A SYMPOSIUM ON RESPIRATORY ENZYMES A Symposium on RESPIRATORY ENZYMES Contributors: Otto Meyerhof Eric G. Ball Fritz Lipmann Kurt G. Stern Fritz Schlenk T. R. HOGNESS Elmer Stotz Carl F. Cori E. A. Evans, Jr. Philip P. Cohen K. A. C. Elliott Dean Burk C. J. Kensler Erwin Haas H. M. Kalckar M. J. Johnson Van R. Potter H. G. Wood R. H. BuRRis C. H. Werkman P. W. Wilson F. F. NoRD Ephraim Shorr a. e. axelrod Frederick Bernheim E. S. G. Barron Fredrick J. Stare The University of Wisconsin Press MADISON Copyright 1942 by the UNIVERSITY OF WISCONSIN All Rights Reserved PRINTED IN THE UNITED STATES OF AMERICA Foreword IN 1897 Buchner published his classical study on alcoholic fermen- tation by cell-free yeast juice. In the same year Eijkman con- cluded that beri-beri among the natives of the Dutch East Indies was caused by a dietary deficiency arising from the use of polished rice. These two discoveries may be said to have initiated the modern investigations in two of the most important fields of biochemistry and medicine: the nature of the respiratory enzymes and the function of the vitamins in cellular metabolism. For the next thirty years research workers in these two fields pursued their investigation almost independently of one another and more or less oblivious to the progress being made in the other's field. Then in the early part of the last decade it was discovered that ribo- flavin was the functional group in a respiratory enzyme and very soon afterward this compound was shown to be vitamin Bg (G). Further discoveries of a similar nature soon demonstrated that the enzyme chemist and the nutritionist were to a great extent prospect- ing the same territory. The time seemed to be ripe, therefore, for the two groups to join in a discussion of the latest advances. Such a meeting on the "Respiratory Enzymes and the Biological Action of the Vitamins" was sponsored jointly by the Universities of Wisconsin and Chicago, institutions that have long been leaders in these fields. This book contains the lectures and discussions given at the Uni- versity of Wisconsin. It deals with the fundamental nature of those enzymes that are intimately connected with the functioning of the vitamins. Informative presentation of the latest developments, in- terpretation of past and present findings, and indication of some of the problems still unsolved in respiratory enzyme research are given by recognized international authorities in the field. Supplementing these explanations of the fundamental nature of respiratory enzymes are discussions applying the findings to specific problems. The Program Committee wishes to thank the many members of the faculty for their cooperation in arranging the meetings held at the University of Wisconsin, the speakers for their papers and dis- cussions, and the Wisconsin Alumni Research Foundation for the grant which made these sessions and the publication of the present volume possible. Address of Welcome C. A. DYKSTRA President of The University of Wisconsin THE University of Wisconsin is a happy host today. It welcomes to its campus scientists from many laboratories who are drawn together for the discussion of common problems and common aims. It recognizes in this symposium the challenge that faces intelligent men of good will everywhere— the great need there is in the con- temporary world for sitting down and reasoning together. From such a process comes progress. We are concerned here with functions which operate in the biological and chemical world. We seek these out and discover how they work so that, knowing about them, we may cooperate with nature for the good of man. This we do by observation, experimenta- tion, analysis, and, finally, the objective setting down of results that may yield a pattern or a principle. As we look about us and see biological specimens called men reacting to special or group inter- ests as passion and selfishness may happen to dictate, we ask our- selves, a bit dismally perhaps, whether the statesmen and public leaders of the world can ever be persuaded to try out the scientific method as an approach to the problems of world organization. We also need desperately a healthy society and a sound international body. Here, today, we pay tribute to the internationalism of science. As we scan our program for the week we are struck by the fact that men from different backgrounds and from many nationalities and races can come together peacefully in a symposium to present the results of long years of human effort in a field of science, check these results, and try to establish what they mean or may mean to life on this planet. Today and right here men labor together who, were they still living in their family homelands, would be enemies, legally and politically. This is the great modem paradox— that as the world of communication has made the globe a unity and as the domains of science, hterature, music, art, commerce, and industry have become international, we have at the same time the phenomenon of a more bitter nationahsm than ever before. Something is wrong that needs vii viii ADDRESS OF WELCOME early correction, and intelligent men must give attention to the chal- lenge. We meet today to talk of many things in the wonderland of science. We have the special vocabulary necessary for the accuracy of our thinking and investigation. This vocabulary is a closed book to the man in the street except for a few words, such as vitamin, for instance. This man in the street, however, does get a partial implication of your work as he hears or reads the advertiser who expounds the merits of certain food products. He may even be led to think that he can be a vigorous and whole man if only he has a box of pills or capsules in his vest pocket. He may even be duped or exploited because of this partial knowledge. We therefore have the obligation in our special fields of science which promise so much to all to attempt such simplification and general statement that those things for which we can vouch will become common knowledge at the earliest possible moment. Just now there is a wide spread of interest in many areas which this symposium deals with. It is a good time, therefore, to capitalize on this popular interest, for we have a receptive pubhc. We of the public are willing and anxious to learn from you. We here at Wisconsin are glad you are with us. We are happy too in the cooperation of our sister institution, the University of Chicago, in the enterprise here represented. Our welcome is genuine, and we wish for the conference unusual and distinguished success. Contents INTERMEDIATE CARBOHYDRATE METABOLISM By Otto Meyerhof, University of Pennsylvania . Introduction, 3 Inadequacy of older views of the relation between aerobic and an- aerobic carbohydrate breakdown, 4 Alternate pathways of oxidation, 4 The oxidation quotient, 5 Competition of oxygen and pyruvate in oxidation of dihydrocozymase, 6 Aerobic phosphorylation, 7 Reversibility of the glycolysis re- actions, 9 Formation of phosphopyruvate from pyruvate through oxalacetate, 12 Inhibitors, 14 OXIDATIVE MECHANISMS IN ANIMAL TISSUES By Eric G. Ball, Harvard Medical School .... Introduction, 16 Oxygen activation and the cyto- chrome system, 17 Substrate activation and the pyri- dine nucleotide and flavoprotein systems, 18 The energy relationships of these systems and the transfer of elec- trons between them, 21 The relative concentrations of cyto- chrome c, diphosphopyridine nu- cleotide, and flavoprotein in animal 16 tissues and their possible signifi- cance, 25 The possibility of the existence of pathways alternate to the cyto- chrome system, 26 Cyanide poisoning of respiration and a theory of its mechanism, 27 An interpretation of respiratory mechanisms in tlie arbacia egg, 30 Azide poisoning of respiration and a possible interpretation of its ac- tion on muscle respiration, 30 DISCUSSION ON HYDROGEN TRANSPORT Van R. Potter, chairman 33 The Possible Role of Intermedi- ary Metabolites as Hydrogen Carriers by K. A. C. Elliott, 33 The Role of the Carriers in Dis- MUTATIONS AND CoUPLED OxiDO- REDUCTIONS, WITH SPECIAL REFER- ENCE TO THE Flavoproteins by E. G. Ball and F. Lipmann, 38 The Physico-Chemical Mechanism OF Hydrogen Transport by Kurt Stem and Erwin Haas, 42 Possibility of a By-Pass around the Cytochrome System by El- mer Stotz, 46 PASTEUR EFFECT By Fritz Lipmann, Massachusetts General Hospital 48 The efficiency of aerobic and an- aerobic metabolism, 50 The metabolic structure of cells, 52 Bacteria, 53 Animal tissues, 56 Interpretation of the Pasteur effect, 59 Equilibrium schemes, 59 Inhibition of the Pasteur effect, 62 Reversible oxidative inhibition of glycolysis in extracts, 65 Thiol influence on fermentation and glycolysis in intact cells, 67 Pasteur effect with very low res- piration, 69 Conclusion and outlook, 70 CTCf /s r^f**j CONTENTS OXIDASES, PEROXIDASES, AND CATALASE By Kurt G. Stern, Yale University School of Medicine .... 74 Introduction, 74 On the mechanics of hemin catal- The common denominator in hemin yses, 86 catalyses, 74 Autoxidizable iron compounds, 96 Enzyme-substrate intermediates, 84 Oxygen transfer in living cells, 99 NICOTINAMIDE NUCLEOTIDE ENZYMES By Fritz Schlenk, School of Medicine, University of Texas 104 Historical introduction, 104 Codehydrogenase I and II, 104 Occurrence, 105 Preparation, 106 Properties, 108 Investigation of the structure of co- dehydrogenase I and II, 109 Reversible reduction, 109 The "model compounds" of Kar- rer, 111 Nicotinamide nucleoside, its prepa- ration and properties, 113 Structure of codehydrogenase I and II, 116 Methods of determination, 120 Apodehydrogenases dependent on nicotinamide nucleotides, 122 Substrate and coenzyme specificity of apodehydrogenases, 123 Mode of action of nicotinamide nucleotide enzymes, 125 Spectrophotometric methods, 126 Biosynthesis of the codehydrogenases I and II, 126 THE FLAVOPROTEINS By T. R. Hogness, University of Chicago 134 Historical introduction, 134 Comparison of activities, 142 The general properties of die known Cytochrome a reductase: test, spec- yellow enzymes, 137 troscopic demonstration, and prop- Dissociation constants, 139 erties, 144 CYTOCHROMES By Elmer Stotz, Harvard University Properties of the cytochrome com- ponents, 149 Cytochrome c: isolation, purifica- tion, structure, and oxidation- reduction potential, 149 Cytochrome b, 153 Cytochrome a, 154 Oxidation-reduction potential of the cytochromes in yeast, 154 Cytochrome oxidase and cytochrome 03, 155 Identity with Warburg's enzyme, 157 Summary of absorption spectra of the cytochromes, 158 Copper-containing oxidase, 159 Soluble cytochrome c peroxidase, 160 149 Physiological reduction of the cyto- chromes, 161 By the succinate system-extra fac- tor, 161 By flavoprotein (cytochrome re- ductase), 163 Catalytic relations of the cyto- chromes and oxidase: oxidation of hydroquinone and p-phenylene dia- mine, 164 Determination and distribution of cytochrome c and cytochrome oxi- dase, 166 Physiological functioning of the cy- tochrome system, 168 Possibility of a by-pass of the cyto- chrome system, 169 CONTENTS xi PHOSPHORYLATION OF CARBOHYDRATES By Carl F. Cori, Washington University School of Medicine, St. Louis 175 Introduction, 175 Transphosphorylation, 181 Uptake of inorganic phosphate, 175 Regeneration of inorganic phosphate. Intramolecular migration of phos- 185 phate, 179 Summary, 188 DISCUSSION ON PHOSPHORYLATION H. M. Kalckar, chairman 190 Myokinase; dephosphorylation By H. M. Kalckar, 190 Dephosphorylating and transphos- phorylating enzymes By Otto Meyerhof, 192 Energy utilization mechanisms By M. J. Johnson, 194 Evidence for acetylphosphate occur- rence By Fritz Lipmann, 195 METABOLIC CYCLES AND DECARBOXYLATION By E. A, Evans, Jr., University of Chicago .... Introduction: cycle reactions in bio- logical systems, 197 Krebs' citric acid cycle, 198 Experimental basis of the citric acid cycle, 198 The "Krebs reaction," 200 The eflFect of malonate on tissue respiration, 201 Evaluation of the reactions of the citric acid cycle, 201 Criticisms of the theory, 202 197 The citric acid cycle in pigeon liver, 203 The synthesis of a-ketoglutaric acid, 203 The assimilation of CO2 in the syn- thesis of a-ketoglutaric acid by pigeon liver, 204 The mechanism of a-ketoglutaric acid synthesis, 204 The nature of CO2 assimilation by pigeon liver, 205 TRANSAMINATION By Philip P. Cohen, University of Wisconsin 210 Types of transamination, 210 Discovery of transamination reac- tion, 211 Substrates active in transamination, 211 Amino and keto acids, 211 Peptides, 212 "Primary" and "secondary" sub- strates, 213 "Catalytic" transamination, 213 Preparation and properties of trans- aminating enzymes, 214 Aminopherases, 214 Transaminase, 215 Substrate specificity, 215 Mechanism of transamination, 216 Transamination in diflFerent tissues: Animal tissues, 218 Malignant and embryonic tissues, 219 Plant tissues, 220 Yeast and bacteria, 221 Transamination in vivo, 221 Influence of various substances on transamination, 221 Inhibitors, 221 Hormones, 222 Carcinogens, 222 Vitamins, 222 Role of transamination in intermedi- ary metabolism, 223 Protein and amino acid synthesis and degradation: animal tissues, plant tissues, transamination and glycolysis, transamination and hy- drogen transport, 223 CONTENTS DISCUSSION ON TUMOR RESPIRATION C. A. Baumann, chairman 229 Characteristics of Tumor Res- piration by K. A. C. Elliott, 229 Phosphorylation Theories and Tumor Metabolism by Van R. Potter, 233 On the Specificity of Glycolysis IN Malignant Liver Tumors as Compared with Homologous Adult or Growing Liver Tissues by Dean Burk, 235 The Effects of Certain Diamines on Enzyme Systems, Correlated WITH THE Carcinogenicity of THE Parent Azo Dyes by C. J. Kensler, 246 DISCUSSION ON BACTERIAL RESPIRATION W. H. Peterson, chairman 252 Criteria for Experiments with Isotopes discussed by H. G. Wood and R. H. Burris, 252 Mechanisms for the Complete Oxidation of Carbohydrates by Aerobic Bacteria discussed by C. H. Werkman, E. S. Guzman Barron, and P. W. Wilson, 258 Reactions in Cell-Free Enzyme Systems Compared with Those IN the Intact Cell discussed by F. F. Nord and P. W. Wilson, 264 DISCUSSION ON ANIMAL TISSUE RESPIRATION C. A. Elvehjem, chairman 268 Factors Affecting the Prepara- tion of Tissue for Metabolic Studies by Ephraim Shorr, 268 Comparison of Slices and Homo- genized Suspensions of Brain Tissue by K. A. C. Elliott, 271 The Homogenized Tissue Tech- nique, THE Dilution Effect and Ion Effects by Van R. Potter, 274 The Stimulatory Effect of Cal- cium UPON the Succinoxidase Ac- tivity OF Rat Tissues by A. E. Axelrod, 275 Tissue Metabolism in Vitro and in Vivo by Frederick Bemheim, 276 Pathways of Carbohydrate Me- tabolism by E. S. Guzman Bar- ron, 278 The Citric Acid Cycle in Tissue Metabolism by Fredrick J. Stare, 280 A SYMPOSIUM ON RESPIRATORY ENZYMES Addresses given at an Institute Held at the University of Wisconsin September 11-17, 1941 Intermediate Carbohydrate Metabolism OTTO MEYERHOF University of Pennsylvania THIS REVIEW OF THE intermediary carbohydrate metabolism must necessarily be treated broadly and generally, for the subject has many diflFerent aspects, and the detailed questions of hydro- gen transport, Pasteur eflFect, pyridine nucleotides, cocarboxylase, metabolic cycles, phosphorylations, indeed all the items which are intrinsic elements of the present picture of carbohydrate breakdown, will be dealt with by competent investigators of these subjects. Moreover, I had the opportunity to discuss the special question of oxidoreduction and dismutation in carbohydrate metabolism at the Chicago congress some months ago. If we take this occasion to look back fifty years and to compare our present knowledge with that which existed at the end of the last century we have reason to be very proud, for at that time this whole field appeared nearly as tabula rasa. But two outstanding achievements had already been accomplished: first, Claude Ber- nard's work on the interconversion of glucose and glycogen in the liver and on the role of blood sugar under normal and diabetic conditions; second, the work of Pasteur on the different microbic fermentations as manifestations of the anaerobic metabolism of these organisms. Nothing was known about the oxidative break- down of sugar. Although lactic acid formation in the blood and especially in the muscles had been observed by Claude Bernard and others, it was not known whether nor how this cleavage was connected with respiration. Since then the interconversion of glycogen and blood sugar have continued to claim the attention of medical investigators, and re- cently, as you know, a highly interesting development was reported: Professor Cori's discovery of glucose-1-monophosphoric acid as inter- mediary. The old problem of diabetes was shifted by the isolation of insulin from the study of blood sugar regulation to the bio- chemical task of studying tissue metabolism under the influence of added hormones. The third old problem of the connection between 4 A SYMPOSIUM ON RESPIRATORY ENZYMES fermentation and respiration remained for a long time a subject of speculation, and even now many a question is unanswered. How- ever, I should like to follow this latter trend of ideas a little more in detail. PfeflFer and Pfliiger, following Pasteur, held a rather simple view of this relationship: the first step of respiration was assumed to be always anaerobic. If no oxygen is present, the products of anaerobic cleavage accumulate: alcohol in yeast and higher plants, lactic acid in the tissues of higher animals and in some bacteria. But if oxygen is present, these products are oxidized to carbon dioxide and water. That this concept required modification became apparent twenty years ago from studies of metabolism of muscle. In 1907 Fletcher and Hopkins (1) showed that under anaerobic conditions frog muscles formed lactic acid steadily during both activity and rest, and that this lactic acid disappeared when oxygen was admitted. Parnas (2), working some years later in the same Cambridge laboratory, claimed to have found that this disappearance was a complete oxidation, thus apparently confirming the views of Pfeffer and Pfliiger. In 1920, because of the controversial state of this question, I repeated the experiments of Parnas, avoiding especially all kinds of irritation or injury of the muscles which would lead to extra-consumption of oxygen (3). Under these conditions much more lactic acid disap- peared in oxygen than could be accounted for by oxidation, and the lactic acid unaccounted for was reconverted into carbohydrate. This was true for the lactic acid formed during activity as well as for that formed during rest. Similarly, it was shown that in equal periods of rest much more lactic acid was formed anaerobically than could be burnt aerobically by the resting respiration. Indeed, the amount of oxygen which failed to be used in a period of anaerobiosis was about the same as the excess consumed after that period. This oxygen was suflBcient only to oxidize from a quarter to a sixth of the lactic acid which disappeared. These facts, which are independent of special interpretations, are sufficient to invalidate the original theory of Pfeffer and Pfliiger in that they show that the oxidative removal of fermentation products is not necessarily identical with the oxidation of these products. But we can pose the more limited question whether the oxidation on the whole attacks the end products of anaerobic breakdown. With respect to lactic acid formed in a preceding anaerobic period, we must surely answer in the affirmative. We know that lactic acid is easily oxidized by way of pyruvic acid. For example, Barron et al. INTERMEDIATE CARBOHYDRATE METABOLISM 5 (4) showed that specially treated, washed bacteria may lose the power to oxidize sugar and other substrates, but retain the power to oxidize lactic to pyruvic acid. Experiments on muscle lead to the same conclusion. After a muscle is poisoned with iodoacetic acid the formation of lactic acid is blocked; at the same time the respiratory quotient drops to 0.7, and is not changed by the addition of sugar, but is brought to 0.95 by the addition of lactic acid. Respiration is increased, and oxygen consumption is essentially equivalent to the disappearance of lactic acid (5). Similar results were obtained by Krebs with respiration of brain and testis after poisoning with iodoacetic acid (6). Since oxidation of sugar is completely checked, no interpretation is possible except that lactic acid is directly oxi- dized. But this is not necessarily the pathway of sugar oxidation in the aerobic steady state. That independent ways of sugar oxidation exist may be gathered from many observations, such as the rapid oxidation of fructose in brain tissue, where, in contrast to glucose (7), it does not give rise to anaerobic lactic acid. Furthermore, Warburg and Christian showed that hexosemonophosphate can be oxidized by the triphosphopyridine nucleotide in yeast extract to phosphogluconic acid (8), and Lipmann demonstrated the complete oxidation to carbon dioxide in this manner (9). On the other hand, the oxidation of sugar by way of pyruvic acid is also firmly established, and in this case the steps up to the forma- tion of the acid are identical in respiration and in anaerobic gly- colysis. As was discovered by Peters (10), pyruvic acid accumulates during oxidation of carbohydrate by cells and tissues in cases of vitamin B^ deficiency, which means that lack of cocarboxylase blocks the oxidative decarboxylation of pyruvic acid. Many other findings, such as the similarity of the oxidation of pyruvic acid to that of sugar in tissue pulps and extracts, point in the same direction, namely, that sugar is oxidized via pyruvic acid (11). Thus several pathways of sugar oxidation exist, the choice of which may depend upon the special set of enzymes in different tissues and also upon hormonal and other controlling influences. All this probably has some bearing on the relationship already mentioned between oxidation and interference with the mechanism of fermentation. I have mentioned before the two possible cases of this relationship— the actual synthesis of split products to the initial substance and the non-formation of the split products during the sta- tionary state of respiration. Without fearing to be accused of a e A SYMPOSIUM ON RESPIRATORY ENZYMES biased judgment I dare say that both cases are characterized by the same numerical relationship— the oxidation quotient, which expresses the ratio of the aerobic disappearance of splitting metabolism in moles sugar to the oxidized sugar equivalents (12). Critics have objected that under extreme conditions this number may range from zero to infinite, but it is equally true, and more important, I think, that under physiological conditions living cells exhibit quo- tients between 3 and 6— approaching 6 more and more as the con- ditions of temperature, oxygen pressure, nutritional state, and milieu become optimal for the cells in question. The same preference for the quotient of 6 was demonstrated by O. Warburg for different warm-blooded tissues where the anaerobic glycolysis is high enough to allow the calculation of the quotient (13). The original concept of a metabolic carbohydrate cycle involved the assumption that in the stationary state the quotient results from a continuous overlapping of anaerobic glycolysis and of oxidative resynthesis of the cleavage products— the endothermic resynthesis made possible by coupling with oxidation. Today it seems possible to refine this scheme and to modify it somewhat without rejecting the main argument. Indeed, in the past fifteen years a tremendous amount of material has been collected to prove that the general concept of these cycles in carbohydrate breakdown holds good, that every oxidative step is coupled with an involuntary phosphoiy- lation, and that the several intermediate stages of the anaerobic breakdown can be reversed by means of the "energy-rich phosphate bonds" (31) created in this way. On the other hand, the original concept of a single complete cycle passing through the stage of lactic acid cannot be exactly true for a very simple reason, which has become clear since 1933; namely, that pyruvic acid is the necessary precursor of lactic acid in glycolysis and of alcohol in yeast fermentation (14). Under anaerobic conditions the reduction of pyruvic to lactic acid is compensated for by the oxidation of phosphoglyceraldehyde to phosphoglyceric acid. The latter, in turn, is decomposed via two intermediaries to pyruvic acid (15). The hydrogen transfer proceeds in both directions by the way of cozymase, the diphosphopyridine nucleotide of War- burg. But if oxygen is present the dihydrocozymase can transfer its two hydrogen atoms to oxygen instead of to pyruvic acid by a long chain of oxidative catalysts : the pheohemin enzyme of Warburg, the three cytochromes, and the flavinproteins; consequently the pyruvic acid INTERMEDIATE CARBOHYDRATE METABOLISM 7 is not reduced. On the contrary, such an oxidation of dihydro- cozymase shifts the equihbrium in the opposite direction, so that lactic acid, if present, would be oxidized by cozymase to pyruvic acid, whereas in the stationary state of sugar oxidation pyruvic acid would be continuously formed by way of phosphoglyceric acid, without a compensating reduction. Therefore only pyruvic acid, and not lactic acid, is formed in the stationary state of oxidation. This interpretation at the same time gives a clue to the oxidation quotient, the numerical relationship between the oxygen consumed and the lactic acid that is prevented from being formed: if one atom of oxygen is required to oxidize the two hydrogen atoms of dihydrocozymase, then this atom prevents one niolecule of pyruvic acid from being reduced to lactic acid or in yeast fermentation to alcohol. Therefore six atoms of oxygen (corre- sponding to the complete oxidation of one molecule of lactic acid) can prevent six molecules of lactic acid from being formed, and we obtain the normal oxidation quotient of 6. Of course this refers only to the principle. The cozymase reoxidized by oxidative catalysts must dehydrogenate other intermediary stages besides triosephos- phate, because every oxidative step in the breakdown of sugar acts in the same way, preventing the formation of one molecule of lactic acid per one atom of oxygen taken up. And this is only one side of the picture. If the breakdown of sugar in oxygen and in nitrogen proceeded with the same speed to the stage of pyruvic acid, and the only difference consisted in the fate of pyruvic acid to be reduced or further oxidized, then the oxidation would not prevent, as it actually does, by this so-called "Pasteur effect," the greater part of sugar from disappearing. But here the concept of metabolic cycles has its place. Actually every oxidative step is coupled with the phosphorylation of the adenylic system, and by this means a corresponding phase of anaerobic breakdown is reversed, so that for every oxygen atom consumed one three-carbon molecule can return to its initial stage as sugar or glycogen. This state of affairs is very neatly shown by the recent experiments of Cori, Kalckar, and co-workers (16) with dialyzed extracts of kidney and heart, and by experiments of Behtzer and Tzibakowa (17) with washed pigeon muscle. Cori and his group found that in the pres- ence of the complete glycolytic coenzyme system the organ extracts oxidize glucose and phosphorylate an excess of it, so that for every hexose molecule burned to carbon dioxide, ten molecules of phos- phate are taken up to form five molecules hexosediphosphate; and 8 A SYMPOSIUM ON RESPIRATORY ENZYMES since the oxidized molecule also had to be phosphorylated, altogether twelve molecules of phosphates are taken up for one molecule glucose or twelve oxygen atoms consumed. Therefore every step of glucose oxidation consisting in an oxidoreduction between cozymase and an oxidizable intermediary is coupled with phosphorylation. Not only is this true for the two steps where it is already known, i.e., the oxida- tion of phosphoglyceraldehyde and that of pyruvic acid, in which Lipmann discovered acetylphosphate as the primary product of oxidation (18), but for every such step an energy-rich phosphate bond is created in adenosinetriphosphate, which enables a synthetic step to take place. The experiments of Belitzer and Tzibakowa are a little dijBFerent, because they added creatine to cut muscle and obtained under these conditions a synthesis of creatinephosphate when lactate, pyruvate, or the four-carbon acids of the Szent-Gyorgy cycle were oxidized. At the most two molecules of creatinephosphate were formed for every oxygen atom taken up. Although the presence of creatine diverts the pathway of synthesis from carbohydrate, the experiments are important in that they demonstrate the uptake of two molecules of phosphate by way of adenosinetriphosphate for one atom of oxygen consumed; this relationship is comparable to the synthesis of creatinephosphate in muscle extract, where two steps of glycolysis are involved in the transfer of phosphate, namely, the oxidoreduction and the dephosphorylation of phosphopyruvic acid (19). Moreover, the reaction studied by Belitzer is closely analogous to the recovery period of the living muscle, especially a muscle which is only slightly fatigued. Here, during oxidative recovery, the oxidation serves mostly for the resynthesis of creatinephosphate, and to a small extent for that of glycogen. If two molecules of creatinephosphate are synthesized for every atom of oxygen taken up, then about 40 per cent of the combustion heat of sugar or lactate is consumed for the endothennic synthesis, a result which comes very close to the efficiency of the oxidative recovery in the living muscle.* But to return from this digression to the significance, already men- tioned, of the experiments for the theory of carbohydrate cycles. One objection may be raised against this interpretation of the Pas- teur effect. Many cases are known where the respiration remains quantitatively the same, while the effect of the respiration on the * Actually the same ratio of two molecules of creatinephosphate syntliesized for one atom of oxygen taken up was found by O. Meyerhof and D. Nachman- .sohn (Biochem. Z., 222, 1, 1930) during recovery of a partially fatigued muscle. INTERMEDIATE CARBOHYDRATE METABOLISM 9 glycolysis is suppressed. In the picture outlined above the oxygen used would automatically eliminate an equivalent lactic acid forma- tion, in so far as the oxygen serves to reoxidize dihydrocozymase. But we must have in mind that the oxygen intervenes only indirectly by way of the oxidizing catalysts. Here the so-called "Pasteur enzyme" assumed by Warburg (20) and demonstrated by Stern and Melnick (21) plays its role in steering the oxidation. All oxidation not going by the way of cozymase would be without "Pasteur effect"; it may be oxidation of non-carbohydrate, which replaces sugar oxidation, or it may be oxidation of sugar by way of triphosphopyridine nu- cleotide. Now we come to the second half of the problem, the actual con- version of lactic acid to glycogen in the oxidative recovery of the Glycogen (starch) HaPO, If Glucose-1-phosphate (Cori-Ester) Dihydroxyacetonephosphate I- (a) -Glycerophosphate (Glycerol -1- H3PO4) Pyruvic acid + H3PO4 ■J' Phosphopyruvic acid (?) (hydrated) H2 J, Phosphoacetic acid -1- CO2 i Acetic acid -f H3PO4 Acetaldehyde -|- CO2 <- Ethyl alcohol d-Glucose + H3PO4 -* Glucose-6-phosphate Fructose-6-phosphate H3P04|f Fructose-l ,6-diphosphate -» d-3-Phosphoglyceraldehyde H3PO4JI d-1 ,3-Diphosphoglyceraldehyde d-l,3-Diphosphoglyceric acid H3PO4II 70% >30% Figure 2. — Oxidation-reduction systems concerned in biological oxidations The source of the potential values used is given in reference 2 except for the diphosphopyridine nucleotide system, which is taken from reference 5. their oxidation-reduction potentials. As shown in Figure 2, we are able to plot fairly accurately according to their potentials all the systems discussed above. Cytochrome oxidase is the chief exception, but presumably we may place it between cytochrome a and oxygen. We thus have interposed between the substrate and oxygen, reading in the order of the potential of their systems, pyridine nucleotides, flavoproteins, cytochrome h, cytochrome c, cytochrome a, and finally cytochrome oxidase. Now if cytochrome h functions in this chain, and it must be remembered that we are not certain that it does, the possibility that another system lies between it and the flavo- 22 A SYMPOSIUM ON RESPIRATORY ENZYMES proteins seems rather small. At any rate, the gap between these two systems in terms of the energy that would be released upon their interaction is only a small fraction of the overall. There seems little doubt, however, that cytochrome c and the flavoprotein systems are necessary links in the chain. No flavoprotein has yet been obtained from animal tissues that will react directly with cytochrome c. This implies that there is some link in the chain between these two con- stituents. Until cytochrome h can be isolated its claim to this posi- tion must remain in dispute. Another system that we have not yet mentioned must also be considered in this connection. Szent- Gyorgyi has suggested that the succinate-fumarate system links the flavoprotein systems to the cytochromes. It will be seen that the potential of the succinate-fumarate system is such that it could play this role. Here again, however, clean-cut proof is lacking, for suc- cinic dehydrogenase and the cytochrome system appear to be intimately tied together and have so far defied separation. Pennit me in passing to call your attention to the position of the methylene blue system. Situated as it is at this crossroad, it is well adapted to react with the flavoproteins on the one hand and on the other to bypass the cytochrome system in reacting with oxygen. The chief pathway, then, by which energy is released in the living cell, so far as we can tell today, appears to be that shown here. The energy liberated when substrates undergo air oxidation is not liberated in one large burst, as was once thought, but is released in stepwise fashion. At least six separate steps seem to be involved. The process is not unlike that of locks in a canal. As each lock is passed in the ascent from a lower to a higher level a certain amount of energy is expended. Similarly, the total energy resulting from the oxidation of foodstuffs is released in small units or parcels, step by step. The amount of free energy released at each step is propor- tional to the difference in potential of the systems comprising the several steps. As indicated in this diagram, the steps involving the cytochromes account for more than two-thirds of the total energy released by this chain. Now also, just as each lock in a canal must be passed in sequence, so here each link in the chain appears to be indispensable. Each component of the chain seems to react readily only with that com- ponent lying immediately above or below it. This marked specificity of interaction is most extraordinary in view of the fact that these substances may react with oxidizing and reducing agents foreign to the living cell. Methylene blue has already been given as an example OXIDATIVE MECHANISMS IN ANIMAL TISSUES 23 of such a reaction occurring with the flavoproteins. Another example is the catalysis by the cytochrome system of the oxidation of certain organic substances, such as p-phenylenediamine. It should be noted, however, that such extraneous substances or their oxidized products may by their lack of specificity react in a way that is harmful to the cell mechanism. Methylene blue, for example, though not reduced by the d-amino acid oxidase system, gradually inactivates it in the presence of light (6). Mr. Kerr, working in my laboratory, has recently been investigating the mode of action by which butter yellow produces liver tumors. He has found that the oxidation catalyzed by a heart muscle preparation of extremely small quanti- ties of p-phenylenediamine (an apparent breakdown product of butter yellow) completely inactivates the succinic oxidase activity of such a preparation. Thus the marked specificity of interaction of these compounds may also serve to prevent unwanted and harmful reactions from occurring within the cell. Next arises the question whether the energy released at each step in this chain is utilized by the living organism and if so, how. At present direct evidence for the utilization of energy furnished by individual oxidative processes such as these is limited to the demon- stration of coupled phosphorylation reactions. Such demonstrations have thus far been largely confined to that portion of the chain in- volving the pyridine nucleotides. Whether the large bulk of energy release that occurs through the cytochromes is useful for phosphor- ylations or for energy-utilizing mechanisms other than phosphoryla- tion has not yet been definitely ascertained. It should be noted, however, that Korr (16) has pointed out that in the fertilized arbacia egg respiratory and functional activity are both inhibited by cyanide. Restoration of the respiratory rate by the addition of a substance such as methylene blue to replace the inactivated cyto- chrome system does not, however, restore functional activity. A point that should perhaps be mentioned in connection with this pathway is the use of the term "hydrogen transport" to describe biological oxidations. E' values of the cytochrome c system exhibit a zero slope in the neutral pH region. Preliminary experiments of the author indicate the same to be true for the cytochrome a and h systems. This indicates that only electron transfer occurs with these systems and that hydrogen is not concerned in the reaction. If the cytochrome oxidase system behaves similarly, only a fraction of the total energy released in biological oxidations involves hydro- gen transport. Thus only the electrons of the substrate can be con- - 0 "50 o Cyfochrome _c_ . X Flavin-Adenine 3 if) "e 100 Dinucleo+ide f|j|jjjj] Di phosphopyridine Nucleotide | | e o o- - U1 -5 50 E - E - ^M nil HEART BRAIN KIDNEY LIVER Figure 3. — The cytochrome c, flavin-adenine dinucleotide, and diphospho- pyridine nucleotide content of four rat tissues Values for the cytochrome c content are recalculated from those given by Stotz (24) by assuming a molecular weight of 13,000. Flavin-adenine dinucleotide values are calculated from the data of Ochoa and Rossiter (22) and Warbrug and Christian (25). Diphosphopyridine nucleotide values are calculated from tlie data of Axelrod and Elvehjem ( 1 ) with the exception of that for heart muscle, which is from a value given by von Euler et al. ( 12 ) . The values given by von Euler et al. are much lower than those reported by otlier workers. 24 OXIDATIVE MECHANISMS IN ANIMAL TISSUES 25 sidered to be transferred to oxygen in an unbroken chain by the various acceptors. Hydrogen as hydrogen ion may enter or be withdrawn from the acid-base continuum at several places in the chain. Also, the various components of the chain may be classified in two groups according to their ability to transport electrons. The cytochromes can transport only one electron for each cycle of oxidation and reduction of their prosthetic groups. The functional groups of the flavoproteins and the pyridine nucleotides are capable, however, of transporting two electrons for each cycle. Thus the possible interaction of the cytochrome system with a flavoprotein would be one where a two-step, one-electron transfer, with the formation of a semiquinone flavoprotein intermediate, might play an important biological role. The ability of the free flavins to under- go such a stepwise oxidation has been amply demonstrated by the work of Michaelis and Schwarzenbach (21). Now a consideration of the pathway just outlined might suggest that tissues contain each of these constituents in somewhat similar amounts. That such is not the case can be seen from Figure 3. The cytochrome c, flavin-adenine dinucleotide, and diphospho- pyridine nucleotide content of four tissues from the rat are here plotted in terms of millimoles per gram of wet tissue. The concentra- tion of diphosphopyridine nucleotide in all four tissues is far greater than that of the other two constituents. Cytochrome c is present in lowest concentration in all these tissues. In liver, for example, the concentration of the pyridine compound, expressed on a millimolar basis, is 340 times that of cytochrome c. Since cyto- chrome c transports only one electron per mole, this ratio becomes 680:1 when expressed in terms of equivalents. In the other tissues the ratio is lower. From such relationships one might conclude that the cytochrome system is far more efficient in the transport of elec- trons than the other systems. Such indeed may be the case. A dif- ferent explanation, however, is supported by more experimental proof, namely, that the pyridine nucleotides and the flavoprotein systems are involved in reactions other than those concerned in the main oxidative pathway. The known role of the pyridine nucleo- tides in certain anaerobic cycles is discussed elsewhere. With re- spect to the flavoproteins, recent studies have indicated that some of them are concerned in oxidative reactions which do not require the cytochrome system. Substrates such as the d-ammo acids, hypo- xanthine, xanthine, and certain aldehydes are so oxidized. The flavo- proteins concerned in these reactions are unusual in that their 26 A SYMPOSIUM ON RESPIRATORY ENZYMES reduced forms react directly with oxygen at a rapid rate. Thus they differ in this respect from the flavoproteins responsible for the oxidation of the reduced pyridine nucleotides. The reduced form of the fZ-amino acid oxidase flavoprotein also differs from all other flavoproteins in that it does not react with methylene blue or with dyestuffs of even higher oxidation-reduction potential (6). This varia- tion in the behavior of different flavoproteins containing the same prosthetic group resembles the variation in the behavior of the various iron porphyrin protein compounds. Obviously the protein partner exerts a marked influence on the behavior of the prosthetic group. Now such flavin systems as those just mentioned are probably of minor importance in furnishing the energy required by the cell; it can be shown that they are not affected by cyanide, which blocks the bulk of the oxygen consumption of the cell. In fact, these systems might be looked upon as incinerators for disposing quickly of unwanted products. Franke and Hasse (13) have termed them "rudimentary." It may be that they represent the earliest types of mechanism to emerge for the furtherance of biological oxidations and thus might be classed as primitive. They differ from cytochrome-linked systems in that hydrogen peroxide appears as a by-product of their reaction with oxygen. Thus there arises the question of the relationship of catalase and peroxidase to such systems. It is interesting to note that, of the tissues examined, the liver, among the richest in catalase, has the highest flavin and the lowest cytochrome c content. Finally, one other interpretation of the relatively low cytochrome c content of tissues must be considered. This is the possible exist- ence of pathways as yet unidentified which parallel the cytochrome system or supplement it. The existence of such unknown pathways has already been postulated as a result of certain experimental data obtained from a study of the action of inhibitors of respiration that are believed to poison cytochrome oxidase. Two of the most com- monly employed inhibitors of this type are cyanide and azide. Let us consider, therefore, two examples of experiments involving the use of these respiratory poisons and examine the validity of the conclusions that may be drawn from such experiments. The respiration of the unfertilized arbacia egg is insensitive to cyanide. Upon fertilization the egg consumes oxygen at a markedly increased rate, and the additional oxygen consumed is found to be cyanide-sensitive. These facts have been interpreted to indicate OXIDATIVE MECHANISMS IN ANIMAL TISSUES 27 that in fertilized and unfertilized eggs the pathways for the oxida- tion of substrates are different. It should be noted, however, that at best the evidence merely indicates the possible existence in un- fertilized eggs of a system alternate to the cytochrome system. Cyanide and azide do not effect the reduction of methylene blue by substrates acting through the pyridine nucleotide and flavoprotein systems. Hence these systems may be functioning in the unfertilized egg. Supporting evidence is furnished by the findings of Krahl et al. (18) that the unfertihzed egg contains flavin-adenine di- nucleotide and is also rich in diphosphopyridine nucleotide (17). Now does the cyanide insensitivity of the unfertilized egg indicate that an iron porphyrin system is not functioning in its respiratory mechanisms? Such an interpretation is indeed possible. One may, for example, postulate that the respiration is of a primitive type and passes directly through a flavoprotein to oxygen. Such a contention is supported by the fact that the presence of cytochrome a, b, or c cannot be demonstrated in the unfertilized arbacia egg. It is possible, however, to demonstrate in the egg the presence of hemin substances (4). Moreover, Krahl and his co-workers (19) have re- cently shown that the eggs contain a substance resembling cyto- chrome oxidase, in amounts equal in activity to the cytochrome oxidase of mammalian tissues. This substance was found to be cyanide-sensitive if it was functioning in an oxidation requiring the addition of cytochrome c. Are we to conclude, then, that this egg "cytochrome oxidase" plays no role in the respiration of the un- fertilized egg because such respiration is cyanide-insensitive? Is it not possible that a reaction can occur directly between "cytochrome oxidase" and flavoprotein in the arbacia egg? In view of the wide variety of properties exhibited by flavoproteins such a reaction might well occur. But it appears that such a postulate is contradicted by the evidence cited above that this egg "cytochrome oxidase" can be inhibited by cyanide. Before we decide what is the correct interpretation of these data, let us review our knowledge about the mechanism of cyanide poison- ing. Cyanide apparently inhibits respiration by reacting with cyto- chrome oxidase, since it prevents the air oxidation of the three re- duced cytochromes. At low concentrations it apparently does not combine with the cytochromes. Presumably it combines with the ferric form of cytochrome oxidase, since it is methemoglobin and not hemoglobin that reacts with cyanide. We may conclude, then, that somehow cyanide prevents the reduction of ferric cytochrome oxi- w ~9 cr a o o c 0) "o ^UJ 6 "o O a) Unfertil Arbacia fvl -XL a. z N lAI XT o a) '' a / o / l_ / -C / -2 L4_ c T3 / '>^ Ql ^2 'S / ^ >> 1 c o 1 •w o ' 1 [ 1 o / f 1 / o 1- o 4-» o 1 < / bCrt 1 / -ro c S'^ c / / / OJ 1 o C '3 o o %, / ^/ T5 U v> CP ' / O/ 'o "rt J$ \ / ^ / 7 'c o L. a, o °\ \ // 1 - in o -*- ja _c c cr o 1^ \ <^ / \9 O o U ) 1 _J \ o lO o "o a < _J \6 JEI O u El _Q c "2 u z u CO 3 £ 2 Ql D d) 0) E O > c .1 ^ u o "> ol tochror tochroi o i_ N >. ">^ tn O U U jo — 1 1 1 1 1 1 > CD O* CD O* o a + q o ' ^ 28 OXIDATIVE MECHANISMS IN ANIMAL TISSUES 29 dase by the cytochromes. Now Barron (8) has presented ample evidence that cyanide hemochromogens are able to act as reversible oxidation-reduction catalysts. If the iron in such cyanide complexes can be reduced, why is the ferric iron of the cyanide complex of cytochrome oxidase not reduced? As one answer to this question I suggested several years ago (3) that the oxidation-reduction potential of the cytochrome oxidase system is lowered in the presence of cyanide below that of cyto- chrome a or c. The reduction of cytochrome oxidase by these com- pounds could then not occur and the respiratory chain as a whole might thus be blocked. The experiments of Barron (8), which showed that the cyanide hemochromogen systems were the most negative of the hemochromogen systems investigated, were cited in support of this idea. Since then Clark and his associates (10) have presented their thorough analysis of such hemochromogen systems. As a result of this study it may be said that the potentials of iron porphyrin systems vary according to the type of nitrogen compound associated with them. Also, if the aflBnity of the nitrogenous compound is great- est for the ferric form, the potential of the system will decrease progressively as the concentration of the coordinating compound is increased. The reverse holds true if the feiTous form displays the greatest affinity. Now assuming that the cytochrome oxidase system behaves in a similar manner, we may picture the effect of cyanide on the potential of the cytochrome oxidase system as shown in Figure 4. Here the cytochrome oxidase system is arbitrarily assigned a potential of 0.5 volts at pH 7.0. The potential of this system is plotted against the log of the concentration of the coordinating com- pound, as is done by Clark et al. (10) for the hemochromogen systems. It is assumed that only the ferric form of cytochrome oxidase reacts with cyanide and that since the first noticeable effects of cyanide poisoning result at concentrations of 10~^ to lO*' molar, it is within this range that a potential shift will first occur.* The potential is assumed to change according to a 0.12 slope, since this value has been found to hold for certain hemochromogen sys- tems (10). It will be seen that, according to such a scheme, when the concentration of cyanide reaches 10'^ molar, the potential of the * As shown by Clark et al. (10) for hemochromogen systems, this point de- pends also upon tlie concentration of the hemochromogen system. The con- centration of the cytochrome oxidase system is of course not known. It is probably of the order of magnitude of the systems studied by Clark et al. (10) if tlie concentration of cytochrome c in the tissues can be used as an index. 30 A SYMPOSIUM ON RESPIRATORY ENZYMES cytochrome oxidase will be depressed below that of the cytochrome a and c systems. In such a tissue as muscle, then, we might expect 100 per cent inhibition of the cytochrome-mediated respiration. It is at such a concentration that cyanide exerts its maximum effect upon respiration. Let us now return to the case of the unfertilized arbacia egg, where the cytochromes appear to be absent. Here cytochrome oxidase may be considered to react directly with a flavoprotein. The addition of cyanide may be pictured as depressing the potential of the cytochrome oxidase in the same manner. In this case, how- ever, the potential of the cytochrome oxidase still remains above that of the flavoprotein with which it normally reacts. Oxidations can therefore still proceed, and no cyanide inhibition is observed. In fact, a stimulation of respiration by cyanide such as was observed by Marsh and Goddard (20) in the fully mature leaf might be encountered if such a set of conditions exists. Barron (7) has shown that the rate of reaction of oxygen with reduced dyestuffs increases as the oxidation-reduction potential of the system decreases. A similar change may occur in the case of the cytochrome oxidase system as its potential is lowered by cyanide. Fertilization of the arbacia egg may then be looked upon as the gearing of the cytochrome oxidase system to another system of higher potential than the flavoprotein, which, though serving to in- crease the respiratory rate, also causes the system to become cyanide-sensitive. The fact that arbacia sperms are rich in cyto- chromes a, b, and c is of interest in this connection. For a further dis- cussion of this aspect see Krahl et al. (19). One other example of an inhibitor study which may be employed is furnished by the work of Stannard (23). This investigator has re- cently presented data to show that resting and stimulated frog muscle respond differently to cyanide and azide. Whereas cyanide inhibits the oxygen consumption of both resting and stimulated muscle, azide inhibits only the oxygen consumption of stimulated muscle. The results have been interpreted to indicate that the path- ways of oxidation are different in resting and in stimulated muscle. As Stannard points out, the cyanide sensitivity of the resting respira- tion seems to preclude the possibility that it represents an independ- ent functioning of a flavoprotein system. Now azide is believed to inhibit respiration also by reacting with cytochrome oxidase. Quantitatively, however, its action is different from that of cyanide. Inhibition of respiration by azide appears to OXIDATIVE MECHANISMS IN ANIMAL TISSUES 31 begin at concentrations higher than those required for cyanide poisoning. As the azide concentration is increased above this initial value, inhibition progresses as rapidly as with cyanide until a concentration of about 10"^ molar is reached. Further increase in azide concentration produces no further effect, and inhibition of respiration remains incomplete, never reaching the maximum value obtained with cyanide. How then can this diflPerence in the be- havior of cyanide and azide be explained if they both act on cyto- chrome oxidase? Is it necessary to postulate that azide and cyanide inactivate separate systems? An explanation of the difference in the behavior of cyanide and azide may be given in terms of their diflFerent effects upon the poten- tial of the cytochrome oxidase system. Azide, like cyanide, is as- sumed to combine with the ferric form of cytochrome oxidase. Un- like cyanide, it is also assumed to combine with the ferrous form. In Figure 4 the effect of azide upon the potential of the cytochrome oxidase system is plotted on the basis of these assumptions. Since azide inhibition first manifests itself at concentrations higher than those for cyanide, its affinity for the ferric form of the oxidase is assumed to be less than that of cyanide. Depression of the potential of the oxidase system is therefore portrayed as starting when the concentration of azide reaches a value between 10 * and 10 "' molar. As the concentration of azide is increased above this value, the potential is assumed to be lowered along the same slope as for the cyanide system. Now since inhibition with azide reaches a maximum at a concentration of 10"^ molar, it is assumed that at this concentra- tion the azide begins to combine with the ferrous form of cyto- chrome oxidase. According to Clark et al. (10), the effect of such a combination on tlie potential of the system will be to alter the slope to a 0.0 value, and thus no further change in potential occurs as more azide is added. On the basis of this assumption it can be seen that, as depicted in Figure 4 the potential of the cytochrome oxidase system in the presence of azide can be depressed only to a level corresponding to that for the cytochrome a system. At this level the cytochrome oxidase may still function to oxidize the cytochromes, but its efficiency will be at least 50 per cent impaired. It is thus possible to conceive of the respiration of resting muscle as being unimpaired by azide, since even at a lower level of efficiency the cytochrome oxidase may still be capable of supplying the oxygen needs of the resting muscle. If, however, increased demands for oxygen are made upon the muscle by stimulation, the impaired 32 A SYMPOSIUM ON RESPIRATORY ENZYMES eflBciency of the cytochrome oxidase system becomes a limiting factor. By this time I can hear you muttering, "Sheer speculation," and I agree with you. I have, however, labored my point concerning the action of such inhibitors because I believe that until we know more about their mode of action any hypotheses of the existence of other oxidative pathways must also be labeled speculation. I do not wish to imply that other unknown pathways do not exist. I heartily agree that they may. I only beg that, in these days of ever-widening use of such inhibitors, more fundamental investigations be made into their mode of action before we becloud the issue with false inter- pretations. REFERENCES 1. AxELROD, A. E., and Elvehjem, C. A., J. Biol. Chem., 131, 77 (1939). 2. Ball, E. G., Symposia on Quantitative Biology, Cold Spring Harbor, 7, 100 (1939). 3. Ball, E. G., Discussion published as part of reference 8. 4. Ball, E. C, and Meyerhof, B., J. Biol. Chem., 134, 483 (1940). 5. Ball, E. C, and Ramsdell, P. A., J. Biol. Chem., 131, 767 (1939). 6. Ball, E. C, and Ramsdell, P. A., unpublished experiments. 7. Barron, E. S. G., J. Biol. Chem., 97, 287 (1932). 8. Barron, E. S. G., Symposia on Quantitative Biology, Cold Spring Harbor, 7, 154 (1939). 9. Barron, E. S. G., and Harrop, G. A., Jr., J. Biol. Chem., 79, 65 (1928). 10. Clark, W. M., Taylor, J. F., Davies, T. H., and Vestling, C. S., J. Biol. Chem., 135, 543 (1940). 11. Dixon, M., and Zervas, L. G., Biochem. J., 34, 371 (1940). 12. von Euler, H., Schlenk, F., Heiwinkel, H., and Hogberg, B., Z. physiol. Chem., 256, 208 (1938). 13. Franke, W., and Hasse, K., Z. physiol. Chem., 249, 231 (1937). 14. Haas, E., Horecker, B. L., and Hogness, T. R., J. Biol. Chem., 136, 747 (1940). 15. Jacobs, H. R., Proc. Soc. Exp. Biol. Med., 37, 407 (1937). 16. KoRR, I. M., Symposia on Quantitative Biology, Cold Spring Harbor, 7, 120 (1939). 17. Krahl, M. E., Verbal communication at a seminar on August 26, 1941, at Marine Biological Laboratory, Woods Hole, Mass. 18. Krahl, M. E., Keltch, A. K., and Clowes, G. H. A., Proc. Soc. Exp. Biol. Med., 45, 719 (1940). 19. Krahl, M. E., Keltch, A. K., Neubeck, C. E., and Clowes, G. H. A., J. Gen. Physiol., 24, 597 (1941). 20. Marsh, P. B., and Goddard, D. R., Am. J. Botany, 26, 724 (1939). 21. Michaelis, L., and Schwarzenbach, G. J., J. Biol. Chem., 123, 527 (1938). 22. Ochoa, S., and Rossiter, R. J., Biochem. J., 33, 2008 (1939). 23. Stannard, J. N., Symposia on Quantitative Biology, Cold Spring Harbor, 7, 394 (1939). 24. Stotz, E., J. Biol. Chem., 131, 555 (1939). 25. Warburg, O., and Christian, W., Biochem. Z., 298, 150 (1938). Ball and Stotz From flavoprotein to cytochrome Meyerhof and Group At the Fountainhead XoKD, Neuberg, axu Kletzien' Was it the weather or the argument? BETWEEN THE SCIENTIFIC SESSIONS Discussion on Hydrogen Transport VAN R. POTTER University of Wisconsin, Chairman Dr. Potter: This afternoon's discussion has been organized with the idea of studying some of the problems which may be considered as open questions at this time. The first speaker will be Dr. Elliott. THE POSSIBLE ROLE OF INTERMEDIARY METABOLITES AS HYDROGEN CARRIERS K. A. C. ELLIOTT Institute of the Pennsylvania Hospital This subject has not been reviewed at length during this sym- posium, and it is impossible to cover it fully in a brief discussion. Some famiharity with the subject will therefore be assumed, and only certain outstanding problems will be discussed. (For details and bibliographies see references 1-3.) The most important theory of intermediary metabolites as hydro- gen carriers is that of Szent-Gyorgyi. It may be represented by the following highly simplified diagram, in which heavy arrows indicate transfers of hydrogen atoms (or electrons) from one substance to the next. OONATORS OXALACETATE FUMARATE REDUCED CYTOCHROME I .2H^ il Jil^ It .2H^ n -^^OXYGEN ^^ ^^, ., MALATE SUCCINATE OXIDIZED CYTOCHROME OXIDIZED DONATORS According to this theory, hydrogen from tissue donators reduces oxalacetate to malate; the malate is reoxidized to oxalacetate and the hydrogen is transferred to fumarate, reducing it to succinate; and the succinate is then reoxidized to fumarate by the cytochrome- cytochrome oxidase system. Malate-oxalacetate and succinate- fumarate thus perform functions similar to that of reduced cyto- chrome-oxidized cytochrome. Known dehydrogenases and cozymase are concerned in the catalysis of the various steps; also, a flavo- protein was believed by the Szent-Gyorgyi school to mediate in the reduction of fumarate by malate. Considerable evidence that this mechanism can function in pigeon breast muscle brei has been adduced by the Szent-Gyorgyi 33 34 A SYMPOSIUM ON RESPIRATORY ENZYMES school, by Stare and Baumann, and, in connection with work on the citric acid cycle, by Krebs and coworkers, and also by various other workers. Krebs has also shown that at least the succinate-fumarate mediation can occur in B. coli (Escherichia coli). If we assume that the evidence for the Szent-Gyorgyi mechanism actually proves that it does operate in pigeon breast muscle (see, however, addendum below), the following questions arise: 1. In what tissues may the mechanism he important? Tissues other than pigeon muscle have not been studied exhaustively from this point of view. One of the main tests for the system consists in finding increased or better maintained respiration when small amounts of fumarate or malate, which are interconvertible by tissue fumarase, are added to the tissue brei. Our experience with this test and that of other workers suggests that the system may be im- portant for the respiration of liver, testis, and possibly kidney, but is not very active in brain or skeletal muscle. However, individual reactions of the system, namely, succinate oxidation and oxalacetate reduction, occur rapidly in all these tissues, and Banga, Cori, and coworkers have shown that the presence of fumarate is necessary for the oxidation of pyruvate by kidney and brain dispersions. Banga showed that it was not easy to remove all the four-carbon acids from tissue. It is therefore possible that when added four- carbon dicarboxylic acids have little effect on the respiration of tissues, these substances may already be present in the tissues in such amounts that their concentration is not a limiting factor of the respiration rate. The volume of respiration passing through the system would be limited by the activity of the relevant enzymes. Dr. Greig and I found that there was sufficient cytochrome-cytochrome oxidase and succinic dehydrogenase activity in many normal tissues to account for all the respiration through succinate-fumarate, but the succinic dehydrogenase activity was quite low in chick embryo, rat thymus, spleen, pancreas, and some tumors. Breusch found that the rates of oxalacetate reduction were high in muscle, liver, and kidney and moderate in brain and pancreas, but negligible in spleen, lung, placenta, peripheral nerves, and certain rat tumors; the rate in embryo was found to be low by Blaszo. 2. Does the four-carbon dicarboxylic acid system always operate in the same waif? Szent-Gyorgyi himself pointed out that not all substances should be expected to utilize the whole system. Lactate, for instance, might be oxidized through fumarate-succinate but DISCUSSION ON HYDROGEN TRANSPORT 35 probably not through oxalacetate-malate, since the redox potential of the latter system is close to that of pyruvate— lactate.* According to the complete theory, fumarate and malate should behave alike, an equilibrium mixture of the two being rapidly produced by the action of fumarase when either is added. But Dr. Libet and I have found that added malate and fumarate (and citrate) have different effects on the repiration of brain suspensions. Greig and Munro found that fumarate, but not malate, caused a lowered respiratory quotient with ox retina and chick embryo. 3. What metabolites are oxidized through the system? The Szent- Gyorgyi school indicated that carbohydrate derivatives— triose- phosphate, alpha-glycerophosphate, and pyruvate— were oxidized through the four-carbon dicarboxylic acid system by tissue sus- pensions. But Greville reported that not more than 70 per cent of the respiration brought about by fumarate catalysis in muscle was due to carbohydrate oxidation. With suspensions of liver from fasted rats, my wife and I found low respiratory quotient values for the extra respiration caused by adding malate. Leloir and Munoz found that added four-carbon dicarboxylic acids increased the rate of butyric acid oxidation by liver suspension. Dewan and Green, with isolated enzyme preparations, showed the oxidation of beta- hydroxybutyrate by fumarate. Annau, and my wife and I, noted * Dr. Eric Ball, Harvard University: My chief objection to the Szent- Gyorgyi theory is that the inclusion of the malate-oxalacetate system in the chain of reactions as it is written seems to be pointless. Attention has been called to tliis fact previously ( 4-6 ) . If we break the scheme into the separate reactions, this becomes evident. Assuming the substrate to be oxidized by means of diphosphopyridine nucleotide (Py(P04)2), we may write the first reaction as fol- lows: ( 1 ) Substrate + Py(P04 ) 2 -> H^PyCPO^ ) 2 + Oxidized Substrate If the pyridine nucleotide is to act as a cyclic catalyst for this reaction, it must be oxidized. According to the Szent-Gyorgyi scheme, tliis reoxidation is brought about by oxalacetate. The reaction may be written: ( 2 ) Oxalacetate -|- H2Py(P04 ) 2 -> Py(P04 ) 2 -f Malate Now if the oxalacetate in turn is to function as a catalyst, it must be regen- erated. This requires that malate be oxidized. The oxidation of malate in the body, however, is known to proceed only through the diphosphopyridine nucleo- tide: ( 3 ) M alate -f Py (PO4 ) 2 ^ H2Py(P04 ) 2 + Oxalacetate This equation is, however, the reverse of equation 2. Thus what is produced is reduced pyridine nucleotide, and we are right where we started when we wrote equation 1. The introduction of the malate-oxalacetate system into diis cycle merely leads us into a blind alley. 36 A SYMPOSIUM ON RESPIRATORY ENZYMES that added four-carbon substances removed, or prevented the for- mation of, acetoacetic acid in liver and kidney brei. Krebs, working vi^ith B. coli [Escherichia coli] indicated that fumarate-succinate mediated the oxidation of glucose, malate, lactate, acetate, glycerol, glyceraldehyde, butyrate, pyruvate, acetoacetate, Z( + )-glutamate and molecular hydrogen. Apparently the oxidation of many types of metabolite can be mediated by four-carbon dicarboxylic acids. Szent-Gyorgyi considered that the four-carbon dicarboxylic acids were quite analogous to coenzymes in their catalytic role. But it must be remembered that they behave also as combustible sub- stances, particularly with kidney cortex slices and also to some ex- tent with brei of various tissues, when they are added in excess. The four-carbon dicarboxyhc acids are not the only substances that can act both as catalysts and as substrates for respiration. Carrier possibihties have been shown for pyruvate-lactate, as well as other alpha-keto— alpha-hydroxy acids, adrenochrome, and transamination reactions. It has been suggested that in B. coli formate-bicarbonate plays a similar role, and in plants catechol derivatives and dihydroxy- maleic acid may be important carriers. Perhaps many other sub- stances may act in this way. In fact, carrier functions could be postu- lated for all reversible oxidation-reduction systems, and we should perhaps think of all oxidizable metabolites as capable of acting as both substrates and carriers in a dynamic oxidation-reduction con- tinuum. Addendum.'* —The Szent-Gyorgyi theoiy was first advanced at a time when the succinic dehydrogenase system was the only one definitely known to reduce cytochrome. Mediation by succinate- fumarate of hydrogen transport between the majority of metabolites and the cytochrome system thus naturally suggested itself. However, the discovery of flavoproteins, which mediate oxidation of the re- duced coenzymes, suggests that a flavoprotein catalyst may bring about more direct oxidation of those substrates that are oxidized through the action of coenzyme-deteiTnined dehydrogenases. Dr. Hogness has described cytochrome c reductase, a flavoprotein from yeast that causes cytochrome c reduction by dihydrocoenzyme II. It seems likely that similar catalysts for the oxidation of both dihydro- coenzymes I and II may occur in animal tissues. In that case there would seem to be no necessity for mediation by fumarate-succinate. (Another flavoprotein in yeast, fumarate reductase, has been de- scribed by Fischer and coworkers. This enzyme causes the reduction " In the liglit of private discussions during the symposium. Dr. EUiott has written an addendum to his remarks. — Ed, DISCUSSION ON HYDROGEN TRANSPORT 37 of fumarate to succinate by certain leuco dyes; it is distinct from succinic dehydrogenase but could perhaps replace the latter enzyme in the Szent-Gyorgyi scheme). While it has been proved that the four-carbon dicarboxylic acids and other metabolites may act as carriers, the previous paragraph indicates that we cannot conclude that fumarate-succinate actually does so. Since malonate inhibits succinate oxidation, the observed inhibition of respiration by malonate has been taken as evidence of the carrier function of succinate-fumarate. But the inhibition of respiration by malonate and the promoting effect of added fumarate may be explained as inhibition and promotion of oxidative metabol- ism through Krebs' cycle. Dr. Potter has emphasized the above points (6), and he points out that carrier functions for four-carbon dicarboxylic acids and other substances have been assumed on the basis of the following four criteria: 1. The compound is a natural constituent of tissues. 2. It can be reduced by tissues at rates compatible with the actual rate of oxidation of the substrate whose oxidation it is presumed to cata- lyze. 3. The reduced compound can be oxidized by the tissue prepa- ration at an adequate rate. 4. The compound is able to stimulate catalytically the rate of hydrogen transport in the system under investigation. But it appears that a fifth requirement is necessary to prove carrier function, namely, the compound must be directly re- duced by one system and directly oxidized by a second system which is not identical with the first. As Dr. Ball and Dr. Potter have pointed out, oxalacetate is reduced by dihydrocoenzyme I and malate is oxidized by coenzyme I, yielding tlie dihydrocoenzyme. That is to say, hydrogen from the donator metabolite is passed to the co- enzyme, producing dihydrocoenzyme, but subsequent reduction of oxalacetate to malate and reoxidation of the latter merely results in producing the dihydrocoenzyme again. Thus no effective trans- port of hydrogen has occurred and, unless another biological mecha- nism for the oxidation of malate is discovered, it seems unnecessary to postulate a carrier function for malate-oxalacetate. It is conceiv- able, however, that the structural relations of enzymes in tissue may render the passage of hydrogen from the donators to fumarate or flavoprotein easier via coenzyme-oxalacetate-malate-coenzyme than directly via coenzyme in one step. It seems likely that in tissues there is a dynamic equilibrium be- tween the oxidized and reduced forms of numerous metabolites and that flavoprotein-cytochrome-cytochrome oxidase mechanisms continually abstract hydrogen (or electrons) from the system, while 38 A SYMPOSIUM ON RESPIRATORY ENZYMES few, if any, of the metabolites can be singled out as specially con- cerned in the hydrogen transport. (This view of the mechanism was suggested to me by Dr. Potter.) REFERENCES 1. Szent-Gyorgyi, a., Studies in Biological Oxidation (Leipzig, 1937). 2. Stare, F. J., and Baumann, C. A., Cold Spring Harbor Symposia on Quan- titative Biology, 7, 277 ( 1939). 3. Elliott, K. A. C, Physiol. Rev., 21, 267 ( 1941). 4. Ball, E. G., Cold Spring Harbor Symposia on Quantitative Biology, 7, 100 (1939). 5. Martius, C, Ergebnisse Enzymforschung, 8, 247 (1939). 6. Potter, Van R., Medicine, W, 441 (1940). Dr. Potter: The second question to be considered is the role of the carriers in dismutations and coupled oxidoreductions.* Since it is agreed that the various metaboHtes may take part in coupled oxido- reductions, it becomes of interest to determine how these reactions may be brought about, and what features these "fermentation" re- actions have in common with the oxidative mechanisms. Dr. Ball and Dr. Lipmann will open the discussion. THE ROLE OF THE CARRIERS IN DISMUTATIONS AND COUPLED OXIDOREDUCTIONS With Special Reference to the Flavoproteins Eric Ball, Harvard University: The chief carriers that play a role in dismutations or coupled oxidoreductions are the pyridine nucleotides and flavoproteins. Since the participation of the pyridine nucleotides in such reactions is more common, examples of dismutations and of coupled oxido- reductions involving diphosphopyridine nucleotide may be given first. The classical example of a dismutation is the so-called Cannizzaro reaction, which may be represented by the following equations: (1) R-CHO + 2e + 2H^->R-CH20H (2) R-CHO-f H2O — 2e-2H*^R-COOH (1) + (2) 2RCHO + H20->R-CH20H-f R-COOH " This term was not included in the original statement of the question. Dr. Barron rose to point out that the term "dismutation" has a very narrow applica- tion and that "coupled oxidoreduction" is the more general expression. DISCUSSION ON HYDROGEN TRANSPORT 39 One molecule of aldehyde undergoes an oxidation to the corre- sponding acid at the expense of another molecule of aldehyde which is reduced to alcohol. An enzyme which catalyzes this type of re- action, called aldehyde mutase, has been found in hver by Dixon and Lutwak-Mann (1). It requires as a coenzyme diphosphopyridine nucleotide, which thus appears to function in the role of carrier of electrons and hydrogen ions from one aldehyde molecule to another. Another example that might be given is the dismutation of triose- phosphate. The role of a carrier in a so-called coupled oxidoreduction diflFers from that in a dismutation reaction only in that electrons and hydrogen ions are transferred between molecules of two different substances. A well-known example of such a reaction (alcohoHc fermentation) may be written as follows if we omit the coupled phosphorylation steps that accompany it: ( 3 ) CHs • CHO -H 2e + 2H" -> CH3CH2OH (4) CHO COOH I I CHOH -1- H2O - 2e - 2H^ -> CHOH HaC-OPOaH, H2COPO3H2 The carrier of the electrons and hydrogen ions between these two aldehyde molecules is also diphosphopyridine nucleotide. The reac- tion is not classed as a dismutation simply because the aldehyde mole- cules involved are not identical. One could give other examples of coupled oxidoreductions in which the reacting molecules are more dissimilar and involve diphosphopyridine nucleotide as a carrier. Such reactions constitute the main type of energy exchange in anaerobic processes. Under aerobic conditions that half of the re- action which involves the loss of electrons and hydrogen ions to the pyridine nucleotide may still occur. The reduced pyridine nucleo- tide, however, under aerobic conditions, loses its electrons and hydrogen ions to a flavoprotein rather than to another substrate molecule. This type of reaction, then, might be hsted as still another class of coupled oxidoreductions. As compared with the pyridine nucleotides, the flavoproteins par- ticipate in only a few direct reactions with substrate molecules. Examples of dismutations or coupled oxidoreduction reactions such as have been given for the pyridine nucleotides are less plentiful for the flavoproteins. One example of a dismutation reaction which may be classed as involving a flavoprotein is that for the substrate 40 A SYMPOSIUM ON RESPIRATORY ENZYMES xanthine. Green (2) has shown that, in the presence of xanthine oxidase, xanthine undergoes the following reactions: (5) Xanthine + 2e + 2H* —> Hypoxanthine ( 6 ) Xanthine — 2e — 2H^ -» Uric Acid 2 Xanthine — > Hypoxanthine + Uric Acid The carrier of the electrons and hydrogen ions between two mole- cules of xanthine may in this case be due to the flavin portion of xanthine oxidase. Examples of oxidoreduction reactions between two dijfferent substrates involving flavoproteins are not known to me. It may be that no such reactions exist, since each flavoprotein appears to be substrate specific. An indirect reaction of a flavoprotein as a carrier between two different substrates is, however, conceivable. This would be the case with two different substrates whose direct reaction occurred only with each of the pyridine nucleotides, as shown by the following equations: Oxidized Substrate A + HjPyCPOi) 2 -^ Py(P04)2 + Substrate A H2Py(P04)2 - 2e - 2H^ -^ Py(P04)2 PyCPOJs + 2e + 2H^ -^ H2Py(P04)3 Substrate B + Py(P04)3 -^ H2Py(P04)3 + Oxidized Substrate B Here in order for oxidized substrate A to react with substrate B, the reduced triphosphopyridine nucleotide must react with diphospho- pyridine nucleotide. Whether such a direct reaction is possible is, I believe, unknown. It is possible, however, that a flavoprotein might act as a carrier of electrons and hydrogen ions between the two pyridine nucleotides. Such a reaction could thus be classed as a coupled oxidoreduction involving a flavoprotein. Fritz Lipmann, Massachusetts General Hospital: Flavoproteins are at present known to react in two ways: (1) with oxygen directly, i.e., as mediators between substrates and oxygen; (2) as mediators between pyridine enzymes and other catalysts or oxygen. The second function has been thoroughly studied in reactions representing anaerobic parts of essentially aerobic reaction chains: the bridging between pyridine nucleotide and methylene blue (Haas, Straub, Green) and between pyridine nucleotide and cytochrome (Haas, Horecker, and Hogness). Although it would be suspected that a flavin mediator is needed in a great variety of purely an- aerobic reactions, especially in those between a pyridine and a non- DISCUSSION ON HYDROGEN TRANSPORT 41 pyridine enzyme, not much is definitely known about such action of flavoproteins. One example, however, is the dismutation of pyruvic acid (Lipmann), in which a flavin component was shown to par- ticipate in a coupled oxidoreduction. Here the flavin mediates between two enzyme systems of different types, i.e., between lactic and pyruvic dehydrogenase: ( 1 ) pyruvate — > lactate 2HT (flavin) pyruvate phosphate — > acetylphosphate + CO2 The presence of a flavoprotein in yeast, which catalyzes the reduc- tion of fumarate (Fischer), suggests the mediator function of such a flavoprotein between pyridine enzymes and fumarate. For example, in the dismutation of fumaric acid (Green) electron transfer pre- sumably occurs between malic dehydrogenase, a pyridine enzyme, and succinic dehydrogenase, a non-pyridine enzyme: ( 2 ) fumarate — » succinate 2H T (flavin) fumarate «^ malate -^ oxalacetate It is probably accidental that in both examples for anaerobic flavin mediation, the coupled oxidoreduction is a so-called dismu- tation. This might serve to show that in most enzymatic dismutations the underlying reaction is, in fact, an oxidoreduction between two enzyme systems catalyzing two fundamentally different reactions. But in dismutation the metabolic substrates for both enzyme sys- tems derive from the same compound. In the cases discussed, the oxidant in the oxidation-reduction reaction is the compound proper, pyruvate or fumarate, and the reductant is a transformation product of the added compound. Oxidant and reductant belong, respectively, to two different oxidation-reduction systems with widely different oxidation-reduction potentials. In dismutation (reaction 1) the lac- tate-pyruvate system of EJ —0.18 volts reacts with the pyruvate... HX —acetate +CO2 system of Eo' below —0.4 volts. In the sec- ond reaction the succinate-fumarate system of Eq' 0.0 volts reacts with the malate-oxalacetate system of Eo' —0.17 volts. These large energy differences between the reacting oxidation-reduction sys- tems explain why dismutation occurs. REFERENCES 1. Dixon, Malcolm, and Lutwak-Mann, Cecilia, Biochem. J., 31, 1347 (1937). 2. Green, Davto Ezra, Biochem. J., 28, 1559 (1934). 42 A SYMPOSIUM ON RESPIRATORY ENZYMES THE PHYSICO-CHEMICAL MECHANISM OF HYDROGEN TRANSPORT Kurt Stern, Yale University: Dr. Ball, on Thursday morning, spoke of a gap between the cyto- chrome oxidase-cytochrome system and the substrate-dehydrogenase systems. It appears that this gap may now be considered filled by several flavoproteins, e.g., Hogness and Haas's cytochrome reductase, Euler and Green's diaphorase (coenzyme factor), or Szent-Gyorgyi's dicarboxyhc acid system, etc. It is incorrect to label the entire process of cell respiration one of hydrogen transport, since the first stages, from oxygen through the four iron atoms of the respiratory ferment and the three cyto- chromes, are concerned exclusively with electron transfers. This brings up a difficulty in the formulation of the elementary steps that connect the iron systems with the dehydrogenase systems. The oxido-reductive changes taking place in the former are one-electron transfers; the dehydrogenation reactions, on the other hand, are formulated as bivalent processes, involving the loss and uptake of two hydrogen atoms per molecule. The most satisfactory way of re- solving this dilemma without invoking the existence of highly problematical trimolecular reactions is to assume, with Michaelis, that the apparently bivalent dehydrogenations are actually two-step processes, involving the transfer of one hydrogen atom or its equiva- lent at a time, with the intermediate formation of semiquinoid radi- cals. It seems to be no mere coincidence that the macromolecular prepa- rations, called cytochrome oxidase by Keilin, contain a number of components: the oxidase, cytochromes a and b, succinic dehydro- genase, etc., which are all concerned with what Oppenheimer calls the "terminal oxidation" of metabolites. It is reasonable to assume that these particles represent functional units which contain these catalysts in a spatial arrangement which facilitates the progress of this important phase of cell respiration in a constant pattern and at a constant and high rate. Dr. Potter: The dilemma of which Dr. Stern has spoken, that is, the mech- anism for getting a bivalent dehydrogenation system to react with a one-electron system, may possibly be resolved by the formation of a complex made up of the proper components of the hydrogen DISCUSSION ON HYDROGEN TRANSPORT 43 (electron) transport system. I believe Mr. Haas has some experi- mental data which have considerable bearing on this question. Erwin Haas, University of Chicago: For a long time the mechanism of respiration was studied only by considering intermediary metabolites. A more direct approach to that problem is now possible, since some of the respiratory enzymes have become available in isolated form and since their functional groups are known. cytochrome c cytochrome reductase HEMIN PROTEIN^ JiL ALLO'XAZINE MONO- PROTEIN NUCLEOTIDE ^ ' -9 K = 10 TRIPHOSPHOPYRIDINE PROTEIN NUCLEOTIDE ' -5 K =10 H GLUCOSE -6 -PHOSPHATE FIGURE 1 K = 10 ' K 10 Figure 1 illustrates a part of the respiratory system which brings about the reduction of cytochrome c by glucose-6-phosphate. The components of the system are arranged in the order in which they react; the details concerning their chemical structure are omitted. Hydrogen or electrons from the glucose are passed on to the pyridine nucleotide, thence to the alloxazine mononucleotide, and then to the iron atom in cytochrome c. In each step a specific protein must be present which, together with the prosthetic group, forms the active 44 A SYMPOSIUM ON RESPIRATORY ENZYMES enzyme. By selecting proper concentrations any reaction in this scheme can be made the hmiting factor. Thus by measuring the rate with which cytochrome c is reduced, it is possible to determine rates of reaction and to demonstrate formation of enzyme complexes for any of the reactions involved. Negelein and Haas have shown by ultraviolet spectroscopy the formation of a complex between protein I, triphosphopyridine nucleotide, and glucose-6-phosphate. With the method indicated above the dissociation of the cytochrome reductase into alloxazine mononucleotide and protein II can be demonstrated if to a small but constant amount of protein II increasing amounts of alloxazine mononucleotide are added. The dissociation constant of cytochrome reductase is small (K = 1 X 10"^ M), and under physiological conditions this enzyme will therefore be present as the practically undissociated complex. The value of the spectrophotometric method for the study of enzyme reactions is well demonstrated here, since 10"'' mg. of flavin are suflBcient for accurate determinations. Not only the relation between the prosthetic group and protein of one enzyme may be studied with this system, but also the interaction of two different enzymes. For example, we can measure the rate of reaction when increasing amounts of dihydrotriphosphopyridine nucleotide are added to a constant amount of cytochrome reductase. The velocity of the reduc- tion of the enzyme is given by (1) d(CR)_ kr(T) (CR) dt ~ {T)+Kd in which ( T) and ( CR) are the concentrations of triphosphopyridine nucleotide and cytochrome reductase, respectively, K is the first- order velocity constant, and Kd the dissociation constant of the pyridine-alloxazine complex. Rate constants and dissociation con- stants have been determined at different temperatures, and by apply- ing the Arrhenius equation the energy of activation is found to be about 10 kg. cal., and the heat of dissociation about 2 kg. cal., for the reaction in which cytochrome reductase is reduced by dihydro- triphosphopyridine nucleotide. Similar experiments have been made for the purpose of studying the oxidation of cytochrome reductase by cytochrome, and again the formation of a complex between the two reaction partners could be established. Furthermore, from the results of the kinetic deter- minations it can be concluded that in the course of the oxidation of alloxazine free radicals are involved. This could almost be antici- DISCUSSION ON HYDROGEN TRANSPORT 45 pated, for the alloxazine undergoes a valence change of two, whereas cytochrome undergoes a valence change of one, as Dr. Stern has just pointed out. To summarize, the facts concerning this part of the respiratory system are, then, as follows: The protein of the cytochrome reductase is bound simultaneously to alloxazine mononucleotide, to triphos- phopyridine nucleotide, and to cytochrome. Approximate values for the different dissociation constants are given in Figure 1. The energy of activation is low, about 10 kg. cal. One may venture to say that the binding forces are of the Van der Waal type rather than ordinary bond forces, which would involve much higher bond energies. The task of the protein may be (a) to estabhsh the proper geo- metrical configuration between the different prosthetic groups; and (b) to aid in the formation and stabihzation of free radicals. The first point may be ofiFered as a working hypothesis to explain the specificity of the proteins, and the second point may explain the tremendous activity of this catalyst. These experiments were done in Professor Hogness' laboratory, with Drs. B. L. Horecker and C. J. Harrer. Mr. Haas:* The formation of complexes in the course of these enzymatic reactions can be demonstrated in two independent ways. 1. Kinetic Measurements.— In agreement with equation 1, the velocity of the reaction is proportional not to (T) or to (CR), as one would expect from an ordinary bimolecular reaction, but to the amount of enzyme present in the form of the complex. 2. Spectroscopic Measurements.— Alloxazine mononucleotide has its maximum absorption at wave length 445 mix; the addition of the protein of the old yellow enzyme causes the maximum absorp- tion to migrate to wave length 465 m[i, and the further addition of triphosphopyridine nucleotide to wave length 475 mpi. Thus the free flavin, the flavin-protein complex, and the flavin-protein-pyridine complex can easily be distinguished by their color. Dr. Potter: Thus far the emphasis has been on the cytochrome system as one of the hnks in the hydrogen transport system. Yet there is an accu- mulating body of evidence which indicates that an alternate path * This statement of Mr. Haas' was made in reply to a question from tfie floor regarding tlie proof of formation of an enzyme-substrate complex.— Ed. 46 A SYMPOSIUM ON RESPIRATORY ENZYMES of hydrogen transport may exist. Dr. Stotz has agreed to open the discussion of this interesting possibiHty. POSSIBILITY OF A BY-PASS AROUND THE CYTOCHROME SYSTEM ELMER STOTZ Harvard University The possibihty of a 'Tsy-pass" around the cytochrome system in certain phases of tissue respiration has eUcited considerable discus- sion of late. The discovery of the flavoproteins, and their function in isolated systems as autoxidizable substances, had raised a doubt as to the exclusive role of the iron system in "oxygen activation." However, this threat to the autonomy of the iron compounds was dispelled by the correction of certain technical points and by the work of Theorell and of Barron. The most recent threats have come from a study of the effects of cyanide and azide ( as oxidase inhibit- ors) on the respiration of "resting" and "stimulated" tissues. The work of Stannard and of Korr on this topic is reviewed in my paper on page 169. Their results may be summarized briefly by the state- ment that the respiration of "resting" tissue is insensitive to azide, whereas that of the stimulated tissue becomes azide-sensitive. Tissues in either phase may, however, be cyanide-sensitive. The 'Iby-pass" theory therefore implies that in the resting tissue an oxidation path- way other than the cytochrome system is functioning. In Korr's terminology, tissues that have been stimulated may liberate sub- stances that "link" or "gear" the reducing systems to the cytochrome system. The by-pass theory, although possibly correct, is in my judgment a rather sweeping conclusion to make from evidence based chiefly on the use of these inhibitors, and without knowledge of the nature of an alternative pathway. Even if we assume an identical action of the inhibitors in isolated systems, as in the muscle and tissue slices, we are far from understanding the mechanism of their action. Stannard believes that cyanide affects not only cytochrome oxidase but also other enzymes essential in respiration, and, to be sure, he offers some evidence for this belief. On the other hand, this does not imply that azide is any more specific than cyanide and that only the effects of the former need be considered. It would seem more conservative at the moment to consider how the differences between cyanide and azide might be explained on the basis of the existing knowledge of the cytochrome-cytochrome DISCUSSION ON HYDROGEN TRANSPORT 47 oxidase system and of the action of nitrogen compounds on iron systems. The relations of cytochrome oxidase and cytochrome c are such in the oxidation of hydroquinone, for instance, that a decrease of oxidase, which would ordinarily cause a decreased hydroquinone oxidation, could be compensated for by an increase in reduced cytochrome. It has been shown that cyanide does in effect "remove" a certain portion of the oxidase. Hence in a resting tissue, where most of the cytochrome is in the oxidized state, a decreased oxidase could be compensated for by increased reduction of the cytochrome. In the more active state of metabohsm, where there is a small reserve of oxidized cytochrome, such compensation is less possible. The more undissociated the oxidase-inhibitor complex the less eflB- cient would be the compensation. Since azide is a "less powerful" oxidase inhibitor than cyanide, and since the greatest sensitivity to azide is found in the "active" state, these factors may be operative. Dr. Ball has considered at some length, in this symposium, the possible differences between cyanide and azide as nitrogenous com- pounds uniting with the oxidase ( Fe) and the lowering of potential caused thereby. He has pictured how a difference in the abihty of the two compounds to unite with the oxidized and reduced forms of the oxidase could lead to differences in the effective potential of the complex formed. Thus the union with cyanide could lead to an oxidase complex with a potential lower than that of cytochrome c, hence possessing a low catalytic power. As a result of combina- tion of higher concentrations of azide with both oxidized and o reduced oxidase, an effective potential might be reached which could be somewhat higher than that of cytochrome c. Although such a complex would be less eflBcient than the original oxidase, it might nevertheless be suflBcient for the low metabolism that exists in the resting state, although insuflBcient for the metabolism of the "active" tissue. This theory was not advanced to overthrow the "by-pass" theory, but only to call attention to other possible explanations for the differences between azide and cyanide. Perhaps these con- siderations should be exploited before postulating, through the mechanism of as yet unknown enzymes, a "by-pass" around the cyto- chrome system in respiration. Pasteur Effect FRITZ LIPMANN Massachusetts General Hospital WITH RESPECT TO their dependence on oxygen supply, organisms may be classified into (1) strict aerobes, equipped only with respiratory metabolic systems, (2) strict anaerobes, equipped only with anaerobic fermentative metabolic systems, and (3) facultative organisms, equipped with both respiratory and fermentative systems. This commonly used classification should not be followed too rigidly, however, for intermediate states between the main classes are com- mon in nature, and adaptive interconversion has been widely ob- served. The organisms in each of the first two groups rely exclusively on one form of energy supply, respiratory or fermentative, respectively. The third group, however, has developed the two mechanisms side by side. It is with this latter group that we shall deal, and more specifically with the interrelation between their respiratory and fermentative mechanisms. Most doubly equipped organisms possess in the Pasteur effect a regulatory device that enables them to use, as occasion demands, either their aerobic or their anaerobic systems. By the operation of this effect their fermentative apparatus is blocked in the presence of sufficient oxygen, and energy is furnished almost exclusively by the far more efficient and powerful respiratory apparatus. When oxygen is lacking, however, the fermentation system is brought into operation. The following example may serve to illustrate the energetic struc- ture of a facultative anaerobic organism. A power plant uses as a source of energy cheap water power; this may be compared to the "cheap" respiratory energy. But because of seasonal variations of flow the water power may not be entirely reliable and hence as a safeguard against a deficiency in the supply of power a more ex- pensively operating steam engine is built into the plant; this may be compared to "expensive" fermentation. For obvious reasons the plant will be equipped with a switch mechanism— its "Pasteur effect" —which keeps the steam engine from functioning so long as the water flow supplies suflBcient energy but throws it into operation when water power is lacking. 48 PASTEUR EFFECT 49 An impressive physiological example of a mechanism utilizing both respiratory and fermentative energy supply is the muscle. Figures 1 and 2, after experiments by Bang (1), reproduce measure- ments of oxygen consumption and lactic acid formation (blood lactate) on human beings during physical work. During a prolonged period of not too hard work (Figure 1) blood lactate at first increases moderately but returns, during the first quarter of the period, almost to the resting level. An anaerobic energy supply is observed only at the beginning, when an adequate oxygen supply is lacking, and until respiration climbs to the equilibrium level. For excessive short-term work the picture is different, as shown in Figure 2. An excessive and long-continued increase of blood lactate signifies a large expenditure of anaerobic energy. In such a situation the adaptation is much too WORK BLOOD LACTIC ACID millimolcs per liter 30 2- I •• 60 Minutes OXYGEN liters per minute Figure 1. — Oxygen consumption and blood lactate with moderate work. The slight initial rise of lactate output coincides with the period of adaptation, before the oxygen consumption rises to the equilibriimi level. (After O. Bang, ref. 1.) 50 A SYMPOSIUM ON RESPIRATORY ENZYMES slow to supply oxygen in time, and the muscle has to rely almost entirely on the anaerobic energy of glycolysis. The Efficiency of Aerobic and Anaerobic Metabolism Fermentations are energy-yielding rearrangements of the atoms constituting the glucose molecule. These are oxidation-reduction reactions in which, after cleavage, one part of the molecule is Figure 2.— Oxygen consumption and blood lactate with strenuous work. The excessive lactate formation signifies predominantly anaerobic energy supply. A moderate rise of oxygen consumption occurs first after completion of the work in tlie period of restitution. The slow fall of lactate in tlie blood indicates a relatively slow removal of lactate by resynthesis or oxidation. (After O. Bang, ref. I.) PASTEUR EFFECT 51 oxidized at the expense of the other part, which accordingly is re- duced. Energetically probably the most eflBcient reaction is one such as the propionic acid fermentation: 3 CgHigOg = 4 CoHjCOOH + 2 CH3COOH + 2 CO2 + 2 H2O. The energetic yield is 61 kg.-cal. of heat per mole of glucose. According to Burk (2), the change in free energy is approximately 18 kg.-cal. higher per mole of glucose than the heat exchange. The probable maximum for a fermentative breakdown of carbohydrate thus amounts to 79 kg.-cal. per mole. This is 11.5 per cent of the 686 kg.-cal. to be obtained by respira- tory breakdown. The more common lactic acid and alcoholic fer- mentations do not reach this maximum but yield only 54 kg.-cal. (36 kg.-cal. plus 18 kg.-cal. entropy change), or 7.9 per cent of the heat of combustion. These values represent the theoretical maximum for fermentation and respiration. To compare the eflBciencies of the two reactions in the cell we must know how much of the energy of each reaction is actually available to the cell. In a recent paper (3) the author has pointed out that from 40 to 70 per cent of the theoretical fermenta- tion energy is utilizable. This is deduced from the fact that in the muscle up to 40 of the 54 kg.-cal. derived from glycolysis can be stored as four energy-rich phosphate bonds in phosphagen, the energy of which is utilizable for muscular work and other purposes. Until fairly recently the view was favored that respiration energy was much less utilizable than fermentation energy— in other words, that fermentation energy was relatively more valuable than would be indicated by a comparison of theoretical caloric yields. Recent results, however, for the conversion of oxidation into phosphate bond-energy strongly indicate that such is not the case. With oxidation of pyruvic acid in the brain, Ochoa (4) found that for each molecule of oxygen consumed four energy-rich phosphate bonds were generated. With carbohydrate oxidation in heart muscle, ac- cording to Belitzer and Tzibakova (5) as many as seven energy-rich phosphate bonds (for nomenclature cf. ref. 3) might be formed per molecule of oxygen. One energy-rich phosphate bond represents from 10 to 12 kg.-cal. of utilizable energy. The six moles of oxy- gen oxidizing one mole of carbohydrate could therefore generate from 11 X (4 to 7) X 6 = 260 to 460 kg.-cal. of utilizable energy, or 40 to 68 per cent of the theoretical. The unexpectedly high yield obtained in these recent experiments shows that there is probably no great diflFerence in utilizability be- tween fermentation and respiration. Therefore it seems permissible 52 A SYMPOSIUM ON RESPIRATORY ENZYMES to take the theoretical caloric value of 54 and 686 kg.-cal., respec- tively, as a basis for comparing their efficiencies, and to conclude that only one-twelfth to one-ninth of the total possible energy is made available to the cell by the anaerobic fermentation of the glucose molecule. From this calculation the superior economy of respiratory me- tabolism becomes evident. To draw the same amount of energy from fermentation as from respiration the cell must use from nine to twelve times as much substrate. In reahty the anaerobic energy is rarely equal to the aerobic. As a rule a fully developed facultative anaerobe uses anaerobically only four to eight times as much sub- strate as aerobically, thus reaching on the average half the energy level of the aerobic state. From these considerations the economical and regulatory aspect of the Pasteur effect becomes evident. Through its operation the voluminous fermentative metabolism is allowed to proceed only in anaerobiosis, as is indicated in the fol- lowing scheme, which represents the increase in glucose utilization following change from aerobic to anaerobic conditions: OXYGEN NITROGEN (Qo, = 6; Q^^ = 0) (Q^^=10) 1. Glucose + 6 Oj -> 6 CO2 + 6 H2O 1. Glucose -> 2 Lactate 2. Glucose — > 2 Lactate 3. Glucose — > 2 Lactate 4. Glucose — > 2 Lactate 5. Glucose — > 2 Lactate Some examples of "ideal" facultative anaerobic cells are given in Table 1. From the Q values the corresponding glucose consumption and the caloric yields are calculated. One cubic millimeter of respira- tion oxygen corresponds to the utilization of 1.34 micrograms of glucose and the yield of 5.2 X 10^^ cal.; one cubic millimeter of fermentation carbon dioxide corresponds to the utilization of 4.03 micrograms of glucose and the yield of 1.2 X 10^^ cal. In the cases cited the large fermentative metabolism disappears completely in aerobiosis. In the experiment with the fish retina almost three-fourths of the aerobic energy is made available through anaerobic metabolism and correspondingly a three- to sevenfold an- aerobic increase of glucose consumption occurs. The Metabolic Structure of Cells In the middle of the metabolic type scale are placed the organ- isms that alternate between anaerobic and aerobic metabolism of PASTEUR EFFECT 53 similar eflBciency. In the upper part are predominantly aerobic types, e.g., kidney and liver and many plant cells with relatively small fermentative capacity, and at the top of the scale is a strict aerobe, azotobacter, with Qog of 2000 and no trace of fermentation. Table 1.— Effect of oxygen on glycolysis Organism or tissue Qo, Qo^ QN=F Substrate consumption Caloric yield Refer- ^ anaerobic . , anaerobic ence rate* — — ratej aerobic aerobic Torula anaerobic . . . — 260 1.04 aerobic ... .-180 18 — 0.31 Embryonic heart anaerobic . . . — 28 0.11 aerobic .... -13.6 0 — 0.018 Pigeon brain anaerobic . . . — 28 0.11 aerobic .... -16 0 — 0.022 Fish retina (30° C.) anaerobic . . . — 29 0.12 aerobic .... - 9.6 1 — 0.017 3.4 6.0 5.0 7.0 0.31 6 0.94 0.33 0.034 7 0.071 0.48 0.034 8 0.083 0.41 0.035 9 0.050 0.70 * mg. glucose per mg. dry weight per hour. f calories per mg. dry weight per hour. On the anaerobic side, below the middle, are a variety of types representing gradations down to exclusively anaerobic life. Here fermentation is partly or wholly persistent in the presence of oxy- gen, and respiration becomes a more or less residual function, BACTERIA The anaerobic type of life is most common among the bacteria, but it occurs frequently in the animal kingdom, esoecially among invertebrates. In all stages of phylogenetic development, life either chooses or is forced to adapt to anaerobic conditions, and similar metabolic arrangements correspond to similar environmental condi- tions. Transition from alternative to exclusive anaerobiosis is well illustrated by a tabulation of the metabohsm of the common yeasts (Table 2), taken from Meyerhof s classical paper (6). The almost exclusively anaerobic cultured yeasts used in the manufacture of alcoholic beverages presumably developed from the 54 A SYMPOSIUM ON RESPIRATORY ENZYMES alternating torula or wild yeast type. Baker's yeast is intermediate, having a fair respiration and partially persistent aerobic fermenta- tion. The metabolic type is not rigidly fixed. With aeration there is adaptation to the respiratory type, and with the exclusion of air the reverse is easily achieved (Table 2). It may be noted that the re- appearance of respiration is accompanied by aerobic repression of fermentation. The history of the manufacture of baker's yeast is an impressive illustration of the economic superiority of aerobic me- taboHsm (10). The earlier "Vienna" procedure of growing yeast Table 2.— Yeast metabolism Type Qo, Q^'f Q^ F (Q^'f-QQ'f)X3 Inhibition Q02 per cent Wild yeast -180 18 260 4 93 Baker's yeast -87 95 274 6.2 65 Brewer's yeast -8 213 233 7.5 8 Same after 15 hours aeration -73 113 193 3.3 42 without agitation has now been almost entirely replaced by the aeration procedure, for it has been found that through aeration the yield can be greatly increased with the same amount of culture fluid. In metabolic terms, the same amount of metabolized substrate yields a larger amount of yeast material with economical respiration than with uneconomical fermentation. As a measure of the Pasteur effect two differently derived units are recorded in the last two columns of Table 2. In the first of the two the Meyerhof Oxidation Quotient is calculated. This relates the disappearance of fermentation to the magnitude of respiration. When three times the difference between fermentation in nitrogen and fermentation in oxygen is divided by the respiration in oxygen, the quotient represents the relation between the glucose equivalent of fermentation and that of respiration. Disregarding underlying theoretical implications, it states how much fermentation glucose is replaced by oxidized glucose when respiration is allowed to occur. Stressing the economical significance of the quotient, we have pro- posed to call it a replacement quotient (3). In the next column the percentage of inhibition is calculated. From the recorded figures one would suspect a relationship of some kind between the magnitude of respiration and the Pasteur effect. This observation originally led Meyerhof to a universal application of his resynthesis theory as an explanation of the Pasteur effect. PASTEUR EFFECT 55 Although a broader discussion of the theory of the Pasteur effect is reserved for a later paragraph, some interesting experiments re- lating to the constancy of the Meyerhof Quotient may be mentioned here. These experiments on the effect on the oxidation quotient of the Z-factor of v. Euler, a factor stimulating only fermentation, are taken from a paper by Meyerhof and Iwasaki (11). See Table 3. For the same yeast, Q values, the oxidation quotient, and the percentage of inhibition are hsted with and without the stimulating Table 3.— Influence of medium on fermentation and oxidation quotient (Reference 11) Inhibition Yeast Medium Qo, QO'F Q^'F O.Q.* per cent Baker's yeast I glucose in phosphate -37.5 44.5 119 6 62 plus wort -37 96 212 9.4 55 Baker's yeast II glucose in phosphate -40.5 105 199 7 47 plus molasses -45.6 228 374 9.6 39 Strain XII sugar in phosphate -10 49 57 2.3 14 plus yeast extract -10.5 137 167 8.5 18 * Meyerhof oxidation quotient. factor. Here with the same respiration but variations in anaerobic and aerobic fermentation, the relative influence of respiration on the disappearance of fermentation increases, whereas the per- centage of inhibition remains constant. These experiments are more consistent with an explanation of the Pasteur effect as an inhibition of some kind resulting from the presence of oxygen but independent of the magnitude of respiration. Similar results were recently re- ported by Burk, Winzler, and du Vigneaud (12), who studied tlie metabolism of a biotin-deficient yeast. They found oxidation quo- tients up to 20, whereas 12 is the theoretical maximum for a re- synthesis theory. The results of these experiments are discussed in some detail in order to draw attention to two lines of approach to an explanation of the phenomenon. The first set of data, showing parallel rates of respiration and fermentation disappearance suggests an interaction between the two lines of reaction which leads to the establishment of a dynamic equilibrium of some kind. The second set of data shows conditions where the aerobic inhibition of increased fermenta- tion is independent of the rate of respiration, which remains prac- tically constant. This suggests a direct inhibition through oxygen. 56 A SYMPOSIUM ON RESPIRATORY ENZYMES Table 4 surveys lactic and propionic acid bacteria and chlorophyll- bearing plant tissue. The three types of lactic acid bacteria are representative examples which duplicate in every respect the metabolic types of the yeasts shown in Table 2. Lactobacillus del- bruckW* deserves special comment. It shows a well-developed Table 4.— Bacterial and plant metaboHsm in glucose Aerobic Material Organism Q05 QO'F QN'F O.Q. inhibi- tion per cent Lactic acid Bacterium cereale -189 49 305 3.9 84 bacteria (13) Bacterium DelbrUckii (hemin-free) -109 79 188 3 58 Lactobacillus casei 0 287 316 9 -3 255 277 22 8 Propionic acid Propionibacterium bacteria (14) pentosaceum -15 4 20 3 80 Lathyrus plant Sprout -3.5 0 1 (15) Leaf -1.4 0 0.8 Algae (15) Chlorella pyrenoidea -5 0 1.6 Coelastrum proboscideum -12.7 0 2.7 Pasteur eflFect without possessing any catalysts of the hemin type. The oxygen-producing plant tissues show only a small fermentative capacity. ANIMAL TISSUES The fairly frequent occurrence among invertebrates of organisms adapted to anaerobic life has been mentioned. Particularly among the worms is found a great variety of partly or wholly anaerobic forms. Here, on a higher phylogenetic level, appears a metabolic stratification similar to that in the yeasts and bacteria. There seems to be a gradual transition from the facultative anaerobic free-living worms to the obligate anaerobic parasitic forms (16). As early as 1909 Lesser (16a) made important quantitative experiments on the interrelation between respiration and fermentation in earthwonn metabolism. He found that fennentation resulted in the con- sumption of from four to six times as much glycogen. These same ratios were later found by Meyerhof to hold likewise for frog * This organism has been referred to by a variety of names in the biochemi- cal literature, including: Bacterium DelbrUckii, Lactobacillus Delbriikii, Bac- terium cereale. Bacillus acidificans longissimus, etc. PASTEUR EFFECT 57 muscle (17). As end products in worm fermentation Lesser found, in addition to lactic acid, large amounts of higher fatty acids, especially valeric acid (cf. also 18 and 19). An increase of glycogen utilization in anaerobiosis is described for many other types of worms (19). The parasitic worms living in the practically oxygen- free intestinal fluids show predominantly anaerobic metabolism. Al- though they are able to respire aerobically, their fermentation does not seem to be inhibited nor their glycogen consumption dimin- ished in oxygen (19). The higher vertebrates and especially the warm-blooded animals must be considered as essentially aerobic organisms. This is not true for all their parts, however, nor under all conditions. The experiment shown in Figure 2 is an example of partial anaerobiosis: the muscle suddenly put under high strain must rely predominantly on a supply of anaerobic energy. In Table 5 metabolic figures for representative Table 5.— Metabolism of animal tissues Refer- Type Tissue RQ Qo, QO'Q QN'G U.Q. ence Largely aerobic Kidney 0.8 -20 0 0-8 80 Liver 0.6-0.9 -12 0-2 0-12 80 Facultative Brain 1 -16 0 26 4.9 36 anaerobic Pituitary - -12 0 13 3.3 81 Testis 0.7-1 -14 8 14 1.3 80 Medulla of kidneys 0.97 - 9 16 23 3.3 20 Mucosa jejuni 0.86 -16 25 23 0 22 Largely anaerobic Sperm, human - 1 6.5 8 82 tissues are assembled. During recent years interesting examples of adult tissues with predominantly anaerobic metabolism have been described. Relatively large and aerobically persistent glycolysis has been found, for example, in the medulla of the kidney and in cartilage (20, 21). Dickens correlates the metabolic pattern with the relatively poor blood supply of these tissues. Large and aerobi- cally persistent glycolysis in the intestinal mucosa recently reported by Dickens and Weil-Malherbe (22) might likewise be correlated with the previously mentioned lack of oxygen in the intestinal fluid surrounding it. These observations show that the pronounced anaerobic me- tabolism of embryonic tissue and of malignant growth (Table 6) is not an isolated phenomenon. Here are tissues, as has been pointed 58 A SYMPOSIUM ON RESPIRATORY ENZYMES out in an earlier paper (23), which through environmental conditions, such as insuflBcient circulation, are forced to rely partly on a supply of anaerobic energy (cf. 20). Recent measurements by Philip (24) on respiration of the early developmental stages of the chick embryo give evidence that this must be the case. This corroborates our earlier findings based on less conclusive experimental data. Philip's remarks may be quoted here: "The study of the early blastomere," he says. Table 6.— Metabolism of tumor tissue ^°^'^'' Refer- Tumor Qo, Q^'g Q^'g CQ. tion "®^^^ ^ , ence per cent Flexner-Jobling carcinoma, rat -15 23 23 0 0 54 - 7 16 24 3 33 Adenocarcinoma, human male - 9 16 29 4.2 42 75 - 1.2 5 12 15 58 0 9 22 M 59 Jensen sarcoma -14 16 34 3.9 53 39 Walker sarcoma 256 ... -22 25 46 2.9 46 39 The average R.Q. for all the tissues was 0.85. "has revealed that oxygen diffusion limits the oxygen consumption in oxygen tensions of the air. This indicates that the early blastomere may actually be in a state of partial anaerobiosis." And later in the same discussion: "The considerations presented suggest that some of the energy used during early periods of growth can be pro- vided to the embryo by anaerobic processes. This condition is prob- ably associated with the rapidly increasing size of the embryo during early periods before the circulation system can function as an adequate oxygenating mechanism." The relatively poor vascu- larization of most tumors is evidence that the last statement holds likewise for malignant growth. We are led then to the conclusion that the high capacity of the anaerobic metabolism present in normal and in malignant growing tissues should be attributed to their partly anaerobic state of life rather than to an unlikely special growth function of glycolysis. A curious phenomenon of hyperfunction of the Pasteur effect is observed in human beings at high altitudes (25). The relatively high lactic acid level of the blood which would be expected at low oxygen pressure is observed only before adaptation occurs (25). PASTEUR EFFECT 59 After adaptation to the new environment the lactate level of the blood becomes normal. Even with exhausting work the lactic acid concentration remains very low, 2 to 3 millimolar, as compared with a blood level of 13 millimolar reached with exhausting work at sea level (Figure 2). Apparently a special mechanism prevents the muscle from utilizing too much of the anaerobic energy supply even at low oxygen pressure. Dill in his book on life at high altitudes (26) makes interesting comments on this phenomenon: "It is as though the body, realizing the delicacy of its situation with regard to oxygen supply, sets up an automatic control over anaerobic work which renders impossible the severe acid-base disturbances which can be voluntarily induced at sea level." Interpretation of the Pasteur Effect During recent years discussion has centered more or less around the question whether the effect depends upon respiratory activity as such, or upon an inhibition produced by the action of oxygen. In the first case the rate of respiration with its output of energy would be a determining factor and a state of dynamic equilibrium would result; part or even all of the respiration energy would be spent or fixed to revert or repress glycolytic breakdown. If, however, the effect is brought about through oxygen, or more specifically by oxidative inhibition of an essential part of the glycolytic enzyme system, then the reaction may be independent of the rate of respira- tion and involve no transfer of energy. equilibrium schemes A fuller understanding of the partial reactions involved in fer- mentation and respiration has given new impulse to the discussion of their interrelationship as manifested in the Pasteur effect. The fact that cozymase and adenylic acid, the two transmitter substances in fermentation, are likewise participants in respiration has given rise to some interesting suggestions. Ball (27) has pointed out that, aerobically, respiratory oxidants such as flavin may compete with pyruvic acid for the reduced cozymase. Pyruvic acid would disappear largely through oxidation rather than through fermentative reduction. Adler and Calvet (28), however, comparing the ratio of oxidized to reduced cozymase in aerobic and in anaerobic baker's yeast found no significant differ- ence, but a ratio of nearly one to one in both cases. Adenylic acid, the other common transfer system, was first linked 60 A SYMPOSIUM ON RESPIRATORY ENZYMES with the Pasteur effect through work done by Ostern and Mann (29). They found that the addition of adenosine triphosphate (ATP) to mashed muscle depressed aerobic glycolysis and raised the Meyer- hof Quotient from 2.2 to 4. Later Lennerstrand (30) discussed the possibility that with aerobic over-phosphorylation of adenylic acid (Ad), the ratio ATP: Ad might become too high to permit adenylic acid to function effectively as a transmitter in fermentation. A scheme based on the recent development of the biochemistry of phosphate turnover has been presented and discussed in detail else- where (3). Figure 3 is taken from this paper. The upper cycle, re- volving clockwise, represents anaerobic glycolysis; the lower cycle, revolving counter-clockwise, represents aerobic resynthesis. It ap- pears that the clockwise run of the glycolytic cycle depends on the outflow of the energy-rich phosphate created in the reaction. An actual reversal of the cycle back to the aldehyde stage may occur when through the aerobic influx of new energy-rich phosphate the carboxyl-bound phosphate in 1,3-diphosphoglyceric acid (Ph- glyceryl-Ph) cannot be removed. In the Meyerhof-Warburg equilibrium reaction, as the diagram shows, inorganic phosphate is bound when the reaction proceeds to the right and is set free when it proceeds to the left. Therefore in- organic phosphate concentration can become a rate-determining factor. Meyerhof et al. (31) and Belitzer (32) have pointed out that in muscle the increased concentration of inorganic phosphate through creatinephosphate breakdown should be regarded as the cause of the release of metabolic activity due to stimulation. Along similar lines, Johnson (33) recently suggested that the lowering of inorganic phosphate concentration might be a possible cause of aerobic inhibition of glycolysis. Inorganic phosphate concentration seems, however, to be high in most cells except in resting muscle, where most of the phosphate is bound to creatine. Phosphate is generally considered to be the intracellular anion. How much of this is really free phosphate and how much is labile phosphate broken down by chemical manipulations remains to be determined (34). In our opinion, the value of such schemes is limited because they disregard controlling factors in the cell which undoubtedly must regulate the routes of phosphate turnover— that is, the synthesis as well as the breakdown of intermediates. The Pasteur effect cannot be due merely to an "open" equilibrium; it must be due to specific transmitter systems. Evidence for this is the fact that the 'linkage" PH-T R I 0 S E 4- PH \ + 2 H ^GLUCOSE < + PH H.DPN PH-G L Y C E R Y L^^PH / t PH-G LYCERATE + AD'^PH / PYRUVIC E N 0 L~PH FUMARATE + PH rH-O L Y c t- /^ /a>-PH— ► 2PYRUV1C / E 0 0 L-~PM 2 PYRUVIC AC. Figure 3. — Schemes for anaerobic breakdown (upper cycle) and aerobic re- syntliesis (lower cycle) of carbohydrate. (From F. Lipmann, Advances in Enzymology, vol. 1, 1941.) 62 A SYMPOSIUM ON RESPIRATORY ENZYMES between respiration and glycolysis may be interrupted without im- pairing the reactions proper. INHIBITION OF THE PASTEUR EFFECT A considerable variety of compounds are able to interrupt the Pasteur eflFect, or the Pasteur reaction, as Warburg (35), after dis- covering the specific action of ethyl carbylamine, first called the phenomenon. Table 7 presents a survey of the agents which have been given the most study and which have proved most effective. Similar effects were found with phenylhydrazine by Dickens (36), with dinitrocresol and dinitrophenol by Dodds and Greville (41), and with HCN on certain plant cells by Genevois (42). Table 7.— Aerobic release of glycolysis , , ., .^ Concentra- Inhibitor tion Qo, QOg R.Q. Tissue Refer- ence Ethyl carbylamine 0 10-3 M - 13 - 14 19 32 Jensen sarcoma 35 Carbon monoxide, light 88% - 11 6 0.72 Allantois 37 Carbon monoxide, dark 88% — 11 11 Oxygen pressure 95% - 25.5 0 1.05 Chorion 38 5% - 22 11 0.7 Phenosafranine 0 - 13 1 1.03 Brain 39 10-6 M - 13 21 0.98 Glutathione 0 -230 74 Yeast 40 2, .5X10-3 M -205 255 The common effect is the release of aerobic glycolytic action up to an anaerobic level, while respiration remains quantitatively un- changed. In Laser's experiments (37, 38) with carbon monoxide and low oxygen pressure the respiratory quotient was lowered, indicating qualitative changes of respiration. With phenosafranine, however, Dickens (39) found that the respiratory quotient of the brain re- mained unity and the manometric and chemical determinations of lactic acid were in excellent agreement. In this case, at least, it seems very probable that the interruption of the Pasteur reaction occurred without a qualitative change of respiration. The action of metal-specific inhibitors has been of great interest. Work in this field has revived discussion of the question whether the effect is dependent on, or independent of, respiratory activity. The old observation that in most tissues cyanide released aerobic PASTEUR EFFECT 63 fermentation by inhibiting respiration was taken as indisputable proof of the dependence of the Pasteur eflFect on the intactness of respiration. Consequently ethyl carbylamine action, affecting only the Pasteur effect, but leaving primary respiration intact, was in- terpreted as inhibition of a reaction linking respiration to glycolysis. A differential inhibition of respiration and Pasteur reaction by carbon monoxide was observed by Warburg (43a) in yeast experi- ments. Mainly interested in the respiratory effect of carbon monoxide, he remarked only incidentally upon the relatively higher sensitivity of the Pasteur reaction. Later, Laser (38) showed that in animal tissues the differences in sensitivity were pronounced. Fre- quently, he found, carbon monoxide had little or no effect on respira- tion but did cause aerobic glycolysis to appear. The release of aerobic glycolysis in animal tissues had been observed by Warburg and Negelein (43), but had been considered as a secondary effect due to inhibited respiration. From some preliminary measurements of the effect of light on aerobic glycolysis in retina in the presence of carbon monoxide, the spectrum of the Atmungsferment was charted. Since Laser (38) had found respiration in retina to be uninfluenced by carbon monoxide, these measurements, as Stern and Melnick (44) recognized, had to be reinterpreted as preliminary measurements of the spectrum of the Pasteur agent-carbon monoxide compound. Stern and Melnick then measured carefully with the Warburg il- lumination technique the relative absorption spectrum of the Pasteur agent-carbon monoxide compound. The decrease in aerobic fer- mentation on irradiation was plotted against wave length. This de- crease may be assumed to be due to the decomposition of the Pasteur agent-carbon monoxide complex. The resulting spectrum was very similar to that of the respiratory enzyme. Such measure- ments were made on retina (44) and yeast (45). Recently Melnick (45a) charted the spectrum of the respiratory enzyme of animal tissue by using heart muscle extracts in which, in contrast to the intact tissue, respiration is sensitive to carbon monoxide (46). The bands developed from these measurements are reproduced in Figure 4. The spectra of the respiratory enzymes in yeast and in animal tissue, respectively, differ greatly, as do those of the Pasteur enzymes. In each case, however, the spectrum of the Pasteur enzyme follows closely that of the respiratory enzyme, deviating only in the absorption at longer wave lengths. The consistent, although small, differences are considered as evidence of the existence of two 64 A SYMPOSIUM ON RESPIRATORY ENZYMES definitely diflFerent enzymes, one catalyzing the final step in the oxidation of metabolites, the other catalyzing the oxidative inactiva- tion of a part of the glycolytic system. This analysis seems a very promising approach at least to an elucidation of the events taking place on the oxygen side. The simi- larity between the respiratory and Pasteur enzymes suggests a direct 400 450 50C ■ ) 550 600 650 RAT TISSUE CYTOCHROME OXIDASE IT (5 oC PASTEUR ENZYME YEAST PASTEUR ENZYME RESPIR^T0RY ENZYME 1— , 1 400 450 500 550 600 m/ Figure 4. — Spectra of the respiratory and Pasteur enzymes in animal tissue and in yeast reaction between oxygen and the transmitter. This is further in- dicated by the difference in the affinity of the two enzymes for oxygen (37, 47, 76). The peculiarities of the carbon monoxide effect on respiration and a change of respiratory quotient at low oxygen tension (37) and with carbon monoxide (38), however, seem to in- dicate that the present interpretation may not represent the final solution. In spite of the interest that attaches to the metal-specific anti- catalysts, it should not be overlooked that such inhibitors as phe- nosafranine, dinitrocresol (41), and glutathione can hardly be con- sidered metal-specific. In an extensive study of the action of pheno- PASTEUR EFFECT 65 safranine and other phenazine derivatives, Dickens (39, 48) presents evidence which suggests that a flavin enzyme may participate in the transmission of the aerobic inhibition. The relationship between flavin and the Pasteur eflFect is indicated also by its occurrence in the hemin-free Lactobacillus delbriickii (13), where flavin is the only respiratory catalyst. The disturbance of the Pasteur effect in brain that attends a lack of ionic balance represents a phenomenon of great complexity. Ash- ford and Dixon (49) observed a profound metabolic change in brain slices suspended in tenth molar potassium chloride. Aerobically they found increased respiration and appearance of glycolysis; and anaerobically, gradual and irreversible disappearance of glycolysis. They correlated the metabolic changes with the well-known increase in cell permeability through potassium ion (50). Dickens and Greville (51) showed subsequently that the potassium effect is spe- cific for brain and is not found in other tissues, and that omission of calcium had a similar effect. Continuing on similar hues, Weil- Malherbe (52) observed definite effects of potassium and also am- monium ions at much lower concentrations than those used by Ashford and Dixon. This effect of electrolyte on brain metabolism signifies a great lability of the Pasteur mechanism. Warburg (54) has emphasized that the Pasteur mechanism is universally very sensitive to un- physiological surroundings. For example, in rat embryo aerobic glycolysis is high in Ringer solution but low or absent in serum or amniotic fluid. Effects of this type must be taken as an indication that aerobic disappearance of glycolysis is the result of an easily disturbed balance of reactions. REVERSIBLE OXIDATIVE INHIBITION OF GLYCOLYSIS IN EXTRACTS In order to approach experimentally the possibility that oxidative inhibition might be the cause of aerobic disappearance of fermenta- tion, I studied some time ago the effect of oxidizing agents on glycolysis and fermentation in extracts (55, 56). It was shown that the fermenting system was inactivated by small amounts of iodine and quinone. By adding indophenols as oxidants inhibition in oxygen was provoked, which disappeared in its absence, when the oxidizing dye was reduced by constituents of the extract and through its enzymatic activity. Two experiments of this type with muscle extract and yeast juice, respectively, are summarized in Table 8. It appears that addition of the dye reproduces a Pasteur effect, which occurs, 66 A SYMPOSIUM ON RESPIRATORY ENZYMES however, with neghgible respiratory activity, demonstrating the possibihty of a reversible oxidative inactivation. The inhibition of glycolysis with iodine was subsequently studied more carefully by Gemmill and Hellerman (57). With concentra- tions just high enough to obtain fairly complete inhibition they were able to recover the activity by adding glutathione or cysteine. This suggested strongly that the oxidative inhibition was due to reversible oxidation of enzyme SH-groups. Rapkine (58, 59) then showed that the oxidoreduction between phosphoglyceraldehyde and pyruvic acid was at least one of the partial reactions being blocked by oxidation, presumably of enzyme-SH. This reaction system could be inactivated by S-S-glutathione and reactivated by SH-glutathione. More recently Rapkine found the same reaction reversibly inacti- Table 8.— Induced Pasteur eflFect in cell extracts Extract Addition COsin O2 O2 Cubic millimeters C02in N2 per hour O.Q. Aerobic inhibi- tion per cent Ref. Yeast 6X10-3Mnaphthol- sulfonate indo- phenol None -49 106 — 1140 710 1120 37 85 0 56 Muscle 10-3 M dichloro- phenol indophenol None - 4 20 — 440 380 425 270 95 0 55 vated by dichlorophenolindophenol (personal communication), which might explain my earlier results with the complete glycolytic sys- tem. With these experiments the possibihty of a reversible oxidative inactivation has become firmly established. It is therefore of little significance for the question at issue that, as was shown by Michaelis and Smythe (60), many dyes, irrespective of oxidation-reduction potential, inhibit irreversibly by various mechanisms, or that naphthol- sulfonate indophenol with different yeast preparations leads earlier to irreversible inactivation than in our experience. Resides the system studied by Rapkine, a number of partial enzymes of glycolysis were found to undergo oxidative inactivation followed by reactivation with glutathione. These reactions are as follows: PASTEUR EFFECT 67 phosphoglyceraldehyde + pyruvate -^ phosphoglycerate -|- lac- tate (58) glycogen + phosphate ±5 glucose-1-phosphate (61, 62) glucose-1-phosphate ?± glucose-6-phosphate (62) adenosinediphosphate + glucose -^ adenyhc acid + glucose-6- phosphate (63) The activation by thiol compounds of glycolysis in extracts, de- scribed by Geiger and Magnes (64) and Michaelis and Runnstrom (65) thus becomes easily understandable. THIOL INFLUENCE ON FERMENTATION AND GLYCOLYSIS IN INTACT CELLS The function of glutathione is not yet well understood. It is present in practically every cell in fairly large amounts. Frequently it has been suggested that it performs the role of an oxidation- reduction buflFer. The very complexity of intracellular metabolism prevents us from making more than vague statements of that type. The protection against oxygen injury which thiol compounds give to strict anaerobes, first observed by Quastel and Stephenson (53), lends support to the assumption that their function is one of stabilization. Observations on intact cells as well as on cell-free enzyme systems suggest a regulatory effect of thiol compounds on glycolysis and fermentation. As yet it is impossible to correlate definitely the action on intact fermenting cells and on fermenting enzyme extracts or partial systems, but a promising approach seems to be opened which is worth very careful consideration. Release of aerobic glycolysis with glutathione was first observed by Bumm and Appel (66) with sliced animal tissues. Soon after- ward Quastel and Wheatley (40) made an interesting study of the effects of glutathione and cysteine on the metabolism of baker's yeast. One of their experiments is included in Table 7 above. Gluta- thione interrupts the Pasteur effect without affecting respiration. They noted that an extract of brewer's yeast had much the same effect as glutathione, which they ascribe to the large content of thiol com- pounds in brewer's yeast. With cysteine the effect on aerobic glycol- ysis was the same, but respiration was markedly inhibited. The respiratory inhibition was specific for glucose and absent when glycerol was used as substrate. More recently Runnstrom and Sperber (67) undertook a study of the cysteine effect. Accompanying 68 A SYMPOSIUM ON RESPIRATORY ENZYMES the release of aerobic fermentation they found an inhibition of the synthesis of higher carbohydrates from glucose. This interesting observation would be still more significant if respiration were not inhibited at the same time. The alternative between glucose fer- mentation and synthesis to glycogen suggests that thiol compounds are able to upset the normal aerobic reaction course from glucose-6- phosphate over Cori-ester to glycogen, forcing the glucose mono- ester into the fermentation cycle. With Propionibacterium pentosaceum a number of interesting observations were made by Fromageot and Chaix (14). Dilute sus- pensions of repeatedly washed bacteria did not ferment in the pres- ence of minute amounts of oxygen. This inhibition was counteracted by very small concentrations of cysteine or hydrogen sulfide. With unwashed and concentrated suspensions small amounts of oxygen did not affect the fermentation, but aerobically fennentation dis- appeared (normal Pasteur effect, Table 4) and was released by thiol compounds. They concluded that normally a substance is present in bacteria which protects the fermentation system against the action of small amounts of oxygen. Since with impoverished or- ganisms protection can be restored with cysteine, they assumed the protecting substance to be a thiol compound. With high oxygen pressure the physiological concentration of the protective system is not high enough to counteract the oxidative inhibition and aerobic disappearance of fermentation; that is, the Pasteur effect occurs. When the concentration of thiol compound is increased, the oxidative inhibition is blocked again, and aerobic fermentation appears. In other words, the occurrence of fermentation depends on the relative concentrations of SH-compound and oxygen, respectively. An observation reported by Dickens (68) with pyocyanine should be mentioned here. In the presence of this dye "anaerobic" glycolysis of sarcoma was inhibited when measured in unpurified nitrogen containing 0.3 per cent oxygen. At the same time a slight color remained, indicating slight reoxidation of the dye. The color and the inhibition disappeared when chromous chloride was used to ab- sorb the traces of oxygen. The parallel between this phenomenon and our dye-induced Pasteur effect in extracts, as well as Froma- geot's effect of low oxygen pressure on propionic acid bacteria, is obvious. Despite the complexity of dye effects on living cells (68, 69), Dickens came to the conclusion that in general there is a tendency for dyes with high oxidation-reduction potential to increase the PASTEUR EFFECT 69 Pasteur eflFect. In harmony with this generahzation is the increase of the Pasteur effect in tumors by ferricyanide (70) and in yeast by indophenols (71). The complexity of the dye effects, however, is il- lustrated by the action of methylene blue, which, according to early observations by Gerard (72), releases aerobic glycolysis in muscle, while it was found by Barron (73) to inhibit aerobic glycolysis in erythrocytes. Nevertheless there seems to be a parallelism between dye action in extracts and in cells and a correlation between thiol and dye effects. PASTEUR EFFECT WITH VERY LOW RESPIRATION That aerobic inhibition of fermentative metabolism is inde- pendent of respiration can be most clearly demonstrated through the occurrence of the Pasteur effect with very low respiration. Aside from the dye-induced Pasteur effect in extracts, some examples of such phenomena in living cells have already been discussed, such as inhibition by traces of oxygen or inhibition in the presence of relatively low respiration in yeast (11, 12). As a rule parallelism between the appearance of anaerobic metabolism and the dis- appearance of respiration is to be expected by the very nature of the phenomenon. The fact that the most-used inhibitors of respira- tion are metal-specific and are likewise more or less pronounced inhibitors of the Pasteur reaction, has greatly complicated the analysis. Inhibitors of respiration which interrupt the chain, not at Table 9.— Pronounced Pasteur effect with very low respiration Aerobic inhibi- Refer- Tissue Addition Q02 Q°^G Q^^G O.Q. tion per cent ence Adenocarcinoma, none -1.2 5 12 15 58 75 human male none 0 9 22 00 59 Brain of rat 10-2 M male- Ihr. -10 2 19 5 90 52 ate 2hr. -5 0.5 16 9 97 3hr. -2 3 16 19.5 82 Embryonic 2.5X10-2 M (1) -1 4 35 72 89 74 chicken heart malonate (2) -4 4 32 20 87 (3) -11 6 31 6.8 80 (4) Ihr. -11 8 6.5 75 2hr. -6 6 32 13 83 in serum- malonate (5) -9.5 5 16 3.5 69 70 A SYMPOSIUM ON RESPIRATORY ENZYMES the end where the iron catalysts are operating but at an earher stage, might be expected under favorable conditions to interrupt respiration without afiFecting the Pasteur reaction. Malonate and maleate, which block the Szent-Gyorgy cycle, might react in this way. Weil-Malherbe (52) has indeed found with maleate poisoning that there is no appreciable aerobic glycolysis in brain when respira- tion has already declined to very low levels. I found, with malonate, similar effects on embryonic heart (74). In Table 9 a survey is given of these and other experiments, where with animal cells a Pasteur effect was found with low respiration. This was observed by Rosen- thal and Lasnitzki (75) with some human cancer without inhibitors and by Kempner and Gaffron (76) with myeloblasts at 6 per cent oxygen pressure. It should be remembered that in Kempner's experi- ments with myeloblasts, while the Pasteur reaction was unaffected, respiration declined with falling oxygen pressure: the Qo^ in 95 per cent oxygen was 8; in 6 per cent oxygen, 3.2. Laser (37) found the reverse with chorion, retina, and mouse liver; that is, little influence of low oxygen tension on respiration but inhibition of the Pasteur reaction. Malonate does not have the effect described above on all tissues. I found with pigeon brain a decrease of respiration accompanied by a large increase of aerobic glycolysis (74). Similar results were reported by Kutscher and Sarreither (77) with skeletal muscle. Conclusion and Outlook It has not been my purpose to give a complete survey of the work in this field. The recent reviews by Burk (83, 84) constitute a com- petent discussion of the problem as a whole, especially with regard to earlier work and thoughts. Our purpose here has been to sum- marize mainly the facts that indicate the occurrence of an oxidative inhibition. In general the evidence may be considered indicative but not conclusive, except in a few instances. The cell may choose to eliminate unneeded anaerobic metabolism by an inhibitory mechanism rather than by a counterforce, but there are indications that such inhibition acts upon reactions directing the internal flow of energy. Substances interrupting the Pasteur linkage likewise in- terrupt synthetic reactions, as has been shown in the case of cysteine (67) and especially dinitrophenol (41). Clifton (78), while in Kluyver's laboratory, made the discovery that dinitrophenol inhibits completely the synthetic processes in microorganisms. His study was based on the work of Barker (85), who demonstrated that with PASTEUR EFFECT 71 "resting" organisms only part of the disappearing non-nitrogenous metabolite could be accounted for by oxidation, while a large part was converted into cell material, presumably carbohydrate. This conversion was completely interrupted in the presence of dinitro- phenol, in which case catabolic breakdown continued until all ma- terial was oxidized (78, 79). Dinitrophenol has therefore become an important tool for the study of the relation between anabolic and catabolic processes, which must be determined by the flow of energy-carrying reactions. In a recent paper (3) where I have discussed the generation and transfer of energy-rich phosphate bonds it is stated that a major part of metabolically yielded energy is converted primarily into phosphate bond energy. An understanding of the means by which the cell directs the flow of energy-rich phosphate bonds into pre- determined reactions should lead to a more precise understanding of the mechanism of regulative cell reactions such as the Pasteur effect and the probable related action of the hormones. REFERENCES 1. Bang, O., Dissertation, Copenhagen, 1935. 2. BuRK, D., Proc. Royal Soc. (London), B 104, 153 (1929). 3. LiPMANN, F., Advances in Enzymology, 1, 99 (New York, 1941). 4. OcHOA, S., J. Biol. Chem., 138, 751 (1941). 5. Belitzer, V. A., and Tzibakova, E. T., Biokimia, 4, 516 (1939). 6. Meyerhof, O., Biochem. Z., 162, 43 (1925). 7. Warburg, O., and Kubowitz, F., Biochem. Z., 189, 242 (1927). 8. LiPMANN, F., Skand. Arch., 76, 255 (1937). 9. Nakashima, M., Biochem. Z., 204, 479 (1928). ■'lO. Warburg, O., Biochem. Z., 189, 350 (1927). 11. Meyerhof, O., and Iwasaki, K., Biochem. Z., 226, 16 (1930). 12. BuRK, D., WiNZLER, R. T., and du Vigneaud, V., J. Biol. Chem., 140, Proc, xxi (1941). 13. Davis, J. C, Biochem. Z., 265, 90; 267, 357 (1933). 14. Fromageot, C, and Chaix, P., Enzymologia, 3, 288 (1937). 15. Genevois, L., Biochem. Z., 186, 461 (1927). 16. v. BuNGE, G., J. Physiol. Chem., 12, 565 (1888). 16a. Lesser, E. J., Ergebnisse d. Physiol, 8, 742 (1909). 17. Meyerhof, O., Pfliigers Archiv., 185, 11 (1920). 18. Slater, W. K., Biochem. J., 19, 604 (1926). 19. V. Brand, T., Ergebnisse Biol., 10, 37 (1934). 20. Dickens, F., and Weil-Malherbe, H., Biochem. J., 30, 659 (1936). 21. Dickens, F., and Weil-Malherbe, H., Nature, 138, 125 (1936). 22. Dickens, F., and Weil-Malherbe, H., Biochem. J., 35, 7 (1941). 23. LiPMANN, F., Biochem. Z., 261, 157 (1933). 24. Philips, F. S., J. Exp. Zoology, 86, 257 (1941). 25. Edwards, H. T., Am. J. Physiol., 116, 367 (1936). 26. Dill, D. B., Life, Heat and Altitude, p. 173 (Harvard University Press, 1938). 72 A SYMPOSIUM ON RESPIRATORY ENZYMES 27. Ball, E. G., Bull. Johns Hopkins Hosp., 65, 253 (1939). 28. Adler, E., and Calvet, F., Arkiv Kemi, Mineral. Geol., 12B, no. 32 (1936). 29. OsTERN, P., and Mann, T., Biochem. Z., 276, 408 (1935). 30. Lennerstrand, A., Naturwissenschaften, 25, 347 (1937). 31. Meyerhof, O., Schulz, W., and Schuster, P., Biochem. Z., 293, 309 (1937). 32. Belitzer, V. A., Enzymologia, 6, 1 (1939). 33. Johnson, M. J., Science, 94, 200 (1941). 34. LiPMANN, F., J. Biol. Chem., 134, 463 (1940). 35. Warburg, O., Biochem. Z., 172, 432 (1926). 36. Dickens, F., Biochem. J., 28, 537 (1934). 37. Laser, H., Biochem. J., 31, 1671 (1937). 38. Laser, H., Biochem. J., 31, 1677 (1937). 39. Dickens, F., Biochem. J., 30, 1233 (1936). 40. Quastel, J. H., and Wheatley, A. H. M., Biochem. J., 26, 2169 (1932). 41. DoDDS, E. C, and Greville, G. D., Lancet, 112: I, 398 ( 1934). 42. Genevois, L., Biochem. Z., 191, 147 (1927). 43. Warburg, O., and Negelein, E., Biochem. Z., 214, 64 (1929). 43a. Warburg, O., Biochem. Z., 177, 471 (1925). 44. Stern, K. G., and Melnick, J. L., J. Biol. Chem., 139, 301 (1941). 45. Melnick, J. L., J. Biol. Ghem., 140, Proc, xc (1941). 45a. Melnick, J. L., Science, 94, 118 (1941). 46. Keilin, D., and Hartree, E. F., Proc. Roy. Soc. (London), B 127, 167 (1939). 47. BuMM, E., Appel, H., and Fehrenbach, K., Z. physiol. Chem., 223, 207 (1934). 48. Dickens, F., and McIlwain, H., Biochem. J., 32, 1615 (1938). 49. AsHFORD, C. A., and Dixon, K. C, Biochem. J., 29, 157 (1935). 50. Dixon, K. C, Nature, 137, 742 (1937); Biol. Rev., 12, 431 (1937). 51. Dickens, F., and Greville, G. D., Biochem. J., 29, 1468 (1935). 52. Weil-Malherbe, H., Biochem. J., 32, 2257 (1938). 53. Quastel, J. H., and Stephenson, M., Biochem. J., 20, 1125 (1926). 54. Warburg, O., StoflFwechsel der Tumoren (Berlin, 1926). 55. LiPMANN, F., Biochem. Z., 265, 133 (1933). 56. LiPMANN, F., Biochem. Z., 268, 205 (1934). 57. Gemmill, C. L., and Hellerman, L., Am. J. Physiol., 120, 522 (1937). 58. Rapkine, L., Biochem. J., 32, 1729 (1938). 59. Rapkine, L., and Trpinac, P., Compt. rend. soc. biol., 130, 1516 (1939). 60. MiCHAELis, L., and Smythe, C. V., J. Biol. Chem., 113, 111 (1936). 61. Gill, D. M., and Lehmann, H., Biochem. J., 33, 1151 (1939). 62. Com, G. T., and Cori, C. F., J. Biol. Chem., 135, 733 (1940). 63. CoLOwicK, S. P., and Kalckab, H. M., J. Biol. Chem., 137, 789 (1941). 64. Geiger, a., and Magnes, J., Biochem. J., 33, 866 (1939). 65. Michaelis, L., and Runnstrom, J., Proc. Soc. Exp. Biol. Med., 32, 343 (1935). 66. BuMM, E., and Appel, H., Z. physiol. Chem., 210, 79 (1932). 67. Runnstrom, J., and Sperber, E., Nature, 141, 689 (1938). 68. Dickens, F., Biochem J., 30, 1064 (1936). 69. Elliott, K. A. C, and Baker, Z., Biochem. J., 29, 2396 (1935). 70. Mendel, B., and Strelitz, F., Nature, 140, 771 (1937). 71. Hoogerheide, J. C, Dissertation, Leiden, 1935. 72. Gerard, R. W., Am. J. Physiol., 97, 523 (1931). 73. Barron, E. S. G., J. Biol. Chem., 81, 445 (1929). 74. LiPMANN, F., unpublished. PASTEUR EFFECT 73 75. Rosenthal, O., and Lasnitzki, A., Biochem. J., 196, 340 (1928). 76. Kempner, W., and Gaffron, M., Am. J. Physiol., 126, 553 (1939). 77. KuTscHER, W., and Sarreither, W., Z. physiol. Chem., 265, 152 (1940). 78. Clifton, C. E., Enzymologia, 4, 246 (1937). 79. DouDOROFF, M., Enzymologia, 9, 59 (1940). 80. Elliott, K. A. C, Greig, M. E., and Benoy, M. P., Biochem. J., 31, 1003 (1937). 81. FujiTA, A., Biochem. Z., 197, 75 (1928). 82. MacLeod, J., Am. J. Physiol., 132, 190 (1941). 83. BuRK, D., Occasional Publications, Am. Assoc. Adv. Sci., 4, 121 (1937). 84. BuRK, D., Cold Spring Harbor Symp., 7, 420 (1939). 85. Barker, H. A., J. Cell. Comp. Physiol., 8, 231 (1936). Oxidases, Peroxidases, and Catalase KURT G. STERN Yale University School of Medicine IN THIS PAPER no attempt will be made to treat the subject in a comprehensive or systematic manner. A number of review ar- ticles and monographs (4, 49, 50, 51, 62, 84, 85) have been written on hemin catalysts and respiratory enzymes, to which the reader is referred for information on historical developments, basic facts, and details. By way of introduction some of the available data on hemin catalyses and on hemin-containing enzymes are given in Tables 1 to 5, which are documented by references to experimental studies. The chief aims of this presentation are to bring out certain funda- mental features shared by all the catalysts under discussion; to analyze critically some of the controversial issues in the field; and to trace some of the more recent developments. To conserve space and to avoid overlappings with other papers, enzymes such as poly- phenoloxidases, which contain copper rather than iron in their prosthetic group, and the cytochromes, which cannot be regarded as independent enzymes, will be considered only in so far as their relationship to the hemin enzymes may require it. The Common Denominator in Hemin Catalyses All reactions and catalyses in which hemins participate are either of the oxidative type or at least involve oxygen as a reactant. Under this heading are grouped a variety of processes, ranging from the transport of molecular oxygen by the respiratory pigments to the activation of oxygen or the transfer of electrons from ferrous to ferric iron. The few reports which have claimed that hemins or porphyrins have promoted hydrolytic reactions remain unconfirmed. The pri- mary step in many hemin catalyses appears to be an interaction be- tween coordinatively linked, porphyrin-bound iron and the bond between two oxygen atoms as it exists either in molecular oxygen or in the forai of a peroxide bridge. This elementary process pre- cedes or, indeed, represents what is mysteriously called the phe- nomenon of "oxygen activation." All that we know about it is that oxygen atoms thus captured acquire a state of high reactivity. The 74 OXIDASES, PEROXIDASES, AND CATALASE 75 explanation that this activation is caused by a "deformation of the electron shells of the oxygen atom" is little more than a clever means of hiding our ignorance about a process which in biological import may be likened to the primary reaction of carbon dioxide with chlorophyll during photosynthesis. It is much easier to comprehend the type of interaction represented by the oxidation of ferrocyto- chrome by the ferri form of the respiratory ferment of Warburg. This reaction must be determined, at least partly, by a difiFerence in oxidation-reduction potential between the two iron compounds, al- though it is not absolutely necessary that the normal potential of the Warburg enzyme be considerably more positive than that of the cytochrome. We know of instances where the reduced form of a reversible system is oxidized partly by the oxidized form of a more negative system, such as the methemoglobin formation by methylene blue. The extent of such an interaction is governed by rigid thermo- dynamic principles only in isolated and homogeneous systems where no side reactions take place. In a living cell, where the ferrous form of the respiratory enzyme is rapidly reoxidized by molecular oxygen and the ferri form of cytochrome is rapidly reduced by Hogness and Haas's "cytochrome reductase," the ratio of the various forms, in the steady state of respiration, may differ considerably from that in the isolated system. The catalytic power of hemins is, in the last analysis, a function of the catalytic power of the central iron atom. The schema, traced in Figure 1, illustrates how the highly complex and specialized hemin protein enzymes stem from iron in its simplest form. In the lowly iron sulfate we already encounter, in a rudimentary form, some of the features which distinguish the oxidases, peroxidases, and catalase. It is fascinating to watch the increase in catalytic activity and the increasing degree of specialization that takes place as the iron atom is riveted into compounds of increasingly complex struc- ture. In many respects hemin occupies a central position in the scheme. In its linkage to the porphyrin skeleton the iron atom reaches a new level of catalytic activity. Under conditions where simple iron salts decompose 10"^ moles of hydrogen peroxide, hemin will split 10"^ moles, representing a thousandfold gain. That the essential feature here is the iron-nitrogen-carbon bridge is indicated by Warburg's classical experiments with charcoal prepared from hemin, where this bridge apparently remains standing among the ruins of the iron-porphyrin ring system and imparts a power to oxidize cysteine and amino acids which is vastly superior to that "J m K^ 4> too ^ , -a o 50 + g^-S- "So"?, •5 _:: h- 1 3^ rt eg . u O-a . -t s P5 o c3 •IH !«! O h .-^ X ®» --^ s ! fe o o o S'fe o o o .3 a o a o o CO o. >^ I— I OJ ■4-> 03 U C5 6 (U W I ^ s »5 (^ W W W ^ . ^ -- - 'C '_ c _u _o _o o g > >. 0) V a a Hi fV -Q-Q a S ^ t^, 4) -5 >i U B Ml a; O ^a o a ti g o _3 o 'a a ga D o 2 - a-s (U O o zi 2r9 52 ^ 'C 'C 'C U ki ti hi (U s ?i ?. a i< 5 -«« a a a a a o o o .2 -O "13 TJ ■i-> a> OJ 4) -— (-( tn tH o o o o l» C/3 Cfi CO Q -^ -^ Tj a a a a O O 4) 0) W ffi W W R a "K o cd aj -Q-a a-n en fc, O y M =*- •sl a^ 4) >5 76 4J C3 y "3 .2 y en 4^ >s o *^ -y 'w' (O tH 0 OT CO t- e< ■* »» t- lO o ■^ '* ^ CO 00 o »o 00 lo O 1— I 05 Ol t~ 05 O «5 IC ifl •*•«*< T}< 03 t3 so &5 to . > T^ O ^ CI fl a cs c? 'flH CO ■"! >' o »o c C^ fa rl i-H i-H O 1 o o o W rg t" a o o o o 2-3-2 -2 Jl 03 w *; .a <« " "-J CO D *J _ cS cS cl •- o.g un 9 I- «5 O ra s ft o o o cc CO 00 i-H ■S Scoo«05is<'-ie<®t c^'S i> »0 'O »0 r-< o § . o iS CO t. 4) CS f ii 8-3 + a >> lU a> X O N o o a .b O cd ^ flj (-, fcn c aj -O "TS a; « j3 jq a;:: o o o o o o o -d -o T3 "TJ -o -d xi (l> 4; 1> 4^ 4^ 4J aj CP 4> ij 4> 4> 4) 4> w w en t/3 to w w a ci c a fl a g I 3 e ^ 4J «j « 9 ^ 03 « OQPh a 'a rl <1) 4J •i>3'2:H to cs o o a 4> fc< .£( ^ aj ^ M a aj3 o O o o ^ *-< ^ g^ ^ a a W ft 0 o O (U 4, § § 2 s a a -goo a o o S o o -S a a ttj 4^ « Q -c -a P o H y 77 a a y y be 60 o o a a a y y -a -a t4 y a. a y a 1"^ ^~ p^ KIZiPH- a a' rf a w) a a a S fl a g' M W) O 60 li iiii ■S S o o t- o a § cj y o o S a 2 2 a 2 ■« £ a a g a y^ 2 S'^ § ■5. a 2 2.S Ji tr.q _, a a ia -o cj •;: a o -2 o 2 :2.« g.o ^a^ Is b >^ a s o a o T) S 03 ft S ^ OS *J o a. .as P5 o 03 • l-( X O IH a. C rH --< 0» o» «5 •* eo in ^ ^ ai toj'2 O $:00C>000 *-> a o s I H m PL, 2 ^ a-^ en t- O O aa> 2 2 a a a a KOhOO a-s a s f^ — 9 i mi 78 -G O O X .Sf >?KQOPh M"0 0 'cj -a a CS "^j 0 0 a 0 0 a 0 N -0 N c3 -0 a ft 0 J3 1 _C t- 'a 'T -q a X c.S ■a^ tJ 4) 0 -d 'a « § ^ ^ ^ OXIDASES, PEROXIDASES, AND CATALASE 79 of ordinary charcoal. From hemin, in turn, branch out even more complex systems of still greater catalytic activity, such as the hemo- chromogens and hemochromogen-charcoal adsorbates, reaching the climax in nature's own products, the hemin enzymes. Upon linking IRON CATALYSTS Homogeneous Sysfems Inorganic Fe- Salts Heferogeneous Sysfems IRON OXIDES (MITTASCH) Iron Salt -Charcoal Adsorbate Charcoal I Inorganic Fe-Complexes Organic Fe-Complexes i Blood or Hemin Charcoal HEMIN Hemm- Charcoal Adsorbate 'N- Bases Hemochromogens — i Hemochrcmogen-Charcoal Adsorbate M/croheferogeneous Sysfems "Globm * Coll. Fe(0H)3 * -* Hemoglobin i ►Active Proteins [enzymes >Catalose, Peroxidase, Oxidase Figure 1. — A family tree of iron catalysts protoferriheme IX to a specific protein, the enzyme catalase is created which is capable of splitting 10^ moles of hydrogen peroxide per sec- ond under optimum conditions, in spite of the fact that because of the screening of the iron atom, the geometrical requirements for success- ful collisions with the substrate have become much more strin- gent (20). This spectacular increase in activity is, however, gained at the price of a decrease in the number of substrates which may be attacked by a given enzyme. Hemin itself shows marked oxidatic, a § O .3 aj o > Ph £ *^ "S w iSi^.a eS fl » .2 M W C ^ 3 o a t^ a s s .a ^ ^ &c CO '^ O 4> -.Ji a « ffi « •73 O s a ■M o a •IH s I CO h '^ g O rH O O o "^ -i ►IS 00 s .- to 00-"«3O'*'OC0O S ■* C<5 1-1 ^ o X «:> I O X o eo o* '© 50 Xi H a ■^ Ph m^Q ^. ^ -a -5 fl x.a .a fl S o 0 0-5 4>4JlUlU003S 80 if II ii a s I* I -a .a fcH tH O tl M U 4) M « d I a « o 2 'hb ^ -3 a "^ Q PL, <1 ®»i>05!-iosao»'5i-ieoM.ao • ■ • • >i o cs A .a >-" ■8 00 s» .S o o o o '^ o o o o •S so 00 (» o< S V Q n 4) ai &i i O 4) _g ars-o la «?<--- a a a a a a jg-^-V ^-^ rs-s-g 2 a a (-. ►^ fci a^ aj ;^iS e2i — «^ yridine yridine □lidazol midazol (5)-Met (5)-Met a "a a Ph Ph 1— 1 1— 1 '^ -^ ilj g g « « ' w) g a 11 1 -^ 2 S « " r1 ^-i ill *J ra iH o " .9:2 s-^-^^ ^ M K S h4 jq o C o P 9 4> Ci a.a.s 4^ o o _0Q _o _w e3 a 81 20 a -S" ^ f1 .5 <^ ■£ t a-S asa Hf! a a a l-H o o o O" "^ "O "O o 0.) a; Ph ^ J2 JD t-ri >-< ^ *-i W O O O ^ [« W CO 01 -^ -^ -^ tf 11 > OS (X »< ®» ^ ^ CO © ?o o ^ GO ^ »o >C "O HH t^ o ts 1— 1 a> ■<-> 03 O ,-v ^3 « liver iver kinse a o &. a o " .a a "a baker' myelo cells ( horse beef li pump o <0 fl v S-^ a V p a v 43 3 O rg ., S3 i 9j -a 'rt a g •s o h u U V ^ a 11 o o o -o ">> s u> O 82 p-l a ^ J3 a (XI O X So O b ^ -2 00 "^ G^ ■^ -l-> ^_^ i a b n- s « O g- O O t> u " ?l a V V lU V V u Z, '" g a a a a a "S a fl o o o o o a Si 2 (H b b b b -S a ^3 C3 ^ ^ J J •^ 9, o ^ Si> CJ o « o y « cl Jj 00 o o o o cyto — >su com] men >5 "^>»>>'^ o u u u u ,_i o> 05 t- 03 CO 00 v> lO "O "o «o o o CO ■* fO •* »« »c «5 o o •>*' (N »c iC >o o o o 1 eo eo (» .2 M tH «- ;s cs 5 f ■jj i> *-' -t^ ^ ^g ^ cS ," ^ -2 "S t- o o t- _a 4) O O rt « N ^ .3 ^ a as *> »- o T2 ■'-' fti X 00 Ol j> CO o o «c so jj 83 't« a a ^ a rH 'k 0) o a o ,_ 00 Cfl -a 1 2 o o 4^ u •^ 1 t< >1 ^— ' »~ V O § 83 84 A SYMPOSIUM ON RESPIRATORY ENZYMES peroxidatic, and catalatic activity. But the enzyme peroxidase has lost the power of splitting hydrogen peroxide without the presence of an oxygen acceptor, and, conversely, the enzyme catalase is ap- parently devoid of any peroxidatic activity. The faculty of reacting with molecular oxygen, finally, is reserved to a small number of autoxidizable hemin proteins, i.e., the respiratory fennents and the Pasteur enzymes. The ultimate in specialization is reached with hemoglobin, which alone of all iron derivatives and hemochromo- gens will combine with but will not be oxidized by molecular oxy- gen. The formally analogous reaction with carbon monoxide is shared by most ferrous complexes, such as ferrous cysteine and ferrohemochromogens of all types. As may be seen from Tables 4 and 5, at least two diflFerent types of hemins occur in hemin enzymes, namely, the red protohemin and mixed-colored pheohemins. Enzyme-Substrate Intermediates It is diflBcult to understand why there should be so much con- troversy over the existence of well-defined intermediates in the course of enzymatic catalyses in general and hemin catalysis in particular. Once it is admitted that enzymes, like all other catalysts, exert their function not by mystical, long-range forces but by actu- ally combining, at some stage of the process, with their substrates and thus creating a new pathway which, although more complex, yields a higher overall rate of reaction, it will depend solely on the lifetime of these intermediates and on their spectroscopic or other properties whether their existence can be detected by experimental means. There are now on record several perfectly clear-cut cases where such enzyme-substrate compounds have been demonstrated by the spectroscope: the intermediate formed in the catalase-ethyl hydrogen peroxide reaction (60), the complexes foiTiied between peroxidase and various proportions of hydrogen peroxide (28), and the less clear-cut observations on complexes between catalase, hydrogen peroxide, and certain inhibitors, such as hydrazine and hydroxy lamine (26). The first-named complex exhibits precisely the behavior that the kinetic theories of Michaelis and Henri postulate for an intermediate enzyme-substrate compound: it is unstable and disappears at the rate at which the end products of the reaction are formed. The peroxidase-hydrogen peroxide complexes are also un- stable, but their decomposition cannot be due to a true peroxidatic reaction, since it occurs also in the absence of oxygen acceptors. However, the recent work by Karush (25) and Chance (12) on OXIDASES, PEROXIDASES, AND CATALASE 85 rapid reactions affords further proof of the existence of enzyme- substrate complexes as intermediates in the peroxidate catalysis. The nature of the ternary complexes formed by catalase, hydrogen peroxide, and some inhibitors is not yet clear. They may represent ternary complexes between enzyme, substrate, and inhibitor, but they may also be binary enzyme-substrate compounds the lifetime of which is prolonged by the presence of the inhibitors. The inter- pretation of the phenomena observed in the foregoing enzyme reactions is reinforced considerably by observations made on such closely related, non-enzymatic models as the methemoglobin- hydrogen peroxide (21) and methemoglobin-ethyl hydrogen per- oxide (58) complexes and the hemin-hydrogen peroxide intermedi- ate (22). If we go back a little, we find that in 1905 Spitalsky observed analogous phenomena in the chromic acid-hydrogen per- oxide catalysis. It is safe to conclude that similar inteimediates arise in all hemin-catalyzed reactions even if they cannot yet be demon- strated experimentally. The writer has just been informed by Professor Hogness that cytochrome peroxidase, too, forms a typical enzyme-substrate complex with hydrogen peroxide. The oxidation of ferrous iron to ferric iron by oxygen is usually regarded as a process fundamentally different from the oxygenation of hemoglobin. However, it would simplify matters a good deal if both types of reactions could be brought to a common denominator. In agreement with Haber* and Warburg the reviewer believes that this common link is that in the ferrous-ferric transformation by molecular oxygen an oxygenated intermediate of a structure analo- gous to oxyhemoglobin is interposed (see Oppenheimer and Stern (49), pp. 14ff.): Fe^+ + O, -» Fe^O^ -» Fe^^\ The reasons for this hypothesis are, first, that molecular oxygen, despite its high potential ( + 0.8 volt), appears to be too sluggish an oxidant to react rapidly with ferrous iron without prior activation. It is logical to assume that the hydrogen peroxide, formed as a result of the oxidation of the ferrous iron, arises by interaction of the ferrous-oxygen intermediate with water molecules or hydrogen ions. Secondly, the well-known competition of molecular oxygen and carbon monoxide for the ferrous fomi of the respiratory ferment, which is governed by Warburg's distribution equations, would be * This concept appears applicable even to the simplest reactions in the gas phase. Thus Haber and Sachsse (16), on the basis of kinetics experiments, conclude tliat during the reaction of sodium vapor with oxygen one sodium atom combines with one oxygen molecule. 86 A SYMPOSIUM ON RESPIRATORY ENZYMES diflBcult to understand if the primary reaction of the enzyme with oxygen were not a reversible equihbrium reaction. This is particu- larly true for the photochemical dissociation of the iron-carbonyl complex, where it is assumed that, under the influence of light, carbon monoxide is reversibly exchanged with oxygen. All phe- nomena may be satisfactorily explained on the basis of the assump- tion that the primary step consists in the formation of a ferrous- oxygen intermediate and that the further course of events is gov- erned by the lifetime and reactivity of this complex. In the instance of oxyhemoglobin the intermediate stage is fixed in a unique manner under the influence of the globin component. In that of the oxidases, the intermediate is very short-lived, with the possible exception of the respiratory ferment in baker's yeast (75), where Warburg has observed an absorption band which he tentatively attributes to an oxygen addition compound of the enzyme. A unified theory of oxidase, peroxidase, and catalase action could, then, be based on postulating, with a fair degree of probability, the formation of "moloxides" and "molperoxides," respectively, as the primary process in the catalysis. The mechanism of the oxidation of hemoglobin to methemoglobin by oxygen deviates from this schema because of the high stability of oxyhemoglobin. The well-defined maximum of reaction velocity at low oxygen pressures is in this instance (10, 48) largely due to a decrease in the concentration of the free, autoxidizable ferroform as the partial pressure of oxygen in the system is increased. The rate of oxidation is proportional to the concentration of reduced hemoglobin and to a function of the oxygen pressure. In this case oxygen represents both a reactant and an inhibitor, not only because of oxyhemoglobin formation but also, perhaps, through breaking of chain reactions. On the Mechanics of Hemin Catalyses When considering reactions promoted by hemins, one almost in- variably encounters the notion that they must all conform to the pattern of the ferri-ferro cycle. There is no question that this is the most handy explanation: everybody knows that the hemin iron may exist in the reduced (Fe+^) and in the oxidized (Fe+++) state and that this transformation may be accomplished in a reversible manner even if no bases are linked to the heme. But like many generaliza- tions, this concept as the only explanation is not only hazardous but definitely too narrow. Let us go back for a moment to the theories that have been ad- OXIDASES, PEROXIDASES, AND CATALASE 87 vanced to explain the catalytic or semi-catalytic effects of iron and other metal ions (37). As in the case of the hemins, we find a widespread inclination to explain everything by a valency change of the metal during catalysis: in the first stage the substrate reduces the trivalent iron (or the bivalent copper) to bivalent iron (or mono- valent copper); in the second stage the oxidizing agent, e.g., oxygen or hydrogen peroxide, regenerates the ferric iron or cupric copper, respectively. There are unquestionably instances where this simple hypothesis will serve to explain all the facts, e.g., in the metal- catalyzed oxidation of phenols by oxygen. But in many other in- stances the hypothesis proves inadequate. Two of the main obstacles are, first, the fact that bivalent iron salts and complexes are so often superior in catalytic activity to the corresponding trivalent compounds and, second, the phenomenon of the "primaerstoss" (alpha-activity). By this we mean the frequent observation that oxidation catalyses, in the presence of ferrous ions, exhibit an initial phase of high velocity which is followed by a steady state of a much lower reaction rate (beta-activity). The transition of ferrous into ferric iron in the course of this process is accompanied by a turnover of many more substrate molecules than would correspond to the number of iron equivalents. Three different theories have been proposed to explain the facts. All postulate the fonnation of labile and highly reactive intermedi- ates which are formulated as peroxides by Manchot, as complex compounds by Wieland, and as free radicals by Christiansen, Baeck- stroem, and Haber. Manchot assumes that only the bivalent iron is capable of forming peroxides of the type Fe^Og or, more recently, of molperoxides of the type Fe'^^HoOg. Such peroxides could oxidize two substrate equivalents and, subsequently, react with hydrogen peroxide present in excess to yield inactive ferric iron, or they could interact with excess hydrogen peroxide to form oxygen while the ferrous iron is regenerated. Wieland, on the other hand, believes that the ferrous iron forms a complex with the substrate or with other substances present in the system with the possible inclusion of the oxidizing agent. By the complex formation tlie substrate hydrogen is "activated" and is thus made accessible to the attack by the oxidant. The alpha-activity ("primaerstoss") is attributed to a temporary protection of the active ferrous iron contained in the complex against the transition into inactive ferric iron. In this way one ferrous ion is enabled to oxidize a larger number of substrate molecules, the oxidized molecules being released from the complex 88 A SYMPOSIUM ON RESPIRATORY ENZYMES and replaced by fresh substrate molecules. In the well-known chain reaction schemas, as developed by a number of authors, the metal is assumed to initiate the reaction by reacting with a substrate molecule and producing a molecule of high energy content ("energy chains" of Christiansen and of Baeckstroem) or monovalent radicals of high reactivity ("radical chains" of Haber and Franck), where- upon the chain is propagated by such intermediary radicals or "hot" molecules without the further participation of the metal. The chain length, i.e., the yield in product molecules per elementary act initiating the process, is determined by the probability with which two of the intermediary radicals or energy-rich molecules will collide and inactivate each other, and by the presence or ab- sence of specific inhibitors, so-called "chain-breakers." It is im- portant to note that in the metal catalysis of sulfite oxidation, cupric copper and ferrous iron but not ferric iron are able to initiate reaction chains (14). In this connection, the work of Haber and Weiss (17) on the decomposition of hydrogen peroxide by ferrous salt and the nature of the alpha -activity is of particular interest. As is well known, a small amount of ferrous salt, when brought together with a large excess of hydrogen peroxide, is oxidized to ferric salt with the simultaneous liberation of oxygen. The yield depends upon the rate at which the two reactants mix; the ratio AHgOa/AFe'^'^ may reach values as high as 15.6. The underlying chain reaction is formulated by the authors as follows: FV" + H2O2 = Fe^^^OH + OH OH + H2O2 = H2O + OH OH + H2O2 = 02 + H2O + OH Fe^^ + OH = Fe^^^OH The last equation depicts a chain-breaking reaction. If all the Fe'^^ has been oxidized, the reaction stops.* It is perfectly true that this process, like many reactions studied by Wieland and his stu- dents, is not a true catalysis but an induced reaction. But it is like the true catalyses in that small amounts of a promoter or inductor bring about the reaction of a disproportionately large number of substrate molecules. On the other hand, we know that all enzymes are slowly but irreversibly "consumed" during the reactions which they catalyze. Only a finite, although large, amount of protein or hydrogen peroxide can be split by a given quantity of proteinase " For an analysis of the catalytic decomposition of hydrogen peroxide by ferric salts in acid solution the reader is referred to the papers by Haber and Weiss (18), and Kuhn and Wassermann (36). OXIDASES, PEROXIDASES, AND CATALASE 89 or catalase, respectively. This observation is usually interpreted in terms of an irreversible destruction of the active protein component of the enzyme in some side reaction, e.g., by denaturation or as a result of attack by other enzymes present as impurities. While this may be so, it is not always easy to distinguish clearly between a true catalysis, where the catalyst is progressively eliminated by side reactions, and an induced reaction, where the inductor is slowly converted into an inactive form. This is particularly true of experi- ments in biological systems. According to all three theories outlined above, the reaction comes to a standstill once all ferrous iron has been converted into ferric iron. A continuation of the process is obviously possible only if the substrate or some other component in the system is able to reduce Fe^^^ back to Fe^^.* One could visualize "hybrid" processes where the ferrous iron, in the main reaction, acts as an inductor and where the inactive ferric iron is slowly reduced to the active ferrous form by some "outsider" such as a thiol and is thereby enabled to start the induced reaction all over again. Such a situation, if encountered in living cells, would probably defy any attempt to distinguish between true and apparent catalysis, especially if the reducer is constantly replenished from suitable precursors, e.g., glutathione from protein breakdown. We go one step further. If iron can break down the potential barrier, shielding stable substrate molecules, either by the ferri- ferro cycle or by the induced reaction mechanisms just mentioned, is it not possible that iron, when linked up in suitable complexes, could bring about changes in certain substrate molecules or initiate chain reactions without itself suffering a change in valency? It is on this possibility that the issue of the mechanism of catalase action largely hinges. It is well established that the iron in catalase exists in a remarkably stable ferri state. The enzyme is invariably isolated from all sources as the ferri foiTn and it defies reduction with activated hydrogen or with hydrosulfite. This property is unique among hemin proteins; it is not shared even by peroxidase, which may be readily reduced to the ferro form by agents such as sodium hydrosulfite. Moreover, it can be shown that the spectroscopically well-defined intermediate arising in the catalase-ethyl hydrogen * Theoretically, ferri ion could act catalytically by being reversibly oxidized to a higher stage of valency. Supporting such a view are the spectroscopic observations of Bohnson and Robertson (J. Amer. Chem. Soc, 45, 2493 (1923), on weak acid solutions of ferric salts in the presence of hydrogen peroxide. The color changes occurring in this system were interpreted by these workers to be due to the formation of ferric acid. See, however, Haber and Weiss ( 18 ) . 90 A SYMPOSIUM ON RESPIRATORY ENZYMES peroxide reaction contains the enzyme in the ferric state (61). Carbon monoxide, which will inhibit many hemin catalyses where ferrous iron is involved, has little or no specific effect on the catalase- hydrogen peroxide reaction (61), other reports to the contrary notwithstanding. Some years ago Haber and Willstaetter (19) pro- posed a chain reaction schema for this catalysis which embodied, as the initial step, a reduction of the enzyme iron to the ferrous form. A little later the present writer modified this schema some- what (55) with a view to avoiding the necessity of assuming a ferri-ferro cycle. The initial step in that schema consisted in the interaction of two adjacent porphyrin-bound ferri atoms with one hydrogen peroxide atom to yield two monovalent OH-radicals which could then propagate the chain. The iron was assumed to remain in the trivalent form throughout. More recently Keilin and Hartree have advanced a different hypothesis, incorporating the idea of the ferri- ferro cycle (27): Step I 4 Fe"*^ + 2 H^O^ = 4 Fe"^ + 4 H" + 2 O, Step II 4 Fe^" + 4 H^ + O^ = 4 Fe^^^ + 2 H2O 2 H2O2 = 2 H2O + O2 The authors state that, in accordance with this schema, the catalysis is greatly inhibited or even suspended in the absence of free oxygen. It will be noted that the hydrogen peroxide is here assigned the role of a specific reducer of the ferri form of catalase, whereas molecular oxygen is considered as the oxidizing agent in the regeneration of the ferri form. The concept of hydrogen peroxide as a reducing agent in itself, although somewhat startling, is not new and finds support in the earlier finding of Kuhn and Wasser- mann (36) that ferric salts, in the presence of such complex formers as a, a'-dipyridyl or o-phenanthroline, are quantitatively reduced by hydrogen peroxide. But the schema is open to other objections, both on theoretical and on experimental grounds. Thus Johnson and van Schouwenburg (24), Weiss and Weil-Malherbe (80), Sumner and Dounce (cf. 68), and the writer (61) failed to confirm the observation of Keilin and Hartree that catalase is inactive under strictly an- aerobic conditions. Furthermore, it seems somewhat strange that, according to the schema of these workers, molecular oxygen should be required for the reoxidation of the ferrous form of catalase in a system containing hydrogen peroxide, which is generally considered a more active oxidizing agent than oxygen, as Dr. M. Gorin has pointed out, in a private communication to the writer. According OXIDASES, PEROXIDASES, AND CATALASE 91 to Dr. Gorin, one would expect, if Keilin's theory were correct, that the reduced form of catalase would accumulate in contact with hydrogen peroxide and in the absence of air. The spectrum of the ferro form of the enzyme, as obtained by treatment with sodium hydrosulfite in the presence of hydrogen sulfide and subsequent removal of the latter, has recently been described by Zeile et al. (87). No spectral change, on the other hand, has as yet been reported for a mixture of catalase and hydrogen peroxide under nitrogen. Attempts to "catch" those ferro-catalase molecules, which might possibly be formed as intermediate products during the catalase- hydrogen peroxide reaction, with the aid of carbon monoxide have been unsuccessful (61). The story of the eflFect of carbon monoxide on catalase under various conditions is just as controversial as the subject of its reaction mechanism (cf. 11, 61, 27). It may suffice here to mention Keihn and Hartree's observations (27) that purified carbon monoxide, in the absence of oxygen, exerts an inhibiting eflFect on the enzyme which is not reheved by light, whereas certain catalase preparations may be made sensitive to carbon monoxide inhibition in the presence of oxygen by adding traces of azide, cysteine, and glutathione. This eflFect of carbon monoxide is stated to be completely relieved in a reversible manner by light. The mechanism of this "sensitization" and the reason why crude enzyme preparations are more readily inhibited by carbon monoxide than chemically purified fractions are still obscure. It would seem that, for the present at least, the observations made with the use of carbon monoxide aflFord no basis for supporting or rejecting Keilin and Hartree's reaction schema. The same is true, in the writer's opinion, of the spectroscopic observations made by Keihn and Hartree (26) on catalase solutions containing sodium azide or hydroxylamine in addition to hydrogen peroxide. Let us now turn to the theoretical objections raised against Keilin and Hartree's schema. As Weiss and Weil-Malherbe (80) point out, an exclusive reoxidation of ferro-catalase by oxygen would in eflFect prevent a decomposition of the hydrogen peroxide. For unless a radical chain mechanism is postulated, the oxygen formed by the reduction of the ferri form of the enzyme by hydrogen peroxide (Step I, p. 90) is quantitatively used up again for the reoxidation of a stoichiometric amount of the ferrous form with a simultaneous reduction of the oxygen to hydrogen peroxide. To avoid this diffi- culty, Keilin and Hartree formulate the oxidation reaction (Step II, p. 90) in such a way that the reduction of oxygen to water does 92 A SYMPOSIUM ON RESPIRATORY ENZYMES not involve the intermediary formation of hydrogen peroxide. This is in opposition to the general views on the mechanism of autoxi- dation and would appear highly improbable, since it postulates a reaction of a very high order. Furthermore, on the basis of Keilin and Hartree's hypothesis, the reaction, when proceeding in nitrogen, should exhibit the characteristics of an autocatalytic reaction be- cause of the production of increasing amounts of oxygen during the process. This, however, is not indicated in the data of these authors. Sumner and Dounce (cf. 68), who also were unable to confirm Keilin and Hartree's observations with respect to the importance of oxygen for catalase action, prefer the following schema, which is based on earlier ideas of Haber, Euler, and Liebermann: Step I Fe-OH + H^O^ = Fe-OOH + H2O Step II Fe-OOH + H2O2 = Fe-OH + H2O + O2 In this schema, catalase in its ferric form is represented by the symbol Fe-OH, which is often employed for methemoglobin, a molecule very similar to catalase. The symbol Fe-OOH represents an intermediary catalase peroxide which is assumed to react with a fresh substrate molecule to yield oxygen, water, and the free ferri form of the enzyme. It will be noted that this hypothetical schema, in contrast to that of Keilin and Hartree, does not involve the ferri- ferro cycle and is in agreement with the spectroscopic observation of a catalase-peroxide complex in the enzyme-ethyl hydrogen per- oxide reaction (58, 60). But the writer doubts very much whether so simple a schema is adequate to explain all the features of the enzymatic catalysis. We must not forget that the decomposition of hydrogen peroxide can be catalyzed by a variety of agencies besides the enzyme, such as ultraviolet light, dust, metallic and non-metallic surfaces, colloidal platinum, inorganic ferri and ferro salts, cobalti salts, etc. We are therefore confronted with the necessity of finding an explanation for the reaction mechanism which will be equally applicable to these various catalysts. The central theme is, of course, the mode in which hydrogen peroxide is transfonned into water and oxygen. The most cogent formulation for this central process has been given by Haber and Weiss (18): ( 1 ) OH -f H2O2 = H2O + HOl (2) Ha + HaOj = O2 -(- H2O + OH In this schema OH and HO. are monovalent radicals. As Haber em- phasizes, in a paper published posthumously (18), the progress of OXIDASES, PEROXIDASES, AND CATALASE 93 the reaction through radicals is the main point of the concept, while the propagation of reaction chains through such radicals, although an interesting phenomenon, is of secondary importance. In fact, the two features are not necessarily associated with each other. The detailed study of the kinetics of the decomposition of hydrogen peroxide by ferrous and ferric salts suggests strongly that under certain conditions the radicals may give rise directly to the formation of the end products, water and oxygen, without initiating a chain by further reacting with hydrogen peroxide and thereby repro- ducing themselves. Thus the catalytic breakdown of hydrogen per- oxide by ferri ions in acid medium does not, in general, represent a chain reaction; the oxygen is released here by the process (3) Fe^^^ + Ha = Fe*" + H^ + O2 Upon slightly changing the experimental conditions, e.g., by increasing the hydrogen peroxide concentration or decreasing the Fe'^^'^, reaction 2 is favored, with the result that chains appear (18). The second question relates to the way in which the central process, consisting of steps 1 and 2, is initiated by the various catalytic agents. The radicals OH and/or HO, can be created only by a monovalent attack on hydrogen peroxide. It is the monovalent character of the primary reaction of the catalyst with the substrate which, according to Haber and Weiss (18, 78), is the common feature of the chemical, photochemical, and electrochemical primary proc- esses. Thus the radicals may arise by the transformation of a metal ion into the one of next higher valency, as in the instance of ferrous ions, (4) Fe"" + n^O. = Fe"^^ + OH- + OH or by the reduction of a higher valent state, as in the case of ferric ions: (5) Fe^*^ + HO=- = Fe^" + Ma Hydrogen peroxide may also be split into two OH radicals by ultraviolet light: (6) H202 + hv = 2 OH In heterogeneous systems, involving metal surfaces, the radical OH is beUeved to result from a simple electron transfer, (7) H,02 + electron („,etai) = OH- + OH While reaction 7 is an expression of the oxidizing action of H2O2, the reducing properties of the molecule are attributed to its anion (HO2"), in accordance with the equation 94 A SYMPOSIUM ON RESPIRATORY ENZYMES ( 8 ) HO2- = Ha + electron Both these functions are related by the dissociation equihbrium (9) H2O2 ^11"^ + HOr (dissociation constant, K) It follows that there is no fundamental difference between the homogeneous and heterogeneous reactions with regard to the ele- mentary process. Weiss (79) has recently discussed the mechanism of catalase and peroxidase action from the same point of view. In his presentation of the central reaction, step 2 is replaced by the equation (2a ) O2- + U2O2 = OH- + OH + O2 where O2- has been substituted for HO^ on the basis of new evidence. These two radicals are related to each other by the electrolytic dis- sociation equilibrium (2b) O2- + H* <^ HO2 (dissociation constant, KHO2). Weiss treats the catalase-hydrogen peroxide reaction as a hetero- geneous catalysis and as an analogon to the catalytic decomposition of the same substrate by colloidal platinum, which has been so carefully studied by Bredig. The similarity is accentuated by the fact that neither catalase nor colloidal platinum shows peroxidase activity toward acceptors of the type of iodide ion or pyrogallol. In both instances the primary process is formulated as a surface reac- tion involving the radicals OH and HO,. The author assumes that the iron atom of the enzyme is alternatively reduced to the ferrous and reoxidized to the ferric form by the hydrogen peroxide during the course of the catalysis. The role of the metal atom in the porphyrin skeleton is regarded as that of facilitating rapid electron transfers, since the valency change of the iron in the porphyrin ring system takes place without appreciable dislocation of heavy par- ticles, in contrast to the situation in the instance of free ferrous and ferric ions, where the water dipoles in the hydration shell must undergo rearrangements upon a change in the charge of the central atom. Furthermore, he believes that the system of conjugated double bonds surrounding the iron in the heme group makes for a rapid "conduction" of the inner electron by virtue of their loosely held n-electrons. In analogy to the action of colloidal platinum on hydrogen per- oxide, as formulated by the same author, the mechanism of catalase action is depicted as follows: OXIDASES, PEROXIDASES, AND CATALASE 95 (10) Fe^^^ + HOr ^ Fe^- + HO2 (11) Fe^^ + H2O2 = Fe^^^ + OH- + OH ( 12 ) Fe+^ + OH = Fe*^^ + OH" The production of molecular oxygen is attributed to reactions 1 and 2a. It will be noted that the process, although involving free radicals, is not a conventional chain reaction, since the enzyme iron plays an active role in three stages: 10, 11, and 12. This is partly based on the finding that in the model system platinum-hydrogen peroxide an average chain length of only about five links may be assumed. It is therefore considered probable that in the enzymatic catalysis, under the usual experimental conditions, the chain reaction prac- tically degenerates to a simple radical reaction. It will be recalled that one of the criticisms directed against the chain reaction theory of Haber and Willstaetter by Haldane was that one would expect the rate to be proportional to the square root of the enzyme con- centration rather than to the enzyme concentration itself, as is actually the case. By postulating very short chains, as Weiss does, the feature of the proportionality between reaction velocity, enzyme concentration, and substrate concentration is retained without sacri- ficing the essential concept of Haber, i.e., the postulate of inter- mediate, monovalent radical formation. The further objection of Haldane that the assumption of the same type of radicals (OH and HO2) in various kinds of enzyme reactions, as was done by Haber and Willstaetter, was in conflict with the well-known specificity of the oxidizing enzymes, is also met if very short chains are as- sumed. In this case the radicals are present only in so low a con- centration that their oxidizing action on acceptors, e.g., iodide ion or oxyphenols, remains below the threshold of sensitivity of the analytical methods. All these considerations refer to the "normal" course of the catalase reaction, i.e., to conditions where the enzyme and the substrate concentration are within the range usually em- ployed in kinetic studies and activity determinations. There can be little doubt that the explosive type of hydrogen peroxide decompo- sition, such as is produced in concentrated peroxide solutions when a relatively large amount of enzyme is added, represents a chain process with a long chain length. There may even be branched chains, such as are assumed to occur during "knocking" in internal combustion engines, when the relatively slow combustion along a flame front of regular rate of progression changes over into detona- tion (cf. Lewis and van Elbe, 40). 96 A SYMPOSIUM ON RESPIRATORY ENZYMES It is not feasible, within the space allotted to this discussion, to cite all the experimental evidence and theoretical arguments in favor of the intermediary formation of free radicals or of the chain reaction character of the hydrogen peroxide catalysis or of oxidative enzyme action in general. It must suffice, therefore, to refer to the original publications where the existence of OH radicals during hydrogen peroxide photolysis (73a) or the formation of the HO2 in reactions with molecular oxygen (78a) have been observed. The experiments of Schwab et al. (53a) on the effect of typical chain- breakers on the catalase reaction did not yield conclusive results. But the experiments of Barron (5) on the effect of antioxidants on the rate of oxidation of unsaturated fatty acids by hemochromo- gens are very suggestive. The pioneer work of Michaelis on the radical nature of the semiquinones, which arise through monovalent reduction or oxidation of reversible dyestuffs makes it all but im- perative to admit the intermediary existence of monovalent radicals (monohydropyridine, monohydroalloxazine, monohydrothiamine, etc.) during cellular respiration, unless reactions of high order are postulated. Those who are interested in the manner in which the action of other oxidative enzymes, e.g., peroxidase or dehydrogenases, may be interpreted in terms of radical chain mechanisms are referred to the papers by Haber and Willstaetter (19), Weiss (79), and the review article by Moelwyn-Hughes (45). AUTOXIDIZABLE IrON COMPOUNDS We return to the starting point of the discussion: What makes the iron atom and the oxygen molecule "click" during the primary process? Why are some iron compounds autoxidizable and others not? No satisfactory answer may at present be given to this ques- tion, which certainly attracts the attention of many workers in the field of biological oxidation. Only partial solutions have been of- fered to this key problem. Ferrous iron is relatively stable in acid solution and is rapidly oxidized to ferric iron by molecular oxygen in alkaline solution. The state of the metal under these two conditions obviously differs in one respect. In the acid medium the iron is present as ferrous ion, whereas in alkaline solution it exists as non-ionized ferrous hydrox- ide. Smythe (54), in an interesting paper, points out that the oxi- dation of ferrous ion involves the separation of a negative charge OXIDASES, PEROXIDASES, AND CATALASE 97 (electron) from a nucleus which already carries two positive charges. Non-ionized ferrous hydroxide, on the other hand, should be more easily oxidized, since this process represents only the separation of the electron from an electroneutral substance. If this reasoning is sound, one would expect that ferrous iron, if built into a non-ionized compound, should be readily autoxidizable regardless of the acidity of the medium. The rate of autoxidation of ferrous complexes would then be expected to be a function of the pH only in so far as the stability of the complex is affected by changes in hydrion concen- tration. In other words, under any given condition the velocity of the reaction with oxygen should be proportional to the concen- tration of the ferrous complex, which in turn is a function of the concentrations of the ferrous ions and the complex-forming anions. Since the latter, within a given pH range, depends on the hydrion concentration, one would expect that within this range the rate of oxidation would vary with the pH. Smythe proceeded to test this working hypothesis on a series of inorganic and organic iron com- plexes of relatively simple configuration. His manometric study of the rate of autoxidation of ferrous pyrophosphate and ferrous meta- phosphate, which are stable over a wide pH range, bore out the prediction. The reaction rate varied strongly with the hydrion con- centration, decreasing markedly in both instances with an increase in acidity. The case of ferrocyanide is of special interest because its structure bears a certain resemblance to the core of iron porphyrin complexes. In both instances the iron atom is surrounded by four groups, containing nitrogen linked to carbon atoms. In the case of the ferrocyanide these groups carry negative charges. To escape from the metal atom, an electron must pierce this "negative atmos- phere," which is obviously difficult. We are not surprised, therefore, to find that the rate of autoxidation of ferrocyanide is extremely small in spite of the fact that the iron is present in covalent rather than in ionic linkage. If it were possible to reduce the negative charge on the groups encircling the iron atom, its reaction with oxygen should be facilitated. Indeed we find that if the charge is reduced by re- placing one of the CN~ groups by ammonia, a compound, penta- cyanoammine-ferroate, is formed which is attacked by oxygen at an appreciable rate. The striking observation of Baudisch and Davidsohn (7) that the oxygen uptake is more rapid at pH 2 than at 7 or 12, in contrast to the findings obtained with iron pyro- or metaphosphate, is explained by Smythe in terms of a suppression of the ionization under the influence of the increased hydrion concen- 98 A SYMPOSIUM ON RESPIRATORY ENZYMES tration and a consequent decrease of the negative charges around the iron atom. However, Smythe's views are not shared by Weiss (78). This author beheves, in direct contradiction to the hypothesis that only coordinatively bound ferrous iron is autoxidizable, that only "free" ferrous ions react, in general, with molecular oxygen. By "free" ions Weiss means iron with incompletely filled electron orbits. The autoxidation of ferrous sulfate is formulated in accordance with Weiss's theory of simple electron transfers as follows: O2- + H* ?^ HOT (13) Fe^^ + O2 -> Fe^^^ + O^-, (14) Fe^^^ + O2- -> Fe^* + O2 (15) Fe^^ + HO2 -^ Fe^^^ +HO2-, HO2- + H^ = H2O2 Equation 15 explains the formation of hydrogen peroxide as an end product of autoxidation processes. If, however, it is not caught in statu nascendi, e.g., by cerium hydroxide, the hydrogen peroxide is relatively rapidly decomposed by reaction with ferrous iron- ferric ions and hydroxyl ions or water being formed as final prod- ucts. During the entire process the oxygen molecule is stepwise reduced by four electrons with the formation of four hydroxyl ions. The sequence given above, according to Weiss, provides a logical explanation for the slow rate at which an acidified ferrous sulfate solution is autoxidized. The latter does not indicate that the inter- action between the ferrous ions and oxygen (reaction 13) is slow but, on the contrary, that the rate of the back reaction 14, which involves a reduction of ferric ions, is quantitatively significant. Under stationary conditions and for a given partial tension of oxy- gen, the rate of reaction 15 is defined by the ratio of the velocities of the partial reactions 13 and 14; in other words, by the ratio [Fe^*] to [Fe^^*]. The greater this ratio the more positive will be the oxidation-reduction potential and the higher will be the rate of ferrous salt oxidation. The essential feature in reactions of this type, according to Weiss, is not the formation of stable ferrous iron com- plexes but the fact that the ferric ions are protected against reduc- tion by the formation of still more stable ferric iron complexes. Thus, in instances where the ferrous complex is more stable than the corre- sponding ferric complex, as in the case of ferrous tridipyridyl sulfate, no autoxidation takes place. A decidedly more mechanistic explanation is offered by Theorell (73) for the failure of cytochrome c to react with molecular oxygen in the physiological pH range: "The heme of the cytochrome is . . . built into the protein component in a manifold way: by means OXIDASES, PEROXIDASES, AND CATALASE 99 of thioether bindings from the side chains of the porphyrin to the protein, and by means of two histidine-imidazole groups strongly bound to Fe on each side of the flat heme disc. Thus the heme group appears to be built into a crevice in the protein molecule. This ex- plains why cytochrome c is not autoxidizable, since oxygen can never approach the iron atom, and why no CO-compounds or cyanide compounds are formed at physiological pH values." Oxygen Transfer in Living Cells In every aerobically living cell we find a number of hemin pro- teins, e.g., a respiratory ferment, three diflFerent cytochromes, cata- lase and/or peroxidase. There is also frequently present what Keilin calls the "unspecific cell hematin." More recently another functional type of hemin enzymes has been found in such cells, which cata- CHAIN OF RESPIRATORY CATALYSTS Op- RESPIRATORY CYTOCHROMES ^CYTOCHROME ^PYRIDINE. ■ ENZYME * A-»C^B REDUCTASE ENZYMES SEQUENCE DURING PASTEUR REACTION PASTEUR ENZYME FERROUS IRON CATALYST FERI^ENTATION ENZYME SYSTEM SUBSTRATES STARCH GLYCOGEN GLUCOSE FRUCTOSE Figure 2. — Function of molecular oxygen in respiration and Pasteur reaction lyzes the inhibiting eflFect of oxygen on fermentation and glycolysis (Pasteur reaction). Some data pertaining to these various iron por- phyrin proteins are given in Tables 4 and 5. Two groups of these intracellular hemin proteins are endowed with the power to react directly with molecular oxygen: the oxygen- transferring enzymes of respiration (for short, "Warburg enzymes") and the aerobic fermentation-inhibiting catalysts (for short, "Pasteur enzymes"). Inasmuch as the Pasteur reaction is the subject of an- other paper, it must suffice here to present only the particular working hypothesis which the writer is advancing on the basis of recent work in this laboratory. See Figure 2. Photochemical work (44, 61) has revealed the pheohemin nature of the prosthetic group of the Pasteur enzymes in rat retina and in 100 A SYMPOSIUM ON RESPIRATORY ENZYMES baker's yeast. It is thus seen that catalysts of the same type, al- though differing in details, have been developed by nature to keep in check both alcoholic and lactic acid fermentation in the presence of oxygen. As Figure 2 indicates, oxygen is "mobilized" for its two chief tasks in aerobic cell life by two autoxidizable hemin proteins (Warburg and Pasteur enzymes) acting in analogous manner but independently of each other. The effect of a variation of the oxygen tension (and, for that matter, of carbon monoxide concentration in inhibitor experiments) on the overall phenomena of cellular respira- tion and aerobic fermentation will depend on the affinity which the iron contained in the two types of enzymes has for these gases. Just as is the case with the respiratory transport protein hemoglobin in various vertebrate species, there will be variations in the gas- dissociation curves for these enzymes from cell to cell. Depending on whether a particular Pasteur catalyst has a higher or a lower aflBnity for oxygen or carbon monoxide than the Warburg ferment in the same cell, we will expect to find respiration more or less readily afiFected by a lowering of oxygen tension or a given ratio of carbon monoxide to oxygen than the Pasteur reaction in that cell. If these gases do not vary in their effect on the two processes, as Warren (unpublished observations) found to be true in the case of bone marrow, it must be because the affinity of the enzymes for the gases is equal. In general it would appear that in higher animal tissues the Warburg ferment has a greater affinity for oxygen and a lesser affinity for carbon monoxide than the Pasteur enzyme in the same tissue (Laser effect, 39); in unicellular systems (certain bac- teria, human myeolocytes) the reverse seems to be true (Kempner effect, 29). It would appear that there are as many (slightly) different Warburg and Pasteur enzymes as there are living forms. Probably the difference resides in the protein rather than in the hemin group of the molecules, except in the case of Azotobacter, which seems to have a Warburg enzyme with a green hemin rather than a mixed- colored or pheohemin in the prosthetic group (47). The reasons for interposing a hypothetical ferrous iron catalyst between the Pasteur enzyme and the fermentation system, as indicated in the figure, cannot be explained here because of lack of space. In closing, some model experiments may be mentioned which are being carried out at present in this laboratory. Some years ago, Lip- mann (41) showed that a Pasteur effect in cell-free systems (yeast and muscle fermentation extracts) may be produced by the addition of suitable reversible dyestuffs of suflBciently high potential (see OXIDASES, PEROXIDASES, AND CATALASE 101 Figure 3). If our concept of the mechanism of the Pasteur reaction is correct, it should be possible to replace the dyestuffs in Lipmann's experiments by an autoxidizable hemochromogen of suitable poten- tial. Pyridine, histidine, and picoline hemochromogen have been tried without positive results. The experiments with the very posi- tive systems nicotine and nicotinic acid amide hemochromogen, on PASTEUR ENZYME FERMENTATION ENZYME SYSTEM SUBSTRATE 2.6-Dichlorophenol GLYCOLYZING O2 * fndophenol > MUSCLE * STARCH Eo=-0.l89(pH74) EXTRACT Naphthol-Sulfonafe- FERMENTING fCLUCOSE O2 — » mdophenol YEAST MACERATION — > < ♦Hexose Diphosphafe Eo=*Ol47v.(pH6.6) EXTRACT [sucrose Nicotine Fern- Hemochromogen Eo=>*0.200v.(pH6.0) FERMENTING YEAST MACERATION EXTRACT ["glucose <^fructose [sucrose Figure 3. — Attempts at reconstruction of Pasteur reaction in cell-free systems the other hand, have yielded some encouraging results which, un- fortunately, did not prove to be reproducible at will. One has the impression that the underlying idea is correct but that some of the factors involved in the complex reaction have thus far evaded adequate control. The problem of the Pasteur reaction, as the writer sees it, is now shifting from the nature of the catalyst to the nature of its substrate. REFERENCES 1. Agner, K., Act. Physiol. Scand., Vol. II. Suppl. 8 (1941). 2. Ahlstrom, L., and v. Euler, H., Z. physiol. Chem., 200, 233 (1931). 3. Altschul, a. M., Abrams, R., and Hogness, T. R., J. Biol. Chem., 136, 777 (1940); and unpublished observations. 4. Barron, E. S. C, Physiol. Rev., 19, 184 (1939). 5. Barron, E. S. C, and Lyman, C. M., J. Biol. Chem., 123, 229 (1937). 6. Barron, E. S. G., de Meio, R. H., and Klemperer, F., J. Biol. Chem., 112, 125 (1935). 7. Baudisch, O., and Davidsohn, D., J. Biol. Chem., 71, 501 (1927). 8. Bergel, F., and Bolz, K., Z. physiol. Chem., 215, 25 (1933). 9. Brann, L., Thesis, Ziirich, 1927. 10. Brooks, J., Proc. Roy. Soc. (London), B 109, 35 (1931); 118, 560 (1935). 102 A SYMPOSIUM ON RESPIRATORY ENZYMES 11. Califano, L., Naturwissenschaften, 22, 249 (1934). 12. Chance, B., J. Biol. Chem., 140, xxiv (1941). 13. V. EuLER, H., RuNEHjELM, D., and Steffenburg, Sv., Arkiv Kemi, Mineral. Geol, Abt. B, JO, 1 (1929). V. EuLER, H., NiLssoN, H., and Runehjelm, D., Sved. Kem. Tidskr., 41, 85 (1929). V. EuLER, H., and Jansson, B., Monatsh. Chem., 53-54, 1014 (1929). 14. Franck, J., and Haber, F., Sitz. Ber. Preuss. Akad. Wiss. Berlin, 1931, p. 250. 15. Franke, W., Liebigs Ann., 498, 129 (1932). 16. Haber, F., and Sachsse, H., Z. physik. Chem., Bodenstein Festschrift, 1931, p. 831. 17. Haber, F., and Weiss, J., Naturwissenschaften, 20, 948 (1932). 18. Haber, F., and Weiss, J., Proc. Roy. Soc. (London), A 147, 332 (1934). 19. Haber, F., and Willstaetter, R., Ber., 64, 2844 (1931). 20. Haldane, J. B. S., Proc. Roy. Soc. (London), B 108, 559 (1931). 21. Haurowitz, F., Z. physiol. Chem., 232, 159 (1935). 22. Haurowitz, F., Enzymologia, 4, 139 (1937). 23. Heubner, W., Naturwissenschaften, 16, 515 (1928). 24. Johnson, F. H., and van Schouwenburg, K. L., Nature, 144, 634 ( 1939). 25. Karush, F., J. Biol. Chem., 140, kvi (1941). 26. Keilin, D., and Hartree, E. F., Proc. Roy. Soc. (London), B 121, 173 (1936). 27. Keilin, D., and Hartree, E. F., Proc. Roy. Soc. (London), B 124, 397 (1938). 28. Keilin, D., and Mann, T., Proc. Roy. Soc. (London), B 122, 119 (1937). 29. Kempner, W., J. Cell. Comp. Physiol., JO, 339 (1937); Am. J. Tubercu- losis, 2, 157 (1939). Kempner, W., and Gaffron, M., Am. J. Physiol., 126, 553 (1939). 30. Krebs, H. a., Biochem. Z., 193, 347 (1928). 31. Kubowitz, F., and Haas, E., Biochem. Z., 255, 347 (1932). 32. Kuhn, R., Brann, L., Seyffert, C, and Furter, M., Ber., 60, 1151 ( 1927). 33. Kuhn, R., and Meyer, K., Naturwissenschaften, 16, 1028 (1928). Meyer, K., J. Biol. Chem., JOS, 25 (1933). 34. Kuhn, R., and Meyer, K., Z. physiol. Chem., J85, 193 (1929). 35. Kuhn, R., and Wassermann, A., Ber., 6J, 1550 (1928). 36. Kuhn, R., and Wassermann, A., Leibigs Ann., 503, 203 (1933). 37. Langenbeck, W., Die Organischen Katalysatoren (Berlin, 1935). 38. Langenbeck, W., Hutschenreuter, R., and Rottig, W., Ber., 65, 1750 ( 1932 ) 39. Laser, H., Biochem. J., 31, 1671, 1677 (1937). 40. Lewis, B., and van Elbe, C, Combustion, Flames and Explosions of Gases (Cambridge, 1938). 41. Lipmann, F., Biochem. Z., 265, 133 (1933); 268, 205 (1934). 42. Lyman, C. M., and Barron, E. S. G., J. Biol. Chem., J2J, 275 (1937). 43. Melnick, J. L., Science, 94, 118 (1941). 44. Melnick, J. L., J. Biol. Chem., 141, 269 (1941). 45. Moelwyn-Hughes, E. A., Ergebnisse d. Enzymforschung, 6, 23 (1937). 46. Negelein, E., Biochem. Z., 243, 386 (1941). 47. Negelein, E., and Gerischer, W., Biochem. Z., 268, 1 (1934). 48. Neh^l, J., and Hastings, A. B., J. Biol. Chem., 63, 479 (1925). 49. Oppenheimer, C, and Stern, K. G., Biological Oxidation (The Hague, 1939). 50. Reid, a., Ergebnisse d. Enzymforschung, J, 325 (1932). 51. Reid, A., Z. Angew. Chem., 47, 515 (1934). OXIDASES, PEROXIDASES, AND CATALASE 103 52. Reuter, F., Willstaedt, H., and Zirm, K. L., Biochem. Z., 261, 353 (1933). 53. Robinson, M. E., Biochem. J., 18, 255 (1934). 53a. Schwab, G. M., Rosenfeld, B., and Rudolph, L., Ber. Deutsch. Chem. Ges., 66, 661 (1933). 54. Smythe, C. v., J. Biol. Chem., 90, 251 (1931). 55. Stern, K. G., Z. physiol. Chem., 209, 176 (1932). 56. Stern, K. G., Z. physiol. Chem., 215, 35 (1933). 57. Stern, K. G., Z. physiol. Chem., 219, 105 (1933). 58. Stern, K. G., Nature, 136, 335 (1935). 59. Stern, K. G., J. Biol. Chem., 112, 661 (1936). 60. Stern, K. G., J. Biol. Chem., 114, 473 ( 1936); Enzymologia, 4, 145 ( 1937). 61. Stern, K. G., J. Gen. Physiol., 20, 631 (1937). 62. Stern, K. G., Yale J. Biol, and Med., 10, 161 ( 1937). 63. Stern, K. G., Cold Spring Harbor Symposia on Quantitative Biology, 7, 312 (1939). 64. Stern, K. G., and Gordon, W., unpublished observations. 65. Stern, K. G., and Kegeles, G., unpublished. 66. Stern, K. G., and Melnick, J. L., J. Biol. Chem., 139, 301 (1941). 67. Stern, K. G., and Wyckoff, R. W. G., J. Biol. Chem., 124, 573 (1938). 68. Sumner, J. B., Advances in Enzymology, J, 163 (1941). 69. Sumner, J. B., and Dounce, A. L., J. Biol. Chem., 121, 417 (1937). 70. Sumner, J. B., and Gralen, N., J. Biol. Chem., 125, 33 (1938). 71. SvEDBERG, Tn., J. Biol. Chem., 103, 311 (1933). 72. Theorell, H., Arkiv Kemi, Mineral. Geol., B 14, No. 20 (1941). 73. Theorell, H., J. Am. Chem. Soc, 63, 1820 ( 1941 ). 73a. Urey, H. C, Dawsey, L. H., and Rice, F. O., J. Am. Chem. Soc, 51, 1371 (1929). Taylor, H. S., and Gould, A., J. Am. Chem. Soc, 55, 859 (1933). 74. Warburg, O., Ergebnisse d. Enzymforschung, 7, 210 (1938). 75. Warburg, O., and Haas, E., Naturwissenschaften, 22, 207 (1934). 76. Warburg, O., and Negelein, E., Biochem. Z., 200, 414 (1928). 77. Warburg, O., and Negelein, E., Biochem. Z., 244, 9 (1932). 78. Weiss,- J., Naturwissenschaften, 23, 64 (1935). 78a. Weiss, J., Trans. Faraday Soc, 31, 688 (1935). 79. Weiss, J., J. Phys. Chem., 41, 1107 (1937). 80. Weiss, J., and Weil-Malherbe, H., Nature, 144, 866 (1939). 81. WiLLSTAETTER, R., and Pollinger, a., Z. physiol. Chem., 130, 281 (1923). 82. Zeile, K., Z. physiol. Chem., 189, 127 (1930). 83. Zeile, K., Z. physiol. Chem., 195, 39 (1931). 84. Zeile, K., Ergebnisse d. Physiol., 35, 498 (1933). 85. Zeile, K., in Oppenheimer Handbuch d. Biochemie, Ergaenzungswerk, 1, 708 (1933) (Jena). 86. Zeile, K., and Hellstroem, H., Z. physiol. Chem., 192, 171 (1930). Stern, K. G., J. Biol. Chem., 121, 561 (1937). 87. Zeile, K., Fawaz, G., and Ellis, V., Z. physiol. Chem., 263, 181 (1940). Nicotinamide Nucleotide Enzymes FRITZ SCHLENK School of Medicine, University of Texas REPORTS OF INVESTIGATIONS in the field of the nicotinamide nu- ^ cleotide enzymes are now so numerous as to make impossible a complete review of the subject here; and in any case it would seem to be unnecessary in view of the comprehensive articles that have been published in recent years (la-k). Emphasis will therefore be laid upon those details that have not been extensively discussed in previous articles and to report some recent advances. CODEHYDROGENASE I AND II Of the two coenzymes in this group, codehydrogenase I (cozymase, diphosphopyridine nucleotide, Co I), was detected in 1906 by Harden as a coenzyme of alcoholic fermentation (2). The other, codehydrogenase II (triphosphopyridine nucleotide, Co II), was dis- covered in 1932 by Warburg (3). In a series of investigations by the von Euler school (1921-34) methods for the purification and deter- mination of cozymase were elaborated. Its classification as a nu- cleotide was ascertained, its codehydrogenase nature was pro- pounded, and the numerous reports denying its existence were refuted (le, f). In 1934 the most noteworthy discovery in the field was made: Warburg and Christian isolated nicotinamide from co- dehydrogenase II (4) and demonstrated its function as part of a hydrogen-transporting coenzyme (5). In 1935 nicotinamide was isolated also from cozymase (6), and soon the very close relationship between these two coenzymes was established by the work of the institutes in Berlin and Stockholm. Both coenzymes are nicotinamide-adenine dinucleotides, the only difference between them being that codehydrogenase II contains three molecules, and codehydrogenase I two molecules, of phosphoric acid. Later the transformation of one coenzyme into the other was achieved by enzymatic dephosphorylation of codehydrogenase II and enzymatic as well as chemical phosphorylation of codehydro- genase I (7). The yield of codehydrogenase II by this reaction is, however, rather low; the isolation from red blood cells is still the 104 NICOTINAMIDE NUCLEOTIDE ENZYMES 105 only method that has been described for preparing it in the pure state. Codehydrogenase II occurs in much lower concentration than codehydrogenase I. In Table 1 are given some typical examples of the occurrence of the two coenzymes. The values are only ap- proximate because crude extracts were used for the determinations. With improved methods of determination it probably will be neces- sary to revise these values considerably, especially those for co- dehydrogenase II. Table 1.— Some examples of the occurrence of the codehydrogenases I and II Codehydrogenase content in micrograms _, . , . , per gram of fresh material Material exammed Co I Co II Bottom yeast >500 <10 Top yeast >500 5-10 Erythrocytes (horse) 100 >12 Liver (rat) >200 30 Muscle (rat) 200 50 Kidney (rat) 160 40 It is remarkable that despite the low content of codehydrogenase II in the source material and the difficulty and lengthiness of the isolation procedure, the preparation of codehydrogenase II in a pure state, the isolation of nicotinamide therefrom, and the demon- stration of its mode of action were carried out by Warburg and Christian in a remarkably short time. This work greatly facilitated the isolation of nicotinamide from cozymase and the subsequent work on this coenzyme in the Stockholm institute. In Tables 2 and 3 the methods of isolating the two compounds are shown schematically. The procedures are diflFerent in the two cases, but typical steps of purification known from earUer work in nucleo- tide chemistry are involved in both instances. Significant steps, some of them new, in the preparation of codehydrogenase II are the following: removal of proteins by acetone; fractionation of the crude mixture of nucleotides as barium salts, yielding adenosine polyphos- phate and coenzyme I as by-products; solution of the coenzyme in methanol-hydrochloric acid and reprecipitation by ethylacetate (5, 8). / '-4 AT-^^-i^ ^^\ C 106 A SYMPOSIUM ON RESPIRATORY ENZYMES The method that led to the isolation of codehydrogenase I in a pure state is based in the initial stage on the purification steps recom- mended by Myrback (le). A definite improvement is the precipita- tion of cozymase by cuprous chloride dissolved in a concentrated Table 2.— Preparation of codehydrogenase II from horse erythrocytes WASHED ERYTHROCYTES hemolysed by water; precipitated with acetone SOLUTION acetone removed by vacuum distillation; fractional precipitation by Hg(0Ac)2 PRECIPITATE (protein) (discarded) FRACTIONS 2 and 3 (contain Co II) +H2S (HgS i ) +acetone FRACTIONS 1 and 4 (discarded) PRECIPITATE (contains Co II; purity about 15%) dissolved in H2O; fractional precipitation by Ba(0H)2 -f alcohol (separation from Co I and cophosphorylase) -|-Hg(0Ac)2; precipitate +H2S; solution -{-acetone PRECIPITATE (contains Co II; purity about 30%) dissolved in HCI-CH3OH; precipitation by ethylacetate CODEHYDROGENASE II (purity about 50%) fractional precipitation by lead acetate-}- alcohol FRACTION 1 (about 60%) FRACTION 2 (about 60%) FRACTIONS 3 and 4 (about 100%) solution of potassium chloride (9). This procedure has been em- ployed in all preparative methods subsequently recommended (10-13). Since pure cozymase does not give a stable precipitate with the cuprous chloride reagent, such precipitation cannot be repeated. Apparently some impurity in the crude cozymase solutions plays an important role in producing a stable precipitate. Decomposition of the precipitate by hydrogen sulfide involves a considerable loss of Table 3— Preparation of codehydrogenase I from yeast YEAST Extracted with HjO at 80-100° C. |+Pb(Ac)2 SOLUTION +Ba(Ac)2+NaOH (pH 8) PRECIPITATE (discarded) SOLUTION +phosphotungstic acid PRECIPITATE (discarded) SOLUTION (discarded) PRECIPITATE suspended in dilute H2SO4 +amyl alcohol-ether H2O-H2SO4 PHASE +Ba(0H)2 (BaS04 i ) +AgN03+NH40H (pH 7.5) AMYL ALCOHOL-ETHER PHASE (discarded) PRECIPITATE |H2S(Ag2S i ) SOLUTION (discarded) SOLUTION l+CuCl in concentrated KCl+HCl SOLUTION (discarded) PRECIPITATE +H2S (CU2S i ) concentration in vacuo alcohol precipitation CODEHYDROGENASE I (60-80 per cent purity) IHjO, Ba(0H)2 (pH 8) SOLUTION +H2SO4 (BaS04 i ) +Pb(Ac)2 PRECIPITATE (discarded) PRECIPITATE (adenylic acid and impurities) SOLUTION fractional precipitation by alcohol FRACTION 1 FRACTION 2 ^^^F.SS:^ a IHoS (PbS i ) IH2S (PbS 1 ) H2S (PbS i ) Ualcohol l+alcohol +alcohol i i i T CODEHYDROGENASE I CODEHYDROGENASE I CODEHYDROGENASE I (80-90 per cent) (90-100 per cent) (90-100 per cent) 107 108 A SYMPOSIUM ON RESPIRATORY ENZYMES the coenzyme by adsorption on the copper sulfide. For desorption the sulfide precipitate is aerated until slight oxidation of copper sulfide is attained. This method of desorption, the details of which were described some years ago (9, 14), has also proved advantageous in preventing losses by adsorption on silver sulfide. The final puri- fication is brought about by fractional barium and lead precipitation. Both codehydrogenases prepared according to the methods given in Tables 2 and 3 contain traces of impurities which complicate their use in the spectrographic methods employed in dehydrogenase investigations. The great stability toward oxidizing agents (le) per- mits the destruction of impurities by treatment with bromine water Table 4.— Properties of codehydrogenases I and II Property Codehydrogenase I Ref. Codehydrogenase II Ref. Empirical formula C21H27O14N7P2 16 CaiHasOnNyPs 5 Molecular weight 663 743 Structural units 1 Mol. nicotinamide 1 Mol. adenine 6 1 Mol. nicotinamide 1 Mol. adenine 5 2 Mol. pentose 17 2 Mol. pentose 5 2 Mol. phosphoric acid 3 Mol. phosphoric acid Base equivalent 1 18 3-4 (?) 5 1Q Stability: J. £7 Oxidized form In 0.1 N HCl at 50% destroyed after 20 50% destroyed after 4 100° C. 8 min. 7.3 min. In 0.1 N NaOH 50% destroyed after 17 min. (20°) 20 50% destroyed after 12 min. (23°) 4 Reduced form In 0.1 N HCl at 20° C. activity disappears 20 activity disappears 5 immediately 10 immediately In 0.1 N NaOH at slight decrease in activ- 21 100° C. ity after 10 min. In 0.1 N NaOH at stable 21 stable 5 20° C. Absorption spectrum: Oxidized form Maximum at 260mjit cm^ E = 3.8X107 ^ |_Mol. li 5 8 10 E-3.5X10' ""' [_Mol._ 5 8 Reduced form fcm"" , r cm 2 Maximum at 260m/u E = 3.3X107 z-z-r Mol. 22 ^ = ^-^X^«' [mo1.J 8 340m/i E = 1.1X10^ ^ [_Mol. 22 E- 1.0X10' ^'f^ _Mol. 8 NICOTINAMIDE NUCLEOTIDE ENZYMES 109 (8). Filtration through a column of activated alumina has also been employed successfully (15). Recently some modifications in the preparation of cozymase as given in Table 3 have been described (10-13), The improvements consist mainly in omitting some of the purification steps, which can be done without complications if a good quality of yeast is used as the source material. B. J. Jandorf has introduced the adsorption of cozymase on charcoal in his method of preparation (13). S. Ochoa has developed a method in which muscle tissue is used as source material (11), and P. Ohlmeyer has described a method for the preparation of dihydrocozymase (10). From 10 kilograms of yeast one gram of almost pure, or 0.5 gram of pure codehydrogenase I is obtained. The yield of codehydrogenase II from 1000 liters of erythrocytes is about 2.5 grams of almost pure, or 1.0 gram of abso- lutely pure preparation. The coenzymes precipitated from aqueous solution by organic solvents are not crystalline. Table 4 shows the composition and properties of codehydrogenase I and II. The most important part of the work on the structure of the co- enzymes has been concerned with the nicotinamide moiety, its mode of action, and the linkage between nicotinamide and the rest of the molecule. This work was begun by Warburg (8) and continued by Karrer and his co-workers (23a-f). Warburg showed first that in codehydrogenase II the nicotinamide reacts with two atoms of hydrogen in the presence of substrate and apoenzyme, forming a dihydro compound. This compound can also be obtained by reduc- tion with hydrosulfite in a shghtly alkaline medium. The reduced coenzyme has an absorption maximum at 340 n\\i, whereas the maxi- mum at 260 miJ, has lost some of its intensity by the reduction (see Figure 1). By reoxidation the original state is restored. Catalytic hydrogenation yields an uptake of six hydrogen atoms by the oxidized coenzymes and of four hydrogen atoms by the biologically reduced coenzymes. Experiments with adenine and its derivatives showed that under the same conditions these compounds are very slowly reduced by catalytic hydrogenation, whereas free nicotinamide exhibits the same properties as the coenzymes upon catalytic reduction (5). It should be remembered that the catalytic reduction which leads to the hexahydro compounds is irreversible, and the products obtained are inactive as coenzymes. These ex- periments demonstrated that the place of the reversible (biological) reduction— i.e., the center of the coenzyme activity— is the nicotin- amide nucleus. / / / / / / / \ \ \ \ \ \ 5 rt r) t— 1 G O s fVJ -13 8 o C) o OJ 4 S \ ^._ 300 320 340 7V< 360 360 Figure 2. — Absorption spectnim of nicotinamide nucleoside Our knowledge of the structure of cozymase is based on the following findings: Besides the isolation of nicotinamide and adenine (6, 9), the isolation of pentosephosphoric acid was attained by acid hydrolysis (17). By means of the periodate method it was showni that the phosphoric acid is linked to the fifth carbon atom of the pentose molecules, as is indicated in formula 6. CONH ^ 2 Mol. FORMULA 6 N I HC II II N C C-NH, I 2 C — N OH OH OH OH III / OHC-C-C-C-CH 0-P=0 III 2 \ H H H OH NICOTINAMIDE NUCLEOTIDE ENZYMES 117 Depending on the experimental conditions, the product from alkaline hydrolysis was nicotinamide (28) or adenosine diphosphoric acid (29), the structure of which is well established by the work of Lohmann (30), Embden (31), and Levene (32). Alkaline hydrolysis in the cold gives, besides nicotinamide, a product of adenosine diphosphate plus pentose (formula 7). This degradation product of cozymase is inactive as coenzyme (33). N = C-NH I I HC C— N\ II II ^CH N — C — N FORMULA 7 CONH. ,^ OH OH H OH OH II 11/ -C— C — C — C — c-o-p-o-p=o I I I I I II \ H H H H H O OH OH OH I I Adeni ne — pentose — P-0-P-peniose -f- II II O O -N The enzymatic hydrolysis gave the two nucleosides: FORMULA 8 ^ ■CONH. .^ I I N H I X- C- I N ■CONH- ■N II ■C-NH. HC I HCOH I HCOH I HC ^C^ O HC- I HCOH I HCOH O HC I CH^OH CH^OH 118 A SYMPOSIUM ON RESPIRATORY ENZYMES In view of these results and the fact that cozymase is a monobasic acid composed of its structural units minus five molecules of water, the structural formula published in 1936 as a working hypothesis (18) (see formula 9) seems well established. It remains, however, to determine whether or not the pentose of the nicotinamide part is identical with cZ-ribose. FORMULA 9 N: H I ■C- CONH. .<^ HC- C = C- I I N N ^C^ •N II C I NH. HC- HCOH I O HCOH I HC O HCOH I HCOH I O HC O CH^O — P — O e — P-O-CH. I OH Concerning the structure of codehydrogenase II the following results may be mentioned: Warburg and his co-workers stated that the molecule consists of nicotinamide, adenine, two molecules of pentose, and three molecules of phosphoric acid. These compounds minus six molecules of water form the coenzyme. The experiments on the linkage of the nicotinamide have already been mentioned. In a very early stage of the work, Theorell, on the basis of cata- phoretic experiments, claimed that the coenzyme is a tetrabasic acid and that the amino group of the adenine is not free (19). Warburg, Christian, and Griese (5) later revised this claim; their titration ex- periments indicate that the coenzyme is a tribasic acid, and they showed that the amino group is free. The finding of Adler and Euler (7) that the enzymatic reaction Co I ^ Co II can take place narrowed considerably the possibilities respecting the structure of codehydrogenase IL The location of the third phosphoric acid group now seems to be the only question concerning the structure that has not been settled. The fact that cozymase contains adenosine diphos- phate as an essential constituent suggested (15) that codehydrogenase NICOTINAMIDE NUCLEOTIDE ENZYMES 119 II might contain adenosine triphosphate as a part of its molecule (formula 10, 1). FORMULA 10 O OH OH © III I Nicofin amide -pentose — P— O— P— O— P — pentose- adenine 11 II II 0 0 0 ®0 OH ® I I 2. Nicotinamide- pentose — P — O — P — pentose- adenine o o 'p^hI This hypothesis, however, was not substantiated by experimental results. Experiments carried out with rather limited amounts of pure coenzyme II showed that it contains no readily hydrolyzable phos- phate, and yields no alkaline degradation product active as a co- phosphorylase (34). Therefore it seems probable that the third phos- phoric acid group is linked to the adenylic acid part of the molecule, perhaps in much the same way as in yeast adenylic acid (formula 10, 2). This phosphoric acid group should block the coenzyme prop- erties of the adenylic acid part when liberated by alkaline hydrolysis. Further experimental work is needed to decide this question. The small amounts of adenosine-5' -phosphoric acid and its homo- logues which occur as contaminants of impure codehydrogenase preparations and which are formed from alkaline hydrolysis of cozymase but not from codehydrogenase II can be traced by a test FORMULA 11 GOGH GOOH GOOH 1 , I /OH _x I /OH GHOH > GH-0-P=0 7 G-0-P=0 I /OH < I ^oH <— N. OH GH 0-P=0 GHOH GH^ OH GOOH GOOH 2 C-P^o'~' 4- ADENYLIC \ p C-OH + ADENOSINE- II Vu ACID II TRIPHOSPHATE GH^^ GH^ ADENOSINE- TRIPHOSPHATE — > ADENYLIC ACID + 2 PHOSPHORIC ACID 120 A SYMPOSIUM ON RESPIRATORY ENZYMES based on the coenzyme properties of adenylic acid and its homo- logues (cophosphorylase) in the enzymatic spHtting of phospho- pyruvic acid (see formula 11). In the absence of a suitable acceptor the phosphoric acid appears in a free state in an amount propor- tional to the amount of cophosphorylase present in the system; by this method 10 to 100 micrograms of adenylic acid can be deter- mined (35). Methods of Determination Only the methods based on the coenzyme properties will be dis- cussed here. For determining the presence of cozymase Harden (2) originally used a press extract of yeast as a source of apoenzymes and a boiled yeast extract as a source of coenzyme. A great im- provement over this method was made when Euler and Myrback found that dried brewer's yeast loses its cozymase by repeated ex- traction with cold water (le, 36). The remaining product was found to contain all the apoenzymes, activators, and coenzymes necessary for fermentation except cozymase. Therefore it was called apo- zymase. The use of this preparation is still the simplest and perhaps the most accurate method of determination, its accuracy being Hm- ited only by the errors of the manometric measurements. Another test system recently elaborated by Jandorf , Klemperer, and Hastings uses, instead of the function of cozymase in fermentation, the co- enzyme properties of cozymase in glycolysis, according to the following scheme (37): Hexose diphosphate -^ phosphoglyceraldehyde + dihydroxy- acetone phosphate Phosphoglyceraldehyde + H3ASO4 -^ arsenophosphoglyceral- dehyde Arsenophosphoglyceraldehyde + DPN -> arsenophospho- glyceric acid + DPN-Hg Arsenophosphoglyceric acid -^ 3-phosphoglyceric acid + H3ASO, Phosphoglyceraldehyde + DPN-Ho -> glycerophosphoric acid + DPN 6) Hexose diphosphate -^ 3-phosphoglyceric acid + glycero- phosphoric acid The amount of phosphoglyceric acid produced in a given time in the presence of bicarbonate buffer can be measured manometrically with the Warburg apparatus. The test system for codehydrogenase II as given by Warburg uses the dehydrogenation of Robison ester NICOTINAMIDE NUCLEOTIDE ENZYMES 121 by codehydrogenase II, Robison ester apodehydrogenase (Zwischen- ferment), and riboflavin enzyme. If the codehydrogenase II is the speed-hmiting factor in this system, the oxygen uptake is a measure of its concentration (5). It is probable that the old yellow enzyme used by Warburg and Christian is an artifact. The cytochrome c reductase of Haas, Horecker, and Hogness (38) seems to be the natural acceptor for the hydrogen of the reduced codehydrogen- ase II. These methods of determination have been used not only to elaborate the purification of the codehydrogenases but also to study their distribution from the standpoint of vitamin research. The codehydrogenases occur in an equilibrium: coenzyme ^ dihydroco- enzyme. Since the stabilities of the oxidized and the reduced forms are different (see Table 4), the proportions of coenzyme and dihy- drocoenzyme can be determined by heat extraction with acid or with alkali, according to Adler and Calvett (39). If the extraction is made with boiling water, the reduced coenzyme is oxidized by air. Therefore no loss occurs if some of the subsequent steps of the preparation are carried out in acid medium. In most tissues somewhat more of the oxidized than of the reduced form of cozymase was found, but in the Jensen rat sarcoma, Euler and his co-workers found a large excess of dihydrocozymase (20, 40). Recent experiments confirm this finding, but apparently it is not true for all cancerous tissues. An excess of dihydrocozymase was repeatedly found in methyl-cholanthrene rat tumor, but not regularly in benzpyrene tumors of mice and in Brown-Pearce rabbit carci- noma (33). The data on human carcinomata are still too limited to permit any conclusion. It must be remembered that all methods of determination in tis- sues involve complications: incompleteness of extraction, limited stability of the codehydrogenases in oxidized and reduced form at high temperature, rapid changes in the equilibrium Co I <=s Co II, enzymatic destruction upon disruption of the cells, and finally errors in the methods of determination. The claims for accuracy which are made by many publications dealing with coenzyme con- tent of tissues seem far too optimistic. The growth-promoting properties of the codehydrogenases for certain microorganisms (Hemophilus influenzae and Hemophilus para-influenzae) can be used for detennining small quantities of these compounds (41). Whereas the number of nicotinamide- requiring microorganisms is rather high, only H. influenzae and 122 A SYMPOSIUM ON RESPIRATORY ENZYMES H. para- influenzae have been found to require cozymase or code- hydrogenase II. According to LwoflF only 0.004 microgram per milli- liter of peptone solution is necessary to produce visible growth under standard conditions. The method does not distinguish between co- dehydrogenases I and II. Dihydrocozymase was found to be in- ferior as a nutrilite to an equivalent amount of cozymase (42). This may be due to a difference in the permeability of the cells to the two compounds. For the examination of pure or almost pure coenzyme prepara- tions the spectrophotometric determination of the dihydro com- pounds is a very accurate method. The solutions must, however, be free from impurities that absorb in the ultraviolet region in which the dihydrocoenzymes exhibit their characteristic absorption. The chemical methods of determination cannot be discussed here in detail. The Apodehydrogenases Since the history of the apodehydrogenases has been reviewed in several comprehensive articles (la, c, g, 43a-d), they will simply be listed here, and our present knowledge about their purification, na- ture, and function will be summarized and their coenzyme speci- ficity discussed. As can be seen from Table 7, some of the apodehydrogenases have been prepared in a pure state and obtained in a crystalline form. The main progress in the field was made by Warburg and his co- workers, especially Christian, Negelein, Gerischer, Haas, Wulff, and Kubowitz. Warburg's methods for purification consist mainly in precipitations of the apoenzymes at the isoelectric point, their fractional precipitation by ammonium sulfate or organic solvents, and removal and inactivation of other enzymes by heat denatura- tion at a temperature of about 50° C. Of special interest is the pre- cipitation of the diphosphoglyceraldehyde apodehydrogenase in a step of its purification by addition of nucleic acid (44), The isolation of the apodehydrogenases in a pure state was im- portant for a detailed study of their relation to coenzymes and sub- strate by the spectrophotometric technique. Whereas the use of crude apoenzymes does not exclude the possibility of more com- plex reactions, a definite conclusion can be drawn respecting the mechanism of a reaction if pure apoenzyme, coenzyme, and sub- strate are used. Our previous conception of the function of cozymase in the dehydrogenation of triosephosphate was greatly revised in Table 7.— The apodehydrogenases dependent on nicotinamide nucleotides Substrate and dehydrogenation Co- product enzyme Note on the apodehydrogenase Ref. l,3-diphosphoglyceraldehydei=^ 1,3-diphosphogly eerie aeid Triosephosphatei^phospho- glyceric acid Lactic acid<=^pyruvic acid Alcohol<=^acetaIdehyde Methyll Propyl plcohoI<=^ Amyl J Form- ) Propion- [aldehyde Valer- J 2 R • CHO+HjO^R COOH +RCn20H Malic acid<=^oxalacetic acid a-glycerophosphate<=^phospho- glyceraldehyde /3-hydroxybutyric acid^aceto- acetic acid Formic acid<=^C02 Glucose-6-monophosphate (Robison ester)— >phospho- hexonic acid Phosphohexonic acid Decarboxylation and dehydro- genation Isocitric acid<^a-keto-/3-carboxy- glutaric acid Glucose^gluconic acid Glutamic acid it Iminoglutaric acid Co I Obtained by Warburg and Christian 4 1 in crystalline state from yeast Co I Crude preparations from animal tis- 70 sues. Identical with 1,3-diphospho- 71 glyceraldehyde apodehydrogenase? Co I Prepared by Straub in a crystalline 72 form from heart Co I Obtained by Negelein and Wulff in 51 crystalline form from yeast Co I Crude preparations from animal tis- 46 sues. Identical with ethyl alcohol apodehydrogenase ? Co I Aldehyde mutase; crude prepara- 73 tions from liver Co I Crude preparations from animal tis- 74 sues Co I Crude preparations from yeast and 75 tissues Co I Crude preparations from animal tis- 76 sues Co I Crude preparations from seeds and 49 B. coli Co II Obtained from yeast in a highly 77 purified state by Negelein and Gerischer Co II Purified preparation obtained from 78 yeast by Warburg and Christian Co II Crude preparations from seeds, ani- 79 mal tissues, and yeast Co I or Crude preparations from liver and 45 Co II yeast Co II For apoenzyme from yeast and B. 80 coli; Co I or For apoenzyme from animal tissues; Coll Co I For apoenzyme from plants 123 124 A SYMPOSIUM ON RESPIRATORY ENZYMES this way by Warburg and Christian. They showed that diphospho- glyceraldehyde rather than triosephosphate is the substrate of cozymase in fermentation and perhaps in glycolysis (44). The main purpose of giving, in Table 7, the degree of purity of the apoenzymes obtained thus far is to afford some idea of the reliability of our knowledge respecting the reactions in question. The most noteworthy fact regarding the apoenzymes is that they are responsible for the specificity of the dehydrogenase, whereas codehydrogenases I and II combined with different proteins form a relatively large number of dehydrogenases. The protein moiety is therefore more specific than the prosthetic group. It was found that both codehydrogenase I and codehydrogenase II can act as coenzymes for glucose apodehydrogenase (45). This apoenzyme, however, has not yet been sufficiently purified, and therefore it is not impossible that a transformation of codehydro- genase I into codehydrogenase II or vice versa may occur in this system. In connection with alcohol apodehydrogenase it should be men- tioned that methyl, propyl, and amyl alcohol can also serve as sub- strate (46). This seems to indicate that the specificity of the apodehydrogenases even toward the substrate is not always an absolute one. Warburg has claimed that the protein part of a given dehydro- genase probably differs for each type of cells. Robison ester apode- hydrogenase from yeast and from rat blood were found to be different, their isoelectric points being at pH 4.8 and 5.8 respectively (lb). In some cases the difference is so great as to suggest that the specificity for the codehydrogenases depends on the source of the apoenzyme. Thus glutamic acid apodehydrogenase from liver was found by Adler to require codehydrogenase II; the apoenzyme from plants uses codehydrogenase I as a prosthetic group (47). Besides the cozymase-dependent glycerophosphate apodehydro- genase, a glycerophosphate dehydrogenase was found to occur in muscle tissue which does not require a coenzyme (48). Similar findings have been reported for formic dehydrogenase. Whereas the dehydrogenase preparations from seeds consist of cozymase plus apoenzyme (49), the corresponding enzyme from Bacterium coli does not need cozymase (50). If there are actually as many variations of apodehydrogenases as there are different types of cells, a further important specificity may be based on this difference. To obtain more evidence, the apode- NICOTINAMIDE NUCLEOTIDE ENZYMES 125 hydrogenases must be prepared from different sources and the re- sulting products compared with one another. It is probable that the differences will consist in a slight variation in the pH for optimum activity or a change in the equilibrium between protein, substrate, and coenzyme rather than in a completely different mode of action. The preparation of some apodehydrogenases in crystalline form has permitted an estimation of the amount present in the cell. They are found in much smaller concentration than the codehydrogenases. O. Meyerhof (Ic) has pointed out that if each cozymase molecule in rabbit muscle or yeast were accompanied by one molecule of each of the apodehydrogenases, the protein of the cells would con- sist exclusively of the apodehydrogenases. The apodehydrogenases combine with the nicotinamide nucleo- tides, the reduced nucleotides, the substrate, and the reaction product. The extent to which the complexes thus formed are dis- sociated detennines the direction and the rate and equilibrium of the reaction. As an example may be mentioned the following system: alcohol, acetaldehyde, cozymase, dihydrocozymase, and apodehy- drogenase. This system was investigated by Negelein and Wulff, who used the pure components (51). They found that under their experimental conditions the concentration of each substance at which the apodehydrogenase is half saturated with it is as follows: cozymase, 0.0001 M.; dihydrocozymase, 0.00003 M.; acetaldehyde, 0.0001 M.; and ethyl alcohol, 0.024 M. The reduced coenzyme and acetaldehyde are bound to a much greater extent than cozymase and alcohol, and under normal conditions the reversible reaction acetal- dehyde + dihydrocozymase ^ alcohol + cozymase proceeds in the direction of alcohol formation. Experiments carried out on other pyridine dehydrogenase systems have yielded similar results. The dehydrogenases requiring codehydrogenase II, however, seem to be less dissociated. The high degree of dissociation between the pyridine nucleotides and their protein parts corresponds to that of the cophosphorylases and their specific proteins. M. Dixon and L. G. Zerfas have pointed out the differences between this group of enzymes and the other type, in which the prosthetic group is in a relatively stable linkage to the protein and the ratio between prosthetic group and protein is 1:1 (52). On the basis of interesting experimental results with artificial hydrogen acceptors they state in their discussion that the use of the term "pyridine-proteid" is misleading and should be dis- continued. It would seem, however, that this suggestion is too radi- 126 A SYMPOSIUM ON RESPIRATORY ENZYMES cal. The work that has been done in this field leaves no doubt that we are dealing with real enzymes, which, after bringing about the dehydrogenation of substrate under biological conditions, are un- changed. Furthermore, at the moment when they exhibit their activity, the nicotinamide nucleotide is combined with its specific protein just as the riboflavin nucleotide is combined with its specific protein. That we deal in the one case with a relatively stable linkage and in the other with a very unstable one is only a difference in degree. Hence a change of the whole nomenclature would seem to be unnecessary and would probably increase the confusion in this field. The physiological reoxidation of the reduced coenzymes is brought about by the oxidized form of the metabolites listed in Table 7, as, for example, acetaldehyde in fermentation or pyruvic acid in glycolysis. In other instances the dihydrocoenzymes are re- oxidized by alloxazine proteids such as diaphorase (84), coenzyme factor (85), cytochrome c reductase (38), or the enzyme recently detected by Altschul, Persky, and Hogness (86). Spectrophometric Methods In studying the function of the nicotinamide nucleotide enzymes we have been greatly aided by the spectrophometric methods de- veloped by Warburg. The basis is the appearance of an absorption band at 340 m\}. when the codehydrogenases are reduced. This absorption band disappears upon reoxidation. Under proper ex- perimental conditions the concentration of coenzyme or apoenzyme, the substrate and acceptor specificity of the dehydrogenases, and the speed of reactions can be studied, as has been done in some laboratories, notably those of Warburg, Euler, and Meyerhof (lb, i; 53). As an example the action of the reducing and the oxidizing fermentation system may be given (44). See Figure 3. Biosynthesis of the Codehydrogenases In view of the relatively complicated configuration of the code- hydrogenases and of their nucleotide character, it is improbable that a satisfactory synthesis will be obtained by present-day methods of organic chemistry. Therefore experiments designed to carry out a biosynthesis from the structural units are very important from a practical point of view. Mention may be made first of the experiments dealing with the NICOTINAMIDE NUCLEOTIDE ENZYMES 127 formation of the codehydrogenases from nicotinamide in the or- ganism. The biological function of the nicotinamide-containing coenzymes and their wide distribution in nature (If) were known before their vitamin properties. There was little doubt that the nicotinamide ingested as a vitamin is used to form the coenzymes. Nevertheless an exact demonstration had to be obtained. k^ 1.6 1.2 - 0.8 0.4 Protein II added Mi N. Figure 3. — Action of oxidizing and reducing fermentation enzymes. Spectro- photometric experiment, absorption at 340 m\i ( dihydrocozymase ) . d = 0.557 cm. Protein I: diphosphoglyceraldehyde apodehydrogenase. Protein II: acetal- dehyde reductase (alcohol apodehydrogenase). The concentrations are as follows: cozymase, 0.183 mg. per ml.; (i-phosphoglyceraldehyde, 0.733 mg. per ml.; acetaldehyde, 1.47 mg. per ml.; apoenzymes (protein I and II), 0.0008 mg. per ml.; orthophosphate, 3.3 X 10 '^ mole per ml.; pyrophosphate, 3.3 X 10"' mole per ml. Experiments on animal tissues were performed by Axelrod and Elvehjem. The cozymase level of pigs and dogs living on a diet low in nicotinamide was found to be decreased in muscle and liver. Upon administration of nicotinamide the normal cozymase content was rapidly restored (56). Later the same effect was observed in pellagrins by Axelrod, Spies, and Elvehjem (58). Experiments on rats also seemed to indicate the same result (57), but according 128 A SYMPOSIUM ON RESPIRATORY ENZYMES to more recent findings the vitamin nature of nicotinamide for the rat is doubtful (81, 82). Dann and Handler recently have shown that nicotinic acid is formed by the chick embryo (83). An important observation made in these experiments was that the organism does not synthesize coenzyme beyond the normal level of coenzyme content under favorable dietary conditions. Even an ad- ministration of nicotinamide far in excess of the normal requirement results in no significant synthesis beyond the normal level. In erythrocytes, however, an increase in coenzyme content upon ad- ministration of an excess of nicotinamide was observed by Kohn and Klein, Vilter and Spies, Axelrod and others (59, 60). Numerous experiments with isolated enzyme preparations from yeast, liver, and muscle have thus far failed to give a noteworthy synthesis of the coenzymes from the structural units. The destruc- tive tendency of these preparations has always been found to be remarkable. It was observed by Euler and co-workers (64—66) that in tissues the coenzyme content after death, and especially upon destruction of the cell structure, decreases rapidly. The significance of this finding for the methods of determination and the precautions necessary have been pointed out repeatedly. Mann and Quastel have recently confirmed the findings of Euler. Of special interest is their finding that free nicotinamide in great excess prevents the postmortem decomposition of cozymase (67), a result which the authors explain by assuming that nicotinamide and cozymase compete for the active center of the nucleotidase which destroys the coenzyme. Lennerstrand found some years ago (68, 69) that a destruction of cozymase by washed dried yeast (apozymase) takes place, but is inhibited by phosphate and hexosediphosphate. Apparently the substrate protects the coenzyme from destruction. It is not yet pos- sible to say what products are formed by the inactivation of cozy- mase by apozymase. If glucose and phosphate are added to the inactivated cozymase plus apozymase, a resynthesis takes place after several hours, and as much as 50 per cent of the cozymase originally present is restored. A similar effect was observed recently with cocarboxylase. When cocarboxylase is incubated with aetiozymase, it is destroyed. In the presence of pyruvate a much slower inactivation takes place (33). It is possible that underlying these findings is an important principle of regulation of the coenzyme level. It seemed possible that a regeneration of cozymase from the nico- NICOTINAMIDE NUCLEOTIDE ENZYMES 129 tinamide nucleoside might take place in a fermentation system, similar to the formation of cocarboxylase from thiamine. Experi- ments designed to carry out such a biosynthesis of cozymase by incubation with yeast preparations in the presence of phosphate and adenosine phosphoric acids have not yet been successful. Never- theless it can be assumed that the nicotinamide nucleoside is an intermediate in the course of biosynthesis of the codehydrogenases. Among the lower organisms of the plant kingdom— for example, Bacterium aerogenes, Torula, and other yeasts— we have examples of biosynthesis of cozymase from very simple nitrogen and carbon sources. Hutchens, Jandorf, and Hastings have shown recently that the protozoon Chilomonas paramecium is capable of synthesizing cozymase in a synthetic medium containing ammonia as the only source of nitrogen and acetate as the sole source of carbon (62). Other microorganisms, such as Staphylococcus aureus (54), Proteus vulgaris (55), and Shigella paradysenteriae (63), require nicotinamide to complete the synthesis of cozymase, a fact which permits these organisms to be used for the bio-assay of nicotinamide. It has been suggested that these bacteria require nicotinamide for growth as a result of their parasitic existence (55). Most striking in this connec- tion is the finding of Lwoff and Lwoff that Hemophilus influenzae and para-influenzae require codehydrogenase I and II (41). This specificity surpasses even that of man and other mammals. Gingrich (42) has shown that the "V" requirement of hemophihc bacteria can be satisfied not only by the oxidized codehydrogenases but also by dihydrocozymase, acid-treated dihydrocozymase (com- pletely inactive as codehydrogenase), and desamino cozymase, an artificial derivative of cozymase (61) in which the adenylic acid is replaced by inosinic acid. Furthermore, the fact that the nicotinam- ide nucleoside has been found (42) to promote the growth of these organisms is highly interesting in that it demonstrates that the only special requirement is the preformed linkage of nicotinamide to the pentose. It is obvious, then, that the nutrihtes required for the biosynthesis of cozymase are related chiefly to the nicotinamide moiety and its linkage to the rest of the molecule. The steps in biosynthesis between simple nitrogen and carbon compounds and the pyridine ring remain a promising field of investigation. REFERENCES 1. a) Thunberg, T., Ergebnisse d. Enzymforschung ( Nord-Weidenhagen ) , 7, 163 (1938). 130 A SYMPOSIUM ON RESPIRATORY ENZYMES b) Warburg, O., Ergebnisse d. Enzymforschung, 7, 210 (1938). c) Meyerhof, O., Ergebnisse d. Physiologic ( Asher-Spiro), 39, 10 (1937). d) V. EuLER, H., Ergebnisse d. Physiologie, 38, 1 (1936). e) Myrback, K., Ergebnisse d. Enzymforschung, 2, 139 (1933). f) Myrback, K., Tabulae Biologicae, 14, 110 (1937). g) Ball, E. G., Bull. Johns Hopkins Hosp., 65, 253 ( 1939). h} ScHLENK, F., and v. Euler, H., Fortschritte d. Chemie org. Naturstoffe (Zechmeister), 1, 99 (1938). i) ScHLENK, F., and Gunther, G., Methodik d. Fermente (Bamann- Myrback) part 7 (1940). k) Baumann, C. a., and Stare, F. J., Physiol. Rev., 19, 353 ( 1939). 2. Harden, A., and Young, W. J., Proc. Roy. Soc. (London) B 77, 405 (1906). Harden, A., Alcoholic Fermentation, 4th ed. (London, 1932). 3. Warburg, O., and Christlvn, W., Biochem. Z., 242, 206 (1931); 254, 438 (1932). 4. Warburg, O., and Christian, W., Biochem. Z., 274, 112 (1934); 275, 464 (1935). 5. Warburg, O., Christian, W., and Griese, W., Biochem. Z., 282, 157 (1935). 6. V. Euler, H., Albers, H., and Schlenk, F., Z. physiol. Chem., 237, I (1935). 7. Adler, E., Elliot, S., and Elliot, L., Enzymologia, 8, 80 (1940). 8. Warburg, O., and Christian, W., Biochem. Z., 287, 291 (1936). 9. V. Euler, H., Albers, H., and Schlenk, F., Z. physiol. Chem., 240, 113 (1936). 10. Ohlmeyer, p., Biochem. Z., 297, 66 (1938). 11. OcHOA, S., Biochem. Z., 292, 68 ( 1937). 12. Williamson, S., and Green, D. E., J. Biol. Chem., 135, 345 ( 1940). 13. Jandorf, B. J., J. Biol. Chem., 138, 305 (1941). 14. Schlenk, F., Svenska Vet. Akad. Arkiv f. Kemi, 12 A, 21 ( 1937). 15. V. Euler, H., and Schlenk, F., Z. physiol. Chem., 246, 64 ( 1937). 16. V. Euler, H., and Schlenk, F., Svensk Kem. Tidskr., 48, 135 (1936). 17. Schlenk, F., Svenska Vet. Akad. Arkiv f. Kemi, 12 B, 20 ( 1936). 18. Schlenk, F., and v. Euler, H., Naturwissenschaften, 24, 794 (1936). 19. Theorell, H., Biochem. Z., 275, 19 (1934). 20. v. Euler, H., Schlenk, F., Heiwinkel, H., and Hogberg, B., Z. physiol. Chem., 256, 208 (1938). 21. Abler, E., Hellstrom, H., and v. Euler, H., Z. physiol. Chem., 242, 225 (1936). 22. v. Euler, H., Adler, E., and Hellstrom, H., Z. physiol. Chem., 241, 239 (1936). 23. a) Karrer, P., Schwarzenbach, G., Benz, F., and Solmssen, U., Ilelv. Chim. Acta, 19, 811, 1028 (1936). b) Karrer, P., Ringier, B. H., BiJCHi, J., Fritzsche, H., and Solmssen, U., Helv. Chim. Acta, 20, 55 (1937). c) Karrer, P., Schwarzenbach, G., and Utzinger, G. E., Helv. Chim. Acta, 20, 720 (1937). d) Karrer, P., and Stare, F. J., Helv. Chim. Acta, 20, 418 (1937). e) Karrer, P., Kahnt, F. W., Epstein, R., Jaffe, W., and Ishh, T., Helv. Chim. Acta, 21, 223 (1938). f ) Karrer, P., Ishii, T., Kahnt, F. W., and van Bergen, I., Helv. Chim. Acta, 21, 1174 (1938). 24. Myrback, K., and Ortenblad, B., Z. physiol. Chem., 233, 87 (1935). NICOTINAMIDE NUCLEOTIDE ENZYMES 131 Karrer, p., Schlenk, F., and v. Euler, H., Svenska Vet. Akad. Arkiv f. Kemi, 12 B, 26 (1936). 25. Haas, E., Biochem. Z., 285, 368 (1936). 26. Schlenk, F., Naturwissenschaften, 28, 46 (1940). 27. Bredereck, H., Beuchelt, H., and Richter, C, Z. physiol. Chem., 244, 102 (1936); Ber. Chem. Ges., 71, 408 (1938). 28. Schlenk, F., v. Euler, H., Heiwinkel, H., Gleim, W., and Nystrom, H., Z. physiol. Chem., 247, 23 (1937). 29. Vestin, R., Schlenk, F., and v. Euler, H., Ber. Chem. Ges., 70, 1369 (1937). 30. Lohmann, K., Biochem. Z., 282, 120 (1935). 31. Embden, G., and Zimmermann, M., Z. physiol. Chem., 167, 114, 137 (1927). 32. Levene, p. a., and Bass, L. W., Nucleic Acids (New York, 1931), pp. 187-192. 33. Schlenk, F., unpublished experiments. 34. Schlenk, F., Hogberg, B., and Tingstam, S., Svenska Vet. Akad. Arkiv f. Kemi, 13 A, 11 (1939). 35. Schlenk, F., and Schlenk, T., J. Biol. Chem., 141, 311 (1941). 36. Myrback, K., Z. physiol. Chem. 177, 158 (1928). Schlenk, F., and Vowles, R. B., Svenska Vet. Akad. Arkiv f. Kemi, 13 B, 19 (1940). 37. Jandorf, B. J., Klemperer, F. W., and Hastings, A. B., J. Biol. Chem., 138, 311 (1941). 38. Haas, E., Horecker, B. L., and Hogness, T. R., J. Biol. Chem., 136, 747 (1940). 39. Adler, E., and Calvett, F., Svenska Vet. Akad. Arkiv f. Kemi, 12 B, 32 (1936). 40. V. Euler, H., Malmberg, M., and Gunther, G., Z. f. Krebsforschung, 45, 425 (1937). 41. LwoFF, A., and Lwoff, M., Comptes rendus Acad. Sci., 203, 896 (1936); Proc. Roy. Soc. (London) B 122, 352, 360 (1937). 42. Gingrich, W. D., and Schlenk, F., unpublished experiments. 43. a) Oppenheimer, C, and Stern, K. G., Biological Oxidation (The Hague, 1939). b) Thunberg, T., Ergebnisse d. Physiologic, 39, 76 ( 1937). c) Franke, W., in Euler, Chemie d. Enzyme, 2, 3 ( 1934). d) Potter, V. R., Medicine, 19, 441 (1940). 44. Warburg, O., and Christian, W., Biochem. Z., 303, 40 ( 1939). 45. Das, N. B., Z. physiol. Chem., 238, 269 (1936). QuiBELL, T. H., Z. physiol. Chem., 251, 102 (1938). 46. Lutwak-Mann, C, Biochem. J., 32, 1364 (1938). 47. V. Euler, H., Adler, E., Gltnther, G., and Das, N. B., Z. physiol. Chem., 254, 61 (1938). 48. Green, D. E., Biochem. J., 30, 629 (1936). 49. Adler, E., and Sreenivasaya, M., Z. physiol. Chem., 249, 24 (1937). 50. Gale, E. F., Biochem. J., 33, 1012 (1939). 51. Negelein, E., and Wulff, H. J., Biochem. Z., 293, 351 (1937). 52. Dixon, M., and Zerfas, L. G., Biochem. J., 34, 371 ( 1940). 53. Meyerhof, O., Ohlmeyer, P., and Mohle, W., Biochem. Z. 297, 90 (1938). 54. Knight, B. C. J. G., Biochem. J., 31, 731, 966 (1937). 55. Fildes, p., Proc. Roy. Soc. (London), B 124, 4 (1937); Brit. J. of Exp. Pathol., 19, 239 (1938). 132 A SYMPOSIUM ON RESPIRATORY ENZYMES 56. AxELROD, A. E., Madden, R. J., and Elvehjem, C. A., J. Biol. Chem., 131, 85 (1939). Dann, W. J., and Handler, P., J. Nutrition, 22, 409 (1941). 57. V. Euler, H., Schlenk, P., Melzer, L., and Hogberg, B., Z. physiol. Chem., 258, 212 (1939). Frost, D. V., and Elvehjem, C. A., J. Biol. Chem., 121, 255 ( 1937). 58. AxELROD, A. E., Spies, T. D., and Elvehjem, C. A., J. Biol. Chem., 138, 667 (1941). 59. Kohn, H. I., and Klein, J. R., J. Biol. Chem., 130, 1 ( 1939); 135, 685 (1940). 60. ViLTER, R. W., ViLTER, S. P., and Spies, T. D., J. Am. Med. Assoc, 112, 420 (1939); Vilter, S. P., Koch, M. B., and Spies, T. D., J. Lab. Clin. Med., 26, 31 (1940). AxELROD, A. E., Gordon, E. S., and Elvehjem, C. A., Am. J. Med. Sci., 199, 697 (1940). 61. Schlenk, F., Hellstrom, H., and v. Euler, H., Ber. Chem. Ges., 71, 1471 (1938). 62. HuTCHENs, J. O., Jandorf, B. J., and Hastings, A. B., J. Biol. Chem., 138, 321 (1941). 63. a) DoRFMAN, A., Koser, S. A., Reames, S. A., Swingle, R. H., and Saun- ders, F., J. Inf. Diseases, 65, 163 ( 1939). b) DoRFMAN, A., Koser, S. A., and Saunders, F., Proc. Soc. Exp. Biol. Med., 43, 434 (1940). c) Saunders, F., Dorfman, A., and Koser, S. A., J. Biol. Chem., 138, 69 (1941). d) Bass, A., Berkman, S., Saunders, F., and Koser, S. A., J. Inf. Diseases, 68, 175 (1941). 64. v. Euler, H., Myrback, K., and Brunius, E., Z. physiol. Chem., 177, 237; 183, 60 (1929). 65. V. Euler, H., and Gunther, G., Z. physiol. Chem., 243, 1 (1936). 66. V. Euler, H., HEivvntNKEL, H., and Schlenk, F., Z. physiol. Chem., 247, IV (1937). 67. Mann, P. J. G., and Quastel, J. H., Biochem. J., 35, 502 (1941). 68. Lennerstrand, A., Biochem. Z., 287, 172 (1936). 69. Lennerstrand, A., Svenska Vet. Akad. Arkiv f. Kemi, 14 B, 1 (1940); 14 A, 16 (1941). 70. Adler, E., and Gunther, G., Z. physiol. Chem., 253, 143 (1938). 71. Meyerhof, O., Kiessling, W., and Schulz, V., Biochem. Z., 292, 25 (1937). 72. Straub, F. B., Biochem. J., 34, 483 (1940). 73. Dixon, M., and Lutwak-Mann, C, Biochem. J., 31, 1347 (1937). Adler, E., v. Euler, H., and GIjnther, G., Svenska Vet. Akad. Arkiv f. Kemi, 12 B, 54 (1938). 74. v. Euler, H., Adler, E., and Gunther, G., Z. physiol. Chem., 249, 1 (1937). Green, D. E., Biochem. J., 30, 2095 (1936). 75. Adler, E., v. Euler, H., and Hughes, W., Z. physiol. Chem., 252, 1 (1938). 76. Green, D. E., Dewan, I. G., and Leloir, L. F., Biochem. J., 31, 934 (1937). 77. Negelein, E., and Gebischer, W., Biochem. Z., 284, 289 (1936). 78. Warburg, O., and Christian, W., Biochem. Z., 292, 287 (1937). 79. Adler, E., v. Euler, H., Gunther, G., and Plass, M., Biochem. J., 33, 1028 (1939). NICOTINAMIDE NUCLEOTIDE ENZYMES 133 80. V. EuLER, H., Adler, E., Gunther, C, and Elliot, L., Enzymologia, 6, 337 (1939). V. EtTLER, H., and Adler, E., Enzymologia, 7, 21 (1939). GiJNTHER, C, Svenska Vet. Akad. Arkiv f. Kemi, 12 A, 23 (1938). 81. Dann, W. J., and Kohn, H. I., J. Biol. Chem., 136, 435 (1940). Dann, W. J., J. Biol. Chem., 141, 803 (1941). 82. Elvehjem, C. a.. Biological Action of the Vitamins. Lecture at the Chicago sessions of the Symposium on Enzymes and Vitamins, 1941. 83. Dann, W. J., and Handler, P., J. Biol. Chem., 140, 935 (1941). 84. Adler, E., v. Euler, H., Gunther, G., and Plass, M., Skand. Arch. Physiol., 82, 6 (1939). LocKHART, E. E., Biochem. J., 33, 613 (1939). Abraham, E. P., and Adler, E., Biochem. J., 34, 119 (1940). 85. CoRRAN, H. S., Green, D. E., and Straxjb, F. B., Biochem. J., 33, 793 (1939). 86. Altschul, a. M., Persky, H., and Hogness, T. R., Science, 94, 349 (1941). The Flavoproteins T. R. HOGNESS University of Chicago r-|-iHE flavoproteins constitute a relatively large class of the _L respiratory enzymes. As we know them today they are charac- terized by having as prosthetic groups either alloxazine mono- nucleotide, i.e., riboflavin phosphate, or alloxazine adenine dinucleotide, which is composed of both riboflavin phosphate and adenylic acid. They are further characterized by their reactivity toward oxygen— some to a very limited extent— when in the re- duced form. Many of them are oxidized by methylene blue; one is oxidized by fumaric acid; and one is specifically oxidized by a known member of the hydrogen transport system: the flavoprotein cytochrome c reductase, in its reduced state, is oxidized by cyto- chrome c. In the oxidized state the flavoproteins as a group can be reduced by a variety of substrates— dihydrotriphosphopyridine nucleotide (cozymase II), dihydrodiphosphopyridine nucleotide (cozymase I), the d!-amino acids, xanthine and other purines, and some of the aldehydes. Each particular flavoprotein, however, is specific toward one (or one class) of the above substrates. The chemical structures of both alloxazine mononucleotide and alloxazine adenine dinucleo- tide are depicted in Figures 1 and 2. Stern and Holiday (1), by spectrographic methods, first found that the prosthetic group of Warburg's old yellow enzyme (2), which will be considered in more detail below, was a derivative of alloxazine, and the structure of the riboflavin phosphate was finally deteiTnined independently by Kuhn (3), Karrer (4), and their collaborators. Theorell (5) demonstrated that the enzyme contained one molecule of phosphate; and later Kuhn, Rudy, and Weygand (6), by synthesizing riboflavin-5-phosphoric acid, demonstrated the posi- tion occupied by the phosphate group. In Figure 1 the hydrogenation of the alloxazine mononucleotide is also indicated. The process of hydrogenation undoubtedly takes place in such a way that one hydrogen atom at a time is transferred from the substrate molecule to the riboflavin phosphate. The forma- tion of a red-colored intermediate in the reduction process consti- tutes much of the evidence in favor of this view. Kuhn and Wagner- 134 THE FLAVOPROTEINS 135 HOCH HOtH pO-PO,H; 'jng H3C HX- :-i^^VV%o 2H H HOCH HOCH ^ H ^fS^y^^r^^ ALLOXAZINE MONONUCLEOTIDE 1^ 6 DIHYDROALLOXAZINE MONONUCLEOTIDE FIGURE 1 Jauregg (7) made the observation that when riboflavin and also alloxazine are reduced in acid solution red-colored intermediates with absorption bands having maxima at 490 m[jL were formed. Haas (8) was able to demonstrate this intermediate formation with the old yellow enzyme and it has also been found in the case of H H H2^0-§-0-P-0-CH2 HOCH 0 0 HO9H HOCH ho6h H3C CH2 NH HOCH HOCH CH2 HC jT c^Vh "%y^^^^^ \ ^N' -C>. ^NH X' H, ALLOXAZINE ADENINE DINUCLEOTIDE FIGURE 2 p^ :o tf ^3 u 3 ^ 4J s 3 3 V a> 2 a -=i C/-V 0^-x a o 4J !«! 83 B3 2 ttJ O 4) •r 3.& 3 i3 fl-d 0 e 4-1 m CI. S c T3 OS a c3 ^ be a 3 .2 'J Xl .'" f-, 'S ^ U 'G '3a CU • .a o = a 2 6cg a O 4) O D So tT g i::^ a a O4 O 4) -G-d .2 o 05 3 ^1 & ~ '<^ I ^ -g-d^T3 o <= S o is c-d q «-^ 9 o CS U So "3 "O . en O s 55 o S-t3 ay -d = 1 "cS -d w> 9 J3 03 m o >jS >< V 4> g ■3 -2.9 13 "d -d a -a c3 1-^ (H (U c 0 -g a; en 0 05 .25 >>.22 05 .2 o 3-S P^-d ^ ^ >> M V O O -d "3^ fl2 o c« -d ti o 3 "d 0 V a 0 t, 0 ^ 0 0 60 0 >> *j X >> 0 0 _q & 0 0 rS-^ ' 'C ft 4^ tn ca . 05 ""^ T3 0 03 tH 0 0 05 0 ^ ffi tn tn - pyridine ->. Cytochrome c > Cytochrome c phosphate Zwischen- nucleotide reductase ferment 144 A SYMPOSIUM ON RESPIRATORY ENZYMES The cytochrome c is oxidized by oxygen through the intermediation of an enzyme or enzymes known as cytochrome oxidase, and by hydrogen peroxide with cytochrome c peroxidase as the catalyst (31). The cytochrome c oxidase, or one component of it, is probably Warburg's oxygen-carrying ferment, but it has not yet been defi- nitely identified. Nor do we yet understand the whole mechanism involved in the oxidation of cytochrome c. The test for determining the relative concentration of the cyto- chrome c reductase involves all these components. If oxidized cyto- chrome c is placed in a solution containing all the other components and if the concentration of the hexose monophosphate is in excess, the cytochrome c will be reduced, and the rate at which it is re- duced will depend upon the concentrations of the other components : Zwischenferment, triphosphopyridine nucleotide and the cytochrome c reductase. The concentrations of all substances can be so adjusted that the logarithmic rate of reduction of the cytochrome c will be pro- portional to the concentration of the reductase. (9) -^ = K(CR) dt In this equation (CyFe+'^+) represents the concentration of the oxi- dized cytochrome c and (CR) the concentration of the cytochrome c reductase. The equation is an empirical one. It is easy to demonstrate that all components of the system must be present before the cytochrome c will be reduced. In the reduced state cytochrome c displays three bands with maxima at about 410, 520, and 550 m^., and any one of these bands may be used to deter- mine the rate of reduction of the cytochrome. For most purposes the band at 550 m[j. is most convenient for analytical purposes, although for very dilute solutions of cytochrome c (10^ M) a wave length of 418 is used. As the cytochrome c is reduced, these bands appear. In the following spectroscopic demonstration, the general method employed in this test is illustrated. If all components except the reductase are added to the buffered solution in the absorption cell, no reduction of the cytochrome c takes place until a solution of the reductase is added to complete the chain of enzymatic reactions. Spectroscopic demonstration: The .apparatus consists of a simple focusing illuminating lantern (Central Scientific Company) slightly modi- fied to hold an absorption cell 2.5 cm. in diameter and 2 cm. thick. A lamp with a ribbon type filament is used, and the filament, turned edge- wise, is used as the slit. By placing a transmission grating in the light path a spectrum is projected on the wall. THE FLAVOPROTEINS 145 Then the cell is filled with buffer, and when all components of the reducing system except the reductase are placed before the filament, no bands appear. However, when the reductase is added, the two visible bands of cytochrome c gradually appear as intense broad black lines. The cytochrome is then reduced. To oxidize the cytochrome c, H2O2 is added. No reaction occurs, i.e., the black bands do not disappear, until the enzyme cytochrome c peroxi- dase is added. Upon the addition of this latter enzyme, the two bands disappear and reappear again as soon as the HgO, is completely reduced. Not only can this system be used to determine the concentration of cytochrome c reductase, but it can serve as an analytical method for the determination of any one of the constituents. In fact, my two collaborators, Haas and Harrer (9), have only recently worked out the conditions necessary for a relatively simple determination for triphosphopyridine nucleotide. By using glucose and adenosine triphosphate instead of hexose monophosphate, this system should serve as a method for determining the relative concentrations of either mutase, the enzyme that converts glucose-1-phosphate to glucose-6-phosphate, or hexokinase, the enzyme that directly phos- phorylates glucose-6-phosphate. Haas and Harrer (9) have also made very accurate measurements on the rate of oxidation of the reductase by cytochrome c and the rate of reduction by dihydrotriphosphopyridine nucleotide under a great variety of conditions, and as a result we have demonstrated by these kinetic measurements that the reductase forms com- plexes with both the dihydrotriphosphopyridine nucleotide and with the cytochrome c. The approximate first-order reaction velocities have been determined, as have also the heat of dissociation of the complexes and the energy of activation of the first-order reactions. The tentative results of these studies are embodied in Table 4. The values given in the table are approximations. Although the data Table 4.— Properties of cytochrome c reductase Reaction Dissociation constant of complex with reductase 25° First-order velocity constant 25° Heat of dissociation Energy of activation Reduction of reduc- tase by TPNH2 Oxidation of reduc- tase by CyFe+++ 1X10-6 moles per liter 7X10-8 moles per liter ca. 2000 min-i ca. 2000 min"! 2 kg. cal. 8 kg. cal. 12 kg. cal. 17 kg. cal. 146 A SYMPOSIUM ON RESPIRATORY ENZYMES are at hand, the final calculations (which involve a series of ap- proximations) have not been made. The dihydrotriphosphopyridine nucleotide forms a stable complex with the reductase, but the com- plex between the reductase and the cytochrome c is a more stable one. In this latter case the complex is one in which both components are proteins; the cytochrome c reductase, with a molecular weight of about 75,000, combines with cytochrome c, with a molecular weight of 13,000. In all probability both the dihydrotriphosphopyridine nucleotide and the cytochrome c, as well as the riboflavin phosphate, are simultaneously attached to the same protein molecule. This picture of the oxidation-reduction complex inspires specula- tion with respect to the manner in which the hydrogen atoms are transferred from the dihydrotriphosphopyridine nucleotide to the riboflavin phosphate constituents of the complex. Because of the complexity of the molecules involved, it is difficult to imagine the "active" positions of each of them approaching close enough for a direct hydrogen transfer. The most likely mechanism seems to be that of ionization, with a hydrogen ion dissociating into the solution, this process followed by an electron transfer from one molecule to the other, and this in turn followed by an attachment of another hydrogen ion from the solution to the new position of the electron. In the oxidation or reduction of cytochrome c we already regard the change taking place as electronic. Cytochrome c reductase bridges one gap in the hydrogen trans- port system— that between cytochrome and triphosphopyridine nucleotide. The other gap, between the cytochrome system and diphosphopyridine nucleotide, is still open, although a beginning toward the solution of this problem has been made. My two col- laborators, Altschul and Persky (32), have found a soluble protein in yeast that is capable of acting as an intermediate in the reduction of cytochrome c by dihydrodiphosphopyridine nucleotide. It is soluble; like other proteins it is precipitated by ammonium sulfate, acetone, and alcohol; it is heat-labile, and can be dialyzed without great loss of activity. It is not reactive toward triphosphopyridine nucleotide. Since its enzymatic function is so nearly like that of cytochrome c reductase, it, too, is probably a flavoprotein, although there is not yet any direct evidence to this effect. Of the flavoproteins which react with either dihydrodiphosphopyr- idine nucleotide or dihydrotriphosphopyridine nucleotide, only Warburg's old yellow enzyme and cytochrome c reductase react directly, and to any appreciable extent, with physiological oxidizing THE FLAVOPROTEINS 147 agents; the reaction between Haas's new yellow enzyme or Straub's yellow enzyme and oxygen is very slow, unless methylene blue is used as a "carrier." Possibly carriers other than methylene blue exist in respiring cells. Xanthine oxidase and aldehyde oxidase react with reducing agents which, while important, do not contribute greatly to the production of energy. Lipmann (33), using the Warburg separation technique, split from a soluble fraction obtained from Bacterium Delbriickii (Lactobacillus delbrilckii) a protein portion which was active as pyruvic acid oxidase only when both thiamine pyrophosphate and alloxazine adenine dinucleotide were added. From this result he postulated that a yellow enzyme possibly oxidizes thiamine pyrophosphate. Green, Knox, and Stumpf (34) have recently reported the finding of another yellow enzyme, the function of which has not yet been determined. Formerly it was assumed that cytochrome b, because of its poten- tial, acted as one of the intermediaries between the flavoproteins and cytochrome c, but the discovery of cytochrome c reductase has obviated the necessity for any such intermediary, although in the intact cell it may act as such (two or more paths of oxidation may be in operation). At present nine flavoproteins are known, not all of which react with both the oxidizing and the reducing agents which are present in the living cell; the discovery of more or these important enzymes will undoubtedly follow. The relatively large riboflavin content of the liver and the kidney and the multiple and complicated bio- chemical functions of these organs alone indicate that we might expect to find many more members of this class of enzymes. REFERENCES 1. Stern, K. G., and Holiday, E. R., Ber., 67, 1104, 1442 (1934). 2. Warburg, O., and Christian, W., Biochem. Z., 254, 438 (1932); 263, 228 (1933); 287, 291, 440 (1936). 3. KuHN, R., et al, Ber., 66, 1034 (1933); 68, 1765 (1935); 69, 1557 (1936). 4. Karrer, p., Helv. Chim. Acta, 18, 69, 72, 426 ( 1935); Ber., 68, 216 ( 1935). 5. Theorell, H., Biochem. Z., 272, 155 (1934). 6. KuHN, R., Rudy, H., and Weygand, F., Ber., 69, 2034 (1936). 7. KuHN, R., and Wagner- J auregg, Th., Ber., 67, 361 (1934). 8. Haas, E., Biochem. Z., 290, 291 (1937). 9. Haas, E., Harrer, C, and Hogness, T. R. (unpublished). 10. Krebs, H. a., Biochem. J., 29, 1620 (1935). 11. Warburg, O., and Christian, W., Biochem. Z., 296, 294; 298, 150 (1938). 12. Haas, E., Biochem. Z., 298, 378 (1938). 13. Straub, F. B., Biochem. J., SS, ISl (1939). 14. Warburg, O., and Christian, W., Biochem. Z., 298, 368 (1938). 15. Ball, E. G., J. Biol. Chem., 128, 51 (1939). 148 A SYMPOSIUM ON RESPIRATORY ENZYMES 16. Fischer, F. G., and Eysenbach, H., Ann., 530, 99 (1937). 17. Fischer, F. G., Roedig, A., and Rauch, K., Naturwissenschaften, 27, 197 (1939). 18. Gordon, A. H., Green, D. E., and Subrahmanyan, V., Biochem. J., 34, 764 (1940). 19. Haas, E., Horecker, B. L., and Hogness, T. R., J. Biol. Chem., 136, 747 (1940). 20. Kalckar, H. M., Chem. Rev., 28, 72 (1941). 21. Green, D. E., Mechanisms of Biological Oxidations (Cambridge Univer- sity Press, 1940). 22. Oppenheimer, C., and Stern, K. G., Biological Oxidation (W. Junk, The Hague, 1939). 23. Barron, E. S. G., and Harrop, G. A., J. Exp. Med., 48, 207 (1928), J. Biol. Chem., 79, 65 (1928). 24. Ball, E. G., and Ramsdell, P. A., J. Biol. Chem., 131, 767 (1939). 25. Morgan, E. J., Stevi'art, C. P., and Hopkins, F. G., Proc. Roy. Soc. (London), B 94, 109 (1922). 26. Theorell, H., Biochem. Z., 275, 37 (1934). 27. Warburg, O., and Christian, W., Biochem. Z., 296, 294 (1938). 28. Warburg, O., and Christian, W., Biochem. Z., 295, 261; 298, 150 (1938). 29. Theorell, H., Biochem. Z., 288, 317 (1936). 30. Szent-Gyorgyi, A., Z. physiol. Chem., 249, 211 (1937). 31. Altschul, a. M., Abrams, R., and Hogness, T, R., J. Biol. Chem., 136, 777 (1940). 32. Altschul, A. M., Persky, H., and Hogness, T. R., Science, 94, 349 ( 1941). 33. Lipmann, F., Nature, 143, 436 (1939). 34. Green, D. E., Knox, W. E., and Stumpf, P. K., J. Biol. Chem., 138, 775 (1941). Wilson and Kalckar: 'The manuscript deadline is HoGNESs AND Elvehjem: "The respiratory eiiz\nie, alloxazine- adenine dinucleotide, . . ." "The vitamin, riboflavin, . . ." OcHOA, Wood, and Carson: "With hea\ v carbon . "But with radioactive carbon . . ." COMMENTS OFF THE RECORD Cytochromes ELMER STOTZ Harvard University"* THE SUBJECT OF THE iron-containing cellular respiration catalysts has demanded the attention of many biochemists. One obvious reason for this is the fact that the cytochromes are so readily detectable by spectroscopic means that their wide distribution and hence their apparent importance were early recognized (1, 2). With increasing knowledge of cellular respiration the unique and funda- mental position of the cytochromes in the respiratory scheme has been emphasized more and more. Although the various substrates of respiration may require many enzymes, coenzymes, and mediators, the individual pathways appear to converge at the cytochrome sys- tem. It is through the ferrous to ferric change of this system that the electrons of the ultimate substrate, hydrogen, come to terms with the ultimate oxidant, oxygen. In proportion to its importance in respiration, perhaps less is known of the cytochromes than of other respiratory components. Cytochrome c has been isolated, but the peculiar linkage of its prosthetic group with the protein is not yet fully understood. Cyto- chromes a and b still remain bands in the absorption spectrum of tissues with little appreciation of their function. Finally, the all- important cytochrome oxidase is still httle more than the insoluble ground residue of tissues. Properties of the Cytochrome Components Cytochrome c— Cytochrome c appears to be quantitatively the most important of the three cytochrome constituents. It was defined spectroscopically by Keilin (2, 3) as that component which in the reduced state has an alpha-band at 5500 A. and a beta-band at 5200 A. It was first isolated in apparently pure form by Theorell (4) in 1936 by a dilute sulfuric acid extraction of defatted beef heart muscle; isolation was followed by ammonium sulfate precipi- tation, barium sulfate adsorption, acetone precipitation, and finally adsorption on cellophane. Shortly thereafter Keilin and Hartree (5) * Contribution from the McLean Hospital, Waverly, Massachusetts, and the Harvard Medical School, Boston. 149 ^ 150 A SYMPOSIUM ON RESPIRATORY ENZYMES described a simpler method of preparation involving a trichloracetic acid extraction of the ground muscle, ammonium sulfate fractiona- tion, and trichloracetic precipitation. Both groups of workers ob- tained a product which had the same absorption bands as the cytochrome c of the intact or phosphate-extracted muscle, and which contained 0.34 per cent iron, resisted further fractionation, and was therefore considered pure. Later Theorell and Akesson (6) obtained by electrophoretic means a product containing as high as 0.43 per cent iron. It would be of interest to know whether this product displayed an equal increase in catalytic activity over the 0.34 per cent iron product. Zeile and Renter (7) calculated a molecular weight of 18,000 for cytochrome c on the basis of its hemin content. Theorell (4), from a study of its diffusion and sedimentation, determined a molecular weight of 16,500. The iron content of 0.34 per cent also yields an equivalent weight of 16,500. The isoelectric point of cytochrome c is approximately at pH 9.8 (4). Ferric cytochrome c is readily reduced by a variety of agents, such as hydrosulfite, ascorbic acid, cysteine, adrenalin, hydroqui- none, p-phenylenediamine, and many leuco dyes, as well as by certain physiological reducing systems to be discussed later. Ferro -cytochrome c is essentially non-autoxidizable in neutral solution, the slow rate being largely inhibited by small amounts of cyanide (8), indicating heavy metal catalysis. It is readily oxidized by ferricyanide and aerobically by oxidase preparations. Below pH 4.0 and above 11.0 the spectrum changes and the substance becomes autoxidizable. A neutral solution of cytochrome c can be boiled, and upon cooling the original spectrum and catalytic properties return. Cytochrome c is likewise stable to dilute acid. It is stable to 0.1 normal potassium hydroxide, but 1.0 normal alkali produces an irreversible change in the spectrum. The substance is then autoxidizable and forms a light- sensitive carbon monoxide compound (9). According to Keilin (10), cytochrome c does not appear to com- bine with hydrogen sulfide, hydrogen cyanide, sodium azide, or hydroxylamine, nor, according to Stern (11), with carbylamine. Ferri-cytochrome c does, however, form a compound with nitric oxide (12). Keilin (13) was unable to detect any change in the spectrum of the c component in the presence of carbon monoxide except in solutions whose pH was above 13.0. Altschul and Hogness (14), using an accurate photoelectric spectrophotometer (15) with a CYTOCHROMES 151 narrow slit, have, however, found evidence of a ferro-cytochrome- carbon monoxide compound throughout the entire pH range. The change in spectrum was reversible; that is, the carbon monoxide could be removed by nitrogen. The carbon monoxide compound was light-sensitive. Keilin and Hartree (10) attribute this finding to the presence of denatured cytochrome or other hematin compounds. At neutral pH they were able to liberate and measure manometri- cally only 10 per cent of the theoretical amount of carbon monoxide that should combine with reduced cytochrome. Potter (16) concluded that cyanide also, contrary to popular belief, forms a complex with ferri-cytochrome c. This conclusion was based not only on the fact that a change in the spectrum was detected but also upon studies on the enzymatic reduction of cyto- chrome c. Since the spectral shift is small, as with carbon monoxide. Potter questions whether one can safely conclude from simple spec- troscopic observation that a given inhibitor has not reacted with cytochrome. Considerable gains have been made in determining the structure of the prosthetic group of cytochrome c and how this might explain the peculiar stability of the heme-protein linkage. Hill and Keilin (17) obtained a porphyrin by hydrochloric acid and sulfur dioxide treatment of cytochrome c which, unlike most porphyrins, was solu- ble in water. Zeile and Piutti (18), in extensive synthetic work, were able to introduce various nitrogen bases into the unsaturated side chains of protoporphyrin and obtain porphyrins whose solubility was similar to that obtained from cytochrome c. When iron was introduced into some of these compounds, they showed the charac- teristic cytochrome c absorption band at 5500 A. Later Zeile and Renter (7) isolated hematoporphyrin from a hydrobromic-acetic acid degradation of cytochrome c. Theorell (19) isolated a sulfur- containing porphyrin and postulated that the vinyl, groups of the hemin are linked to amino acids of the protein by thio-ether bonds. Upon demonstrating later that such a porphyrin could arise by con- densation of hemato -porphyrin with cysteine during the course of the cytochrome hydrolysis, Theorell (20) explained that either nitro- gen or oxygen as well as sulfur might form the connecting link. Zeile and Meyer (21) offer support to the sulfur-bridge theory in obtaining the sulfur-containing porphyrin under conditions of hydrolysis in which a condensation of porphyrin with free cysteine would be very unlikely. A tentative structure of cytochrome c is illustrated in Figure 1. 152 A SYMPOSIUM ON RESPIRATORY ENZYMES Cytochrome c has, as compared with other biological systems, a very high oxidation-reduction potential. As early as 1932 CooHdge (22), using an impure preparation from yeast, reported a potential. The rather unsatisfactory potential with the electrode could be stabilized with hydroquinone. By adding oxidants or reductants to a point where the spectrum of the cytoclirome changed, an Eo' of + 0.260 V. at pH 7.0 was recorded. Lower values were obtained at pH 5.0. In 1934 Green (23) determined the potential of an impure PROTEIN HX H.G rNJH2 COOH CH r^Hp COOH C7H,oO±C±2H ± O — I S CH-CH CH2 COOH Figure 1. — Structure of cytochrome c CYTOCHROMES 153 yeast cytochrome c, but found a much lower value of Eg' — +0.127 V, between pH 4.6 and 7.1. Finally, Wurmser and Filitti-Wurmser (24) in France and Stotz, Sidwell and Hogness (25) in Chicago, measured the potential of pure cytochrome c isolated from heart muscle. The former measured the equilibrium potential in mixtures of reduced and oxidized cytochrome c, the proportion being deter- mined spectrophotometrically. They obtained the value Eo' = + 0.254 V. between pH 5.0 and 8.0. Stotz et al. used a purely spectrophotometric method and obtained a value of +0.262 v. in the same pH range. The spectrophotometric method consisted in measuring accurately the amounts of oxidized and reduced indicator and cytochrome in equilibrium with each other. The potential of the indicator being known, the potential of the cytochrome could be readily calculated. The results recorded by the two groups were reached independently and represent good agreement. They are both in essential agreement with the potential of +0.27 v. reported by Ball (26), who was able to estimate the potentials of the three cytochromes as they existed in a heart muscle extract. At a physio- logical pH, therefore, cytochrome c has about the same potential as the hydroquinone-quinone system. Cytochrome h.—in the reduced state this cytochrome component possesses an alpha-absorption band at 5640 A. and a beta-band at 5300 A. It appears to be more closely bound to the insoluble material in tissue extracts than is cytochrome c. Nevertheless Ya- kushiji and Mori (27) claim to have isolated cytochrome Z? in a soluble form. It seems doubtful from their method of preparation whether the product obtained could be an undenatured cytochrome h. In some of the original extracts the reduced band is not at 5640 A., but as purification proceeds this band is shifted to the normal posi- tion of reduced cytochrome h. They believe that the hemin portion of their product is ordinary protohemin. Since several other proteins combined with protohemin to form spectroscopically and catalyti- cally similar hemochromogens, it is difficult to believe that these workers actually obtained cytochrome h. Judged from its behavior in tissue extracts, cytochrome h appears to be a thermolabile hemin-protein complex. Unlike cytochrome c, the h component is autoxidizable. Since this component reaches equilibrium with other reversible systems in a heart muscle extract. Ball (26) was able to estimate its potential as —0.04 v., the lowest of the cytochrome components. It does not combine with carbon monoxide or other respiratory inhibitors (10). 154 A SYMPOSIUM ON RESPIRATORY ENZYMES Cytochrome a.— Cytochrome a is another component whose prop- erties can be judged only in a crude tissue extract. It was originally designated as the component which in the reduced state possessed an alpha-band at 6000-6050 A. Upon more careful analysis, Ball (26) and Keilin and Hartree (10) discovered that this band was not homogeneous to reduction or to various reagents; that is, the absorption in this region must be attributed to more than one sub- stance. Their results might be interpreted to mean that the portion of the band nearer 6050 A. is to be attributed to cytochrome a. Because of this complication, the properties of cytochrome a have not been definitely established. Keilin and Hartree (10) conclude that it does not combine with carbon monoxide or cyanide. It is reduced by the same agents as the other cytochromes. Its potential has been estimated by Ball (26) as +0.29 v., a value which is perhaps less certain than those of the other cytochromes. Very recently Yakusizi and Okunuki (28) claim to have isolated cytochrome a from heart muscle. The muscle pulp was extracted with sodium cholate and alkaline phosphate. Ammonium sulfate fractionations followed by redissolving in the cholate mixture yielded a product which was, in the oxidized state, of a red-brown color. When it was reduced it was green, indicating that the prosthetic group was of the "mixed" or Spirographis hemin type. The reduced compound showed a strong absorption at 6050 A. and a weak band at 5130 A., carbon monoxide having no effect on the spectrum. Reduced cytochrome c was partially oxidized by the oxidized form of this compound. The latter properties are in agreement with our concept of cytochrome a, and this important finding should be con- firmed and extended. There is some evidence that the potentials of the yeast cyto- chromes differ from those of the heart cytochromes. But the only measurement on a pure component has been with heart cyto- chrome c. It will be recalled that Green (23) obtained a value of Eo' = +.127 V. for yeast cytochrome c, and the recent work of Baumberger (29) is in agreement with this finding. The latter work, however, requires certain comments. Baumberger was able to meas- ure simultaneously the light absorption at various wave lengths (photoelectrically) and the Ei, levels of a yeast suspension. The suspension was vigorously stirred by oxygen-nitrogen mixtures, which likewise eventually established a constant Ei, level. By vary- ing the gas mixture and hence the Eh level, the presence or absence of the cytochrome bands could be determined by changing the wave CYTOCHROMES 155 length of the incident hght and observing the galvanometer deflec- tions of the photoelectric device. In this way he arrived at the above potential for cytochrome c. Perhaps more startling was his finding that all three cytochromes appeared to have the same potential; that is, at a given Eh level all were equally reduced. This is distinctly a contradiction of Ball's finding that in a heart muscle extract the cytochrome potentials diflfered markedly. In Ball's work, however, the relative degree of reduction of the cytochromes was measured when in equilibrium with systems of known potential and systems known to react with the cytochromes, whereas in Baumberger's work the normal re- ductants within a more organized structure establish the equilibrium. A legitimate question concerning these experiments would be whether it can be assumed that the potential recorded by a platinum electrode in a yeast suspension is the same as that existing within the cell. And, furthermore, is it not likely that different points in the organized cells actually have very different potentials? The data might indeed be taken as evidence for the latter hypothesis. Baum- berger suggests the possibility of a molecular aggregate of the three cytochromes which is oxidized or reduced as a whole or in which the three cytochromes do have the same potential. In a more disorganized structure, such as a heart muscle extract, these relations apparently do not exist, and it is not unlikely that this explains the great difference between the frequency of oxida- tion and reduction of cytochromes in the intact yeast as compared with that in extracts. Certain it is that in the study of tissue respira- tion the problem of adsorption and dependence of function on organized structure is met most frequently in the consideration of the cytochrome system (see "Oxidation and Cell Structure" in Korr, 30). Cytochrome Oxidase (Cytochrome a^).—\t may be recalled that in 1924 Warburg (31), upon observing the cyanide sensitivity of cellu- lar respiration in conjunction with the catalytic behavior of the hemin-charcoal model, gave the name Atmungsjerment to the cata- lytically active iron compounds involved in cellular respiration. This study continued with measurements of the inhibition of yeast respiration by carbon monoxide and its reversibility by light (32, 33). By measuring this effect at various light frequencies, the relative carbon monoxide spectrum and later the absolute carbon monoxide spectrum of the Atmungsferment were determined (34, 35). It was renamed the Sauerstoffiibertragendes Ferment or oxygen- 156 CYTOCHROMES 157 transferring enzyme. The spectrum was obviously that of a hemin- containing compound which resembled in type that of Spirographis hemin (36). The alpha-band of the reduced carbon monoxide com- plex lies at 5920 A. and the gamma-band at 4320 A. Finally, in the highly respiring Bacterium Pasteurianum (Acetobacter pasteuria- num), under anaerobic conditions, a weak band was observed at 5890 A. which was attributed by Warburg and Negelein (37) to the reduced form of the oxygen-transferring enzyme itself, since carbon monoxide shifted the band to 5920 A. On the other hand, cyanide produced a band at 6390 A. which, since it may be observed even in the simultaneous presence of the 5890 A. band, need not be a derivative of the oxygen-transferring enzyme. Keilin (38) believes, however, that the 5890 A. band is only a degradation product of cytochrome a and is seen only in certain bacteria. In fact, Fujita and Kodama (39) observed the 5890 A. band in bacteria only when the cytochrome a band was absent and have named this band cyto- chrome flj. Certainly the best criteria by which to establish the identity of a compound with Warburg's oxygen-transferring enzyme would be the positions of the carbon monoxide absorption bands (5920 and 5320 A.). Keilin's work (40) with "indophenol oxidase" pointed to the identity of this enzyme with the Warburg enzyme. The oxidase brought about the aerobic oxidation of the cytochromes, was in- hibited by cyanide, and showed a Hght-reversible inhibition with carbon monoxide. Because of the similarity it has been generally believed that War- burg's "oxygen-transmitting enzyme" and Keilin's presently-called cytochrome oxidase are identical. Until recently, however, Keilin had not observed a band that he could attribute to the carbon monoxide complex of the oxidase. In 1939 Keihn and Hartree (10) believed that they had identified spectroscopically in heart muscle extracts a new cytochrome, a^, which might be identical with the oxidase. They concluded that the cytochrome a band at 6000-6050 A. is actually due to com- ponents a and a^, since upon addition of carbon monoxide this band divides and a new one appears at 5900 A. With the aid of strong cane sugar or glycerine solutions, or bile salts, to clarify the solutions for spectroscopic examination, they were able to examine further the Soret or gamma-bands of the cytochromes. Simultaneously with the above shift, a portion of the 4480 A. band is shifted to 4320 A. These two new bands represent the carbon monoxide complex of 158 A SYMPOSIUM ON RESPIRATORY ENZYMES the new component, a^. Since these positions correspond to those of the Warburg enzyme, a^ may be identical with this enzyme. On the other hand, the alpha-band of the compound itself is claimed to be at 6000 A., whereas Warburg and co-workers (41, 42, 43) believe their compound to have a band at 5890 A. This compound (also called cytochrome a^) likewise reacts with oxygen, and formation of its cyanide compound prevents reoxidation of the other cytochromes. Furthermore, its carbon monoxide compound has an alpha-band at 5920 A. Cytochrome a^, in either the reduced or the oxidized form, com- bines with potassium cyanide. The cyanide complex of the reduced form is readily autoxidizable, whereas that of the ferric form is not easily reduced. Ferric cytochrome a^ also reacts with hydrogen sulfide, sodium azide, and hydroxylamine. It is thermolabile and easily destroyed by organic solvents, acids, or alkalies. Cytochrome flg is reduced along with the other cytochromes by e.g., succinate. It is also autoxidizable. It is therefore believed that a^ is also identi- cal with cytochrome oxidase. The authors have themselves offered certain objections to the above conclusion: 1. It was impossible to demonstrate the reduction of cytochrome flg by added reduced cytochrome c, but technically these experi- ments were not satisfactory. In this connection it is interesting that Ball (26) noticed that the portion of the 6000-6050 A. band at- tributed to ^3 by Keilin has a higher potential than any of the other cytochromes, an expected but not an essential condition for oxidase function. 2. The carbon monoxide compound of ferro -cytochrome flg did not appear to be sensitive to light anaerobically. The effect of light, however, may become apparent only in the presence of oxygen, which oxidizes the a^ component and thereby prevents its reaction with carbon monoxide. Such an explanation of the light effect could also explain its property of relieving inhibition of carbon monoxide. 3. Finally, it has been found possible in the presence of carbon monoxide to oxidize cytochromes a, b, and c by air while the spec- trum of the carbon monoxide complex of reduced a^ remains visible. It is therefore difficult to explain the oxidation of a, h, and c through the flg component. In general, it may be said that the Keilin and Hartree paper by no means clarifies the whole problem of the identity of cytochrome oxidase and oxygen-transferring enzyme or of either one with the CYTOCHROMES 159 new flg component. The situation is understandable, however, when one considers the number of hematin compounds which exist in tissue preparations and the fact that apparently the same component may vary slightly in the position of its absorption band in diflFerent biological materials. An experiment that has been much needed has finally been pub- lished as a short note by Melnick (44), namely, the photochemical determination of the carbon monoxide spectrum of cytochrome oxidase. This was accomplished by employing a phosphate extract of heart muscle with succinate as substrate. On the assumption that the oxidase is the only functional substance present in the prepara- tion which forms a light-dissociable carbon monoxide complex, the spectrum measured should be, by Keilin's own definition, cyto- chrome oxidase. The spectrum obtained was that of a pheohemin compound. The alpha-band was located at 5890 A., which agrees very well with that of the oxygen-transferring enzyme in yeast and bacteria, as well as with the carbon monoxide complex of cyto- chrome Og seen directly. But in the case of the gamma- or Soret- band, Melnick finds a band at 4500 A., which does not agree with the carbon monoxide band of the oxygen-transmitting enzyme (War- burg) nor with that of a^ (Keilin). The whole situation therefore remains clouded and awaits chemical separation and identification for its clarification. Keihn and Hartree (10) believe that because of the association of cytochromes a and a^ the two components are intimately related. They may have an identical heme nucleus, since on alkali denatura- tion and addition of pyridine they yield the same hemochromogen. They are both sensitive to heat, alcohol, acetone, and extreme changes in pH. It has been suggested that cytochrome oxidase may be a copper protein. The evidence is quite indirect, such as the wide distribution of copper, the ability of copper salts to oxidize cytochrome c, and certain similarities between cytochrome oxidase and the copper- containing polyphenol oxidase (45, 46). It may be pointed out that copper does appear to be essential in the formation of cytochrome oxidase. Cohen and Elvehjem (47) have found it to be essential for the regeneration of cytochrome a and oxidase in anemic rats, and Yoshikawa (48) finds it to be a stimulus to the oxidase activity of yeast cultures. Most conclusive is the recent work of Schultze (49, 50), who showed that copper was necessary for the maintenance and formation of cytochrome oxidase in rat liver and heart, and that 160 A SYMPOSIUM ON RESPIRATORY ENZYMES the regeneration of cytochrome oxidase in the bone marrow of anemic rats was extremely rapid following administration of copper. Important as these findings are, they do not constitute proof that cytochrome oxidase contains the copper any more than similar find- ings with hemoglobin regeneration. Graubard (51) claims actually to have isolated a water-soluble cytochrome oxidase from uterus. It is very labile and its action is inhibited by copper inhibitors. He therefore claims that at least this oxidase contains copper as the active metal. Certainly any claims for the identity of a water-soluble or copper-containing oxidase with cytochrome oxidase must be supported by many experiments giving evidence that the material conforms with the existing definition of cytochrome oxidase. The work of Altschul, Abrams, and Hogness (52, 53) in connec- tion with the oxidase is of particular interest. They first reported the isolation from yeast of a soluble cytochrome oxidase which aerobically oxidized cytochrome c, and which was inhibited by cyanide and carbon monoxide. Upon concentration, it was noted, the activity of the enzyme was inhibited by catalase. It was then found that during reduction of the substrate, cytochrome c, hydro- gen peroxide was produced as a contaminant and oxidation was actually a catalyzed oxidation by peroxide. This enzyme, being specific, is now called cytochrome c peroxidase. Because so little is known about the physiological mechanism of the aerobic oxidation of the cytochromes and of the function of hydrogen peroxide, this very active enzyme may. be of no small importance physiologically. From the standpoint of isolation, therefore, we still have only cytochrome c. The other cytochromes and the oxidase are still as- sociated with insoluble particles and have resisted separation. The usual opalescent alkaline phosphate extract of heart muscle, which contains these substances, has been the subject of a physico-chemical investigation by Stern (54). Observations of this material, in the ultracentrifuge and electrophoresis reveal properties similar to those of other macro-molecular materials, such as fractions from Rous chicken sarcomata. Such a dispersed suspension of particles con- tains lipids, nucleic acid, hemin, and other constituents. Stern feels that it is largely a matter of definition whether such a "mono- dispersed" suspension should be called a mechanical dispersion or a true solution, and he predicts that when the individual components are isolated they will no longer display their characteristic biological orientation. This view does not and should not discourage attempts CYTOCHROMES 161 to isolate the individual components. Although such heart muscle preparations have thus far resisted fractionation, one factor, the diaphorase or coenzyme factor (a flavoprotein) has been separated from such a preparation (55). Euler and Hellstrom (56) claim to have efiFected a separation of the cytochromes by an ammonium sulfate fractionation of a sodium cholate clarified preparation. The first fraction to precipitate con- tained cytochromes a and h, the second only cytochrome h, and the third, cytochrome c. The activity of some of these fractions toward succinate does not, however, bear out the claim for any extensive fractionation (see Keihn and Hartree, 9). Such precipitates simply resuspend in buffers to yield the usual opalescent preparations. Nevertheless the clarifying or so-called peptizing action of the bile salts is interesting. I have found (8) in a few experiments that sodium desoxycholate at a neutral pH yields a virtually clear solu- tion of the original turbid oxidase preparation. Fractional salt pre- cipitations have not, however, yielded any striking results. In view of the action of bile salts with lipids, it is possible that the diflBculties of separating the cytochrome components may lie in their associa- tion with or their presence as lipo-proteins. Reduction of the Cytochromes by Other Respiratory Systems The fundamental position of the cytochromes in cellular respira- tion is emphasized by the fact that tissue respiration is so completely blocked by cyanide (57), and by the experiment of Haas (58) in which it was demonstrated that the rate of alternate oxidation and reduction of cytochrome c in intact yeast cells could account for all the oxygen consumption of the yeast. The succinate-succinic dehydrogenase system has long been recog- nized as a reducing system for the cytochromes, and this connects the important Szent-Gyorgyi— Krebs cycle with the cytochromes. On the other hand, the details of this reduction are by no means clear. Thus in 1939 Hopkins, Lutwak-Mann, and Morgan (59) prepared a succinic dehydrogenase from heart muscle which with succinate did not reduce cytochrome c, but did nevertheless reduce methylene blue. The preparation was made in such a way (with alcohol treat- ments) that no cytochrome oxidase activity remained. Their result suggests another intermediate between succinic dehydrogenase and cytochrome c. Stern and Melnick (54, 60) in their ultracentrifuge studies found that the sedimented material showed typical succinic dehydrogenase 162 A SYMPOSIUM ON RESPIRATORY ENZYMES activity toward methylene blue and oxidase activity toward p-phenylenediamine or hydroquinone. It lacked, however, the ca- pacity to oxidize succinate, apparently being unable to reduce cytochrome c. When supernatant fluid from the ultracentrifuge run was added, aerobic activity of the preparation toward succinate was restored. The unknown material is evidently a substance of lower molecular weight (estimated at 140,000), is heat-labile, and is re- moved by trichloracetic acid, hence is probably a protein. It can- not be identified with aluminum (61), catalase, or the Straub flavo- protein (55). Recently Keilin and Hartree (9) have tested and analyzed the eflFects of several factors on succinate and p-phenylenediamine oxi- dation by typical heart muscle extracts. Among their findings were these: 1. Narcotics inhibited the oxidation of succinate by the cyto- chrome system more strongly than the oxidation of methylene blue. Since not only reduction of the cytochrome components was in- hibited, but also oxidation of cytochrome h, it is possible that these findings are related to the function of cytochrome h in succinate oxidation. 2. Preparations treated with alcohol modified irreversibly the spectrum of cytochromes a^, a, and h and destroyed oxidase activity. Such preparations did not reduce cytochrome c, although they retained their ability to reduce methylene blue. 3. Treatment with acetic acid (pH 5.0 for one hour) did not affect p-phenylene- diamine oxidation, but destroyed the ability of the succinate system to reduce cytochrome c. Again, methylene blue reduction was still possible. Spectroscopically, the absorption bands of the cytochromes were normal, except that cytochrome h appeared to be no longer autoxidizable. Evidently cytochrome h had undergone some change. 4. Treatment with pancreatin gave a preparation similar to the acid-treated preparation. The results suggest that cytochrome h may be the labile link between succinic dehydrogenase and cytochrome c, although the possibility that a flavin or another hematin is a link is by no means excluded. It may be recalled that the potential of cytochrome h (—0.04 V.) places it in a favorable position as such a link. Keilin and Hartree have suggested as an alternative the possibility that the failure to react with cytochrome c "may be due to an irreversible change in the colloidal structure of the preparation accompanied by a loss of accessibility of the succinic system to c, which is a non- diffusible protein while it remains still accesible to small and dif- fusible molecules of methylene blue." CYTOCHROMES 163 It was long desirable to find a link between the di- and tri- phosphopyridine nucleotides and the cytochrome system. For some time it has been known that the old yellow enzyme, i.e., the Warburg-Christian flavoprotein (62), was readily reduced by tri- phosphopyridine nucleotide (63), but only very slowly oxidized by cytochrome c (64). This link has recently been established by the excellent isolation work of Haas, Horecker, and Hogness (65), who isolated a flavo (mononucleotide) -protein which rapidly reduces cytochrome c. It has been given the functional name "cytochrome reductase." The coenzyme II dependent systems have thus been satisfactorily linked to the iron-containing system. Cytochrome c reductase does not hnk reduced coenzyme I with cytochrome c (66). Haas, Horecker, and Hogness (65) believe that since the reductase loses its power to reduce cytochrome c when subjected to chemical treatments common to the preparation of the old Warburg- Christian flavoprotein, it is very probable that the latter represents a denatured product of cytochrome reductase. The story is less satisfactory than in the case of coenzyme I. In 1937 Adler, Euler, and Hellstrom (67) discovered an enzyme which they called "diaphorase," and independently Green and Dewan (68) investigated what was apparently the same enzyme, which they called "coenzyme factor." This factor appeared to link coenzyme I with the cytochrome system. Of significance for the present discus- sion is the fact that Green and Dewan (68) stated that cytochromes a and b but not c were involved in the reaction. The absence of cytochrome c in their preparation has been challenged by Haw- thorne and Harrison (69) and by Lockhart and Potter (66). The latter authors have in fact demonstrated that cytochrome c is a link in the aerobic oxidation of coenzyme I. They also have used two types of "diaphorase" preparation, one of which catalyzed the coenzyme I reduction of cytochrome c and the other did not. Both, however, could catalyze the reduction of methylene blue. This recalls the findings in connection with succinate reduction of cytochrome c; in fact, Lockhart and Potter have noted that the preparation unable to reduce cytochrome c contained no cytochrome b. Thus comes the suggestion but not the proof that cytochrome b may be involved as a link between coenzyme I dependent systems and cytochrome c. Straub (55) has isolated from heart muscle a flavoprotein which is considered to be identical with the coenzyme factor or diaphorase. Corran, Green, and Straub (70), in studying the catalytic proporties of this flavoprotein, find that its reduced form is only slowly autoxi- 164 A SYMPOSIUM ON RESPIRATORY ENZYMES dizable and that it does not react with cytochrome c but only with "carriers" such as methylene blue. It is not yet known whether this protein requires another link (such as cytochrome h) for cytochrome c reduction, or whether it is actually a denatured product which has lost its ability to reduce cytochrome c, analogous to the old Warburg-Christian yellow enzyme and the newer cytochrome re- ductase. Catalytic Relations of the Cytochromes and Oxidase "Indophenol oxidase" was long recognized as the substance in tissues which produced the aerobic oxidation of Nadi reagent or p-phenylenediamine (3). This name was retained in spite of the finding of Keilin (71) that the addition of cytochrome c accelerated the oxidation of cysteine by "indophenol oxidase" and a similar finding by Stotz, Harrer, Schultze, and King (72) with respect to ascorbic acid. When pure cytochrome c became available, it was not difficult to study the relation of oxidase and cytochrome c in the oxidation of various substrates. The high oxidation-reduction potential of cytochrome c that had been noted suggested that the action of indophenol oxidase was due to an unspecific reduction of the cytochrome c which it contained, followed by an aerobic cata- lyzed oxidation of the reduced cytochrome. KeiHn and Hartree (73), upon noting the accelerating efi^ect of cytochrome c on the oxidation of several substrates, renamed the oxidase "cytochrome oxidase." Stotz, Sidwell, and Hogness (74) had come to the same conclusion and had prepared an oxidase which was largely free of cytochromes Table 1.— The sensitivity of hydroquinone and p-phenylenediamine oxidations to cyanide* (Reference 74) Percentage of inhibition Cyanide concentration mM X 106 total Hydroquinone p-phenylenediamine oxidation oxidation 0 0 0 20 57 52 40 82 70 60 92 77 100 99 82 120 100 83 240 100 85 T = 38° C; pH, 7.15; 23.5 X 10"* mM cytochrome c total. CYTOCHROMES 165 c and h. This preparation was essentially unable to oxidize either hydroquinone or p-phenylenediamine without the addition of cyto- chrome c. A study of the cyanide sensitivity of hydroquinone and p-phenylenediamine oxidation (see Table 1) suggested, because of the redox potential relations of the cytochromes and the substrates, that hydroquinone oxidation involved only the oxidase and cyto- chrome c, whereas p-phenylenediamine could be independently oxidized by cytochrome h as well. The cyanide-resistant portion of p-phenylenediamine oxidation is probably due to the autoxidizable, cyanide-resistant cytochrome h. The oxidation of hydroquinone is a function of both the oxidase and cytochrome c; hence its oxidation by tissue extracts is not an absolute method for determining either substance. The effect of cytochrome c in accelerating the rate of hydroquinone oxidation by a heart muscle oxidase preparation is shown in Figure 3. 500 400 300 200 100 Velocity (cmm. 02/hr.) h)iS^o -^Mr O addition of cytochrome clone A oddition of cytochrome to heated oxidase ■ addition of cytochrome to crude oxidase preparation I I I I I I I I 1 0 10 20 30 40 50 60 70 80 Added cytochrome c (mM. X|0^) 90 100 Figure 3. — The oxidation of hydroquinone by the oxidase-cytochrome c system. T = 38°C., pH 7.15, hydroquinone 0.033 mM. total. 166 A SYMPOSIUM ON RESPIRATORY ENZYMES The reduction of cytochrome c by hydroquinone is very rapid; hence the rate-controHing reaction here is the oxidation of the re- duced cytochrome. Several curves such as those in Figure 3 have been found to comply with the laws of a typical enzyme-substrate complex (75), indicating that the oxidase and cytochrome c form such a complex. A study of such curves in relation to the effect of cyanide and carbon monoxide led to the conclusion that the action of these inhibitors on the reaction was concerned with the oxidase component (75). On the other hand, in the presence of an excess of cytochrome c the velocity of oxidation of hydroquinone (correcting for autoxida- tion) was directly proportional to the amount of oxidase added. This offers a method of estimating, in arbitrary units, the cytochrome oxidase activity of tissues. From hydroquinone and p-phenylenediamine tests it appears that successive acetic acid precipitations remove the larger part of the cytochrome c and some of the cytochrome h. By two precipitations with acetic acid and a long dialysis an active oxidase preparation can be obtained which shows very little cytochrome c or h, al- though the oxidase activity is likewise greatly diminished. Determination and Distribution of Cytochrome c and Cytochrome Oxidase Junowicz-Kochalaty and Hogness (76) have developed a method for estimating cytochrome c in tissues. Relatively large amounts (100 grams of tissue) are worked up through the initial steps of Keilin's isolation procedure (5) to the point where traces of hemo- globin and myoglobin are the principal colored impurities. The cytochrome c is then measured spectrophotometrically. The use of measurements at three wave lengths permits of calculations to cor- rect for the hemoglobin and myoglobin. These authors found pigeon breast muscle and beef heart muscle high in cytochrome c, tumor tissue very low. Stotz (77) has developed a method for determining cytochrome c in rat tissues. The ground tissue is extracted with trichloracetic acid, the extract neutralized to eliminate further inactive protein, and the cytochrome precipitated by phosphotungstic acid. After solution in dilute ammonia, the phosphotungstate is eliminated with barium. The final solutions are tested manometrically for their power to accelerate oxidation of hydroquinone by a heart muscle oxidase prep- aration. A calibration curve must be prepared, pure cytochrome c being used. CYTOCHROMES 167 In applying this test to various rat tissues it was found that agree- ment between animals was reasonably good; the order of activity of the tissues was the same in all the rats studied. The average results of the cytochrome c estimations in the tissues of ten rats are re- corded in Table 2. Table 2.— Comparison of cytochrome oxidase and cytochrome c activities of rat tissues (Reference 77) Oxidase Cytochrome c Units per mg. dry ' mg. per g. tissue dry tissue Heart 9.7 Heart 2.34 Kidney 4.7 Kidney 1.36 Brain 3.5 Skeletal muscle 0 . 68 Skeletal muscle 2.3 Brain 0.35 Liver 1.7 Liver 0 . 24 Spleen 1.6 Spleen 0.21 Lung 1.3 Lung 0.14 Testis 1.1 Embryo (early) 0.03 Diaphragm muscle 0.72 Embryo (late) 0.18 Large intestine 0.36 Tumor R-256 0.02 Embryo (early, late) 1.1 Tumor R-39 0.03 Tumor R-256 2.9 Tumor spontaneous 0.01 Tumor spontaneous 2.4 Most striking is the low cytochrome c content of embryos and the tumors studied. It may be noted that the cytochrome c content of the embryos increases just before birth. With the present knowl- edge of the importance of cytochrome c in respiration, the low c content of these tissues may be at least one factor responsible for the aerobic glycolysis of these tissues. Potter and Dubois (78) have likewise developed a micro method for estimating cytochrome c, the actual deteiTnination being made with a photoelectric spectrophotometer; the light absorption is measured before and after specific enzymatic reduction with succi- nate. The values they obtained with rat tissues are in remarkably good agreement with those recorded by Stotz. They have found a low c content in several tumors thus far studied. The manometric estimation of oxidase activity in rat tissues may be carried out on ground dialyzed tissue in the presence of excess cytochrome with hydroquinone as a substrate (77). Schultze (49) uses in addition semicarbazide, which maintains a constant rate of oxidation over a longer period, a helpful modification. 168 A SYMPOSIUM ON RESPIRATORY ENZYMES The distribution of oxidase in rat tissues is also illustrated in Table 2. Perhaps the most interesting thing about these data is the fact that the oxidase activity of the tissues parallels quite closely their cytochrome c content, although in the case of the tumors and em- bryonic tissue there is no lack of cytochrome oxidase. This paral- lelism may indicate a close chemical similarity between, or common origin of, cytochrome c and cytochrome oxidase. A decrease in oxidase activity has been noted in various tissues of the rat in anemia (49, 50) and in guinea pig tissues during acute scurvy (79). Physiological Functioning of the Cytochrome System Flexner and Stiehler (80, 81) have made some very interesting observations on the funct. nal development of the cytochrome sys- tem. By histochemical methods they studied the changes in the chorioid plexus of the fetal p'g, especially during the phase when the spinal fluid changes from an ultra-filtrate to a secretion. Oxidase activity was measured by blue-staining with dimethyl p-phenylene- diamine and alpha-naphthol, a test which of course measures the combined oxidase and cytochrome activity. Oxidation-reduction potentials were estimated by introduction of oxidation-reduction in- dicators intravascularly or supravitally. It was found that previous to the secretory phase the concentration of "indophenol oxidase" was the same in the epithelium and stroma and that there was no potential difference between the two. In the secretory phase, how- ever, the oxidase was concentrated in the epithelium and a poten- tial difference developed, more positive in the epithelium. It was concluded that the functional changes occurring with the onset of secretion are correlated with the potential difference set up as a result of the selective development of the oxidase (cytochrome) sys- tem in the epithelium. The selective transference of dyes across the secretory plexus was abolished by cyanide, and this was associated with a loss of the potential difference between epithelium and stroma. Flexner, Flexner, and Straus (82) have likewise studied the cyto- chrome system in the cerebral cortex of the fetal pig. During the first half of gestation p-phenylenediamine was not actively oxidized, but during the second half an active oxidation system was present. Employing the test for cytochrome oxidase (using excess cyto- chrome c), they found that there was no lack of oxidase in any phase of gestation and hence the change in p-phenylenediamine CYTOCHROMES 169 oxidation was due primarily to development of cytochrome c. Nevertheless the Qo^ of the tissue increased little during this period and at all times was over 90 per cent sensitive to 0.001 M. cyanide. Thus there is an apparent anomaly in that cyanide, the inhibitoiy action of which is generally considered to be on the cytochrome oxidase system, is inhibiting a respiratory system which does not even contain a complete cytochrome system. This is not the first instance in which the "single point of attack theory" of cyanide has been questioned. Parallel experiments with cyanide and azide (see Stannard, 83) on intact tissues have also suggested that cyanide probably combines with one or more other enzymes involved in respiration. The question arises, therefore, what is the nature of the "oxygen-activating" system in the absence of cytochrome c? If cyanide can also block an enzyme systeoi near the "dehydrogenase end" of respiration, then flavin enzymes as well as cytochrome b (which is relatively cyanide-stable aij^cX autoxidizable) become pos- sibilities for this role, despite the fact that their oxidation is gener- ally considered to be through the oxidase system. A somewhat similar situation arises in Stannard's experiments (83) on the "resting" and "activity" oxygen consumption of frog muscle. It was found that the resting metabolism was insensitive to azide, although sensitive to cyanide. On the other hand, the activity metabolism (caffeine or electrical stimulation) was greatly inhibited by azide. When the azide concentration in the stimulated prepara- tion was increased, the inhibition *came to an abrupt stop, leaving a respiration equal to that of the resting state. The latter finding excludes the possibility that the azide stability of the resting respira- tion can be explained by failure of the inhibitor to penetrate or its absence in a form capable of exerting its typical inhibition on the oxidase system. Stannard (84, 85), who has studied the anaerobic glycolysis of frog muscle, does not believe that the azide insensi- tivity of resting muscle can be explained by the lack of substrate for "saturation" of the oxidizing enzymes. He has concluded that the oxygen transfer for the resting metabolism and for the extra metab- olism resulting from stimulation are due to different systems. The extra metabolism due to stimulation was apparently by way of the cytochrome system, but the nature of the "oxygen-activating" en- zymes of the resting metabolism remain unknown. Recently Korr (30) has carried out experiments on the metabolism of slices of mammalian tissues during rest and during a stimulated phase. Salivary glands stimulated by acetylcholine or adrenalin, 170 A SYMPOSIUM ON RESPIRATORY ENZYMES pancreas by secretin, and myometrium by oxytocin all show that large increases of respiration attend the change from a state of rest to one of activity. The finding of azide-stable and relatively cyanide-stable respiration, combined with spectroscopic observa- tion of the cytochrome bands, indicated that the resting metabolism was not proceeding through the cytochrome system, but that the extra metabolism that followed stimulation was mediated through this system. It is important to note that in the resting state the cytochrome system was nevertheless "available" for the oxidation of p-phenylenediamine. It was postulated that the cytochrome system becomes "geared" or "linked" to the substrate-dehydrogenase system by an agent capable of reducing the cytochrome that is made avail- able upon stimulation of the cells. Again there is no clue to the nature of the oxidizing enzymes that replace the cytochrome sys- tem in the resting cells. Perhaps the most serious obstacle to the immediate acceptance of Stannard's and Korr's conclusions is the possibility that at the low levels of oxygen consumption the oxidase system is only partially saturated with substrate and hence considerable amounts could be blocked by inhibitors without affecting the oxygen consumption of the tissue. Such a view has been expressed by Warburg (86), and Commoner (87) has actually demonstrated that cyanide inhibition of yeast respiration is dependent on substrate respiration. The early work of Keilin showed that whereas in the resting muscle the cytochromes were oxidized, when it became active the bands of the reduced cytochromes appeared. Thus during activity the ratio of reduced to oxidized cytochrome c is higher. It is known that reduced cytochrome c forms a complex with the oxidase (74, 75). Since under conditions of activity the oxidase is relatively more saturated with its substrate (reduced cytochrome c), the respiration of this system should, according to the "under-saturation" concept, be more sensitive to the same concentration of oxidase inhibitor. The "activity" respiration should then be totally sensitive to azide if azide only affects the rate of oxidation of the cytochrome. But Stannard's experiment demonstrated that during activity only the extra oxygen consumption caused by the activity was azide- sensitive. The experiments of Stotz, Altschul, and Hogness (75) on the rela- tions of the oxidase and of cytochrome c on hydroquinone oxidation showed that the rate of oxidation was a function of both compo- nents. Thus at a fixed concentration of reduced cytochrome c (oxi- CYTOCHROMES 171 dase under-saturated) cyanide does produce an inhibition of hydro- quinone oxidation. But if the cytochromes are chiefly in the oxidized state, as they are in the resting tissue, such a decrease in oxygen consumption could be compensated for by increased reduction of the cytochromes. It is therefore important to know whether in the experiments of Stannard and of Korr azide does or does not pro- duce an increased state of reduction of the cytochromes in the resting tissue. The differences between the characteristics of the respiration of the eggs of various species before and after fertihzation have been extensively studied and have been related to the function of the cytochrome system. The results are at present very difficult to inter- pret, largely because of the tests employed for detecting the oxidase or the cytochrome and because of the problem of permeability of the cells to inhibitors. For example, it appears that in the un- fertilized eggs of the sea urchin cytochrome has never been ob- served spectroscopically, and its respiration is not inhibited by cyanide or azide. After fertilization, however, the respiration is typical of one proceeding through the cytochrome system (see Sha- piro, 88). Korr believes the respiration of the unfertilized sea urchin egg to be another case of functional inactivity of the cytochrome system, since these eggs can oxidize p-phenylenediamine (see Dis- cussion in Shapiro, 88). Allen (89), on the other hand, using grass- hopper eggs, has shown that the activity metabolism is definitely connected with the new development, rather than with "gearing," of the cytochrome system. A recent paper by Krahl, Keltch, Neubeck, and Clowes (90) on the cytochrome system of sea urchin eggs demonstrates that there is a complete absence of cytochrome c in the unfertilized eggs, but no lack of cytochrome oxidase. On the other hand, even in the fertilized eggs, which are relatively more cyanide- and azide-sensitive, they were still unable to detect any cytochrome c. The establishment of the presence of cytochrome oxidase in the unfertilized eggs, al- though it does not constitute proof, is strong evidence that this com- pound functions in the respiration of these cells. The authors favor the hypothesis suggested earlier (91) and described by Ball (page 29) that the action of cyanide and azide on cellular respiration may be related to the formation of a complex with the oxidase possessing a lower potential than the original oxidase (92, 93). The sensitivity or stability of a respiration to azide or cyanide then depends on the potential of the oxidase-inhibitor complex formed and the potentials 172 A SYMPOSIUM ON RESPIRATORY ENZYMES of the available reducing systems. With a given lower potential of the oxidase-inhibitor complex, the lower the potential of the reduc- ing system the less sensitive would be the respiration to the inhibitor. On this basis, a system in which cytochrome c (with a relatively high potential) is the normal reducing agent of the oxidase would be expected to show a high sensitivity to oxidase inhibitors. In the absence of cytochrome c this theory provides an explanation of cyanide or azide stability even when the cyanide may be combining with the oxidase component. Such a theory deserves consideration in the interpretation of all experiments purporting to demonstrate the "by-passing" or non-functioning of the cytochrome-cytochrome oxidase system. The frontiers of research in the cytochrome problem therefore ap- pear to consist on the biochemical side in the isolation and proper- ties of the individual components, and on the physiological side in the mode of function or non-function of this system in various states of metabolism of tissues. REFERENCES 1. MacMunn, C. a., J. Physiol., 8, 57 (1887). 2. Keilin, D., Proc. Roy. Soc. (London), B QH, 312 (1925). 3. Keilin, D., Proc. Roy. Soc. (London), B 100, 129 (1926). 4. Theorell, H., Biochem. Z., 2%S, 207 (1936). 5. Keilin, D., and Hartree, E. F., Proc. Roy. Soc. (London), B 122, 298 (1937). 6. Theorell, H., and Akesson, A., Science, dO, 67 ( 1939). 7. Zeile, K., and Reuter, F., Z. physiol. Chem., 221, 101 (1933). 8. Stotz, E., unpublished experiments. 9. Keilin, D., and Hartree, E. F., Proc. Roy. Soc. (London), B 12Q, 277 (1940). 10. Keilin, D., and Hartree, E. F., Proc. Roy. Soc. (London), B 127, 167 (1940). 11. Stern, K. G., Discussion, Symposia on Quantitative Biology, Cold Spring Harbor, V77, 119 (1939). 12. Keilin, D., and Hartree, E. F., Proc. Roy. Soc. (London), B 122, 298 (1937). 13. Keilin, D., Ergebnisse d. Enzymforschung, 2, 239 (1933). 14. Altsciiul, a. M., and Hogness, T. R., J. Biol. Chem., 124, 25 (1938). 15. Hogness, T. R., Zscheile, F. P., Jr., and Sidwell, A. E., Jr., J. Phys. Chem., 41, 379 (1937). 16. Potter, V. R., J. Biol. Chem., i37, 13 (1941). 17. Hill, R., and Keilin, D., Proc. Roy. Soc. (London), B 107, 286 (1930). 18. Zeile, K., and Piutti, P., Z. physiol. Chem., 2m, 52 (1933). 19. Theorell, H., Biochem. Z., 2^8, 242 (1938). 20. Theorell, H., Biochem. Z., ^01, 201 (1939). 21. Zeile, K., and Meyer, H., Z. physiol. Chem., 262, 178 (1939). 22. CooLusGE, T. B., J. Biol. Chem., 0^, 755 (1932). 23. Green, D. E., Proc. Roy. Soc. (London), B 114, 423 (1934). CYTOCHROMES 173 24. WuRMSER, R., and Filitti-Wurmser, S., Compte Rend. Soc. Biol., 127, 471 (1938). 25. Stotz, E., Sedwell, A. E., Jr., and Hogness, T. R., J. Biol. Chem., 124, 11 (1938). 26. Ball, E. C, Biochem. Z., 295, 262 (1938). 27. Yakushiji, E., and Mori, T., Acta Phytochim., 10, 113 (1937). 28. Yakusizi, E., and Okunuki, K., Proc. Imp. Acad. (Tokyo), 17, 38 (1941). 29. Baumberger, J. P., Symposia on Quantitative Biology, Cold Spring Harbor, VII, 195 (1939). 30. KoRR, I. M., Symposia on Quantitative Biology, Cold Spring Harbor, VII, 74 (1939). 31. Warburg, O., Biochem. Z., 152, 479 (1924). 32. Warburg, O., Biochem. Z., 177, All (1926). 33. Warburg, O., Biochem. Z., 189, 354 (1927). 34. Warburg, O., and Negelein, E., Biochem. Z., 214, 64 ( 1929). 35. KuBOwiTZ, P., and Haas, E., Biochem. Z., 255, 247 (1932). 36. Warburg, O., and Negelein, E., Biochem. Z., 244, 9 (1932). 37. Warburg, O., and Negelein, E., Biochem. Z., 262, 237 (1933). 38. Keilin, D., Nature, 132, 783 (1933). 39. FujiTA, A., and Kodama, T., Biochem. Z., 273, 186 (1934). 40. Keilin, D., Proc. Roy. Soc. (London), B 104, 206 (1928). 41. Warburg, O., and Negelein, E., Biochem. Z., 262, 237 (1933). 42. Warburg, O., Negelein, E., and Haas, E., Biochem. Z., 266, 1 (1933). 43. Warburg, O., and Haas, E., Naturwissenschaften, 22, 207 (1934). 44. Melnick, J. L., Science, 94, 118 (1941). 45. KuBowiTz, F., Biochem. Z., 292, 221 (1937). 46. Keilin, D., and Mann, T., Proc. Roy. Soc. (London), B 125, 187 (1938). 47. Cohen, E., and Elvehjem, C. A., J. Biol. Chem., 107, 97 (1934). 48. YosHiKAWA, H., J. Biochem. (Japan), 25, 627 (1927). 49. ScHULTZE, M. O., J. Biol. Chem., 129, 729 (1939). 50. ScHULTZE, M. O., J. Biol. Chem., 138, 219 (1939). .51. Graubard, M., Am. J. Physiol., 131, 584 (1941). 52. Altschul, a. M., Abrams, R., and Hogness, T. R., J. Biol. Chem., 130, 427 (1939). 53. Altschul, A. M., Abrams, R., and Hogness, T. R., J. Biol. Chem., 136, 777 (1940). 54. Stern, K. C, Symposia on Quantitative Biology, Cold Spring Harbor, VII, 312 (1939). 55. Straub, F. B., Biochem. J., 33, 787 (1939). 56. v. EuLER, H., and Hellstrom, H., Z. physiol. Chem., 260, 163 (1939). 57. Alt, H. L., Biochem. Z., 221, 498 (1930). 58. Haas, E., Naturwissenschaften, 22, 207 (1934). 59. Hopkins, F. C, Lutwak-Mann, C, and Morgan, E. J., Nature, 143, 556 (1939). 60. Stern, K. C, and Melnick, J. L., Nature, 144, 330 (1939). 61. Horecker, B. L., Stotz, E., and Hogness, T. R., J. Biol. Chem., 128, 251 (1939). 62. Warburg, O., and Christian, W., Biochem. Z., 254, 438 (1932); 257, 492 (1933). 63. Warburg, O., and Christian, W., Biochem. Z., 266, 377 (1933). 64. Theorell, H., Biochem. Z., 288, 317 (1936). 65. Haas, E., Horecker, B. L., and Hogness, T. R., J. Biol. Chem., 136, 747 (1940). 66. LocKHART, E. E., and Potter, V. R., J. Biol. Chem., 137, 1 (1941). 174 A SYMPOSIUM ON RESPIRATORY ENZYMES 67. Adler, E., v. Euler, H., and Hellstrom, H., Arkiv. Kemi, Mineral. Geol., 12B, No. 38 (1937). 68. Green, D. E., and Dewan, J. C, Biochem. J., 32, 626 (1938). 69. Hawthorne, J. R., and Harrison, D. C, Biochem. J., 33, 1573 (1939). 70. CoRRAN, H. S., Green, D. E., and Straus, F. B., Biochem. J., 33, 793 (1939). 71. Keilin, D., Proc. Roy. Soc. (London), B 106, 418 (1930). 72. Stotz, E., Harrer, C. J., Schultze, M. O., and King, G. G., J. Biol. Chem., 122, 407 (1938). 73. Keilin, D., and Hartree, E. F., Proc. Roy. Soc. (London), B 125, 171 (1938). 74. Stotz, E., Sidwell, A. E., and Hogness, T. R., J. Biol. Chem., 124, 733 (1938). 75. Stotz, E., Altschul, A. M., and Hogness, T. R., J. Biol. Ghem., 124, 745 (1938). 76. Junowicz-Kochalaty, R., and Hogness, T. R., J. Biol. Ghem., 129, 569 (1939). 77. Stotz, E., J. Biol. Ghem., 131, 555 (1939). 78. Potter, V. R., and Dubois, K. P., J. Biol. Ghem., 140, Scientific Proceedings XXXV, cii (1941). 79. Harrer, G. J., and King, G. G., J. Biol. Ghem., 138, 111 (1941). 80. Stiehler, R. D., and Flexner, L. B., J. Biol. Ghem., 126, 603 (1938). 81. Flexner, L. B., and Stiehler, R. D., J. Biol. Ghem., 126, 619 (1938). 82. Flexner, J. B., Flexner, L. B., and Strauss, W. L., Jr., Proc. Am. Physiol. Soc, 1941, p. 90. 83. Stannard, J. N., Symposia on Quantitative Biology, Gold Spring Harbor, VII, 394 (1939). 84. Stannard, J. N., Am. J. Physiol., 122, 379 (1938). 85. Stannard, J. N., Am. J. Physiol., 126, 196 (1939). 86. Warburg, O., Biochem. Z., 189, 354 (1927). 87. Commoner, B., J. Gell. Gomp. Physiol., 13, 121 (1939). 88. Shapiro, H., Symposia on Quantitative Biology, Gold Spring Harbor, VII, 406 (1939). 89. Allen, T. H., J. Gell. Gomp. Physiol., 16, 149 (1940). 90. Krahl, M. E., Keltch, A. K., Neubeck, G. E., and Glowes, G. H. A., J. Gen. Physiol., 24, 597 (1941). 91. Ball, E. G., in discussion of the paper by E. S. G. Barron, Symposia on Quantitative Biology, Gold Spring Harbor, VII, 154 (1939). 92. Barron, E. S. G., J. Biol. Ghem., 121, 285 (1937). 93. Glark, W. M., Taylor, J. H., Davies, T. H., and Vestling, G. S., J. Biol. Ghem., 135, 543 (1940). Phosphorylation of Carbohydrates CARL F. CORI Washington University School of Medicine, St. Louis THE METABOLISM of Carbohydrate in animal tissues is made up of a series of enzymatic reactions in which phosphate plays an essential role. What is usually referred to as the phosphate cycle can be divided into four parts: the uptake of inorganic phosphate, the intramolecular migration of phosphate groups, the transfer of phosphate groups from one molecule to another (transphosphoryla- tion), and the regeneration of inorganic phosphate. Uptake of Inorganic Phosphate The only reaction leading to the uptake of inorganic phosphate that is definitely known to be enzymatic is the phosphorylation of glycogen and starch. The uptake of inorganic phosphate which is associated with the oxidation of phosphoglyceraldehyde is presum- ably non-enzymatic, and the same may be true of the uptake of inorganic phosphate associated with the oxidation of pyruvate. In the phosphorylation of glycogen the C— O— C bond of the 1—4 glucosidic chain is replaced by the C— O— P bond of glucose-1- phosphate. This reaction is reversible, and from the position of the equilibrium it may be calculated that the change in free energy is very small. This may be interpreted to mean that the ester linkage in glucose-1-phosphate is nearly equivalent to the glucosidic linkage in the large polysaccharide molecule. When phosphate is replaced by water, as in the hydrolysis of glycogen or starch by diastase, the reaction seems to be largely irreversible; that is, the end products of diastatic activity, maltose and glucose, even when added to diastase in high concentrations, are not polymerized to glycogen or starch. Glucose, in order to undergo enzymatic polymerization, must first be phosphorylated. Reversi- bility is thus clearly connected with the introduction of a phosphate group into the polysaccharide molecule. The position of the equi- librium at physiological pH is about 77 per cent to the glycogen side (see Table 2), and is determined by the concentration of the divalent ions of orthophosphate and glucose-1-phosphate. Since the 175 xs^^-^iy 176 A SYMPOSIUM ON RESPIRATORY ENZYMES second dissociation constant of these two acids is different, the posi- tion of the equihbrium changes with pH (1, 2). The enzymatic phosphorylation of glycogen and starch is of in- terest from the standpoint of the configuration of these polysac- charide molecules. There is considerable evidence, based on the hydrolysis of methylated starch and glycogen, that these polysac- charides contain other than the prevalent 1—4 glucosidic linkages. At points at which a branching of chains occurs, a 1—6 glucosidic linkage has been postulated. Starch appears to be made up of relatively straight chains consisting of 24 to 30 glucose units; this explains its ability to assume a crystalline structure and to exhibit well-defined x-ray diffraction patterns. Glycogen seems to be made up of relatively short chains (consisting of 12 to 18 glucose units), with many branchings which give it the properties of greater solu- bility and lack of crystallizability. Judging from the properties of various plant starches, transitions exist between these two extremes. These various forms of polysaccharide raise the problem of enzyme specificity. The questions are whether one and the same enzyme splits (or builds up) both the 1—4 and the 1—6 glucosidic linkages and what determines the special configuration of the polysaccharide synthesized. It is to be noted here that phosphorylases cannot poly- merize glucose-1-phosphate unless a small amount of polysaccharide is added to prime the reaction (1). However, the nature of the poly- saccharide synthesized seems to be solely determined by the type of phosphorylase used and not by the nature of the activating poly- saccharide. For example, muscle phosphorylase, when primed with hver glycogen, synthesizes a typical starch in vitro, and liver phos- phorylase, when primed with plant starch, synthesizes glycogen (3). Important also in this connection is the fact that the same enzyme, muscle phosphorylase, can synthesize both starch and glycogen, the former in vitro and the latter in the intact cell. The difference in activity of the muscle phosphorylase in the two situations has not been explained, but it suggests that unknown environmental factors and perhaps the physical state of the enzyme have something to do with the nature of the polysaccharide which is formed. As has been stated, the phosphorylation of glycogen is an enzy- matic reaction which provides for the entrance of inorganic phos- phate into the phosphate cycle. All other reactions leading to the uptake of inorganic phosphate are linked with oxidations. The reac- tion between orthophosphate and phosphoglyceraldehyde during oxidation of the latter to phosphoglyceric acid has been elucidated PHOSPHORYLATION OF CARBOHYDRATES 177 by Warburg and his school (4). Lipmann (5) has described a bac- terial enzyme system in which inorganic phosphate is taken up and acetylphosphate is formed as an intermediate of pyruvate oxidation. In both cases the phosphate group taken up during oxidation is transferred by the adenylic acid system to suitable phosphate ac- ceptors. The oxidation of pyruvate in various animal tissues is also linked with the uptake of inorganic phosphate, and the same has been shown to be true for certain steps of the citric acid cycle, particularly for the oxidation of succinic to fumaric acid (6). The primary phos- phorylation products formed in these cases have not been identified. In dialyzed and suitably supplemented tissue dispersions or extracts, the inorganic phosphate taken up during oxidation of pyruvate is transferred by the adenylic acid system to glucose, which is con- verted to hexosediphosphate. It was noted, however, that when no glucose was added to the system, a small amount of an easily hydrolyzable phosphorus compound was formed. This compound has recently been identified in our laboratory in collaboration with Dr. Ochoa. In large-scale experiments with dialyzed rat liver dis- persion and with glutamate, pyruvate, or succinate as oxidizable substrate we have isolated inorganic pyrophosphate as the crystal- hne sodium salt. The orthophosphate which disappeared corres- ponded in amount to the pyrophosphate formed. It is too early to evaluate the significance of this observation. There is the possibility that the pyrophosphate group occurs in some or- ganic combination which is split during the process of isolation. It is fairly certain, however, that the pyrophosphate does not originate from adenosinetriphosphate, since we have found no enzyme in liver preparations which splits added adenylpyrophosphate to adenylic acid and inorganic pyrophosphate. Another possibility is that the pyrophosphate group has nothing to do with the primary phos- phorylation product which is formed during the oxidation of the substrate, but is the result of phosphorylation of orthophosphate. The phosphorylation of glucose and other phosphate acceptors which is connected with the oxidation of pyruvate has been termed "aerobic phosphorylation" to signify that the energy for the forma- tion of the phosphate bond comes from oxidations. Table 1 illustrates the quantitative relationship between oxygen consumption and phosphorylation. This experiment shows that the dialyzed heart extract supple- mented with magnesium ions, inorganic phosphate, and a trace of 178 A SYMPOSIUM ON RESPIRATORY ENZYMES adenylic acid has practically no basal oxygen consumption and that addition of 2 micromoles of succinic acid, a catalytic amount, has very little effect on oxygen consumption. The addition of 50 micro- moles of glucose had a very marked effect on oxygen consumption, Table 1.— Glucose balance in dialyzed heart muscle extract* (1 cc. of extract supplemented with Mg"*"*" ions, inorganic phosphate, and a trace of adenylic acid. Incubated 60 minutes at 37° C. All values are expressed in micromoles.) Addition Glucose Oxygen Phosphate Total glucose disappearing consumed esterified accounted for None 1.5 2 succinate 2.6 50 glucose + 2 succinate 24.4 22.3 36.1 3.5+18 = 21.5 *J. Biol. Chem., 137, 343 (1941). and there can be no doubt that glucose was the substrate under- going oxidation. Detennination of the respiratory quotient in other experiments showed that it was unity for added glucose and 1.25 for added pyruvate. It is to be noted that an oxygen consumption of 22 micromoles corresponds to one-sixth as much glucose, that is to 3.5 micromoles, while the glucose which actually disappeared ac- cording to sugar analysis was 24.4 micromoles. The glucose which disappeared without being oxidized was largely recovered as hexose- diphosphate, 36 micromoles of phosphate esterified corresponding to 18 micromoles of glucose. Lactic acid formation was not deter- mined. The balance indicates that for each mole of glucose oxidized an additional 6 moles of glucose disappear, 5 of which are present as phosphate ester. This means that about one atom of phosphate is esterified for each atom of oxygen consumed. Ochoa (7) and Belitzer and Tsibakova (8) observed even higher ratios, namely, from 2 to 3 atoms of phosphate esterified for each atom of oxygen consumed. This would indicate that not only the primary removal of hydrogen from the substrate but also one or even two subsequent hydrogen transfers over intermediate catalysts may cause phosphorylation. Aerobic phosphorylation is a mechanism by which oxidative energy is utilized in the cell. The oxidative energy is converted into phosphate bond energy, to use Lipmann's (5) terminology, and the adenylic acid system serves as the mediator of this energy transfer. When glucose is the phosphate acceptor, the system, once started, is PHOSPHORYLATION OF CARBOHYDRATES 179 self -perpetuating. The phosphorylation of glucose enables it to un- dergo oxidation by way of triosephosphate and pyruvate, and this oxidation causes further phosphorylation of glucose, thus providing new substrate for oxidation and so on. During recovery of muscle from work, oxidative energy is also converted into phosphate bond energy; that is, the phosphocreatine which breaks down during muscular contraction is reformed largely at the expense of oxidation, and the phosphorylation of glucose which is supplied by the blood stream provides the necessary sub- strate for the resynthesis of the glycogen lost during contraction. Intramolecular Migration of Phosphate Groups The first enzymatic reaction of this type was described by Meyer- hof and Kiessling (9), namely, the conversion of glyceric acid-3- to glyceric acid-2-phosphate. This reaction was shown to be reversible, and the assumption that one is dealing with an intramolecular mi- gration of the phosphate group was confirmed by the use of radio- active phosphorus. Another reaction of this type is the conversion of glucose-1- to glucose-6-phosphate. This reaction was at first regarded as irre- versible, but more recent work has shovsni that it can be reversed under suitable experimental conditions; that is, glucose-6-phosphate can be converted to glucose-1-phosphate and then to glycogen (10). To study the equilibrium of this reaction it was necessary to separate the enzyme which catalyzes this reaction from interfering enzymes which upset the equilibrium by acting either on glucose-1- or glu- cose-6-phosphate. With such a purified enzyme preparation 94 per cent of added glucose-1-phosphate is converted to an ester which was isolated and identified as glucose-6-phosphate. Fructose-6-phosphate was absent, because Lohmann's enzyme (11), which catalyzes the reversible reaction between glucose-6- and fructose-6-phosphate, had been removed. Conversely, when pure glucose-6-phosphate was added to the enzyme, 6 per cent of glucose-1-phosphate was formed; that is, the same equilibrium was reached from either side. From the equihbrium constant (K = 15.7 at pH 7 and 25° C.) it may be calculated that the change in standard free energy amounts to about —1600 calories. When barium ions and phosphorylase were added to this system, glucose-6-phosphate was converted to glycogen. As shown in Table 2, the position of the equilibrium of the second reaction is unfavor- able for glycogen synthesis, because only a smaU amount of glucose- 180 A SYMPOSIUM ON RESPIRATORY ENZYMES Table 2.— Position of equilibria at pH 7 and 25° Glycogen + inorganic phosphate (77 per cent)<=^gIucose-l-phosphate (23 per cent) (1) Glucose-1-phosphate (6 per cent) i=^glucose-6-phosphate (94 per cent) (2) Glucose-6-phosphate (70 per cent)<=^fructose-6-phosphate (30 per cent) (3) 1-phosphate is formed. The overall reaction cannot progress very far to the glycogen side, but when barium ions are added, which cause precipitation of the inorganic phosphate set free when re- action 1 goes to the left, up to 40 per cent of added glucose-6- phosphate can be converted to glycogen. Such an experiment is shown in Table 3. Table 3.— Glycogen formation from barium salt of glucose-6- phosphate with dialyzed muscle extract (Extract + Mg++ ions + catalytic amounts of glycogen and adenylic acid. Incubated at 30° C.) Type of extract Time of incubation 6-ester converted to glycogen minutes per cent Original 60 5.0 180 15.1 Concentrated sevenfold 10 3.1 60 9.9 180 33.4 Concentrated, but without addition of adenylic acid and glycogen 180 1.2 Muscle extract, prepared in the usual manner and dialyzed free of inorganic phosphate, shows only a slight activity when the barium salt of glucose-6-phosphate is added. When the same extract was concentrated sevenfold by freezing and by drying in vacuo, the activity was markedly increased, since up to 33 per cent of the added glucose-6-phosphate was converted to glycogen. The glycogen formation was measured by (1) the amount of inorganic phosphate set free, which precipitates as the barium salt because of the addition of barium ions; (2) the increase in glycogen, which is somewhat less than the increase calculated from the inorganic phosphate because of the presence of diastase in the extract; and (3) the color reaction with iodine, which is blue for the polysaccharide formed by muscle phosphorylase. As a control procedure, the addition of adenylic acid and of the glycogen necessary for the priming of the reaction PHOSPHORYLATION OF CARBOHYDRATES 181 was omitted, and it may be seen that no conversion of glucose-6- phosphate to glycogen took place. Colowick and Sutherland have also been able to convert glucose to glycogen in vitro by the addition of hexokinase and adenosinetri- phosphate to the concentrated muscle extract. It can be shown that in this case glucose is first converted to glucose-6-phosphate at the expense of the labile phosphate groups of adenosinetriphosphate. Glucose-6-phosphate is thus an important intermediate of carbo- hydrate metabolism; it is formed from glycogen via glucose-1- phosphate and it can also be formed by direct phosphorylation of glucose in position 6. The latter reaction is the so-called hexokinase reaction which was described by Meyerhof (12) and v. Euler and Adler (13). Transphosphorylation All known transphosphorylation reactions involve adenosinemono- or adenosinediphosphate; these nucleotides act in catalytic amounts as acceptors of phosphate from such substances as phosphopyruvate, acetylphosphate, and 1,3-diphosphoglycerate and are thus converted to adenosinedi- and triphosphate respectively. These polyphosphates then serve in a second enzyme reaction as phosphate donors to a number of organic molecules, such as glucose, fructose, mannose, fructose-6-phosphate, glycerol, creatine, adenosine, and probably others. The enzymes which transfer the phosphate group from the polyphosphates are specific with respect to the acceptors. For ex- ample, yeast contains enzymes that phosphorylate glucose, fructose, mannose, and adenosine, but no enzyme that phosphorylates crea- tine. Extract of skeletal muscle has little or no hexokinase activity. Since this can hardly be true of intact muscle, it may be due to destruction, inhibition, or poor extractability of the enzyme. Other tissues, such as brain, heart, liver, kidney, and retina yield active extracts. The hexokinase in these tissue extracts has not been separated from other enzymes, but in the case of yeast such a separation has been effected. Colowick and Kalckar (14) have recently studied the reaction between adenosinetriphosphate and glucose with purified yeast hexokinase. Only one of the labile phosphate groups of adenosine- triphosphate is transferred to glucose; that is, the reaction products that were identified are adenosinediphosphate and glucose-6-phos- phate. When adenosinediphosphate is substituted for adenosine- triphosphate, no reaction with glucose takes place, but when a heat- 182 A SYMPOSIUM ON RESPIRATORY ENZYMES stable protein of muscle is added to the system, hexokinase is able to transfer the labile phosphate group of adenosinediphosphate, the reaction products in this case being adenylic acid and glucose-6- phosphate. This is illustrated in Tables 4 and 5. Table 4.— Reaction of adenosinetriphosphate with glucose (Hexokinase and Mg++ in all samples. T = 30°C.) Time Po Pio Pio-Po Addition (min.) (7) (7) (7) Adenosinetriphosphate 0 4.0 68.0 64.0 Adenosinetriphosphate 5 8.9 70.0 61.1 Adenosinetriphosphate+glucose, 2 mg 5 8.8 43.8 35.0 Adenosinetriphosphate 15 9.0 68.0 59.0 Adenosinetriphosphate+glucose, 2 mg 15 9.5 43.4 33.9 The diflFerence between the Pq and Pjo value (initial and ten-minute hydrolysis values in normal sulfuric acid) corresponds to the amount of labile phosphate added as adenosinetriphosphate. With a purified hexokinase preparation of yeast, no appreciable reaction takes place Table 5.— Necessity of heat-stable muscle protein for reaction of adenosinediphosphate with glucose (Hexokinase and Mg++ in all samples. Time, 5 minutes. T = 30°C.) Addition Pio— Po (7) AdenosineM'phosphate 51.8 Adenosine/rzphosphate-l- glucose 28.4 Adenosine2 Phosphate 2 Phosphoglycerate tt 2 Phosphopyruvate IT Lactate;=^2 Pyruvate Phosphate transfer - -8H +4 phosphate Summary This brief presentation of the phosphate cycle is of course far from a complete picture. Its elements are the individual enzymatic reactions. If an enzyme has been isolated from the tissues, if the reaction product or products have been identified, if the kinetics of the reaction and the role of coenzymes, activators, and inhibitors are known, we are, I believe, on solid ground. Several of the indi- vidual enzymatic reactions comprising the phosphate cycle have been studied in this manner; others remain to be studied in greater detail. It is also possible to combine a number of individual enzymatic reactions in the test tube and to reproduce overall effects, such as the polymerization of glucose to glycogen or the conversion of fructose to glucose. The coupling between respiration and phosphorylation has given us an insight into the mechanism of energy transfer in the cell. There is still an essential element lacking in this picture, which in the absence of a better definition might be called the regulatory function of the cell. There can be no doubt that mechanisms exist in the intact cell which regulate the rate and direction of individual enzymatic reactions and which lead to a high degree of integration PHOSPHORYLATION OF CARBOHYDRATES 189 of overall eflFects. The next approach is, perhaps, a study of the mechanisms underlying this regulatory function. The mechanism of action of a number of hormones is still obscure; all that can be said at present is that they are part of the regulatory mechanism of the cell rather than essential constituents of enzyme systems. REFERENCES 1. Com, G. T., and Com, C. F., J. Biol. Chem., 135, 733 (1940). 2. Hanes, C. S., Proc. Roy. Soc. (London), B 129, 174 (1940). 3. Baer, R. S., and Com, C. F., J. Biol. Chem., 140, 111 (1941). 4. Warburg, O., and Christian, W., Biochem. Z., 303, 40 (1939). 5. LiPMANN, F., Advances in Enzymology, 1, 99 (New York, 1941). 6. CoLowicK, S. P., Kalckar, H. M., and Com, C. F., J. Biol. Chem., 137, 343 (1941). 7. OcHOA, S., J. Biol. Chem., 138, 751 (1941). 8. Belitzer, V. A., and Tsibakova, E. F., Biokimia, 4, 516 (1939). 9. Meyerhof, O., and Kxessling, W., Biochem. Z., 276, 239 (1935). 10. Sutherland, E. W., Colowick, S. P., and Com, C. F., J. Biol. Chem., 140, 309 (1941). 11. LoHMANN, K., Biochem. Z., 262, 137 (1933). 12. Meyerhof, O., Biochem. Z., 183, 176 (1927); Natunvissenschaften, 23, 850 (1935). 13. V. EuLER, H., and Adler, E., Z. physiol. Chem., 235, 122 (1935). 14. Colowick, S. P., and Kalckar, H. M., J. Biol. Chem., 137, 789 (1941); 140, xxix (1941). 15. Com, G. T., and Cori, C. F., J. Biol. Chem., 116, 119, 129 (1936). 16. Kalckar, H. M., Biochem. J., 33, 631 (1939). 17. PiLLAi, R. K., Biochem. J., 32, 1087 (1938). 18. Parnas, J. K., Handbuch der Enzymologie, II, 902 (Leipzig, 1940). 19. Needham, J., Shen, S. C, Needham, D. M., and Lawrence, H. S. C, Nature, 147, 766 (1941). 20. Geiger, a., Biochem. J., 34, 465 (1940). 21. OcHOA, S., J. Biol. Chem., 141, 245 (1941). Discussion on Phosphorylation H. M. KALCKAR Washington University, St. Louis, Chairman Dr. Kalckar: It seems advisable to begin this discussion on phosphorylation with a brief survey of the major present-day problems in this field. The mechanism of the compulsory coupling between oxidation of triose and phosphorylation of adenosine nucleotides was solved when Warburg and Christian isolated and crystallized the catalyst of the phosphotriose oxidation and Negelein and Bromel isolated the 1,3-diphosphoglyceric acid. Warburg and his collaborators also demonstrated the transfer of the phosphate from the carboxyl group to adenosinediphosphate. Both the oxidation process (1) and the phosphate transfer are easily reversed reactions: O 0 0 I II II II (1) c=0-t-HO— P— 0'+ pyridinium ^ — C— O— P— 0'+ reduced pyridine I I I O' O' 0 0 o (2) — C — P — 0'+ adenosinediphosphate "'<=± — C + adenosine?r?phosphate"" 1 I I O O' O' These two reversible reactions explain completely the findings of Needham, Meyerhof, and others that the utilization of the labile phosphate in adenosinetriphosphate determines the extent of oxida- tion of phosphotriose. Lipmann's observations of the formation of acetylphosphate in the bacterial pyruvate oxidation furnished another important ex- ample of carbonyl oxidation coupled with phosphate uptake. Carbonyl oxidation is not the only type of oxidation that is coupled with phosphorylation. Oxidations of fumaric acid or even of succinic acid are coupled with phosphorylations, but the mechanism of these reactions has not been clarified. It has been mentioned that the extent of triose oxidation was auto- matically regulated by the removal of labile phosphate in adenosine- 190 DISCUSSION ON PHOSPHORYLATION 191 triphosphate and it is therefore of interest to discuss briefly the problems relating to adenylpyrophosphate utilization. The utilization of adenylpyrophosphate takes place by easily reversed reactions or by irreversible degradation. The phosphorylation of carboxylate or amidine ions represents easily reversed dephosphorylations of adenyl- pyrophosphate. The phosphorylations of the hydroxy groups of monohexoses or glycerol and the mineralization of phosphate repre- sent the irreversible degradations of adenylpyrophosphate. These irreversible strongly exergonic reactions are probably essential in the energy transformations. The phosphorylation of glucose by adenylpyrophosphate has been observed by Euler and Adler in brewer's yeast, and by Meyerhof in baker's yeast and in extracts of animal tissue. Strangely enough, the enzyme called hexokinase does not occur in muscle extracts. This may be related to the fact that the concentration of free glucose inside the muscle cell is very low. Perhaps phosphorylation of glu- cose in the skeletal muscles takes place in the so-called cell mem- brane and only there. I think that there is every reason to look for the hexokinase enzyme in the water-insoluble residue. Recently Colowick and I found that phosphorylation of glucose by adenylpy- rophosphate when adenosinerf/phosphate is the phosphate donor needed, besides hexokinase, another protein which occurred only in skeletal muscle and was therefore called myokinase. Myokinase is active in amounts smaller than one microgram per milliliter and is extremely resistant to acid treatment. I have found recently that my- okinase is also necessary for the dephosphorylation of adenosine- diphosphate in muscle tissue. The experiments were carried out with suspensions of myosin as described by Engelhardt and Ljubimova. The exact function of myokinase is not known. Dr. Johnson has suggested that myokinase catalyzes a transfer of phosphate from one molecule of adenosinediphosphate to another molecule, thus forming from two moles of adenosinediphosphate one mole of adeno- sinetriphosphate and one mole of adenosinemonophosphate. One might also consider the formation of a dinucleotide between adeno- sinetriphosphate and adenylic acid of the type that has been isolated by Kiessling and Meyerhof. Myokinase might function not only as a catalyst but also as a phosphate transfer system. The inhibition of myokinase by adenylic acid may also indicate a reversed reaction of myokinase""with adenosinediphosphate. Adenylpyrophosphate phosphorylates glucose or fructose to 192 A SYMPOSIUM ON RESPIRATORY ENZYMES hexose-6-phosphate. Sutherland, Colowick, and Cori have shown recently that this ester can be converted to hexose-1-phosphate and thus to starch or glycogen. The phosphorylation of hexosemonophosphate to hexosediphos- phate has been observed in several tissues and also in muscle extracts. This reaction is probably strongly exergonic, hke the hexo- kinase reaction. Meyerhof and Lohmann have shown that the dephosphorylation of adenylpyrophosphate to adenosinedi- or monophosphate and orthophosphate is one of the strongest exergonic reactions in bio- logical systems. The enzyme is present in large amounts in muscle extracts but the myosin fraction contains also large amounts of this enzyme, and Engelhardt and Ljubimova have found that even highly purified myosin contains adenosinetriphosphatase. The utilization of adenylpyrophosphate by these irreversible reac- tions probably determines the extent of numerous important oxida- tions. The great question is how the irreversible utilization of adenylpyrophosphate is regulated. The existence of adenylpyro- phosphate in resting muscles with very low metabolism clearly shows that the enzymes which catalyze the breakdown of pyrophosphates must be in a more or less inactivated state and are fully activated only under certain conditions. The state of structural proteins such as myosin might determine the extent of the activation of the en- zymes which catalyze the degradations of adenylpyrophosphate. Discharged structural proteins might even be enabled to use the large amount of energy liberated by pyrophosphate degradations for recharging by participating in the transfer of phosphate. We do not yet know how irreversible enzymatic reactions are regulated or how they are directed in order to transfonn chemical energy into mechanic^ energy. It has become clear, however, that the irreversible dephosphorylations of adenylpyrophosphate repre- sent some of the most important degradations in biological systems. Otto Meyerhof, University of Pennsylvania: May I discuss three of the points that have been proposed by Dr. Kalckar: 1. Because of the specific structure of the living cell, the metab- olism can be regulated differently from the metabolic processes in extracts or other preparations of dead cells. At least one such dif- ference can easily be explained by the insufiicient stability of the extracted enzyme. I refer to the accumulation of hexosediphosphate in yeast preparations in contrast to the balance of formation and DISCUSSION ON PHOSPHORYLATION 193 decomposition of this ester in living yeast. Indeed the occurrence of the same phosphorylated intermediaries in hving cells as in extracts is clearly demonstrated by the experiments of Miss Macfarlane, Werkman, Dische, and others. But the adenylpyrophosphatase, which is responsible for the fermentation of hexosediphosphate in the absence of a stoichiometric amount of phosphate acceptors, is weak- ened even by drying the yeast, more by incubation and extraction of the dried yeast, and it can be completely destroyed by precipita- tion with acetone. On the other hand, we may assume that in the living cell excess phosphorylation of sugar occurs in connection with growth. The starting point of these syntheses may be some other ester instead of hexosediphosphate; nevertheless the autocatalytical increase of this ester in a fermenting yeast extract can be taken as a good model for material growth brought about by the energy of fermentation; indeed, such an extract, containing sugar, will not start fermentation unless it is "inoculated" by a trace of hexosediphosphate, which then "grows" at the expense of sugar. 2. At least some of the experiments of different authors quoted by Dr. Lipmann seem not too reliable in regard to abnormally high oxidation quotients. I refer especially to cases with Qo, values of 0 to 1. Under such conditions the manometric method is not accurate enough, and a completely stationary state during the time of the experiment is not assured. Furthermore, it seems to me, the distribution of oxidized to re- duced cozymase in toto cannot be used as the basis for deciding how far the hydrogen transfer by means of cozymase is affected by respiration. This transfer occurs by means of the bound cozymase of specific enzymes, and the oxidative state of such a compound may well be altered by the oxygen transfer from oxidative catalysts with- out an appreciable change in the overall distribution of oxidized to reduced cozymase. 3. The interesting finding of Dr. Kalckar that the action of hexo- kinase of yeast is supplemented by a heat-stable enzyme of muscle, "myokinase," by which adenosinediphosphate transfers its labile phosphate group to glucose, points to a difference, already well known, between the enzymes in muscle and yeast preparations. While the former (extract, acetone powder, etc.) retains the ability to use adenylic acid as the phosphorylating coenzyme even after long dialysis and "ageing," in yeast preparations the reaction step, (2) P-acceptor + adenosinediphosphate ±5 P-acceptor — phos- phate + adenylic acid. 194 A SYMPOSIUM ON RESPIRATORY ENZYMES is easily destroyed, and only the first reaction step of transphos- phorylation, (1) P-acceptor + adenosinetriphosphate ±^ P-acceptor — phos- phate + adenosinediphosphate is still occurring. A more detailed study of the different enzyme proteins in these phosphorylating reactions along the Hnes suggested by Colowick and Kalckar seems highly desirable. M. J. Johnson, University of Wisconsin: The present status of our knowledge on biological phosphoryla- tions raises the question whether energy of phosphorylation is the sole form in which energy from food oxidation (or fermentation) is made available for metabolic reactions. In other words, is phos- phorylation the only device employed by the cell to utilize oxidative energy? The only answer that can be made at present is that phosphory- lation is the only mechanism we know of by which a part of the energy derivable from the burning of food material can be made available for endergonic life processes. There appears to be no a priori reason to suspect the existence of another mechanism. In fact, the simplicity of a single-mechanism hypothesis seems very attrac- tive. On the other hand, we have no reason to doubt the existence of other "energy-fixing" reactions. The question is entirely open. There is, however, at least one fragment of evidence that definitely favors the single-mechanism hypothesis. Yeast growing anaerobically in a simple medium must derive all the energy for its growth reac- tions from the conversion of glucose into ethyl alcohol and carbon dioxide. Unless our present views of the mechanism of this conver- sion are erroneous, no energy-fixing mechanism other than phos- phorylation exists in alcoholic fermentation. Therefore the only form in which energy from the fermentation process is available to the organism is energy of phosphorylation. It follows, of course, that the yeast cell is capable of utilizing, directly or indirectly, energy of phosphorylation for all its metabolic energy requirements. Hence any other energy-fixing reactions which might be postulated to occur during aerobic sugar breakdown are at least not essential for yeast growth. DISCUSSION ON PHOSPHORYLATION 195 Fritz Lipmann, Massachusetts General Hospital: On several occasions during this symposium acetylphosphate has been mentioned as an intermediate in carbohydrate breakdown. Since I have not yet published a complete account of my experi- ments, I should like to take this opportunity to summarize the evi- dence so far accumulated for the formation of acetylphosphate as an intermediate in pyruvic acid oxidation. Early in my study of the oxidation of pyruvic acid in lactic acid bacteria it was observed that inorganic phosphate was an integral part of the pyruvic acid oxida- tion system (1). In partial explanation of its necessity it was shown that with oxidation of pyruvic acid, phosphate could be transferred to adenylic acid to form adenosinepolyphosphate (2). In other words, pyruvic acid oxidation generated energy-rich phosphate bonds (terminology of ref. 3). In analogy to Negelein and Bromel's phos- phoglycerylphosphate (4), acetylphosphate was then suspected to be the phosphorylated intermediate between pyruvate and adenosine polyphosphate. In confirmation it was found that synthetical acetyl- phosphate enzymatically transferred phosphate to adenylic acid with the formation of adenosinetriphosphate (5). This observation led to a reinvestigation of the phosphate turnover in pyruvic acid oxidation. In the bacterial metabolism pyruvate is fortunately not oxidized farther than the acetate stage. Under favorable conditions, therefore, one could hope to demonstrate the accumulation of a phosphorylated precursor of acetic acid. Earlier attempts had always failed to disclose any disappearance of inorganic phosphate during the oxidation of pyruvate. Now that acetylphosphate was suspected as intermediate, its stability was studied, and it was found that its formation would have been over- looked because in the course of all the known procedures of phos- phate assay it would have been decomposed to inorganic phosphate. Therefore the problem arose of finding a method for the detennina- tion of inorganic phosphate in the presence of a compound of the stability of acetylphosphate, since this compound does not with- stand the alkalinity of the magnesia mixture tolerated by creatine- phosphate, nor the acidity of the molybdate reagent. It appeared possible, however, to completely precipitate inorganic phosphate as calcium phosphate in dilute alcohol at a pH of 8, where acetylphos- phate is stable. The use of this more delicate precipitation procedure showed indeed that large amounts of inorganic phosphate disap- peared during pyruvate oxidation with or without fluoride (6). Phosphorylation, oxygen consumption, and pyruvate disappearance 196 A SYMPOSIUM ON RESPIRATORY ENZYMES occurred in definite proportions and in good agreement with the equation: CH3 • CO • COOH + HO • PO3H2 + O - CH3 • CO • OPO3H3 + CO2 + H,0 Recently the formation of a labile phosphate compound could be demonstrated directly by the use of the method with which creatine- phosphate was originally discovered both by Fiske and Subbarow and by the Eggletons. This method consists in reading the develop- ment of the blue color against a standard of inorganic phosphate, the increase in color with time being proportional to the breakdown of the labile phosphate compound. The half-decomposition time of our phosphorylation product (and of acetylphosphate) is about one minute at room temperature. This rapid decomposition of the nat- ural as well as the synthetic product is due not merely to the acidity of the solution but largely to a catalytic effect of molybdate. Re- cently silver fractions containing the labile phosphate have been obtained from trichloroacetic acid extracts. Although, as should be emphasized, pure preparations of the phosphorylation product have not yet been obtained,* the reported results may be taken as fair evidence that the intermediate formed during pyruvic acid oxidation is in fact acetylphosphate. REFERENCES 1. LiPMANN, F., Enzymologia, 4, 65 (1937). 2. LiPMANN, F., Nature, 143, 281 (1939). OcHOA, S., J. Biol. Chem., 138, 751 (1941). 3. LiPMANN, F., Advances in Enzymology, 1, 99 (1941). 4. Negelein, E., and Bromel, H., Biochem. Z., 300, 225 (1939). 5. LiPMANN, F., Nature, 144, 381 (1939). 6. LiPMANN, F., J. Biol. Chem., 134, 463 (1940). * Since submitting this paper we have succeeded in isolating silver acetyl- phosphate from the impure silver fractions. SCHLENK Stern Werkmax Haas CLOSE-UPS OF FOUR PARTICIPANTS Metabolic Cycles and Decarboxylation E. A. EVANS, JR. University of Chicago^ IF I WERE to adhere rigidly to a discussion of the topic assigned me I should be compelled to mention almost every aspect of our present knowledge of intermediary metabolism. I am claim- ing, therefore, the traditional privilege of discussing those matters that seem of particular interest and importance. From what we know of the chemical constitution of most cells, it is to be expected that those interactions of cell constituents by which the organism obtains the energy necessary for continued existence should exhibit certain characteristics of continuity and recurrence, as does the cell itself. The utilization of foodstuffs by the cell frequently involves a cyclic chain of chemical transforma- tions in which certain cell constituents, usually present in small and apparently constant amounts, facilitate the transformation of larger quantities of other metabolites in reactions releasing energy or leading to the formation of the actual protoplasmic fabric of the cell itself. The synthesis of urea, the transforaiation of glycogen into lactic acid, and, in a broader sense, the transport of the respiratory gases by the blood are familiar and typical examples of the cyclic mechanisms by which the organism maintains the balance necessary for its existence in the midst of the dynamic processes by which it functions. In the past few years such a cyclic series of reactions has been proposed for the mechanism of oxidation of carbohydrate in various tissues— more specifically for the oxidation of pyruvic acid in volun- tary musculature. This scheme, the so-called citric acid cycle, was proposed by Krebs (1, 2, 3, 4). In its original form the theory was concerned with the oxidation of pyruvic acid by minced pigeon breast muscle, and although the generalizations of the theory have been extended in part to other tissues and species (5, 6), the most convincing and complete data are those derived from suspensions of muscle tissue. * The original work reported in this paper was aided in part by grants from tlie John and Mary R. Markle Foundation and from tlie Dr. Wallace C. and Clara A. Abbott Memorial Fund of the University of Chicago. 197 198 A SYMPOSIUM ON RESPIRATORY ENZYMES A great variety of chemical transformations involving pyruvic acid has been demonstrated in different species of Hving cells, iso- lated enzyme systems, etc. (representative reactions are listed in Table 1 which is not, however, complete). In some cases several of these diverse ways of treating pyruvic acid can be demonstrated to occur in the same cell. In such circumstances there is considerable advantage in a working hypothesis in which reactions involving the same compound are regarded as components of an integrated sys- tem. In a tissue such as muscle, where specialization of function might be expected to reflect a similar specialization of metabohsm, an hypothesis of this type should be especially fruitful. The citric acid cycle represents, then, an attempt to summarize the available information with respect to the metabolism of pyruvic acid in pigeon breast muscle. Such a summary is of value only so long as it does not conflict with experimental observations and to the extent that it adequately represents the experimental foundation on which it rests. At the moment there are no observations that can be regarded as invalidating the theory, although objections to one or another feature of the scheme have been made. At the risk of adding little that is new to what you already know, I should like to discuss several aspects of the citric acid theory, particularly in regard to its experi- mental basis. The cycle itself is shown diagrammatically in Figure 1. The essential experimental support for the theory can be sum- marized in the following equations: In the presence of malonate: ( 1 ) Fumarate + pyruvate + 2 O2 — > succinate + 3 CO2 + H2O (2) Malate + pyruvate + 2 O2 -^ succinate -\- 3 CO2 + 2 H2O (3) Oxalacetate + pyruvate -f- IJ2 O2 -^ succinate + 3 CO2 + H2O (4) Citrate + O2 -^ succinate + 2 CO2 +H2O (5) a-Ketoglutarate + )k O^-^ succinate + CO2 + H2O In the absence of malonate: (6) Succinate + M O2 -^ fumarate -|- H2O ( 7 ) Pyruvic acid + 2Yi O2 -> 3 CO2 -|- H2O ( Equations 1 -|- 6 ) In nitrogen: ( 8 ) 2 Oxalacetate -j- pyruvate — > citrate + CO2 + malate (9) Oxalacetate + citrate — > a-ketoglutarate -|- CO2 + malate. The demonstration, in suspensions of muscle tissue, of the stoi- chiometric relationships expressed in these equations constitutes the experimental proof on which the citric acid cycle is based. This demonstration has two aspects— one concerned with the qualitative presence of these reactions, i.e., with the existence of the necessary METABOLIC CYCLES AND DECARBOXYLATION Table 1.— Reactions of pyruvic acid (P.A.) 199 Reaction studied Tissue used P.A. -^ acetaldehyde + CO2 P.A. + ^02 -^ acetic acid + CO2 2P.A. + H2O —* lactic acid + acetic acid + CO2 P.A. + 2H ^ lactic acid P.A. + oxalacetic acid — > citric acid 2P.A. -^ acetic acid + formic acid P.A. + glutamic acid — > alanine + a-ketoglutaric acid 2P.A. — > acetoacetic acid 2 P.A. —* ^-hydroxybutyric acid P.A. — > alanine P.A. + acetate — > acetopyruvate — > acetoacetate P.A. -tC02 —* a-ketoglutarate 2P.A. — » acetylmethylcarbinol yeast brain, gonococcus gonococcus, brain, liver muscle, tumor muscle, liver, kidney streptococcus muscle liver muscle muscle liver liver muscle enzymic channels through which these reactions may flow, and the other with their quantitative capacity to perform the task imposed upon them by their postulated role in the cycle. To the extent that a suspension of minced pigeon breast muscle retains the chemical reactions involved in its respiration in the intact state, the reactions summarized in equation 7 may be regarded as most closely approxi- "TRIOSE" i -2H- PYRUVATE OXALACETATE j | +Y^p-2H- CITRATE cii-ACONlTATE ISO-CITRATE + H20-2H- o(-KETOGLUTARATE i J. Pl_l ^ frUMARATE SUCCINATE i FUMARATE I H-MALATE + H^0-2H OXALACETATE ' U H-MALATE -2H-1 -2H -iQ2- — ^ ■^H,0 -2H- Figure 1. — Scheme of the oxidative breakdown of carbohydrate in pigeon breast muscle 200 A SYMPOSIUM ON RESPIRATORY ENZYMES mating the physiological state, since they occur when pyruvate is added to the tissue in the presence of an adequate supply of oxygen and in a medium approximately physiological in ionic concentration and pH. In these circumstances the stoichiometric relationships of equation 7 are very closely realized, and it seems certain that the total oxygen uptake of the tissue is utilized for the oxidation of pyruvic acid completely to carbon dioxide and water. Pyruvic acid will also disappear in the absence of oxygen. The quantity con- cerned is about one-tenth that removed aerobically, and although lactic acid, acetic acid, carbon dioxide, succinic acid, and beta- hydroxybutyric acid have been recognized as products of this anaer- obic reaction, the details are still obscure. Equations 1 to 5 deal with reactions carried out in the presence of malonic acid. Equation 1 is the most important reaction of the theory, since it demonstrates unequivocally an oxidative formation of succinic acid from fumaric acid. In the presence of proper con- centrations of malonic acid the oxidation of pyruvate by pigeon breast muscle can be almost entirely inhibited. When fumarate is added, however, pyruvate is oxidized and in accordance with equa- tion 1 we find, per mole of fumarate added, one mole of pyruvate and 2 moles of oxygen consumed, and 1 mole of succinate and 3 moles of carbon dioxide formed. A direct reduction of fumarate to succinate under these circumstances is inhibited by the malonic acid present, as can be shown by anaerobic experiments; that is, we find much more succinate formed from fumarate in the malonate- poisoned muscle in oxygen than in nitrogen. Since it is impossible to explain the conversion of fumarate to succinate in the presence of malonate by anaerobic reduction, a second mechanism must exist which is oxidative and unaffected by malonate and which results in the transformation of fumarate into succinate. It should be em- phasized that we find in this transformation the stoichiometric re- lationships of equation 1. The citric acid cycle postulates that this oxidative transformation involves the intermediate formation of citric acid. Regardless of the nature of the intermediates, however, any explanation of the oxidation of pyruvic acid by muscle must account for reaction 1, i.e., what might be teiTned the Krebs reaction. The demonstration of the stoichiometric relationships of equation 1 depends upon the recognition of essential experimental conditions. In view of the fundamental nature of this reaction it may be per- missible to discuss a few matters of experimental detail. The working plan of the experiments summarized in equations 1 METABOLIC CYCLES AND DECARBOXYLATION 201 to 5 involves interruption of the cycle at the succinate-fumarate stage by poisoning with malonic acid. Implicit here is the assumption that malonate acts specifically on this particular reaction, i.e., the oxida- tion of succinate to fumarate. The inhibition of succinic dehydro- genase by malonic acid in isolated enzyme preparations has been recognized and demonstrated by various investigators (7, 8). The inhibition is competitive in nature; that is, it depends not upon the absolute concentration of malonate but on the relative quantities of succinate and malonate. Since the basis of this competitive inhibition is the resemblance in chemical structure between succinic and malonic acids, it follows that malonate will probably inhibit the enzymic transformation of any substrate bearing some chemical resemblance to its own structure. However, the available data suggest that the effect of the poison is most pronounced with the succinic dehydrogenase. Expressed quantitatively, with 1 : 10 suspen- sions of pigeon breast muscle in calcium-free phosphate sahne, 0.001 M malonate inhibits pyruvate utilization about 20 per cent. Higher concentrations of malonate, around 0.025 M, inhibit to more than 90 per cent. As equation 1 indicates, fumaric acid will abolish the inhibitory eflFect of malonate in muscle tissue with the simultaneous utilization of one mole of pyruvate and 2 moles of oxygen to give 1 mole of succinate and 3 moles of carbon dioxide. The extent of this fumarate eflFect, however, depends upon the relative concentrations of malo- nate and fumarate. Krebs, in a long series of experiments (2), has shown very clearly that the demonstration of the stoichiometric relationships of equation 1 depends upon the presence of an ade- quate concentration of malonic acid. When the malonate concentra- tion is too low, as, for example, with 0.001 M malonate in the presence of 0.0025 M fumarate, the gradual conversion of fumaric acid to succinic acid will so increase the succinic acid concentration as compared with that of the malonate that succinic acid will be oxidized to fumaric acid at a rate sufficient to provide for the continuation of the cycle at the full rate. On the other hand, with 0.025 M malonate the stoichiometric relationships of equation 1 are realized. It is obvious, then, that the relationships expressed in equations 1 to 5 can be demonstrated only in the presence of ade- quate concentrations of malonic acid. When this is done, the experi- mental data are in fairly close approximation to the expected values. In similar experiments with pig heart muscle, Smyth (5) has ob- served an occasional failure with citric acid to obtain the quantities 202 A SYMPOSIUM ON RESPIRATORY ENZYMES involved in reaction 4. He oflFers several possible explanations for these occasional failures, although the matter must still be con- sidered unsettled. In terms of the postulated cycle, the possible participation of intermediate steps such as those listed in equations 1 to 5 can also be critically examined from the standpoint of the minimum rates at which they occur. This has aheady been discussed in considerable detail elsewhere (9). If we assume that the citric acid cycle repre- sents the entire respiratory process involved in the oxidation of pyruvic acid in pigeon breast muscle, then the total oxygen uptake of the tissue represents a total of five consecutive reactions of the cycle. Under circumstances in which individual reactions of the cycle are isolated— for example, when alpha-ketoglutarate is added to the malonate-poisoned tissue— the rate of oxygen uptake for the conversion of alpha-ketoglutarate to succinate cannot be less than one-fifth that for pyruvate oxidation. If the reaction velocity is below this minimum requirement, the step can be excluded as an inter- mediate in the overall reaction. On the other hand, if the isolated reaction proceeds with a velocity greater than the minimum rate, it may still be considered as an intermediate, since in the intact system the concentration of the particular enzyme concerned may exceed the quantity of the substrate normally available. When the reactions summarized in equations 1 to 5 are examined from this viewpoint, we find in each case that the rate of the isolated step proceeds with the necessary velocity consonant with its composing part of the cycle. The existence of what I have termed the Krebs reaction is gener- ally accepted: most critical comment on the citric acid hypothesis centers around the postulated intermediate foi*mation of citric acid. It has been argued that failure of this substance to accumulate in large amounts during pyruvate oxidation in various tissues is con- trary to its postulated intermediate role. However, the accumulation of an intermediate must represent a balance between its synthesis and its removal, and the accumulation of large amounts of citric acid would be much more difiicult to reconcile with the premise that it is an active intermediate substance than would the fact that it accumulates to only a limited extent. In addition, Krebs has demon- strated an anaerobic reaction between citrate and oxalacetate to yield alpha-ketoglutarate and malate (equation 9). Although there is reason to question the validity of such anaerobic experiments as the basis for assigning the role of hydrogen carriers to the oxalacetate- METABOLIC CYCLES AND DECARBOXYLATION 203 malate system in the aerobic cycle, this reaction may also explain why citrate does not accumulate to any extent. On the other hand, the only reason we have for including citric acid in the cycle is the rapidity of its formation from oxalacetate and pyruvate as indicated in equation 8 and the fact that its oxidative breakdown will yield alpha-ketoglutarate and succinate at the required rate. It has been suggested, first by Breusch (10) and later by others, that citrate formation represents a side reaction in pyruvate metabolism— a means for disposing of excessive amounts of the dicarboxylic acids. There is, however, no experimental support for this concept at the present time; the simpler hypothesis, in which citric acid is regarded as an intermediate in the cycle, is probably to be preferred as a basis for further experiments. It should be emphasized that succinate is formed in malonate-poisoned muscle at the proper rate only by those compounds listed as intermediates, although a great number of other possible substances have been examined for their ability to do this. It seems improbable, therefore, that the quantitative formation of succinate from citrate in the malonate-poisoned muscle is without significance for the role of citrate in muscle metabolism; Smyth's failure to obtain consistent results with citrate in pig heart muscle may be due to our ignorance of the essential conditions for citrate oxidation, as he suggests. With other tissues the evidence for similar mechanisms of pyru- vate oxidation is much less complete. It has been shown in pigeon hver, for example, that the enzyme systems necessary for the reac- tions of the citric acid cycle are present (6). Whetlier or not they play any considerable role in its metabolic function remains a question for further study. We do have evidence that alpha- ketoglutaric acid formation can occur in this tissue independently of the reactions of the citric acid cycle (9, 11, 12). Demonstration of the synthesis of alpha-ketoglutaric acid from pyruvic acid in pigeon liver has been followed by efforts to ascer- tain the mechanism involved. The reaction proceeds in the presence of malonic acid and without the addition of any of the four-carbon dicarboxylic acids, presumably by a direct utilization of the three- carbon compound. The scheme of Toeniessen and Brinkmann (13) for succinate synthesis from pyruvate in muscle, involving the inter- mediate formation of diketo-adipic acid, could be disregarded, since pigeon liver is incapable of metabolizing formic acid at a rate neces- sary for this mechanism to operate (6). Since pigeon Hver can form citric acid from oxalacetic acid and 204 A SYMPOSIUM ON RESPIRATORY ENZYMES can also oxidize citric acid to alpha-ketoglutaric acid, the reaction of Wood and Werkman involving the carboxylation of pyruvic acid to yield oxalacetate (14) offers an attractive hypothesis for alpha-ketoglutarate synthesis, i.e., oxalacetate + pyruvate -^ citrate -> alpha-ketoglutarate. Slotin and I carried out experiments to test this hypothesis, using carbon dioxide containing radioactive C" and found that the succinic acid formed on the addition of pyruvate to malonate-poisoned pigeon liver, presumably through the inter- mediate formation of alpha-ketoglutaric acid, was devoid of radio- activity. However, when we extended these experiments to the iso- lation of alpha-ketoglutaric acid as the dinitrophenylhydrazone, we could clearly observe the utilization of carbon dioxide in the forma- tion of radioactive alpha-ketoglutarate (15). As the earlier succinate experiments suggested, and as was confirmed later by direct exami- nation of the isolated keto acid, the radioactivity is confined entirely to the carboxyl group of this compound. The fact that carbon dioxide is used in the synthesis of alpha- ketoglutarate from pyruvate in pigeon liver and that the assimilated carbon dioxide is present entirely in the carboxyl group alpha to the carbonyl oxygen, is now firmly established in view of the iden- tical results of Wood, Werkman, Hemingway, and Nier (12), using the stable C^^, and of Slotin and me, with the short-lived radioactive C" (11, 15). These data cannot be reconciled with the intermedi- ate formation of citric acid, since any symmetrical intermediate mole- cule of this type would yield alpha-ketoglutarate with radio- activity at both carboxyls. Therefore the original conception of alpha-ketoglutarate synthesis occurring by way of pyruvate and carbon dioxide condensation to oxalacetate, and the subsequent reaction of this dicarboxylic acid with another mole of pyruvate to yield citric acid, must be abandoned. Support for such a view is also derived from experiments in which non-radioactive citrate was added to liver suspension in radioactive bicarbonate medium during the synthesis of alpha-ketoglutaric acid. Table 2 lists ex- periments showing that the addition of 25 mg. of sodium citrate affects neither the yield of alpha-ketoglutaric acid nor the ratio of activity per mg. carbon of the alpha-ketoglutarate to that of the medium. If citrate is an inteiTnediate in the formation of alpha- ketoglutaric acid from pyruvate and carbon dioxide, the addition of non-radioactive citrate to the synthesizing tissue should yield alpha-ketoglutaric acid in which the radioactivity had been con- siderably diluted. Failure to demonstrate such an effect may be METABOLIC CYCLES AND DECARBOXYLATION 205 considered additional evidence for the nonparticipation of citric acid in the synthesis of alpha-ketoglutaric acid from pyruvate and carbon dioxide. Table 2.— The e£Fect of citrate on the radioactivity of synthesized alpha-ketoglutaric acid* Experiment Sodium citrate added a-Ketoglutarate synthesized Activity of a-ketoglutarate 1 mg. 0 mg. 31.4 per mg. C 0.148 25 28.8 0.113 2 0 36.9 0.116 25 39.8 0.116 * Experimental conditions were similar to those in the experiments reported in Table 3. The sodium citrate was added simultaneously with the pyruvic acid. Wood and Werkman and their collaborators (12) have suggested that isocitric acid rather than citrate serves as an intermediate, such a scheme retaining in skeleton form the original suggestion that alpha-ketoglutarate was produced by way of the citric acid cycle. In view of the demonstrated equilibrium, in most tissues, between citrate and isocitrate (16), additional evidence is required to estab- lish this hypothesis. The exclusion of citric acid as an intermediate in alpha-ketoglutarate synthesis might well lead to an examination of whether oxalacetate itself is formed as the primary product of carbon dioxide assimilation in pigeon liver. The great chemical reactivity and instability of this compound make it improbable that it should accumulate to any appreciable extent under our experi- mental conditions. Though we have studied a wide variety of conditions, we have failed to obtain any evidence that the compound was present during the course of alpha-ketoglutarate synthesis; or, to express the matter more precisely in experimental terms, we have failed to observe the accumulation of any substance that would release radioactive carbon dioxide on treatment with aniline citrate. Certainly it would be desirable to have more convincing evidence that oxalacetic acid is the primary product of carbon dioxide as- similation in pigeon liver. There seems little doubt that the utilization of carbon dioxide by pigeon liver is a process of considerable magnitude. Table 3 gives data from a series of typical experiments which indicate that as much as 5 per cent of the original activity of the added inorganic 206 A SYMPOSIUM ON RESPIRATORY ENZYMES carbonate appears in the synthesized alpha-ketoglutaric acid. Un- fortunately these data are difficult to interpret in terms of any precise mechanism of alpha-ketoglutarate formation. Since all the Table 3.— Synthesis of alpha-ketoglutarate in radioactive bicarbonate medium* Terminal Ratio Ratio a-Keto- Exp. glutarate synthe- sized Activity t of a-ketoglutarate Total activity of medium activity of medium per mg. inorganic carbon a-ketoglutarate activity medium activ- ity per mg. C Total per mg.C original activ- ity of medium a-ketoglutarate per mg. C mg. 1 31.4 0.471 0.036 8.74 0.078 0.054 2.0 2 27.5 4.09 0.362 67.0 1.13 0.061 3.1 3 55.0 3.85 0.171 0.161J 0.178i 79.8 1.17 0.048 6.9 * 7.6 gm. minced pigeon liver in 50 ml. calcium-free bicarbonate saline, pH. 7.4, containing radioactive C"; 1.7 ml. 0.1 M malonate; 10 ml. 0.02 M pyruvate; malonate added directly to tissue; pyruvate added after 10 minutes; temperature, 40° C; gas phase, 5 per cent carbon dioxide; experimental period, 40 minutes. t Activities are expressed in divisions per second (Lauritsen electroscope) ; and are corrected for decay so that the values are comparable. X Activities after successive recrystallizations of alpha-ketoglutarate dinitrophenyl- hydrazone. radioactivity is lost on oxidation of alpha-ketoglutaric acid to suc- cinic acid, we can definitely conclude that not more than one carbon atom of the five of alpha-ketoglutarate is derived from the medium. From the quantity of radioactivity present in the alpha-ketoglutarate we can calculate (the details of these very approximate calculations are given elsewhere [11]) that approximately one carbon atom in ten of the alpha-ketoglutarate is derived from the inorganic car- bonate of the medium. Since this figure involves very approximate calculations, and is of value only in suggesting the order of magni- tude of the reaction, we are inclined to believe, as a working hypothesis, that one mole of carbon dioxide from the medium is assimilated per mole of alpha-ketoglutarate synthesized. The synthesis of alpha-ketoglutarate is by no means the limit of carbon dioxide assimilation. Under our experimental conditions we find that alpha-ketoglutaric acid represents only about 25 per cent of the radioactivity which has been assimilated, that is, activity which can no longer be released as carbon dioxide on the addition METABOLIC CYCLES AND DECARBOXYLATION 207 of strong acid. Part of this non-a-ketoglutarate radioactivity can be released as carbon dioxide on treatment with ninhydrin at 100° C, and with chloramine-T (Table 4). This suggests that part of the assimilated carbon dioxide has been converted into an amino acid. Beyond this, however, we know nothing of the chemical nature of the compounds concerned. Table 4.— Effect of ninhydrin and chloramine T on non-a- ketoglutarate radioactivity* Total a-Keto- Total activity Activity of Activity of Experiment glutarate activity of medium CO2 released CO2 released synthesized of a-keto- after CO2 by ninhydrin by chloramine-T glutarate removal mg. 1 35.0 5.18 31.8 7.8 — 2 39.0 6.2 31.2 — 10.39 3 53.0 1.2 4.2 — 0.78 4 31.6 0.22 0.929 0.209 — * Experimental conditions are similar to those described in Table 3. The experiments with chloramine-T and with ninhydrin were carried out on aliquots of the metaphos- phoric deproteinized reaction mixture under the conditions described, respectively, by P. P. Cohen in Biochem. J., 33, 551 (1939) and by D. D. Van Slyke and R. T. Dillon in Comp. rend, du Lab. Carlsberg, Ser. Chim., 22, 480 (1937). In all experiments involving the use of radioactive isotopes or tracer elements it is a matter of primary concern whether the ap- pearance of the tagged atom in a product represents a true meta- bolic reaction or has been introduced by a nonspecific exchange reaction. Oxalacetic acid, for example, breaks down to pyruvic acid and carbon dioxide in tissues, and it might be argued that any reaction involving this compound would yield a radioactive end product if the reaction were carried out in the presence of radio- active carbon dioxide. Such a possibility is difficult to control. How- ever, the very large amounts of carbon dioxide assimilated argue against this explanation of the presence of radioactivity in the alpha-ketoglutaric acid. Also suggestive are experiments in which the possibility of an exchange between the carboxyl group of oxal- acetic acid and carbon dioxide of the medium has been examined. In muscle, alpha-ketoglutaric acid is synthesized by the condensation of oxalacetate with pyruvate to form citrate and the subsequent oxidation of this compound to alpha-ketoglutarate; i.e., in terms of the citric acid cycle this is what occurs. Now, a formation of oxal- 208 A SYMPOSIUM ON RESPIRATORY ENZYMES acetic acid from pyruvate and carbon dioxide in pigeon breast muscle does not occur, inasmuch as malonate will completely inhibit pyruvate utilization in the absence of added oxalacetic or other dicarboxylic acids. Under these circumstances the synthesis of muscle alpha-ketoglutarate in a radioactive bicarbonate medium would give rise to radioactive alpha-ketoglutarate only if a process of exchange between intermediates in the reaction (such as oxal- acetate) and the carbonate of the medium took place. If the quantity of radioactivity found in muscle alpha-ketoglutarate is similar in magnitude to that found in the alpha-ketoglutarate synthesized by liver, it would eliminate the necessity for proposing a stoichiometric utilization of carbon dioxide in the process. When this reaction is studied in radioactive bicarbonate medium, the alpha-ketoglutarate formed is devoid of radioactivity. There is apparently no appreciable interchange between the carbon dioxide of the medium and the carboxyl group of oxalacetic acid or other intermediates. A similar negative result was obtained in studies of the equilibrium between pyruvic acid and carbon dioxide in the presence of carboxylase in which radioactive carbon dioxide and pressures as high as 300 atmospheres were used. Under these circumstances no interchange of carbon dioxide and the carboxyl of the pyruvate could be demon- strated. In view of these data, it seems most probable that the assimilation of carbon dioxide by minced liver is a true metabolic reaction, although the mechanism is still obscure. The most promising line of attack of these problems lies probably in the attempt to isolate or simplify the enzyme systems involved. Recently attention has been directed again (17, 18) to the enzymic formation of acetylmethylcarbinol from pyruvate as a possible step in pyruvate oxidation. This reaction has two features of interest: first, it represents an anaerobic, i.e., non-oxidative, decarboxylation of pyruvate; secondly, it poses again the question of acetaldehyde as an intermediate. It is still too early to determine the relative importance of this reaction in pyruvate metabolism of muscle and other tissues. If it should prove to be of significance, the resulting clarity will undoubtedly be suflBcient compensation for the revision that will have to be made in our current beliefs. REFERENCES 1. Krebs, H. a., and Johnson, W. A., Enzymologia, 4, 148 (1937). 2. Krebs, H. A., and Eggleston, L. V., Biochem. J., 34, 442 (1940). 3. Krebs, H. A., Biochem. J., 34, 460 (1940). 4. Krebs, H. A., Biochem. J., 34, 775 (1940). METABOLIC CYCLES AND DECARBOXYLATION 209 5. Smyth, D. H., Biochem. J., 34, 1046 (1940). 6. Evans, E. A., Jr., Biochem. J., 34, 829 (1940). 7. Thunberg, T., Skand. Arch. Physiol., 24, 23 (1910). 8. QuASTEL, J. H., and Wooldridge, W. R., Biochem. J., 22, 689 (1928). 9. Evans, E. A., Jr., BuU. Johns Hopkins Hosp., 69, 225 (1941). 10. Breusch, F. L., Biochem. J., 33, 1757 (1939). 11. Evans, E. A., Jr., and Slotin, L., J. Biol. Chem., 141, 439 (1941). 12. Wood, H. G., Werkman, C. H., Hemingway, A., and Nier, A. O., J. Biol. Chem. 139, 483 (1941). 13. ToENiEssEN, E., and Brinkmann, E., Z. physiol. Chem., 187, 137 (1930). 14. Wood, H. C, and Werkman, C. H., J. Bact, 30, 332 (1935); Biochem. J., 30, 48 (1936); 32, 1262 (1938); 34, 129 (1940). Wood, H. C, Werkman, C. H., Hemingway, A., and Nier, A. O., J. Biol. Chem., 135, 789 (1940); 139, 365 (1941); 139, 377 (1941). 15. Evans, E. A., Jr., and Slotin, L., J. Biol. Chem., 136, 301 (1940). 16. Johnson, W. A., Biochem. J., 33, 1046 (1939). 17. Tanko, B., Munk, L., and Abonyi, I., Z. physiol. Chem., 264, 91 (1940). 18. Green, D. E., Westerfeld, W. W., Vennesland, B., and Knox, W. E., J. Biol. Chem., 140, 683 (1941). Transamination PHILIP p. COHEN University of Wisconsin IN VIEW OF THE FACT that Braunstciii (1) has recently reviewed in some detail the results of experiments on transamination in his laboratory, no attempt will be made here to present tabular details of these studies. Unfortunately most of the publications issued since the review by Braunstein have not been available to the writer except in Chemical Abstracts, to which reference is made in the bibliography. Many of the papers published in this field, being of a preliminary sort, are of uncertain value. Most investigators seem to be interested merely in demonstrating that transamination does or does not occur in a given tissue. Unfortunately the analytical methods usually em- ployed permit little more than a qualitative demonstration of the reaction. Thus, in spite of a widespread interest in transamination, few reliable quantitative data are available that can be used for evaluating the role of this reaction in intermediary metabolism. Types of Transamination Transamination is a reaction between an alpha-amino and an alpha-keto acid resulting in the transfer of the amino group from the former to the latter. When the reaction is enzymatically cata- lyzed, the end products formed are an alpha-amino and an alpha- keto acid, the former corresponding in structure to the original alpha-keto acid, and the latter to the original alpha-amino acid (Reaction 1). (1) Ri-CH(NH2)-COOH-f-R.COCOOH^RrCOCOOH + R2-CH (NH2)-COOH When carried out at boiling temperatures, the end products of transamination are usually an aldehyde corresponding to the original alpha-amino acid, an alpha-amino acid corresponding to the original alpha-keto acid, and carbon dioxide (Reaction 2). (2) Ri-CH(NH2)COOH + RrCOCOOH^R,-CHO -f- RjCHNHj-COOH -f- CO2 Reaction 2 has been studied in some detail by Herbst, who has recently reviewed this subject (2). 210 TRANSAMINATION 211 The present review will deal chiefly with the system represented by Reaction 1, and the use of the term "transamination" will be Hmited to this reversible, enzyme-catalyzed system. Discovery of Transamination Reaction Transamination as a biological reaction was first recognized by Needham "(3), who found that glutamic and aspartic acids disap- peared anaerobically when added to pigeon breast muscle without a decrease in amino nitrogen. She could find no increase in urea or ammonia nitrogen, but did observe an increase of succinic acid. She suggested that "possibly a combination of the amino group with some reactive carbohydrate residue takes place; then when splitting and oxidation occur the amino group is retained in the form of a new amino acid." Unrecognized evidence for transamination was published from Szent-Gyorgyi's laboratory in 1936 and 1937. It was observed that the rate at which oxalacetic acid disappeared was greatly increased when glutamic acid was added to pigeon breast muscle (4) and to certain enzyme preparations from the same tissue (5). Credit for the discovery of transamination goes to Braunstein and Kritzmann (6), who carried out the first detailed investigation of this reaction. They succeeded in showing that the reaction, termed "Umaminierung" by them, took place in pigeon breast muscle. The system represented in Reaction 3 was studied, and they were able to show its reversibility by isolation and chemical identification of the alanine and glutamic acid formed. (3) Z( + ) -glutamic acid -j- pyruvic acid <^ a-ketoglutaric acid -f Z(-j-) -alanine. Substrates Aciive in Transamination Investigations by Braunstein and Kritzmann (7) led them to con- clude that the enzymatic transfer of amino groups in pigeon breast muscle takes place between any alpha-amino acid (with the possible exception of glycine) and the dicarboxylic acids, alpha-ketoglutaric and oxalacetic, as well as between the dicarboxylic amino acids, glutamic and aspartic, and various alpha-keto acids. No amino group transfer was observed in the following systems: (1) between mono- carboxylic alpha-amino acids and monocarboxylic alpha-keto acids; (2) from amines and peptides; or (3) from alpha-amino dicarboxylic acids to ketones, hydroxy-ketones, or aldehydes. The analytical method employed in this study for measuring the formation or dis- 212 A SYMPOSIUM ON RESPIRATORY ENZYMES appearance of glutamic acid was an adaptation of the Foreman (8) method, which has recently been shown by Zorn (9) to be unreliable for studying transamination. In contrast to certain of the above findings Cohen (10) found that of twenty-one different alpha-amino acids studied in pigeon breast muscle Z(— )-aspartic acid and Z( + )-alanine were the most active in forming glutamic acid in the presence of alpha-ketoglutaric acid. Alpha-aminobutyric acid and ?( + ) -valine were slightly active, but none of the remaining amino acids was appreciably active. Of a series of alpha-keto acids studied, oxalacetic and pyruvic acids were the most active in causing the anaerobic disappearance of glutamic acid. Alpha-ketobutyric and mesoxalic acids were slightly active, but no activity was observed with alpha-ketovaleric, alpha-keto- caproic, acetoacetic, and laevulic acids. Moreover, no transamination was observed between alpha-ketoglutaric acid and a variety of amino compounds other than alpha-amino acids. The formation and disappearance of glutamic acid were determined by the method of Cohen (11). The specificity and accuracy of this method has recently been confirmed (12, 13). d-Amino Acids.— Amino acids of the d series are not active in transamination (10, 14, 15, 16). Activity with (i-amino acids has been reported by Braunstein (1), Braunstein and Azarkh (17), and Euler et at. (18). Peptides.— The role of peptides as transamination substrates is of considerable interest. As noted above, Braunstein and Kritzmann (7) observed no activity with various peptides. In the case of trans- aminase preparations Cohen (14) was unable to demonstrate any appreciable transamination between glutathione and oxalacetic acid. On the other hand Agren (19) reported that in minced cattle diaphragm muscle transamination takes place between alpha- ketoglutaric acid and the peptides glycylaminobenzoic acid and valylglycine. Valylglycine was found to be as active as alanine (about 30 per cent transamination), while glycylaminobenzoic acid was less active. Agren employed the same method as Braunstein and Kritzmann for measuring transamination. In view of the latter work- ers' observation that glycine interfered with this detemiination (1) by being carried down in the dicarboxylic acid amino nitrogen fraction, it is possible that the results obtained by Agren with glycine dipeptides are due to this fact. No control experiments of glycine dipeptides plus tissue are reported. "Primary" and "Secondary" Substrates.— According to Braunstein TRANSAMINATION 213 the substrates available for transamination can be considered as consisting of two types: 1. "Primary" or "active" substrates. These have a high affinity for the enzyme and react with mono- or dibasic acids after adsorp- tion. Primary substrates are the dibasic alpha-amino or alpha-keto acids, and include compounds such as cysteic acid. 2. "Secondary" or "passive" substrates. This group includes all the alpha-amino and alpha-keto monobasic acids which have no direct affinity for the enzyme and serve only as "reaction partners" for the primary substrates. This idea of Braunstein can be extended to mean that only the dibasic alpha-amino or -keto acids are activated by the enzyme. This would suggest that the dibasic alpha-amino and -keto acids act as prosthetic groups, which in the presence of the enzyine react with the secondary substrates. That is, the secondary substrates have an affinity for the enzyme only when the "primary" substrates have become activated. This concept is not unreasonable, but it appears unnecessary to classify the substrates as active or passive. From Braunstein's point of view this is essential, since he has to account for the activity of the large number of monobasic alpha-amino and -keto acids which he reports to be active in transamination. From the writer's point of view this concept is unnecessary, since his experi- mental data lead to the conclusion that transamination is essentially a limited reaction and concerned chiefly with the dibasic alpha- amino and -keto acids. The activity of other compounds can be explained on the basis of different affinities for the enzyme. Thus with large amounts of tissue and long incubation periods other amino acids show a small amount of activity (20). "Catalytic" Transamination.— Braunstein (1) and Braunstein and Kritzmann (21) have reported that the addition of small amounts of a dibasic alpha-amino or alpha-keto acid (as little as M/16000) causes transamination to take place in pigeon breast muscle brei between lysine and pyruvic acid. The latter system alone is inactive. With glutamic acid, the reaction is pictured as proceeding in the following manner: ( 3 ) glutamic acid + pyruvic acid — > a-ketoglutaric acid + alanine ( 4 ) a-ketoglutaric acid -f lysine —> glutamic acid -f 2-keto-6-aminocaproic acid The net effect of this would be: ( 5 ) lysine + pyruvic acid —> 2-keto-6-aminocaproic acid -|- alanine The effectiveness of so low a concentration as M/16000 is remark- 214 A SYMPOSIUM ON RESPIRATORY ENZYMES able in view of the fact that, as Braunstein himself points out, pigeon breast muscle brei has a much higher content of dibasic alpha- amino and -keto acids normally present (11). Yet Braunstein reports that reaction 5 will not take place without the addition of small amounts of catalyst. The writer (10, 14) was unable to demonstrate a similar catalytic eflFect with either pigeon breast muscle or purified transaminase by using the system: ( 6 ) l{ — )-aspartic acid + pyinivic acid ^ oxalacetic acid + H + ) -alanine. Braunstein (1) and Bychkov (22) have reported that cysteic acid and phosphoserine are active in transamination in pigeon breast muscle. Using purified transaminase, the writer confirmed the ac- tivity of cysteic acid, but phosphoserine was found to be inactive. Of interest, however, is the report by Braunstein (1) that neither of these two compounds is active with purified enzymes (glutamic and aspartic aminopherases). Preparation and Properties of Transaminating Enzymes Kritzmann (23, 24, 25) has described in some detail the prepara- tion and properties of purified transaminating enzymes from pigeon breast and pig heart muscles. According to her, two distinct systems exist, one of which is concerned with glutamic acid (and alpha- ketoglutaric acid) and the other with aspartic acid (and oxalacetic acid). The former enzyme is called glutamic aminopherase and the latter aspartic aminopherase. Both enzymes are reported to require co-factors, present in muscle kochsaft, whose chemical constitutions are still unknown but which are similar if not identical for the two systems. Aspartic aminopherase is thought to be a more labile system, since it is claimed that muscle suspensions lose their trans- aminating activity on dilution more rapidly with aspartic acid than with glutamic acid. The following are some of the properties of the aminopherases found by Kritzmann (25): 1. Purification by adsorption, salting out, or dialysis results in inactivation. 2. Reactivation follows on addi- tion of boiled muscle extracts or ultrafiltrates. 3. To be effective, the glutamic-aspartic aminopherases must contain a thermostable, low molecular weight activator or coenzyme. 4. Denaturation by acetone, ethyl alcohol, or methyl alcohol leads to an irreversible in- activation. 5. Heating at 80° C. for five minutes causes a 50 per cent decrease in activity. 6. pH activity range is 5.5-8.5, with an optimum TRANSAMINATION 215 at 7.4. 7. Glutamic aminopherase is best obtained from pig heart muscle, aspartic aminopherase from coarsely ground pea seedlings. If the latter are finely ground, both aspartic and glutamic amino- pherases are obtained. Transaminating enzyme preparations from pig heart and pigeon breast muscle have recently been studied by Cohen (14, 26). It was found that the activity of these preparations was greatest with the systems, Z( + ) -glutamic acid plus oxalacetic acid, and alpha-keto- glutaric acid plus /(—) -aspartic acid. That is, the enzyme was most active in catalyzing a reaction in which both glutamic and aspartic acids (and the corresponding alpha-keto acids) were substrates. The addition of pigeon breast muscle kochsaft was without influence on the rates of reactions 3, 6 or 7. (7) /( + ) -glutamic acid -f- oxalacetic acid ^ a-ketoglutaric acid -f Z( — ) -aspartic acid Reaction 7 was catalyzed at a rapid rate, the Qt values being of the order of 1600.* The position of equilibrium for this system was far to the right, with an equilibrium constant of about 3. Re- action 3 was catalyzed at a much slower rate by transaminase, the Qt values being of the order of 300. The equilibrium constant was about 1. Reaction 6 was not catalyzed by transaminase. The following are some properties of transaminase: 1. The enzyme is best prepared from pig heart muscle and pigeon breast muscle. 2. Transaminase can be dried by rapid lyophilization at low tem- peratures. Such preparations remain active for as long as six weeks at room temperature (63). 3. Purification by salting out or dialysis results in inactivation. However, solutions of the enzyme can be further purified by adsorption on calcium phosphate (63). 4. Trans- aminase has an optimum activity at 40° C. and at pH 7.5. The Michaelis constant with the substrates glutamic and oxalacetic acids is 0.0138 M. 5. Muscle kochsaft, diphosphopyridine nucleotide, thia- min, and cocarboxylase are without influence. V. Euler et al. (27) have stated that the transaminating enzyme does not require coenzymes or apparently any other cofactor dis- sociable at a neutral reaction. Substrate Specificity.— Ex-periments with purified enzyme prepara- tions (transaminase) showed essentially the same substrate specificity as with pigeon breast muscle (14). Thus with alpha-ketoglutaric * Qt = Qtransamiiiation = microlitcrs substratc transaminated per mg. dry weight per hour. 216 A SYMPOSIUM ON RESPIRATORY ENZYMES acid only Z(— )-aspartic acid, Z(+) -alanine, and Z( — )-cysteic acid were active, while with oxalacetic acid only Z( + ) -glutamic and Z(— )-cysteic acids reacted. A few experiments have been reported by Braunstein (1) in which glutamic acid formation from alpha - ketoglutaric acid and diflferent amino acids was measured with both pigeon breast muscle and purified enzyme preparations. Of interest are the findings that while Z( + ) -alanine shows a slightly smaller activity with the pigeon breast muscle, /( + ) -valine, Z( — ) -leu- cine, and Z( + )-isoleucine react to only about one-third the extent with purified enzyme. These differences, no doubt, reflect the in- adequacy of the analytical method employed in these investiga- tions. From the studies with purified transaminase (12, 14) it appeared that the chief substrates for this enzyme are those represented in reaction 7. Aside from pyruvic acid and /(-{-) -alanine, no monobasic alpha-amino or alpha-keto acids were found to be active. However, dibasic alpha-keto and alpha-amino acids other than those shown in reaction 7 are active. Thus ?(— )-cysteic acid will react with both oxalacetic and alpha-ketoglutaric acids. Glutathione does not react with oxalacetic acid. Mechanism of Transamination The mechanism by which the intermolecular transfer of the amino group takes place is not known. Following Herbst's (2) idea of the mechanism of non-enzymatic transamination, Braunstein and Kritzmann (6) have postulated the formation of an intermediate Schiff's base, and picture the reaction as follows: SCHEME I R R. R R HpO. CO + HpN-C-H -^^^^ C=N-C-H I I I I COOH COOH COOH COOH 11 R R, R R, H-C-NH2 + CO ^^^^ H-C-N=C I I II COOH COOH COOH COOH TRANSAMINATION 217 The inability of purified transaminase preparations to dehydrogen- ate glutamic or aspartic acids in the presence of suitable hydrogen acceptors, and the absence of free ammonia during the course of the reaction suggest that the amino group is transferred through an intermediate complex of the SchiflF's base type. Karrer et al. (28) have investigated the metabolism of octopin, following the suggestion of Knoop and Martins (29) that octopin represents a naturally occurring intermediate compound of trans- amination. Karrer et al. found that fresh Hver brei (which they state contains an Z-amino acid dehydrogenase) was capable of de- hydrogenating octopin. d-Amino acid dehydrogenase preparations were not active, indicating that the alanine was of the Z-form. Since the biological synthesis of octopin involves an intermediate reduc- tion of a SchifiF's base to an a, a'-imino dicarboxylic acid, the authors conclude that transamination may involve intermediate reduction of the Schiffs base with subsequent dehydrogenation. The reaction is pictured as follows: SCHEME II R R, I I H-C-NHp + 0=0 I I COOH COOH -^^ H-C— N=C COOH COOH + H; R H R, R R, I I C=0 + HpN-C-H I I COOH COOH H-C— N— C-H I I COOH COOH -Hz R R, +H2O I I ^-^- C=N— CH I I COOH COOH It has the advantage of providing a mechanism for the explana- tion of the double bond shift between the nitrogen and the two alpha carbon atoms. As previously pointed out, the evidence for the existence of a 218 A SYMPOSIUM ON RESPIRATORY ENZYMES cofactor for transaminase is still inconclusive. Should an oxidizable and reducible cofactor prove to be involved, its function might be that of acting as the hydrogen acceptor and donator in the scheme of Karrer et al. (Scheme II). However, as has been said, none of the known cofactors has any influence on transaminase activity, not- withstanding the fact that the method of its preparation is such as to remove practically all the known cofactors with the possible exception of flavinadenine dinucleotide. Of interest in this connec- tion is the writer's unpublished observation that the activity of transaminase preparations, fractionated by various methods for pur- poses of purification, is associated with those fractions showing a green fluorescence, similar to that obtained with flavoproteins. Transamination in Different Tissues As has been said, much of the available data on transamination is of a qualitative nature. Not only must it be demonstrated that transamination occurs in a given tissue, but the rate of the reaction in terms of unit weight of that tissue must be known. Thus a sig- nificant amount of transamination may be shown to take place in certain cases with large amounts of tissue and long incubation periods, but calculation of the rates in terms of Qt would reveal a value so low as to cast doubt on the significance of this reaction in the metabolism of that tissue. It is thus essential to have accurate data on the rate of transamination in different tissues before as- signing to it a role in intermediary metabolism. Unfortunately very few such data are available. Animal Ti55we5.— Transamination in different animal tissues was first studied by Kritzmann (30). Using the system glutamic acid plus pyruvic acid, she reported transaminase activity in liver, kidney, skeletal muscle, heart muscle, and brain, but none in smooth muscle (chicken gizzard), lung, erythrocytes, and yeast. There was ques- tionable activity in the case of malignant tissue. Values of Qt calculated from these data are of the order of 1.5-2.0 for the more active tissues. Similar studies were carried out on a variety of tissues by Cohen (31). A quantitative study of the rate of transamination in different rat tissues was recently carried out by Cohen and Hekhuis (15). As can be seen from Table 1, the rates, expressed in terms of Qt, are very high in most tissues with the substrates glutamic acid plus oxalacetic acid. The Qt values are higher than the succinoxidase TRANSAMINATION 219 Q values for the same tissues (32) and in most instances exceed the cytochrome oxidase Q values (32, 33). The Qt values with the systems glutamic acid plus pyruvic acid, and aspartic acid plus pyruvic acid, are comparatively low in all tissues studied except liver. The higher rates of transamination in liver are probably due to the fact that pyruvic acid is converted into oxalacetic and alpha-ketoglutaric acids in this tissue (15, 34, 35, 36). Table 1.— Values of Qt in difiFerent rat tissues Glutamic acid Glutamic acid Aspartic acid Tissue Oxalacetic acid PjTuvic acid Pyruvic acid Heart muscle . . ."' . . 425 7 7 Skeletal muscle . . . . 316 13 1 Brain . , 260 2 8 Liver . . 245 46 10 Kidney . . 245 3 3 It is apparent from the data in Table 1 that transamination is chiefly concerned with the substrates glutamic acid plus oxalacetic acid. The metabolic importance of this will be discussed later. Breusch (37) reported some experiments on transamination in different tissues with the system glutamic acid plus oxalacetic acid. The following Qt values for various minced cat tissues were cal- culated from his data: muscle, 19.7; liver, 14.1; kidney, 22.5; lung, 22.5; embryonic muscle, 14.1; brain, 19.7; and washed erythrocytes, 2.8. These values are all much lower than those reported for cat tis- sues by Cohen and Hekhuis (16). Breusch claims that lung tissue is an excellent source of transaminase, since it is free of many dehydrogenase systems. Cohen and Hekhuis, on the contrary (15), found lung to have a low transaminase activity. The method of oxalacetic acid determination used by Breusch (4) for measuring transamination is far from satisfactory and probably accounts for the results obtained. Karayagina (38) found that ?(— )-aspartic acid reacts with pyruvic acid to form alanine in skeletal and cardiac muscles, in liver, kidney, and brain, but not in testes. Malignant and Embryonic Tissues.— Euler, Giinther, and Fors- man (39) and Braunstein and Azarkh (17) reported low transamina- tion values for tumors. The former workers used a quaUtative ana- 220 A SYMPOSIUM ON RESPIRATORY ENZYMES lytical method for measuring the disappearance of oxalacetic acid. The latter workers studied chiefly the reactions, glutamic acid plus pyruvic acid, and alpha-ketoglutaric acid plus diflFerent amino acids. Quantitative studies of the rate of transamination in different tumors and embryonic tissue have been carried out by Cohen and Hekhuis (16). In their study a series of mouse tumors and cat em- bryonic tissues were shown to have low rates of transamination as compared with normal adult tissue. Further, the writer has shown (63) that the transaminase activity of liver from rats fed dimethyl- aminoazobenzene decreases progressively to a low value in the hver tumors arising in these animals. It thus appears that rapid growth, or increased protein synthesis, is associated with a low transaminase activity. Plant Tissues.— Euler et al. (27) reported transaminase activity in higher plants, but no data were pubhshed in support of this state- ment. Virtanen and Laine (41) state that transamination between oxalacetic acid and alanine takes place in crushed pea plants. Transaminase activity in extracts prepared from pea, lupine, and pumpkin seedlings has been reported by Kritzmann (24). Cedran- golo and Carandante (42) studied transamination in leguminous and graminaceous plants. Dialyzed extracts of seeds and sprouts were used and were prepared by the same method as that employed by Adler and Sreenivasaya (43) for the preparation of formico- dehydrogenase. The systems Z(— )-aspartic acid plus pyruvic acid, Z(— )-aspartic acid plus alpha-ketoglutaric acid, and alpha-keto- glutaric acid plus Z( + ) -alanine were investigated. Experimental data are not presented in this paper, but rather the relative veloci- ties of the above reactions in graminacae seeds and sprouts as com- pared with leguminous seeds and sprouts, which are assigned an arbitrary value of 100. The graminacae extracts are reported to be 1.5 to 2.5 times as active as those from legumes. According to these authors, the lack of aspartic acid utilization by graminacae, re- ported by Virtanen, cannot be due to the lack of a transaminating enzyme. Of interest is the finding of these workers that dialyzed plant extracts are active. As previously mentioned, Kritzmann (25) reported that dialysis inactivates transaminating enzymes from plant sources. Experiments with Chlorella (63) by the author failed to show any evidence of transaminase activity with the systems alpha- ketoglutaric acid plus Z(+) -alanine, and alpha-ketoglutaric acid plus Z( — ) -aspartic acid. TRANSAMINATION 221 Wyss (44) observed formation of considerable amounts of aspartic acid when oxalacetic acid was added to crushed pea nodules. The amount of aspartic acid found was somewhat greater when alanine was also added. Whether the aspartic acid formation is due to transamination or to some other reaction is not certain. Yeast and Bacteria.— Adler, Giinther, and Everett (45) reported transaminase activity in yeast extracts. Adler et al. (46) state that Bacterium coli {Escherichia coli) suspensions form oxalacetic acid from alpha-ketoglutaric acid and aspartic acid, and that lactic acid bacilli are capable of transamination, though at a slower rate than Esch. coli. No experimental data are given in either of the above papers. Experiments by the writer (63) showed that Lebedev juice pre- pared from brewer's yeast was active in catalyzing the reaction alpha-ketoglutaric acid plus Z(—) -aspartic acid, but not the reaction, alpha-ketoglutaric acid plus Z(-f )-alanine. Baker's yeast showed no activity with either system. Experiments with Esch. coli suspensions (63) demonstrated that no transamination occurred between Z( + )- glutamic acid and pyruvic acid, but did take place between Z( + ) -glutamic acid and oxalacetic acid. Qt for the latter reaction was 17.2. Transamination in uiuo. —Kritzmann (40) found that intravenous, intramuscular,- or intraperitoneal injection of glutamic acid into rabbits, pigeons, and white mice causes the rapid appearance of alanine in the blood and the tissues. The pyruvic acid necessary for this reaction is endogenous. The injection of alpha-ketoglutaric acid plus alanine resulted in the formation of glutamic acid. No transamination occurs in the blood itself. The failure of blood cells to catalyze transamination has also been observed by the writer (63). Influence of Various Substances on Transamination Inhibitors.— None of the well-known inhibitors has any marked effect on transamination with the exception of cyanide, which in high concentrations (0.05 M) causes up to 80 per cent inhibition. This is in all probability due to the formation of cyanohydrin. How- ever, at 0.001 M concentration cyanide still causes an inhibition of about 30 per cent, indicating that an effect on the enzyme system may occur. Malonate, pyrophosphate, sodium fluoride, iodoacetate, bromoacetate, arsenious oxide, and octyl alcohol have little or no influence on transamination in pigeon breast muscle (10). With transaminase no inhibition of transamination between glu- 222 A SYMPOSIUM ON RESPIRATORY ENZYMES tamic acid and oxalacetic acid was noted in the presence of malo- nate, succinate, pyrophosphate, and citrate (14). In contrast to the above findings with di- and tribasic acids Braunstein (1) reported that small concentrations of dibasic acids competitively inhibit "catalytic" transamination. The latter refers to the catalytic effect of small concentrations of a dibasic alpha- amino or alpha-keto acid on transamination between a monobasic alpha-amino acid and a monobasic alpha-keto acid. The author's attempts to corroborate this finding of Braunstein's have not been successful (10, 14). Vysshepan (47) found that the activity of glutamic aminopherase was inhibited by the following reagents (molar concentrations): quinone (0.01); potassium cyanide (0.01); glutathione (0.002-0.004); cations of calcium, barium, and strontium (0.02); mercury and silver (0.0001). Reagents which are relatively harmless are narcotics, so- dium fluoride, monoiodo- and monobromo-acetic acids, arsenite, arsenate and selenite; the anions, chloride, bromide, iodide, acetate, nitrate, carbonate, sulfate (0.01); ascorbic acid (0.01), hydrogen sulfide, cysteine, ferrous ion, semicarbazide, phenylhydrazine and hydroxylamine. Sober and Cohen (64) observed no inhibition of transaminase by glutathione (0.005 M.). It is apparent that thus far no specific inhibitor is available for transaminase. The discovery of such a compound would aid greatly in elucidating the physiological role of transamination. JF/ormones.— Transamination with purified transaminase and the system glutamic acid plus oxalacetic acid is uninfluenced by high concentration of the following: insulin (crystalline and zinc com- pounds), desoxycorticosterone, cortical extract, anterior pituitary extract, estradiol, androsterone, testosterone, and stilbestrol (63). Carcinogens.— No effect on transaminase activity was observed with methylcholanthrene and dimethylaminoazobenzene (63). Vitamins.— Thiamin and cocarboxylase are without influence on transaminase activity (63). However, a decrease in activity was ob- served in minced breast muscle from B^ deficient pigeons with the system glutamic acid plus pyruvic acid (63). Similar findings have recently been published by Kritzmann (48). Barron (49), on the other hand, found no decrease in transamination in liver from Bj deficient rats. These results are not necessarily contradictory, since Ban'on and his coworkers employed somewhat different experi- mental conditions (personal communication). Investigations are at TRANSAMINATION 223 present under way by the writer to determine the influence of other vitamin deficiencies on this reaction. Role of Transamination in Intermediary Metabolism The exact role which transamination plays in intermediary metabo- lism is still not clear. The substrates of this reaction are highly reactive and participate in many diflFerent rapid metabolic proc- esses. Obviously if transamination is to play a significant role in cellular metabolism it must be shown to proceed at a rate rapid enough to be quantitatively significant. From Table 1 it is apparent that truly rapid rates are seen only with the system glutamic acid plus oxalacetic acid. The rapid rates at which oxalacetic, alpha- ketoglutaric, and pyruvic acids participate in non-transaminating reactions make it highly doubtful whether reactions 3 and 6 ever proceed fast enough to participate in the metabolism of these com- pounds. This may not apply to liver and pigeon breast muscle. Among the possible metabolic reactions which transamination may influence are those of protein and amino acid synthesis and degrada- tion, glycolysis, and hydrogen transport. protein and amino ACm SYNTHESIS AND DEGRADATION Animal Tissues.— An attractive theory of amino acid synthesis and degradation in plant and animal tissues has been proposed by Braunstein (1). According to this theory amino acids are synthesized or degraded by the transamination reaction in conjunction with the glutamic dehydrogenase system of Euler et al. (27) and Dewan (50). The latter system serves the two functions of synthesizing glutamic acid for transamination with alpha-keto acids to yield new amino acids, and of oxidizing the glutamic acid formed from alpha- ketoglutaric acid and different amino acids. Both enzyme systems are present in most tissues, although in varying amounts. Thus transaminase is higher in muscle than in liver, whereas glutamic dehydrogenase is higher in liver than in muscle. Braunstein and Bychkov (51) have reported production of am- monia from Z( + ) -alanine when the latter is incubated with alpha- ketoglutaric acid, glutamic aminopherase, glutamic dehydrogenase, pyocyanine, and cozymase. About 12.5 per cent of the theoretical yield was realized after three hours' incubation. These workers pointed out that the above system is a cell-free model of Z-amino acid dehydrogenase, which to date has not been demonstrated to 224 A SYMPOSIUM ON RESPIRATORY ENZYMES represent a single enzyme system, as is the case with ci-amino acid dehydrogenase. The mechanism of deamination of Z-amino acids in Hver and kidney can be explained on the basis of the above theory. However, the failure of the other tissues to oxidize amino acids other than glutamic acid to any appreciable extent is difficult to explain, since, as was said above, the two necessary enzyme systems are present in most tissues. Inasmuch as the theory rests in large measure on Braunstein's claim that all amino acids are active in transamination, unequivocal proof for this would seem desirable. As previously pointed out, experiments by the writer have shown transamination to be a limited reaction rather than a general one. Further, the author has found that homogenized liver and kidney, fortified with cozymase and methylene blue, failed to show any appreciable yields of ammonia from various combinations of alpha-ketoglutaric acid and Z- amino acids (63). A possible metabolic relationship between transamination and protein synthesis in animal tissues has been pointed out by Linder- str0m-Lang (52). On the basis of the plausible assumption that the synthesis of protein occurs by a metabolic mechanism other than the reversal of proteolysis, he postulates the following scheme: 0 OH SCHEME III -H2O R-C-C-OH + HpNR, I H O OH li I ^ R-C-C-N-R, I I H H -H2 O OH NHp O H \^ II I R-C-C-N-R, I H + R-C-C=N-R, .Glutamic , Acid NH2 OH I / R-C-C=N-R, H ex Ketoglutaric Acid TRANSAMINATION 225 As can be seen from this scheme, a keto-aldehyde reacts with an amino acid to form a ketonic SchifiTs base. The latter then reacts with glutamic acid via the transamination reaction to yield a pep- tide. Agren (19) has recently published unconvincing evidence in support of this. On the basis of the above reaction it would be expected that in tissues where rapid protein synthesis was taking place, e.g., embry- onic and tumor tissue, the transaminase activity would be higher than in normal adult tissues. Actually the reverse has been found to be the case (16). Thus it was observed that in tumors and em- bryonic tissue the transaminase activity was low as compared with normal adult tissues. This apparent inverse relationship between protein synthesis and transamination suggests that the latter re- action may serve as a controlling mechanism in protein synthesis. Plant Tissues.— The possible role of transamination in plant pro- tein synthesis is suggested from the following scheme of Virtanen and Laine (41) for leguminous plants: SCHEME IV Ng ^ Hydroxylamlne^^ ^(Oxime of Oxalaceiic Acid Carbohydrate — ^Oxalacefic Acidj^ +H2 \ \p l(-)Aspar+ic Acid + a Keto Acid 'Oxalaceiic Acid + a Amino Acid Experimental evidence for the above scheme has been reported by Virtanen and Laine (41, 53) and has been critically examined by Wilson (44). However, careful quantitative studies on trans- amination in plant tissues have not as yet been carried out. Until this is done the role of transamination in plant tissues will continue to remain obscure. OTHER REACTIONS Transamination and Glycolysis— Krehs (54) and Weil-Malherbe (55) observed that glycolysis in retina and brain tissue was in- 226 A SYMPOSIUM ON RESPIRATORY ENZYMES hibited by glutamic acid. This eflFect has been studied in more detail recently by Grodzensky (56), who was able to show that anaerobic glycolysis in pigeon breast muscle was also inhibited (20-50 per cent) by glutamic acid. Further, he was able to demon- strate that the inhibition was due to the conversion of pyruvic acid to alanine by transamination of the former with glutamic acid. The influence of glutamic acid on glycolysis in tumors has been investigated by the author (63). Inhibitions of the order of 10-15 per cent were observed. These results are in keeping with the finding that tumors have a low transaminase content (16). Transamination and Hydrogen Transporf.— Transamination is not only a very rapid reaction but it is also concerned chiefly with those substances that are known to play key roles in intermediary metabo- lism. Thus oxalacetic, alpha-ketoglutaric, glutamic, and aspartic acids all catalytically influence respiration (3, 4, 57, 58, 59). Further, glutamic acid has been found to function as a hydrogen carrier not only because its dehydrogenase can act with both di- and triphos- phopyridine nucleotides (27, 60) and so can couple with other di- and triphosphopyridine nucleotide-catalyzed systems (60, 61), but also because of its role in a dismutation reaction involving alpha- ketoglutaric acid and ammonia (62). It would thus appear that the chief role of transamination may be that of rapidly interconverting certain of the respiratory mediators. That the transamination reaction is fast enough to compete suc- cessfully with other metabolic reactions involving the same sub- strates has been previously indicated. REFERENCES 1. Braunstein, a. E., Enzymologia, 7, 25 (1939). 2. Herbst, R. M., Symposia on Quantitative Biology, 6, 32 (1938). 3. Needham, D. M., Biochem. J., 24, 208 (1930). 4. Annau, E., Banga, I., Blazso, A., Bruckner, V., Laki, K., Straltb, F. B., and Szent-Gyorgyi, A., Z. physiol. Chem., 244, 105 (1936). 5. Banga, I., and Szent-Gyorgyi, A., Z. physiol. Chem., 245, 118 (1937). 6. Braunstein, A. E., and Kritzmann, M. C, Enzymologia, 2, 129 (1937). 7. Braunstein, A. E., and Kritzmann, M. G., Biochimia, U.S.S.R., 3, 603 (1938). 8. Foreman, F. W., Biochem. J., 8, 463 (1914). 9. ZoRN, K., Z. physiol. Chem., 266, 239 (1940). 10. Cohen, P. P., Biochem. J., 33, 1478 ( 1939). 11. Cohen, P. P., Biochem. J., 33, 551 (1940). 12. Macara, T. J. R., and Plimmer, R. H. A., Biochem. J., 34, 1431 ( 1940). 13. Woodward, G. E., Reinhart, F. E., and Dohan, J. S., J. Biol. Chem., 138, 677 (1941). 14. Cohen, P. P., J. Biol. Chem., 136, 565 (1940). TRANSAMINATION 227 15. Cohen, P. P., and Hekhxhs, G. L., J. Biol. Chem., 140, 711 (1941). 16. Cohen, P. P., and Hekhuis, G. L., Cancer Research, J, 620 (1941). 17. Braunstein, a. E., and Azarkh, R. M., Nature, 144, 669 (1939). 18. V. EuLER, H., Hellstrom, H., Gunther, G., Elliott, L., and Elliott, S., Z. physiol. Chem., 259, 201 (1939). 19. Agren, G., Acta Physiol. Scand., 1, 233 (1940). 20. Cohen, P. P., Proc. Am. Soc. Biol. Chem., J. Biol. Chem., 133, xx ( 1940). 21. Braunstein, A. E., and Kritzmann, M. G., Biochimia, U.S.S.R., 4, 168 ( 1939). Chem. Abstr. 34, 1694 ( 1940). 22. Bychkov, S. M., Biochimia, U.S.S.R., 4, 189 (1939). Chem. Abstr., 34, 1694 (1940). 23. Kritzmann, M. G., Biochimia, U.S.S.R., 3, 603 (1938). 24. Kritzmann, M. G., Nature, 143, 603 (1939). 25. Kritzmann, M. G., Biochimia, U.S.S.R., 4, 667 (1939). Chem. Abstr., 34, 5865 (1940). 26. Cohen, P. P., J. Biol. Chem., 136, 585 (1940). 27. V. EuLER, H., Adler, E., Gijnther, G., and Das, N. B., Z. physiol. Chem., 254, 61 (1938). 28. Karrer, p., Koenig, H., and Legler, R., Helv. Chim. Acta., 24, 127 (1940). 29. Knoop, F., and Martius, C., Z. physiol. Chem., 254, I ( 1938). 30. Kritzmann, M. G., Enzvmologia, 5, 44 (1938). 31. Cohen, P. P., Am. J. Physiol., 126, 467 (1939). 32. Elliott, K. A. C, and Grieg, M. E., Biochem. J., 32, 1407 (1938). 33. ScHULZE, M. O., J. Biol. Chem., 129, 727 (1939). 34. Evans, E. A., Jr., and Slotin, L., J. Biol. Chem., 136, 301 (1940). 35. Krebs, H. a., and Eggleston, L. V., Biochem. J., 34, 1383 (1940). 36. Wood, H. G., Werkman, C. H., Hemingway, A., and Nier, A. O., J. Biol. Chem., 139, 483 (1941). 37. Breusch, F. L., Biochem. J.. 33, 1757 (1939). 38. Karayagina, M. K., Biochimia, U.S.S.R., 4, 168 (1939). Chem. Abstr., 34, 1694 (1940). 39. V. EuLER, H., GiJNTHER, G., and Forsman, N., Z. f. Krebsforsch., 49, 46 (1939). 40. Kritzmann, M. G., Biochimia, U.S.S.R., 4, 184 (1939). Chem. Abstr., 34, 1694 (1940). 41. ViRTANEN, A. I., and Laine, T., Nature, 141, 748 (1938). 42. Cedrangolo, F., and Carandante, G., Boll. soc. ital. biol. sper., 15, 482 (1940). 43. Adler, E., and Sreenivasaya, M., Z. physiol. Chem., 249, 24 (1937). 44. Wyss, O., quoted by Perry W. Wilson in The Biochemistry of Symbiotic Nitrogen Fixation ( University of Wisconsin Press, Madison, Wisconsin, 1940), p. 175. 45. Adler, E., GiJNTHER, G., and Everett, J. E., Z. physiol. Chem., 254, 27 (1938). 46. Adler, E., Hellstrom, V., GIjnther, G., and v. Euler, H., Z. physiol. Chem., 255, 14 (1938). 47. Vysshepan, E. D., Biochimia, U.S.S.R., 5, 271 (1940). Chem. Abstr., 35, 4788 (1941). 48. Kritzmann, M. G., Biochimia, U.S.S.R., 5, 281 (1940). Chem. Abstr., 35, 4788 (1941). 49. Barron, E. S. G., Ann. Rev. Biochem., JO, 15 (1941). 50. Dewan, J. G., Biochem. J., 32, 1378 (1938). 51. Braunstein, A. E., and Bychkov, S. M., Nature, 144, 751 (1939). 228 A SYMPOSIUM ON RESPIRATORY ENZYMES 52. Linderstr0m-Lang, K., Ann. Rev. Biochem., 7, 37 (1939). 53. ViRTANEN, A. I., and Laine, T., Biochem. J., 33, 412 (1939). 54. Krebs, H. a., Biochem. J., 29, 1951 (1935). 55. Weil-Malherbe, H., Biochem. J., 32, 2257 (1938). 56. Grodzensky, D. E., Bull. biol. med. expd. U.S.S.R., 9, 116 (1940). Chem. Abstr., 35, 2535 (1941). 57. Krebs, H. A., Biochem. J., 34, 775 (1940). 58. Krebs, H. A., and Eggleston, L. V., Biochem. J., 34, 442 (1940). 59. Baumann, C. a., and Stare, F. J., J. Biol. Chem., 133, 183 (1940). 60. Adler, E., v. Euler, H., GiJNTHER, G., and Plass, M., Biochem. J., 33, 1028 (1939). 61. Dewan, J. G., Biochem. J., 33, 549 (1939). 62. Krebs, H. A., and Cohen, P. P., Biochem. J., 33, 1895 (1939). 63. Cohen, P. P., unpublished studies. 64. Sober, E. K., and Cohen, P. P., unpublished studies. BuRK AND Elliott continue the tumor discussion. LlPMANN AND MeYEHHOF Whv is the Pasteur effect? Neuberg and Cori Who is con\incing whom? DISCUSSIONS NOT ON THE AGENDA Discussion on Tumor Respiration C. A. BAUMANN, University of Wisconsin, Chairman CHARACTERISTICS OF TUMOR RESPIRATION K. A. C. ELLIOTT Institute of the Pennsylvania Hospital, Philadelphia Having been out of the cancer research field for some time, I am not in a position to discuss many recent developments. But it may be useful to start this discussion with a resume of the generahzations and theories concerning the metabolism of cancer tissue that have been proposed from time to time. My remarks may appear somewhat pessimistic, for it does not seem to me that any definite and peculiar characteristic of tumor metabolism has been proved. However, later speakers who are familiar with more modern work will perhaps strike a more optimistic note. Warburg's First Theory.— As is well known, Warburg and his school discovered that, whereas slices of a number of normal tissues produce lactic acid from glucose or glycogen rapidly in the absence of oxygen (anaerobic glycolysis), cancer tissue slices show an un- usual ability to continue production of lactic acid from glucose in the presence of oxygen (aerobic glycolysis). In his early studies Warburg happened to study tumors which showed very low oxygen uptake rates. He therefore concluded that in cancer tissue the respiratory mechanism was impaired and that glycolysis took its place as a means for producing energy. Later, however, Warburg observed, and emphasized, the fact that the oxygen uptake rate of most tumor tissues under good conditions is not usually lower than that of many normal tissues. Warburg's Second Theory.— Warburg concluded that since the respiration rate of cancer tissue might be normal while rapid aerobic glycolysis continued, there must be something wrong with the type of respiration in the tumors. Some sort of damage to the respira- tory mechanism must have occurred in the production of tumor cells which caused a loss of eflBciency of respiration in suppressing gly- colysis. High anaerobic glycolysis appeared to be a general property of growing or multiplying tissues, since it was found in embryo tissue and testis, but in these normal tissues glycolysis was largely abolished when respiration occurred, that is, in the presence of oxygen. In 229 ^30 A SYMPOSIUM ON RESPIRATORY ENZYMES the preface to the Enghsh edition of The Metabolism of Tumors ( 1), which should be consulted for details of Warburg's work on tumors, Warburg stated that "interference with the respiration in growing cells is, from the standpoint of the physiology of metabolism, the cause of tumors. If the respiration of a growing cell is disturbed, as a rule the cell dies. If it does not die, a tumor cell results. This is no theory, but a comprehensive summary of all the measurements at present available." At present, however, the idea that respiration itself rather than the presence of oxygen inhibits glycolysis is frequently questioned. High anaerobic glycolysis is not a property of growing tissue alone, since it is found, for instance, in adult brain and in glycogen-rich livers. More important, the high aerobic glycolysis is not found in tumors alone. Warburg himself showed that mammalian retina glycolyzes very rapidly, aerobically as well as anaerobically, though he considered that some sort of damage occurred to this delicate tissue in preparing it for in vitro work. Gyorgy and co-workers (2) and Dickens and Weil-Malherbe (3) found high aerobic glycolysis in kidney medulla, and the latter authors (4) have recently found the same for jejunal mucous membrane. Testis and several other non-cancerous tissues are now known to show a moderately high aerobic glycolysis, and my co-workers and I have found that a fairly high rate of aerobic glycolysis occurs during the first few minutes of an experiment with various other normal tissues. Murphy and Hawkins (5), on the other hand, reported little or no aerobic glycolysis with some spontaneous tumors occurring in mice. Con- cerning the normal tissues which show continuous aerobic glycolysis, that is, an "anaerobic type of metabolism" by which energy is pro- duced by the anaerobic method even in the presence of oxygen, Dickens (3) considered that "the cause of an anaerobic type of metabolism is in all such cases merely a disparity between blood supply, i.e., oxygen supply, and energy requirements in vivo." Con- sideration of aerobic glycolysis as opposed to anaerobic glycolysis does not enter into the following theory of Dickens. Dickens' Theory.— Dickens and Simer (6) arranged the normal tissues into three groups. One group, which included kidney cortex and liver, showed little anaerobic glycolysis and respired with a rather low respiratory quotient and so apparently metabolized little carbohydrate. Another group, including brain, retina, chorion, and embryo tissue, showed high anaerobic glycolysis and a respiratory quotient of unity in glucose-containing medium. These tissues appar- DISCUSSION ON TUMOR RESPIRATION 231 ently respired at the expense of carbohydrate, and in the absence of oxygen the carbohydrate metaboHsm led to lactate accumulation. A third group, including spleen, testis, and submaxillary gland, formed an intennediate group. But cancer tissue seemed to consti- tute a separate class, having a high glycolysis and a low respiratory quotient. It was concluded that in cancer tissue the mechanism for starting carbohydrate metabolism is present, even over-developed, but the mechanism for oxidizing the split products of carbohydrate is lacking. However, Dr. Baker and I (7) found specimens of tumors which respired with respiratory quotient values ranging up to unity and pointed out that other workers, including Dickens himself, had found similar high values. On the other hand, we found values of somewhat less than unity for retina and brain. We concluded therefore, that Dickens' generalization was not valid.* Dickens himself, with Weil-Malherbe (4), has now found that normal jejunal mucosa, a rapidly respiring tissue, has a low respiratory quotient with high aerobic and anaerobic glycolysis. A similar type of metabolism was found with synovial membrane by Bywaters (8). Dickens has therefore abandoned his generahzation, since these are normal tissues showing the behavior that was believed to be specific for tumors. It thus appears that no characteristic of metabolism is truly spe- cific for cancer tissue. Nevertheless, most cancer tissues show only a moderately high respiration rate, a moderately low respiratory quotient, and a definitely high sustained aerobic and anaerobic glycolysis, whereas few normal tissues show all these characteristics together. Orr and Stickland (9) very recently found that tumors oc- curring in the livers of rats fed butter yellow possess the power, as do most cancer tissues, to form lactate from glucose, anaerobi- cally and aerobically, and differ distinctly in this respect from non- cancerous liver tissue. Dr. Dean Burk [see below, page 242] has * Dickens at first defended his tlieory by attacking our value for the respiratory quotient of brain sUces, while ignoring the rest and ridicuhng "a complicated theory of salt eflFects" which did not appear in our paper. Our values for the respiratory quotient of brain slices were supposed to disagree with otliers in the literature, but careful study showed that objections could be raised to all the results cited, except perhaps Dickens' own. Varying results could be due to the fact that early aerobic glycolysis by brain liberates carbon dioxide from the medium, and this must be carefully controlled if incorrectly high or low respira- tory quotient values are to be avoided. Observing this precaution, we ourselves later obtained slightly higher values ( 18 ) . Our figure for retina has been con- firmed by Dixon (19). 232 A SYMPOSIUM ON RESPIRATORY ENZYMES expressed the view that, even as the histology of cancer is suflBciently characteristic to enable a pathologist to recognize the tissue as cancer usually, but not always, so also the experienced student of tissue metabolism could almost always state correctly what is or is not cancer tissue from results of metabolic measurements on tissue sHces. Impaired Respiration Mechanisms.— Warburg's idea that the res- piration of cancer tissue is in some manner deranged may still be true, and the idea has inspired a number of studies of individual respiration mechanisms in cancer tissue. Dr. Benoy, Dr. Baker, and I (10) found the succinic oxidase system inactive in sHces of certain tumors, and Dr. Greig and 1(11) showed that most tumor suspensions tested were low in succinic dehydrogenase and the cytochrome- cytochrome oxidase systems. I (12) found that tumor breis added to liver or other tissue breis (except heart) rapidly destroyed the suc- cinic oxidase system of the liver or other tissues. Stotz (13) and Potter and DuBois (14) found low cytochrome c content in a number of cancer tissues. Banga (15) found that certain tumors are scarcely able to reduce added oxalacetate; this and the lack of succinic dehydrogenase would indicate impaired catalysis by the mechanism postulated by Szent-Gyorgyi. However, Dr. Greig and I found some tumors with a fair amount of succinic dehydrogenase and some normal tissues with little. Some normal tissues, especially pancreas, and also commercial trypsin, would inhibit liver succinoxidase; the inhibition might be a purely in vitro effect and auto-digestion might partly account for low suc- cinoxidase values found for tumor tissues. Potter and DuBois found at least one normal tissue (lung) with as low a cytochrome c content as tumor tissue, and Breusch (16) found that the rate of oxalacetate reduction was negligible also with some normal tissues, namely spleen, lung, placenta, and peripheral nerves. Dr. Baker and I (17) found that the effects of a number of dyes on the metabolism of tumor tissue were different from their effects on any of the normal tissues tested. Nothing further has come of these observations. Altogether the various results suggest that cancer tissue tends to differ from normal tissues in its respiratory mechanisms, but no very well-defined difference has yet been disclosed. REFERENCES 1. Warburg, O., The Metabolism of Tumors, translated by F. Dickens ( Richard Smith, Inc., New York, 1930). 2. Gyorgy, p., Keller, W., and Brehme, T., Biochem. Z., 200, 356 (1928). DISCUSSION ON TUMOR RESPIRATION 233 3. Dickens, F., and Weil-Malherbe, H., Biochem. J., 30, 659 (1936). 4. Dickens, F., and Weil-Malherbe, H., Biochem. J., 35, 7 ( 1941). 5. Murphy, J. B., and Hawkins, J. A., J. Gen. Physiol., 8, 115 (1925). 6. Dickens, F., and Simer, F., Biochem. J., 24, 1301 (1930); 25, 985 (1931). 7. Elliott, K. A. C, and Baker, Z., Biochem. J., 29, 2433 (1935). 8. Bywaters, E. G. L., J. Path. Bact., 44, 247 (1937). 9. Orr, J. W., and Stickland, L. H., Biochem. J., 35, 479 (1941). 10. Elliott, K. A. C., Benoy, M. P., and Baker, Z., Biochem. J., 29, 1937 (1935). 11. Elliott, K. A. C., and Greig, M. E., Biochem. J., 32, 1407 (1938). 12. Elliott, K. A. C., Biochem. J., 34, 1134 (1940). 13. Stotz, E., J. Biol. Chem., 131, 555 ( 1939). 14. Potter, V. R., and DuBois, K. P., J. Biol. Chem., 140, cii (1941). 15. Banga, I., Z. physiol. Chem., 244, 130 (1936). 16. Breusch, F. L., Biochem. J., 33, 1757 (1939). 17. Elliott, K. A. C, and Baker, Z., Biochem. J., 29, 2396 ( 1935). 18. Elliott, K. A. C, Greig, M. E., and Benoy, M. P., Biochem. J., 31, 1003 (1937). 19. Dixon, M., Biochem. J., 31, 924 ( 1937). PHOSPHORYLATION THEORIES AND TUMOR METABOLISM VAN R. POTTER McArdle Memorial Laboratory, University of Wisconsin I should like to submit briefly for your consideration a working hypothesis concerning the metabolism of tumor tissue. This hypothe- sis is grounded in the Embden-Meyerhof scheme of carbohydrate breakdown, the Warburg descriptions of tumor metabolism and the concept of phosphate energy transfer recently put forth by Johnson. It is difficult if not impossible actually to prove such theories, and hence one must be satisfied with data that do not prove but are merely compatible with the given concept. Only after a great mass of circumstantial evidence has accumulated can we begin to have confidence in the theory. During the accumulation of these data the hypothesis is necessarily modified in the light of incompatible data. It is such incompatible data which probably will soon be brought to bear upon the hypothesis I am about to present. According to this hypothesis, tumor tissue uses its adenosine tri- phosphate (ATP) reservoir for but two main purposes, growth and glucose phosphorylation, in contrast with most other tissues, which in addition have function and thus do work and presumably split ATP in doing it. Since the growth stimulus is ever present in tumor tissue, inorganic phosphate is released in large enough quantities to permit rapid glycolysis (in the sense of carbohydrate cleavage), 234 A SYMPOSIUM ON RESPIRATORY ENZYMES yet the growth does not deplete the ATP to such an extent that glucose cannot be phosphorylated and hence glycolyzed. In other tissues glycolysis is slowed down either by a depletion of inorganic phosphate, as in resting tissue, or by a depletion of ATP, as in dying tissue. (Liver and kidney under anaerobic conditions deplete ATP by tending to maintain function and are hence unable to phos- phorylate glucose; therefore glycolysis stops and death occurs.) In the tumor tissue the glycolytic rate is so rapid in relation to the oxidative mechanisms that lactate accumulates. The oxygen uptake is limited by the amount of the oxidative enzymes present, but since there is an excess of substrates, the Q02 is higher than might be expected on the basis of the Qoo of normal tissue, in which the oxidative enzymes are present in excess and the Q02 is limited by the amount of substrate furnished by glycolysis. In the tumor tissue the growth process outpaces the synthesis of the oxidative enzymes, and the latter become diluted as compared with their concentration in other active tissues. From this it would follow that growth may not require as high a level of oxidative enzymes as does function. Our experimental results are being reported elsewhere. At this point it may be said that one component of the oxidative mech- anism, namely, cytochrome c, appears to have been established as definitely lower in the various types of tumor tissue than in normal tissues. Assays on the succinoxidase system are at present being carried out. Preliminary experiments with rapidly growing hver support the idea that growth outpaces the synthesis of the oxidative enzymes temporarily in this tissue. DISCUSSION ON TUMOR RESPIRATION 235 ON THE SPECIFICITY OF GLYCOLYSIS IN MALIGNANT LIVER TUMORS AS COMPARED WITH HOMOLOGOUS ADULT OR GROWING LIVER TISSUES* DEAN BURK National Cancer Institute, National Institute of Health, U. S. Public Health Service, and Cornell University Medical College In this discussion I wish to focus attention upon one particular aspect of tumor and growth metabohsm that is simple but far- reaching in implication. It is a problem more of comparative bio- chemistry than of intermediate metabolism proper, and concerns the origin of tumor metabolism. The question I wish to pose, and hope to succeed in answering here, is whether the large glycolysis of tumors is necessarily an expression and requirement of their extensive and usually rapid growth. It has been widely held, since the middle period of Warburg's tumor work (1925), that growing tissues in general have a high anaerobic (and sometimes aerobic) glycolytic activity. This very active metabolism has in turn been attributed to extra and special metabolic requirements of the growth process. If it could be found that certain growing tissues do not exhibit marked glycolysis, then it might well be said that the glycol- ysis of tumors is not necessarily a consequence merely of extensive growth, but that it has a more specific and characteristic significance for tumor metabolism than has been recognized or acknowledged. Before presenting data bearing directly on the foregoing question, a related aspect of the problem of the origin of tumor glycolysis, and of suitable criteria for ascertaining significant differences between normal and tumor metabolism, should be discussed by way of back- ground (cf. also ref. 17). In recent years Berenblum, Chain, and Heatley (1) have made the claim that "valid comparisons can only be made between any particular tumor and the normal tissue from which it is derived" (la, p. 370). This emphatic assertion, which I believe it is very desirable to contravert at this early stage in its possible development, is surely dogmatic and arbitrary to say the least, for there are many valuable comparisons to be drawn between tumor materials and adult tissues widely separated from them embryologically, as well as between tumors and tissues as closely homologous as possible. I for one would not undertake to say which type of comparison would, in fact, be the more profitable in the long run, let alone advocate the exclusion of either one. Certainly * For much valuable help in the preparation of this manuscript I am greatly indebted to Miss Juliet M. Spangler, Senior Cancer Aide, 236 A SYMPOSIUM ON RESPIRATORY ENZYMES both types of comparison, involving non-homologous as well as homologous contrasts and similarities, must be made. In furtherance of their position, Berenblum, Chain, and Heatley advance the view that "the tumors which have hitherto been found to have a glycolyzing type of metabolism associated with a low R.Q, possess these properties in virtue of their origin from noiTnal tissues which also possessed these metabolic characters" (Id, p. 138). In experimental support of this view they reported, following Crab- tree's earlier measurements showing that whole skin undergoes little alteration of metabolism when it becomes papillomatous, that noniial skin epithelium and Shope papilloma of the domestic rabbit also possess essentially the same quantitative metabolism in regard to aerobic and anaerobic glycolysis, respiration and respiratory quo- tient. Unfortunately these data, although interesting enough in themselves, have no great bearing on the really pertinent problem as to the difference (or similarity) between a definitely malignant tumor and a closely homologous normal tissue; indeed the support- ing experiments are themselves somewhat unsatisfactory, being diffi- cult to analyze because of the unorthodox technical method em- ployed and the fact that Q values were based on nucleic acid- phosphorus content instead of on dry weight. Certainly there are advantages in the use of the nucleic acid-phosphorus criterion, but it is unfortunate that the dry weight values were not at least reported so that the reader could make Q value comparisons by the standard methods and check, in particular, the bare and doubtful state- ment (lb, le) that the normal skin epithelium and Shope papilloma metabolic values were "very similar to those for many skin carci- nomas quoted in the literature." (By certain inferences, the anaerobic glycolysis of the skin epithelium and papilloma studied would appear to have been at most Q^^a = 1 to 3, or quite low for the usual malig- nant tumor.) The two criticisms of the unorthodox (however correct) procedures employed are admittedly minor as compared with the fact that the Shope papilloma, as such, is not malignant, nor was it so described. Table 1, now presented for discussion, provides, in regard to primary rat liver tumors, not merely one but several types of homol- ogous tissue, including adult normal liver and two types of growing liver, regenerating and embryonic. These materials will provide, I believe, as pertinent cases as are yet available for the comparison of a malignant tumor with an homologous, in fact identical, tissue of origin (liver). DISCUSSION ON TUMOR RESPIRATION 237 The data in Table 1 are the result of the collaboration of many investigators, as indicated in the footnotes, and have been or are being detailed elsewhere under respective authorships. They are brought together here for the purpose of a broad and unified inter- pretation and discussion at this meeting. Adult Liver The azo dye tumors reported upon in the table ofiFer an excellent opportunity to determine not only whether their metabolism is dif- ferent from the tissue of origin but also, if it is, to ascertain at what stage or stages of tumor development the altered metabolism ap- pears. The hepatomas, adenocarcinomas, metastases therefrom, and necrotic material thereof, obtained from rats fed butter yellow, and also the mouse tumor transplant derived originally from o-amino- azotoluene feeding, all show the high anaerobic glycolysis and low or intermediate respiratory quotient characteristic of malignant tumors, and a considerably increased aerobic glycolysis as compared with either normal rat or mouse liver. The respiration is not changed significantly. There is a definite but relatively small anaerobic lactic acid formation in "prc-cancerous," cirrhotic hver as compared with normal liver; there is likewise a slight but quite definite increase in "normal" lobes adjacent to tumor-bearing lobes and in livers of rats protected against tumor formation and extensive liver damage by butter yellow feeding. The aerobic glycolysis increases rather abruptly, essentially at the onset of gross tumor formation, and cer- tainly more abruptly than in the case of the anaerobic glycolysis, where there is a small, but perfectly definite, several-fold increase in the pre-cancerous liver as compared with the normal. The results on tumor in Table 1 in good part confirm and extend the well-known results of Nakatani, Nakano, and Ohara ( 10), who obtained, in fact, relatively more pre-cancerous change in anaerobic glycolysis, though hkewise none in the aerobic. For comparison with absolute and not relative Q values, it is necessary to reduce the Japanese values by four- to five-tenths to put them on an initial dry weight basis corresponding to those in Table 1. Orr and Stickland ( 13, and previous preliminary communications) reported, contrary to the results of Nakatani et al. and our own, that they did not observe any change in the glycolytic metabolism of liver tissue in the pre-cancerous stage of butter yellow treatment. For the basis of their comparisons, however, they reported that their normal hvers yielded anaerobic glycolysis Q values of 2-16 238 A SYMPOSIUM ON RESPIRATORY ENZYMES (presumably 1-8-12, if corrected to initial dry weights). These results for normal liver, with the possible exception of some very early work of Rosenthal ( 14), are practically unique in the literature of liver metabolism. In our own experiments, Nakatani et ah, and (so far as I know) essentially all others, anaerobic glycolysis Q values of more than 2 (initial dry weight basis) have never been consistently reported. In our experience with many hundreds of normal rat livers under a great variety of dietary conditions, Q^^a values of 1 or considerably less were regularly obtained in rats weighing over 50 grams (using the ordinary manometric methods, and with varying periods of oxygenation between the killing of the rat and the estab- lishment of anaerobiosis). Without attempting to account at this Table 1.— A comparison of the metabolism* of various rat liver tumors with various homologous liver tissues (normal, embryonic, aged, cirrhotic, regenerating) Rat Tissue Weight (grams) Q% Q02 R.Q. QN^A M.O.Q U Average Normal Liver (2-24 mos.) 50-350 1.5 6.0 0.70 1.0 -0.3 -11.0 "Butter Yellow" Liverf 75-250 "Normal," yeast protected 1.6 4.9 0.98 2.1 0.3 - 7.7 "Normal" lobe, adjacent to tu- mor lobe 1.3 7.7 0.82 2.3 0.4 -13.1 Cirrhotic 1.5 6.5 0.82 3.1 0.7 - 9.9 Necrotic tumor 3.9 4.2 0.77 6.0 1.5 - 2.4 (Mouse transplantf) 2.1 4.7 0.66 7.7 1.2 - 1.7 Metastases (to omentum and mesentery) 2.5 5.0 0.83 8.7 3.7 - 1.3 Adenocarcinoma-hepatoma 3.3 7.3 0.84 10.0 2.8 - 4.6 Hepatoma 6.0 6.4 0.87 12.1 2.9 - 0.7 * Q values based on initial dry weights (original data of Tamiya (16) based on final dry weights and here factored by 50 per cent to reduce to approximate initial dry weight; cf. refs 5 and 8c).Qoj,Q'^2AandQ°2^ = inm.' oxygen consumption, anaerobic and aerobic acid production/mg. initial dry weight of tissue per hr.; R.Q. = respiratory quotient = Qo2/Qc02; M.O.Q. = Meyerhof oxidation quotient = 3(QN2A.— Q°2a)/Qoj; tJ = fermentation excess = Q'^'^a — 2Q02. t Sample of average data taken with O. K. Behrens and K. Sugiura (cf. ref. 4 et seq.). Up to 30 specimens of each tissue type examined metabolically. Rats fed 0.06 per cent p-dimethylaminoazobenzene (butter yellow) for 150-200 days on brown rice-carrot diet. "Yeast-protected" with 5-15 per cent added dried brewer's yeast. J 33d generation, subcutaneously transplanted, o-aminoazotoluene-induced liver carcinoma 1 in dba strain. Mouse obtained from Dr. H. B. Andervont through Dr P. M. West; liver of same animal: QN'a = 0.8; QO2a = 1.0. DISCUSSION ON TUMOR RESPIRATION Table 1. — continued 239 Rat Tissue Weight Q°U Qo2 RQ- QN^A M.O.Q. U (grams) Regenerating Livers§ Days Per cent regeneration of unexcised livers 2 180 361 0.8 4.5 0.38 0.6 -0.1 - 8.4 11 280 300 0.5 5.2 0.72 0.7 0.1 - 9.7 10 330 205 0.7 2.3 0.67 0.6 -0.1 - 4.0 1 170 76 3.6 0.88 0.7 - 6.5 3 280 34 1.5 6.4 0.64 1.8 0.1 -11.0 Embryonic and Post-embryonic Liver** Liver Per cent Rat age weight body (grams) weight Foetal 0.060 7.4 0.80 8.9 Foetal 0.093 8.1 1.09 10.0 Foetal 0.110 8.7 1.26 0.6 6.0 1.00 8.1 1.3 - 3.9 Foetal 0.185 8.7 2.12 0.6 6.6 0.98 8.0 1.2 - 5.2 8 hours 0.240 5.3 4.6 6.4 Iday 0.186 3.7 4.7 2.7 1 day 0.215 3.9 5.5 1.2 6.9 0.66 2.3 0.4 -11.5 3 days 0.262 4.8 5.4 1.7 9 days 0.350 2.5 14.0 1.0 21 days 1.052 3.5 30.0 1.7 5.7 0.52 0.8 0.1 -10.6 Embryonic Chicken Liver (16) Embryo age Liver weight 5 days 0.048 mg. 9.5 0.093 7.6 0.131 0.7 7.7 5.4 1.8 -10.0 6 days 0.398 5.0 0.580 0.0 6.8 4.0 1.8 - 9.6 7 days 0.720 4.3 0.960 0.0 6.1 4.0 2.0 - 8.2 8 days 1 . 40 3.4 2.25 2.8 (Adult) 8000.00 0.0 7.2 1.4 0.6 -13.0 Chick Bone Marrow Erythro- blaststt 1.9 4.2 0.84 8.2 4.5 - 0.2 Rabbit Bone Marrow Erythroid Cells (nucleated) (18) 0.0 9.0 7.0 2.3 -11.0 § Sample of data taken with J. Blanchard, C. Povolny, J. Norris, and J. Saxton (12). Rats 10-600 days old (15-400 grams body wt.); regeneration 1-11 days after original 65 per cent hepatectomy (left and median lobes extirpated). ** Data taken with J. Norris (11, 12). tt Average of data taken with H. Sprince, E. A. Kabat, and J. Furth on chickens treated with acetylphenylhydrazine to produce hyperplastic (not leukotic) bone mar- row, with some leukogenic cells but mainly erythroblasts. To be published. 240 A SYMPOSIUM ON RESPIRATORY ENZYMES time for the exceptional results of Orr and Stickland with normal hvers, it can very definitely be stated on the basis of our results and those of Nakatani et ah, that there is a one or more fold increase in the anaerobic glycolytic metabolism (and likewise glucolytic, for those interested in this distinction) of liver tissue in the pre-cancerous stage of butter yellow treatment. That this increase, unless due to a rather inconceivable difference in rat strains, "must have been due to chance," as proposed by Orr and Stickland (13, p. 486), is out of the question, and we prefer the alternate view that the very unusual magnitude and spread of normal liver values obtained by Orr and Stickland have served to confuse rather than to clarify the results they observed with the pre-cancerous livers. A second and regular point of difference in the pre-cancerous livers observed by us was that the initial rate of glycolysis was better maintained over a period of several hours, whereas in the normal livers the Q^^A values dropped to zero or a few tenths in the course of an hour or two, and this relative effect was even more striking when the Q values were based on chemically measured lactic acid rather than on manometric acid production. It is conceivable, in the absence of information to the contrary, that the Orr and Stickland determina- tions on normal liver do not refer to measurements over sustained periods of time (hours), and that in some way the high normal values reported by them involve incidental aspects of initial or preparatory phases of technique, in some measure connected, to be sure, with the glycogen content of livers, as they demonstrated; I hesitate to suggest explicitly the trite explanation of extensive damage, but evidently some factor is operating to give them profoundly atypical (this is not to say incorrect) values for normal liver that certainly make comparisons with other kinds of liver material difficult. Regardless of the foregoing discrepancy of result in regard to pre-cancerous livers, all investigators agree that lactic acid formation from glucose by malignant tumors induced by butter yellow is strikingly different from that of normal liver and definitely or con- siderably increased over any form of pre-cancerous liver, and that the same is true in less marked degree with respect to aerobic lactic acid formation. In general, the anaerobic and aerobic lactic acid productions by these hepatomas are, on an absolute basis, inter- mediate between those of most rat, human, and chicken malignant tumors, on the one hand, studied by Warburg and many others afterward, and, on the other hand, those of certain mouse tumors studied originally by Murphy and Hawkins, and by Crabtree and DISCUSSION ON TUMOR RESPIRATION 241 Cramer and others later. The fermentation excess, U, is in fact not positive but zero or shghtly negative, and the Meyerhof oxidation quotient is nearer 3 than 6. In regard to the low or intermediate respiratory quotient of the azo dye tumors of Table 1, I might comment, in view of the frank discussion and expression of personal opinion desired here, that in the recent discussions in Nature on the metabolism of tumors by Dick- ens (7), Boyland (2), Berenblum, Chain, and Heatley (lb), and Dickens and Weil-Malherbe (8b), I agree in general with the com- ments of Dickens and disagree with the other commentators where they take exception, for in my judgment they fail to introduce the proper quantitative perspective. However, I do not feel that Dickens has been correct, during the past decade, in his view that "cancer tissue has a respiratory quotient indicating that the oxidation of carbohydrate is abnormal" (7, p. 512). I prefer to regard the low or intermediate respiratory quotient exhibited by the majority of malignant tumors as being unchanged from the similar low or intermediate respiratory quotient values of the great majority of normal adult tissues, those, in fact, cited by him over a decade ago. So why refer to them as "abnormal"? Why not consider them as simply unaltered? It is the glycolytic capacity, not the respiratory quotient, of tumors which by and large has changed or is "abnormal" or different from normal adult tissue; it is in most growing normal tissues that the respiratory quotient has tended to rise to or attain unity, and the oxidation of carbohydrate to become relatively more pronounced, and also the capacity for glycolysis (mainly anaerobic). General confusion on these matters has led some, including Dr. Elliott, in his intentionally pessimistic comments this morning, to suggest, with reference to the very recent paper of Dickens and Weil-Malherbe (8c) that the high glycolysis and low respiratory quotient found by them for jejunum mucosa put the metabolism of this tissue into the class of malignant tumor metabolism; but the very high absolute Q value for respiration (about equal to the high anaerobic glycolysis Q value) and the absence of a Pasteur effect make this designation, in my opinion, quite impossible. Likewise, if not one or two but a sufficient number of metabolic criteria (absolute and relative) are considered, it is impossible to agree with the recent tendency (la, and possibly 7) to regard cartilage (data, 6, 8a) and synovial membrane (data, 6b) as, like the alleged skin epithehum (1), providing rather good examples of normal tissues with "mahgnant tumor metabolism," even after appropriate correction for inert ma- 242 A SYMPOSIUM ON RESPIRATORY ENZYMES terials in these tissues and, I must add, also in the tumors taken for comparison! I know of no normal tissue whose metabolism, fully regarded, need as yet be confused with that of malignant tumors.* Further background for the foregoing interpretation of quite recent data is detailed elsewhere (3, 4, 5). Growing Liver One type of homologous tissue has been presented, but it might still be argued that the normal liver, although homologous, was not a growing tissue and not as comparable with liver tumor as might * I may reiterate a statement I have already made on many occasions, namely, that I believe that tlie metabolic diagnosis of malignant tumor as compared with normal tissue may be correlated with pathologic diagnosis in well over 95 per cent of tested cases (and, I venture to say, as yet untested cases), upon due consideration of the absolute as well as the relative magnitudes of, first and fore- most, anaerobic glycolysis (8-20 ± ) and of respiratory quotient (0.75 - 0.9 ± ) ; secondly, respiration (2 - 10 ± ), and aerobic glycolysis (0 - 15 ±); and thirdly the derived quotients, absolute Pasteur effect (8 - 15 ± ), Meyerhof oxidation quotient (M.O.Q. ) (3-6 ±), fermentation excess (U) ( — 5 to + 25 ± ) etc., (Q values based on initial dry weights); and fourthly quite possibly the new criterion developed by Salter et al. (15) in regard to separation of certain tumors from their homologues on the basis of differential oxidation of glucose and succinate. Non-tumor tissues can be excluded from malignant tumor tissue designation by one or more of these metabolic criteria; thus, to consider previously debated cases: for the kidney medulla, too high an R.Q.; cartilage, synovial membrane, and (presumably) skin epithelium, too low an anaerobic glycolysis ( Q^^a ) or respiration ( Q02 ) even with reasonable correction for inter- cellular substance and inert components; retina and jejunal mucosa, too high a respiration and in tlie latter case also M.O.Q. =: 0 (no Pasteur effect). The recent discussion, pro or con ( Ic, Id, 7, 2, lb, 8b), and emphasis laid, on aerobic glycolysis would in my opinion be much better transferred to anaero- bic glycolysis, which without exception, to my knowledge, is always considerable in malignant tumors. From my point of view aerobic glycolysis is almost in- variably merely an expression ( consequence ) of how much anaerobic glycolysis goes on in relation to how much oxygen consumption is occurring in tlie par- ticular tissue under examination (3b, 4). In malignant tumors for example, it can be said that the anaerobic glycolysis values are so high relative to the respira- tion that the latter is unable to inhibit completely the glycolysis under aerobic conditions, even with extensive operation of the Pasteur effect ( M.O.Q. = 3 — 6) ( 5 ) ; tlie aerobic glycolysis thus resulting is a quantity dependent upon two rather independent functions, oxygen consumption and anaerobic glycolysis. I might add, parenthetically, that most of the aerobic glycolysis values reported as zero in the literature ( including the often quoted mouse data of Murphy and Hawk- ins ) are in fact definitely positive due to a methodological error of not correcting the calculations for the fact that the R.Q. is ordinarily definitely less than unity, and hence the aerobic glycolysis greater than otherwise calculated. Unfortunately I cannot go here more deeply into details of elaboration needed to treat adequately the subjects discussed in this footnote and the two sentences tliat gave rise to it, but shall do so when the butter yellow tumor data summarized in Table 1 are described at length. DISCUSSION ON TUMOR RESPIRATION 243 be desired. Two more types of homologous tissue are presented in Table 1 by the data on regenerating hver and embryonic hver, both of which tissues may at certain stages attain growth rates even greater than that of liver tumor. The data on regenerating liver are very striking in that they show, as compared with normal adult liver, no appreciable alteration in any of the metabolic values studied. In the case of the very young rats, the regenerating Q^^a value is slightly increased, but mainly as a matter of neonatal age rather than of regeneration. Regenerating liver is indeed a remarkable case, demonstrating that tissue growth may take place without appreciable glycolysis, and at the expense, even, of unchanged oxygen consump- tion; for the growth increase in liver tissue (on a water-free basis) may attain 50 to 100 per cent per day during the most active phases of regeneration at about two to three days after partial hepatectomy, when mitotic figures are most numerous, several being visible on a high power field, or even more than would be found with a butter yellow liver tumor. Orr and Stickland (13) reported that the glycoly- sis of regenerating liver was not appreciably different from that of the normal livers they examined, but in these data the issue is again confused and rendered indefinite by their exceedingly high and variable normal liver values. A third type of homologous tissue is the rapidly growing embry- onic Hver. Tamiya (16) showed over a decade ago that there is a marked rise in the anaerobic glycolysis of chicken embryo livers the younger the embryo and the smaller the liver (at least back to a very early stage or microscopic size). The recent experiments of J. Langdon Norris (11, 12) give essentially the same results for embryonic rat livers as for the embryonic chicken livers. Contrary, however, to opinion held since the work of Tamiya (16) and Hawkins (9) that embryonic liver in particular and growing tissues in general show considerable anaerobic glycolysis, the histopathologic sections taken by Norris (illustrated elsewhere, 11, 12) show that the smaller the embryonic livers and the greater the anaerobic glycolysis, the greater and in parallel manner is the extent of haematopoesis, which in the extreme may amount to an estimated 70 to 80 per cent of the liver, involving mainly erythropoietic cells (with some megakary- ocytes and myelogenic cells). The metabolism of these nucleated erythropoietic cells has not yet been measured directly, but two comparable types of nucleated erythroid cells are available for com- parison, namely, the chick bone marrow erythroblasts produced by acetylphenylhydrazine injections and the normal rabbit bone mar- 244 A SYMPOSIUM ON RESPIRATORY ENZYMES row erythroid cells. In both these cases the anaerobic glycolysis Q values are of the same order of magnitude (7-10) as that of the embryonic Hvers containing a high percentage of red cells, and the other metabolic values are hkewise comparable. Direct measurement of the metabolism of the red cells in the embryonic livers offers con- siderable diflBculty experimentally; moreover, when they are ob- tained, the measurements might still be somewhat uncertain because of possible secondary effects of the technical methods employed to separate them from the liver. But it is felt that in the light of the two quite comparable cases offered and a considerable background of knowledge regarding the metabolism of blood cells generally, there is no reason to doubt that the rather high glycolytic metabohsm of embryonic hvers is due to the erythropoietic element and not to the true liver cells. In other words, neither the embryonic liver per se, nor the regenerating Hver, nor adult normal liver, nor in fact any healthy liver, growing or otherwise, possesses a noteworthy glycolysis, in contrast with the various malignant hepatomas, where a large glycolytic capacity obtains. The case of Berenblum, Chain, and Heatley ( lb) that "when a tumour is compared with the tissue from which it is derived, there are no metabolic characteristic differences or pecuHarities between the carbohydrate metabolism of the two" is clearly not valid; nor, for lack of evidence, is the more general contention quoted earlier that tumors glycolyze by virtue of their origin from normal tissues which also possess this metabohc charac- ter* (Id, p. 138). The question whether the increased formation of lactic acid in the hepatic tumors— or in tumors generally— is necessarily an expres- sion or requirement of the growth involved, as is commonly believed, is thus answered in the negative by the experiments briefly described in Table 1. It may be concluded, more generally, that growth does not necessarily require glycolysis, and may on occasion be main- tained at the highest levels on an essentially aerobic non-fermenta- tive metabolism. If it should be asked why malignant tumors possess glycolysis if not because of growth, I would venture the opinion that the glycolysis is better correlated with the more primitive organization or lesser differentiation involved. Most of the foregoing discussion has been concerned with com- * I do agree with Berenblum et al. — and not with Dickens, as aheady indi- cated— tliat the medium-low respiratory quotient of tumors may well be derived from the tissues of origin, in the sense that it remains by and large unchanged in tumors. DISCUSSION ON TUMOR RESPIRATION 245 parative biochemistry, not intermediate metabolism. The studies made in recent years by Elhott and by Potter, and especially this year by Salter and collaborators (15), on the deficient (cytochrome- succinate) oxidation systems in many tumors as compared with tissues of varying degrees of homologousness, will undoubtedly lead the way in indicating by what mechanisms, if any, the respiration systems of tumors permit high glycolysis to occur. But that is a story for the future. The paper of Mr. Kensler now to follow, based on his study with Dr. C. P. Rhoads (9a) on the action of butter yellow intermediates (free radicals) in correlating metaboHsm and carcinogenic action, will present an exceedingly promising pioneer work that opens up an entirely new approach to the connection be- tween tumor metabolism and tumor genesis, an approach that will, I believe, command the admiration of us all. REFERENCES 1. Berenblum, I., Chain, E., and Heatley, N. G., (a) Amer. J. Cancer, S8, 367 (1940); (b) Nature, 145, 778 (1940); (c) Ann. Rep. Brit. Emp. Can- cer Campaign, 16, 215 (1939); (d) ibid., 17, 135 (1940); (e) Abstr. 3d Inter. Cancer Congress, Atlantic City, p. 127 ( 1939). 2. BoYLAND, E., Nature, 145, 512 (1940). 3. BuRK, Dean, (a) Occas. Publ. Amer. Assoc. Adv. Sci., No. 4, 121 (1937); (b) Cold Spring Harbor Symposia on Quantitative Biology, 7, 420 (1939). 4. BuRK, Dean, Behrens, O. K., and Sugiura, K., Cancer Res., 1, 733 (1941) et seq. 5. Burk, Dean, Springe, H., Spangler, J. M., Kabat, A. E., Furth, J., and Claude, A., J. National Cancer Institute, 2, 201 (1941). 6. Bywaters, E. G. L., (1) Nature, 138, 30 (1936); (b) J. Path. Bact, 44, 247 (1937). 7. Dickens, F., Nature, 145, 512 (1940). 8. Digkens, F., and Weil-Malherbe, H., (a) Nature, 138, 125 (1936); (b) ibid., 145, 779 ( 1940); and (c) Biochem. J., 35, 7 ( 1941 ). 9. Hawkins, J. A., J. Gen. Physiol., 9, 111 ( 1926). 9a. Kensler, C. J., Dexter, S. O., and Rhoads, C. P., Cancer Res. 2, 1 (1942). 10. Nakatani, M., Nakano, K., and Ohara, Y., Gann, 32, 240 (1938). 11. Norris, J. Langdon, "Metabolism of rat livers, with particular reference to embryonic liver and malignancy," Polk Prize Paper, Cornell University Medical College, Sept., 1941. 12. Norris, J. L., Blanchard, J., and Povolny, C, "Regeneration of rat liver at different ages and metabolism of embryonic neonatal and regenerating rat liver," Amer. J. Path., Opie Dedication Number (in press). 13. Orr, J. W., and Stickland, L. H., Biochem. J., 35, 479 ( 1941). 14. Rosenthal, O., Biochem. Z., 207, 263 (1929). 15. Salter, W. T., Craig, F. N., and Bassett, A. M., Cancer Res., 1, 751 (1941); cf. 1, 869 (1941). 16. Tamiya, C, Biochem. Z., 189, 175 ( 1927). 17. Voegtlin, Carl, Physiol. Rev., 17, 92 (1937). 18. Warren, C, Amer. J. Physiol., 131, 176 (1940). 246 A SYMPOSIUM ON RESPIRATORY ENZYMES THE EFFECTS OF CERTAIN DIAMINES ON ENZYME SYS- TEMS, CORRELATED WITH THE CARCINOGENICITY OF THE PARENT AZO DYES C. J. KENSLER Memorial Hospital, New York In 1935 Hashimoto (1) reported the isolation of acetyl-2-methyl-p- phenylenediamine from the urine of rats fed the carcinogen, o- aminoazotoluene. Another azo carcinogen, dimethylaminoazoben- zene (butter yellow), the metabolism of which has been studied by Stevenson, Dobriner, and Rhoads (2), has also been found to be split in vivo at the azo linkage. The urine of animals fed butter yellow contained aminophenol and p-phenylenediamine. The free and acetylated forms of both compounds were found. No dimethyl-p- phenylenediamine was isolated, but it may be assumed until evi- dence to the contrary is presented that it is a precursor of the excreted p-phenylenediamine. Orthoamidoazotoluene and dimethyl- aminoazobenzene are the only carcinogenic azo compounds whose metabolic breakdown has been studied. A previous report (3) from this laboratory presented evidence that the concentration of diphosphopyridine nucleotide (Coenzyme I) in the livers of rats fed dimethylaminoazobenzene is 60 per cent less than in the livers of rats fed the same basal diet without the carcinogen. This fact suggested that the administered dimethyl- aminoazobenzene, or some metabolic breakdown product of it, might depress in vitro the activity of a fermenting system from yeast in which diphosphopyridine nucleotide is the limiting factor. Experi- ment proved the validity of this suggestion. The methods and the results obtained are presented in detail in a report which will appear shortly (4). In brief, it was found that whereas the original carcinogen, butter yellow, did not inhibit fermentation at all, the isolated derivative of butter yellow, p-phenylenediamine, was strongly inhibitory, and dimethyl-p-phenylenediamine, the supposed precursor of the simpler compound, had an even more powerful toxic eflFect. In view of the results obtained with butter yellow and its break- down products, it seemed desirable to test on the same feraienting system the eflFect of chemically related substances. Dr. Leonor Michaelis, who had previously prepared a large number of methyl derivatives of p-phenylenediamine as a part of his general study of two-step oxidations, generously provided samples of these com- pounds for use in this investigation. Their inhibitory eflPects on the DISCUSSION ON TUMOR RESPIRATION 247 rate of fermentation in the yeast system were measured. In Table 1 are presented the results of tests of only those substances of which the carcinogenic potency of the primary azo compound for rat livers has been tested. However, in further tests of the inhibitory effect on the yeast system of seventeen of the methyl derivatives of p-phenylenediamine, as well as of those hsted, the inhibition was found to correlate closely with the stabihty of the semiquinone intermediary oxidation product as reported by Michaelis, Schubert, and Granick (5). In the experiments with the yeast system it was noted, further- more, that those compounds which were toxic and which formed stable free radicals (semiquinones) were oxidized readily by the yeast apozymase used. The substances which were shown by Michaelis and his associates not to form stable free radicals were not so oxidized. This suggested, therefore, that the stable free radi- cal or some further oxidation product of it was responsible for the inhibition observed. Among the possible end products of the oxidation that might be responsible for the inhibition are quinone, methylamine, dimethyl- amine, and formaldehyde. Of these compounds only quinone was found to have significant toxicity. This compound, however, was less than half as toxic as the diamine in equivalent concentrations. Hydro- gen peroxide, which can be formed under certain conditions of oxidation, was also proved to be non-toxic. The acetylated p-phenyl- enediamine, which is quite stable and is a metaboHte of dimethyl- aminoazobenzene, was found to be non-toxic. This inhibition of the diphosphopyridine nucleotide system by the p-aromatic diamines made it desirable to test another yeast enzyme system. Similar results were obtained when a carboxylase- cocarboxylase system was employed, except for one difference. The addition of reducing agents such as cysteine, glutathione, and as- corbic acid reduced markedly the toxicity of the p-aromatic diamines for the diphosphopyridine nucleotide system. In the carboxylase- cocarboxylase system, on the other hand, to which the same sub- stances are inhibitory and where there is no complicating catalytic oxidation of the inhibiting agent, the addition of the reducing sub- stances prevents any inhibition at all. In experiments with both the diphosphopyridine nucleotide sys- tem and the cocarboxylase system, where the inhibition by the toxic compounds was large in the presence of low coenzyme concentra- tions, it was found that the addition of large amounts of coenzyme Table 1* Parent molecule X=CH3 Carcino- genic potency (rat liver) Ref. Split product X=CH3 X ■N=N •N< X X 4,5'-Dimethyl-iV-iV-diinetliylamino- azobenzene X H2N N<^ -N=N- Aminoazobenzene -NH, 10 HoN NH2 X X 2,3'-Dimethylaminoazobenzene NH2 + 11 H2N X NH2 N=N' ■N< + 4,6'-Dimetliyl-iV-A^-dimethylamino- azobenzene X H2N ■N< < •N-:N- ++ 12 iV-A^-dimethylaminoazobenzene H2N X •N< + + 4-Methyl-A^-iV-dimethylaminoazo- benzene H2N- X X H CH3OC N- n/ * The table on these facing pages presents the results of tests showing the apparent correlation between the carcinogenic properties of the parent azo com- pounds and the toxic properties of the p-aromatic diamine split product of these molecules. 248 Stability of free radical of split product Oxidation by cytochrome- oxidase system* cmm. O2 per 10 minutes Percentage Inhibition of Enzyme Systems diphospho- pyridine nucleotide t cocarboxy- lasef respirationj oxygen consump- tion** 5 minutes 2 0 9 10 0 4-8 hours 160 38 51 16 0 4-8 hours 164 65 67 39 21 2 days 176 83 75 45 32 7 days 165 92 81 41 41 7 days 165 92 81 41 41 very unstable 0 2 5 20 4 * M/50 concentration used. No inhibition of the cytochrome-oxidase system with M/50 p-phenylenediamine results with any of the split products at concentrations of 5X10-*M. t System: washed yeast; compounds at a concentration of 5X10"'* M. t System: rat liver slices, 5-hour experiment; compounds at a concentration of 1X10-3 M. ** System: rat liver brei, 2-hoiu- experiment; compounds at a concentration of 5X10-«M. 249 250 A SYMPOSIUM ON RESPIRATORY ENZYMES (500-1000 micrograms) prevented the inhibition. But when large amounts of coenzyme were added, the inhibition was prevented only in the system in which it was the limiting factor; that is, the addition of large amounts of diphosphopyridine nucleotide main- tained activity only in the yeast-fermenting system and did not reduce the toxicity of the diamino compounds in the cocarboxylase system; and, conversely, the addition of large amounts of cocar- boxylase maintained activity only in the carboxylase system. These experiments indicate that the inhibition in both systems is a com- petitive one and further suggest that the action of the diamino compounds is on the protein enzyme component rather than on the coenzyme. Rat liver cell (slice) and rat hver suspension (brei) oxidations are also inhibited by the p-aromatic diamines. The toxicity gradient of these compounds to surviving liver tissue is similar to that ob- served in the yeast systems. To detect the protection of these rat liver systems by the addition of reducing agents is not practicable, for it was observed that the presence of these diamino compounds in the crude liver enzyme systems catalyzed the oxidation of the re- ducing agents, cysteine and ascorbic acid. Those diamino compounds that are most toxic act as stronger catalysts of the oxidation of the re- ducing agents than do the less toxic compounds in the presence of a liver suspension. It was important to establish the fact that the p-aromatic diamines are not simply non-specific enzyme poisons. The (Z-amino acid oxidase, tyrosinase, cytochrome oxidase, and acid and alkaline phosphatase enzymes were not inhibited by dimethyl- p-phenylenediamine in equivalent concentrations (5 X 10"^ molar). In these experiments (see Table 1) the gradations of the toxicity of the diamino split products of methyl derivatives of aminoazo- benzene parallel the carcinogenic potency of the parent molecules. But inasmuch as the list of compounds of this series which have been tested for carcinogenic power is small, further animal experiments with other related compounds are needed to determine whether this apparent correlation is a true one. Furthermore, the evidence that the production of p-aromatic diamino split products from the parent azo molecule in the liver of the rat is concerned in the resulting production of hepatic cancer is entirely indirect. Although the mechanism of the inhibition of the protein enz)Tnes in these several systems by the p-aromatic diamine is obscure, it appears that the formation of an oxidation product or products of the reduced compounds is essential to secure this effect. The indirect DISCUSSION ON TUMOR RESPIRATION 251 evidence available suggests, in addition, that the inhibition may be produced by oxidation of, or combination with, sulfhydryl groups on the protein enzyme. For example, in the diphosphopyridine nucleo- tide system iodoacetate and alloxan, which are known to react with sulfhydryl groups, also inhibit fermentation. The alloxan inhibition is, like that caused by the p-aromatic diamines, a competitive one. Vassel (6) has found that dimethyl-p-phenylenediamine will con- dense with cysteine in strongly acid solutions in the presence of Fe+++ and ZnCL to yield a colored (blue) product. In several experiments done in our laboratory we found that in the presence of a liver suspension (brei) an orange product is formed when both dimethyl-p-phenylenediamine and cysteine are added. This product is not formed on the addition of either alone. The work of White (7) has shown that organic sulfur (cystine, methionine) can counteract the growth-inhibiting properties of both the azo and hydrocarbon carcinogens and supports the view that the carcinogen may combine with the sulfhydryl groups of proteins. Further evidence is provided by the studies of L. F. Fieser and his collaborators (8). These concern experiments with the hydro- carbon carcinogens in which an entirely different approach was fol- lowed. The results suggest that the action of the hydrocarbon carcin- ogens may be on an S— S link of a protein and serve to emphasize the need for further information on the mode of action of the p-aromatic diamines in catalytically active systems. REFERENCES 1. Hashimoto, Y., Gann, 29, 306 (1935). 2. Stevenson, E. S., Dobriner, K., and Rhoads, C. P., in press. 3. Kensler, C. J., SuGiuRA, K., and Rhoads, C. P., Science, 91, 623 (1940). 4. Kensler, C. J., and Rhoads, C. P., Journal of Cancer Research, 2, 1 (1942). 5. Michaelis, L., Schubert, M. P., and Granick, S., J. Amer. Chem. Soc, 61, 1981 (1939). 6. Vassel, R., Jour. Biol. Chem., 140, 323 (1941). 7. White, J., J. Natl. Cancer Inst., 1, 337 (1940). 8. Fieser, L. F., Production of Cancer by Folynuclear Hydrocarbons ( Univer- sity of Pennsylvania Press). 9. Nagao, N., Gann, S5, 20 ( 1941 ) . 10. Sasaki, R., and Yosida, T., Virchow's Archiv., 295, 175 ( 1935). 11. Yosida, T., Virchow's Archiv., 283, 29 (1932). 12. KiNosiTA, R., Trans. Soc. Path. Jap., 27, 665 (1937). Discussion on Bacterial Respiration W. H. PETERSON, University of Wisconsin, Chairman CRITERIA FOR EXPERIMENTS WITH ISOTOPES H. G. Wood, Iowa State College: The criteria that are to be used in considering the rehabihty of isotopic work will vary considerably with the type of investigation. For example, in studies with carbon of atomic weight 13 in which one wishes to determine qualitatively whether or not carbon dioxide is fixed in a biological reaction, the criteria are relatively simple. The reaction is simply conducted in an atmosphere containing car- bon dioxide in which the content of C^^ has been increased artifi- cially above that of carbon found in nature. Carbon in nature contains about 1.1 per cent C^^. The carbon dioxide used for a tracer usually contains from 5 to 15 per cent C^^, the concentration depending on the method used to obtain the heavy isotope. To determine whether or not carbon dioxide is incorporated into an organic compound, it is necessary only to free the reaction mixture of carbon dioxide, convert the organic compounds to carbon dioxide, and determine its content of C^^ in a mass spectrometer. If the C^^ content of the reaction mixture is significantly above 1.1 per cent (the natural complement of C^^), carbon dioxide has been incor- porated into organic compounds. This is obviously the case, since there is no other source of carbon than carbon dioxide with a content of C^^ above the normal. The question is, what reliable conclusion can be drawn from such an experiment? Clearly, the only conclusion that can be made is that carbon dioxide is fixed, but no idea is given as to the reaction involved or the compounds concerned. In the early isotopic work such experiments were given undue significance. For example, this type of experiment was conducted by Ruben and Kamen with pigeon liver, Escherichia coli, and a number of other heterotrophic systems and it was found that a small amount of radioactive carbon dioxide was fixed. These experiments have been cited as examples of assimila- tion of carbon dioxide by heterotrophic organisms. There is no general agreement on a definition of assimilation, but it is certain that to 252 DISCUSSION ON BACTERIAL RESPIRATION 253 many the term implies construction of cell material, enzymes, and perhaps organic compounds involving a carbon chain. Clearly this simple demonstration of carbon dioxide fixation is not a reliable cri- terion of assimilation so defined. Obviously the carbon dioxide could have been fixed in the simple one-carbon compound urea or formic acid, a fixation far different from that which the term "assimilation of carbon dioxide" suggests to most people. Assuming that one has completed a simple demonstration of fixa- tion of carbon dioxide in a biological process, what additional criteria must be met then to determine the mechanism of fixation? It is here that the real problems of isotopic work are met, and the problems are by no means simple. Isotopes are a very valuable tool to the investigator, but even isotopes involve many uncertainties. Frankly, the field of isotopic investigation is not fully enough developed to permit a clear understanding of all the criteria that must be met. Therefore, in this short presentation only a few of the possible sources of error will be presented as a starting point for discussion. In any investigation, whether it involves isotopes or not, the gen- eral experimental procedure must be reliable if results are to be valid. In fact, assuming that one is cooperating with a competent physicist, the actual isotopic analysis and separation of isotopes will be the minor problem; the general experimental procedure will offer the real problems. A specific experiment may be cited as an example of a case in which the error resulted from the general procedure and not from any isotopic considerations. In the propionic acid fermen- tation, propionic acid is formed in which carbon dioxide-carbon is incorporated. To obtain information on the possible mechanism of the reactions concerned in the fixation of carbon dioxide in propionic acid, it was necessaiy to determine the location of the fixed carbon in the propionic acid molecule. Thus the problem was to select a reliable chemical reaction for the degradation of the molecule. In doing this Carson and his co-workers degraded radioactive propionic acid obtained from the propionic acid fermentation by alkaline permanganate oxidation and obtained oxalic acid and carbon diox- ide: CH3CH2COOH -> COOHCOOH + CO2 From 70 to 75 per cent of the radioactive carbon was found in the oxalate fraction and 25 per cent in the carbonate. Since the workers believed that the carbonate arose from the carboxyl carbon and the oxalate from the alpha and beta carbons of propionic acid, it ap- 254 A SYMPOSIUM ON RESPIRATORY ENZYMES peared that the carbon dioxide was equally distributed among the three carbons in the chain. In other words, it seeemed that the pro- pionic acid might be synthesized entirely from carbon dioxide, which would be a most remarkable accomplishment for a typically hetero- trophic propionic acid organism. It was soon proved, however, that the reaction was not a reliable method of decarboxylating propionic acid. This was done by checking the reaction with synthetic pro- pionic acid, which contained C^^ in the carboxyl group. Further, by use of a rehable degradation reaction, all the fixed carbon in the biologically formed acid was shown to be in the carboxyl group. Carson and his co-workers obtained the same results when they re- checked their previous procedures. With this example we may pass on to consideration of the case, in which it will be assumed the criteria described above have been met, i.e., ( 1) a reliable determination of C^^ has shown there is fixa- tion of carbon dioxide; (2) the experimental procedure has been a good one for demonstrating the desired results; and (3) the com- pound or compounds containing the fixed carbon have been isolated and degraded by reliable chemical reactions; thus the location of the fixed carbon in the respective compounds is known. It is true that only by the use of isotopic carbon could such in- formation be obtained, but even then there may be much uncer- tainty respecting the mechanism of fixation of carbon dioxide. For example, in bacterial glucose fermentations by Staphylococcus, Streptococcus, and Proteus, it has been shown by Slade et al. that the carbon dioxide is fixed in the carboxyl groups of the lactate and suc- cinate; with Aerohacter and Clostridium welchii, there is fixation in the carboxyl group of the acetic acid as well. With pigeon liver on pyruvate, there is fixation in the carboxyl groups of malate, fumarate, succinate, lactate, and alpha-ketoglutarate. Many of these fixations are believed to occur initially by three- and one-carbon addition through the following reaction: CO2 -1- CH3 • CO • COOH = COOH • CH2 • CO • COOH Particularly it is believed that this reaction is instrumental in the formation of four-carbon dicarboxylic acids. The mechanism of the fixation in lactate and acetate is still largely unknown. But what criteria can be used in determining the reliability of the suggested fixation by three- and one-carbon addition? At present there is no completely reliable criteria; but this represents our interpretation of the present known experimental facts. Thus far the synthesis of oxalacetate from pyruvic acid and carbon DISCUSSION ON BACTERIAL RESPIRATION 255 dioxide has not been accomplished, but Krampitz in our laboratories has done the next best thing. With a preparation of Micrococcus, which decarboxylates oxalacetate to pyruvate, he has shown the catalysis of the exchange of C^'^-carbon dioxide with the carboxyl group of oxalacetate, i.e., incubation of oxalacetate with C^^-carbon dioxide and the enzyme gave oxalacetate containing C^^ in the carboxyl group. During this exchange the oxalacetate is apparently broken down to a three-carbon compound and again resynthesized, permitting the entrance of C^ ^-carbon dioxide. It is suggested that the three-carbon compound involved in the fixation reaction is not pyruvic acid, as such, but a derivative of this compound that is foraied during the decarboxylation of oxalacetate. This is the first direct evidence, i.e., evidence with oxalacetate as such, that has been obtained as proof of the reaction. Finally, I should like to comment on one further consideration. Assuming that we know the mechanism of the reactions concerned in fixation of carbon dioxide, we are still faced with one very impor- tant question, namely, is the fixation reaction an essential step in the dissimilation or is it the result of an exchange reaction, for example? Time does not permit full development of this subject. Furthermore, since the problem has not yet been given much consideration in most investigations, satisfactory criteria have not been devised for de- termining whether a reaction is essential or not. An example may serve to illustrate the point. In the Krebs cycle, as Dr. Evans pointed out in his lecture, the pyruvate is believed to be oxidized after union with oxalacetate through a cyclic conversion. The oxalacetate, it is believed, arises from the fixation reaction through union of pyru- vate and carbon dioxide. The fixation reaction would then be an essential reaction in the oxidation of pyruvate, since it would supply the necessary oxalacetate. It is possible, however, that the oxalacetate cannot be formed by this reaction but is formed by some other reaction of pyruvate. The oxalacetate thus formed might react with carbon dioxide by an exchange reaction as studied by Krampitz. In this case one would arrive at oxalacetate containing heavy carbon just as he would if the oxalacetate was formed by union of pyruvate and carbon dioxide. Isotopic analysis would not serve to differentiate the mechanisms. It is evident that a false importance may be ascribed to carbon dioxide fixation reactions, for fixed carbon dioxide may result from a non-essential exchange reaction. Frankly, I do not believe this to be the case, but I can cite only indirect criteria to support my opinion, such as the fact that carbon dioxide is necessary for growth of bacteria and in the reduction of methylene blue by 256 A SYMPOSIUM ON RESPIRATORY ENZYMES dehydrogenase. If carbon dioxide was concerned only in a non-essen- tial exchange, it would not be required. R. H. BuRRis, Columbia University: The criteria that must be met in the use of isotopes as tracers have been well covered by Dr. Wood in the preceding discussion, so we may consider certain other problems connected with isotopic tracers. The question is frequently heard, "Which tracers are more suitable for biological problems, radioactive or stable isotopes?" There is no blanket answer, for the suitability of a given tracer depends not only upon the properties and availability of the given element, but also upon the type of experiment in which the tracer is to be employed. In quahtative studies the radioactive isotopes of reasonably long half -life possess many advantages; they can usually be detected in greater dilution and with greater facility than can the stable isotopes. In addition, they may often be traced directly in vivo without la- borious fractionation of the organism. For example, shortly after the ingestion of a radioactive sodium salt radioactivity may be de- tected in any part of the body if it is brought into proximity to a Geiger counter chamber. Radioactive substances with sufficiently in- tense radiations will reveal their distribution by forming radiographs on photographic plates. In this manner the pattern of radioactive substances in bones has been demonstrated following the feeding of isotopes. Radioactive isotopes have been prepared in much greater variety than have the stable isotopes. More than two hundred radioactive isotopes have been produced. Many, of course, are of no biological significance; others have not been prepared in suitably high con- centrations for tracer studies; and unfortunately some of the ele- ments of greatest biological interest form isotopes of very short half- life. The following radioactive elements have been used in biological investigations: Isotope Half-life Isotope Half-life ,W 85 days i,CP« 37 minutes eC^i 20.5 minutes j^K" 12.4 hours eC^^ 100-1000 years 2oCa*^ 2.5 hours tN^^ 10.5 minutes scFe^^ 47 days 9F18 112 minutes 3,Br8o 44 j^q^j-s nNa^* 14.8 hours g.Br^^ 34 j^q^^s 15P'' 14.3 days ggP^s 25 minutes leS^^ 88 days DISCUSSION ON BACTERIAL RESPIRATION 257 The chief feature in favor of the stable isotopes is, of course, their stabihty. The extreme speed necessary in handhng radioactive iso- topes of short half-hfe, such as C" and W^, has led to some un- fortunate errors, as has been pointed out by Dr. Wood. With the stable isotopes time is not a factor, and compounds can be separated and analyzed at leisure. When isotopes become available for general distribution, the stable forms may be stocked vi^ithout decomposition. The objection that radiations from radioactive tracers may injure tissues is scarcely a serious one, since in most cases there is a wide margin between the level of radioactivity necessary for measurement and the level that will be injurious. With the stable tracers the ques- tion of radiation does not arise. Only in the case of deuterium, with its marked differences in properties from its analogue, have injury eflFects been noted when high concentrations of a stable isotope were present. Although the stable isotopes have not been prepared in as great variety as the radioactive isotopes, the elements of chief interest to the biologist have been concentrated. H- (deuterium), C^^, N^^, O^^, and S^* have been concentrated by Urey. Deuterium is an item of commerce, and N^^ and C" may be on the market soon. For quantitative experimentation the stable isotopes are far su- perior to the radioactive isotopes. Rittenberg has found that in the analysis of N^^ with the mass spectrometer he can expect a precision of ± .003 atom per cent N^^. To obtain comparable precision in meas- uring a radioactive substance with a Geiger counter it would be necessary to count in the order of a million impulses. If a Geiger counter capable of handling a thousand impulses a minute were available, and the sample under examination were suflBciently con- centrated to give this output of charged particles, sixteen and two- thirds hours would be required to register a million counts. It is obvious that radioactive nitrogen with a half -life of 10.5 minutes or radioactive carbon with a half-life of 20.5 minutes could not be measured with precision. Nor is short half -life the only factor that interferes with the quantitative measurement of radioactive isotopes; many of the radioactive isotopes with a long half -life (C", S^^, Fe^^, etc.) emit such soft radiations that measurement is extremely diffi- cult, and background counts constitute a considerable percentage of the total counts. Radioactive phosphorus is an ideal tracer with a long half -life and an intense radiation, and in this case precise quan- titative measurements can be obtained. But with the elements of chief interest to the biologist— carbon, hydrogen, and nitrogen— the 258 A SYMPOSIUM ON RESPIRATORY ENZYMES stable isotopes are much more favored for quantitative studies than the radioactive elements. The application of isotopes to problems involving respiratory en- zymes has not been extensive. The metabolism of lactic acid and the assimilation of carbon dioxide have been traced with stable and radioactive carbon isotopes. The use of radioactive P^^ in the study of the role of phosphorus in respiration has attracted a number of investigators. In the near future, undoubtedly, isotopes will be em- ployed in many other studies of this nature. MECHANISMS FOR THE COMPLETE OXIDATION OF CARBOHYDRATES BY AEROBIC BACTERIA C. H. Werkman, Iowa State College: This topic is a broad one that has not been exhaustively inves- tigated because it offers serious technical diflBculties. Bacteria seem to possess a distinct and quite troublesome cell wall whose behavior as regards changes in permeability is far from clear. Realiz- ing the danger of drawing conclusions from the use of unnatural systems, we have attempted, by correlating the results obtained with juices and whole cells, to circumvent the pitfall. But to do so would apparently require an uncanny ability, since neither the work with juices nor the cell suspension could be accepted as portraying processes that occur in the living, reproducing organism. In the face of this difficulty of preparing juices or cell suspensions that behave naturally, the task of elucidating the natural processes in bacterial respiration is laden with danger. Our present purpose is to encour- age a free discussion of the possibilities in bacterial respiration and to outline a working hypothesis. There is abundant evidence that phosphorus plays an important, if not essential, role in bacterial respiration. Early workers, including Virtanen, have shown the ability of bacteria to form phosphorylated esters. Wiggert in our laboratory showed an uptake of phosphorus by living bacteria. That the principles of the Embden-Meyerhof scheme of glucolysis operate in bacterial metabolism was first given substantial support in 1936, when the characteristic intermediate of that scheme, phosphoglyceric acid, was isolated from representa- tive types of bacteria, both aerobic as well as anaerobic (Werkman et al., Stone and Werkman). Utter in our laboratory has shown the occurrence of the aldolase reaction by means of bacterial juices. In this reaction hexosediphosphate is converted into phosphoglyceralde- hyde and dihydroxyacetone phosphate. Furthermore, he has estab- DISCUSSION ON BACTERIAL RESPIRATION 259 lished the dissimilation of phosphoglyceric to pyruvic acid through phosphopyruvic. Pyruvic acid has been shown to be a general intermediate in the dissimilation of glucose by bacteria. In the case of anoxybiontic metabolism the pyruvic acid is converted into a variety of products; in the case of oxybiontic (aerobic) metabolism, however, it appears that pyruvic acid initiates the changes of terminal respiration. These are the changes of the whole aerobic dissimilation, and provide for the oxidation of pyruvic acid to carbon dioxide and water. With respect to bacteria there is evidence that the final processes of respiration involve the cytochrome-cytochrome oxidase system. Both cytochrome and its oxidase have been shown to be present in many aerobic bacteria, and in no case known has cytochrome failed in an aerobic species. There is no doubt that bacteria possess a cytochrome mechanism, although many questions respecting the details of its operation remain to be answered. We have found, for instance, an acetone-resistant cytochrome oxidase not reported in animal tissues, and bacterial cytochrome may have a lower potential than that in animal tissue. Between the terminal stage of glycolysis and the initial stage of the cytochrome mechanism there is a portion in the spectrum of respiration that has received only passing investigation so far as bacteria are concerned. This may be referred to as the four-carbon dicarboxylic acid portion. The Szent-Gyorgyi and Krebs schools have pioneered in this work as it relates to animal tissue, and our lead is taken from their work. It is this four-carbon acid portion which we wish to discuss first. Mr. Krampitz in our laboratory has been wrestling with these secrets of nature. A few experiments and results will be cited here for pur- poses of discussion. It is hardly necessary to mention that our remarks are preliminary. Shortly after Szent-Gyorgyi formulated his theory of the role of the four-carbon acid in cellular physiology, an attempt was made to apply it to bacteria. Before any catalytic eflFect of the four- carbon acids can be shown, the preparation of enzymes must be made deficient in four-carbon acids by washing bacteria free from these acids. Using Micrococcus lysodeikticus to make our prepara- tions, we did obtain stimulation with fumaric acid varying from 35 to 120 per cent as measured by oxygen uptake. Such stimulation was not, however, obtained consistently, and therefore we attempted to increase the cell permeabihty by acetone treatment, hoping to re- 260 A SYMPOSIUM ON RESIRPATORY ENZYMES move the four-carbon acids more effectively. But this preparation did not function in all respects like normal cells, failing to oxidize glucose and succinate among others, although it did oxidize the four- carbon acids (fumaric and malic) and lactate and pyruvate to acetic acid. Since succinate was not oxidized, we believed that the cyto- chrome system might have been injured, inasmuch as the hydrogen, according to Szent-Gyorgyi, is transmitted through succinate to the cytochrome system and thence to oxygen. When methylene blue or cresyl blue was added to the system as carriers of hydrogen to re- place the cytochrome system, no oxidation of succinate occurred. Injury to Havoprotein action appeared ruled out because of the rapid turnover of the substrates that were oxidized, i.e., fumarate, malate, etc. Since acetone is known to destroy cytochrome oxidase, it was difficult to understand the rapid attack on fumarate, malate, lactate, and pyruvate. It was shown spectroscopically that the preparation contained a cytochrome oxidase resistant to the acetone treatment. These results were enough to indicate that investigations of bacterial respiration were to prove interesting as well as a bit troublesome. One of our principal objectives has been to integrate the hetero- trophic assimilation of carbon dioxide and respiration. We have therefore run a number of experiments to gain a better insight into the mechanism of bacterial respiration. The evidence so far ac- cumulated seems to indicate that the Szent-Gyorgyi cycle does not function in bacterial respiration. Glucose, for instance, is not at- tacked by Micrococcus lysodeikticus anaerobically by cell suspen- sions or the acetone preparation. If the Szent-Gyorgyi system were operating, malic acid present would be oxidized anaerobically, since fumarase would provide fumaric acid as a hydrogen acceptor for the system. This does not occur. Thus far the evidence favors the occurrence in principle of the Krebs citric acid cycle in bacteria. As the Krebs cycle is presented, every alpha-keto acid with the single exception of pyruvic acid is oxidatively decarboxylated. This point should be further investigated to determine whether pyruvic acid is not also oxidatively decar- boxylated, not necessarily to acetic acid, but to some two-carbon compound which is able to condense with oxalacetic acid to initiate the citric acid cycle. Acetic acid is commonly found as an end-prod- uct when juices or perhaps injured cells are employed. Maintenance of an adequate supply of oxalacetic acid is a requirement of the Krebs scheme, and this is assured by regeneration in the cycle and by the utilization of carbon dioxide through the Wood and Werk- DISCUSSION ON BACTERIAL RESPIRATION 261 man reaction. We have found some evidence for the formation of citric acid in the presence of acetic acid by Micrococcus lysodeik- ticus. On the other hand, many bacteria do not appear to metaboHze citric acid; this is difficult to reconcile with the Krebs cycle as pro- posed, although it is probable that it is not citric acid as such which is the intermediate. At present we are at work on C" acetic acid as a tracer. Certain experimental evidence with the four-carbon acids has at times sug- gested that we are dealing with phosphorylated compounds, prob- ably of a very labile nature. Thus an acetone preparation is made magnesium- and cocarboxylase-deficient by alkaline phosphate washing of the cells as determined by testing on pyruvic acid. The deficient preparation does not decarboxylate oxalacetic acid; how- ever, the addition of magnesium ions completely restores the ac- tivity. The reaction yields carbon dioxide and pyruvic acid. When malic acid replaces oxalacetic acid, the deficient preparation under the same conditions does not oxidize malic acid to carbon dioxide and pyruvic acid, but only to oxalacetate, which accumulates and does not inhibit the bacterial malic dehydrogenase as it does tissue dehydrogenase. The complete preparation (deficient plus magnesium ions or the unwashed acetone preparation) oxidizes malate to carbon dioxide and pyruvic acid with traces of oxalacetic acid. Thus with laboratory oxalacetic acid the reaction goes to carbon dioxide and pyruvate, whereas with "physiological" oxalacetic acid (from malate) the reaction appears to maintain an equilibrium. Is the "physiological" oxalacetic acid different from that prepared in the laboratory, possibly a phosphorylated compound? We investi- gated the problem and at one time thought that it was. Definite stimulation by phosphate has been demonstrated for fumaric or malic acid. As yet we have not shown a phosphate uptake, or isolated an organic phosphate; however, we may be dealing with a labile carbonyl phosphate in the sense of Lipmann. It is suggested that the carbonyl group is bound in the physiological oxalacetate formed from malic acid, since traces of oxalacetate are known to inhibit malic oxidation in tissue. If malate is oxidized in an atmosphere of C^^Oa and O2, the oxalacetic acid contains C^^ in the carboxyl group adjacent to the methylene group. No chemical exchange takes place. The enzymatic decarboxylation of oxalacetic acid in the presence of heavy carbon dioxide also yields heavy carbon oxalacetic acid. This is a form of carbon dioxide utilization. 262 A SYMPOSIUM ON RESPIRATORY ENZYMES E. S. GuzMAPf Barron, University of Chicago: Until a few years ago seme investigators in the field of oxidation- reductions tended to devote their efforts exclusively to animal, plant, or bacterial oxidations. Many papers published recently showing the variety of oxidation mechanisms, even in the oxidations involving simply an electron transfer, have demonstrated the necessity of in- tegrating the facts obtained with these different kinds of living be- ings. For such comparative studies, work with bacteria has been fruitful not only because it is possible to obtain suspensions or ex- tracts with which quantitative studies can be perfoiTned but also because in a single species, say hemolytic streptococci, a variety of oxidation mechanisms may be found in different strains. If we take the component of oxidation enzyme systems closest to molecular oxygen, the iron porphyrins, we may divide bacteria into two groups: cytochrome-containing bacteria (including most of the so-called aerobic bacteria) and cytochrome-lacking bacteria (the so- called anaerobic bacteria). Species of the two groups may produce identical oxidations. They may oxidize, for example, lactate or glycerol. The rxidation of lactate and glycerol by cytochrome-con- taining bacteria (Staphylococcus) is completely inhibited by cyanide, whereas the same oxidations by cytochrome-lacking bacteria are cyanide-insensitive. Obviously iron porphyrins take part in the oxidation of lactate and glycerol by cytochrome-containing bacteria, whereas in cytochrome-lacking bacteria the oxidation proceeds through different channels (flavin nucleotides). In the field of phosphorylative oxidations the laboratory of Werk- man has demonstrated that there exist in bacteria the different phosphorylations observed in the breakdown of carbohydrate by muscle or yeast extracts. In our laboratory it has been found that the oxidation of glycerol by hemolytic streptococci does not take place in the absence of phosphates. This does not mean that phos- phorylation is essential for glucose oxidation; it is known that the breakdown of carbohydrate by molds proceeds without phosphoryla- tion. It is assumed that in animal tissues carbohydrate metabolism starts with the fermentation process ending in lactate, whereas in yeast it ends in the fonnation of alcohol. In bacteria the fermentation process may end in the formation of either lactate or of alcohol or in the formation of both end products, as Friedemann has shown. Fermen- tation may be absent altogether, as in glucose-non-fermenting bac- DISCUSSION ON BACTERIAL RESPIRATION 263 teria. These bacteria either oxidize glucose directly without previous phosphorylation {Pseudomonas aeruginosa) or oxidize directly phos- phorylated hexose (hexose monophosphate and diphosphate). The same variety of mechanisms is found in the oxidation of pyru- vate (CH3COCOOH + 'AO^ = CH3COOH + CO2): it requires an iron porphyrin catalyst in gonococci; it proceeds without iron por- phyrin in Bacterium Delbriickii, as Lipmann has shown. d-Amino acid oxidase, isolated by Warburg and Christian, is an alloxazin dinucleotide protein, the oxidation of alanine to pyruvate being cyanide-insensitive; this oxidation when performed by cytochrome- containing bacteria requires iron porphyrin as a component of the enzyme system because the oxidation is completely inhibited by cyanide. These examples are presented as proof of the existence of multiple mechanisms of oxidation. A comprehensive study of biologic oxida- tion-reduction demands, therefore, a continuous and simultaneous attention to the oxidation mechanisms throughout living cells. P. W. Wilson, University of Wisconsin: A discussion of oxidations by aerobic bacteria should certainly include reference to the fact that one of the most actively respiring tissues known belongs to this group of organisms. I refer, of course, to the extremely high rate of respiration possessed by certain cultures of Azotobacter, the free-living nitrogen-fixing bacteria. One of our distinguished guests and participants. Dr. Otto Meyerhof, first called attention to this several years ago when he reported Q02 values of 500 to 8600 for Azotobacter chroococcum on glucose at 28° C. (1, 2). The extremely high values were obtained with very young cultures so diluted that the total dry weight involved was less than 10 micro- grams. Its estimation may have been subject to some error, but it is probable that young cultures of this organism have a Q02 value of at least 5000. Since Meyerhof and his collaborators made these experiments, important advances have been made toward developing cultural conditions that are optimum for growth and nitrogen fixation by Azotobacter. Bates of fixation are consistently obtained in experi- ments today which are several times greater than those reported several years ago. For example, Azotobacter vinelandii can fix as much as 20-30 mg. of nitrogen in 24r-36 hours instead of 4-5 mg. in one or two weeks which was the characteristic result of most of the 264 A SYMPOSIUM ON RESPIRATORY ENZYMES earlier studies (3). It is of interest to examine such cultures for re- spiratory activity. My associate. Dr. R. H. Burris, has determined the Q02 values of a 12-hour culture of Azotobacter agilis on a number of substrates; his findings were as follows: endogenous, 28; glucose, 29; lactate, 129; arabinose, 29; acetate, 1109; ethyl alcohol, 1240. When a 48-hour culture was used, these values were greatly reduced. More recently Mr. Joe Wilson in our laboratory has made similar observations with Azotobacter vinelandii. In most of these studies a 24-hour culture was diluted, and the oxygen uptake measured im- mediately for a period of 60 minutes. The substrate was sucrose, the temperature 30° C, the pH 7.0. Air was used as the gas phase so that opportunity for growth existed, but no detectable fixation of nitrogen occurred during this short experimental period. To avoid the error attached to estimation of dry weight, he calculated the rate of respiration on the basis of cell nitrogen (4). Since these cells of Azotobacter contain about 10 per cent nitrogen, such Q02 (N) values are on the average about 10 times as great as the Qoo based on dry weight. Under these conditions the values of the Q02 (N) ranged from 25,000 to 30,000. These data emphasize the extremely high rate of respiration of difiFerent species of Azotobacter and suggest that this organism may well provide an excellent source for the preparation and isolation of different enzyme systems concerned with tlie transfer of hydrogen from substrate to molecular oxygen. With this in mind we are in- vestigating methods for growing Azotobacter on a scale considerably greater than any previously attempted. In a pilot plant designed for yeast production we have succeeded in producing several pounds of moist azotobacter cells during a growth period of 24 to 30 hours. REFERENCES 1. Meyerhof, O., and Burk, D., Z. physik. Chem., 139A, 117 (1928). 2. Meyerhof, O., and Schulz, W., Biochem. Z., 250, 35 (1930). 3. Wilson, J. B., and Wilson, P. W., Jour. Bact., 42, 141 (1941). 4. Burris, R. H., and Wilson, P. W., Proc. Soc. Exp. Biol. Med., 45, 721 ( 1940). REACTIONS IN CELL-FREE ENZYME SYSTEMS COMPARED WITH THOSE IN THE INTACT CELL F. F. NoRD, Fordham University: In interpreting the results of investigations of bacterial metabolism a few principles should be mentioned which, in addition to the very recent use of tracers, appear to have been applied in approaching DISCUSSION ON BACTERIAL RESPIRATION 265 the problems with which we are confronted in extracts and in Hving cells: 1. In extracts a disturbance of the ratio of the various components and an effect upon the total enzyme system occurs automatically, which may cause an accumulation and even a stabilization of tran- sient products. 2. By selective poisoning of parts of the enzyme system numerous facts have been established. 3. In hving cell processes, because of the introduction of reagents not akin to the whole system, a supposed or possible intermediary product is removed and thereby excluded from the reaction se- quence. For example: 1. Amino acid oxidase is capable of deaminating I- or d-amino acids in tissue slices. In the case of injured or denatured tissues, however, natural amino acids are no longer deaminated. 2. The extent to which the carrier enzymes are dispersed may change under the influence of various factors. Moreover, in the cells disperse particles of the various protoplasmic substances actually possess widely differing pH values. 3. Even Dr. and Mrs. Cori declare, in accordance with the afore- mentioned fact, that the conditions for glycogen synthesis are much more favorable in the intact cell than in tissue extracts, where, they state, they have obtained starch.* 4. Experiments show that in the case of Corynebacterium diph- theriae, strains gravis and mitis, conditions are comparable. 5. There is no stoichiometrical relationship between the carbon dioxide evolved and the actual decrease in inorganic phosphorus (in the living cell). 6. When carbon dioxide evolution was compared with energy liberated as heat by living yeast cells and by Lebedew extract, it was noted that the heat of reaction in the course of fermentation changed continually, indicating that fermentation with living cells does not proceed according to a fixed scheme. The thermochemical course of fennentation with juices shows, in contrast, that at least two different conversions occur: (1) fermenta- tion of free sugar in the presence of free phosphate (inhibited by phloridzin) and (2) the subsequent fermentation of the residual substrate in the absence of free phosphate (not inhibited by phlorid- zin). * Compare G. J. Goepfert, Brewers Digest, 16, No. 6 (1941). 266 A SYMPOSIUM ON RESPIRATORY ENZYMES P. W. Wilson, University of Wisconsin: It is only recently that most investigators of bacterial metabolism have had the opportunity to choose between cell extracts and intact cells. Because of their small size, bacteria as a group have resisted attempts to destroy their cellular integrity. In the past, if studies on bacterial enzyme systems were to be made, the investigator was in the position of Kipling's thief who took the hot stove because there was nothing else that season— he had to use intact cells. Although this limitation has undoubtedly complicated the investigations and the interpretation of the results, it has not been insurmountable. By use of the so-called "resting cell" technique, great strides have been made toward an understanding of the biochemistry of bacteria. With respect to individual enzyme systems the greatest success has at- tended studies in which isolation of the reaction is made possible through choice of substrate rather than enzyme system, for example, hydrogenase and hydrogenlyase. When less specific substrates are employed, great care must be exercised that the interpretation is not oversimplified by applying information gained in what appears to be an analogous study made with the more purified enzyme prep- arations. With certain bacterial enzyme systems the investigator is denied even the use of non-proliferating cells. For example, studies on nitrogen fixation by Azotobacter must usually be made with growing cultures : on the symbiotic system with a very complex association of bacteria and host plant. Despite these technical handicaps, the de- velopment of certain methods (1, 2) during the last decade has enabled investigators to secure what Burk has recently described as "the most intimate information that we possess on the mechanism of fixation, and in particular on the nature of the first crucial step in- volved" (3). In recent years special techniques have been developed which allow cell-free extracts containing a variety of enzymes to be pre- pared from bacterial species. While it is gratifying that this first step toward isolation of individual enzymes has been made, the imme- diate practical value of the achievement has not been great. In most cases the net result of the separation has been the verification of a portion of the knowledge previously obtained with the resting cells. Recently we have prepared a cell-free Azotobacter "juice" which contains among other enzymes very powerful preparations of hydro- genase and oxalacetic decarboxylase. Our satisfaction over this ac- DISCUSSION ON BACTERIAL RESPIRATION 267 complishment is somewhat lessened by our recognition that we can do Httle with the extract that we have not already done with intact resting and acetone-treated cells. Nevertheless, we continue studies with it in the hope that a clue to the mechanism of nitrogen fixation may be furnished by the extract that has been successfully hidden in the intact organism. Meanwhile, however, the physical-chemical studies on the intact growing bacteria will not be neglected, since they have already proved their value. REFERENCES 1. BuRK, D., Ergebnisse d. Enzymforschung, 3, 23 (1934). 2. Wilson, P. W., The Biochemistry of Symbiotic Nitrogen Fixation (University of Wisconsin Press, 1940 ) . 3. BuRK, D., and Burris, R. H., Ann. Rev. Biochem., 10, 587 (1941). Discussion on Animal Tissue Respiration C. A. ELVEHJEM, University of Wisconsin, Chairman FACTORS AFFECTING THE PREPARATION OF TISSUE FOR METABOLIC STUDIES EPHRAIM SHORR Cornell Medical College, New York My comments will be restricted to certain difficulties encountered in the preparation of tissue for in vitro studies of metabolism. They touch on the various methods of preparing tissue, such as the slice and mince method and on the influence of certain chemical changes that are inevitable during the handling of the tissue prior to the experimental run. The Slice Method— Most workers use the limiting formula of War- burg without testing the permissible thickness of the specific tissue with which they are working. The general tendency is to get a slice as thin as possible. Histological studies, as well as comparative studies of the rate of respiration, show that this is not altogether wise. The superficial layers can be shown to undergo degeneration to variable depth. The thinner the slice the larger is the proportion of damaged tissue. This is particularly important for tissues such as cardiac muscle, where a whole large cell unit at the surface must inevitably undergo degeneration. Thicker slices can be shown to have a higher rate of respiration than very thin ones. The maximum thickness which is permissible is therefore better. Not infrequently thicknesses which exceed the formula behave very well. This points to the possibility that there may be mechanisms for maintaining oxygen pressure other than the gradient set up by the tension in the solution— perhaps the iron-carrying compounds of the tissue, which serve as a storehouse. The slice method is of course best adapted to parenchymatous organs, least well to muscle. Unfortunately cardiac muscle does not lend itself to dissection as does skeletal. However, the individual muscle cells are much shorter than skeletal muscle cells, hence the degeneration occurring at the surface is not extensive enough to do much harm. As regards the brain, there seems to be little evidence that the more convenient method of chopping the tissue with a razor is less good than slicing. 268 X DISCUSSION ON ANIMAL TISSUE RESPIRATION 269 The Muscle Strip Technique.— In most experiments reported in the literature where skeletal muscle has been employed, it has either been minced or chopped with scissors if larger animals have been used, or the diaphragm of a small animal, such as the rat, has been employed intact. The chopping or mincing is extremely destructive, sets up abnormal chemical processes in the presence of oxygen, such as aerobic glycolysis, and permits brief survival. The rat diaphragm, while excellent for many purposes, is small, permits of few con- comitant chemical measurements, and has only a limited usefulness because of the fact that many metabolic conditions, such as diabetes, cannot be brought about in this animal. The dog is a much better experimental animal in these respects. The neck muscles of the dog are ideal for obtaining, by careful dissection, long muscle strips for in vitro studies. Individual fibers can be teased out intact. These maintain their histological integrity for long periods. They respond to electrical stimulation for hours unchanged, and give reproducible results as far as work and heat production. Furthermore, enough material can be obtained to allow for extensive chemical balances. It is, I am sure, the method to be used when this type of tissue is employed with in vitro studies. Minced Tissue.— Here the choice lies between using a Latapie mincer or a homogenizing apparatus described by Potter. Histo- logical examination of the tissue obtained by these two methods shows that with the homogenizer the tissue is completely and uni- formly disintegrated, whereas with the Latapie the destruction is not complete and the tissue is a mixture of disintegrated and intact cells. The amount of destruction differs with the organ, and these differences are paralleled by differences in the rate of respiration. For example, skeletal muscle because of its long fiber is completely destroyed by both methods and the rate of respiration is the same. With the parenchymatous organs, such as the liver and kidney, a definite difference in respiration results from the two methods, the respiration of the homogenized tissue being lower than that of the Latapie. The same is true of cardiac muscle. The higher respiration may be due to the presence of the intact cells in the tissue put through the Latapie. Obviously tissues minced by these two methods are not comparable with respect to respiration, and other differences may exist. This discrepancy should be borne in mind whenever the two methods yield different results. It would certainly seem desir- able to decide on one or the other. My preference would be homog- 270 A SYMPOSIUM ON RESPIRATORY ENZYMES enization because of its uniformity. With this latter method, over- heating of the solution during the stirring should be rigorously avoided. With minced tissue from the Latapie, two methods of obtaining aliquots are commonly employed. The minced tissue may be weighed out as such or stirred up in a cool solution and pipetted out. My preference is for the latter method, which is a simple one if a wide-mouth pipette is used. It is also much more rapid and yields somewhat more uniform respiration in duplicates and triplicates. With the other method it seems difficult to keep all the tissue under the same conditions of cooling, a desideratum when things are hap- pening as rapidly as they do in such a tissue. Furthermore, I have no confidence in single experiments and so would urge that triplicate de- terminations be carried out. Complicating Factors Arising between Removal of the Tissue and the Experimental RMn.— During this inevitable unphysiological pe- riod in which the tissue is anaerobic, many breakdown processes occur. These lead to the accumulation of a number of metabolites which can influence the results, especially during the early part of the experiment. Whatever the speed of preparation, these changes cannot be avoided. Those I have had to deal with have been the accumulation of lactic acid, and striking changes in the hexosemono- phosphates, adenosinetriphosphate, and phosphocreatine. The ac- cumulation of lactic acid influences markedly the initial rate of res- piration, and may suppress the effect of added lactate. In short experiments it may lead to erroneous conclusions respecting the rates of respiration of individual tissues. The literature contains not a few instances of such misconceptions. It can be dealt with either by reduc- ing the lactate content prior to the experiment, by aerating the tissue in a Ringer solution long enough to bring the content to the normal value, or by allowing the excess lactic acid to be dealt with in the micro respiration vessels for an hour or so before the experimental run. The former appears to be much less time-consuming and just as effective. Oxygen is bubbled vigorously through the solution con- taining the tissue. This is kept at room temperature. As regards changes in the organic phosphate compounds during this interval of preparation, it is very important to restore the normal relationships existing prior to the experiment; otherwise the recovery process may be taking place during part of the experimental period, confusing the results of concomitant chemical balance experiments. This is particularly important for muscle tissue, and it is necessary to de- termine, for each tissue, how long an equilibration is required to DISCUSSION ON ANIMAL TISSUE RESPIRATION 271 permit complete recovery. Undoubtedly other accumulations and disturbances occur which we have not detected. This is worthy of investigation for any specific system under study. The Escape Phenomenon.— Where in vitro experiments are pro- longed, another type of phenomenon must be borne in mind. We have shown that diabetic tissue on prolonged survival in vitro grad- ually regains its ability to oxidize carbohydrate. A complete restora- tion takes place in four hours at 41° C, in ten hours at 37.5° C. With cardiac muscle a definite elevation in the respiratory quotient of normal tissue is found as early as the second hour. This change has been attributed to the release of the tissue from certain influences carried over from the intact animal. The change can be checked by the use of other than inorganic phosphate buffers. Among these is beta-glycerophosphate. In prolonged experiments with any tissue this phenomenon should be borne in mind, since, it may be asso- ciated, as in this instance, with an entirely different type of me- tabolism toward the end of the experiment than at the start. Miscellaneous.— In obtaining respiratory quotients it has been our experience that for vessels of any given size more reliable results are obtained when the respiratory exchange is large. When the total oxygen consumption and carbon dioxide production is small, the results are likely to be unreliable, and generally the respiratoiy quotient is erroneously high. In vessels of 20 to 24 cc. capacity, if the oxygen consumption in the period of observation is less than 150 cmm., we are likely to obtain incorrect respiratory quotients. It is recommended that for each size and type of vessel used, studies be made to determine the amount of oxygen consumption that yields a reliable respiratory quotient. COMPARISON OF SLICES AND HOMOGENIZED SUSPEN- SIONS OF BRAIN TISSUE K. A. C. ELLIOTT Institute of the Pennsylvania Hospital, Philadelphia Warburg introduced the technique of using slices of tissue for metabolic studies, and the method has been used by many other workers. Slices of many tissues can be prepared without disrupting the majority of the cells; gases and substrates can diffuse in and out of thin slices rapidly enough not to limit the rates of metabolic processes. With various mashed and ground preparations, respiration has been found to occur less rapidly, and it is commonly believed that a closer approach to physiological conditions is obtained with 272 A SYMPOSIUM ON RESPIRATORY ENZYMES slices than with tissue breis. While slices have been used in attempts to determine what a tissue actually does, breis have been useful in discovering essential substances and the interrelation of reactions. The disintegration of tissue permits various essential materials to be more readily diluted by the suspending medium or destroyed by enzymes, so that their concentration falls below the optimal and their addition produces striking efiFects often not found with slices. The preparation of slices from some tissues, especially brain, is slow and delicate work, sampling is inaccurate, and one cannot be sure that the individual slices do not vary in activity. Dr. Libet and I (1) have recently studied the respiration of brain suspensions and we find that two types of suspension can be obtained. One type behaves as disintegrated tissue; the other type, when prepared under proper conditions, behaves very similarly to slices, shows a com- parable respiration rate, and is a suitable and very convenient preparation for the study of brain metabolism. The first type of suspension is obtained when brain tissue is homogenized by the apparatus of Potter and Elvehjem (2) in hypo- tonic medium, dilute phosphate buffer solution. Such suspensions respire at a low rate which may be increased up to 65 per cent by adding salt or sugar after homogenization. With such preparations it is therefore necessary that the osmotic pressure be equal in the control and experimental flasks when the effect of added substances is tested. These hypotonic suspensions have largely lost the power to utilize glucose. They show considerable effects when tissue extract or substances like fumarate are added. The rate of respiration per unit weight of tissue increases with increasing tissue concentration. The second type of suspension is obtained when the medium in which the brain is homogenized contains sufficient salt, sucrose, or glucose to make the osmotic pressure equal to that of serum. The respiration rate of such suspensions is up to 400 per cent greater than the rate of tissue homogenized in hypotonic medium. (Isotonic urea behaves like hypotonic solution.) Suspensions of whole brain (con- taining a large amount of white matter, which respires only slowly), prepared in 0.13 M sodium chloride-0.017 M phosphate buffer solu- tion, respire on the average at 71 per cent of the rate of an equal tissue weight of slices of pure gray matter in the same medium. Such suspensions are much less affected by the addition of tissue extract, fumarate, etc., than are hypotonic suspensions, and their respiration rate per unit weight is practically independent of the tissue concentration. Isotonic suspensions respire with the same res- DISCUSSION ON ANIMAL TISSUE RESPIRATION 273 piratory quotient as slices, and are similarly affected by additions of various substrates and by variations in the ionic content of the med- ium. An exception to this rule is that the respiration of slices, unlike that of suspensions, is less well maintained in isotonic sucrose than in isotonic saline solution. When isotonic suspensions are used, air may be used in reaction flasks instead of oxygen. This is an advantage, since it has been found that, while the respiration rates in the presence of air and in the presence of oxygen are identical for 90 minutes, the rate in the presence of oxygen falls off much more rapidly there- after. A similar slowly appearing toxic effect of oxygen can also be demonstrated with slices, but with slices the presence of oxygen is necessary, since otherwise the initial respiration rate is limited by inadequate diffusion of oxygen into the tissue. After homogenizing fresh brain in isotonic medium, the rate of respiration falls off very rapidly at first, thereafter less rapidly. Slices and isotonic suspensions prepared from slices do not show the initial very rapid decrease in rate. This suggests that a specially labile part of the respiratory activity can be observed with suspensions of fresh tissue, but is lost during the slow process of preparing for experi- ments on slices. Salts induce specific as well as osmotic effects. Isotonic solutions of sodium chloride, nitrate, sulfate, and especially phosphate increase the initial respiration rate appreciably when added to suspensions in isotonic sucrose. Bicarbonate in physiological concentration has no special effect. Calcium and magnesium ions have inhibitory effects on the respiration but cause better maintenance of rate. Potassium in high concentrations is also inhibitory. The initial inhibitory ef- fect of magnesium in the concentration found in serum is fairly small, and this ion is known to take part in reactions of carbohydrate metabolism. For studies of brain respiration, tissue homogenized in isotonic sodium chloride medium containing magnesium is therefore recommended. However, respiratory activities measured on suspensions or slices in any given medium cannot be considered to represent the true physiological activity of brain tissue in vivo until more is known of the effective concentrations of ions in the cells' immediate environ- ment in vivo and until more is known of the differences that may exist between brain in its normal physiologically active condition and the tissue which has been subjected to the abnormal injuries and stimuli of in vitro work. The marked difference between tissue homogenized in isotonic 274 A SYMPOSIUM ON RESPIRATORY ENZYMES and in hypotonic media occurs to some extent with certain tissues other than brain, but not with all tissues. PreHminary studies would have to be made on each tissue to which the method of isotonic suspensions is to be applied before it could be assumed that the behavior of suspensions would be comparable to that of slices of the particular tissue. REFERENCES 1. Elliott, K. A. C, and Libet, B., in press. 2. Potter, V. R., and Elvehjem, C. A., J. Biol. Chem. 114, 495 (1936). THE HOMOGENIZED TISSUE TECHNIQUE, THE DILUTION EFFECT AND ION EFFECTS VAN R. POTTER McArdle Me7norial Laboratory, University of Wisconsin We have developed the homogenized tissue technique for the study of isolated phases of metabolic activity. Since the enzymes that catalyze biological oxidations are in many cases extremely la- bile, we have used the device of homogenization to effect a physical isolation of a particular enzyme system by dilution. We believe that by so doing we can retain the original activity of the tissue and thus develop assay methods for specific enzymes, where separation by chemical treatment, such as fractional precipitation, could not result in 100 per cent yields and hence would be useless for assay purposes. Since certain components of various enzyme systems are readily soluble and are capable of diffusing away from each other or from solid phases, such as cytochrome oxidase, it is necessary to fortify the homogenate with these diffusible components. Whether it is necessary to add these accessoiy factors is determined by measur- ing the enz)Tne activity at various dilutions. If the measured effect is proportional to the amount of enzyme used, fortification is unneces- sary. With regard to choice of buffer, it should be pointed out that since we are now dealing in tenns of intracellular components, buf- fers based on extracellular fluids (such as serum) may not necessarily be optimum. A system which may illustrate these points is the succinoxidase system. It is inhibited by chloride ions, hence these are omitted from the buffer medium. It requires at least three soluble components for maximal activity, namely, cytochrome c, calcium ions, and aluminum ions. The dehydrogenase and the cytochrome oxidase appear to be associated with solid particles of protoplasm. When the dissociable DISCUSSION ON ANIMAL TISSUE RESPIRATION 275 factors are supplied, the activity of the system is just as great as that of intact tissue. More complex are the coenzyme systems in which two additional factors are soluble, namely, the dehydrogenase and the coenzyme. However, by means of fortification with the appropriate coenzyme and inhibition of the coenzyme nucleotidase it should be quite feasible to study these complicated systems in tissue homogenates. THE STIMULATORY EFFECT OF CALCIUM UPON THE SUCCINOXIDASE ACTIVITY OF RAT TISSUES A. E. AXELROD University of Wisconsin Variations in the ionic composition of the medium in which the surviving tissue respires are known to exert profound eflFects upon the extent of the respiration. Our attention was drawn to these ionic effects by the observation that calcium (as calcium chloride) stimulates markedly the succinoxidase activity of minced rat liver. This phenomenon was investigated further in the succinoxidase sys- tem of tissue homogenates prepared according to Potter and Elve- hjem. The following results were obtained. In the absence of added cytochrome c the succinoxidase activity of minced liver was in- creased 43 to 80 per cent by the addition of 20 micrograms of cal- cium. With homogenized liver (40 mg. per flask) the addition of 20 micrograms of calcium resulted in increases of 93 and 48 per cent in the absence and presence, respectively, of added cytochrome c (3 X 10'* mole per flask). The succinoxidase activity of homoge- nized kidney cortex (20 mg. per flask) was stimulated 40 per cent in the presence of added calcium. Added cytochrome c did not affect the magnitude of the stimulatory effect of calcium in this tissue. The most pronounced effect of calcium was observed in the case of homogenized heart tissue (20 mg. per flask), in which the addition of 20 micrograms of calcium in the presence of 3 X 10"* mole of cytochrome c caused an increase of 200 per cent in the succinoxidase activity. Under our experimental conditions the addition of 20 micrograms of calcium always yielded the maximum stimulatory effect. In many cases the addition of smaller amounts of calcium (as little as 1 or 2 micrograms) resulted in a marked acceleration of succinoxidase activity. In only a few isolated cases was a calcium effect observed in brain and skeletal muscle. The apphcation of the homogenized tissue technique to the study 276 A SYMPOSIUM ON RESPIRATORY ENZYMES of isolated enzyme systems has been discussed by Dr. Potter, who pointed out that all precautions must be taken to assure complete restoration of the diffusable factors necessary for the maximal ac- tivity of the enzyme system under consideration. The stimulatory eflFect of calcium upon the succinoxidase system in tissue homo- genates necessitates the addition of calcium to such a system if the maximal activity is to be attained. A dissociable complex involving calcium is indicated. The addition of aluminum ions has been shown by Dr. Potter to overcome an effect resulting from further dilution of the tissue. It thus becomes apparent that when either minced liver preparations or tissue homogenates are employed as the source of the succinoxidase system, it is necessary to add calcium and, under certain conditions, aluminum in order to eliminate these ions as possible limiting factors which may affect the validity of the succin- oxidase assay. TISSUE METABOLISM IN VITRO AND IN VIVO FREDERICK BERNHEIM School of Medicine, Duke University The work of Battelli and Stern in the early part of the century may be said to have begun the work on tissue metabolism in vitro which has culminated in the isolation of a number of dehydrogenases. Some of the enzymes which they showed to be present in tissue suspensions have since been shown to be of importance in the economy of the animal. In particular, all subsequent work on the physiology and pharmacology of alcohol has shown that its fate in the body can be accounted for by the activity of the alcohol oxidase of liver which they discovered. Some time later Warburg studied the effect of cyanide on isolated tissues, and from this work came the discovery of cytochrome oxidase and the cytochromes. Again, work on the pharmacology of cyanide has shown that its action on the animal can be explained on the basis of its inhibition of the cytochrome oxidase. In these two early groups of experiments the correlation between in vitro and in vivo results is good. With more recent work sirnilar correlations have either not been made or have not been satisfactor)^ The justification for working with broken cell suspensions or cell extracts is that it is only by this means that intracellular enzymes can be studied. The results thus obtained can indicate only that the cell or cell catalyst has certain potentialities under the given set of condi- tions. Under no circumstances are such results in themselves evi- DISCUSSION ON ANIMAL TISSUE RESPIRATION 277 dence for the normal activity of the cell. Attempts to make the condi- tions more "normal" by adding physiological salt solutions to tissue suspensions is essentially paradoxical, for it is known that the ionic environment inside the cell diflFers markedly from that outside, and thus, for the study of intracellular enzymes, physiological salt solu- tions are unphysiological. If better results are obtained by the addi- tion of such solutions to broken cell suspensions, it may simply mean that the salts are acting on the residual penneabilities of partially damaged cells rather than on the enzymes directly. The physiological significance of results obtained in vitro must be obtained by physiological means, i.e., by experiments on the whole animal. The classical method for studying intermediary metabolism in which substance x is fed and substance y is isolated in the urine does not ordinarily give enough detailed information about the fate of the substance. Correlation with the data obtained with isolated tissue suspensions is therefore often impossible. With the increasing use of labeled atoms animal experiments are yielding more precise results and thus offer promise of better correlation between in vitro and in vivo results. The story of sarcosine illustrates this point. When it is fed to a rabbit with benzoic acid, an increased amount of hip- puric acid is excreted. This shows that sarcosine in some way gives rise to glycine. When sarcosine with labeled nitrogen is fed, the hippuric acid excreted contains the labeled nitrogen. This shows that sarcosine must be demethylated in the body but not deaminated. When sarcosine is added to liver suspensions an extra oxygen uptake occurs and glycine is formed. The sarcosine undergoes an oxidative demethylation. In this series of experiments the in vitro results serve to elucidate the mechanism of a reaction known to occur in the intact animal. Only in this way is the physiological significance of results with tissue experiments established. Workers in tissue respiration tend to look for reactions of general significance. Results obtained with yeast or hashed pigeon breast muscle are applied with great facility to mammalian tissue. The omnipresence of the cytochrome oxidase and succinoxidase encour- ages this attitude, but does not justify it. Animals that differ so profoundly, physiologically, pharmacologically, and in all other ways, undoubtedly show differences in their respiratory mechanisms. Perhaps in the future more information will be obtained if these differences between animals are emphasized rather than minimized in an attempt to provide universal mechanisms that may have only superficial similarity. 278 A SYMPOSIUM ON RESPIRATORY ENZYMES PATHWAYS OF CARBOHYDRATE METABOLISM E. S. GUZMAN BARRON University of Chicago Whether the metabohsm of carbohydrates by animal tissues is always accomplished according to Embden-Meyerhof's and Cori's schemes of phosphorylations and oxidation-reductions has not yet been established. In fact, if the normal pathway is obstructed, the breakdown of carbohydrates might proceed through other pathways. Direct oxidation of glucose, of hexosemonophosphate, and of phos- phoglycerate might be the accessory pathways, although none of these has yet been shown to occur in animal tissues. The oxidation of a single substance may proceed via different enzyme systems; thus in sea urchin eggs (with no succinodehydrogenase and no cyto- chromes) the metabolism of carbohydrates undoubtedly proceeds through different pathways than in sperm (with succinodehydro- genase and cytochromes). The interesting findings of Korr (oxida- tions not inhibited by azide in resting cells and inliibited by azide when the cells are in active work) can be presented as examples in favor of this opinion. The existence of these multiple pathways makes possible the orientation of reactions which occur continually in living cells, the metabolism of pyruvate being the clearest example. This is illus- trated in the scheme shown on page 279 in which only the pertinent steps are reproduced. Pyruvate is an extremely reactive substance, and seventeen differ- ent pathways of its metabolism are known to exist in living cells. In animal tissues, in the absence of oxygen, part of the pyruvate formed during the breakdown of carbohydrate is reduced by dihy- drodiphosphopyridine nucleotide (Py(P04)2Ho) to lactate, a reac- tion the extent of which represents the degree of anoxia; part of it may be reduced to alanine (Warburg and Christian's d-amino acid oxidase) or may be used for transaminations or dismutations. In the presence of oxygen, pyruvate activated by diphosphothiamine- protein may be oxidized to acetylphosphate; may, through conden- sation reactions, be responsible for the synthesis of alpha-ketoglu- tarate, cisaconitate, acetoacetate, acetylmethylcarbinol, or carbo- hydrate. It might, as postulated by Wood and Werkman, combine with carbon dioxide to give oxalacetate, an important reaction still eluding direct demonstration, which is the base of many hypotheses for the breakdown and synthesis of carbohydrate. Oxalacetate formed in this way may be reduced by Py(P04)8Hg to malate, thus DISCUSSION ON ANIMAL TISSUE RESPIRATION 279 starting the oxidative pathway through Szent-Gyorgyi's cycle, or it may produce phosphopyruvic acid and thus start the synthesis of carbohydrate. These manifold reactions of pyruvate, all of them present in different degrees in animal tissues, show that the orienta- tion of reactions during the metabolism of carbohydrates is ex- tremely complex, and certainly diflFerent from tissue to tissue. ^, oxidation Glucose > ,^ II 1 , . oxidation Hexose monophosphate it Hexose diphosphate it Phosphoglyceraldehyde + Py (PO4) 2 it Phosphoglycerate + Py (PO4) 2H2 it t I Phosphopyruvate < Py(P04)2H2 i Lactate ^^ Pyruvate oxidation Diphosphothiamine-proteins O 4- '■ *^ - >> c3 -a is ^ 9, 03 S 4> H ^ ."S o a < ^■ Malate Fumarate Succinate Fe^"*"*" cytochrome c Cytochrome oxidase IT o. 280 A SYMPOSIUM ON RESPIRATORY ENZYMES THE CITRIC ACID CYCLE IN TISSUE METABOLISM FREDRICK J. STARE Washington University School of Medicine, St. Louis This is an appropriate time to mention the reasons why some of us do not believe in the importance, or even the presence, of citric acid as a component of the metaboHc cycle generally termed the "citric acid cycle." The citric acid cycle is based on carefully established experimental facts. Some of these were mentioned this morning by Dr. Evans. We question, however, the application of these facts, particularly the presence of citrate, to a cycle of importance in re- spiring muscle. It was only with high concentrations of pyruvate that Krebs was able to demonstrate increases in citrate, and these increases were of a small order, varying from 1 to 15 per cent. In tissue and body fluids, pyruvate occurs in a far lower concentration. It was only with high concentrations of pyruvate, citrate, and malonate that an increase in succinate was detected. I know of no evidence, but I believe that a high concentration of glutamate, and other related compounds not included in the citric acid cycle, would also yield increases in succinate in the presence of a high malonate concentration. The citric acid cycle assumes that malonate completely prevents the anaerobic reduction of oxalacetate to succinate, but there is no adequate proof for this assumption. Citrate is definitely a weaker substance, as compared with the other members of the cycle, in increasing or prolonging the oxygen consumption of respiring muscle or in eflFecting pyruvate removal by respiring muscle. In fact, citrate frequently inhibits such effects. Malonate in a concentration of 0.005 M which is 5 to 6 times less than the concentration used by Krebs, inhibits oxygen uptake and pyruvate utilization on an average of 70 to 75 per cent (pigeon breast muscle). It completely and always inhibits any catalytic effect that citric acid may show. Yet any other member of the cycle when added in an equal concentration of 0.005 M will completely over- come the malonate inhibition of oxygen uptake and of pyruvate removal. If the latter two depend in any way on a mechanism in- volving citric acid, the malonate should stop It because citric acid activity is always inhibited by malonate. Recently, in Dr. Barron's laboratory, Lipton, Goldinger, and I have studied pyruvate and citrate utilization in respiring pigeon muscle tissue. According to the citric acid cycle, each molecule of DISCUSSION ON ANIMAL TISSUE RESPIRATION 281 pyruvate is converted to citrate in the course of its oxidation. The cycle imphes that either citrate should be oxidized as rapidly as pyruvate, or that if citrate is not oxidized as rapidly as pyruvate, but still is a stage in the removal of pyruvate, it should accumulate in quantities suflBcient to account for the difference in the rates of utihzation of the two compounds. We found that pyruvate is oxidized at a far greater rate than citrate, and that citrate does not accumu- late. Our experimental observations favor a cycle involving a conver- sion of pyruvic acid to alpha-ketoglutaric acid, without citrate as an intermediary, followed by the Szent-Gyorgyi series of conversions of the dicarboxylic acids to oxalacetate. The occasional catalysis of respiration observed when citrate is added to muscle is probably not due to citrate itself but rather to alpha-ketoglutarate and the four-carbon acids which may be formed from it. Citrate may serve as a "stockroom" for the essential catalysts, exerting an effect on respiration only when they are low. Its synthesis from pyruvate may represent an unusual side reaction which under normal conditions is of little significance. However, under specific experimental condi- tions in vitro, such as large amounts of pyruvate, the rate of their reaction may be accelerated. A similar condition, with a high level of pyruvate, appears to exist in vivo. Thus Sober, Lipton, and Elve- hjem found that in the recovery from acute thiamine deficiency large amounts of citrate are excreted. In concluding these remarks, may I emphasize that these criti- cisms of the citric acid cycle are directed against a citric acid cycle which contains citric acid; they do not apply to a citric acid cycle which contains no citric acid. Proponents of the citric acid cycle quite properly spend most of their energy proving that some sort of a cycle exists rather than attempting to answer the question whether citric acid is, or is not, a member of it. Ifo ' '^i