, 1942). These facts would fall in line with our theory of Immuno-catalysis that viruses, like other antigens, function as catalysts in producing specific antibodies. In this role, they do not produce a permanent genetic change in the globulin synthe- sizing cellular enzyme system, but direct certain steps in globulin synthesis to yield antibody globulin. This directive influence continues to function so long as antigen persists (Libby, 1947) as a component 102 IMMUNO-CATALYSIS of the complex enzyme system involved in globulin synthesis. Antigens thus function as extraneous catalytic factors, and not as specific modi- fiers of the cellular proteinases. Dougherty, White and Chase (1945) compared the antibody con- tent of normal and malignant lymphocytes. Hemolytic filtrate of Staphylococcus aureus was used as antigen in the immunization of mice. Tumors were transplanted into mice in the middle of, before and following the courses of immunization lasting from 15 to 30 days. Extracts of normal and malignant lymphocytes, and blood sera were assayed for their antihemolytic antibody titers. They reported that the growth of an antibody-containing tumor transplant in normal mice was accompanied by the development of antibody-containing malignant cells. The normal lymphocytes of the host animal receiving this trans- plant likewise contained antibody, "As the growth of the tumor pro- ceeds, increasing number of antibody-containing lymphocytes are formed. Thus, the total quantity of available antibody in the lymphoid structure is dependent upon the number of antibody containing lymphocytes. This type of antibody production, i.e., multiplication of pre-existing cells containing immune globulin, may be a prominent mechanism of antibody production in the normal organism." On the basis of the above quotation one may be tempted to postulate that the lymphocytes have acquired the ability to synthesize antibody in the absence of antigen. However, there is no statement concerning the question of the number of transplants one can make before the antibody formation ceases. In conversation. Dr. A. White informed me that the third or fourth transplant ceases to show the presence of anti- body. It would thus indicate that successive generations of lympho- cytes are non-homogeneous with respect to their content of antigen. In the third transplant only a few lymphocytes will contain antigen. That is, the antigen will be so diluted by this time that it can no longer produce a measurable amount of antibody. e. Consideration of Antigens as Toxic Agents Causing the Muta- tion of Globulin Synthesizing Enzyme System. Another point which may be worthy of consideration is the question of whether or not anti- gens function as toxic agents or inhibitors of the enzyme system in- volved in the synthesis of globulins. If they function as toxic agents one may assume that they bring about, as stated above, the second type of change in a cell which is of degradative nature. Since this type of MECHANISM OF ANTIBODY FORMATION 103 change, unlike the changes ensuing from cross-fertilizing genetic changes, is invariably brought about by agents genetically, or species specifically, not related to the cells undergoing this change, and since antigens are foreign to the host undergoing immunization, one may postulate similarities between the action of antigens and muta- genic chemical agents. Such a change of degradative nature could be expected to yield mutant cells which, one may assume, produce specific antibodies as abnormal by-products. As in all degradatively produced mutant cells, such a postulated change in host cells would be expected to be of permanent and inheritable character, and would, superficially, support the idea of life-long lasting immunity to certain viruses claimed by Burnet, et al. and implied by their theory. Since, however, immunity to the dominant number of toxic and non-toxic antigens is of temporary nature, one must assume that antibody producing cells or their enzymes do not undergo a degradative mutation preceding or concomitant wdth active immunization. Furthermore, toxic agents which induce muta- tions, are known to suppress or interfere with the metabolism or synthe- sis of specific cell metabolites or components. In contrast, immunity to antigens is accompanied by hyperglobulinemia and not by hypo- globulinemia, indicating that antigens do not suppress the synthesis of globulins. Certain antigens produce toxic effects and pathological conditions in a host followed by antibody production or immunity. Non-antigenic drugs or toxic agents also produce noxious effects on cells (for example, inhibition of enzyme activities) resulting in the development of re- sistance to these effects. Unlike the latter effects, certain antigenic toxins, for example, produce their effect functioning as enzymes and destroying the host tissue components, e.g., lecithinase, proteolytic, necrotic and hemolytic actions of bacterial products and snake venoms. However, the host system is not known to develop a non-immune resistance'*' to these toxins in a manner comparable, for example, to the resistance manifested by a bacterium to a drug. Bacterial drug *If a resistance in a host can be shown to have resuhed from the action of the toxins of infectious agents, it is not improbable that the action of the toxins, hke those of the antibacterial drugs (Sevag, 1946; Sevag and Gots, 1948; Steers and Sevag, 1949; Sevag and Steers, 1949), have permanently abolished the receptor sites of the host tissues. This would result in a life-long resistance to the toxins of the infectious agents concerned. 104 IMMUNO-CATALYSIS resistance does not involve a continued production of an antibody-like substance capable of combining with a drug to neutralize its toxic effect. On the other hand the only resistance a host develops to an antigenic toxin is believed to be by means of the production of specifically combining and neutralizing antibodies during a limited period. Consideration of the above discussed questions seems to show that the phenomenon of immunity is conditioned by the presence of anti- genic units but that this conditioning disappears when the antigenic units are eliminated from the host system. 6. On Jordan's Autocatalytic Theory of Antibody Formation Jordan (1944) has proposed an autocatalytic theory of antibody production. According to this theory, the multiplication of a gene or a virus only in the presence of itself, is a basic autocatalytic phenomenon in all living processes including the formation of antibody. The mole- cules, or fractions of molecules which possess the ability for parallel orientation and are capable of multiplication are identical and exhibit affinities for each other. A molecule M, which is capable of multipli- cation, may be composed of /^l, /a2, iS fractions each of which pos- sesses the property of attracting others to itself. These molecular frac- tions per se, in solution, by themselves possess very little tendency to combine with each other; they do so, however, when M is present form- ing another molecule identical with M. The attraction among the fractions jul, /x2, ju.3 is attributed to resonance of the fractions. That is, two substances are said to attract each other through resonance when they possess groupings identical or nearly so. Jordan had earlier ex- pressed similar views concerning this question which had been denied by Pauling and Delbriick (1940) on grounds that the resonance energy would be so small as to be ineffective. Reasoning in similar fashion, Jordan (1940) had postulated that antibodies and specific antigens are identical, a postulate which is related to the concept of Biichner (1893). This likewise has been refuted by the chemical- immunological evidence supplied by Haurowitz, et al. (1942). The latter investigators using iodo-proteins, bromo-protein and arsonil- azoprotein as antigens showed that none of the antibodies contained the MECHANISM OF ANTIBODY FORMATION 105 determinant group of tKe antigen or a serologically related group in the repective antibody molecules. The concept of the identity of antigen and antibody postulated by Jordan must be rejected also for genetic reasons, and for reasons that an antigen is antigenic by virtue of being species specifically unrelated to the antibody globulin elaborated by the immunized host. Jordan (1944) assumes that antibody formation does not take place during a period of from six to 48 hours following the injection of anti- gen. This assumption is to be denied for it has been observed (Ham- burger, 1902; Oerskov and Anderson, 1938; Ramon, 1928, etc.) that the production of antibody can be demonstrated three to four hours after the injection of antigen. On the basis of this assumption Jordan postulates that antigens initially produce only small amounts of anti- bodies, and that the subsequent increase in the amount of antibody results from an autocatalytic process mediated by the antibody mole- cule itself. As stated above, a measurable amount of antibody has been demonstrated a few hours after the injection of antigen. There is no doubt that non-measurable amounts of antibody must have been synthesized at a still earlier period, indicating a rapid output of anti- body by a process which, in our opinion, is possible if antigen func- tions as a catalyst. The autocatalytic concept of Jordan implies that the antibody pro- duction should be a continuous process. A second implication of Jordan's concept is that the introduction of an antibody obtained from an actively immunized animal into another animal of the same species should autocatalytically produce new antibody in amounts many-fold greater than the amount introduced. The production of antibody postulated in this manner in a normal individual would be comparable to the production of an infection in a host by an infectious agent de- rived from a disease bearing host. Jordan considers this possibility in the above described manner, but was unable to cite a single experi- mental observation. As is well known, passive immunizations are of very short duration, and the experiments of Heidelberger, Schoenheimer and their associates (1942), as discussed in a preceding section, have shown that an antibody derived from a rabbit and introduced into a normal rabbit enters into metabolic reactions resulting in its disappearance. 106 IMMUNO-CATALYSIS 7. The Concepts of "Proteinogen" and Enzyme Precur- sors Considered in the Light of the Specificities of Antigens and Antibodies and Their Digestion Products Northrop (1946, 1948) has conceived the idea that normal proteins, enzymes, antibodies, viruses are derived from a "proteinogen" or "ur- protein" by a purely catalytic, or autocatalytic reaction which does not require energy. The synthesis of the "proteinogen" molecule, re- quiring energy from another source, is, likewise, assumed to be auto- catalytic. The proteinogen molecule possesses the general chemical structure characteristic of the species. The formation of the above mentioned biologically and chemically specific proteins from the master proteinogen molecule is considered as analogous to the formation of pepsin and trypsin from their respective precursors. Assumption is also made that by such transformations from proteinogen to specific pro- teins the latter acquire specific enzyme and immunological properties. If we understand Northrop's postulates correctly, they imply that the derivation of the specialized functions of various proteins found within a kind is associated, or brought about, by an autocatalytic reaction which is capable of converting proteinogen molecules into many differ- ent species of protein molecules. On this basis, enzymes, such as lysozyme, d-ribonuclease, cytochrome c, with molecular weights of about 15,600, myoglobin (mol. wt. 15,900), lactalbumin (mol. wt. 17,400) etc. which are structurally, enzymatically and serologically different from species specifically related proteins have acquired these specific properties during their autocatalytic geneses from pro- teinogen molecules. Since, as stated by Northrop, the molecular weight of the proteinogen molecule is equal to or greater than that of the protein derived, and since the autocatalysis in these conversions does not require energy, it must be assumed that: (a) there must be as many specific proteinogens as there are specific proteins. For example, molecular weights for certain specific proteins are: pepsinogen, 42,000; pepsin, 38,000; chymotrypsinogen and chymotrypsin, about 40,000; and trypsinogen and trypsin, 35,000; (b) it must be assumed that the master proteinogen molecule must have a molecular size larger than the largest specific molecule found in a species, so as to be autocatalytically convertible to all molecular sizes; or (c) there must be a proteinogen MECHANISM OF ANTIBODY FORMATION 107 molecule which autocatalytically can be polymerized or aggregated to yield the largest molecular species one can find in a species. a. Consideration of Enzyme Precursors. Pepsinogen is changed into pepsin under the following conditions: (a) by hydrogen ion on the acid side of pH 6.0 (maximal activity at pH 2.0); (b) by addition of pepsin; or, (c) increasing the salt concentration to increase the rate of conversion. This change from pepsinogen to pepsin in each case is associated with the dissociation of a polypeptide, with a molecular weight of 6,000, which is capable of recombining with pepsin at pH 5 to 6 to form a dissociable pepsin-inhibitor complex. In this combina- tion the activity of pepsin is blocked but not destroyed. On long stand- ing with pepsin between pH 2.0 and 5.0, the inhibitor is destroyed. The pepsin inhibitor has exposed basic groups. The change from chymotrypsinogen to chymotrypsin by slightly acid solution, or by trypsin, would appear to be the result of a limited hydrolytic reaction without any apparent effect on the molecular weight of the inactive protein. According to Northrop (1937), the chymotrypsin molecule contains five amino groups* more than chy- motrypsinogen, indicating the opening of a peptide ring. There is also an associated shift in isoelectric point from pH 5.0 to 5.4. No change was observed in elementary analysis, and the tyrosine plus tryptophane content. In the change from trypsinogen into trypsin, no measurable increase in amino groups was detected. As suggested, it might be that the hydrolysis of a peptide link, too small to be detected, might have taken place. The change from trypsinogen to trypsin takes place at pH 7.0 to 9.0 without the aid of any outside activator, at pH 3.0 to 4.0 by means of a mold kinase (Kunitz, 1938), in the presence of enteroki- nase at pH 6 to 9.0 where spontaneous activation of trypsinogen occurs readily, or, simply, in the presence of calcium ion at pH 8.0 and 5°C (Kunitz, 1945). Under this latter condition, trypsinogen is quantita- tively converted into trypsin vdthout the formation of "inert protein" formed under other conditions. In the above cited typical examples of enzyme precursors which are studied more completely than any other crystalline enzymes isolated, the change from inactive to active form involves treatments which are * Amino nitrogen as per cent of total nitrogen: chymotrypsinogen, 4.7; chymo- trypsin, 6.0 (Northrop, Kunitz and Herriott, 1948). 108 IMMUNO-CATALYSIS found not to cause deep structural variation. One may therefore, perhaps, ask the question of whether or not these changes can be taken as proof that these enzyme precursors are really the forms which are synthesized first. It would seem that the inactive forms could easily be assumed to have been derived from the active forms under in vitro or in vivo environmental influences.''' The ease with which pep- sin can reversibly combine with pepsin inhibitor recovered from pep- sinogen complex (Pepsin + pepsin inhibitor ^ "pepsinogen"), and the ease with which both chymotrypsinogen and trypsinogen are changed to their respective active forms with reagents such as calcium ion, or, spontaneously, at neutrality, permits, as an alternative, the as- sumption that the enzyme molecule is synthesized first and subse- quently converted into certain reversibly inactive forms. As will be discussed later in this work, serum proteolytic enzyme which, among other proteins, also digests fibrin clot exists in an inactive form in normal serum. The treatment of serum with chloroform, ether, ethyl alcohol, etc., or dialysis, liberate the active enzyme; the enzyme may also be liberated spontaneously, on standing in the cold. Thus, it is most likely that these treatments dissociate a substance from an inactive complex which blocked the activity of the serum protease. Northrop (1946) mentions the serological behavior of the precursor and the respective active enzyme as proof that the latter is derived from the former, or that a protein may be formed by an autocatalytic re- *One must ask the question of whether chymotrypsinogen is a precursor of the chymotrypsins or an artifact derived from native chymotrypsin as the result of the conditions used for their isolation from the source material. Similar question may he raised with respect to the relationship between trypsinogen and trypsin. In relation to these questions we must be aware of the fact that there is no evidence that chymotrypsinogen (and trypsinogen) fer se exist in the pancreatic tissue. All that we know is that they are obtained from extracts made by means of strongly acid, 0.25N H2SO4, solutions at a temperature of 5°C. After several fractionations from solutions on the acid side, a crystalhne material from a solution at pH 5.0 is obtained. A solution of these crystals on adjusting to pH 7.6 with a trace of trypsin, yields the crystalline chymotrypsin (Northrop, et al. 1948). The intriguing question is the nature of the possible effect of the cold strong acid on the raw source of these enzymes. Is it not possible that under the above prepara- tive conditions these enzymes undergo dehydration of the free amino and carboxyl groups yielding chymotrypsinogen (and trypsinogen) which reverse to the active forms spontaneously in weakly acid or alkaline solutions with or without the aid of activating agents? While correcting the page proof, S. D. Elliott (J. Exp. Med., 92:201-218, 1950) reported that the maximal production of the precursor of streptococcal proteinase oc- curred only in slightly acid, and none in neutral or alkaline environment. MECHANISM OF ANTIBODY FORMATION 109 action from another protein. Ten Broeck (1934) reported that guinea pigs receiving chymotrypsinogen were sensitized against the so-called precursor itself but not against chymo trypsin, and guinea pigs receiving chymotrypsin were sensitized against itself, but not against chymotryp- sinogen. But, he also reported that in some cases there were cross- reactions, particularly between the chymotrypsin and chymotryp- sinogen, indicating a close serological relationship between the inactive and active proteins. In the absence of minute amounts of trypsin, the change from chymotrypsinogen to chymotrypsin may occur spontaneously, slowly at pH 5.0, and faster in slightly acid or weakly alkaline solutions. The change from trypsinogen to trypsin is reported to occur in the presence of calcium ion and at pH 7.0 to 9.0 without the aid of any outside activator. Under these conditions, the increase of five free amino groups in chymotrypsin does not appear to indicate an opening up of stable peptide linkages as would result from the characteristic action of proteolytic enzymes. These changes may involve salt-like reactions with basic or acidic groups, or result perhaps from a reaction involving hydrogen bonding. Under such conditions a serological difference between the two forms of protein is conceivable (see Pauling in Land- steiner, 1945). As such, chymotrypsinogen, for example, could be a derivative of chymotrypsin and not a precursor. In relation to this, it may be pointed out that an antigen can be modified in similar manner to yield an immunologically different protein derivative. Landsteiner, et al. (1932) reported a serological difference between an antigen with a terminal -COOH group and that having -CONH2 as terminal group instead. Or immune sera prepared with a methyl ester antigen precipitated the homologous azoprotein but not that made from the parent, p-aminobenzoic acid. If the ester was hydrolyzed by gentle alkaline treatment of the azoprotein a strong reaction with serum for p-aminobenzoic antigen (Landsteiner, et al. 1927) took place. Chow and Goebel (1935) showed that acetylation of the amino groups of an antibody abolished its reactivity with a specific polysaccharide. Con- versely, the esterification of the -COOH group of the polysaccharide deprived it of its reactivity with the native antibody. Saponification of the esterified carbohydrate restored the activity to react with the specific antibody. In the above cited instances with the inactive and active forms of 110 IMMUNO-CATALYSIS enzymes, and in the latter instances with antigens and antibodies, we are observing reactions with superficial groups and not deep seated chemical changes in the structure of protein molecules. In immunologic respects, the case of trypsinogen and trypsin, or chymotrypsinogen and chymotrypsin, does not appear to deviate from the above basic processes, b. Derivatives of Chymotrypsin and Their Properties. In a com- prehensive investigation, Jacobson (1947) reported that the tryptic activation of chymotrypsinogen at 0°C consists of the following trans- formations: trypsin CO Chymotrypsinogen > Tr-Chymotrypsin; one peptide bond is (minute amounts) split. (a) trypsin ■^ 8-Chymotrypsin; one peptide bond is split. (2) TT-Chymotrypsin simultaneous competitive , reactions (b) spontaneous > a-Chymotrypsin; three peptide bonds are (or autolytic) split. Note: In connection with the above spontaneous conversion of TT-chymotrypsin to a-chymotrypsin it may be recalled that Kunitz and Northrop (1935) had found that chymotrypsinogen spontaneously (without the minute amounts of trypsin) undergoes a change, leading to the formation of chymotrypsin. There was no marked pH optimum for this change, but it occurred faster in weakly acid or alkaline solu- tions independent of the effect of trypsin. TT-Chymotryfsin is an hitherto unrecognized chymotrypsin, and is not isolated. B-Chymotrypsin is likewise unrecognized hitherto and is not crystal- lized. Specific activities: Jacobson reported that the specific activity of TT-Chymotrypsin is 2 to 2.5 fold > a-chymotrypsin, and that of 8-chymo- trypsin is 1.5 fold > a-chymotrypsin. It is most interesting to note that in the conversion of inactive chy- motrypsinogen into the above three chymotrypsins, the splitting of one "peptide bond" yields the most active enzyme, 7r-chymotrypsin, and that a-chymotrypsin, resulting from the splitting of five (?) "peptide bonds" in chymotrypsinogen, is the least active form. MECHANISM OF ANTIBODY FORMATION 111 According to Jacobson, the number of titratable groups per a-chymo- trypsin molecule is greater by six to nine than that per chymotrypsino- gen molecule, indicating a probable hydrolysis of four peptide bonds during the conversion of chymotrypsinogen to a-chy mo trypsin. He considers several possibilities to support or refute his assumption that the above transformations involve the splitting of peptide bonds. He considered for chymotrypsinogen a protein structure as formu- lated by Mirsky and Pauling (1936). According to these authors, the native globular unconjugated proteins consist of an uninterrupted polypeptide chain which is continuous throughout the molecule. The polypeptide chain folds into an uniquely defined configuration. This configuration is maintained by means of hydrogen-bonding betv^^een the peptide nitrogen and oxygen atoms and also between the free amino acid carboxyl groups of diamino and dicarboxylic amino acids. In the case of chymotrypsinogen, the hydrolysis of peptide bonds, leading to chymotrypsin with six to nine titratable groups more than the original molecule, must occur near the ends of the peptide chain. The hydroly- sis of certain terminal amide bonds by trypsin was taken into considera- tion in view of the fact that trypsin is able to hydrolyze the amide bond in benzoyl-glycyl-arginamide, benzoyl-arginamide, benzoyl-glycyl- lysin-amide and benzoyl-lysin-amide (Bergmann, et al. 1939). Such a hydrolysis of a protein should yield ammonia. The finding of Butler (1941) that at the maximum activation by trypsin 1 NHs per 12 chymotrypsinogens is produced, and no increase in NHs took place after 66 hours, was not considered of any significance by Jacobson. It would, however, seem that the ease with which the amides are hydro- lyzable might be in line not only wdth the catalytic conversion of chymotrypsinogen into chymotrypsin by trypsin, but also, particularly, with the spontaneous (wathout the presence of minute amounts of trypsin) change of chymotrypsinogen into chymotrypsin as has been observed. According to Jacobson, the splitting of a terminal amide bond from chymotrypsinogen does not occur to any significant extent during the tryptic hydrolysis leading to 7r-chymotrypsin. The splitting off of an amino acid a-amide during the autolytic (or spontaneous) conversion of TT-chymotrypsin leading to a-chymotrypsin is not impossible if such bonds exist in chymotrypsinogen since a-chymotrypsin is reported to possess such an activity (Bergmann and Fruton, 1938; Fruton and 112 IMMUNO-CATALYSIS Bergmann, 1939), however, Butler's results do not indicate that the hydrolysis of amide bonds occur to any significant extent during the activation process. In view of these experimental material, Jacobson does not seem to favor, on the basis of the fonnulation of Mirsky and Pauling, that chymotrypsinogen contains or consists of one unin- terrupted peptide chain. Assumption is made that if proteins contain several peptide chains, the breaking of about four peptide bonds in chymotrypsinogen with the result of an increase of six to nine titratable groups when a-chymotrypsin is formed, may be explained. This could result from the proteolytic rupture of peptide bonds in the peptide chain or chains of chymotrypsinogen and of the proteins (one of which is TT-chymotrypsin) which are intermediaries of chymotrypsinogen and a-chymotrypsin, without necessitating the formation of non-protein- nitrogen. This, however, leaves unexplained the origin and nature of 2.7 per cent non-protein-nitrogen formed during these reactions, part of which at least can be accounted for as ammonia nitrogen as Butler (1941) has found. Jacobson arbitrarily chooses a molecular weight of 36,000 (36,700 by Brand and Kassel, 1941) for chymotrypsinogen. Since he found the production of 2.7 per cent non-protein-nitrogen during the conversion of chymotrypsinogen into a-chymotrypsin, the molecular weight of the latter was calculated to be at least 35,000. These values are basically diJfferent from those of 32,000 to 36,000 for chymotrypsinogen, and 40,000 for a-chymotrypsin as measured by Kunitz and Northrop (1935), and Kunitz (1939). Consequently, Jacobson's chymotrypsin would appear to be a derivative of diminished molecular weight, and that of Kunitz and Northrop a chymotrypsinogen derivative of about 1 1 per cent increased molecular weight. This discrepancy may, how- ever, be due to the inconstancy of the molecular state of the active enzyme. In connection with the above observations of Jacobson, it is to be noted that earlier Kunitz (1938) had prepared crystalline a- and y-chy- motrypsins from chymotrypsin. These two proteins did not differ in activity from the parent chymotrypsin of 40,000 molecular weight. The molecular weights of a- and y-chymotrypsins were reported to be, respectively, 27,000 and 30,000. It is interesting that a non-essential part or component corresponding to a molecular weight of 10,000 to 13,000 can be split off the chymotrypsin molecule (or complex) with- MECHANISM OF ANTIBODY FORMATION 113 out affecting the degree of enzyme activity and immunological speci- ficity (Northrop, et ah 1948). It would be interesting to know how far one can reduce the molecular size of the smallest of the three chymo- trypsins without affecting its serological and enzymatic properties. c. Virus-Host Relationship. The experimental demonstration of the independence of the biophysical and serological specificities of virus preparations derived from different species of hosts may, at present, for technical reasons, be difficult to achieve, particularly with animal viruses of greater complexity and particle size which have not lent themselves to purification, for example, by crystallization. The failure to separate the host components from such virus prepara- tions may easily give rise to assumptions that these viruses carry the specificity of tissues from which they are derived. It has been reported (Knight, 1946), for example, that the virus of influenza contains a specific normal component which is characteristic of each individual host from which the virus is obtained. However, this component ex- hibits a serological specificity distinct from that of the virus. On the other hand, similar studies with plant viruses which are simpler in composition and are obtainable in crystalline form have shown the biophysical and serological individuality of the viruses irrespective of the plant hosts in which they are multiplied (Bawden and Pirie, 1944, 1946; Gaw and Stanley, 1947; Malkiel, 1947). It has been found that the purified preparations of tomato bushy stunt virus de- rived from the leaf sap and the fibrous residues of infected tomato plants did not differ in any significant manner. Purified tobacco mosaic virus and the rib-grass strain of tobacco mosaic virus, in these respects, showed identical properties. Malkiel (1947) reported that there was no serological difference found for either tobacco mosaic virus or rib- grass virus when each was isolated from widely separated plant species. Immunochemical studies failed to indicate the presence of any normal host protein as a component of any of the virus preparations. These findings do not offer any support to the assumption that the viruses are autocatalytically derived from a normal proteinogen possessing the species characteristics of the host in which the virus multiplies. And there are, at present, no data to indicate that normal host pro- teins or the viruses derived from the host can be modified chemically or otherwise, to show any serological interrelationship suggestive of the autocatalytic origin of viruses from the host proteins. 114 IMMUNO-CATALYSIS Any body of experimental data offered for the formidation of a concept dealing with the mechanism of the m-ultiplication of a viral or bacterial parasite, and that of the synthesis of viral or bacterial pro- teins, must take into consideration the following well-kno%vn facts: C«) no living cell as yet has been found to synthesize a protein which is not species specifically related to itself; and (I?) no living unit with distinct genetic makeup has been demonstrated to exist which is capable of directing, in accordance with the image of its own species specific characteristics, the synthetic facilities of another living unit belonging to an entirely different species. d. Structural Specificity of Polypeptides of Low Molecular Weight. The above considerations suggest that structural specificity is introduced into protein molecules during their synthesis from smaller building blocks. What is more fundamental is the fact that enzymes of the smallest unit of molecular weight (15,000), that is the smallest unit which cannot reversibly be split, possess biological specificities. Polypeptides with a molecular weight of 6,000, such as trypsin and pepsin inhibitors, and d-glutamic acid poh^peptide of B. anthracis possess specific serological (Ivanovics and Bruckner, 1937, 1940) and, other activities. Pepsin inhibitor, for example, specifically combines with pepsin, but has no demonstrable effect on the activity of crystalline trypsin and on the milk clotting activity of crystalline chymotrypsin or commercial rennet. Trypsin inhibitor isolated from trypsinogen crystals (inhibitor-free trypsinogen crystals have not been obtained, Kunitz and Northrop, 1936) inactivates trypsin, activates chymotrypsinogen to chymotrypsin, but has no effect on the milk clotting action of pepsin. These findings show that polypeptide mole- cules of 6,000 molecular weight show a high degree of structural specificity. Irrespective of their origin, these characteristics are in the polypeptide molecules, indicating that they are introduced there during their synthesis as independent units, or as parts of certain larger protein molecules, split therefrom, perhaps, by enzyme hydrolysis. In this connection, it is of interest to note that Kunitz (1945, 1946, 1947) isolated a trypsin inhibitor of globulin nature from soy bean which formed a crystalline inactive complex with trypsin. Lineweaver and Murray (1947) isolated from egg-white an ovomucoid (mol. wt., 29,000) which caused 50 per cent inhibition of trypsin at equimolar concentrations. It would be of pertinent interest to learn if these latter MECHANISM OF ANTIBODY FORMATION 115 two inhibitors can be subjected to hydrolytic cleavage to yield poly- peptides comparable with those isolated from pancreas and serum (Schmitz, 1938). Landsteiner and Chase (1933) showed that the reaction between the antigen, sheep serum, and its homologous antibody could be in- hibited by a readily dialyzable albumose obtained from the products of peptic digestion of the coagulated sheep serum antigen. Landsteiner (1942) reported also that the products of silk hydrolysis consisting of peptides having molecular weights from 600 to 1,000 were capable of inhibiting the reactions of precipitin sera for silk. From these results Landsteiner inferred that silk fibroin contains determinant structures of not more than eight to 12 amino acids. Holiday (1939) immunized rabbits with horse serum albumin which was purified by electrophoresis and shown to be homogeneous in the ultracentrifuge. The purified antigen was digested with 1/4000 parts by weight of pepsin at pH 2.0 for periods of five and 30 minutes. In electrophoretic analysis the five minute digest, showed two com- ponents with mobilities different from that of the intact antigen, and by ultracentrifugation these components were inferred to have a size 14 that of the original molecule. The 30 minute digest showed the components to be of Vs the size of the intact antigen. The V4 size components showed practically undiminished precipitating activity, and Vs size components showed no precipitating activity but partially inhibited the reaction between the whole antigen and antibody. The 14 and Vs size digestion products would have molecular weights of about 17,000 and 8,500 respectively. Holiday concluded that at the stage of division into eight parts there is still evidence of affinity be- tween the digestion products and antibody. In this connection it is interesting to observe that non-protein products formed by peptic diges- tion of ovalbumin, according to Tiselius and Ericksson-Quensel (1939) had an average molecular weight of 1,080 (see also Haugaard and Roberts, 1942). Winnick (1944) reported that partial hydrolysis products from the action on casein of chymotrypsin, trypsin, pepsin, ficin or papain yielded products with an average molecular weight ranging from 600 to 450. Serological activities of these products were not studied. The above facts concerning the structural specificities of polypeptide molecules of as low a molecular weight as 600 to 1 ,000 seem to show 116 IMMUNO-CATALYSIS conclusively that these specificities are inherent in the reactions in- volved in their synthesis. It is difficult to conceive that a master proteinogen molecule can contain hundreds or thousands of speci- ficities, and the cells of each organ or specialized structure of the body must all contain the same proteinogen molecule. Our existing knowl- edge does not lend any support to these assumptions. e. Does Antigen Function as Coenzyme in the Production of Antibody.'* Northrop applies the proteinogen concept to the production of highly specific antibody proteins. He is of the opinion that antibodies are formed in a manner comparable to the formation of "adaptive enzymes." We have already discussed this process and concluded that it does not involve the synthesis of a new protein. The formation of an antibody, on the other hand, is a synthesis of a new protein as will be discussed below. What is more surprising is the analogy drawn between an antigen and a coenzyme; the former in the role of the latter is sup- posed to be capable of autocatalytically modifying serum proteins so that an antibody, instead of the normal serum protein, is formed. The hypothetical implications of the experiments of Pauling and Campbell (1942) may have been held in view in making this postulate. These authors reported the manufacture of "antibody" in vitro from denatured serum globulin in the presence of certain haptens. The claimed "anti- body" formed a precipitate with haptenic agent. There is no doubt that such an observation with great potentialities has been the object of many unsuccessful experiments in various laboratories, but only few have published their findings. In similar experiments, Haurowitz et ah (1946) found that the precipitates obtained in such experiments are not due to the effect of antibodies but are brought about by the non-specific flocculation of globulins charged positively at pH 5.0 to 5.5 by negatively charged azoproteins. Kuzin and Nervaeva (1947) could not detect the formation in vitro, using the method of Pauling and Campbell, of antibody to- polysaccharides from Type III pneumococcus, gum arable, Shigella dysenteriae, S. foradysenteriae Flexner, and Stre'ptococcus hemolyticus. Also, no antibodies were observed in solutions against dye haptens. Similarly, no antitoxin formation was observed during the dehydration of serum proteins in the presence of the toxin from Corynehacterium diph- theriae. In an earlier section of this work, we arrived at the conclusion that MECHANISM OF ANTIBODY FORMATION 117 all antigens fulfill the requirements of the criteria of ideal catalysis, they must, therefore, be considered as catalysts. Since all proteins are antigenic, they exhibit, therefore, biocatalytic activities. In Nor- throp's postulates, our concept of antigens acting as catalysts has under- gone a variation by which antigens have been considered as acting as coenzymes. This variation contains certain concepts of the mechanism of antibody formation fundamentally different from ours. His antigenic coenzyme is supposed to be capable of initiating autocatalytic con- version of normal to antibody globulin. In contrast, our concept deals with a mechanism of the development of the specificity of antibody globulin structure preceding the completion of the synthesis of a normal globulin molecule, under the directive influence of the active units of the antigenic molecule. In other words, the sfecifLcity of an antibody molecule is the consequence of specific cellular synthethic fwcesses catalytically inodified hy an antigen to conform with the configuration of certain active groups of the antigenic molecule. The use of the term coenzyme to account for the postulated role of antigen in connection with the formation of specific antibodies intro- duces confusion. A comparison, therefore, of the role of coenzymes in respiratory processes, and the role of antigens in the formation of antibodies seems to be required. Haptens, which constitute the prosthetic groups of conjugated antigens, are incapable of inciting the formation of specific antibodies without combination with a protein. This would indicate that the catalytic role of an antigen resides in the protein molecule and that only in this combination can a hapten exercise its determinant function. The coenzyme groups of the respiratory enzymes are likewise inactive when not in combination with specific proteins. The same coenzyme group seems to function exclusively with certain specific proteins, and not with all proteins, yielding a group of related enzymes, for example, flavoproteins, each characterized by distinctive specificity of action. This specificity resides in the protein molecule and not in the coenzyme group. On the other hand, in immune reactions if the conditions permit, the same hapten can be combined with any antigenic protein to demonstrate its determinant characters. In this respect, a hapten is not selective. In these combinations, both the protein and the hapten produce anti- bodies, respectively specific. In conjugated enzymes, neither the specific protein nor the coenzyme 118 IMMUNO-CATALYSIS group are by themselves active; they are interdependent. In conjugated antigens, vi^hile the determinant character of the hapten is dependent on its combination wdth a protein, the antigenicity of the latter is independent of the haptenic group. These basic differences in the relationship of the coenzyme group to the specific protein in conjugated enzymes on the one hand, and that of a determinant prosthetic group to the protein component of a conjugated antigen on the other, makes the assumption that antigens function as coenzymes en- dowed with autocatalytic powers unwieldy. Furthermore, the results of studies with bacteria and other cells fail to show that a coenzyme group is capable, adaptively or otherwise, of initiating the synthesis of a new protein in a cell specifically reactive with the particular coenzyme where there is none to be found to start with. In contrast, an antibody is produced in vivo only when there is an antigen present. This contrast likewise shows the lack of an experimental basis for an analogy between a coenzyme and an antigen. f. Specificity of Cleavage Products Derived from Antibody. How far down in molecular size an antibody molecule can be brought proteolytically, or otherwise, before it loses every vestige of reactivity with homologous antigen is a pertinent fundamental question which has not as yet been adaquately investigated. An antigenic protein hydrolytically split into polypeptide units would lose the ability of stimulating antibody formation, but would maintain the ability to combine with antibody and thus inhibit its reaction with whole antigen. The antibody combining abilities of smaller molecular entities are therefore inherent in these as well as whole molecules from which they are derived. Since these non-antigenic haptens are serologically specific, they must differ in this respect from the corresponding entities present in other proteins derived from the same species. It is reasonable therefore, to assume that there must function enzyme reactions specific for the synthesis of each protein and also for the synthesis of its haptenic parts which are obtainable by the hydrolytic cleavage of the whole molecule. In a consideration of the structural difference between y-globulin and antibody globulin (page 139), it was pointed out that diphtheria antitoxin (mol. wt., 184,000) on digestion with proteolytic enzyme gave a crystalline product (mol. wt., 90,000) 90 per cent precipitable- with diphtheria toxin (Northrop, 1942). Immunologically, this product. MECHANISM OF ANTIBODY FORMATION 119 possessing Vi the size of the original antitoxin molecule, was distinct from the normal serum proteins. It would thus seem that the antitoxin freed from the normal serum components behaved as a new species of protein molecule. This is understandable if we accept the interpretation that antigens function as catalysts, similar in this respect to other specific enzymes involved in protein syntheses. Related to the above finding, are the results of a study by Weil, Parfentjev and Bowman (1938). They reported that antibody globulin can be subjected to proteolytic digestion in a manner whereby it almost completely loses its ability to incite the formation of specific antibodies, without suffering loss in antitoxic properties. Diphtheria antitoxic globulin after partial peptic digestion at pH 4 and 4.5 was from 70 to 80 per cent non-coagulable by heat. The antitoxic component of the digest was separated by ammonium sulfate fractionation and purified by dialysis. The traces of remaining pepsin were eliminated by a freshly prepared suspension of calcium phosphate. This preparation was compared, in various serological and animal tests, with normal and immune horse plasma, and antitoxin globulin salted out with am- monium sulfate. Absorption experiments demonstrated directly that the antitoxic property was in the hydrolytically cleaved part of the solution and not in the unchanged fraction. The results of experiments showed that unimpaired— and even increased— antibody-function was possessed by this derivative of the antibody molecule. This derivative was so far removed that its property to function itself as an antigen was eliminated, and wdth it all traces of its original connection, detectable by immunological methods. The above results show that an antibody molecule can be stepwise split into various derivatives. In the first example, its serological relationship to normal serum proteins was eliminated without a loss in antigenic and antitoxic properties. In the second example, its antigenic relationship was eliminated without a loss of antitoxic activity. These facts show strongly that the groups responsible for the properties of antitoxin to combine with and neutralize toxin are distinct molecular entities derivable from the larger whole molecules. In a similar study, Peterman (1946) reported that the early effect of papain or bromelin on horse diphtheria antitoxin and beef serum globulin is quite similar to that of pepsin and trypsin. The molecules are split into halves. On prolonged digestion, these halves are split into 1 20 IMMUNO-CATALYSIS quarters. The observed molecular weights of the quarters were 47,000, though, on the basis of the molecular weight of 153,000 of the original globulin particle, they should have been only 38,000. Although 90 per cent of the globulin was split into quarters, much antibody activity was retained. The antitoxin activity of the whole digest was from 70 to 90 per cent of the undigested antitoxin when papain was used, and 100 per cent of the activity was retained when digestion was carried out with bromelin. In another study, Peterman (1946) found that human immune gamma globulin is split by pepsin into molecules of half size. Further digestion yielded smaller particles and ultimately dialyzable fragments. Immunological assays showed that in the digests containing half size antibodies the typhoid "O" agglutinin was lost, while the concentration of "H" agglutinin, of diphtheria and strepto- coccus antitoxin and of anti-influenza A remained substantially un- changed. In view of the above findings, one may expect that the specificity of antibody molecules may reside in still smaller units derivable from the quarter size digestion products. Dialyzable units of antibody may fail to agglutinate bacteria, or precipitate antigens, but they may inhibit the reactions between whole antigen and whole antibody. Our information concerning the structure of proteins may be ex- panded if the smaller proteolytic or hydrolytic cleavage products are isolated and characterized chemically and serologically. One may be able profitably to employ the technique developed by Tiselius (1947) and Martin and Synge (1945) for this purpose. The results of such studies may have specific bearing on the debate on the valency of anti- body molecules, and on the mechanism of antibody synthesis. Conclusion. A consideration of the above experimental facts shows clearly that biological specificities are inherent in the respective chemical structures of proteins. The discussed material can be sum- marized in the following manner: (a) The enzymes possessing minimum unit molecular weights of about 15,000, that is the smallest molecular weight entities which cannot reversibly be split or dissociated, are serologically and en- zymatically distinct and different from the other proteins related to them species specifically. (b) Polypeptides with molecular weights of about 6,000 show MECHANISM OF ANTIBODY FORMATION 121 specific combining activities with enzymes, and haptenic qualities. Hydrolytic cleavage products which are dialyzable, having molecular weights ranging from 600 to 1,000, exercise serologically specific activities demonstrated by the whole original molecules from which they are derived. (c) Antibodies of Vi to Vs size of the original molecule show activities comparable to original activities. Antibodies which proteolyt- ically can be reduced to V2 size with a loss of their species relationship to normal proteins, and antibodies which proteolytically can be changed to such forms that represent loss of their antigenic qualities without loss of their antitoxic activities, show clearly that these bio- logical specificities are inherent in their respective chemical structures. It can therefore be inferred that these specificities arise from specific synthetic enzyme reactions. The summation of the specificities of the smaller structural units represent the various specificities of the whole protein molecule. The concept of the formation of various proteins from proteinogen, as discussed at the beginning of this section, does not appear to be capable of accounting for the above enumerated speci- ficities of structural units. In other words, our view is that the S'pecificity of an antibody molecule is the consequence of Sfecific cellular synthetic 'processes catalyticaily modified hy an antigen to conform with the configuration of certain active groups of the antigenic molecule. The change a certain specific protein undergoes from its inactive to its active state, for example, trypsinogen to trypsin, whether by a shift in the pH of the surrounding medium, by an ion effect, by non- specific agents such as concentrated solutions of ammonium or magnesium sulfate, or by the presence of an enzyme to accelerate the spontaneously occurring change, in an autocatalytic manner or other- wise, appear to be a superficial chemical change, and do not involve deep-seated changes to merit the assumption that a new protein is derived from another protein. The enzyme precursor idea does not appear satisfactory as the basis for such fundamental questions pertain- ing to the synthesis of enzymes, antibodies, viruses and the like. The results obtained from immunological and chemical studies appear to contradict the postulate that one protein is autocatalytically formed from another protein, or from a "master proteinogen." 122 IMMUNO-CATALYSIS 8. Anti-Antibodies According to the theories of Breinl and Haurowitz, Mudd, Pauhng, and Alexander, etc., the antigens modify the configuration of the normal globulins during their synthesis, resulting in the formation of specific antibody globulins. The antigens exercise this influence to this extent and apparently not further. The statements by Breinl and Haurowitz to the effect that "the orientation of amino acids in the formation of antibodies . . . "; and the "coupling of amino acids to the 'peptide occurs in an orienting environm-ent, namely, the antigen- protoplasm interface." (Mudd), need not therefore be interpreted to imply that amino acids in an antibody molecule occupy diff^erent positions in the peptide chain than those of normal globulins. It may simply mean, as stated by Mudd, that "the chemical groupings coupled to the molecule undergoing synthesis at the antigen surface must he adapted spatially." The studies and conclusions of Landsteiner and van der Scheer (1928) were instrumental in the formulation of the above concept. They demonstrated that the steric configuration of antigens exercise distinctive influence on the serological specificity of the antibodies against them; 1- and d-antigenic isomers react specifically with the respective antibodies. These findings brought definite proof for the view that the steric configuration of the antigenic groups is one of the factors determining serological specificity. The mere difference in the position of H, OH and COOH in 1- and d-antigenic isomers sufficed to alter the specificity of the antibody. They also suggested that the steric configuration may play a significant part in the serological spec- ificity of carbohydrates such as those discovered in bacterial anti- gens. This view has been amply confirmed by the extensive studies of Avery and Goebel, et al. Since this subject will be discussed in a latter part of this study, it suffices here to mention that the antibody against galacto-albumin reacted with solutions of galacto-globulin and galac- to-albumin. These antibodies, however, failed to precipitate gluco- albumin. The antibody produced against the antigen containing a-glucoside reacted selectively with this antigen, and not with the antigen containing a /3-glucoside group, and vice versa. A difference in the spatial arrangement of the same polar group therefore exercised MECHANISM OF ANTIBODY FORMATION 123 definitive influence on the serological specificity of the antibodies produced. The optical antipodes, as well as space isomers differ chemically and physicochemically, exercising different degrees of attractive and repulsive forces. Atoms or atomic groups falling within the influ- ence of these forces are either repulsed out of the sphere of influence or are attracted to combine. The optical and space isomers thus mani- fest different properties in combining with proteins, functioning as a substrate for an enzyme, in bacteriostatic action, in enzyme inhibiting effect, in forming racemates, etc. During the synthesis of globulin molecules under the influence of antigens containing polar groups, or optical and space isomers, such influences do, no doubt, come into play. It could thus be assumed that during the synthesis of the globu- lin molecules the atomic groups rotate their position in space in ac- cordance with the configuration of the antigenic molecule, resulting in the formation of antibody as modified globulin. a. Presence of Serologically Reactive Basic Amino Groups in Antibody Globulins. Mudd and Joffe (1933) found that when antisera were treated with formaldehyde before combination with anti- gen, agglutination was consistently inhibited. The antibodies treated with formaldehyde showed decreased basicity, which was evidenced by a shift in the isoelectric point of the sensitizing film toward the acid side by about 0.6 to 0.8 pH unit, and reduced agglutinating tendency of the sensitizing film. Chow and Goebel (1935) investigated chemically and serologically the function of the basic amino groups in Type I anti- pneumococcus precipitin. Eighty-eight per cent of the purified anti- body they studied was precipitable by Type I specific carbohydrate. The remaining 12 per cent protein was considered as representing anti- bodies against cellular proteins of the pneumococcus and for the somatic carbohydrate present in the original serum. The isoelectric point of the precipitin was unusually alkaline, i.e., pH 7.6. This was attributed to the presence of either a relatively high ratio of amino to carboxyl groups or to the presence of a higher percentage of basic amino acids in the antibody molecule than the globulin of normal serum would contain. The distribution of basic amino acids in the globulin of normal and immune horse serum, aside from a slightly higher lysine content of the latter, showed no essential difference. Acetylation of antibody revealed the presence of one acetyl group for 1 24 IMMUNO-CATALYSIS every primary amino group in the original antibody molecule. Acet- ylated antibody lost to a great extent its capacity to precipitate the type specific polysaccharide. When relatively high concentrations of type specific carbohydrate were added to the acetylated antibody, a precipi- tate was formed. The original antibody gave H — | — \- precipitate with 1:1,024,000 carbohydrate dilution; in contrast, acetylated antibody gave H — I — h with 1:4000 and + precipitate with 1:8000 carbohy- drate dilution. These results were interpreted to indicate, besides the amino groups, the presence of certain other reactive groups likewise involved in the precipitation reaction. Since formalized antibody protein gave no precipitation test the amino groups in the native antibody were believed to be concerned in the union of the antibody with the carbohydrate. Deformalization of the antibody restored its serological specificity. The loss of serological specificity of the formalized antibody was considered presumably due to the conversion of the -NHg to -N^CHs group. Conversely the esterification of the — COOH group of the antigen carbohydrate de- prived it of its reactivity with the native antibody. Saponification of the esterified carbohydrate fully restored its reactivity. Though the saponified carbohydrate still contained methyl groups (as a result of the treatment of the carbohydrate with diazo-methane in the esteri- fication experiment) bound both on the primary -NH2 group of the parent carbohydrate and as a methoxyl group attached probably to one of the hydroxyl groups, unlike the inert ester, it reacted readily with antisera. These results showed that a free -COOH group in the carbohydrate was of primary importance in rendering it serologically reactive. In other words, the union of the free -COOH group in carbohydrate with a free — NH2 group in the antibody was considered possibly necessary for the serological reactions. The normal globulin molecule does not possess this serologically reactive "free amino group." The presence of a serologically reactive "free amino" group in anti- body globulin and not in normal globulin was also shown by Goebel and Hotchkiss (1937) with substances containing organic acid radi- cals. They found that artificial azoprotein antigens containing organic acid radicals, quite unrelated in chemical constitution to glucuronide or galacturonide, precipitated in antipneumococcal horse sera. Antigens prepared from j^-aminobenzene carboxylic and sulfonic acids, though not reactive in normal horse serum, precipitated in antipneumococcus MECHANISM OF ANTIBODY FORMATION 125 sera. The precipitation of these antigens in Type III immune horse sera was inhibited indiscriminately by the sodium salt of any one of the uncombined acidic derivatives (p-aminobenzene sulfonic acid, |7-ami- nobenzene carboxylic acid, p-aminobenzylglucuronide, |9-aminobenzyl- galacturonide as inhibiting substances against the test antigens prepared from them and chick serum). The above cited substances containing — COOH and — SO3H groups, possess but one property in common, namely, acidic groups of divergent nature. As stated by Goebel and Hotchldss the reaction of these antigens in antipneumococcal horse sera represents a neutraliza- tion of the charge of basic groups of the antibody protein by the acid groups of the non-specific haptens, followed by precipitation. b. The Directive Influence of Optically Active Catalysts in Producing Optically Active Substances. The presence of serologi- cally reactive basic amino groups in antibody globulin may appear to indicate that during its synthesis the position of the basic — NH2 groups had been specifically oriented by the presence of -COOH polar groups of the pneumococcal carbohydrate, or synthetic glucoproteins. In contrast, during the synthesis of normal globulin, the position of the amino groups not being directed by foreign influences, do not occupy any specific position with reference to foreign substances. It can be visualized that, similar to the orienting influence of — COOH groups in the pneumococcal carbohydrates, optically active groups in d- and 1-antigens (Landsteiner and van der Scheer, 1928, 1929) may specifically orient certain groups during the synthesis of the antibody globulin molecule to conform with the optical configuration of the antigens. The question as to whether d- and 1-antigens, acting as optically active catalysts, have produced optically active specific groups in the antibody molecule is of general interest; to this question we have at present, no direct answer. It has been observed by numer- ous investigators that in general, it is characteristic of living cells to metabolize or synthesize only one mirror image of optically active substances. This fact has been known since the time of Pasteur and the literature is rich in confirmatory evidence which we need not dis- cuss here. In relation to the question raised above whether optically active antigens produce antibodies with specifically oriented optically ac- tive groups responsible for the serological reactivity, it might likewise 126 IMMUNO-CATALYSIS be inquired as to whether the cellular enzymes which catalyze one- sided asymmetric synthesis or metabolism (i.e., either 1- or d-isomer) of optically active substances are themselves optically active.* It has been claimed (Helferich, et ah, 1938) that ;8-d-glucosidase and 13-d- fructosidase contain carbohydrate-like substances of glucose nature; and it is postulated that these enzymes by virtue of the carbohydrate groups, perhaps with iS-con figuration, combine with and catalyze mirror image substrates. The active group of the enzyme is assumed to form a racemate with the substrate in accordance with the "Lock and Key" idea of Emil Fischer (Emil Fischer, 1894, 1898; see also Lettre, 1937). These claims and postulates though of utmost theo- retical interest are not as yet experimentally demonstrated. On the other hand, it has been definitely shown that optically ac- tive d- and 1-active organic catalysts orient specifically the configura- tion of the atomic groups on asymmetric carbon atoms produced during the synthesis of optically active substances from optically inactive start- ing materials. By an accurate consideration of these laboratory proc- esses, which take place continuously in the living cell, we might be able to understand at least the principles involved in asymmetric synthesis. t Rosenthaler (1908, 1909, 1913) obtained the first proof that an enzyme in emulsin which he presumed to be optically active could * Pasteur suggested that every asymmetry owes its existence to some asymmetric forces operating at the moment at which the asymmetry appeared. Emil Fischer ( 1 894, 1898) stated that one active molecule gives rise to another, and the formation of sugar in plants takes place by chlorophyll which he assumed to be optically active. The optical activity of chlorophyll has recently been demonstrated by Stoll and Wiedemann (1933). tin evaluating the reports concerning the optical purity of an enzymic system or asymmetric synthesis with a preponderance of one of the antipodes, the following considerations must be kept in mind. From the standpoint of true catalysis there is no difference in the energy required for the interconversion or the formation of the 1- and d-antipode of a given substance in equal concentrations. Analyzing the principles of the asymmetric synthetic action of enzymes Werner Kuhn (1936) stated that from the thermodynamic point of view the optically active state is not a state of equilibrium as compared with the racemic state. The mixing of the two antipodes to form a race- mate liberates energy, while their separation requires an expenditure of energy. According to Kuhn, in certain instances, the preponderance of optical purity in a system, despite the gradual decrease of optical activity in a single enzymatic process, can be explained by "stereo-autonomic" behavior of some substances. That is, this behavior conditions stable optical purity of the synthetic product. It is knovm that d-mandelonitrile, synthesized by the plant, is stabilized in tlie form of the less soluble and easily crystallizable fi-gentiobioside (natural amygdalin). In contrast the more soluble gentiobioside of 1-mandelonitrile remains in solution. The excess 1-mandelo- MECHANISM OF ANTIBODY FORMATION 127 influence the combination of benzaldehyde and hydrogen cyanide so that the reaction took place one-sidedly, the mandelonitrile formed being dextrorotatory. The extent of the asymmetric synthesis must have been considerable for the dextrorotatory nitrile was formed in large excess over its antipode since 1-mandelic acid obtained on hydrol- ysis of the nitrile was optically pure after two crystallizations from benzene. Emulsin CeHgCHO+HCN >C6H5CHCOH)CN d-Mandelonitrile Hydrolysis C6H5CHCOH)CN >C6H5CHCOH)COOH d-Mandelonitrile 1-Mandelic acid This was the first definite example of an asymmetric synthesis carried out in the laboratory by an enzyme. Rosenthaler (1913) found in the leaves of Taraktogenos Bluenci HSSK also an enzyme, called l-oxynitriluse, which brought about the formation of 1-man- delonitrile from benzaldehyde and hydrogen cyanide, as distinct from the common d-oxynitrilase which forms d-nitrile. The emulsin from cherries was found by Krieble (1913) to yield sometimes a d- and sometimes an 1-nitrile. Working with bacteria as an enzyme source Mayer (1926) was able to obtain from phenylglyoxylic acid and hydrocyanic acid 1-mandelic acid by Bacterium ascendens, and d-mandelic acid by lactic acid bacteria. Bredig and Fiske (1912) made similar observations using optically active organic catalysts. They dissolved 50 ml. of benzaldehyde (0.5 mole) in 170 ml. of chloroform (as solvent) and treated the solution with anhydrous hydrocyanic gas (0.5 mole). After one hour at 25°C., nitrile can thereupon be racemized according to the requirements of true catalysis; that is, the d-nitrile originating from excess free 1-nitrile is again transformed into ig-gentiobioside and the process continues until all the 1-nitrile is converted into the less soluble glucoside or d-mandelonitrile. The preponderance of optical purity in an enzymatic sjmthetic process is observed also: (a) when widely different velocities in the formation of the two optical isomers determine the degree of preponderance of the d- or 1-isomer in the substance syn- thesized; (b) an increase of optical purity can be observed when the reaction is in- terrupted before the optical activity disappears as the result of gradual racemization (Langenbeck and Triem, 1936); (c) by the elimination from the living system without utilization (Abderhalden and Samuely, 1906), or by more rapid metabolism of the unnatural optical isomer CKrebs, 1933). 128 IMMUNO-CATALYSIS it was treated with 0.5 g. (0.001 5M) quinine or quinidine alkaloid as catalysts. After 24 hours the reaction mixture was treated with 100 ml. of 4 N aqueous sulfuric acid and shaken for five minutes to remove the catalyst alkaloid. A study of the chloroform solution showed that the alkaloid had acted as a specific directive catalyst in the formation of nitrile. Laevorotatory quinine caused the formation preponderantly of a dextrorotatory nitrile, from which laevo- and ^dextrorotatory man- delic acids were obtained on hydrolysis. In contrast, ^dextrorotatory quinidine gave preponderantly kevorotatory nitrile: CgHsCHO-fHCN 1 -quinine CeHgCHOHCN Cdextro:) Mandelonitrile Hydrolysis CeHgCHOHCOOH, dextro- Mandelic acid, laevo- d-quinidine CflHfiCHOHCN Claevo') Mandelonitrile -48.5% -51.5% Hydrolysis C0H5CHOHCOOH. dextro-54.3% Mandelic acid, Zaevo— 45.7% Analogous results were obtained when cinnamic aldehyde, anis- aldehyde, citral, piperonal and acetaldehyde were used in the place of benzaldehyde. Bredig and Fiske believed that alkaloids and cyanhydrin form a complex compound as a necessary condition for catalytic synthesis. They found that the alkaloids disappeared from the aqueous solution into the chloroform or toluene solution of cyanhydrin; only cyan- hydrins appeared to participate in such a combination. McKenzie (1936), commenting on the findings of Bredig and Fiske, cited Isobel A. Smith's interpretation of the findings of Traube and Onodera (1914) who found that alkaloids of high molecular weight, such as atropine or quinine, formed colloidal solutions in water, while their salts formed true solutions. Accordingly it appeared to Smith that MECHANISM OF ANTIBODY FORMATION 129 herein in all probability lay the asymmetric synthesis; in the chloro- form solution these alkaloids formed colloidal adsorption compounds acting in a capacity very similar to that of an enzyme. In this connection it may be mentioned that Bredig and Gerstner (1932) made the interesting observation that by introducing the di- ethylamino group into cotton fibre the latter was rendered catalyti- cally active in effecting the synthesis of 1-mandelonitrile from benzal- dehyde and hydrogen cyanide. Rosen thaler (1909) extending the scope of his observations later obtained also 1-, d-, and i- (inactive) forms of nitrile from numerous aldehydes and hydrocyanic acid. Acetaldehyde, isobutyr aldehyde, heptaldehyde, furfural, o-methoxybenzaldehyde, anisaldehyde, cin- namic aldehyde, etc. yielded d-nitriles; only citral and o-phthalylal- dehyde yielded 1-nitrile; chloral, salicylaldehyde, m- and p-oxybenzal- dehyde, p-nitrobenzaldehyde, methyl-ethylketone yielded optically inactive nitriles. These results were interpreted to indicate that either the emulsin preparation contained more than one optically active enzyme, or there exists a certain type of substrate specificity in the synthesis of stereochemical isomers. The results obtained by Bredig and Minaeff (1932) with organic optically active catalysts in some respects were analogous to the above observations made by Rosen- thaler (1909). Table III CBredig and Minaeff) Catalysts Synthetic product d-Quinidine 1-Quinine Cinnamic Aldehyde Cinnamic Aldehyde Anisaldehyde Citral Piperonal Acetaldehyde present present present present present present l-rotatory d-rotatory l-rotatory d-rotatory d-rotatory l-rotatory The above results show, as Rosenthaler found, that the specificity of substrate may likewise play a role in defining the optical configura- tion of the final reaction product. 1 30 IMMUNO-CATALYSIS Granting that d- and 1-active antigens might catalyze the synthesis of homologous antibodies, comparable to the catalytic synthesis of d- and 1-active substances respectively by 1- and d-active organic cata- lysts, it would then appear reasonable to expect that hypothetically d-antigen catalyzes the synthesis of 1-antibody and 1-antigen that of d-antibody. The serological reaction between the optically active anti- gens and the homologous antibodies would then be comparable with the formation of a racemate such as d-cinchonine-1-tartrate. This racemate is a relatively more insoluble molecular combination than either of the components, which fact might be compared with the formation of an insoluble precipitate by the reaction between the antigen and antibody. The protein molecule is made of optically active amino acids which have the 1-configuration. Lettre (1937) postulated that in the anti- body molecule the serologically determinant group as the prosthetic group is the mirror image of the naturally occurring 1-active groups of the normal globulin molecule. A simple protein acting as antigen would produce accordingly an antibody containing as many d-active prosthetic groups as there are antigenic 1-active groups in the simple protein. Though this concept might appear to account for the formation of hypothetical d-active antibodies against 1-active antigens, difficulty arises when we attempt to explain the nature of the active groups in the antibody molecule produced against d-antigens. Since the antibody against the latter would hypothetically be expected to have the 1-configuration, it is difficult to conceive that there can be two kinds of 1-active globulins to account on one hand for the serological reactivity of the antibody against the d-antigen, and a second one corresponding to normal globulin made of 1-active amino acids on the other. Antibody Carbohydrate as a Possible Seat of Specificity. Perhaps it is pertinent to the question discussed above to consider the possible role of the carbohydrate contained in the antibody globulin on the specificity of serological reactions. As far as we know, this question has never been raised or studied before. It has been shown by Hewitt (1938) that normal serum globulins contain chemically bound carbohydrate groups. Working with normal serum fractions which were distinguishable by biological, chemical, and physical methods he found that seroglycoid, globoglycoid and pseudoglobulin, respectively, MECHANISM OF ANTIBODY FORMATION 131 contained 5.6, 6.2 and 2.2 per cent galactosemannose and 2.7, 2.9 and 1.1 per cent glucosamine. Euglobulin, pseudoglobulin, globoglycoid and seroglycoid, respectively, contained 4.1, 3.3, 9.3, and 8.4 per cent polysaccharide. Remington and Van der Ende (1940) reviewed the problem by pre- paring crystalline albumin, crystalline globoglycoid, and seroglycoid and seromucoid from horse serum. According to them seromucoid contained 10.7 per cent of a polysaccharide consisting of N-acetyl- d-glucosamine, d-mannose and d-galactose in equimolecular pro- portions. Crystalline albumin and globoglycoid are essentially identical and differ from seroglycoid. The latter, although closely related to seromucoid, was not identical with its immunological properties. The purified antitoxic pseudoglobulin contained 2.6 per cent of carbohydrate calculated as mannose-galactose-glucosamine. Peterman and Pappenheimer (1941) showed that after digesting diphtheria anti- toxic material with pepsin the carbohydrate moiety remaining with the antitoxic portion of the molecule amounted to 3.8 per cent. Northrop (1942) working with crystalline diphtheria antitoxin showed the pres- ence of not less than 2 per cent carbohydrate as part of the molecule. Immune y-globulins from human, horse and bovine sources were analyzed by Smith, Green and Bartner (1946) and were found to contain from 1.26 to 1.5 per cent hexamine. The ratio of hexamine to hexose was 1 to 2. During the synthesis of antibody globulin resulting in the orienta- tion of various groups in accordance with the configuration of the antigenic molecule one could expect that the polar groups in the carbohydrate moiety of the molecule may likewise undergo similar orientation. The specificity of the antibody wdth reference to the optical configuration of the homologous antigen might in particular be a property of the carbohydrate group of the antibody molecule. In this connection the important question is whether or not d- and 1-active, or a- and /3-antigens produce, respectively, 1- and d-active groups, or groups with a- and ^^-configurations in the antibody carbo- hydrate group, in a manner comparable to the action of d- and 1-ac- tive organic catalysts discussed above. At present we have no answer to any one of these questions. c. Experiments Dealing Directly with the Question of Anti- Antibody Formation. If the synthesized antibody molecule possesses 132 IMMUNO-CATALYSIS a stereochemical conespondence with the antigen, as appears to be the case from various studies, for the reason that its spatial configuration has been oriented to fit the spatial configuration and the affinities of the polar groups and of antigenic optical antipodes and space isomers, it would appear reasonable to expect that the differences between anti- bodies against 1- and d-antigens are sufficiently great to render them antigenically distinguishable. At least it would seem that these anti- bodies should spatially be sufficiently different to make them antigeni- cally different from the normal globulins. Pauling's concept of the antibody molecule that "only in the con- figuration of the chain, that is, in the way that the chain is coiled in the molecule," could not, in our opinion, be viewed simply as a physical change not involving a change in the spatial position of the atoms or groups. For it is stated by Pauling that the atoms and groups which form the surface of the antigen will attract certain complementary parts (positively and negatively charged groups). This may imply that various groups in the globulin molecule may undergo spatial rearrangement as a consequence of the said polar property of groups to yield a new configuration in the coiled chain end of the molecules. It would therefore appear that if a difference in the position of H and OH in R-a-glucoside or R-a-galactoside antigens and those in d- and 1-tartranilic acid antigens are capable of producing specific antibodies, it is reasonable to ask why possible changes in the spatial arrangement of the polar groups, NH2, COOH, OH, CONH, SH, largely responsible for the chemical reactivity of the proteins, should not be considered also as possible factors in the formation of anti- antibodies against these antibodies. To obtain an answer to the question whether or not anti-antibodies could be produced numerous investigations have been carried out since the time of Ehrlich. The results, however, have not been of a defini- tive nature to effect a clear cut differentiation between the antigenicity of normal and antibody globulins derived from a given species of animal. The reason for failure to do so apparently is to be found, as recent studies have shown, in the fact that the antibody globulin molecule is composed in part of an inactive portion which is in- distinguishable from normal globulin. The only clear proof that diphtheria antitoxic activity may be exhibited by a protein which is antigenically distinct from normal globulin is provided by Northrop MECHANISM OF ANTIBODY FORMATION 133 (1942) who isolated the antitoxin in crystalline form after enzymic digestion of the original complex antitoxic globulin and by eliminating the inactive component. Despite the undecisive nature of the results of the investigations carried out previous to that of Northrop it might not be in vain to present briefly the nature and the type of these experiments. It is to be noted that because of its theoretical importance this question has been an object of study by older immunologists, namely, Bordet, Ehrlich, Sachs, Morgenroth, Pfeiffer, Friedberger, etc., and immunol- ogists of our times, Landsteiner, Heidelberger, Marrack, and others and their collaborators. Ehrlich and Morgenroth (1901) carried out cross immunization experiments with ox, goat and rabbit blood cells, and then, with the adsorption technique, they arrived at the conclusion that plural im- mune bodies are produced by injections of ox and goat blood. By eff^ecting differentiation of various groups of immune bodies by means of the anti-immune bodies they believed they had shown the formation of anti-antibodies. Bordet (1899) demonstrated the formation of anti-hemagglutinins (anti-amboceptors). Fowl's serum injected intraperitoneally into a rabbit produced an anti-immune body which protected rabbit red blood corpuscles against the hemolytic action of fowl's serum. In a subsequent extensive study Bordet (1904) observed that red blood corpuscles of diff"erent species of animals, sensitized by appro- priate hemolytic sera obtained from a species B, lose their sensitivity to complement when treated with the anti-amboceptor. In certain in- stances it was not necessary to use for immunization the serum con- taining specific amboceptor; it was sufficient to inject the animal, namely the guinea pig, with normal serum of species B, which Bordet stated, contains normal amboceptors. (See also Muir and Browning, 1906). Friedberger and Moreschi (1908) stated that in their earlier ex'peri- ments they believed they had demonstrated the formation of anti- amboceptors by immunizing goats with hemolytic immune rabbit serum against goat red corpuscles. In later experiments, contrary to their expectations, when the serum of a rabbit immunized with goat's serum hemolytic for rabbit corpuscles was added to rabbit red corpuscles sensitized wdth goat amboceptor the hemolysis in the presence of 1 34 IMMUNO-CATALYSIS complement was not prevented. On the contrary, they observed an acceleration of hemolysis. In a subsequent study Moreschi (1908) observed that red blood corpuscles, which were sensitized with a non-agglutinating dose of a corresponding immune serum and w^ashed free of excess serum, treated with a small amount of a precipitin containing serum underwent readily a strong agglutination. He also observed the same effect in bacterial agglutinations. Altman (1912) using sensitized and washed red blood cells as immunizing agents obtained results similar to those of Friedberger and Moreschi (1908). The formation of anti-antibodies capable of neutralizing the lytic action of antibacterial immune sera were reported by Pfeiffer and Friedberger (1903). It was, however, immaterial whether the rabbits were immunized with normal or artificially immunized goat serum. Pfeiffer and Friedberger interpreting their results were inclined to as- sume that various antibodies supplied by the same species of animal, speaking in Ehrlichian terms, possess a common group, which char- acterized the species of animal from which it originates, and that the anti-immune-serum in some way must be related to this group. Dehne and Hamburger (1904) had shown that normal horse serum precipitinogen as a normal constitutent of horse serum is closely as- sociated with the horse antitoxins. Kraus and Pribram (1905) re- ported that horse sera containing a high agglutinin titer for typhoid bacilli were completely absorbed out by anti-horse rabbit serum. Ex- periments with anti-cholera immune horse serum yielded similar re- sults. The anti-bacterial horse antibodies and normal horse serum precipitinogens were shown to possess a common combining group. Landsteiner and Prasek (1911) immunized rabbits against goat serum. The immune serum reacted strongly with goat serum. The immune rabbit serum inhibited goat agglutinins, particularly the bacterial agglutinins. These facts indicated that precipitins acted as anti-agglutinins. d. Non-Identity of the Combining Sites for Antigen and Anti- body in an Antitoxin Molecule. Eisler (1920) found that horse serum antitoxin (tetanus, diphtheria) could be precipitated by a rabbit immune serum against horse serum even after combination with toxin. Smith and Marrack (1930) showed that antitoxin and toxin-antitoxin floc- cules react like pseudoglobulin with anti-pseudoglobulin serum. They MECHANISM OF ANTIBODY FORMATION 135 Stated that since antitoxin, when precipitated by a precipitin, can still combine with toxin, different groups must be involved in the two re- actions. Eagle (1936) stated, as previous investigators had shown, that a molecule of antibody can function either as an antibody or as an antigen. In the former capacity, it combines with homologous anti- gen; in the latter capacity the antibody molecule is itself precipitated by an antibody to serum protein. Experimental details of Eagle's studies differ from those of Eisler, Smith and Marrack in that the latter did not saturate antitoxin vdth toxin or precipitin prior to the addition of precipitin or toxin. In Eagle's experiments horse antitoxin precipitated by a large excess of a rabbit antiserum against horse serum and presumably saturated with precipitin, nevertheless retained almost its original neutralizing activity. Antitoxin so completely saturated with toxin as to form a highly toxic compound combined with a rabbit pre- cipitin against horse serum. From these and numerous other serological data, it is to be seen that antibody globulin molecules possess specific reactive groups not present in the normal globulin molecules. e. Further Experimental Data Concerning the Presence of a Com- mon Group in DiflFerent Antibodies from a Species of Animal. Ando and his associates (1937, 1938) immunized rabbits, ponies and horses with diphtheria toxin-antitoxin flocculi, and with the specific precipitates of pneumococcal type specific carbohydrate and its antibody in its purest form. The serological tests were carried out by the optimal- proportions-method of Dean and Webb and the absorption technique. They found that the precipitin-versus-antitoxin absorbed diphtheria antitoxin in the range of optimal proportions and also at higher ratios. Pneumococcal antibody, however, was absorbed by precipitin-versus- antitoxins only within the range of the mixtures containing far less amounts of anti-pneumococcal serum than that in the optimal propor- tion. Precipitin-versus-antibacterial antibody absorbed anti-pneumococ- cal antibody, but diphtheria antitoxin was absorbed by the same precipitin only in the secondary zone and not in the primary or main zone. These results were interpreted by the authors to signify that various antisera contain larger amounts of antibody against the homol- ogous than the heterologous immunizing flocculi. In these studies as indicated above the horse antisera against different flocculi reacted vdth homologous as well as heterologous antisera. Treffers and Heidelberger (1941) examined the properties of spe- 136 IMMUNO-CATALYSIS cific precipitates both as immunizing and as test antigens to deter- mine quantitatively the chemical relationship, not only of antibodies to the different serological types of pneumococcus, but also of antibodies directed against a somatic component, the so-called C polysaccharide of the pneumococcal cell. A specific precipitate derived from anti-pneumococcus Type II horse serum was used as antigen for the injections of rabbits and the result- ing sera were tested against the specific precipitates obtained with: (a) Pneumococcus Type I, II and Group C carbohydrate and ho- mologous horse sera; (b) H. Influenza Type B polysaccharide and anti-influenza Type B horse serum; (c) Diphtheria crude toxoid and horse diphtheria antitoxin; (d) Crystalline egg albumin and homologous horse antiserum; and (e) Specific precipitates from Type II anti-pneumococcal rabbit serum and from Type I anti-pneumococcal pig serum. The results showed that "after absorption with egg albumin-anti- egg albumin (horse) specific precipitate and removal in this way of roughly one-half of the antibody the anti-specific precipitate rabbit serum still resembled the unabsorbed serum in its quantitative re- activity toward a high ratio pneumococcal specific precipitate when compared at the same antibody content. "The co-existence of separate groupings which function as antigen and antibody on the same molecule of horse globulin is indicated by experiments showing pneumococcus Type II horse antibody precipi- tated with an excess of anti-specific precipitate rabbit serum can still combine with Type II specific carbohydrate, and that the horse anti- body in a solution of the egg albumin-anti-egg albumin (horse) specific precipitate in an excess of egg albumin still reacts with the anti-specific rabbit serum." Tests were carried out to examine whether the antibody specificity would affect in any way their function as test antigens. Neither by the sensitive quantitative agglutinin method nor by passive anaphylaxis tests in the guinea pig could any variation be demonstrated due to the specific antibody groupings. These and other findings led the authors to the conclusion that the antigenic reactions of this representative water-insoluble group of antibodies engendered in the horse are es- sentially similar. These findings confirm in a more rigorous quantita- MECHANISM OF ANTIBODY FORMATION 137 tive manner the conclusion of previous investigators that the various specific antibodies in a species of animal are closely related anti- genically. Thus the only antigenic specificity demonstrable for the antibodies investigated was that due to their common origin, and "the groupings/' as stated by the authors, "responsible for their antibody function, constitute either a small part of the total protein molecule or else are non-antigenic." f. A Comment on the Use of Antigen-Antibody Complex for the Production of Anti-Antibodies. According to our present concept the determinant groups in antigenic molecules which specifically stimulate the production of antibodies in vivo are identical with the groups which combine with the antibodies produced. It would therefore be expected when these groups of antigens are saturated with antibodies the treated antigens would be deprived of their capacity to develop specific antibodies. Olitzki (1935) reasoning in this manner reviewed the previous studies concerning this question and found it controversial. He then set out to clarify this controversy. He reported that the immunization of rabbits with E. ty-phosa sensi- tized with anti-serum to 0-antigen did not reduce the production of H-agglutinin, but reduced the production of 0-agglutinin to nearly 20 per cent of the normal immunization effect by an unsensitized vaccine. Sensitization of the vaccine by H-serum reduced only the production of H-agglutinin to 20 per cent of the normal immunization effect. Sensitization of the vaccine by both H- and O- serum reduced the production of both types of agglutinins. Olitzki found that the injection of sensitized antigen together with larger amounts of free antibody (immune serum) suppressed the formation of agglutinins completely. The effect was the same whether the excess serum was administered before or after the inoculation of the sensitized vaccine. Phage lyzed bacteria produced at least the same quantities of anti- bodies as whole bacteria. But after sensitization with the same quantity of immune serum, phage lyzed bacteria were completely deprived of their ability to produce antibodies, while the phage untreated bacteria remained effective. It is interesting to note in the results of OHtzki that after injection of the same quantity of free immune-agglutinins as used for sensitiza- tion the free agglutinins could be recovered in the serum. Especially the H-agglutinins were demonstrable in the serum on the next day 138 IMMUNO-CATALYSIS after the injection and persisted there during a period of ten days. However, the same amounts of antibodies injected after being fixed by the antigen do not appear in the serum. It is evident from the above results that the combining groups of an antigen, when free, are able to produce in vivo antibodies with specific combining groups. When the combining or antigenic groups of a mole- cule are saturated with antibody and kept saturated in vivo with excess antibody they are unable to incite the formation of antibodies. Ap- parently the antigen -f- antibody ;=^ antigen-antibody complex equilib- rium is completely shifted to the right in the presence of an excess amount of antibody. The question may therefore be asked whether the reverse process does or does not take place. Namely, if the combining groups of an antibody are saturated with antigen the former may fail to incite the production of specific anti-antibodies, or antibodies to the specific configuration or combining groups of antibodies involved in the anti- gen-antibody reaction. The saturation of the combining groups of anti- bodies would not block the species specific antigenicity of the antibody globulins which would account for the formation of antibodies (when combined with antigen) which are reactive with normal globulins from the same species. In view of the above consideration one may question the validity of the conclusions drawn concerning the absence of anti- antibody formation in experiments involving the use of the above mentioned antigen-antibody complexes (Ando and associates, 1937, 1938; Treffers and Heidelberger, 1941) where combining groups of both reactants are mutually saturated, and therefore are incapable to negotiate the reactions required for the formation of anti-antibodies. g. Acquisition of New Antigenic Specificity by Antitoxins. The question of whether or not antibody has acquired a new specificity antigenically distinct from the corresponding normal globulins has been investigated by various workers. Weil, Parfentjev and Bowman (1938) made the following observations: (a) antibody response in rabbits, im- munized with pepsin-treated antitoxin, is slight and delayed; (b) tests for cutaneous hypersensitivity, and precipitin and complement fixation tests all showed that the treatment with pepsin had eliminated the original antigenicity of the antitoxin; and (c) guinea pigs sensitized to pepsin-treated antitoxin were shocked by normal horse serum, but not by the homologous antigen. The failure to shock the guinea pig MECHANISM OF ANTIBODY FORMATION 139 by pepsin-treated antitoxin has been suggested by other investigators to be due to the hmited amount of antigen that could be injected. Coghill, et al. (1940) reported that diphtheria antitoxin which had been despeciated by treatment with taka-diastase did not kill guinea pigs sensitized passively to normal horse serum, and only a few despeciated sera gave any anaphylactic reaction at all. The guinea pigs sensitized with serum treated with taka-diastase could be shocked but only with large doses of the same antigen. Kass, Scherago and Weaver (1942) reported that ileum segments from guinea pig sensitized passively to normal horse pseudo-globulin failed to respond to sera digested vdth enzyme. Christensen and Kerrick (1948) confirmed this observation and found that guinea pigs sensitized to pepsin-treated serum will react when tested with the same antigen, but also show that this cannot be due to the remaining horse serum specificity, because pepsin-treated serum is significantly more effective in eliciting anaphylactic shock in guinea pigs sensitized to pepsin-treated serum than in guinea pigs sensitized to normal horse serum. They concluded that "pepsin-treated antitoxin not only has retained some of its original horse serum specificity, but in addition as a result of the process of purification, has acquired a new antigenic specificity." In this study there is not a detailed description of the purification of the pepsin- treated antitoxin, and, therefore, the presence of normal serum com- ponents is not excluded. Neither is there any evidence to show that treatment of normal antitoxic globulins with pepsin accords new specificity to these serum components. The new antigenicity which these investigators have observed does not, therefore, appear to be due to peptic digestion, nor to the process of purification, but rather to the nature of the original antitoxin molecule. h. Crystalline Diphtheria Antitoxin Antigenically Distinct from Normal Serum Components. The inability of the previous investi- gators to differentiate the antigenic specificity of the antibody from that of normal globulin may perhaps be attributed to the fact that even the antibodies precipitated with their homologous antigens contain as integral portions of their molecules components indistinguishable from normal serum globulins. It has been shown by Parfentjev (1936, 1938), Pope (1938, 1939), and others that treatment of diphtheria antitoxin with pepsin under certain conditions splits the antitoxin into an active and an inactive portion. This observation has been 1 40 IMMUNO-CATALYSIS substantiated by Pope and Healey (1938, 1939) with ultracentrifugal experiments. Pope further showed that after treatment of antitoxic pseudoglobulin with pepsin at pH 4.2, the antitoxic product is no longer coagulable at 58 °C., whereas the inactive split component is completely coagulable at this temperature. Peterman and Pappen- heimer (1941) studied the physico-chemical properties of an antitoxic pseudoglobulin preparation and found it to be homogeneous by sedi- mentation, electrophoresis and diffusion criteria. Its molecular weight (184,000) was nearly the same as that of normal horse pseudo- globulin, but its mobility was different from those of any of the protein components of normal serum. This antitoxic pseudoglobulin contained 86,000 antitoxic units per gram, and was only 43.5 per cent specifically precipitable by toxin. After digestion with papain at pH 4.2 and the coagulation of the split product at 58°C. the antitoxic material re- maining in the supernatant was almost homogeneous. Its molecular weight now was 113,000 and it contained 135,000 antitoxic units per gram. In a comprehensive study Northrop (1942) reported the isolation of crystalline diphtheria antitoxin from trypsin digested antitoxic material. The crystalline antitoxin has a molecular weight of 90,000, which is very nearly one-half of the molecular weight of the original antitoxic horse serum pseudoglobulin, and smaller than the antitoxin obtained by Peterman and Pappenheimer. The antitoxin of Northrop was strictly homogeneous in the ultracentrifuge with a sedimentation con- stant of 5.7X10~^^ (Rothen, 1942). The material showed only one boundary in the electrophoresis cell at pH 7.3 or 3.0. It was 90 per cent or more precipitable by diphtheria toxin and had about 700-900 anti- toxic units per milligram protein nitrogen by the flocculation test and about 700 units per milligram by the animal test. Immunologically the crystalline antitoxin was found to behave as an antigenic entity distinct from the normal horse serum proteins (Ten Broeck). The serum of a rabbit immunized against normal horse serum gave a precipitate with 1/4000 ml. normal horse serum (con- taining about 0.002 mg. of protein nitrogen) but gave no precipitate with 1 ml. of a solution of purified antibody containing 1/10 mg. of protein nitrogen. Guinea pigs sensitized by the subcutaneous injection of purified antibody containing 0.003 mg. N gave a typical anaphylactic reaction MECHANISM OF ANTIBODY FORMATION 141 three weeks later when antibody protein containing 0.05 mg. N was injected intravenously. Similarly sensitized guinea pigs failed to react to normal horse plasma diluted to 1:10, but four out of six reacted to normal undiluted plasma containing 0.5 mg. of protein nitrogen, which probably, as stated by Northrop, indicates the presence of minute amounts of normal protein in the purified antibody preparation. Part 111 Antibody as a Specific Enzyme Inhibitor THE THEORY formulated in this treatise considers humoral immunity as resulting from a biocatalytic process. This theory links the anti- gens with enzymes. As discussed in Part I, the antigenicity of all the simple and conjugated proteins is due to the catalytic activity of the pro- tein molecule proper.* Similarly the catalytic activity of simple protein enzymes such as pepsin, trypsin, urease, d-ribonuclease, etc., and that of conjugated protein enzymes such as heme-, copper-, alloxazine-, thiamine pyrophosphate- and pyridine-proteins is dependent on the protein molecule proper. The prosthetic groups fer se (haptens of con- jugated antigens, and coenzyme groups of conjugated protein enzymes) are incapable of catalytic activity either as antigens or as enzymes. In both instances the protein molecule, as part of the antigen or the enzyme must therefore be looked upon as the basic biocatalytic unit. Since 'practically all -proteins are antigenic the conclusion appears to he inescapable that all proteins are endowed with catalytic activity of the particular kind under discussion.^ There is no doubt that there will be arguments raised against the inclusion of antigens among the enzymes. Such protests, however, must *By definition an antigen must exhibit two properties: (1) the power of stimulat- ing the production of antibody in vivo; (2) the power of combining specifically with this homologous antibody. Antigenicity has been ascribed to materials which give in vitro precipitation of complement fixation (Wassermann "antigen") but do not produce in vivo measurable antibodies. This loose usage of the term "antigen" may cause some confusion. In this treatise the term "antigen" implies the original defini- tion as given above. There may be non-protein substances which may prove to be truly antigenic, e.g. acetylated polysaccharide. If this does occur it is difficult to know whether the anti- genicity is due to this polysaccharide alone or due to the combination of this poly- saccharide and an in vivo protein to form a complete antigen. tBergmann (1938) stated that "The essential substances transmitted from one generation of cells to the next must be enzymes, and that they have to be enzymes gifted with the capability of synthesizing individual proteins by predetermined sequence of specificity reactions. . . . Therefore the proteinases owe their existence 143 1 44 IMMUNO-CATALYSIS find their answer in the concept as to what constitutes an enzyme. If we confine the term enzyme to certain famiHar substances of biological origin capable of catalytically breaking down in vitro or in vivo certain selected substances, or capable of synthesizing higher complexes out of simple substrates, we will be setting artificial barriers or placing the known enzymes as an exclusive class behind closed bars. If on the other hand, we define an enzyme (or an antigen) as any protein capable of performing a specialized physiological function in accord- ance with the well known criteria of ideal catalysis, a comprehensive theory of biocatalysis is provided which links antigens, enzymes, vita- mins and hormones* and possibly other still unknown substances of similar role. The knowledge gained from the study of one family of biocatalysts will be useful in elucidating the mechanism of certain un- explained processes associated with other biocatalysts. A. NATURE OF THE ANALOGY BETWEEN IMMUNE AND ENZYME REACTIONS For over half a century immunologists have observed an analogy between immune and enzyme reactions. This analogy has been based on the fact that immune bodies react with antigens or their parts (haptens) with a high degree of specificity which favorably compares with the specificity exhibited between enzymes and their substrates. The origin of the above cited analogy between immune and enzyme reactions goes back to as early as 1890. During this period notable ad- vances were made in the study of the specificity of immune reactions, chemistry of proteins, carbohydrates, and the enzyme reactions by Emil Fischer, Behring, Ehrlich, Bordet, Calmette, Kitasato and many other investigators. In 1890 Behring discovered that the serum of an animal immunized against diphtheria was capable, when injected into a fresh animal, of conferring immunity upon the latter, which, failing the use to the preexistence of other proteinases. There is, in Hfe, a practically endless sequence of sequence reactions, in which one proteinase synthesizes the next by a predetermined reaction, and so forth. The sequence breaks off whenever a proteinase has sjTithesized a protein that does not possess enzymatic properties." The definition, as suggested by us, embracing the versatile properties of the protein molecule is not in strict agreement with the implications suggested by the last sentence of the above quotation from Bergmann. *The reader is referred to an excellent review on the studies of the antigenic properties of hormones by Thompson (1941). ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 145 of immune serum, died from the effects of diphtheria toxin it received. Similar results were obtained with tetanus by Behring and Kitasato. In 1891 Ehrlich treated animals with increasing doses of ricin and abrin and found that the toxin was neutralized m vitro when added to the treated animal's serum, proof of neutralization being offered by the fact that when certain proportions of toxin and immune serum were mixed in vitro, these mixtures were innocuous when injected into animals. He proved that the neutralizing action of immune serum upon each of these toxins was specific, that is, the antiserum for abrin did not neutralize ricin, and vice versa. To Ehrlich was also reserved the elucidation of the nature of the acquired immunity to toxins of snakes through the formation of antitoxins in the bodies of the toxin- treated animals, though the immunization against the toxins of snake venom had already been practiced by Sewall (1887), Calmette (1894), and by Eraser (1895). These observations led Ehrlich to the conclusion that "the power of toxin to combine with antibodies must depend upon a specific atom- group in the toxin-complex possessing a maximal specific relation to definite atom-groups of the antitoxin complex, so that it rapidly unites therewith, like a lock and key," a figure borrowed from Emil Fischer in describing the action of specific ferments.* The observed striking analogy between the specificities of immune and enzyme reactions must have made a deep impression on the minds * Pasteur found that Penicillium glaucum decomposed d-tartaric acid and left the 1-tartaric acid intact. E. Fischer (1894, 1898) observed that yeast fermented d-hexoses, i.e., glucose, mannose, galactose and fructose, but not the 1-hexoses; a- and ^-methyl- d-glucosides were hydrolyzed by enzymes; in contrast a- and yg-methyl-l-glucosides were not attacked. Whilst a-methyl-d-glucoside was hydrolized by a-glucosidase (maltase) only, /J-methyl-d-glucoside was attacked by emulsin. ^-d-Glucoside was hydrolyzed by /J'd-glucosidase, and /J-d-galactoside by yg-d-galactosidase. This was explained by the fact that glucose and galactose differed stereochemically from each other at the 4-carbon atom. These facts led E. Fischer to believe that there existed a specific enzyme for each substrate. Maltose was believed to be hydrolyzed only by maltase, and in fact, glucoside, disaccharide, trisaccharide and polysaccharide were each believed to be attacked by a specific enzyme. The specific behavior of an enzyme towards the optical antipodes he believed to be conditioned by the optical configura- tion of the enzymes, which was compared with the lock-and-key system; that is, it can function only when it fits the lock. Recent observations, however, have shown that a-glucosidase of bottom beer yeast is capable of hydrolyzing a-methylglucoside and maltose (a-glucosido<|4-glucose) (Willstatter, Kuhn and Sobotka, 1924). Likewise saccharase, hydrolyzing saccharose, is capable of hydrolyzing the fructose linkage of raffinose (/j-d-fructose<; >a-glu- cose^a-galactose) (Kuhn, 1923). Emulsin from sweet almond hydrolizes yQ-d-glucoside and its 6-brom-hydrin derivative, ^-d-isorhamnoside, yg-d-xyloside, jg-d-galactoside 146 IMMUNO-CATALYSIS of the investigators of that early period and must have stimulated research along these lines. The early literature would seem to reveal the fact that these investigators experimented on the theory that the antitoxins played the role of specific enzyme inhibitors. For during this period, immunologists and enzyme chemists discovered numerous "normal" antitoxins, hemagglutinins and hemolysins, as well as natu- rally occurring enzyme inhibitors. Along with these findings, numerous investigators concerned themselves with the production of antibodies against known enzymes, using enzymically active preparations as antigens. They are: anti-emulsin (Hildebrandt, 1893), antiserum in- hibiting the liquefaction of gelatin by staphylococcus and B. 'pyo- cyaneus (Von Dungem, 1899), anti-rennin (Morgenroth, 1899; Kor- schun, 1902), anti-cyanurase (Morgenroth, 1900), anti-trypsin (Achalme, 1901), anti-coagulin (Wendelstadt, 1901; Bordet and Gen- gou, 1901), anti-inulase (Saiki, 1907), antiserum inhibiting the pro- duction of pigment by fyocyaneus (Gheorghievsky, 1899), anti-hlastic immunity against anthrax bacillus (Ascoli, 1908), anti-serum inhibit- ing the pneumococcal enzymes (Dochez and Avery, 1916), and anti- serum inhibiting the formation of methemoglobin by pneumococcus (Cole, cited by Dochez and Avery, 1916). The above citations show that during the period of 1890-1907, along with the pioneering observations on the toxin-antitoxin reactions, there was considerable activity in the study of anti-enzyme immunity. Un- fortunately researches in this direction slowed down to such a negligible rate that until a decade ago investigations reminiscent of the activity of the earlier workers are rarely encountered. There has lately come about a renewed interest in anti-enzyme immunity due, probably, to the isolation of known enzymes in crystalline form. The recent findings with crystalline enzymes have established, beyond doubt, the antigenic- ity of enzymes. In fact the antibody against urease has been shown and a-1-arabinoside. There is no evidence as to the presence of individual enzymes for each of the above substrates. A ^-d-glucosidase is beheved to be responsible for the hydrolysis of all of them (R. Weidenhagen, 1932; Helferich, 1933). According to Weidenhagen (1940) there are only glucosidases, the principal spec- ificity of which is confined to the constitution and configuration of the glycosidic sugars. The nature of the glucosidic pairs, whether sugar or aglycon, for the hydrolytic process as such are unessential, and exercise only a relative degree of specificity, par- ticularly on the rate of hydrolysis. For instance, the ratio of speeds of hydrolysis between phenyl-^Q-d-glucoside and phenyl-jg-d-xyloside is 150 to I; that of saligenin- ^-d-galactoside to methyl-/3-d-galactoside is 60 to 1. It is stated by Weidenhagen that "there is no special key for each lock but a master key for a group of locks." ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 147 by Sumner and his associates to affect urease in a manner comparable in every respect to those exercised by antitoxins on toxins of various origin. 1. Serological Specificity of Stereoisomeric Conjugated- Protein Antigens Since the statement made by Ehrlich that the specificity of toxin- antitoxin reactions is analogous to the specificity of fermentation re- actions there has not been any systematic study to define clearly what this analogy consists of. During the last two decades the investigations with artificial conjugated antigens have yielded valuable data which have defined satisfactorily the role of determinant (prosthetic) groups on the specificity of the reactions with these antigens. Pioneering work in this field has been carried out by Landsteiner and his associates, and Avery, Goebel, and associates. Landsteiner (1936) formulated his findings as follows: There are striking differences among aro- matic compounds, isomeric with respect to the position of substitu- ents in the benzene ring, and the gradation in their cross reactions is to be taken as a definite indication that spatial structure, as well as chemical constitution in the ordinary sense of the word, plays an im- portant role in serum reactions, "as had been shown for enzymatic processes by E. Fischer, and frequently confirmed. Hypothetically, Fischer's conception had been applied by Ehrlich to the investigation of serum reactions." On immunizing animals with azo-proteins of d- and l-phenyl-(para- aminobenzoylamino)-acetic acids: H H ~\ I I Protein— N = N—< >— C— N— C O COOH d-isomer H COOH Protein— N = N—< ^C— N— C O H 1-isomer 148 IMMUNO-CATALYSIS Landsteiner and van der Scheer (1928) obtained two different anti- sera which distinguished between the two isomeric antigens, demon- strating that a change in the spatial arrangement of atoms or radicals linked to one asymmetric carbon atom had resulted in the alteration of serological specificity of antigens. They also demonstrated that d- and 1-para-aminotartranilic acids coupled with proteins stimulated the formation of antibodies which reacted distinctively with stereoiso- meric antigens. There was practically no cross-reaction. The 1- and d- immune sera gave rather weak group reactions with the meso-antigen. The m-immune serum gave practically no group reactions against 1- and d-antigens. Tartaric acid anti-sera also reacted with d- and 1-maleic acid anti- gens, the d-serum chiefly with d-, the 1-serum with 1-compound, in agreement with the configurational correspondence demonstrated in the optical activity (rotation of polarized light in the same direction). On the basis of these and numerous other similar findings Landsteiner frequently referred to the analogy, as stated above, between immune and enzyme reactions. Avery, Goebel, and associates published a series of papers dealing with the influence of spatial configuration of synthetic sugar- protein antigens on the specificity of serological reactions. Avery and Goebel (1929) prepared the azoprotein of p-aminophenol-^-glucoside and p-aminophenol-)8-galactoside : CHoOH ^ o. ^ OC6H4N = N— Protein H OH OH H H CH2OH OH OH H /I- ' H -O. H p-Aminophenol-/3-glucoside Antigen OC6H4N = N— Protein OH H H OH H p-Aminophenol-/3-galactoside Antigen ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 149 These hexosides were diazotized and coupled, respectively, with horse serum globulin and crystalline egg albumin. Gluco-globulin antiserum prepared by the immunization of rabbits reacted not only with gluco-globulin but also with gluco-albumin. Anti-serum reactive with gluco-albumin showed no reaction with galacto-albumin. The serum of rabbits immunized with galacto- globulin reacted with the solutions of galacto-globulin and galacto- albumin in approximately equal titre. These antibodies, however, failed to precipitate gluco-albumin. These data provided proof that the con- figuration of the sugar radical, regardless of the nature of the protein to which it was attached, was a determining factor in the serological specificity of these conjugated antigens. The antisera prepared with sugar proteins also contained anti- protein antibodies which exhibited the protein specificity of the species from which they were derived. They could be removed from the immune sera by specific absorption without loss in the titre of the coexisting anti-carbohydrate antibodies. Conversely, the anti-carbo- hydrate antibodies could be specifically inhibited from reacting by the addition of the homologous glucoside without diminishing the activity of the anti-protein precipitins present in the same serum. Avery, Goebel and Babers (1932) also showed that, though the antisera against azo-proteins of p-aminophenol-a-glucoside cross-reacted with i8-antigen, a differentiation in the specificity of a- and /^-antigens could be affected by cross-inhibiting tests. The reaction of an immune serum with its homologous antigen was specifically inhibited only by the homologous glucoside, while the cross-reaction between this serum and the heterologous antigen was completely inhibited by either gluco- side. This lack of reciprocal inhibition of the precipitins in a- and 13- antisera was interpreted, by the authors, as further evidence of the lack of the immunological identity of the two isomeric glucosides. In studies with azo-proteins of a-maltoside, ^-cellobioside, i8-lacto- side and /?-gentiobioside, Goebel, Avery and Babers (1934) demon- strated that the specificity of serological cross-reactions of disaccharide antigens is determined by the stereochemical configuration of the terminal hexose molecule; that only the test antigen containing the maltoside (terminal hexose, a-glucose*) reacted only in a-antiserum. * Strictly speaking the terminal hexoses in the above disaccharides are no longer free molecules and therefore should be named hexosides. 1 50 IMMUNO-CATALYSIS The two test antigens containing the cellobioside and gentiobioside (terminal hexose, /?-glucose) reacted only in )8-antisemm, and finally the lactoside test antigen (terminal hexose, /^-galactose) reacted only in )8-galactoside antiserum. CH2OH CH2OH 0— J OC6H4N=N— Protein H CH2OH OH H OH p-Aminophenol-j3-maltoside Antigen OH CH2OH p-Aminophenol-j3-cellobioside Antigen J_OC6H4N=N--Protein CH2OH OH J 8^ OH CH2OH p-Aminophenol-/?-Iactoside Antigen O — CH O*^ lJ_OC6H4N=N--Protein OH LXOC6H4N=N— Protein p-Aminophenol-j3-gentiobioside Antigen ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 151 2. Specificity of Enzymes Which Catalyze Stereoisomeric Substrates The above cited facts may be considered some of the most striking examples as characteristic of the role of determinant groups in general, and for the highly specific influence of the spatial configuration of glucosides in particular, on the specificity of immune reactions. These findings have shown beyond doubt the sensitivity of the response of the cytoplasm-antigen vi^hich is involved in the synthesis of the immune globulins. These facts have no doubt been instrumental in the emphasis of the oft referred analogy betu^een immune and enzyme reactions. This analogy is clearly shown in the animal system which demonstrates enzyme specificity in the synthesis of a particular optical isomer as well as in the production of specific antibodies against optical antipodes. The analogy between these two types of biocatalytic reactions offers indeed a very attractive prospect of comparing directly the role of the participants in immune reactions wdth those of enzyme reactions. In comparison with the role of the stereochemical and optical con- figuration of the carbohydrate groups of the synthetic gluco-protein antigens on the specificity of antibodies, we will discuss below the specificity of carbohydrases. According to the specificity theory of Weidenhagen (1932, 1940) the specificity of carbohydrases is char- acterized by their ability to hydrolyze or catalyze carbohydrates of three structural types. They are: a. Sugar Isomerism, that is, the sugars differing at C-atoms. For example, the difference between glucose and galactose at the fourth carbon-atom; b. Constitutional Isomerism, that is, the isomerism between aldoses and ketoses. For example, the difference between glucose and fructose; c. Ring Isomerism, or a- and ^S-isomerism. For discussion of the possible analogy between the specificity of the glycoside— azoprotein antigens and the specificity of the carbohydrases which hydrolyze di- and tri-saccharides, the following table is con- structed. Table IV shows that a disaccharide such as saccharose containing a- and )8-hexosides are hydrolyzed by both the a- and )8-glucosidases. Each glucosidase can attack, however, only one of the linkages and 152 IMMUNO-CATALYSIS Table IV a- and l^-Specificity of Glucosidases Glycosides Hydrolyzed by: a-Glycosides a-Heteroglucoside (hetero^aglucon, a non- sugar group) a-glucosidase Maltose i=a-glucosido-4-glucose a-glucosidase Saccharose :=a-glucoside-^-fructoside a-glucosidase and /3-fructosidase Turanose=:a-glucosido-fructoside a-glucosidase Melezitose=:a-glucosido-/3-fructosido-a-glu- coside a-glucosidase ^-Glycosides /?-Heteroglucoside ^-glucosidase Cellobiose=y8-glucosido-4-glucose ^-glucosidase Gentiobiose=/3-glucosido-6-glucose ^-glucosidase Salicin^a yg-glucoside (glucose -fsaligenin) ^-glucosidase a-Galactosides a-Heterogalactoside agalactosidase Melibiose = a-galactosido-6-glucose agalactosidase Raffinose=yS-fructosido-a-glucosido-6-a-galac- toside a-galactosidase and ^-f ructosidase ^-Galacto sides ^-Heterogalactoside /3-galactosidase Lactose =: yS-galactoside-4-glucose /3-galactosidase fi-Fructofuranosides Saccharose yg-fructosidase (invertase) Raffinose /5-fructosidase Gentianose=/3-fructosido-ci-glucosido-6- yS-glucoside yg-fructosidase Inulin^A polymer of ^-fructose /3-fructosidase Glucose and galactose in the above combinations possess pyranose structure and fructose furanose structure, and all have d-configuration. ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 153 not the other. a-Glucosidase will attack the a- and not the ^-linkage and /?-fructosidase will attack the 13- and not the a-linkage in saccha- rose. This is illustrated in the following manner: Saccharose =a-Glucose^ y/5-Fructose f '^ a-Glucosidase /3-Fructosidase Raffinose=o-Galactose Antibody globulin Enzyme+substrate ^» Reaction products It is apparent from the above schema that in the in vivo mechanism the following pairs have the same roles: (a) Antigen and Enzyme (b) Globulin Factorsf and Substrate (c) Antibodies and Reaction Products ^Combination between toxin and antitoxin. fThe term Globulin Factors does not stand for normal globulins. It stands for peptides or amino acids which are utilized for the synthesis of antibody globulins. ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 157 The specific catalytic action of antigens on globulin factors or a cer- tain group of amino acids, or polypeptides, etc., produces the antibody globulin molecule. Antibody is therefore a final reaction product with specific affinity for the antigen. Similarly the specific catalytic action of an enzyme on its substrate produces final reaction products with spe- cific inhibitory affinity for the enzyme. Antibodies being final reaction products of a chain of catalytic re- actions—associated with the synthesis of globulin— should be looked upon as specific inhibitors of antigens. As inhibitors of antigens, or as inhibitors of the biological activity (toxicity, etc.) of antigens they are comparable to the specific enzyme inhibitors which result from the action of enzymes on substrates. This is particularly true when the enzyme reactions take place in a complex protein environment as shown by Northrop (1922) and others. The combination between an antigen and its homologous antibody is a stable one and usually results, under optimal conditions, in a pre- cipitation or agglutination reaction. This reaction is reversible under certain not too mild conditions, such as heating at 50°-60°C., disso- ciation by 10 per cent sodium chloride or the separation of the anti- body from its combination with antigen by dilute acids. The action of antibody on antigen does not produce a permanent chemical change in the latter. The neutralization of the toxic properties of toxins and venoms is not one of destruction but is a process of blocking of active groups in the same way that an inhibitor blocks enzymically active groups. Marrack's contention that the action of antigens on antibodies is to produce some change in antibodies resulting in insolubility con- cerns a secondary and physico-chemical property of the antigen- antibody complex rather than a change produced in the antibody mole- cule. For antibodies freed from their combination with antigens have been shown to exercise a neutralizing property identical in every respect with that exercised previous to their combination with antigen. The stable union between an antibody and antigen is therefore a stoi- chiometrical reaction representing a simple combination between multivalent radicals, and it obeys the mass action law, as often empha- sized by Heidelberger and Kendall (1935; Heidelberger, 1939; Ken- dall, 1942). In contrast, an enzyme molecule does not form a stable combination with its substrate. The life of the combination between an enzyme 158 IMMUNO-CATALYSIS substrate is often less than one millionth of a second. That is, in one second more than one million molecules come into contact with a single enzyme molecule during which time they are catalyzed and metabolized. Such a continuous combination, dissociation and destruc- tion of the substrate at tremendous speed by its specific enzyme is in no way comparable to the antigen and antibody reaction. It can be seen readily that the action of an enzyme on its substrate is dynamic and a catalytic reaction, and that of antigen with antibody is non- catalytic and results in a static state. This process has a true counterpart in the enzyme reactions.* In practically all enzyme reactions whether the substrates are anabolized or catabolized by the action of enzymes, certain inhibitory *In in vitro experiments the combination between an antigen and antibody may or may not result in precipitation. If the number of antigen molecules is in large excess a precipitation may fail to occur (certain bacterial carbohydrates or small molecular weight proteins reacting with their respective antibodies). However, there are antigen- antibody reactions in which precipitation occurs even in the region of excess antigen (a large molecular weight protein antigen reacting with its specific antibody). In inhibition experiments, reactions which involve the participation of antibody and of small molecular weight haptens, a precipitation fails to occur, despite the complete neutralization of the combining groups of the two reactants. The phenomenon of precipitation does not therefore accompany all the phases of, or various types of, antigen-antibody reactions. The counterparts of the various phases of or of various types of antigen-antibody reactions are encountered in reactions involving a combination between an enzyme and its reaction products. These inhibitors combine with and completely inhibit the activity of the enzymes. Such inhibitions (or combinations), however, may or may not result in the formation of an insoluble enzyme-inhibitor complex. As in antigen- antibody reactions the failure to form an insoluble complex appears to depend on the molecular size of the inhibitor as well as the ratio of the number of molecules of the two reactants. As discussed below the enzyme pepsin is inhibited by pepsin inhibitor; the degree of inhibition is dependent on the concentration of the inhibitor. When, however, one molecule of inhibitor combines with one molecule of pepsin the enzyme is completely inhibited as a consequence of the formation of a soluble inhibitor-pepsin complex. As in certain antigen-antibody reactions, the combination between the inhibitor and pepsin does not result in a precipitation. On the other hand a solution of trypsin inhibitor mixed with a solution or trypsin of equal molecular strength completely inactivates trypsin with the formation of a crystalline compound with trypsin. Straus and Goldstein (1943) have reported a study on the zone behavior of enzymes. Departing from the classical treatment of the kinetics of enzyme reactions, they show that under a number of common conditions systems involving the participa- tion of an enzyme, specific substrate and an inhibitor, the enzyme-substrate, enzyme- inhibitor systems behave in three distinct ways depending upon the concentrations of the reactants and the dissociation constants of the system. An important practical con- sequence of the theory of zone behavior concerns the effect of diluting a mixture of enzyme and inhibitor (or substrate). They show that dilution is a crucial operation which significantly afiFects the subsequent experimental observations. For a fuller understanding of this study the reader is referred to the original article. ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 159 reaction products are formed in due course. By virtue of the structural similarity to the substrates from which they are formed, these products exercise specific affinity for the enzymes in accordance with the mass action law, in the same manner as antigens combine with antibodies. The degree of inhibition is dependent on the degree of affinity for the enzyme and the amount of the inhibitor produced. The consequences of the reaction of the enzyme molecule with its reaction product as inhibitor, is comparable in every respect with the reaction of antigen with the antibody which has resulted from the action of antigen in vivo. The inhibitor produced by one enzyme will not inhibit another enzyme. It is therefore highly specific in the same way that an antibody will not combine with a heterologous antigen. 5. Zinsser's View on the Formation and the Role of Antitoxins Northrop (1922) in his studies on the kinetics of the action of pepsin and trypsin on protein observed certain divergences which he explained by the production of hydrolytic products acting as inhibitors. These inhibitory products combining with the enzyme maintained an equilibrium which obeyed the ordinary mass action law. The diges- tion of protein by pepsin, combined with such substances as peptone (the word "peptone" used broadly as signifying substances in solution with which the pepsin combines without hydrolyzing them), was not proportional to the total concentration of the pepsin. He therefore, believed that as the pepsin digested the protein, peptone-like substances are formed with which pepsin combined and with which it then maintained an equilibrium following the mass action law. In adding increasing amounts of peptone to pepsin solution, as a result of this manner of combination, the first amounts added inactivated more pepsin than the latter additions, which Northrop pointed out is in very striking analogy to the manner in which antitoxin and toxin react. In the case of trypsin as it acted on the undigested protein solution trypsin inhibiting substances were also formed. The activity curve showed that, at first, there was a very rapid drop of activity which gradually became slower. Hans Zinsser (1923) in his discussion of toxin and antitoxin stated that a good deal of light may be hoped for in regard to the nature of 160 IMMUNO-CATALYSIS antigen-antibody unions from investigations upon the nature of enzyme reactions. Discussing the above cited resuhs of Northrop with a view as to the relation of such findings to the toxin-antitoxin reaction he made the following statement: "These experiments of Northrop do not, of course, solve the question of antigen-antibody unions, but they do serve to bring the analogy of toxin-antitoxin relations much closer to laws governing the union of enzyme with substrate. Moreover, they show that enzyme is actually used up in its reactions, just as toxin is used up in its reactions with antitoxin, and that equilibrium follow- ing the mass action law may be a definite factor in the quantitative rela- tions governing the reactions. It is not at all impossible that the general laws which govern reactions between trypsin and its inhibiting sub- stances are similar to those which govern the toxin-antitoxin reaction. Moreover, while it is a dangerous subject upon which to theorize, it is yet not utterly impossible that the toxins are closely analogous to enzymes, and that they produce in their action upon cells products of injury which, passing into the circulation, become the specific inhibitors of the toxin or the antitoxin. We do not wish to dignify this with the label of a theory, but in subjects as vague as the origin and biological meaning of antitoxins, we must grasp at every straw that may suggest experimental paths for enlightenment." Since Zinsser expressed the above view two decades ago our knowl- edge of the mechanism of enzyme and immune reactions has greatly advanced. Numerous enzymes have been isolated in crystalline and highly pure form, and their chemical nature has been clarified. With the same pace the chemical nature of whole antigens, haptens and antibodies has been quite extensively studied. The reactions of antigen and antibody, principally through the studies of Heidelberger and Kendall, have been quantitated and related to the mass action law. Reevaluation and correlation of these findings appear to lend support to the view expressed by Zinsser and define the nature of the analogy between immune and enzyme reactions observed long ago by Ehrlich. B. THE FORMATION OF SPECIFIC INHIBITORS IN ENZYME REACTIONS As stated above, by the action of an enzyme on a substrate specific inhibitors, structurally related to the substrate, are generally pro- ANTIBODY AS A SPECIFIC EIS^ZYME INHIBITOR 161 duced. This structural relationship between the substrate and the inhibitor is the reason for the affinity an inhibitor possesses for the enzyme, and thereby causes the inhibition of the enzyme. In the follow- ing pages the formation of several such specific enzyme inhibitors will be considered. 1. Pepsin Inhibitor It has been shown by Herriott (1938, 1941) that pepsinogen auto- catalytically is transformed into pepsin at pH 4.6; the pepsin formed catalyzes the reaction. During this transformation there are simultane- ously produced certain polypeptides one of which has a powerful inhibiting action on pepsin at pH 5.0-6.0. The inhibitor has been isolated in the form of spheroids which change later on to rosettes of tiny needles. This inhibitor on combining with pepsin between pH 5.0 and 6.0 forms a dissociable inhihitor-fe-psin complex. The reversible combination of pepsin vdth the inhibitor follows quantitatively the simple mass law equation. The molecular weight of the inhibitor as determined by diffusion and combining equivalence with pepsin lies somewhere between 4000 and 10,000. Since 0.000,25 mg. of inhibitor nitrogen is approximately equivalent to 0.0012 mg. of pepsin nitrogen in the inhibitor-pepsin complex the ratio of mg. of inhibitor N to mg. of pepsin N is 1 :4.8, or one molecule of pepsin reacts wdth one molecule of inhibitor. This complex dissociates upon dilution and upon acidification (in a manner which appears to be comparable to the dissociation of anti- gen-antibody complex under the same conditions). It does not combine in acid solution. The entire reaction is presented by Herriott as follows: Pepsin pH<5.4 Pepsin Pepsinogen > Pepsin-Inhibitor < > -f- Compound pH>5.4 Inhibitor— >X The first reaction from pepsinogen to pepsin-inhibitor compound is catalyzed by pepsin, while the second reaction from the compound to free pepsin and the inhibitor is a reversible dissociation. The third reaction is the destruction of the inhibitor on long standing with pepsin between pH 2.0 and 5.0. The pepsin inhibitor is a polypeptide which has basic groups which are exposed, since it is precipitated by many reagents used to precipi- 162 IMMUNO-CATALYSIS tate basic substances, namely, tungstic, phosphotungstic, flavianic, picric, and picrolonic acids. The inhibitor is not precipitable with trichloracetic acid. The inhibition of pepsin by this inhibitor is demonstrated by the rennet method (decrease in milk clotting activity of pepsin) which is carried out at pH 5.8. The pepsin inhibitor has no demonstrable effect on the activity of crystalline trypsin, on the milk clotting activity of crystalline chymo- trypsin or commercial rennet. Conversely, the crystalline trypsin in- hibitor has no effect on the milk clotting action of pepsin. This indicates a high degree of specificity, that usually is associated with en- zymes, also exists among some inhibitors of enzymes. Bovine pepsin was inhibited to the same degree as swine pepsin, but chicken pepsin was not inhibited at all. On the other hand, a crude inhibitor solution prepared from chicken pepsinogen inhibited both swine and bovine pepsin, but had no effect on the chicken pepsin. According to Bourdillon (1945) the action of pepsin (and other proteolytic enzymes) on antitoxin (diphtheria) pseudoglobulin yields in addition to a heat labile and low molecular weight nitrogenous material, a new protein of reduced molecular weight (60 per cent of the native substance) which still has all the characters of a native sub- stance. Split antitoxin is only slowly hydrolyzed by pepsin in moder- ately acid medium and is thus able to form reversible compounds of appreciable stability with pepsin. This complex contains from two to three molecules of pepsin to one of antitoxin. The split antitoxin-pepsin complex is markedly similar to the edestin-pepsin combination. In both cases, maximum precipitation occurs at about pH 4.0 and the two pro- teins combine with pepsin in approximately equal amounts. Both com- plexes are richer in pepsin when formed in the presence of excess pepsin. 2. Trypsin Inhibitor of Pancreatic Extracts Kunitz and Northrop (1936) isolated the trypsin inhibitor from trypsinogen crystals. The inhibitor is believed to play a very important part in regulating the activation of trypsinogen, and in partly activated pancreatic extracts more or less active trypsin occurs in the form of an inactive compound with the inhibitor. Like pepsin inhibitor, the ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 163 trypsin inhibitor is a polypeptide with a molecular weight of 6000. It is likewise not precipitable with 2.5 per cent trichloracetic acid, either hot or cold, nor by boiling in water. A solution of the inhibitor mixed with a solution of trypsin of equal molecular strength at pH 7.0 in 30 minutes at 6°C. completely inactivates trypsin. The inhibitor forms a compound with trypsin which is dissociable at pH 1 .0. The inactivating effect of the inhibitor is demonstrated in experiments on the digestion of casein, the digestion of sturin, or the activation of chymotrypsinogen into chymotrypsin, or trypsinogen into trypsin in the presence of trypsin or salt, and the clotting of blood. The substance also inhibits chymotrypsin but to a less marked extent. One molecule of inhibitor combining with one molecule of trypsin results in the formation of inhibitor-trypsin complex in the form of hexagonal, many-faced crystals with a molecular weight of 40,000. This combination is split by trichloracetic acid which precipitates trypsin and leaves the inhibitor in solution. 3. Trypsin Inhibitor of Blood Serum It is known that normal serum also contains trypsin inhibitor which blocks the activity of trypsin. Schmitz (1938) isolated this inhibitor in various ways. On precipitating the serum proteins with trichloracetic acid the inhibitor remained in the supernatant acid solution. It was separated by ultrafiltration from the proteins. After precipitation of serum proteins with acetone the inhibitor was found in the acetone solution. The behavior of this inhibitor against trypsin was analogous to the inhibitor isolated from pancreas by Kunitz and Northrop (1936). 4. Inhibition of Carbohydrases by the Reaction Products A disaccharide or glucoside such as sucrose, with specific affinity for invertase, on hydrolysis yields reaction products which manifest inhibitory affinities for the enzyme. Thus fructose and glucose, the products of hydrolysis, inhibit invertase markedly (Henri, 1902; Michaelis and Menten, 1913). a-Methylglucoside which has the same configuration as the a-glucose is also found to inhibit invertase (Michaelis and Rona, 1914). On the other hand, those disaccharides and glucosides, such as maltose, lactose and /?-methylglucoside, possess- 164 IMMUNO-CATALYSIS ing configuration different from sucrose, lack affinity for invertase and therefore are not hydrolyzed by invertase and do not inhibit the activity. Kuhn (1925) and van Klinkenberg (1932) have further shown that the inhibition of an enzyme by its reaction product is highly specific, and that this specific inhibition is dependent on the configuration of the reaction product in the same manner that the serological specific- ity of the reactions between conjugated sugar-protein antigen and its homologous antibody is dependent on the configuration of the sugar molecule in the antigen. Kuhn (1925) determined the inhibitory effect of stereoisomeric carbohydrates on the activity of various amylases. He found that /?- amylase present in malt was inhibited most strongly by )8-maltose. On the other hand a-amylase present in takadiastase (also in pan- creas) was inhibited by a,^-maltose twice as strongly as by i8-maltose. These findings are significant in view of the fact that a-amylase hy- drolyzing starch produces a-maltose and i8-amylase produces ^-malt- ose as a reaction product (starch consists of 36 per cent a-starch and 64 per cent /?-starch). These data show that various amylases are most strongly inhibited by those maltoses which they themselves produce specifically. The following table is constructed from the results obtained by Kuhn which show clearly that strong inhibition is caused by i8-malt- ose and yS-glucose in contrast to a negligible inhibitory effect exer- cised by their a-isomers.* Table VI Inhibition of Malt /3-Amylase hy Stereoisomeric Sugars Per cent Inhibition by: Substrate a-Glucose ^-Glucose a-Makose /?-Makose a-, ^-Mahose Soluble starch Amylose 5, IP 4,0.4 37,39 48,44 2, 14 55,50 60,49 31,36 40,55 ^The 5 and 11, etc. pair of figures represent per cent inhibitions calculated from results obtained by Kuhn respectively after 12 and 25-minute reaction periods. * Hunter and Downs (1945) reported that the action of arginase upon arginine at pH 8.4 is inhibited by all a-amino acids of the naturally occurring configuration, but not by d-a-amino acids, amino acids having the amino group in other than the o-position, urea, or native proteins. ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 165 5. Inhibition of the Oxidation of /^-Hydroxybutyric Acid by Its Reaction Product Acetoacetic Acid and Related Acids Jowett and Quastel (1935) found that fatty acids are oxidized at considerable rates by slices of liver in vitro and give as one of their oxidation products acetoacetic acid, as in the body. Jowett and Quastel (1935) studied the rate of acetoacetic acid production from butyric, crotonic and ^-hydroxybutyric acid under various conditions. On the basis of the results obtained they proposed a mechanism of the oxida- tion of these acids to acetoacetic acid as follows: Crotonate >iAcetoacetate±^^-hydroxybutyrate Butyrate ^ They found that 0.00 IM benzoate, cinnamate and phenylpropionate strongly inhibit the oxidation of butyric and crotonic acids to acetoacetic acid. Cinnamate and propionate inhibit oxidation of /?-hydroxybutyric acid to a much smaller extent than the oxidation of butyric and crotonic acids. Green, Dewan and Leloir (1937) studied the nature of the enzyme system responsible for the oxidation of j8-hydroxybutyric acid to ace- toacetic acid. The dehydrogenase system was prepared from heart muscle, and coenzyme I was found to be an indispensable component of the system. This system specifically catalyzed the oxidation of l-/?-hydroxybutyrate to acetoacetate. They also showed that the change from y8-hydroxybutyrate to acetoacetate is reversible. The highly specific nature of the dehydrogenase system was evidenced by the fact that it did not catalyze the oxidation of iS-hydroxypropionic, a-hydroxybutyric, crotonic, y-hydroxybutyric, butyric and acetic acids. The above specific dehydrogenase system was inhibited by 0.03 M iodoacetate, pyruvate and oxalacetate 56, 48 and 71 per cent, re- spectively. Under identical conditions the inhibition by 0.012 M of the reaction product, acetoacetic acid, was 55 per cent. Green et at., as cited above, appear to have shown that the tissue slices used by Quastel and his associates represent a mixture of en- 166 IMMUNO-CATALYSIS zyme systems capable of oxidizing several of the fatty acids mentioned above. In evaluating the inhibitions of the tissue enzyme systems it is clear that the inhibition of the oxidation of butyric and crotonic acids by benzoate, cinnamate and phenylpropionate, and relative absence of inhibition of the oxidation of ^-hydroxybutyric acid by these inhibitors indicate their affinity for one enzyme system and not for the other. The inhibition of the specific oxidation of )3-hydroxybutyric acid by the enzyme system used by Green, et al. by iodoacetate, pyruvate, oxalacetate, etc. as well as by the reaction product acetoacetic acid must be attributed to their common affinity for the same enzyme system. Such common affinities for an enzyme are exhibited by substances which are structurally similar to the specific substrates or their reaction products. 6. Inhibition of Succinoxidase by the Oxidation Product of Succinic Acid, and by Structurally Related Acids That acids structurally related to a substrate and to the reaction prod- ucts inhibit the particular enzyme system has been reported by other investigators. Weil-Malhebre (1937) reported that in the presence of an enzyme preparation from fresh ox heart succinic acid consumed (during a two-hour period) 311 fi\ of O2, 1-a-hydroxyglutaric acid, 10 fj}, a-glycerophosphate, 29 /xl and d(— )glutamic acid 45 lA; l-(+) glutamic acid and ;8-glycerophosphate were unreactive. In agreement with the earlier observations of Quastel and Wooldridge (1928) he found that malonic acid* inhibited the enzyme completely, whereas maleic acid did not inhibit at all. a-Ketoglutaric acid (M/20) was found to inhibit specifically the succinic dehydrogenase 40-50 per cent. Other keto acids, such as pyruvic or 2-ketogluconic acids or other substances of related constitution, e.g., hydroxyglutaric or glutamic acids, exercised no inhibition. Potter and Elvehjem (1937) studied the succinoxidase system of the brain, liver and kidney tissues of rats and chickens. The inhibitory ^That malonic acid is a biological metabolite has been demonstrated by Raistrick (1938). It is a component of, and is liberated by alkaline hydrolysis of a high molec- ular weight polysaccharide which is produced from glucose by a strain of Penicillium luteum. Vennesland and Evans (1944) reported that the oxidation by tissue of oxalacetic acid, a derivative of succinate, yielded malonate. Malonate, therefore, in- hibits the very enzyme which produces it. ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 167 effect of 0.034 M oxalic, malonic, glutaric, adipic, 1-aspartic, 1-malic and fumaric acids on the oxidation of succinic acid was determined. Malonic and oxalic acids exercised most marked inhibition, while glutaric, adipic and aspartic acids inhibited about 25 per cent, and fumaric acid 24 per cent. The last is the first oxidation product of suc- cinic acid. In these cases also the inhibition of succinic acid oxidation by malonic acid would seem to be easily explainable on the basis of an affinity between succinic acid dehydrogenase and malonic acid, owing to the similarity of configuration. Since malonic acid cannot be de- hydrogenated, unlike the activated succinic acid, it does not readily dis- sociate from the enzyme and consequently blocks the enzyme from re- acting with succinic acid. The inhibition of succinoxidase by the other acids must likewise be attributed to its property of combining with com- pounds possessing affinities similar to succinic acid. 7. Inhibitory Effect of the Salts of Organic Acids on the Oxidation of Tyrosine by Tyrosinase With the object of finding an analogy to the inhibition phenomenon observed in precipitin reactions with organic acids, isomeric with or structurally related to those coupled vdth proteins, Landsteiner and van der Scheer (1927) examined the effect of 41 sodium salts of organic acids on the action of the oxidase of potatoes and mushrooms on tyrosine (p-hydroxyphenyl-a-aminopropionic acid). In a general way they stated that carboxylic acids of benzene and other cylic com- pounds acted more strongly than fatty acids, and that stronger inhibi- tion was caused by the meta and para substituted acids than by the ortho substituted. 8. Inhibition of Lactic Acid Dehydrogenase by the Reaction Product Pyruvic Acid Green and Brosteaux (1936) in their study of the mechanism of the dehydrogenation of lactic acid by enzyme solutions prepared from animal tissues observed that there was a rapid oxygen uptake by the system during the first five to 10 minutes, and then the rate fell 168 IMMUNO-CATALYSIS off sharply. The fall of the rate of oxygen uptake was so marked that at the end of one hour the final uptake was only slightly greater than at the end of the first few minutes. This effect was held to be due either to (a) the rapid destruction of the enzyme, or that (b) some product of the reaction inhibited the enzyme very strongly. The product of the reaction was assumed to be pyruvic acid and this assumption was confirmed by the fact that by the addition of cyanide to the system and thus by the formation of cyanhydrin, the accumulation of pyruvic acid was prevented. In the presence of an optimal concentration of cyanide the rate of the oxidation of lactate was increased markedly. The beneficial effect of cyanide was formulated as follows: CH3COCOOH+HCN ^ CH3C-COOH l\ I CN OH To confirm their assumption they added an excess of pyruvate and abolished the effect of cyanide; addition of a small amount of pyru- vate had no effect. That the rapid fall of the oxygen uptake by lactate was due to pyruvate as reaction product was also demonstrated by adding 0.05 M pyruvate to the system at the beginning of the experi- ment in the absence of cyanide. This concentration of pyruvate com- pletely inhibited the oxidation of lactic acid. While 0.04 M pyruvate inhibited the oxidation of l(+)-lactate 100 per cent, 0.03 M tartronate inhibited 60 per cent by virtue of configurational similarity to both the substrate and the reaction product. The enzyme system studied by Green and Bostreaux shows a spec- ificity similar to the high degree of serological specificity shown by antibodies against the optical, or d- and 1-isomers of haptens coupled with proteins. Though it oxidizes Mactate rapidly it has no effect on d-lactate, nor does a 0.03 M concentration of the latter exercise an inhibitory effect on the enzyme activity. 9. Inhibition of the Dehydrogenases of Succinic, Lactic and Mahc Acids by Their Reaction Products Das (1937a) investigated the affinity of lactic dehydrogenase towards both malic and lactic acids. The enzyme systems he used were ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 169 prepared from pigeon breast muscle, pig kidney and pig liver. He found that the "maximum concentration" (critical concentration)* of lactic acid is about five to eight times greater than the maximum con- centration of malic acid, thus showing that the affinity of the enzyme is much higher towards malic acid than towards lactic acid. The question of whether the oxidation product of one substrate inhibits the oxidation of the other substrate was also studied: i.e., does oxalacetic acid, the oxidation product of malic acid, inhibit the dehydrogenation of lactic acid, and conversely does pyruvic acid, the oxidation product of lactic acid, inhibit the dehydrogenation of malic acid? In the former case the lactic acid dehydrogenation was inhibited to an extent of 50 to 60 per cent (by 1 mg. of oxalacetic acid), using the maximum concentration of lactic acid, while with the same con- centration of oxalacetic acid the dehydrogenation of malic acid was also inhibited to the same extent. But pyruvic acid only inhibited the dehydrogenation of lactic acid and not the dehydrogenation of malic acid. This inability of pyruvic acid to inhibit the dehydrogenation of malic acid was explained by the fact that the affinity of the enzyme is much higher for malic acid than for lactic acid. Das in a subsequent study (1937b) investigated the mechanism of inhibition of reversible dehydrogenation-hydrogenation systems by their respective reaction products. The experimental conditions were arranged so as to obtain 50 per cent inhibition. He found, for example, that the dehydrogenation of lactic acid to pyruvic acid was inhibited 50 per cent by 0.3 mg. of pyruvic acid. Conversely the hydrogenation of pyruvic acid to lactic acid was inhibited 50 per cent by 3.0 mg. of lactic acid. The reaction was presented as follows: 0.3 mg. of pyruvic acid (50%) t Lactic acid ^ pyruvic acid i 3.0 mg. of lactic acid (50%) In the same way the dehydrogenation of malic acid to oxalacetic acid was inhibited 50 per cent by 0.02 mg. of oxalacetic acid, and the hy- drogenation of the latter to malic acid was inhibited 50 per cent by * "Maximum concentration" is the concentration at which the enzyme is most active. 1 70 IMMUNO-CATALYSIS 3.5 mg. of malic acid. Similar effects were obtained in experiments on the reversible dehydrogenation of succinic acid to fumaric acid. Oxalacetic acid inhibited both the dehydrogenation of succinic acid and the hydrogenation of fumaric acid. Malonic acid inhibited the dehydrogenation of lactic, malic and succinic acids and the hydrogen- ation of their respective reaction products. Das emphasized the point that the enzyme has greater affinity for the oxidized forms (i.e., p)n:uvic acid and oxalacetic acid) than for their reduced forms (i.e., lactic and malic acids). This was evident from the comparatively strong inhibition of the dehydrogenation of the reduced forms by pyruvic and oxalacetic acids. 10. Inhibition of Carboxylase by Acetaldehyde, the De- carboxylation Product of Pyruvic Acid Since the discovery of carboxylase by Neuberg it has been variously reported that acetaldehyde resulting from the decarboxylation of pyruvic acid inhibits the activity of carboxylase. Lohmann and Schuster (1937) in their study on the isolation and properties of cocarboxylase determined the inhibitory effect of acetaldehyde in a system containing a yeast suspension, washed with alkaline phosphate, purified cocarboxylase and magnesium. During a 30 minute reaction period at optimal pH of 6.2 to 6.4 the inhibition of the decarboxylation of pyruvic acid by 2, 4 and 8 mg. acetaldehyde was respectively 25 to 30, 50 to 55 and 64 to 67 per cent. 11. Inhibition of the Oxidation of Purines by the Re- action Products of Purine Structure There exists in milk and in most living tissues an enzyme capable of oxidizing the purine bases hypoxan thine and xanthine. Dixon and Thurlow (1924) studied this system in some detail. They used a preparation from milk. They reported that uric acid, a purine base, inhibited the oxidation of both hypoxanthine and xanthine, and the oxidation of hypoxanthine to xanthine. Other structurally related purine bases, such as guanine, adenine and xanthine itself were found to produce this inhibitory effect upon the oxidation of hypoxanthine. ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 171 The effect was, however, found to be remarkably specific for caffeine, but the pyrimidines uracil, cytosine and thymine showed no inhibitory effect, f^istidine, which contains the imidazole ring, as in the purines, was also found not to have any effect. From the study of the kinetics of the reaction with hypoxanthine they concluded that the inhibitions of the oxidation of hypoxanthine by uric acid and by xanthine are essentially the same in nature. Dixon (1926) in a later study determined the specificity of xanthine oxidase. This enzyme, with a high degree of specificity, oxidized only hypoxanthine and xanthine, and adenine to a slight extent, and most aldehydes. It had no action on (a) guanine, alloxan, quinoline, mor- phine, ricin; (b) tryptophane, ketones, benzylamine, peptone; (c) formate, acetate, lactate, succinate, citrate, glutaminate; (d) caffeine, theobromine, uracil, thymine, cytosine, histidine, uric acid; and (e) glycine, tyrosine, alanine, serine, leucine and aspartic acid. Coombs (1927) extending the studies on the specificity of xanthine oxidase found that the enzyme had no oxidizing action on 3-, 8-, and 9- methylxanthines, 1,3- and 3,8-dimethylxanthines, 1- and 7-meth- ylguanine, 1 ,7-dimethylguanine and benziminazole. Beside hypo- xanthine and xanthine 6,8-dihydroxypurine and 2-thioxanthine were readily oxidizable. The reaction products were uric and thiouric acids, respectively. His studies further showed that the introduction of a single methyl group in either ring entirely prevented the oxidation. In inhibition experiments hypoxanthine, xanthine, 3-methyl- xanthine, uric acid, adenine 6,8-dihydroxypurine, guanine, 1- and 7-methylguanine and 1,7-dimethylguanine were strongly adsorbed on the enzyme; 8- and 9-methylxanthine and alloxan adsorbed only to a small extent. Dimethyl- and trimethylxanthines and pyrimidines ad- sorbed only to a very slight extent. It is evident that the purine ring, and not the pyrimidine nor the imidazole ring, exhibits specific affinity for the enzyme. The results of this study showed clearly that the degree of adsorp- tion on, or the specific affinity for the enzyme was not only responsible for the oxidation of the purine bases, but also for the inhibition of the enzyme. The results also show that the simpler the structure of the purine compound, the greater is the degree of affinity exhibited for the enzyme. 172 IMMUNO-CATALYSIS The oxidation of hypoxanthine H— N— C = H— N— C = H— Ni— eC^O H H— C C— N H ->0 = C C— N H -_).0 = C2 ^C— 'N CH ^ N— C— N Hypoxanthine H— N— C— N Xanthine ^ CH H- «c=o -N3— "G— ^N— H Uric Acid shows that the point of attack is at carbon 8 in the purine ring when xanthine is oxidized to uric acid and at carbon 2 when hypoxanthine is oxidized to xanthine. The fact that uric acid as a reaction product inhibits the oxidation of hypoxanthine and xanthine shows that it reacts with the same active grouping of the enzyme molecule which combines also with xanthine or hypoxanthine. As a result of this re- action the two xanthines become activated but uric acid not being activated by the enzyme can act only as a competitive inhibitor. 12. Inhibition of the Hydrolysis of Guaninedesoxyribose by the Reaction Product Guanine Klein (1935) prepared an enzyme from spleen which was highly specific in hydrolyzing purine nucleosides. This enzyme preparation did not act on pyrimidine nucleosides, purine or pyrimidine nucleo- tides, or on d-ribose- or desoxyribose-nucleic acids. To determine the specific affinity of nucleosidase for various sub- stances he resorted to inhibition experiments, using guanine desoxy- riboside as substrate. enzyme Guaninedesoxyriboside >guanine-|-desoxyribose The experimental procedure yielded 51 per cent hydrolysis of the nucleoside. Studying the effect of 18 different substances he found that the addition of 1 mg. each of guanine, hypoxanthine and ade- nine to the reaction system produced, respectively, 47, 53 and 21 per cent inhibition. Under identical conditions xanthine produced 10 per cent inhibition and uric acid had no effect whatsoever. It is evi- ANTIBODY AS A SPECIFIC ENZYME INHIBITOR 173 dent that guanine as the reaction product of hydrolysis is a strong inhibitor of nucleosidase. While adenine exercises slight inhibitory effect, its deamination product, hypoxanthine, exercises greater inhibi- tion. In contrast, the deamination of guanine to xanthine deprives the inhibitor of its effect on the enzyme. Five mg. of desoxyribose, one of the reaction products of hydrolysis, exercised only 10 per cent inhibition. Five mg. of each of the following substances: fructose, glucose, uracilriboside, cytosinedesoxyriboside, yeast and muscle adenylic acid, yeast and thymus nucleic acids, riboseguanilic acid were entirely ineffective. 13. Conclusion In the preceding pages numerous experimental facts have been dis- cussed showing how in reaction systems catalyzed by enzymes, substances are formed as final reaction products which specifically combine with the enzymes and inhibit their activity. This specific inhibition appears to compare with the inhibition of the biological activity of antigenic substances by antibodies which as final reaction products result from the catalytic action of antigens in vivo. It would have been more expedient perhaps if we could have found examples from synthetic processes catalyzed by enzymes for a direct comparison with the production of antibodies resulting from the catalytic influence of antigens during the synthesis of globulins. As our knowledge of the mechanism of complex synthetic enzyme processes is very fragmentary, such a comparison was not possible. However, since as both processes are catalyzed, and in each case the final reaction products act as inhibitors on their respective enzyme systems, we believe the comparison is well within the limits of sound reasoning. Our view concerning the synthesis and production of antibody as catalyzed by antigen differs from the theoretical basis of the experi- ments carried out by Pauling and Campbell (1942). We believe that antigen catalyzes and directs the synthesis of globulins from amino acids or polypeptides (see pp. 120, 156) in a specific direction to yield the homologous antibody globulin. Pauling and Campbell, on the other hand, have experimented udth and advanced the idea that the complete globulin molecules can be converted into antibody molecules. 1 74 IMMUNO-CATALYSIS According to them the end chains of the globulin molecule can be uncoiled by a denaturing agent or condition, and during the process of recoiling, antigenic substances may intervene and impress their con- figurations on the recoiling end chains, yielding specific "antibodies." A similar experiment has been carried out by Bacon (1943) whereby whole plasma was dehydrated from the frozen state in the presence of an antigenic substance. The above investigators claimed to have obtained specifically reactive final "antibody" products. While none of these investigators have since produced additional corroborative results, as it was discussed earlier (page 116), various investigators have repeated these experiments and failed to confirm the above results. Part IV Anti-Enzyme Immunity A. ANALYSIS OF CERTAIN "CONTROVERSIAL" ASPECTS OF ANTI-ENZYME IMMUNITY THE FORMATION and existence of anti-enzymes in animal systems or the production of antibodies against enzymes, by parenteral injection into animals of enzyme preparations, must be demonstrated by the same critical tests employed in toxin-antitoxin or other antigen- antibody reactions. The immune sera against enzymes must be specifi- cally produced and must manifest specific serological properties or must specifically inhibit the activity of the homologous enzymes in vitro as well as in animals, whenever one or both of these tests are experimentally possible. If such anti-enzyme sera satisfy the known criteria of immune reactions, the antigenicity of enzyme proteins and, therefore, the existence of enzyme antibodies can be considered as an established fact. The experimental data concerning anti-enzyme immunity have been divided into two categories to enable us to effect a reasonable comparison of both the enzyme and other immune processes, and the development of the various phases of the concept of "Immuno-cataly- sis," in a logical order. We have already described in Part I of this treatise the preparation of numerous immune sera against crystalline enzymes which were tested by precipitation and anaphylactic reactions. The experimental data to be presented in this part of the treatise deal with the results obtained from experiments carried out in animals and in vitro with respect to the inhibition of the activities of enzymes by the homolo- gous immune sera. However, before undertaking the presentation of the data certain controversial questions as to the existence of anti- enzymes must be analyzed in the light of available experimental facts. 175 1 76 IMMUNO-CATALYSIS 1. Analysis of Bayliss' Objections Against the Existence of Anti-Enzymes There have been pubhshcd numerous studies in support of or against the existence of anti-enzyme immunity. Some of these studies date back to the last decade of the 19th and early part of the twentieth century. Numerous studies have been made and reported on this subject since. Bayliss' has been one of the chief antagonists of anti-enzyme immunity. One will find the list of his objections in the first and through the fourth edition of his Principles of General Physiology and also in all of the editions of his monograph on the Nature of Enzyme Activity. Since, through these publications, Bay- liss has, perhaps, been very influential in shaping the reserve or per- haps the censorious attitude maintained by certain workers in the field of immunology regarding anti-enzyme immunity, it is necessary that his objections be analyzed. In his monograph on The Nature of Enzyme Activity Bayliss (1925) has presented the following four principal objections: First: "The facts that enzymes are not proteins and that evidence is accumulating to show that their chemical constitution is of a sim- pler nature than was supposed at one time, are, 'prima facie, grounds for doubting their capacity of acting as antigens." We quote also the following paragraph from Wells' (1929) Chemical Aspects of Im- munity which expresses practically the same view as the above, i.e., "At first both* were believed to be proteins; now both are considered by many not to be proteins, but molecular complexes of nearly equally great dimensions." The first objection of Bayliss, though the most important of the four, is the weakest. Studies carried out during the last twenty years have confirmed beyond doubt the older view of the protein nature of enzymes by isolating and subjecting to critical experiments numerous crystalline proteins with enzyme activities, such as urease, amylase, trypsin, pepsin, papain, d-ribonuclease, catalase and numerous other enzymes. These facts obviate Bayliss' principal objection. Second: Discussing the merits of certain studies on anti-enzymes Bayliss claimed to have demonstrated that the absence of the proper *Toxins and enzymes. ANTI-ENZYME IMMUNITY 177 control of H^ ion concentration in these studies may have resulted in decreased enzyme activity, which fact was considered by him as responsible for an anti-enzyme effect. In the subsequent discussions of the experimental data published by various investigators, this ques- tion will be carefully analyzed and it will be shown that Bayliss' own data do not contradict but confirm the anti-enzyme effect as an immune reaction. Third: "I have already pointed out that many of the 'anti' effects shown by serum can be accounted for by change of reaction, but this fact does not seem capable of explaining the increase of such effects stated to be produced by injection of enzymes. It is to be remembered, however, that when the normal blood already shows such properties, it is practically impossible to be certain that an increase following an injection is not due to a spontaneous change. Natural variations are, in fact, very considerable." The implications of the above objection of Bayliss are vague. Since no specific experimental facts are mentioned in conjunction with the statement of this objection we refrain from analyzing it; we believe, however, that the experimental data to be discussed at length will show that the properties of anti-enzymes cannot be accounted for by ascribing them to non-immunological natural variations. Fourth: "A further fact to be kept in mind is that substances capa- ble of taking up enzymes by adsorption produce a diminution of their action merely by reducing their concentration." The examples he cited were: the adsorption of trypsin on charcoal; saponin prevents the adsorption and anti-action in a similar way to that in which it pre- vents the inactivation of rennet by shaking; the anti-tryptic action of serum has been shown to be associated with the albumin fraction of the proteins, not with the globulin fraction, as is usual with true anti- bodies. The probable interpretation of this fact, according to Bayliss, is "that the effect is due to an adsorption of the enzymes, this dimin- ishing the effective concentration." He mentioned also the fact that when raw serum or egg-albumin is acted on by trypsin, it is found that no effect appears to be produced for some hours, and that gradu- ally the enzyme begins to act and regains its power. This phenomenon is explained by him as follows: 'The raw protein, for some reason not as yet clear, is difficult of attack, but adsorbs the enzyme. As it is slowly attacked and converted into products which have no ad- 178 IMMUNO-CATALYSIS sorbent properties, more and more of the enzyme is set free to act." He stated further: "There is no doubt of the existence of substances which have a markedly inhibiting action on certain enzymes, although it leads to confusion if they are called 'anti-enzyme,' since there is no evidence that they can be produced in response to the injection of these enzymes into organisms." The question of non-specific adsorption of enzymes on various ad- sorbents as a possible cause of the diminished enzyme activity and of thereby accounting for the immune anti-enzyme inhibitory effect will be discussed under the heading of non-specific adsorptions. We will therefore discuss here the nature of the normal enzyme inhibitors found in the living cell and compare their properties with those of anti-enzyme antibodies. a. The Nature of the Trypsin Inhibitor Present in Sera in Re- lation to Anti-Enzyme Antibody. The anti-tryptic action of certain normal sera, referred to by Bayliss, appears to us to correspond to the trypsin inhibitor of low molecular weight and of polypeptide nature which has been studied by Schmitz (1938). As discussed previously Schmitz isolated it from horse serum and characterized it as being similar to that crystallized from pancreatic extract by Kunitz and Northrop (1936). Both of these inhibitors are polypeptides of about 5000 molecular weight, which are not comparable at all with the anti- trypsin immune globulin described by Ten Broeck (1934). Antitrypsin antibody falls into the class of globulins having a molecular weight of about 150,000 to 160,000. It is also to be noted that the antitrypsin antibody is not present in the albumin fraction of the serum; in con- trast, the trypsin inhibitor present in normal serum, as emphasized by Bayliss and demonstrated by Schmitz, is solely in the albumin fraction of serum. These facts make the further discussion of this subject un- necessary. However, in this connection it is of interest to note an observation by Maschmann (1937). He found that the proteolytic activity of the toxin of CI. ferfringens was not inhibited at all by normal horse serum which was shown to exercise an inhibitory effect on other bacterial proteolytic enzymes. The proteolytic activity of the toxin, on the other hand, was strongly inhibited by antitoxic horse serum. Smith and Lindsley (1939) studied the inhibition of the proteolytic enzymes of bacteria by immune sera. They separated the constitu- ANTI-ENZYME IMMUNITY 179 ents of normal serum by electrophoresis in the Tiselius apparatus. The inhibitor was found primarily in the albumin fraction and the antiproteinase antibody in the globulin fraction. Smith and Lindsley (1939) likewise found that while the trypsin inhibitor in normal serum had no inhibitory action on the proteinases of pathogenic CI. histolyticum, CI. welchii and CI. oedetnatis-maligni, antiserum against the enzyme of CI. histolyticum strongly inhibited the enzyme activity. Immune sera against the enzymes of the latter two organisms were not prepared. b. The Nature of the Trypsin Inhibitor in Egg White. Balls and Swenson (1934; see also Balls and Hoover, 1940) isolated an in- hibitor present in egg white; the purer form dialyzed slowly through collodion. It gave positive biuret and Millon tests, but was negative in a test with sodium nitroprusside, in a Molisch test and with Feh- ling's solution. A solution of 10 mg. per ml. gave no visible precipitate with picric, trichloracetic, tannic acids, or mercuric chloride. With phosphotungstic acid, or with two volumes of saturated ammonium sulfate, a heavy precipitate formed. The total nitrogen content was 10.55 per cent; it was resistant to heat. These data show that the properties of the inhibitor are comparable to those of peptone-like substances, or to the trypsin inhibitor found in pancreas (Kunitz and Northrop) or in blood serum (Schmitz). Balls and Swenson (1934) stated that they had corroborated the earlier findings of Delezenne and Pozerski (1903) that this inhibitor combines with kinase, because additional amounts of kinase decrease the inhibition. Furthermore, they found that the reversal of inhibition by kinase becomes more marked as the amount of inhibitor is de- creased. On the other hand, additional amounts of inactive enzyme (trypsinogen or protrypsin) have also the effect of removing the inhi- bition. The inhibition is therefore a reversible reaction. The interpretation of the above mentioned reactions was based on the then current idea that "trypsin-kinase" was a definite compound. Since the publication of the above paper, various enzyme components of pancreatic extracts have been obtained in crystalline form and studied for their interactions (Kunitz, 1939; Kunitz and Northrop, 1936). The results of these studies may be compared to those ob- tained by Balls and Swenson with the inhibitor from egg white. According to Kunitz and Northrop (1936) and Northrop (1939) 1 80 IMMUNO-CATALYSIS enterokinase activates crude chymotrypsinogen to chymotrypsin. In contrast, crystalline trypsin fails to effect this activation. This failure was attributed to the presence of a trypsin inhibitor in the crude chymotrypsinogen which combines with trypsin and inhibits its ac- tivating property. After one crystallization of the crude chymotryp- sinogen, however, the inhibitor remains in the mother liquor; trypsin then is capable of activating chymotrypsinogen to chymotrypsin. The crude chymotrypsinogen also contains trypsinogen, which, acted upon in the same manner (removal of inhibitor) by entero- kinase, is transformed into trypsin. A sufficient amount of active tryp- sin thus formed is capable of overcoming the inhibitory action of the solution, resulting in the activation of chymotrypsinogen. The same result is obtained by adding enough trypsin even in the presence of the inhibitor. Balls and Swenson (1934) and Balls (personal communication) stated that the trypsin-inhibitor of egg white was incapable of inhibit- ing the activity of papain. One positive effect on papain was attrib- uted to something else, since the search for the same positive effect with other papain preparations utterly failed. Ross and Tracy (1942) have studied the effect of the inhibitor of egg white on the digestion of casein by chymotrypsin. They observed only 15 to 20 per cent inhibi- tion. However, if the inhibitor was first incubated with casein (at 37.5°C. for 40 minutes) before adding the enzyme, no inhibitory effect was observed. The delayed action of trypsin on raw egg white described by Bayliss could not be due to this inhibitor for the following reasons. The tryp- sin used by him was capable of hydrolyzing heat denatured egg white. Since the inhibitor as found by Balls and Swenson is resistant to heat, it should be still present in the heat denatured egg white in an active form and therefore inhibit the active trypsin. Since heat denatured egg white exercised no inhibitory effect on trypsin, the delayed action of trypsin on raw egg white must be due to some other factor. The possible nature of this factor is discussed below. c. The Resistance of "Living" Protein to the Action of Proteolytic Enzymes. It has been known for a long time that all living cells are resistant to proteolytic enzymes. Dead organisims are rapidly digested by them. Several hypotheses have been advanced to explain this fact. ANTI-ENZYME IMMUNITY 181 Among others, the following may be mentioned: The presence of enzyme inhibitors (or anti-enzymes) in the living cells; the presence of passive or active protective membranes; selective impermeability of the cell membranes to the proteolytic enzymes; the existence of re- pulsive forces between the cell membrane and the proteolytic en- zymes, etc. After a critical analysis of these hypotheses Fermi (1910) rejected them and arrived at the conclusion that the resistance of "living" proteins is due to a difference of "biochemical constitution" between the proteins of living and dead cells. Northrop (1926, 1939) stated that he confirmed the conclusion of Fermi. Of the above hy- potheses, the impermeability of the living and the permeability of the dead cells to proteolytic enzymes appears to have been one of the most important aspects. Northrop stated that "in every case that as long as the cell was alive, no detectable quantity of enzyme was taken up; whereas when the cell dies, the enzyme was rapidly removed from solution and concentrated in the cell." In the opinion of Northrop the digestion of heat-killed organisms may be accounted for by assuming a change in the chemical nature of the proteins, or the destruction of the anti-enzymes. These objections cannot, however, account for the digestion of the organisms (earth- worm and mealworm) when killed simply by mechanical injury. It is known that there is an inhibitor in organisms such as the earthworm (Luvtbricus terrestris). The presence of this inhibitor might be consid- ered as the reason for the resistance of the living earthworm to the action of trypsin. This can, however, be ruled out in view of the fact that in the presence of an amount of trypsin in excess of the amount required to neutralize this inhibitor the tissue of the earthworm killed by cutting is found to be digestible (Northrop, 1926). In the light of what has been said above, the delayed action of tryp- sin or raw egg albumin in Bayliss' experiment does not appear to be due to an inactivation by adsorption. The gradual "reversal of ac- tivity" beginning with the tenth hour of the reaction period, as de- scribed by Bayliss, and that at the end of the 70th hour, the activity of trypsin on raw egg albumin and on egg albumin denatured at IOO°C. was equal, may indicate a gradual denaturation of raw egg white by dilution with nine volumes of water and standing in contact with the reaction system for such a long time. The denaturation of 182 IMMUNO-CATALYSIS egg white with relative ease would appear to account for the facts described by Bayliss.* The ability of trypsin to digest the type specific M protein from living streptococci is reported by Lancefield (1941). She stated that the M substance in the living bacteria is readily accessible to the action of proteolytic enzymes without injury to other vital functions (viability, virulence and the ability of the multiplying streptococci to synthesize M) of the living cells. This, she concluded, may be due to its possible location near the outer surface of the streptococci. M is an unsymmetrical protein of about 41,000 molecular weight (Pappenheimer, et at. 1942). The ratio of major to minor axis is at least 20 to 1, which puts it into the class of "denatured" or unfolded proteins. Since the process of extraction from streptococci (16 hours at 56°C. with 0.05 N HCl containing 2 per cent sodium chloride) and subsequent treatments with alcohol, etc. do not affect the sero- logical type specificity of M, it is possible that the constitution of this protein in its natural environment may be similar to that of the isolated form. These facts may perhaps explain why M, of all the other cellular proteins, is apparently selectively digested by trypsin. Impermeability of living cells to proteolytic enzymes as the cause of the resistance of "living" proteins to these enzymes does not ac- count for the resistance of extracellular native proteins to proteolytic digestion. Anson and Mirsky (1933), and Anson (1938) reported that hemoglobin denatured by salicylate is digested by trypsin which does not attack native hemoglobin. The denatured hemoglobin was in- soluble under the same conditions under which native hemoglobin was soluble; it had the parahematin type of spectrum which is also given by a solution of hemin in pyridine. When the denaturation of hemoglobin by salicylate is reversed, the original properties of native hemoglobin are restored, t Bawden and Pirie (1937) also found that crystalline tobacco mosaic virus is resistant to proteolytic enzymes. They stated that no enzyme *In this connection an observation by Pozerski and Guelin Cl938a) is of impor- tance. They found that raw egg white exercised no inhibitory effect on the gelatinase activity of B. histolyticum. In contrast the proteolytic activity of this organism was strongly inhibited by immune horse serum against B. histolyticum. This serum exer- cised no inhibitory action on the proteolytic activity of closely related B. sforogenes. fin this connection it is to be noted that active bacteriophage and botulinal toxin similar to native hemoglobin, are resistant to proteolytic enzymes (Kalmanson and Bronfenbrenner, 1943^. ANTI-ENZYME IMMUNITY 183 preparation has yet been found that attacks purified virus prepara- tions at an appreciable rate, or that has any permanent effect on their infectivity. They tried the proteolytic effect of trypsin, pepsin, papain and autolyzed preparations of kidney at a number of pH values around that optimal for enzymic activity, but in no case were they able to observe any proteolytic activity on the living virus. In con- trast, all these enzymes were shown to be strongly active proteolyti- cally when tested against the heat denatured virus which was rapidly hydrolyzed. In the presence of a large amount of trypsin the infectivity of the purified virus preparations was reversibly inactivated as a result of a possible virus-trypsin complex formation. This inactivation of the infectivity was observed to occur immediately after the virus and enzyme were mixed, and no further loss followed incubation. By precipitation with acid or dilute ammonium sulfate solution the virus was recovered with its full activity from such non-infective mixtures. Similarly, when various amounts of a solution of papain were added to constant amounts of virus, a papain-virus precipitate was obtained. From these precipitates the active enzyme was recovered by extraction at pH 3.3. Virus was also recovered without undergoing any change in chemical, infective or serological properties. According to Kleczowski (1944) pepsin combines with virus X (without affecting infectivity, suggesting that different parts of the virus particles are involved in combination) and casein, which are substrates for its proteolytic activity, but not with tobacco mosaic virus, which is not a substrate. Tobacco mosaic virus denatured by heat is readily hydrolyzed by pepsin and combines with pepsin almost to the same extent as potato virus X. On the other hand, more trypsin com- bined with tobacco mosaic virus (with loss of infectivity), which is not a substrate for its proteolytic activity, than with potato virus X, which is a substrate. The combination of trypsin with tobacco mosaic virus could account for the reversible inhibition of infectivity of the virus by trypsin. This combination protects trypsin from spontaneous inactiva- tion at pH 7.0. Invertase does not combine with potato virus X, with tobacco mosaic virus, whether heat denatured or not, or with casein. d. The Inhibition of Enzymes by Enzymes and Viruses. Bayliss emphasized also the presence of normal enzyme inhibitors in living cells, particularly a proteolytic enzyme inhibitor present in the worm 184 IMMUNO-CATALYSIS ascaris. Weinland (1903) isolated a substance from ascaris capable of inhibiting both peptic and tryptic activity. He concluded that the inhibitor acted by combining with the enzymes and that its function was to prevent the destruction of the parasite by the host's digestive juices. In the Hght of the findings of Fermi, Northrop and others, discussed above, that "hving" cells or living protein molecules are not attacked by proteolytic enzymes, Weinland's conclusions do not ap- pear to be acceptable. Furthermore Sang's (1938) findings throw a different light on the subject. Though Sang's experimental findings corroborate the principal part of Weinland's findings, his data suggest that the antipeptic and antitryptic substance of ascaris is a protease.* He extracted it by grinding the cut up worms to a mass in a mortar and extracted the ground mass with water for 3 days under sterile conditions. The extract filtered through a Berkefeld filter was found to be stable on boiling in acid solution. In fact the proteolytic activity of the preparation increased by coagulating and removing the ex- traneous proteins. On boiling from five to 36 hours the activity was lost with the disappearance of the biuret reaction. It was completely pre- cipitable by less than half saturation with ammonium sulfate. This, with its ready diffusibility through parchment and its precipitability by 70 to 80 per cent alcohol made it probable that this substance was an enzyme of small molecular weight or was associated with a sub- stance of the order of a primary albuminose. Sang called this substance ascarase. This preparation exercised proteolytic activity on casein and pep- tone in buffered reaction systems at an optimum pH of 5 to 7. Since it was also found that the maximum inhibition of trypsin by this sub- stance lay within the same pH range. Sang held it probable that the two enzymes combine with each other at a pH range of their maximum activity as they would with their normal substrates. This combina- tion between the two enzymes was rapidly formed and easily reversible. It was found that ascarase was partly destroyed by standing in solution at neutrality, while in its combination with trypsin it was not destroyed. The study of the kinetics of numerous reaction systems revealed that * According to Hamed and Nash C1932) extracts of Ascaris lumhricoides contain both a trypsin inhibitor and a proteolytic enzyme of optimal activity at pH 8 and 37.5 °C. They found that this trypsin inhibitor (free from the worm proteinase) protected isoelectric insulin at pH 8 and 37.5°C. against the proteolytic action of strong pancreatic solutions. ANTI-ENZYME IMMUNITY 185 the inhibitory and the proteolytic powers of ascarase ran parallel. It inhibited trypsin and pepsin to the same degree, but had no effect on papain. Northrop (1933) found that crystalline edestin adsorbs pepsin from a solution, removing it completely at pH 4.0. The pepsin content of "edestin-pepsin" complex was found to be as high as 50 per cent. In this form the pepsin activity was not arrested. A suspension of "edestin- pepsin" complex on standing at room temperature gradually dissolved, eventually completely, so that the final solution consisted of the digested edestin containing the original quantity of pepsin. A combination between crystalline d-ribonuclease and tobacco mosaic virus resulting in the reversible inactivation of the virus was reported by Loring (1942). A virus-enzyme complex formed long fibre-like particles, which on analysis proved to contain about 14 per cent d-ribonuclease. This complex was dissociable completely, liberating the virus in fully active form. The dissociation took place by diluting the solution of the complex from 1 to 500 to 1 to 1000 times. The liberation of virus from the virus-enzyme complex was achieved also by sedimenting it at high speed and redissolving the pellets. On repeating this process a few times 92 per cent of the virus was recovered. 2. The Question of Non-Specific Adsorption of Toxins or Enzymes on Coexisting Protein-Anti-Protein Precipitates The diminution of the activity of enzymes in the presence of ho- mologous immune sera was assumed by Bayliss to be due also to non- specific adsorptions on extraneous protein impurities in the form of antigen-antibody precipitates. This question of non-specific adsorption as the cause of serological reactions has often been raised since the early beginnings of the science of immunology and has been critically considered in numerous studies. However, since there still seems to be an element of doubt in the mind of many workers regarding the exist- ence of immune anti-enzymes, it is necessary that the available data regarding this question be further analyzed. At first a few examples of adsorptions of enzymes on materials, such as were mentioned by Bayliss, will be given. There is no doubt 186 IMMUNO-CATALYSIS that enzymes adsorb on charcoal, alumina, Fuller's earth, etc. in a manner similar to the adsorption of toxins on these adsorbents. However, one very seldom carries out immune reactions in the pres- ence of such adsorbents; secondly, it does not necessarily lead to the inactivation of the enzymes or toxins. The decrease of the activity or absence of inactivation in the adsorbed state depends on what particu- lar groups of the enzymes or toxins have been blocked by the adsorption process. Griffin and Nelson (1916) found that invertase adsorbed on alumi- num hydroxide or on small amounts of charcoal is just as active enzy- matically as the solution of free enzyme. The presence of egg-albumin in invertase solution likewise does not affect the enzyme activity. Michaelis (1921) reported that invertase adsorbed on Fuller's earth or iron oxide, likewise, is just as active as the free enzyme. Fructose, mannose, lactose, a- and )8-methylglucoside fail to elute the enzyme from the adsorbent. Sucrose and raffinose which are hydrolyzed, and maltose which is not hydrolyzable by invertase, slowly promote the elution of the enzyme. These facts show that the elution effects of the former two substances are not related to their being substrates for the enzyme. On the other hand, those substances, glucose, fructose, man- nose, a-methylglucoside, which show a great affinity for and thereby inhibit invertase activity, fail to effect the elution of the enzyme. According to Kleczowski (1944) invertase adsorbed on charcoal can be set free by casein, and it can be extracted by tobacco mosaic virus, but not by sucrose. Oparin and Kurssanow (1929) precipitated invertase by tannic acid; in this state invertase was inactive. It is interesting to note that by shaking the tannic acid precipitate of the invertase preparation with egg albumin or peptone they stated they had formed egg albumin- tannate, or peptone-tannate precipitates, setting free the invertase in an active form. Evidently egg-albumin-tannate, or peptone-tannate precipitates are incapable of adsorbing the liberated invertase. Freund (1931) reported that tannin detoxifies diphtheria and tetanus toxins in solution or adsorbed on collodion particles. This effect oper- ates in vitro as well as in vivo. Tannin combining with the toxins produces a precipitate. The degree of detoxification runs parallel with the amount of precipitate formed. However, this effect is reversible at pH 8, or in high dilutions, and the toxin is recovered in an intact ANTI-ENZYME IMMUNITY 187 form. Klopstock and Neter (1933) reported that tannin detoxifies cohra venom and ricin in vivo, as well as their hemolytic effect in vitro. However, these inhibitions by tannin were reversible by simple dilu- tion of the detoxified reaction mixtures. It is apparent from these studies that the inhibitory action of such substances is easily reversible, and has nothing in common with the inhibitions brought about by immune bodies. Liiers and Albrecht (1926) in a study on immune anti-amylase, in- vestigated this particular question of non-specific adsorption of an enzyme on a coexisting antigen-antibody precipitate to meet this specific objection of Bayliss. The amylase activity in the presence of egg albumin-anti-egg-albumin precipitate showed no decrease. Since data of sufl&cient scope regarding the non-specific adsorbability of enzymes in an antigen-antibody reaction environment are not avail- able, we will discuss as a means of comparison the data regarding the non-specific adsorbability of toxins on serological precipitates. Such a comparison between toxins and enzymes is justified on the basis of the resemblances of their physico-chemical properties, as well as the enzymic properties manifested by certain toxins. These resemblances have been observed for many years (Wells, 1929). The same difficulty is encountered in isolating toxins and enzymes; both are approximately of similar molecular dimensions and are non-dialyzable through animal or other membranes; they pass through porcelain filters; they possess similar adsorption properties; neither will stand boiling and most forms are destroyed at 80° instantly or in a short time; left standing in solu- tion for some time they gradually lose their specific properties, the toxin becoming toxoid and the enzyme a fermentoid. Besides the above physico-chemical resemblances between the toxins and enzymes they exercise two other properties which bring them closer. The first is the high degree of activity of toxins which compare favorably with those of enzymes. The second is the fact that toxins often show enzyme activities, whereas certain enzymes show toxic activities. Eaton (1936) found that tetanus toxin of certain purity is fatal in a quantity of 0.4 microgram per kgm. body weight of a guinea pig. Sommer (1937) found that 0.2 microgram of botulinus toxin per kgm. of the body weight of mouse is fatal. Eaton (1936) and Pappen- heimer (1937) found that 0.4 microgram of diphtheria toxin per kgm. body weight of guinea pig is fatal. 188 IMMUNO-CATALYSIS As will be discussed in detail in the latter part of this study, the toxic effects of several toxins are correlated with their proteolytic or leci- thinase activity. On this and on the grounds cited above, a comparison of the adsorbability of toxins with that of enzymes appears to be justified. a. Adsorption of Toxins on Red Blood Cells and Charcoal with- out Loss of Activity. Sbarsky (1923), and Sbarsky and Michlin (1923) stated they had found that inactivated diphtheria toxin adsorbs on red blood cells in vivo and in vitro, and that there is a relationship between the adsorption property of the cells and the susceptibility of the animal to diphtheria. They found that diphtheria toxin is adsorbed by the red blood cells of various animals in the following order of adsorptive power expressed in percentages: rat, 14; rabbit, 75.8; horse, 88; guinea pig, 91.8; hen, 93; pigeon, 95. The above observation appears to have been corroborated by the findings of Dujarric de La Riviere and Kossovitch (1932). They stated that red blood corpuscles adsorb diphtheria toxin and that the adsorbing ability of the red blood cells varies among different animals. The blood plasma did not fix, or it fixed only a very small amount of toxin. Levine (1939) was able to remove staphylococcal exotoxin quantitatively by adsorbing on red blood cells. Eisler (1922) studied the reactivity of toxins and antitoxins after adsorbing them on charcoal. The activity of vibrio toxin adsorbed on charcoal was not altered. In this state, it was capable of combining with an equivalent amount of neutralizing antitoxin. In contrast, vibrio toxin adsorbed on red blood cells was incapable of combining with antitoxin. Diphtheria and tetanus toxin adsorbed on charcoal manifested weaker combining properties for the respective homologous antitoxins. For this reason, they required greater amounts of antitoxin for neutralization. Eisler correlated this behavior of tetanus or diph- theria toxin adsorbed on charcoal wdth the fact that in serum treatment they require a larger amount of antitoxin than the equivalent amount. Despite the weakening of the combining properties of charcoal ad- sorbed toxins, their injurious effect on tissue was not diminished or altered. b. Failure of Protein-Anti-Protein Precipitates to Adsorb Toxins. Maloney and Weld (1925) investigated the neutralization of toxin with specific antitoxin in the presence of other bacterial impurities ANTI-ENZYME IMMUNITY 189 and their antibodies. Horses were immunized against diphtheria cul- ture fluid filtered through a Berkefeld candle. The immune horse sera contained agglutinins against the diphtheria bacilli. A mixture of diphtheria agglutinating sera (which contained no antitoxin) and crude toxin gave a precipitate. After centrifuging the reaction mixture the supernatant showed a marked drop in agglutinin content. This same supernatant showed no measurable fall in the toxin content deter- mined by skin tests on guinea pigs. Conversely, a plasma containing both agglutinin and antitoxin after flocculation of antitoxin with toxin showed no change in the agglutinin titer of the centrifuged super- natant. Schmidt (1926) carried out similar tests as described above. After treating a diphtheria agglutinating serum with crude toxin and removing the floccules by centrifuging, he compared the toxicity of the supernatant with a control sample of toxin treated with saline and found no difference in their strength. These facts show that the protein impurity present in the toxin solution reacts with its homolo- gous antibody, and that the resulting precipitate does not adsorb detectable amounts of the toxin present in the mixture. In an aggluti- nating serum which also contains antitoxin the two reactions, toxin- antitoxin and agglutinogen-agglutinin, take place independently. Marrack and Smith (1931) tested the mutual adsorbability of an azo-globulin dye solution (prepared by linking purified horse serum pseudo-globulin, or crystalline egg-albumin with diazotized atoxyl) with diphtheria toxin-antitoxin floccules. Diphtheria toxin was mixed with azo-globulin solution and then an equivalent amount of antitoxin was added. The mixture was kept for three hours at 45 °C. and over- night in the ice-chest. The centrifuged precipitate, dissolved in 0.01 N NaOH solution, was completely colorless. No azo-globulin was therefore carried down with the toxin-antitoxin floccules. c. Failure of Protein-Anti-Protein Precipitates to Adsorb Non- specific Colored Proteins. Marrack and Smith (1931a) mixed anti- pseudoglobulin serum with azo-egg albumin solution; then one-third the optimum proportion of horse pseudo-globulin was added. The mix- ture was kept five hours at room temperature and in the ice-chest over- night. The precipitate on dissolving in 0.01 N NaOH solution was completely colorless. No dye azo-albumin was therefore carried down by the pseudoglobulin-anti-pseudoglobulin precipitate. The absence of azo-protein in the precipitate was determined spectroscopically with 190 IMMUNO-CATALYSIS ultraviolet radiation of 3600 A. At this wave length only azo-protein was measured. Over the range of concentrations measured Beer's law was found to be obeyed whether other proteins were present or not (see further, Smith and Marrack, 1930; Marrack and Smith, 1931b; Marrack, 1938b). Heidelberger and Landsteiner (1923) in a study on the specificity of hemoglobins obtained from various species of animals, compared the color of precipitates produced in the following combinations: (1) Hemoglobin-j-anti-hemoglobin serum^red precipitate; (2) Hemoglobin -[-horse serum-)-anti-horse serum=rpure white precipitate; (3) Hemoglobin -[-human serum -|-anti-human serum^pure white precipi- tate; (4) Hemoglobin -[-donkey serum-[-anti-donkey serum=pure white precipi- tate; The absence of any color (hemoglobin) in the precipitates of (2) to (4) combinations demonstrated clearly that hemoglobin was not dragged down non-specifically by these three different serum-anti- serum precipitates. Haurowitz and Breinl (1933) precipitated 400 ml. of 1 : 1000 normal horse serum with 40 ml. of anti-horse rabbit serum in the presence of colored beef atoxyl-albumin containing 8.24 mg. of protein with a 61.2 microgram arsenic content. After 24 hours the precipitate was centrifuged and washed once with saline. The precipitate was pure white and free from dye. The dry weight of the precipitate was 55.2 mg. It contained less than 0.1 microgram of arsenic or at the most 0.03 per cent atoxyl-protein. Ten ml. of human serum (1 : 100 dilution) was treated with atoxyl-horse globulin, containing 3.0 mg. of protein with 16.7 microgram of arsenic, and treated with 1 ml. of anti-human serum. The precipitate was pure white and free from arsenic. d. Failure of Protein-Anti-Protein Precipitates to Adsorb Non- specific Proteins. Heidelberger and Kendall (1935) in their quan- titative studies on the mechanism of precipitation reaction subjected the question of the effect of non-specific proteins on the amount of antigen-antibody precipitates to a critical study. They found that at 0° or 37° the ratio of nitrogen to Type III pneumococcus polysaccharide in the precipitate was of the same order whether the precipitation was carried out in the whole serum in which the antibody constituted about 15 per cent of the total protein, or in the antibody solution prepared ANTI-ENZYME IMMUNITY 191 from it, in which antibody was 50 to 60 per cent of the total protein. In other parallel experiments, carbohydrate precipitation determinations were studied with antibody solution alone, and with antibody solu- tion to which an equal volume of normal horse serum had been added. The non-specific protein had no effect upon the amount of antibody nitrogen precipitated under any conditions of temperature investigated. e. Failure of Agglutinated Bacteria to Adsorb Non-Specific Proteins. Heidelberger and Kabat (1934, 1936, 1937) investigated the question as to whether non-specific proteins were adsorbed on bacteria in an agglutinating system. They found that the amount of antibody nitrogen taken out by the bacterial after 48 hours at 0°, with occasional stirring, was independent of the volume just as in the precipitation reaction. Thus the antibody nitrogen removed by agglutination was independent of the concentration of antibody nitrogen in the super- natant. Pneumococcal cells (Type I), agglutinated with a considerable ex- cess of antiserum, were washed with saline until the supernatant con- tained no agglutinin. The cells agglutinated in the region of excess antibody, they believed, would still have available on the surface of the particles some of the specifically reactive groupings of the originally multivalent antibody. These particles, then, should be able to combine with Type 1 pneumococcal carbohydrate on the surface of freshly added unsensitized Type I pneumococcal cells, and reagglutination should take place to form larger aggregates. This assumption was veri- fied and the effect was found to be specific, since it was not given by pneumococcal Types II or III, or by Type I R cells under identical con- ditions of salt concentration. Reagglutination was, moreover, produced almost as completely by suitable amounts of Type I carbohydrate in solution, so that the conclusion appeared inescapable that these par- ticulations, as well as the original antigen-antibody combination, were a chemical, and not a non-specific adsorption process. While the finely and uniformly resuspended pneumococcus-agglu- tinin complex was found to reagglutinate completely with the addition of 0.01 mg. and partially with 0.0001 mg. of Type I pneumococcal polysaccharide, no visible reaction was observed with 0.1 mg. of Type II pneumococcal polysaccharide. When the upper fluid containing the Type II cells in suspension was decanted away from the mass of the Type I pneumococcus-antibody agglutinate, and the decanted suspen- 192 IMMUNO-CATALYSIS sion (containing Type II cells) was treated with Type II antiserum agglutination took place. Conclusion In the preceding pages certain controversial aspects of the question of the existence of anti-enzymes were analyzed. The points considered no doubt apply primarily to impure enzyme preparations. In the light of the above cited facts showing the absence of non-specific adsorption of proteins on antigen-antibody precipitates,* the objections in this regard appear to have been merely of a speculative nature. Further- more, in studies in which crystalline or highly purified enzymes have been used these objections are stripped of their power and are only of historical interest. 3. Effect of pH of Optimal Enzyme Activity on the Nature and Extent of the Antigen-Antibody or Enzyme- Anti-Enzyme Combinations Each enzyme exercises its greatest activity at a narrow range of pH. Pepsin, for example, is most active at pH 2 to 3. In contrast trypsin and papain function best at neutrality. ^-Amylase is most active at pH 5.2; /?-fructosidase (invertase), pH 4.5; a-glucosidase, pH 7.5 to 6.5; in contrast, i8-glucosidase functions best at pH 3 to 6. The pH of the optimal activity of bacterial enzymes varies. The cytochrome-cyto- chrome oxidase systems function at neutrality; on the other hand, carboxylases are most active at or around pH 6. In the study of the nature of the combination between any one particular enzyme and its homologous antibody it might appear necessary to work in a region of acidity where the enzyme is most active. However, the pH of optimal enzyme activity might not be the favorable one for complete enzyme- antienzyme combination. It is necessary that the test be carried out at a pH favorable for immune reactions. As a direct result of the chosen pH the activity of the enzyme might be less than the optimum. This, however, is unavoidable if we are interested in demonstrating the inhibition of the activity of an enzyme by its homologous antibody. *This statement does not apply to the combination between the antigen-antibody complex and the proteins comprised in complement. In complement fixation the com- plement combines non-specifically with various antigen-antibody complexes. ANTI-ENZYME IMMUNITY 193 Schubert (1933), in a study on the retardation of invertase activity by homologous immune sera, found that from pH 3 to pH 4.2, normal and immune sera produce the same effect, the inhibiting property of immune serum vanishing. The presence of either normal or immune se- rum in this acid region enables the invertase to retain 90 to 92 per cent of its activity at pH 4.5 to pH 5. From pH 5 on, on the alkaline side, invertase activity is not altered in the presence of normal serum. In contrast, the retardation by immune serum between the range of pH 4.5 to pH 6.5, respectively, rose from to 25.6 per cent. (At pH 6.5 the activity of invertase per se was 50 per cent of its activity at its opti- mum pH of 4.5.) These findings showed that in the acid region the combination between the invertase and anti-invertase was completely prevented, and that within a range of pH 5 to 6.5 union took place producing retardation of the invertase activity. In view of these facts, a review of the literature regarding the effect of pH on antigen-antibody combination is necessary. Mason (1922) stated that precipitation reactions occurred between pH 4.5 and 9.5 and outside this range, immune precipitates dissolved. Schmidt (1930) found that diphtheria toxin and antitoxin flocculated rapidly at pH 5.0 to pH 8; outside of this range the toxin is partly destroyed and the flocculation is slow or does not take place. Bayne-Jones (1924), studying the effect of pH 4.5, 6, 8 and 9 on the rate of toxin-antitoxin flocculation, stated that the results of titra- tions at pH values beyond the range of 6.4 to 8.4 were irregular and not significant. When the toxic broth was made more acid or alkaline than pH 6.4 to 8.4 non-specific precipitates were produced in the broth, which obscured any flocculation due to the union of toxin and anti- toxin. Brown (1934), in a study of the effect of pH, ranging from 8.0 to 4.77, on the optimal flocculation values, stated that the greater the pH, the more pneumococcal Type I and Type II polysaccharide is necessary to combine with the antibody. Appreciable effect both on the amount and speed of flocculation was noted below pH 6.26. The pH effects were very similar to those of increasing salt concentration; that is, the more salt in the flocculation mixture, the more antibody was necessary to form a stable compound with the polysaccharide. Marrack and Smith (1930, 1931) stated that they observed no effect at pH values from 8.0 to 6.6 on the ratio of antibody to antigen in the ] 94 IMMUNO-CATALYSIS precipitates. Heidelberger and Kendall (1935) in their study on the quantitative theory of the precipitin reaction found no difference, within the range of pH 6.7 to 7.9, on the weight of antigen-antibody precipitates. These investigators made no mention of the effect of lower pH values. Pressman, et at. (1948) reported that the precipitation of p-azosuc- cinanilate ovalbumin antiserum with homologous antigen, and that of the protein antigen of antiserum specific to the p-azophenylarsonate ion and the p-(p-azophenyl)-phenylarsonate ion is optimum at pH values of 7.4 and 8.1 This range of pH is also optimal for the precipita- tion of other azoprotein antigens with negatively charged haptenic groups. Optimal pH for complete agglutination of sensitized cells has been found to correspond with the iso-electric pH of the antibody globulin. At the isoelectric pH of the antibody globulin, the largest amount of sensitizing antibody is found to be absorbed by the cells (Coulter, 1920; DeKruif and Northrop, 1922-1923; Jolfe, 1935). It is apparent from these studies that the most favorable pH for complete antigen-antibody combination is the region near to neu- trality. It is desirable, however, that in an investigation of enzyme- anti-enzyme reactions, this aspect of the question be carefully studied. As an interesting illustration of the importance of pH on immune reactions the findings of Northrop (1932) from a study of the serologi- cal behavior of pepsin are described here. He immunized rabbits using intraperitoneal injections with either active or inactive pepsin prepara- tions. The immune serum against active pepsin gave a precipitate with active pepsin in a dilution of about 1 : 2000, and with inactivated pepsin in a dilution of 1:16. The serum against alkali inactivated pepsin gave a precipitate with both the inactivated and active pepsin in a dilution of about 1 :8. The inhibition of the activity of pepsin (tested with a casein solu- tion) by either of the immune sera was about the same and not very much greater than the inhibition by normal serum. The inhibitory effect, measured by the rate of hydrolysis of gelatin with a small amount of pepsin to which increasing amounts of the various sera were added, was about 40 per cent with serum against active pepsin in a dilution of 1:4, and about 20 per cent with denatured pepsin serum, and still less with normal serum. The weak inhibitory effect ANTI-ENZYME IMMUNITY 195 of antipepsin immune serum was explained by Northrop by the fact that the active pepsin injected must have been almost instantly in- activated in the animal body, due to the neutral reaction of the circulat- ing body fluids. For it was shown that pepsin is instandy more than half inactivated at pH 6. It would therefore seem that it is almost impossible to prevent the denaturation of active pepsin in the animal system; antibody may be formed against the denatured enzyme. B. A CRITICAL CONSIDERATION OF THE ANTIGEN-ANTI- BODY REACTIONS IN RELATION TO THE INHIBITION OF ENZYMES BY SPECIFIC ANTIBODIES I. An Analysis of the Reactions of the Active Groups of Proteins in Relation to their Biological Specificities The ability of enzyme proteins to stimulate the formation of specific antibodies is now an established fact. There are, however, questions pertaining to the inhibition of the enzymes by respective antibodies which require consideration. The complexity of these questions and the inadequacy of the available data do not permit us at present to treat them satisfactorily. One can only raise questions and discuss them as plausibly as possible. These questions may be formulated in the following manner: (1 ) Are the SH, -S-S-, NH2, tyrosine, etc., groups which are studied in relation with the inactivation or activation of certain enzyme pro- teins the sole determinant factors? Or are the reactions involving these groups simply superficial manifestations of other more significant changes which the protein molecule undergoes which we have not as yet been able to determine. (2) Do the anti-enzyme antibody molecules contain specific com- bining sites elicited in response to the stimulation of the active groups of enzymes? (3) Do the inhibitions of enzymes by homologous antibodies result directly from the interaction of the respective specifically reactive groups of enzymes and antibodies in a manner identical with those of other antigens and antibodies? Or are these inhibitions due to second- ary effects resulting from the formation of antigen-antibody complexes? (4) Are the inhibitions due to the mechanical blocking of the en- 196 IMMUNO-CATALYSIS counter of the substrates with the enzymes? That is, does the combina- tion of the antibody with the enzyme block, without combining with, the active areas of the enzyme preventing the contact between the sub- strate and these areas? a. Studies on the Reactive Groups of Proteins. In recent years, certain phases of the first question have been studied and continue to be a hvely subject of interest. Let us briefly discuss the results of such investigations. Investigators have pursued the line of reasoning that, if the biological activity of a purified protein is first lost by the action of group-specific reagents, and then, after reversal of the reaction, re- gained, the group in question is assumed to play a positive role in the biological activity of the said protein. Of the several types of groups located in the side-groupings of the protein molecules SH, NH2 and phenolic groups have received most consideration as possible active "centers" of enzymes, toxins, hormones, viruses, and antigens. The subject has been variously reviewed during recent years (Herriott, 1947; Olcott and Fraenkel-Conrat, 1947; Anson, 1945; French and Edsall, 1945; Landsteiner, 1945), and we can only briefly refer to the pertinent phases of the subject. It should at the start be pointed out that there is as yet no clear demonstration which of a given number of specific groups are essentially related to the biological activity of a protein. The data so far available would seem to indicate that a certain number of a given group situated near or at a certain configurational position, more than others, may play a positive role in regulating the activity of a given species of proteins. b. Reaction of Formaldehyde with Proteins and Amino Acids. It is a classical immunological fact that the action of formaldehyde converts toxins into toxoids, products which are deprived of toxicity without loss of antigenic potency and capacity to combine with antitoxins. French and Edsall (1945) subjected a large number of studies on the reactions of formaldehyde with amino acids and proteins to a comprehensive treatment. It shows that formaldehyde frequently enters into an addi- tion reaction with a compound containing an active hydrogen atom with the formation of a mono-hydroxy-methyl compound, R-CH2OH, which can enter into a condensation reaction forming a methylene bridge, R-CH2-R'. Thus, numerous groups found in amino acids, peptides and proteins are capable of undergoing addition and con- densation reactions with formaldehyde. ANTI-ENZYME IMMUNITY 197 With the amino group it forms mono- and dihydroxymethyl deriva- tives, R-NHCH2OH, R-NH(CH20H)2; with the amide group, hydroxymethyl, R-CO.NHCH2OH, and, at elevated temperatures, methylene diamide, CH2(R-CONH)2; with one imino group, hydroxymethyl, R.i-N-CHoOH, and with two imino groups, meth- ylene compound, (R2-N)2CH2; with the peptide linkage, hydroxy- methyl groups, -C0.N(CH20H)-; with the guanidino group, a drastically changed arginine derivative; with the hydroxy Qalcoholic) group, acetals, R-O-CH2OH, and hemi-acetals, (R-COO)2CH2; with the sidfhydryl (SH) group, thio analogs of acetals, R-S-CH2OH, and hemiacetals, (R-S)2CH2. Action of formaldehyde on tryptophane, tyrosine, phenylalanine, and histidine leads to the formation of addi- tional rings. Fraenkel-Conrat and Olcott (1948) reported that under conditions of pH and temperature which are used for tanning and for the prepara- tion of toxoids and vaccines, formaldehyde introduces a methylene bridge -CH2 between amines on the one hand and the reactive groups of phenolic and imidazole rings on the other. The linkage is resistant to acid hydrolysis. Condensations joining indoles, amines, and formaldehyde under similar conditions may occur with the -NH group of the indole ring. Very careful analytical determinations (Nitschmann and Hadorn, 1943) have shown that formaldehyde can partially (ca. 35 per cent or more) be removed from formolized casein by prolonged washing at room temperature. In combination with a protein, formaldehyde, there- fore, exists in a form which is less firmly bound, or readily removable, and in another form which cannot be readily removed. The point was made that it is impossible, at present, to differentiate the chemical groups involved in these two forms of binding. Velluz (1938) reported that in tetanus toxin formaldehyde combines with tryptophane forming an irreversible heterocyclic three ring com- pound, transforming the toxin into a new antigen. Pappenheimer (1938) reported that diphtheria toxin treated with low concentrations of formalin in alkaline solution forms an irreversible combination. The number of acetylated free amino groups corresponded to the number of £-amino groups of lysine present (5.3 per cent). In converting the toxin into toxoid about 40 per cent of the total free amino nitrogen was found still free. Eaton (1937) reported that 30 per cent of the amino 1 98 IMMUNO-CATALYSIS nitrogen of the toxin is slowly and irreversibly bound. In view of the irreversibility of the union of the toxin with formaldehyde, Pappen- heimer considered this reaction not to be due to the mere formation of methylene linkages to the nitrogen. Hydroxymethyl compounds re- sulting from this reaction with free amino groups are unstable combina- tions (French and Edsall, 1945) and should therefore be reversible. While with low concentrations of formaldehyde the toxin did not lose the ability to flocculate with antitoxin, the higher concentrations caused the destruction of antigenic properties. Horsfall (1934), and Jacobs and Sommers (1939) also reported serologically demonstrable changes with formolized proteins involving combinations other than with free amino groups. The reviewers of the reactions with formaldehyde point out that the types of reactions it enters into is governed by the H+ and 0H~ con- centration, period of treatment and the type of protein treated. In neutral solution, the immediate reaction with proteins is a reversible combination with the free amino groups. With longer time of reaction, as in the case of the preparation of toxoids and vaccines, the formalde- hyde slowly becomes more firmly bound with decrease of amino nitro- gen; under these conditions, not only the amino, but also the indole, amide, and guanidyl groups react with formaldehyde, forming cross- linkages and more stable combinations and, therefore, antigenically modified proteins. There has been indication that formaldehyde reacts also with sulfhydryl groups (Anson, 1945). In view of the many complications resulting from the action of formaldehyde on the various groups in the protein molecule, it is impossible to correlate any specific group in the protein molecule with its loss of viral, enzyme, toxic, serological and antigenic properties. c. Acylation of Proteins. The principal reagents used for the acyla- tion of proteins are ketene [(HoC-C-A)] and acetic anhydride (CH3CO)20. Ketene has been extensively used with aqueous protein solutions. At a pH above 5.0, ketene reacts with NH2, SH and tyrosine-OH groups in proteins. With NH2 groups it yields pro- tein-NH-COCH3. Acylation at pH 5.5 of the amino groups of pepsin causes no inactiva- tion (Herriott, 1935). The number of amino plus tyrosine groups covered was less than the number of acetyl groups, indicating that some other protein groups had reacted. It has been found that ketene reacts ANTI-ENZYME IMMUNITY 199 also with tryptophane (Herriott, 1935), and -SH groups (Fraenkel- Conrat, 1944). With native egg albumin, the reaction is faster with the NH2 than the SH group. In lactogenic hormone (Li and Kalman, 1946) the phenolic groups react with ketene faster than do free amino groups; in insulin amino groups are attacked more rapidly than phenol groups (Stem and White, 1938); and in parathyroid hormone (Wood and Ross, 1942) both the amino and phenolic groups are only about 40 per cent acetylated. In diphtheria toxin (Pappenheimer, 1938) the presence of amino groups of different reactivity with ketene has been indicated. Extensive treatment of tobacco mosaic virus with ketene fails to acetylate all the amino groups (Miller and Stanley, 1941). Boor and Miller (1939) found that freshly ketenized gonococcus still retained enough toxin to kill two out of six mice, but after standing a week in the cold, the preparation killed six out of six mice, indicating the reversibility of the reaction of ketene with gonococcus. After short acylation at pH 6 to 7 with ketene (Goldie, 1937; Pappenheimer, 1938) diphtheria toxin lost its toxicity without losing its ability to combine and flocculate with antitoxin. During the detoxication a number of free amino groups were acetylated, corresponding closely to the number of e-amino groups of lysine present (5.3 per cent). Acylation also of the tyrosine-OH groups caused the loss of the ability of the toxin to combine with antitoxin. Little and Caldwell (1942, 1943) found that acetylation of the amino groups of a-amylase with ketene deprived it of enzyme activity. Sulfhydryl and phenolic groups (tyrosine) groups were of little, if any, significance. Amylase was also inactivated by formaldehyde, phenylisocyanate and nitrous acid, which indicated that the primary amino groups were involved. On the other hand, Sizer (1945) finds that the action of ketene, nitrous acid, phenylisocyanate, formaldehyde, oxidants and reductants on chymo- trypsin causes no inactivation, indicating that the primary amino, sulfhydryl or disulfide groups are not required for chymotrypsin activity. In tobacco mosaic virus, ketene or phenylisocyanate have been reported (Schramm and Miiller, 1940, 1942) to react first with the free amino groups; later the phenol and indole groups are affected. Only the reaction with phenol and indole groups has been observed to be associated with loss of infectivity. The disappearance of the NH2 group with these reagents is said to be without any effect on the in- fectivity of the virus. Alkaline treatment is expected to hydrolyze the 200 IMMUNO-CATALYSIS acetylated phenolic groups. Failure of this treatment to reactivate the acetylated virus, however, led them to conclude that the inactivation of the virus was not due to acetylation of the phenolic groups. Miller and Stanley (1941) reported that about 70 per cent of the amino groups and 20 per cent of the phenol groups of tobacco mosaic virus could be acetylated without loss of activity. Discussing the results of various studies, Olcott and Fraenkel-Conrat (1947) point out the possibility that also some of the aliphatic hydroxyl groups, and groups other than those concerned above, might be in- volved in the treatment of proteins by ketene. It would thus appear that ketene falls short of being a good protein reagent: its action is not specific; it falls short of acetylating completely any or all of the reactive groups; it involves difficult analytical manipulation; it tends to surface denature sensitive proteins; it appears to be extremely toxic, and the racemization of asymmetric carbon atoms of proteins has been indicated. Under these circumstances, the effects resulting from the ketenization of proteins leaves unexplained the specific relationship of various groups with the biological specificities of proteins. d. lodination of Proteins. Iodine has often been used as an oxidiz- ing agent in the study of HS-enzymes. In dilute acid solutions, high iodine concentrations specifically oxidize SH-groups. In neutral and alkaline solutions, iodine substitution in the tyrosine groups of proteins occurs. In either treatment, both oxidation and substitution can take place. Upon iodination of proteins in strong ammoniacal solutions, under conditions which have been employed in studying the antigenic properties of proteins, additional impairment such as loss of species specific properties of the whole molecule would be expected to occur. It has been found that the rate of iodination is associated with the degree of denaturation. The rate of iodination of native protein is slower. In urea solution, which favors denaturation, iodination is more rapid, indicating an increased availability of the tyrosine groups for iodination with increasing denaturation. lodination of imidazole groups (histidine) of proteins, and indole groups (tryptophane) may occur in excess of iodine and prolonged treatment. lodination of horse serum globulin in alkaline medium, when presumably all tyrosine is substituted in the 3 and 5 positions, caused loss simultaneously both of the ability to react with, and the ability to produce anti-species antibodies (Kleczowski, 1940b). On the other ANTI-ENZYME IMMUNITY 201 hand, the treatment of proteins with phenyHsocyanate, forming R— NH— CO— NH— <( y derivatives, or with formaldehyde only slightly reduced their affinity toward homologous antibodies. This was interpreted to indicate that amino groups of the antigenic proteins are not involved in combination with antibody (Kleczowski, 1940). As may be recalled from the preceding discussion on acetylation of proteins, the acetylation of the OH groups of tyrosine in diphtheria toxin was reported to cause the loss of the toxin's ability to combine with antitoxin (Pappenheimer, 1938). Due to technical difficulties, the loss or presence of the original species specificity of the acetylated toxin could not be determined. In connection with the results with iodinated proteins of Kleczowski, it must be remembered that substi- tution in positions 3 and 5 of tyrosine in proteins with diazonium haptenic radicals does not cause loss of the ability of the substituted proteins to produce anti-species antibodies. Anson and Stanley (1941) reported that the treatment of tobacco mosaic virus with iodine, causing abolition of the sulfhydryl groups does not inactivate the viral activity. If enough iodine is added to the virus, or if the reaction is carried out at a sufficiently high temperature, converting the tyrosine groups into di-iodotyrosine groups, the viral activity is lost. Sizer (1945) found that likewdse the activities of chymo- trypsin and phosphatase are destroyed if tyrosine groups are destroyed. Strong oxidants, likewise iodine, inactivated chymotrypsin, involving the oxidation of its tyrosine groups. Herriott (1937) reported that iodinated pepsin is practically inactive when the number of iodine atoms per molecule of pepsin is 35 to 40. Since there are 16 tyrosine residues (mols) per mole of pepsin, the number of iodine atoms found in iodinated pepsin corresponds to complete di-substitution of all the tyrosine molecules, or three to eight atoms of iodine more than required by theory. In evaluating the above considered data one must, no doubt, keep in mind that in addition to the substitution of iodine in the tyrosine molecules, other substitutions and the oxidation of SH groups, under the experimental conditions used, and also the denaturation of the protein molecule, would be expected to occur. e. Reactions with Sulfhydryl and Disulfide Groups of Proteins. It is known that the -SH ^ -S-S- relationship in certain enzyme 202 IMMUNO-CATALYSIS proteins plays a role. This relationship to enzyme activity has been discussed by Barron (1943), and to the denaturation and properties of proteins, by Anson (1945). The role of the SH group in the activi- ties of succinoxidase, phosphoglyceraldehyde dehydrogenase, phospho- glucomutase, and the pyruvate oxidase system has been emphasized. It is common practice to activate papain by certain reducing agents, suggesting, perhaps, a role for SH groups in this enzyme. /^-Amylase (from barley and malted barley) activity was found by Weill and Cald- w^ell (1945) to be associated with SH groups. The inactivation of the enzyme by dilute iodine, ferricyanide and cupric ion was reversed by hydrogen sulfide. The enzyme was irreversibly inactivated by treatment with iodoacetamide, which, as is known, forms an irreversible deriva- tive with the SH groups. Micheel and Bischoff (1937) reported that the reduction of crotoxin by means of cuprous oxide, cysteine and a stream of oxygen inactivates the venom by converting the R-S-S-R to RSH groupings. Sulfu- rous acid alone was capable of inactivating the venom in this manner. A crystalline product derived from crotoxin (rattlesnake venom) was found to contain 4 per cent sulfur existing in -S-S- linkages. (Slotta and Fraenkel-Conrat, 1938). Reduction with cysteine of this product likewise destroyed reversibly the toxic activity in a manner comparable to the reversible inactivation of insulin (du Vigneaud et at, 1931- 1932). De (1940) observed that cobra hemolysin is reversibly inacti- vated by cuprous oxide or organo-mercurials and reactivated by hydro- gen sulfide and reduced glutathione. Benzoquinone depressed the activity of purified hemolysin in 10 minutes; this was partially regen- erated by hydrogen sulfide. Henry (1939) had found that reduced glutathione given either in vitro or in vivo counteracted the anticoagu- lant activity of cobra venom. Tetsch and Wolff (1937) analyzed various animal and insect toxins and found the following correlation between the sulfur content and toxicity. Table VII Fatal Species % Sulfur y poison/g. mouse Bee poison 2.6 10.0 Scorpion poison 3.8 0.7 Cobra venom 5.5 0.15 to 0.12 Crotoxin (rattlesnake) 3.6 0.7 ANTI-ENZYME IMMUNITY 203 Velluz (1938) observed that carbon disulfide detoxifies tetanus toxin, not diphtheria toxin, the difference being ascribed to the absence of SH groups in the latter. Diphtheria toxin contains 0.75 per cent sulfur, but fails to give the nitroprusside test for SH (Pappenheimer, 1937). This amount of sulfur corresponds to about 20 molecules of cysteine per molecule of toxin of 72,000 molecular weight. Pillemer, et al. (1938) found that urease oxidized by cuprous oxide and air is reactivable with reducing agents, H2S and KCN. Although oxidized urease has lost its specific reactivity with antiserum to crys- talline active urease, oxidized and reduced urease elicited similar anti- bodies. This was explained by the fact that whole blood was unable, but tissue extracts were capable of reactivating oxidized urease, which indicated that the oxidized urease is made specifically antigenic in the animal tissue. Pillemer, et al. (1939) converted keratins (wool, chicken feathers, human hair) into their sulfhydryl derivatives (kerateines), and the latter were oxidized to obtain di-thiol derivatives (meta- keratin). Immunologically it was found that species specificity is an individual characteristic of the keratin. Optimal species specificity was demonstrated only when the reduced keratin (kerateine) was allowed to react with the antiserum prepared by the injection of the homologous kerateine. Thus, immunological differences among the keratins were detectable depending on the state of oxidation or reduction of the SH groups in the proteins. Using a similar approach, Ecker and Pille- mer (1940) found immunological species differences between the ocular lens proteins of chicken and fish (pike). Similar proteins from swine and sheep lenses behaved as if identical. Reduced swdne lens protein contained 8.5 per cent sulfur, and those from the other three animals from 4.3 to 5.2 per cent. In anaphylactic tests, Markin and Kyes (1939) found species differences between the lens proteins of dog and beef and that from the pigeon lens. The above results would seem to show that sulfhydryl and disulfide groupings play significant roles in certain enzyme reactions as well as in the specificity of certain antigenic proteins and toxins. 204 IMMUNO-CATALYSIS 2. Identity of the Nature of the Inhibition of Enzymes by Homologous Antibodies with Antigen-Antibody Reactions A consideration of the results of various studies on the specific side groupings of proteins in relation to their various biological activities fails to establish a direct causal relationship. It is true that modifica- tions in certain groups result in reversible inactivations, but what other changes underlie such modifications are not known. It must be remem- bered that the integrity of the whole protein molecules to which various side groups are attached, and not the groups by themselves, determines the complex mechanism of the biological activities of pro- teins. For example, the relation of the SH groups in succinoxidase does not bear even a suggestion of similarity to their role in the activa- tion of papain or the activity of pneumococcal hemolysin. We do not know as yet what particular configuration or groupings in one protein enable the heme group to function as cytochrome oxidase, and in another protein enables it to act as a catalase, or cytochrome c. What makes a certain protein to enable pyridine-adenine-dinucleotide to function as coenzyme for lactic acid dehydrogenase and another protein enables it to act as coenzyme for phosphoglyceraldehyde dehydrog- enase? Similar difficulties arise when we consider the groupings involved in the combinations between antigens and antibody. There is no doubt that certain specific groupings in these reactants make them mutually attractive and a union takes place. The suggestion of Heidelberger and Kendall (1929), and the experimental data provided by the studies of Chow and Goebel (1935), and Goebel and Hotchkiss (1937), as we have discussed elsewhere, contribute to the support of the idea that the positively charged -NHs groups in antibodies and the negatively charged -COOH groups in antigens might be involved in the com- binations between antigens and antibodies. According to Pauling (1945) hydrogen bond linkages may arise from reactions involving positively charged imino, =NH, and negatively charged carboxyl, -COO~, groups forming R-N-H .... O-CO-R bonding. It is sug- gested that the forces of these bondings play a role in keeping the proteins in their specific configurations. From the results of haptenic inhibitions of antigen and antibody reactions, it has been suggested (Pressman, Bryden and Pauling, 1948) that the principal forces of ANTI-ENZYME IMMUNITY 205 attraction between a positive charge in the antibody and negative charge of the carboxyl of a haptene, presumably forming a hydrogen bond, are responsible for their union. In his discussion on the denatura- tion and properties of various protein groups, Anson (1945) considers the hydrogen bond theory plausible on general chemical grounds, but it has not as yet any direct experimental basis. It is evident also in these reactions between the positively charged ammonium groups of antibodies and negatively charged carboxyl groups of antigens, with or without the formation of hydrogen bonds, that these side groupings do not themselves carry the specificity of the reactions of the complex reactants to which these side groupings are attached. There is no specific information as to what properties of the configuration of an antigen enable its carboxyl groups to react specifi- cally with the ammonium or imino groups of the homologous anti- body, and do not permit them to react with the same groups in heterol- ogous antibodies. The question of whether an antibody which specifically inhibits the biological activity of an antigen contains specific configuration evolved in response to the stimulus by the active groups of the specific protein molecule, or whether the observed inhibition is the result of a secondary reaction associated with the union of antigen and antibody may now be considered. It is a known fact that the toxicities of anti- gens can be eliminated with non-specific agents without loss of their ability to produce antibodies possessing the ability of combining with both the inactive and active forms of the antigen. The unanswered question is the mechanism by which inactivated antigen is capable of producing an antibody which neutralizes the toxicity or the enzyme activity of the antigen. Ramon (1943) inquiring into this question experimented with papain. As Achalme (1901) had previously re- ported, Ramon observed that papain acts like certain bacterial toxins and venoms. The papain solutions filtered through porcelain and in- jected subcutaneously into guinea pigs, rabbits, or a horse produced local disorders. Oedema and inflammation were followed by the forma- tion of a scar. Papain in large amounts caused the death of animals. Treated with formaldehyde at 45 °C. papain lost its in vitro enzyme activity and in vivo toxicity. Detoxified papain produced antibody which flocculated and neutralized the toxicity and the known activity of the enzyme, in a manner comparable to the properties of diphtheria 206 IMMUNO-CATALYSIS toxin, toxoid and antitoxin. He was unable to offer any explanation for this anomalous reaction. Stanley (1936) made similar observations with tobacco mosaic virus and arrived at the conclusion that the precipitin reaction which has been used as a measure of virus activity, may not be used unreservedly as a measure of virus activity, for in the case of inactive protein there is no correlation between the precipitin titer and virus activity. His inactivated virus produced immune sera which neutralized the viral activity. In the consideration of any of the systems belonging to the above category, we must look into the presence or absence of a direct rela- tionship between enzyme and antibody producing activities of the proteins and their abilities to combine with respective antibodies with or without the neutralization of their biological activities. While com- bination of an antigen with its homologous antibody and other com- binations with serologically non-specific inhibitors are stoichiometrical reactions, the mechanism which governs the enzyme and antibody producing activities of proteins are of catalytic nature. Stable stoichio- metrical combinations with the specific active groups would be ex- pected to block simultaneously all of the above named activities of a given protein. It is, perhaps, for this reason also that in the form of antigen-antibody complex an antigen may fail to produce a satis- factory, if any, amount of antibody. In other words, when the specific groups are completely blocked by specific antibody not only enzyme activity, but also its ability to catalyze the production of specific anti- body is inhibited. After reviewing the literature and on the basis of his own experimental findings, Olitzki (1935) reported that when the receptor groups of an antigen are saturated with antibody the treated antigen is deprived of its capacity to develop specific antibodies. He found that the injection of sensitized antigen together with free antibody suppresses the formation of antibodies to 10 to 20 per cent of the amount obtained with injections of antigens without serum and the rate of reproduction is much slower. When larger amounts of antibodies are added, then the formation of agglutinins can be completely stopped. As sensitized bacteria in vivo and in vitro can be phagocytized, sensitized pneumococci and toxin- antitoxin complexes can be attacked by proteolytic enzymes liberating the cells or toxins without loss of activity, the failure of antigen in the form of the antigen-antibody complex to stimulate the production of ANTI-ENZYME IMMUNITY 207 antibody would not appear to be due to an in vivo blocking of the antigenic groups from exercising their activity. There seems, therefore, to exist a striking interrelationship between the specific combining abilities of antigens and their catalytic activities. There is, however, the fact that when an antigen is proteolytically reduced to split prod- ucts the enzymic and antigenic properties are lost without parallel loss of combining abilities. The split products merely function as hap- tens. Here again, the point to be remembered is that the catalytic ac- tivity of the intact molecule is the principal factor which enables its potential haptenic groups to invoke complementary parts in the anti- body molecule during its synthesis. The interrelationships cited do not appear to account for the loss of enzyme (or toxic) activity without loss of the ability to produce neutralizing antibody when the toxin or enzyme is treated with non- specific agents such as ketene or formaldehyde. This discrepancy would seem to be more apparent than real. The following possibilities can be considered to account for this discrepancy: First; The neutralization of the enzyme or toxic activity of a protein by its specific antibody may be a secondary effect related to the union of the protein as antigen with the antibody which has been produced in response to the inactivated but antigenic protein. If formaldehyde, ketene etc., can cause inactivations, one could postulate that inactiva- tion of the enzyme and toxic properties would automatically result when the antigen-antibody combining reaction occurs. Such combina- tions produce neutralization of the respective negative and positive charges, and a decrease in the energy content of the reactive groups. Consequently, possible deleterious effects on other regions of the active protein molecules might transform the enzyme molecule into an inactive form in the combined state.* There are no experimental data to support or to refute such an interpretation of the observed effect. The discussed properties are so interrelated that it is difficult to characterize which is cause, which is effect. As in the oxidation of SH groups, or the reduction of -S-S- groups resulting in the ^Stanley (1936) reported that the precipitation reaction, which has been used as a measure of virus activity, may not be used unreservedly for this purpose. Kassanis (1943) reported that normal and heterologous sera cause marked neutral- ization of plant viruses— tobacco mosaic virus, tomato bushy stunt virus and two cultures of tobacco necrosis viruses. The additional specific effect of homologous antisera was small in comparison. Unless sera were kept frozen their non-specific neu- 208 IMMUNO-CATALYSIS inactivation of the enzymes or toxins, one cannot define what con- figurational and other more critical changes in the whole molecule precede or follow such reactions. We must, therefore, inquire into other possibilities which may account for the above mentioned dis- crepancy. Second: The second possibility is the reactivation in vivo of the in vitro inactivated enzymes and toxins, etc. The amount of the reactivated molecules produced at a given time might be insufficient to produce noticeable toxic effects but sufficient to produce specific antibodies. The data concerning this point are not as yet adequate, but whatever can be cited is strongly suggestive of this process. For example, urease oxidized with dilute iodine is inactive. This inactiva- tion, probably involving the oxidation of SH groups, is reversed by sulfhydryl groups. As discussed above, Pillemer, et al. (1938) showed that although oxidized urease has lost its specificity to react with antiserum to crys- talline active urease, both the oxidized and the reduced urease produced a similar antibody. The ability of the oxidized urease to produce such a specific antibody was attributed to the fact that tissue extracts were ca- pable of reactivating oxidized urease, showing that the oxidized urease regains its lost specificity in the animal system. Similarly, the detoxifi- cation of snake venom involves the reduction of di-thiol groups into sulfhydryl groups. In vivo the reduced molecule can regain to some degree the original form. The reaction of formaldehyde Math the imino, amino, amide, hydroxyl and sulfhydryl groups are reversible reactions, though some are more stable than others. Which of these reactions predominate in the conversion of toxin into toxoid is not known. It is assumed, but contested, that formaldehyde reacts prin- cipally with amino groups. If this is the predominant reaction the resulting hydroxymethyl compounds are unstable and reversible. The toxoid molecules can be considered subject to reversion in vivo to the original molecular form. Acetylation with ketene likewise would seem tralizing power fell rapidly on stirring. All heterologous antisera reduced infectivity more than normal sera stored similarly. Precipitating antibodies did not appear to be responsible for neutralization. No correlation was found between precipitation titre and neutralization power, and removal of precipitins did not affect neutralization power. Only quantitative differ- ences were found in behavior between homologous and other sera; the infectivity of all virus serum mixtures was regained by dilution. ANTI-ENZYME IMMUNITY 209 to produce unstable derivatives. As referred to above, Boor and Miller (1939) found that ketenized gonococci after standing a week in the cold fully regained their toxicity to mice. In the light of v\'hat we know about the above discussed reversible inactivations, one is inclined to justify the assumption that this process occurs provided the protein under consideration has not been subjected to greater changes. The many observed inhibitions of enzymes and toxins by their specific antibodies may lend themselves to two types of interpretations. These inhibitions may be due to direct combination between the spe- cific groupings in the enzyme and anti-enzyme antibody molecules. If such is the case it is immaterial whether or not the substrate molecule can penetrate the immune complex to reach the site of the active enzyme groups. The substrate could not be activated under these con- ditions. Or, the antibody molecule may not contain groups specific for the enzyme, but the antibody molecules occupying positions on the surface of the enzyme may be so closely packed that the substrate molecules are incapable of reaching the site of the active enzyme or toxin groups. There are no experimental data to show that antibody molecules on the surface of antigen constitute a mechanical barrier to the sub- strate molecules. Before such an argument can be deemed worthy of consideration it would be desirable that experimental results be pro- vided. One may perhaps be able to work out conditions for the treat- ment of the enzyme-antienzyme complexes with acetylating and other agents containing tagged isotopic atoms. On separating the antigen from the antibody, the relative amounts of the tagged agents in each component could be estimated. Using the size of the reagents as a measure for the space available between the antibody molecules on the surface of antigen one may be able to gain some information concern- ing this question. On the basis of various experimental data it would seem difficult to conceive that antibodies combining with an enzyme molecule con- stitute a wall impermeable to the specific substrates. In the reaction of diphtheria toxin with antitoxin, in the presence of extreme excess of the latter, at least eight molecules have been shown to combine with one molecule of toxin. The composition of toxin-antitoxin soluble com- plex in the zone of toxin excess has been reported to consist of only one 210 IMMUNO-CATALYSIS antitoxin molecule for two of toxin (Pappenheimer, 1940; Pappen- heimer, Lundgren and Williams, 1940). Boyd (1947) has tabulated molecular compositions of various antigen-antibody complexes. In the presence of extreme excess of antibody, serum albumin can combine with six rabbit antibody molecules; one molecule of thyroglobulin (mol. wt. 650,000) combines with forty rabbit antibody molecules; ovalbumin with four (horse) or five rabbit antibody molecules; and nine hundred rabbit antibody molecules with one of tobacco mosaic virus. The high molecular composition of certain antibodies with the respective antigens is understandable if we take into consideration the molecular weights of these antigens. The molecular weight of tobacco mosaic virus is 60 million. The molecular weights of diph- theria toxin and serum albumin are about 70,000, and that of ovalbu- min about 40,000. Since the size of rabbit antibody globulin to various antigens is constant, the critical factor in the molecular composition of antigen and antibody complexes is the molecular size and the surface areas of antigens. The ratios of the molecular weight of the tobacco mosaic virus to that of the serum albumin or diphtheria toxin is about 860. One virus particle combining with 900 rabbit antibody molecules would be equivalent to a combination of one molecule of antibody and one virus unit of 70,000 molecular weight. This approximation would be valid if the shape of the virus and that of the serum albumin or diphtheria toxin are of comparable dimensions. The ratio of the major to minor axis of the diphtheria toxin is 4.7, and that of the virus is 18. Weight for weight, virus particles may therefore possess a three times greater surface area. Using this rough relationship, one can see that only a very few antibody molecules can combine with one sub- molecular unit of virus antigen. Under these conditions, the formation around the antigen molecule of a barrier impermeable to a substrate does not appear to be probable. Let us examine other data which would seem to throw some light on this question. It had been assumed that when antibodies combine with microorganisms, causing clumping or agglutination, they simply reduce the effective surface relationship between the enzyme and sub- strate (Taliaferro, 1948). Under these conditions it had been assumed that substrates are prevented from reaching the enzyme sites. Sevag and Miller (1948) found that agglutinated pneumococci with or with- ANTI-ENZYME IMMUNITY 211 out complement use just as much oxygen as non-agglutinated or control systems. The results with E. typhosa, 0-901 strain, were similar to those obtained with pneumococci. These findings show that the layer of antibody specifically deposited on the surface or cell-wall of microorganisms, with or without agglutination, do not constitute a mechanical or physical barrier to the penetration of glucose or glycerin to the active sites of enzymes if the latter are not inactivated by an antigen-antibody combination. In a study with Salmonella Harris (1948) obtained results similar to that discussed above. The question of whether or not protein or starch molecules can squeeze themselves to the sites of the specific enzymes of agglutinated bacteria through the layer of deposited antibody molecules remains to be asked. In this connection an observation by Feiner, et al. (1946) is of con- siderable interest. They studied the ability of lysozyme to attack a sub- strate (antigen) precipitated by homologous antibody. They observed an unmistakable difference in the appearance between the untreated and enzyme-treated immune precipitates. The control, untreated pre- cipitates remained as opaque white pellets, whereas in the treated series they appeared as translucent vacuolated material closely adherent to the bottom of the tube and markedly diminished in size. Antibody was released as a result of lysozyme action. They concluded that lyso- zyme is capable of attacking the organism, or its