ISOTOPES IN BIOCHEMISTRY Reports of Conferences held at The Ciba Foundation and published by The Blakiston Company : — TOXEMIAS OF PREGNANCY Human and Veterinary LIVER DISEASE In the press STEROID HORMONES AND TUMOUR GROWTH . . . Ready end 1951 METABOLISM OF STEROIDS AND THEIR ESTIMATION Ready end 1951 Other volumes in preparation GIBA FOUNDATION CONFERENCE ON ISOTOPES IN BIOCHEMISTRY Consulting Editors J. N. DAVIDSON, M.D., D.sc, f.r.s.e. L. H. GRAY, M.A., Ph.D. A. S. McFARLANE, m.\., b.Sc, m.b., ch.B. A. NEUBERGER, m.d., Ph.D., f.r.s. G. POPJAK, M.D.,D.Sc. C. RIMINGTON, M.A.,Ph.D. Editor for the Ciba Foundation G. E. W. WOLSTENHOLME, o.b.e., m.a., m.b., B.ch. With 79 Illustrations THE BLAKISTON COMPANY PHILADELPHIA 1951 Aix Rights Reserved Printed in Great Britain Published in London by J. dh A. Churchill, Ltd. 104, Gloucester Place, W.l FOREWORD hy Sir Charles Harington, f.r.s., m.a., Ph.D. The advent of an entirely new tool of research is always something of an event in science. There can be few more striking illustrations of this than the story of the advances in knowledge that have been made with the aid of isotopes. The speed of development of this branch of research is due to the realization of some of the potentialities of the method even before separated isotopes were available, and to the eagerness with which such pioneers as Hevesy exploited the possibilities as soon as the discovery of induced radioactivity made the preparation of minimal amounts of radioactive isotopes possible ; it was still further increased by the dramatic circumstances of the war-time work on nuclear fission, which brought the technique within the reach of laboratories in all branches of science. So far as biological science is concerned it was obvious at the outset that the tracer isotope technique would find its widest application in biochemistry ; it offered indeed a method of attack on problems of intermediary metabolism and biosynthesis which had hitherto been unapproachable. It was unfortunate, though perhaps only natural, that the sudden general availability of isotopes for biochemical research should have led to sonje rather hasty experimenta- tion in which the more obvious problems were studied without deep thought. Such work, although it may provide quick answers to some questions of interest, is far from representing the best use of the new method. As time has passed, there has been more opportunity for reflection, and the result of this is apparent in the nature of the biochemical problems that are now being tackled with the vi Foreword aid of isotopes as tracers. All of these problems require powers of biochemical interpretation ; many of them demand also new work in organic chemistry, both for the synthesis of the specifically labelled precursors whose metabolic fate it is desired to study, and for the stepwise degradation of the biosynthetic product, so that the process of its formation can be analysed. Again new methods of isolation of intermediary metabolites may have to be devised. Thus the full develop- ment of the isotope technique in this field can only be achieved with the aid of concurrent advances in more con- ventional chemical and biochemical methods. The collection of papers in the present volume, contributed at an informal meeting by a number of leading workers, affords as good an illustration as could be wished of the stage of isotope research in biochemistry that has now been reached by the process of development that has been outlined. The Ciba Foundation are to be congratulated on performing a useful service in assembling a number of contributions which are not only of the greatest intrinsic interest but, in so far as they indicate the potentialities of isotope research in biochemistry, will serve as an inspiration for the further cultivation of this productive field of work. CONTENTS PAGE Foreword by Sir Charles Harington, PhD, frs, Director, National Institute for Medical Research, London . . . . v Opening Address A. S. McFarlane, ma, mb, chB, National Institute for Medical Research, London ...... 1 Part I— STEROIDS Chairman: A. S. McFarlane, ma, bsc, mb, chB Editor: A. S. McFarlane, ma, bsc, mb, chB Metabolism of i*G-labeiIed steroids C. P. Leblond, PhD, Dept. of Anatomy, McGill University, Montreal ......... 4 Discussion R. J. BoscoTT, C. P. Leblond, M. Marois, C. Heidel- BERGER ......... 12 High cholesterol content of human spleen K. Bernhard, Physiologisch-chemisches Institut, University of Basle ......... 14 Discussion D. Rittenberg, K. Bernhard, K. Bloch ... 16 The biosynthesis of radioactive cholesterol by surviving liver slices S. GuRiN, PhD, and R. O. Brady, md, Dept. of Physiological Chemistry, University of Pennsylvania .... 17 Discussion K. Bloch, W. G. Dauben, D. Rittenberg ... 24 Studies with deuterium steroids T. F. Gallagher, PhD, Sloan-Kettering Institute, New York 28 Discussion R. J. BoscoTT, T. F. Gallagher, D. Rittenberg, K. Bloch 38 Vll viii Contents PAGE Part II— HEMOGLOBIN AND METABOLIC DERIVATIVES Chairman : [Sir Charles Harington, PhD, frs Editor : C. Rimington, PhD The biosynthetic mechanism of porphyrin formation D. Shemin, PhD, and J. Wittenberg, phD. Dept. of Bio- chemistry, Columbia University, New York ... 41 Discussion A. Neuberger, D. Shemin, H. G. Wood, K. Bloch, C. Rimington, D. Rittenberg, S. Gurin .... 64 Studies on mammalian red cells A. Neuberger, md, PhD, frs. National Institute for Medical Research, London . . . . . . .68 Discussion B. Thorell, a. Neuberger, A. Wormall, D. Ritten- berg, P. L. Mollison ...... 82 Preliminary investigations for a study of energy utilized by the surviving fowl erythrocyte in haem synthesis C. Rimington, PhD, Dept. of Chemical Pathology, University College Hospital, London ...... 86 Discussion D. Shemin, C. Rimington, H. G. Wood, H. A. Sloviter 89 Iron metabolism in pathological conditions A. Vannotti, md, Clinique M^dicale Universitaire, Hopital Cantonal, Lausanne ....... 90 Discussion A. S. McFarlane, a. Vannotti, K. Aterman, J. E. Falk 94 Part III— USE OF TRACERS IN THE STUDY OF BIOLOGICAL EFFECTS OF RADIATION Chairman: Sir John Cockroft, frs Editor : L. H. Gray, ma, PhD The modification of X-ray sensitivity by chemicals A. HoLLAENDER, PhD, G. E. Stapleton, and W. T. Burnett, Jr., Division of Biology, Oak Ridge National Laboratory, Tenn 96 Contents ix PAGE Discussion L. H. Gray, A. Hollaender, M. Gordon, B. E. Holmes, H. A. Sloviter, a. S. Parkes, J. F. Loutit, L. F. Lamerton ......... 109 Effect of X-rays on nucleic acid and protein synthesis in the Jensen rat sarcoma Barbara E. Holmes, PhD, Dept. of Radiotherapeutics, University of Cambridge . . . . . . .114 Discussion R. J. BoscoTT, B. E. Holmes, G. B. Brown . . . 120 Radiation dose in tracer experiments involving auto- radiography S. R. Pelc, PhD, Radiotherapeutic Research Unit of the Medical Research Council, Hammersmith Hospital, London 122 Discussion L. F. Lamerton, S. R. Pelc, L. H. Gray, A. Hollaender, B. E. Holmes 136 Synthesis of deoxyribose nucleic acid and nuclear incor- poration of ^^S as shown by autoradio graphs Alma Howard, PhD, and S. R. Pelc, PhD, Radiotherapeutic Research Unit of the Medical Research Council, Hammer- smith Hospital, London . . . . . . .138 Discussion C. Heidelberger, a. Howard, C. P. Leblond, G. PopjAK, S. R. Pelc, G. B. Brown, B. E. Holmes, A. S. McFarlane ........ 148 Part IV— NUCLEIC ACIDS Chairman : A. S. McFarlane, ma, b.sc, mb, chB Editor : J. N. Davidson, md, dsc, frs.ed The biosynthesis of pyrimidines irt vitro D. Wright Wilson, Pho, Dept. of Physical Chemistry, University of Pennsylvania ...... 152 Discussion E. Hammarsten, K. Aterman, A. Neuberger, D. W. Wilson ......... 162 Studies with organic- and bio- synthetic nucleosides and nucleotides G. B. Brown, phD, Sloan-Kettering Institute, New York . 164 X Contents PAGE Discussion C. Heidelberger, E. Hammarsten, G. B. Brown . . 171 The use of radiophosphorus in the study of the nucleic acids J. N. Davidson, md, dsc, frs.ed, Biochemistry Dept., University of Glasgow . . . . . . .175 Rate of synthesis and quantitative variations of the ribo- nucleic acid during the grov^th of a culture of Polytomella coeca R. Jeener, Dept. of Animal Physiology, Universite Libre de Bruxelles (presented by D. Szafarz) .... 184 Discussion G. B. Brown, J. N. Davidson, C. P. Leblond, D. Szafarz, G. PopjAK ......... 188 Part V— PROTEINS AND AMINO-ACIDS Chairman : A. S. McFarlane, ma, bsc, mb, chb Editor : A. Neuberger, md, phd, frs A method for the evaluation of the rate of protein synthesis in man D. RiTTENBERG, PhD, Dept. of Biochemistry, Columbia University, New York . . . . . . .190 Discussion A. Neuberger, D. Rittenberg*, . • • . . 201 Turnover rates during formation of proteins and poly- nucleotides in regenerating tissues E. Hammarsten, md, Dept. of Biochemistry, Karolinska Institutet, Stockholm ....... 203 Discussion D. Rittenberg, E. Ha:mmarsten, G. B. Brown . .211 Synthesis of phenylalanine and tyrosine in yeast K. Bloch, PhD, Dept. of Biochemistry, University of Chicago 213 Discussion R. Bentley, H. R. V. Arnstein, E. E. Pochin, D. Rittenberg, K. Block, H. A. Krebs, H. G. Wood . 224 Contents xi PAGE Part VI— CARBOHYDRATE AND FATTY ACID METABOLISM Chairman : F. G. Young, PhD, dsc, fric, frs Editor : G. Popjak, md, dsc. A study of acetone metabolism, using glycogen and serine as indicators, and the role of C^- compounds in metabolism H. G. Wood, phd, Dept. of Biochemistry, Western Reserve University, Cleveland, Ohio ...... 227 Discussion H. A. Krebs, H. G. Wood, D. Rittenberg, K. Bloch . 244 Asymmetric citric acid C. Heidelberger, PhD, and V. R. Potter, PhD, McArdle Memorial Laboratory, University of Wisconsin . . . 246 Discussion H. R. V. Arnstein, C. Heidelberger, R. Bentley, H. A. Krebs 256 Mode of formation of fatty acids from acetate and glucose, as studied in the mammary gland G. PopjAK, MD, DSc, National Institute for Medical Research, London ........ 258 Discussion S. J. Folley, G. Popjak, S. Gurin, K. Bloch, R. Eraser, F. G. Young, H. A. Krebs ...... 276 List of those participating in or attending the Conference, 12th to 15th March, 1951 G. W. Anner H. R. V. Arnstein K. Aterman R. Bentley K. Bernhard K. Bloch . R. J. BOSCOTT J. C. BOURSNELL G. B. Brown L. BUGNARD C. Chagas . I. Clark Sir John Cockroft J. W. Cornforth J. COURSAGET W. G. Dauben . J. N. Davidson T. E. DE ESTON V. R. DE ESTON . J. E. Falk s. j. folley Russell Eraser T. F. Gallagher R. Glascock M. Gordon C. H. Gray L. H. Gray J. Gross Ciba Ltd., Basle National Institute for Medical Research, London University of Birmingham National Institute for Medical Research, London University of Basle I University of Chicago University of Birmingham University of Cambridge Sloan- Kettering Institute, New York Institut General d'Hygi^ne, Paris Rio de Janeiro Columbia University Atomic Energy Research Establishment, Harwell National Institute for Medical Research, London Lab. des Isotopes, Hopital Necker, Paris University of California University of Glasgow Sao Paulo, Brazil Sao Paulo, Brazil University College Hospital, London National Institute for Research in Dairying, Reading Postgraduate Medical School, London Sloan-Kettering Institute, New York National Institute for Research in Dairying, Reading University of California King's College Hospital, London Radiotherapeutic Research LTnit of the Medical Research Council, London National Institute for Medical Research, I^ondon, and McGill L^niversity, Montreal xui XIV List of Conference Attendance S. GURIN . E. Hammarsten Sir Charles Harington C. Heidelberger A. F. Hensox H. P. HiMSWORTH a. hollaender Barbara E. Holmes Alma Howard . A. St. G. Huggett E. KODICEK H. A. Krebs T. H. Kritchevsky L. F. Lamerton C. P. Leblond . J. F. LOUTIT A. S. McFarlaxe M. Marois A. J. P. Martin . R. H. Mole p. l. mollison . Helen M. Muir A. Neuberger . A. S. Parkes S. R. Pelc J. C. Perrone . Rosalind Pitt Rivers E. E. POCHIN G. POPJAK C. RiMINGTON D. RlTTE^JBERG . University of Pennsylvania Karolinska Institutet, Stoekholm National Institute for Medical Research, London McArdle Memorial Laboratory, University of Wisconsin Imperial Chemical Industries, Ltd., Wehvyn Medical Research Council, London Oak Ridge National Laboratory, Tenn. University of Cambridge Radiotherapeutic Research L^nit of the Medical Research Council, London St. Mary's Hospital, I^ondon Dunn Nutritional Laboratory, University of Cambridge University of Sheffield Sloan-Kettering Institute, New York Royal Cancer Hospital, London McGill L^niversity, Montreal Radiobiological Research Unit, Harwell National Institute for Medical Research, London College de France, Paris National Institute for Medical Research, London Atomic Energy Research Establishment, Harwell Postgraduate Medical School, London National Institute for Medical Research, London National Institute for Medical Research, London National Institute for Medical Research, London Radiotherapeutic Research Unit of the Medical Research Council, London National Institute for Medical Research, London, and Brazil National Institute for Medical Research, I^ondon University College Hospital, London National Institute for Medical Research, London University College Hospital, London Columbia University, New York List or Conference Attendance XV D. SUEMIN H. G. B. Slack H. A. Sloviter A. SOMERVILLE D. SZAFARZ C. Terner B. Thorell A. Vannotti N. Veall . H. Vetter E. B. Vischer A. Wettstein W. F. Widdas D. W. Wilson T. H. Wilson H. G. Wood P. F. Work A. WORMALL F. G. Young Columbia University, New York National Institute for Medical Research, London National Institute for Medical Research, London, and L^niversity of Pennsylvania. Imperial Chemical Industries, Ltd., Wehvyn Universite Libre de Bruxelles National Institute for Research in Dairying, Reading Institute for Cell Research and Genetics, Stockholm Hopital Cantonal, Lausanne Radiotherapeutic Research Unit of the Medical Research Council, London 2nd Med. Univ. Clinic, Vienna Imperial Chemical Industries, Ltd., Welwyn Ciba Ltd., Basle St.^Mary's Hospital, London University of Pennsylvania University of Oxford and University of Pennsylvania W^estern Reserve University, Cleveland National Institute for Medical Research, London St. Bartholomew's Hospital, London University of Cambridge CHAIRMAN'S OPENING REMARKS A. S. McFARLANE With some notable exceptions, the biological use of isotopes has lagged in Europe by comparison with America, and this is a matter which has to be put right if the high standard of European biological science is to be maintained. A step towards doing this can be made by means of conferences with our American colleagues and to-day we make an important start in this direction. We are helped by having the example of several similar post-war U.S. conferences to follow — notably the meetings held at Wisconsin in 1947, and at Cold Spring Harbor in 1948, and we are additionally fortunate in being the guests of the Ciba Foundation. This will be the tenth conference held by the Foundation and already a high reputation has been established for efficiency. You will note a main idea behind their organization, namely, strict limitation of membership to people who are actively engaged in research in the subject under discussion. Proceedings will be recorded, but nothing will be published without the approval of the speaker concerned. In the privacy of this room, and in the company of kindred workers, it is hoped that discussion will be full and only restrained by considerations of time. Everyone will agree, I am sure, that such a conference is held most opportunely. Almost without exception, the Ciba Foundation's invitations have been accepted and difficulties of travelling largely ignored. One^ is tempted to enquire why leading isotope users — all of them busy people — are prepared to travel vast distances to be here. What is this conference likely to offer that cannot be obtained at home by reading the literature? In the few moments at my disposal I would like to try to answer this. Biological application of isotopes has brought to light many surprises, with which we are all familiar, but I think most ISOTOPES 1 2 2 Chairman's Opening Remarks people will agree that the result to date is by no means a simplification of oiir former conceptions of how the animal body works. It appears rather that the use of isotopes has uncovered problems of greater complexity than the ones they were intended to solve. Perhaps this is not quite a fair generalization since at least in the case of two elements, namely iron and iodine, relatively clear-cut metabolic pictures have emerged — and for the same reasons, namely that the body reserves of both elements are relatively small, and in each case one metabolic pathway predominates over all others. In the case of almost all other elements and especially, of course, with compounds of carbon complex interconnected chains of biochemical events have come to light. Where these chains cross and so-called "metabolic pools" exist, the investigator's field of interest is necessarily broadened beyond that of his immediate problem; for example, the expert in porphyrin synthesis is compelled to take an interest in the broader field of amino-acid metabolism. It seems that more than ever before the specialist has to pay attention to his general biochemistry, and herein presumably lies one reason why he welcom.es such an opportunity as this to meet experts in other branches of biochemistry. The second reason arises from the fact that the literature is remarkably disjointed and incomplete, and it is extremely difficult for the reader to correlate the findings of one tracer user with those of another. For instance, in spite of du Vigneaud's demonstration that doubling the dietary level of methionine in the rat increases the proportion of amino-acid oxidized nine-fold, some workers still employ doses of labelled compounds which are comparable to the dietary intake. Others use the dubious procedure of starving or otherwise depleting their animals beforehand in order to enhance the isotope uptake. Far too often, also, samples are taken — especially of blood — which are large enough to disturb signi- ficantly the physiological balance. These difficulties are all avoidable by the preparation of labelled material of high specific activity, by refinements in Chairman's Opening Remarks 3 counting technique or by the use of larger experimental animals. There will be advantages also in using one, or at most a few, kinds of animal, since species differences may be surprisingly great. Whereas in normal man 90 per cent of a dose of radioactive-iron is utilized for haemoglobin production, the same figure in the rat or dog never rises above 60 per cent. There is also little uniformity in the manner of administering labelled substances. Whereas Whipple and his collaborators feed amino-acids to their dogs and find peak values for free labelled amino-acids in the plasma at 4-6 hours, Borsook uses intravenous injection and finds that the free amino-acid has largely disappeared from the plasma in 10 minutes. Even the method of expressing results is sometimes confusing. Whereas some plasma protein investigators give their results in the form of the proportion of amino-acid in the protein which was derived from the labelled amino-acid in the diet, others give the proportion of total administered isotopic amino-acid which appears in a given weight of plasma protein — and the figures cannot easily be related. These inconsistencies which may appear to be of a minor technical nature stand, nevertheless, in the way of an alto- gether wider view of metabolic events. If our discussions, ranging as they will over several important fields and con- tributed to by leading protagonists in each, can serve to smooth out the technical difficulties while at the same time keeping in view the broader principles, the effort made by many to attend the Conference and by the Ciba Foundation to arrange it will be amply repaid. PART I -STEROIDS METABOLISM OF i^C-LABELLED STEROIDS* C. P. LEBLOND The basic postulate of tracer work is that the changes exhibited by the radioactive material also take place normally under physiological conditions. This is true only if the amount of radioactivity is kept low enough to avoid radio- chemical effects, and if the amount of the labelled substance is such that it can mix with the same substance in body fluids and tissues without significantly altering the physio- logical level of the material under investigation. In the case of ^^C labelled steroids, the first requirement is easily satisfied but not the second. "f It is indeed quite difficult to trace physiological amounts of ^^C-labelled substances which exhibit their biological activity in minute doses, as is the case with oestrogens. The problem, on the other hand, is somewhat easier in the case of substances active in larger doses, for example progesterone, with which we may hope to obtain in the future truly physiological results. At any rate, a greater amount of radioactivity per mg. of substance, i.e. a higher specific activity, may be expected from (1) the use of ♦Report from the Depts. of Biochemistry (R. D. H. Heard, F. Peron, J. Saffron, L. Thompson and C. Yates) and of Anatomy (R. C. Greuheh and C. P. Leblond), McGill University, Montreal. T Substances labelled with radiocarbon must of necessity contain a laroje amount of carrier since the half life of ^^C is about 5,500 years, and therefore, one millicurie must contain at least enough atoms to have 3-6x10' atoms exploding every second over thousands of years. Thus, 1 mc. of radiocarbon as barium carbonate must weigh at least 3 mg. In practice, the available radiocarbon weighs several times that amount, due to contamination with non-radioactive carbon in the course of preparation. As a result, the dose of a ^^C-labelled substance which is administered for tracing purposes will often be greater than the amoimt of the natural substance present in body fluids. 4 Metabolism of ^^C-Labelled Steroids 5 radiocarbon uncontaminated with non-radioactive carrier, and (2) the labelHng of as many carbon atoms as possible. Two types of methods for tracing radioactive material in the body exist. The first is by the Geiger Counter tube, by which quantitative estimates are obtained with extracts of tissues and excreta. The other technique consists of the recognition of radioactivity by means of its effect on photographic emulsion. In this case, the detector is the photographic grain, the size of which makes it suitable for the microscopic localization of radioactivity. The autoradiograph technique consists of placing emulsion in intimate contact with a thin section of tissues of animals sacrificed at various intervals after administration of radioactive material. Recently, a similar method has been applied to the localization of radio- active spots in paper chromatograms. This method, which makes it possible to identify substances with a precision com- parable to that of melting point determinations, can be used to locate infinitesimal amounts of a substance provided it contains radioactivity and a sufficiently long exposure is used. However, the use of both the Geiger counter and radio- autographic techniques is complicated by the low specific activity of the carbon present in radioactive steroids. Thus, in using the Geiger counter, the easiest method with the soft jS-rays of ^*C is to employ preparations of infinite thickness, for which corrections for absorption are unnecessary. But this method has a low sensitivity, and to detect the small amounts present in tissues, thin samples may have to be used, and complex corrections for absorption may have to be made. In autoradiography the main difficulty is that the preparations require very long exposures, a year in the case of the slide illustrated in Figs. 1 and 2. A number of radioactive steroids have been synthesized: progesterone labelled with ^*C in positions 21 or 3; cestrone labelled in position 16; and deoxycorticosterone acetate (DCA), labelled in positions 21 and 3. Thus, the chemical part of the research is well advanced, more so than the biological. There have been a few reports on the use of 6 C. P. Leblond radioactive steroids: (1) We examined the distribution of radio-iodine-labelled oestradiol and the outstanding findings (presented in May, 1947, before the American Association for Cancer Research) were the major role played by the gastro- intestinal system in promptly disposing of the material, and the lack of any remarkable concentration in the secondary sex organs with the exception of the mammary glands. Unfortunately, when the di-iodo- and mono-iodo-a-oestradiol used were assayed in spayed rats by the vaginal smear test, which is sensitive to 0-1 fig. of oestradiol, neither of these substances displayed oestrogenic activity in doses up to 100 fig. Thus, some doubt has been cast on the significance of the results obtained with these substances (Albert, Heard, Leblond and Saffran, 1949). (2) Twombly, McClintock and Engelman (1948) obtained similar results with equilin bro- minated with radio-bromine. (3) A recent paper by Riegel, Hartop and Kittinger (1950) deals with the metabolism of [21-^^C]-progesterone in mice and rats. Their main findings are that a considerable proportion of the material is recover- able as labelled COg in the expired air. Some is excreted in the faeces but none in the urine. Finally, a significant con- centration of material is present in the pituitary, adrenals and ovaries. Like the above, the results to be reported here are of a preliminary nature and refer to pilot experiments only. Three steroids have been investigated so far: [16-^*C]-oestrone, [21-^*C]-progesterone, and [21-^*C]-deoxycorticosterone ace- tate (DCA). Methods and Techniques (1 ) Preparation of Radioactive Steroids. Radioactive oestrone was prepared by Heard, Saffran and Thompson by opening ring D, addition of ^*C-diazomethane to the resulting aliphatic side chain through the Arndt and Eisert reaction and pyrolysis of the corresponding homo-di-acid back to the original ketone. Radio-progesterone and radio-deoxycorticosterone were obtained by Heard and Yates. [21-^*C]-progesterone Metabolism of i*C-Labelled Steroids 7 was prepared from [21-^*C]-21-diazoprogesterone by the action of hydriodic acid, and [21-^^C]-DCA from the same labelled intermediate by treatment with acetic acid. 21- Diazoprogesterone was prepared from 3-keto-J*-etiocholenone chloride by an improved Arndt-Eisert synthesis. (2) Mode of Administration. After trying a variety of solvents and doses, the following procedure was adopted. A dose of 1 mg. of radioactive steroid is dissolved in 0-25 ml. ethyl laurate and injected at 9 p.m. intramuscularly in the neighbourhood of the head of the femur. Immediately after injection the animals are placed in a metabolism chamber without food or water. The air circulating through the chamber is first freed of CO2 by being passed through barium hydroxide solutions. The air coming from the chamber is circulated through two sodium hydroxide bottles for trapping the CO2 exhaled by the animals. Contact of the feet of the animals with the floor of the chamber is prevented by a wire sheet, in such a way that the urine falls to the bottom but the faeces are retained on the wire. All the animals used for complete distribution studies were sacrificed twelve hours after injection by exsanguination from the inferior vena cava under ether anaesthesia. In addition, 5 mice were given 1 mg. of oestrone, and the faeces and urine were collected for 144 hours, after which the animals were sacrificed. CO2 was collected for the first 12 hours and again for 12 hours starting at 97 hours after injection. Five mice were similarly treated with progesterone and 5 other mice with DCA. (3) Preparation of Samples of Tissues and Excreta for Counting. In the animals sacrificed 12 hours after injection, about 40 tissues and excreta were examined with the Geiger counter. In the animals sacrificed at 144 hours, only plasma, kidney, adrenal, thyroid, hypophysis, liver and gall bladder were investigated. The organs were removed from the animals and the excreta digested in sodium hydroxide, plated on a brass planchette and the plates carbonated. The more reliable results were obtained with infinitely thick preparations 8 C. P. Leblond The technical details will be published elsewhere. The results were expressed as the percentage of the recovered dose and therefore did not include the amount retained at the site of injection. The concentrations were also recorded as the number of counts per mg. of organ over the number of counts recovered in tissues and excreta per mg. of body weight. (4) Preparation of Autoradiographs. The tissues removed from the animals for autoradiography were fixed in 10 per cent formalin at pH 5-6. The tissues were cut on the freezing microtome and coated with photographic emulsion according to a previously described method. The slides were then placed in a light-tight box containing a dehydrating agent and were stored in a low temperature cabinet for several months. The preparations were developed and fixed according to routine photographic procedures. They were then mounted under balsam and protected with a coverslip (Leblond, Percival and Gross, 1948). Results (1) Distribution of (Estrone. The distribution of radioactive oestrone in about 40 organs and tissues was examined in two tumour-bearing CgH mice. Animal number 1 had several very large tumours, and animal number 2, a rather small one. The plasma contained a fair amount of radioactivity, approx- imately equal to that found in the red blood cells. An out- standing feature of the metabolism was the major role played by the gastro-intestinal tract. There was little radioactivity in the stomach but a considerable amount in the intestine and faeces. In animal number 2, the bile was highly radio- active, thus indicating a major role of biliary excretion. The secondary sex organs, uterus, vagina and mammary gland contained a small amount of radioactivity. Urinary excretion of oestrogens has received much emphasis in the past. Indeed, 27 per cent of the dose was present in the urine of the first animal and 10 per cent in the second. A fair concentration was found in the kidney, especially in animal number 1. Metabolism of ^C-Labelled Steroids 9 The nature of the radioactive material present in urine and faeces was especially investigated in the animals sacrificed at 144 hours. Thus, in the urine only 15 per cent of the activity was removable by ether and, after acid hydrolysis in the autoclave, only another 10 per cent became ether- soluble, the rest, or roughly 75 per cent, being water-soluble. The situation in the faeces was very similar, with only a small portion soluble in ether. In the long duration experiment with 5 mice the loss of CO2 was 2-0 per cent in the first 12 hours following injection. Another estimation at 97 hours after administration gave an elimination of 1 • 1 per cent for a 12-hour period. (2) Distribution of Progesterone. A complete distribution of radioactivity after administration of radioactive progesterone was determined in one tumour-bearing C3H mouse. The following points were observed: (a) the radioactivity in the plasma was not high and was within the range of that of the red blood cells. The blood level may have been maintained for a very long period since it remained appreciable at 144 hours. (6) A major role of biliary excretion was indicated by a high concentration in the bile and large amounts in the intestine at 12 hours after injection, (c) However, urinary excretion tended to be more pronounced than faecal excretion, at least during the first 12 hours following injection, (d) As in the case of oestrone, partition of urine and faeces revealed that the greatest proportion of the excreted material was water-soluble and remained so after hydrolysis, (e) Neither secondary sex organs, nor endocrines showed a high concentra- tion of radioactivity. (/) The loss of radioactivity as COg was 2-5 per cent within 12 hours after injection in the long duration experiment with 5 mice. Another estimation at 97 hours after administration gave an elimination of • 9 per cent for a 12-hour period. In another experiment, two mice were castrated and three weeks later one of them received 2 /xg of oestradiol a day for 3 days, after which they both received the standard dose of radioactive progesterone. The cestradiol-treated animal was 10 C. p. Leblond the only one to have a concentration above background in the uterus. The autoradiographic detection of pure progesterone has not yet yielded results. However, impure progesterone containing methyl 3-keto-J*-etiocholenate showed interesting autoradiographic reactions. The most precise localization was observed in liver and uterus (Figs. 1 and 2). (3) Distribution of Deoxycorticosterone Acetate. The results at 12 hours after injection were very similar to those reported above. However, with this substance the urinarv loss was greater than the fascal loss. The intestinal content of radio- activity at 12 hours after injection was correspondingly lower than with other steroids. The material in urine and faeces was mainly water-soluble. The smallest concentration to be found in any organ was in the adrenal. The radioactive material persisted in blood until 144 hours after injection, while it had vanished from other organs and tissues. The loss as CO2 was 6-7 per cent within 12 hours after injection in the 5 mice in the long duration experiment. At 97 hours no activity could be detected in the CO2 over a 12-hour period. Discussion The most clear cut results were those obtained in relation to excretory processes. Fgecal excretion played an outstand- ing role, especially with cestrone and progesterone; and, since the highest concentrations by far were found in the bile, there was little doubt that the bile was the source of the material excreted in the faeces through the intestine (as demonstrated by Albert et al. for the iodo-oestradiols). The fact that the faecal material, as well as the urinary material, was mainly insoluble in fat solvents even after hvdrolvsis indicated that the metabolism of oestrone, progesterone and deoxycorticos- terone in mice differed from what was known to occur in man. This was indeed found to be the case in the work reported by Gallagher in this volume. The excretion as COg was appre- ciable, but smaller than that by faecal and urinary routes. The fact that we obtained smaller CO2 excretion and greater Fig. 1 . Unstained autoradiographs of kidney (upper rioht), uterus (U, upper centre), and liver (L, lower). Low magnification. A light reaction is present in kidney and most of the uterus. A pronounced reaction is present in liver tissues. The reaction is minimal near the hepatic vein and portal veins, and maximal half way between the two. ■'■■■^' ■-?'fe^- Fig. 2. Unstained autoradiograph of uterus. Higher magni- fication. An intense reaction of the endometrium is apparent. To face page 10] Metabolism of ^*C-Labelled Steroids 11 urinary elimination of progesterone than did Riegel and co-workers could probably be explained by the difference in injection routes, since these authors used intraperitoneal administration. The absorption through the liver which occurred in this case would lead to a greater breakdown to CO2 and a greater excretion into the faeces through the bile. It is indeed well known that progesterone and its derivatives are detoxified in the liver (Leblond, 1942). The radioactivity figures for organs were not very informa- tive. They may be expected to become most useful when two further steps in the progress of this investigation are taken: the chemical analysis of the nature of the radioactive material present, a study which will require the use of such techniques as paper chromatography; and the autoradiographic localiza- tion of the radioactive material. Thus, in the case of the uterus, sections of which are shown in Figs. 1 and 2, the radioactivity content as estimated with the Geiger counter did not significantly differ from that of other organs. Never- theless, some radioactive material was selectivelv concentrated in the epithelial cells. Summary Preliminary experiments have been carried out with oestrone, progesterone and deoxycorticosterone acetate labelled with ^*C. The results obtained and some possibilities for future work are discussed. Acknowledgemeni This work was carried out with the help of grants from the National Cancer Institute of the U.S. Public Health Service (Department of Biochemistry) and of the National Cancer Institute of Canada (Department of Anatomy). REFERENCES Albert, S., Heard, R. D. H., Leblond, C. P., and Saffran, J. (Id49). J. biol. Chetn., 177, 247. Leblond, C. P. (1942). Arner. J. med. Sci., 204, 566. 12 C. P. Leblond Leblond, C. p., Percival, W. L., and Gross, J. (1948). Proc. Soc. exp. Biol, iV.y., 67, 74. RiEGEL, B., Hartop, W. L., Jun., and Kittinger, G. W. (1950). Endocrinology, 47, 311. TwoMBLY, G. H., McClintock, L., and Engelman, M. (1948). Amer. J. Obstet. Gynec, 56, 260. DISCUSSION Boscott: Have you tried tritiated oestrone or oestradiol? It pre- sumably should be available from J«-oestrone, which Dr. Djerassi of Syntex Ltd., Mexico, has recently synthesized from cholesterol. Also, Dr. Miescher's work on the J ^ ^-progesterone series should give another route of tritiated progesterone, deoxycorticosterone and so forth. Would tritium have any advantage over "C? Leblond: These possibilities have not been examined, since there is a considerable difference in mass between hydrogen and tritium, and physiological behaviour may be different. Even with deuterium results different from those obtained w4th hydrogen have been recorded, for instance in the case of water metabolism (Barbour). Boscott: Tritiated hexoestrol has been recently synthesized at Oak Ridge.* Have you heard anything about the results? Leblond: I am not acquainted with the biological results. Marois: La belle communication dc Dr. Leblond apportc de nombreux faits importants. II en est un que je veux souligner particulierement: e'est la forte radioactivite de I'hypophyse chez des animaux traits par la progesterone marquee. Cette observation merite d'etre rapproch^e des travaux poursuivis depuis plusieurs ann^es au College de France dans le laboratoire de M. R. Courrier. En 1944, F. Joliot, R. Courrier, A. Horeau et P. Siie (C. R. Acad. Sci., Paris, 1944, 218, 769) apportent la premiere observation de la penetration dans I'hypophyse d'une hormone dont la secretion est controlee par cette glande. II s'agissait de thyroxine marquee par I'^^il (A. Horeau et P. Siie, 1945, Bull. Soc. Chim. biol., Paris, 27, 483). Depuis M. R. Courrier a confirmee et etendu ces resultats. En 1949 (R. Courrier, A. Horeau, M. Marois, F. Morel, 1949, C. R. Soc. Biol., Paris, 143, 935; et R. Courrier, A. Horeau, J. Jaques, M. Marois, F. Morel et P. Siie, 1949, C. R. Acad. Sci., Paris, 229, 275) nous avons utilise des doses physiologiques de thyroxine et tenu compte du volume des espaces extracellulaires de I'hypophyse. Ce volume fut evalu^ a I'aide du Na radioactif. Nous avons pu alors determiner la quantite d'hormone qui penetre dans les cellules: elle est de I'ordre de quelques milliemes de gamma dans une hypophyse de lapin, deux heures apres I'injection intraveineuse de thyroxine. Cette quantite augmente avec la dose administree. II semble se produire une saturation de la glande a partir de 250 y/kg. *D. L. Williams, A. R. Ronzio. Unclassified publication, AECU-714 LADC-750: Micro-syntheses with tracer elements, XVI. The synthesis of hexoestrol labelled with tritium. Metabolism of i*C-Labelled Steroids 13 En 1951 (R. Courrier, A. Horeau, M. Marois et F. Morel, 1951, C. R. Acad. Sci., Paris, 232, 776) une thyroxine de tres forte radioactivite sp^cifique (30 millicuries sur 12-5 mg. de thyroxine; 70 coups/minute par TTTT^Trp: de gamma au moment de I'experience) fut injeetee par voie veineuse a diverses especes animates. Les animaux traites furent sacrifies deux heures apres et Ton preleva chez eux sang, plasma, ante- hypophyse, post-hypophyse et differents organes. Les resultats peuvent se resumer ainsi: (1) La penetration de la thyroxine dans I'hypophyse varie selon les especes. (2) Chez les animaux oil le phenomene est decelable, e'est dans Yhypophyse posterieure que la thyroxine penetre davantage. II reste a determiner la signification physiologique de ce phenomene. II serait interessant d'etablir si la progesterone et I'acetate de desoxy- corticosterone radioactifs retrouves par M. Leblond dans I'hypophyse totale, se concentrent dans la posthypophyse. Heidelberger: In 1947 Jones and I did some work on the carcinogenic hydrocarbon [9:10-^*C] dibenzanthracene, and the results of our studies on excretion were very similar to those of Dr. Leblond for the steroids. I wonder whether this might be a general phenomenon, based not upon biological properties but largely upon the solubility of the compound. It is a large water-insoluble molecule. The difference in excretion pattern you observe with the deoxycorticosterone might be due primarily to its greater water-solubility. I would also like to say that we have used paper chromatographic methods for separating derivatives of carcinogenic hydrocarbons, based on a modification of Zaffaroni's procedure. Leblond: That water-insoluble molecules are more likely to be excreted in the bile, and water-soluble ones in the urine is true. Thus, thyroxine is excreted in the bile and diiodotyrosine in the urine. However, it is only part of the story. For instance, the water-soluble iodine ion is excreted to some extent in the bile. HIGH CHOLESTEROL CONTENT OF HUMAN SPLEEN KARL BERNHARD The lipid diseases (lipoidoses) are generally considered to he due to disorders of the lipid metabolism of the whole organism (Pick) in which storage of lipids occurs in different organs. There may be involvement of bone marrow, skin, mucous membranes, viscera and other organs or tissues. One may imagine that the cells involved produce the excess lipids themselves. To approach this problem we investigated some time ago a case of lipoidosis. The 25-year-old female patient showed a large splenic tumour extending to the iliac crest. The tumour proved to have a smooth surface and to be sensitive to pressure. The family history revealed that three brothers and sisters who died during childhood all showed splenic enlargement, as does also a 36-year-old living sister. As splenectomy was considered necessary in our patient's case, we thought it possible to investigate certain aspects of the metabolism of the spleen by injecting DgO, so as to yield a concentration of heavy water in the body fluid of the patient. We injected 25 ml. of DgO (99 -7 per cent) twice with a short interval, and also had the patient drink over a period of four days 8 pints (4 litres) of water containing 5 atom per cent deuterium. Splenectomy was performed on the fifth day of the experi- ment. Unfortunately the patient's condition did not allow us to take samples for the isolation of blood lipids. The body fluid then contained 0-56 atom per cent D. The spleen weighed 715 g. and consisted of 29-9 per cent dry matter and 1-49 per cent ash. The fresh material was reduced to small pieces, dried with acetone and then thoroughly extracted with ether. This extraction yielded 54-9 g. of ether-soluble 14 High Cholesterol Content of Human Spleen 15 material with a phosphorus content of 1 • 9 per cent and a nitrogen content of 2 • 97 per cent. The ether-soluble fraction represented 25 • 6 per cent of the dry weight. There followed an extraction with warm methanol and chloroform. This procedure yielded 44-4 g. of lipids, or 20-7 per cent of the dry weight. The lipids resulting from the ether extraction were separated in the usual way and showed the following composition: fats and unsaponifiable matter, 25-6 g.; phospholipids, 21-0 g.; protagon, 8-3 g. The fats and unsaponifiable matter together amounted to 11-9 per cent of the dry residue. The iodine number was 93 • 0, the acid index 34 • 4 and the saponi- fication value 58-3. By shaking the ethereal solution with alkali we obtained 1-32 g. of free fatty acids, and by saponi- fication 19-8 g. (=9-6 per cent) of cholesterol and 2-58 g. of fatty acids from neutral fat. The cholesterol after crystalliza- tion had a melting point of 148°. The estimation of deuterium in this material showed • 04 atom per cent D or a D-value of 7. The D-content of the fatty acids was somewhat lower (0-03 per cent). The high content of cholesterol in the spleen, amounting to 9 • 6 per cent of the dry residue, is striking. In normal spleens we have found that cholesterol represents, for example, • 84 and 1 • 05 per cent of the dry residue. Other authors have obtained similar results. However, Bloom and Kern (1926) found cholesterol in the spleen of a patient suffering from Xiemann-Pick's disease to be 8-31 per cent of the dry residue. Other values mentioned in the literature for abnormal spleens are, however, much lower (1 to 1-9 per cent). Bloch, Borek and Rittenberg (1946) were able to demonstrate in vitro that tissues- of the spleen of normal rats were not able to form cholesterol from acetate. Our observation, that a few days after giving heavy water the cholesterol present in very high concentration in the diseased spleen already contains deuterium, indicates that under such conditions a synthesis of cholesterol takes place in the spleen. The storage of this steroid would appear to result from an increased synthetic capacity of the diseased organ. 16 Karl Bernhard The working up of the remaining hpid fractions yielded large amounts of sphingomyelin and lecithin. The above observations point to the importance of deu- terium as a tool in investigating clinical problems, especially in relation to lipoidosis. REFERENCES Block, K., Borek, E., and Rittenberg, D. (1946). J. biol. Chem., 162, 441. Bloom, W., and Kern, Ruth (1926). Arch, intern. Med., 39, 456. DISCUSSION Rittenberg: If I understand Dr. Bernhard correctly his half-time for the steroid in this particular organ was 8 days? That is what I observed for serum cholesterol in the human, and I wonder whether the identity of the two figures may not indicate a mobile equilibrium between the spleen and the circulating plasma cholesterol? Bernhard: Unfortunately we could not get enough blood from the patient for the isolation of plasma cholesterol. After all, the cholesterol content of the spleen was 9 per cent under these conditions. I think there was certainly much more cholesterol in the spleen than in the blood. I believe that there really is a synthesis in the spleen and not a transfer of cholesterol from the blood to the tumour. Block: Some years ago Dr. Rittenberg and I did some experiments, which were never published, in which the half-life time of various organs of the rat was studied by administration of deuterio-acetate. A very high deuterium concentration was found in the cholesterol of the spleen. The values for the spleen were very similar to those in liver and in plasma, so that Rittenberg's suggestion of a rapid equilibrium between plasma and spleen cholesterol sounds very reasonable. THE BIOSYNTHESIS OF RADIOACTIVE CHOLESTEROL BY SURVIVING LIVER SLICES SAMUEL GURIN and ROSCOE 0. BRADY It is now established that surviving shces of rat Uver are able to utilize carbon atoms of acetate (Block, Borek and Rittenberg, 1946), acetone (Borek and Rittenberg, 1949), pyruvate (Bloch, 1948; Brady and Gurin, 1950a), butyrate, hexanoate and octanoate (Brady and Gurin, 1950a), as well as the isopropyl fragments of isovalerate (Brady and Gurin, 1951) for the biosynthesis of cholesterol. Although the results obtained with ^^C or ^^C may be regarded as a reflection of incorporation into, rather than net synthesis of cholesterol, nevertheless, the information so derived is of considerable value in attempting to understand the mechanism of this biosynthesis. Until recently, it has been believed that only those pre- cursors which are capable of being degraded to acetate or two-carbon fragments are utilized by mammalian tissue for the biosynthesis of cholesterol. Since the initial demon- stration by Bloch, Borek and Rittenberg (1946) that liver slices can transform labelled acetate into cholesterol, abundant confirmatory evidence has been reported from several laboratories. Little and Bloch (1950) have studied the incorporation of [1-^*C] acetate, [2-^*0] acetate, as well as acetate labelled with ^^C and ^^C into cholesterol. The methyl carbon was shown to be incorporated into positions 18, 19, 26, 27 and probably 17, while carbon atoms 25 and probably 10 were derived from the carboxyl of acetate. The ratio of CH3 to COOK carbons incorporated into cholesterol was found to be 1 • 27. It is apparent that some decarboxylation occurs during the process. More carboxyl groups of acetate are incorporated into the nucleus than into the iso-octyl side ISOTOPES 17 3 18 Samuel Gurin and Roscoe O. Brady chain of cholesterol. The ratio of CH3/COOH groups incor- porated into the nucleus was calculated to be 1 • 1 while the corresponding value for the side chain was 1-67. Carboxyl-labelled short chain fatty acids are also utilized by liver slices for the biosynthesis of cholesterol (Table I). This presumably occurs by preliminary fragmentation into two-carbon units since there is abundant evidence that fatty acids are converted to acetoacetate by way of two-carbon Table I Biosynthesis of Cholesterol from Short Chain Fatty Acids Substrate Radioactivity administered Radioactivity recovered in cholesterol Cpm per mg.C. Total Cpm per mg. C. Per cent i^CHgCOOXa . . . CH3(CH2)2i4COONa . CH3(CH,)4i4COOXa . CH3(CH2)6"COONa . 6,000 2,000 3,500 1,450 10,500 11,000 26,000 5,800 246 49 58 84 2-7 1-3 1-2 2-5 fragments. Evidence has also been obtained suggesting that short chain fatty acids are to a considerable extent degraded into two-carbon units prior to their incorporation into long chain fatty acids (Brady and Gurin, 1950a). We have confirmed the finding that the methyl groups of acetate contribute 1 • 3 times as much isotope to cholesterol as do the carboxyl groups (Table II). It will be observed that acetaldehyde is somewhat more efficiently utilized by liver slices than is acetate. We have also found this to be true for the biosynthesis of long chain fatty acids by surviving liver slices (Brady and Gurin, 1951). Under similar conditions, three acetoacetate preparations labelled with ^^C in the carboxyl, carbonyl, and methyl - methylene positions respectively, were found to be incor- porated into cholesterol (Table III). In each case aliquots of the same liver were incubated with suitably labelled acetate as controls. Approximately three times as much acetate is Biosynthesis of Cholesterol 19 utilized for cholesterol synthesis as is acetoacetate. Whether this result is simply an expression of differences of rates of diffusion into the cell is uncertain. It is clear, however, that ace- toacetate is utilized for this biosynthesis without preliminary Table II Biosynthesis of Cholesterol by Liver Slices No* Substrate Radioactivity administered Radioactivity recovered as cholesterol Substrate incorpor- ated fiM-f C.P.M.I Total C.P.M.I Per 1 mg. C. counts mg. C. cent "CHgCOOXa 3,150 5,500 323 7-4 8-6 CHgi^COOXa 2,600 4.600 204 5-6 6-5 2 i^CHgCOOXa 3,150 5,500 333 7-6 8-8 CHgi^COOXa 2,600 4,600 207 5-7 6-6 3 "CH3I4CHO 6,500 21,000 2260 14 26 i^CHg^COOXa 9» »» 1560 9-7 17 4 i^CHgi^CHO >> j> 2500 12 28 i^CHgi^COOXa »» >5 1790 8-5 19 CHgl^COCHg 6,400 20,000 165 0-9 1-5 CHgi^COOXa 5,000 8,800 196 2-5 2-9 *A11 the experiments designated by the same number were performed on aliquots of pooled liver slices. tCalculated for 10 mg. of cholesterol digitonide. Table III Biosynthesis of Cholesterol by Liver Slices No* Substrate Radioactivity Administered Radioactivity Recovered as Cholesterol Substrate Incorporated^ cpm/mg. C Total Counts cpm/mg. C Per cent 1 2 3 CHgi^COCHgCOOXa CHgi^COOXa i^CHgCQi^CHjCOOXa "CHgCOOXa CHgCOCHgi^coOXa CHgi^COOXa 2,600 5,000 580 4,800 1,550 6,000 12,400 8,800 1,260 8,500 3,360 10,500 83 208 59 890 37 246 1-8 5-6 9-7 17 1-3 2-7 1-4 3-5 5-2 15-6 10 31 *A11 the experiments designated by the same number were performed on aliquots of pooled liver slices. tCalculated for 10 mg. of cholesterol digitonide. 20 Samuel Gurin and Roscoe O. Brady cleavage into two-carbon fragments since it has been pre- viously established by Buchanan, Sakami and Gurin (1947) that singly labelled acetoacetate is not significantly rando- mized by liver slices. Furthermore, labelled acetoacetate is neither incorporated into long chain fatty acids by liver slices nor oxidized to any significant extent by this tissue. If labelled acetoacetate is incubated under comparable con- ditions with carrier non-labelled acetate, a small amount of ^*C can be recovered in the isolated acetate. This recovery of ^*C is much too small, however, to account for the incor- poration of acetoacetate into cholesterol. Although acetate may furnish most of the carbons of cholesterol, some of it is incorporated by way of acetoacetate. It has been shown by Zabin and Bloch (1950) as well as Price and Rittenberg (1950) that [2-i^C] acetone is converted by the rat to cholesterol and to two-carbon fragments. We have recently confirmed the fact that liver slices not only are capable of converting labelled acetone to cholesterol, but can convert it to acetate. If carrier non-labelled acetate is incubated with [2-^*C] acetone in the presence of liver slices, acetate containing appreciable radioactivity is recovered. If, from the same experiment, one examines the ^*C activity of the isolated long chain fatty acids, it is clear that much of the ^*C of the fatty acids could not have been derived from the acetate. It is reasonable to assume, therefore, that acetone may be degraded to a metabolically active two-carbon fragment which is secondarily converted to acetate. Since acetone is a relatively poor source of acetoacetate via fixation of COg, it is unlikely that very significant amounts of acetone are converted to cholesterol by way of acetoacetate. Zabin and Bloch (1950) reported that, in the whole animal, methyl labelled isovalerate is a better precursor of cholesterol than is acetate. Upon incubation with liver slices, methyl- labelled isovalerate was somewhat poorer than acetate in this respect. The reasons for this discrepancy are not clear since isovalerate appears to diffuse readily into the cell. Since the isopropyl moiety of isovalerate or leucine can react rapidly Biosynthesis of Cholesterol 21 with CO2 to form acetoacetate (Coon, 1950) and can, in addition, be cleaved to two-carbon fragments, it is not sur- prising that the intact animal can readily utilize the terminal carbon atoms of iso valerate for the biosynthesis of cholesterol. Preliminary experiments with labelled vinylacetic acid indicate a limited incorporation, while aldol is utilized about as well as acetoacetate. In contrast to aldol, ^*C labelled crotonic aldehyde contributes no carbon atoms to cholesterol under these experimental conditions. Liver slices are more- over unable to utilize labelled formate or formaldehyde for this biosynthesis (Table IV). Table IV Biosynthesis of Cholesterol by Liver Slices Exp. No. 1 2 3 4 Substrate CHgi^COONa CHgi^COOXa 14CH3I4CHOH14CH2CHO CHgi^COONa H^^COOXa CHgi^COONa Radioactivity Administered cpmlmg. C Total Counts 3000 8700 6000 11000 2000 2700 9000 16900 3750 8200 9000 16900 12000 9600 15000 42000 9000 16900 Radioactivity Recovered as Cholesterol cpmlmg. C* 20, 29 303, 265 33, 795, 450 178 570 30, 950 Substrate Incorporate 0-3, 4-2, 3- 0-7, 3, 4' 3 0-4, 8-8 ♦Each pair of results represents a separate experiment with an acetate control. For the time being, any theories concerning the mechanism of biosynthesis of cholesterol must be based upon the utiliza- tion of acetate or acetaldehyde, acetoacetate, and metabolites which can be derived from these substances. Although it is attractive to speculate about three-carbon or one-carbon precursors there is no supporting evidence in their favour. Pyruvate is poorly utilized in the in vitro system while acetone apparently goes through a two-carbon intermediate. The ready incorporation of the isopropyl fragment of isovalerate 22 Samuel Gurin and Roscoe O. Brady can be explained on the basis of its ready conversion to aceto- acetate as well as to acetate. It is of interest that incubation of liver slices with labelled acetate in the presence of non- radioactive isovalerate fails to stimulate the incorporation of ^*C acetate into cholesterol. If the synthesis of isopropyl fragments is required for this biosynthesis, it is not a limiting- reaction. Although hepatic biosynthesis of long chain fatty acids from two-carbon fragments or other precursors is practically ■T-l CII^(CH2)2_5C00E CH COUH CH,CHO CH COCH CH COCOOH in Runo'^-'Uun ^ Blocked ^ Diabetes -CH2-5o- -f .^ QO*^ > ^AcNHR CH 80CH 50( ' 4' 'MT ^ JVCH^CHIIH^COGH C CH ) ^CHCH^CHNH^C OCH jWK < CHOLESTEROL Fig. 1. Metabolic interrelations of lipids in liver. abolished in alloxanized rats and depancreatized cats, liver slices obtained from such animals retain their ability to syn- thesize cholesterol at a normal or perhaps accelerated rate (Brady and Gurin, 19506). It was suggested, at the same time, that the synthesis of fatty acids in the liver proceeds by a process independent of fatty acid breakdown (Fig. 1). In the diabetic state, the synthetic process is blocked, hence the accumulation of fat in such livers cannot be ascribed to hepatic synthesis. Since surviving liver tissue obtained from Houssay cats (Brady, Lukens and Gurin, 1951a) is capable of incorporating labelled acetate into long chain fatty acids Biosynthesis of Cholesterol 23 at a normal rate, it has become clear that the pituitary secretes a principle which is capable of inhibiting hepatic synthesis of fat. Preliminary results indicate that purified growth hormone is at least one of the factors responsible for this inhibition (Brady, Lukens and Gurin, 19516). Cortisone is also effective in this respect. Since the diabetic liver is characterized by an accelerated oxidation of fattv acids to two-carbon fragments, which can no longer be reconverted to fatty acids, these fragments must accumulate or be diverted into other metabolic pathways. In as much as both acetoacetate and two-carbon fragments are utilized in the biosynthesis of cholesterol, it is perhaps not surprising that hypercholesterolaemia is so prominent a feature of the diabetic state. Although it is generally believed that acetoacetate is primarily synthesized in liver (and to a minor extent in kidney), the evidence that acetoacetate may be required for the biosynthesis of cholesterol implies that all tissues capable of synthesizing cholesterol under in vitro conditions must have an obligatory synthetic mechanism for the production of acetoacetate. We have recently found this to be true for testicular (rabbit) and adrenal (human) tissue. Incubation of either of these tissues with radioactive acetate in the presence of carrier non-radioactive acetoacetate results in the formation of acetoacetate containing a significant proportion of the administered ^^C. It is of interest that Jowett and Quastel (1935) were able to demonstrate a net synthesis of acetoacetate in testis. Whether this will prove to be true for all other tissues capable of synthesizing cholesterol must await further experimentation. REFERENCES Bloch, K. (1948). Cold Spr. Harb. Sym. quant. Biol., 13, 29. Block, K., Borek, E., and Rittenberg, D. (1946). J. biol. Chem.y 162, 441. Borek, E., and Rittenberg, D. (1949). J. biol. Chem., 179, 843. Brady, R. O., and Gurin, S. (1950a). J. biol. Chem., 186, 461. Brady, R. O., and Gurin, S. (19506). J. biol. Chem., 187, 589. 24 Samuel Gurin and Roscoe O. Brady Brady, R. O., and Gurin, S. (1951). J. biol. Chem., (in press). Brady, R. O., Lukens, F. D. W., and Gurin, S. (1951a). Science (in press). Brady, R. O., Lukens, F. D. W., and Gurin, S. (19516). Unpublished. Buchanan, J. M., Sakami, W., and Gurin, S. (1947). J. biol. Chem., 169, 411. Coon, M. J. (1950). J. biol. Chem., 187, 71. JowETT, M., and Quastel, J. H. (1935). Biochem. J., 29, 2181. Little, H. N., and Block, K. (1950). J. biol. Chem., 183, 33. Price, T. D., and Rittenberg, D. (1950). J. biol. Chem., 185, 449. Zabin, I., and Block, K. (1950). J. biol. Chem., 185, 131. DISCUSSION Bloch: We have come to the same conclusion as Dr. Gurin, namely that acetoacetate is an intermediate in the conversion of acetate to cholesterol. Dr. Zabin and I have carried out some experiments with butyrate and isovalerate and have in both cases been able to ascribe the conversion of these compounds to cholesterol to the intermediary formation of acetoacetate. I would like to present also some highly speculative material on the mechanism of cholesterol biosynthesis. The evidence that indeed all the carbon atoms of the sterols can be derived from acetic acid is now quite satisfactory. In considering various schemes for the condensation of acetate units, one paper came to our attention which dealt with the biosynthesis of an isoprenoid compound, namely natural rubber. Bonner and Arreguin incubated isolated guayule leaves with various substrates and found that acetate, acetone, acetoacetate, and /S-dimethyl acrylate were potent carbon sources for natural rubber. On the basis of these results Bonner and Arreguin postulated the following mechanism for the formation of an isoprene unit from acetate: condensation of 2 mols of acetate to acetoacetate, decarboxylation to acetone, which then would condense with a further molecule of acetate to j8-dimethyl acrylate, and then finally a conversion to an isoprene unit. Note that an isoprene unit if synthesized in this fashion from acetate will contain three methyl carbons of acetate and two carboxyl carbons of acetate. There have been many speculations (Channon, Heilbron, Robinson) on the possible role of the isoprenoid hydrocarbon squalene as a precursor of cholesterol. Last year, as a result of some discussions with Dr. Gallagher, Dr. Langdon and I tried again to fit the squalene hypothesis into the existing knowledge on the biosynthesis of cholesterol. This polyisoprenoid hydrocarbon can be arranged as shown in Fig. 1, and a structure is obtained which, by the formation of various cross-linkages, could form the polycyclic steroid structure. Providing that squalene is formed as postulated, namely by the condensation of acetate to acetoacetate, acetone, and so forth, then one should obtain a squalene molecule with the isotope distribution shown in Fig. 1, the circles indi- cating methyl carbons and the crosses carboxyl carbons of acetate. In Biosynthesis of Cholesterol 25 the same figure is given the distribution of methyl and earboxyl carbons of acetate in cholesterol, as it has been found by direct isolation. In all cases the observed isotope distribution in cholesterol and that postulated for squalene coincide. Squalene contains six isoprene units and each isoprene unit will contain 3 methyl carbons and 2 earboxyl carbons of acetate, if syn- c o •C o O: ACETATE CH3 X-. ACETATE COOH CHOl£STEROL Cz? H46 B y o c o C c'" ^c^ C o r^o^ I c ^ o SQUALENE C30 ^50 Fig. 1. A. Distribution of acetate carbon found inlcholesterol. B. Postulated distributionTof acetate carbon in squalene. thesized by the mechanism postulated before. Squalene will therefore contain six times as many methyl carbons and six times as many earboxyl carbons, i.e. methyl carbons and earboxyl carbons of acetate will be present in a ratio of 18:12. In the hypothetical conversion of squalene to cholesterol, 3 branched carbon atoms must be lost, and the ratio of methyl to earboxyl carbon atoms would, by the loss of 3 26 Samuel Gurin and Roscoe O. Brady methyl carbons of acetate, change from 18:12 to 15:12. We have shown by the use of doubly-labelled acetate ^^CHg^^coOH, that the ratio of these two carbon atoms in cholesterol is equal to 1-25, or 15:12. It appears to us that a mechanism which involves the forma- tion of cross-linkages in a folded chain would, on the basis of present evidence, not be unreasonable. I may further mention that squalene has recently been shown to be much more widely distributed than was originally assumed. It has now been isolated from human sebum, and therefore the possibility that the function of this hydrocarbon is that of a precursor for the steroids would appear to be worth considering. Dauben: It is very interesting to hear about the role of squalene in cholesterol synthesis. Dr. Chaikoff and I in Berkeley have arrived at the same conclusion from experiments using dimethylacrylic acid in the slice and in the intact animal. We have gone one step farther than Dr. Bloch. We have synthesized radioactive squalene and now have it in the animal to find out if the specifically labelled squalene will come out with the right distribution of marked carbons. For example, if you label it properly and it folds around you should either get 4-C or 6-C labelled cholesterol. Rittenberg: We have been interested in problems quite similar to those Dr. Gurin has just reported on. We have attempted to investigate the role of a more highly reduced Cg unit than acetate, and have chosen ethyl alcohol as a possible substrate (Curran, G. L., and Rittenberg, D., 1951, J. biol. Chem., 190, 17). We tried ethyl alcohol because we thought it was a much more pleasant compound to incubate with a tissue slice than acetaldehyde. We wanted to know whether acetalde- hyde is utilized in cholesterol synthesis by some direct route or whether it is first converted to acetate. Of course, such problems cannot be attacked with ^^C labelling and instead we labelled the a carbon of ethyl alcohol with deuterium. We found that while the carbon atoms are utilized, neither in the in vitro nor the in vivo systems does the deuterium of ethyl alcohol yield deuterio cholesterol. We therefore consider it unlikely that ethyl alcohol or acetaldehyde can be inter- mediates. Curran in my laboratory has been working with carbonyl labelled acetoacetate and he finds results similar to those reported by Dr. Gurin, that is, that acetoacetate is utilized for cholesterol formation. However, by a procedure very similar to that which Dr. Gurin reports he finds a very considerably higher conversion of acetoacetate to acetate, of the order of 5 per cent conversion. Of course this raises the question whether the utilization goes by way of acetate, because a 5 per cent conversion is more than adequate to supply all the Cg units required for cholesterol synthesis. He has therefore carried out a dilution experiment to test whether acetate is the intermediate, by incubating i*C-labelled acetone and normal acetate. For comparison, let us consider a similar experi- ment using carbonyl labelled acetone, which goes very effectively to acetate. If you incubate one equivalent of acetone with one equivalent of normal acetate you dilute the ^^C activity of the cholesterol by a factor of about 10 or 20, indicating very strongly, as Dr. Bloch's work Biosynthesis of Cholesterol 27 and my own have shown some time ago, that acetone is metabolized via acetate. If, however, you do the same experiment \Wth carbonyl labelled acetoacetate the dilution is only about 2. Curran is inclined to believe that the utilization of acetoacetate goes by two pathways: one through acetate, and another pathway through acetoacetate. This raises the possibility that the biosynthesis of cholesterol goes by a mechanism in which the acetoacetate is used to form the beginning of the molecule, and then repeated additions of acetate complete the synthesis. STUDIES WITH DEUTERIUM STEROIDS T. F. GALLAGHER The metabolic studies have been a joint project with my colleague, Dr. Konrad Dobriner. Infra-red spectrometry has contributed greatly to the strictly chemical studies and I am indebted to Dr. Dobriner for his help in this phase of the programme. The synthesis of the isotopically labelled steroids was the result of the collaboration of my co-workers, Dr. David K. Fukushima, Dr. Theodore Kritchevsky and Dr. Bernard Koechlin in the Division of Steroid Biochemistry. The deuterium analyses were carried out by Mr. Robert W. Jailer and the tritium studies were done with the co-operation of Dr. Max Eidinoff. The primary aim of our group has been to investigate the steroid metabolism of man. For this purpose, we have pre- pared and studied steroid hormones containing both stable and radioactive isotopes. Since we wished to study normal as well as diseased subjects, much of our work has been done with steroids containing deuterium in stable positions in the molecule. Radioactive hormones have been used in patients with limited life expectancy but as yet their use in normal subjects has not been permitted. The deuterium steroids, on the other hand, offer no radiation hazard; they can be prepared relatively simply from a variety of substances closely related to the hormones, and qualitative detection of the isotope by infra-red spectrometry is relatively easy and is accomplished without loss of material. Certain of the chemical investigations have led to interesting results quite apart from their biological implications. The present report will be restricted to studies with the isotopes of hydrogen and will be presented as a survey of some problems which have engaged the attention of our group, rather than a report of an integrated investigation. 28 Studies with Deuterium Steroids 29 It is already clear from the many studies of Rittenberg and his associates that deuterium is an excellent tracer in bio- chemical studies and this has been confirmed in our investiga- tions. There is little basis for the view still held by some workers that the chemically stable hydrogen atoms of organic compounds exchange easily with the medium or participate in obscure reactions which lead to loss of hydrogen without essential alteration of the molecule. On the contrary, it is important that even when a chemically labile deuterium atom is present in an intermediate, this isotope is not neces- sarily lost in a metabolic transformation and, indeed, a high percentage is frequently retained in the excretory end product. The use of this element as a tracer is therefore fully justified and in many instances can furnish information unobtainable with other tracers. The hormones, moreover, do not suffer extensive dilution from either endogenous production or from the diet, and thus in many instances it is not obligatory to work with extremely low concentrations of the isotope^^ Finally, there has accumulated a considerable body of know- ledge of the metabolism and fate of some of the important steroid hormones and this has served as a guide and control of these initial studies (Dobriner and Lieberman, 1950). Many deuterium-labelled steroids can be prepared by reduction of an unsaturated compound with the hydrogen isotope, and a properly chosen intermediate can furnish a number of compounds for metabolic study. By hydro- genating lithochol-11-enic acid in solution in acetic acid-c? with deuterium gas in the presence of a platinum catalyst, the saturated acid with 4 • 03 atoms per cent excess deuterium was obtained. This was converted to pregnanolone by known methods and the pregnanolone served as a good intermediate for the preparation of the steroid hormones and metabolites shown in Fig. 1. The products served nicely for tracer experiments since they contained a useful concentration of isotope and their chemical preparation was relatively simple. We sought, however, a more generally applicable reaction which could be used with substances closely related chemically 30 T. F. Gallagher to the hormones we wished to prepare. The platinum- catalysed exchange reaction first studied in sterols by Bloch and Rittenberg seemed to offer the means for this end. It was found that, with suitable compounds, the incorporation of isotope was satisfactory and the yield was good. The best yields were obtained with unsaturated ketones of the type exemplified by androst-4-ene-3:17-dione. This compound upon heating in 70 per cent deuterated acetic acid and OH dz- !I,I2 TESTOSTERONE CH3 I c=o o?^ CH. I c=o ^.^ ,^^.AJ d2- 11,12 PROGESTERONE HO CH3 I c = o dz-M,I2 ETIOCHOLANOLONE CHzOAc I c=o rs^ j^ .^:" 17* - HYDROXY- d2- 11,12 PROGESTERONE d2- 11,12 Substonce "s" ocetote Fig. 1. Steroid hormones and metabolites prepared from d2-ll,12-pregnanolone. deuterium oxide in the presence of an active catalyst was recovered in over 90 per cent yield with approximately 2 • 5 atoms per cent excess deuterium. Similar results were obtained with other ketones. The reaction thus constitutes an excellent preparative procedure for certain compounds closely related to the steroid hormones. The synthesis of deuterium testosterone from androst-4-ene-3:17-dione labelled by exchange is shown in Fig. 2. The results obtained in the catalysed exchange with cestrone are illustrated in Fig. 3. The incorporation of radioactive tritium into steroid hormones Studies with Deuterium Steroids 31 -J CH3-C00D ^f^T_J KOH ^jT^JLj 1 90% Yield ETHYL ORTHOFORMATE + HCL L1AIH4 d- TESTOSTERONE 0.79 gm. oloms D per mole Fig. 2. Synthesis of d-testosterone from androst-4-ene-3:17- dione. II Pt. 150° DgO- CD3COOD 05N KOH, 'E Hr. Reflux 5.73 D 2N HCI^ 6 Hr. Reflux 4.28 D 3.20 D /m.p. 257-259° ["^In ' "•" ISKChloroform) 6 2810* 2050 (Ethanol) < Acefatej 3.25 m.p. 126 5-127.5° I [«.] = + l36(Chloroform) Ve2680,2750 = '''*5 (Ethanol) Fig. 3. Preparation of d-oestrone by the platinum catalysed exchange reaction. 32 T. F. Gallagher by the exchange reaction is Hkewise an extremely useful procedure as illustrated by the preparation of radioactive cortisone acetate illustrated in Figs. 4 and 5. The amount of isotope incorporated by the exchange CH-7 c = o CH3 c = o AcO- 0. 2 DAY 150" Pr 70% CH3C00H*-H^0 2 HOURS REFLUX WITH KOH HO- 1.02X10'" DiSINT./min/m.m Fig. 4. Preparation of radioactive 3.-hydrox\pregnane-ll:20- dione by the platinum catalysed exchange reaction. HO' X CH- KOH HO' 3 STEPS 3 90X10^ DiSINT/min./mm C0RTIS0^JE ACETATE Egjeo^'^-'OO m p 245-246° U]d= +I86°(ACET0NE) , oo y ,^9 h.cmt / / [/]^+2ie-(CHCL3) 2 82X10 D,SINT/mm/m, Fig. 5. Preparation of radioactive cortisone acetate. reaction, while useful for the biological studies, was not as high as we had anticipated from the results of Bloch and Rittenberg (1943) with cholesterol. These authors had found that cholesterol stably labelled with deuterium by the Studies with Deuterium Steroids 33 exchange reaction had an approximately equal distribution of the isotope in the nucleus and in the side chain. We re-investigated this distribution by somewhat different ana- lytical procedures and confirmed the general conclusion of Bloch and Rittenberg. However, the isotope was more precisely localized than might have been anticipated and, in fact, was nearly exclusively situated upon three carbon atoms, C-6, C-26 and C-27. The evidence for this conclusion is CH^COO CHOLESTEROL ACETATE 2.48 D 1.34 D (1.14 D losl) COOCH-. CHjCOO'*--,-^''^:^-^ CH3COO DEHYDROISOANDROSTERONE ACETATE I.I7D (I 31 D lost) 1.19 D (1.29 D lost) 1.20 D (1.28 D lost) Fig. 6. Distribution of the isotope in cholesterol obtained by the platinum catalysed exchange reaction. presented schematically in Fig. 6. The acetate dibromide of cholesterol from the exchange reaction was oxidized with chromic acid and from the reaction a sample of 3/3-hydroxy- chol-5-enic acid was isolated. This compound contained 1 • 20 atoms of deuterium per molecule as compared with 2 • 48 atoms of deuterium per molecule in the cholesterol. The iso- tope lost, 1 • 28 atoms of deuterium per molecule, must have been attached to the i^opropyl group removed by oxidation, or to C-24, the carbon atom which is now the carboxyl group. While the possibility exists that the isotope was attached to ISOTOPES 34 T. F. Gallagher C-24, in view of the fact that there was none at C-23 or C-22 this possibihty seems unreasonable. Since there is con- siderably more than one atom it cannot all be at the tertiary carbon C-25. Thus the methyl groups are a highly probable situation and we have concluded that these must be the loci of the side chain deuterium. The remainder of the isotope (approximately 1 • 2 atoms per molecule) was found in the vicinity of the unsaturated bond and secondary alcohol group as shown by the following experiments. Oxidation of cholesterol to cholestenone and equilibration with base in aqueous alcohol resulted in loss of 1-14 atomiS of deuterium per molecule (Fig. 6). This isotope lost must have been on carbons 2, 3, 4 or 6, since only hydro- gen from these positions would be lost in the oxidation or exchanged with the medium after oxidation upon treatment with base. Thus more than 97 per cent of the isotope must have been on carbons 2, 3, 4, 6, 26 and 27. Further localization of the ring isotope was possible. Another sample of cholesterol with a different isotopic content, but similar distribution, was reduced with ordinary hydrogen in the usual way and the saturated cholestanol isolated. Oxidation to cholestanone caused a minor loss and this must be ascribed to a small amount of deuterium attached to C-3. More important, however, was the finding that refluxing with base in aqueous alcohol did not result in any further loss of isotope. This clearly proved that in the deuterated choles- terol there was no isotope at either C-2 or C-4, and since there was no hydrogen at C-5 the location of the deuterium in the ring system of the cholesterol obtained from the exchange reaction was in small measure at C-3, and in preponderant amount at C-6. We thus find that the platinum-catalysed exchange reaction yielded a deuterated cholesterol in which practically all of the isotope was on one ring carbon and in the i^opropyl group at the end of the side chain. The results clearly suggest that the localization of isotope will be different with different steroids and that a random distribution in an unsymmetrical molecule is unlikely. The Studies with Deuterium Steroids 35 localization of the isotope proves that the union of catalyst and the organic molecule is a highly specific one dictated by the various features of the molecule. More important than this for biochemical purposes, the studies show that isotope in a chemically labile position may still contribute very significant information, since Bloch, using cholesterol from the exchange reaction, was able to demonstrate the conversion of cholesterol to pregnanediol (Bloch, 1945). Since almost surely this transformation was accomplished through an intermediate such as progesterone or a similar a/S-unsaturated ketone, the chemically labile deuterium of the ring system was not lost and was in fact retained to a very considerable extent. Recal- culation of the data of Bloch, using the isotope distribution we have found, indicates that 80 per cent of the urinary pregnanediol was derived from the serum cholesterol. The isotopically-labelled steroid hormones permit study of the following points which could not be discovered by alternative methods: (1) It is possible to distinguish clearly between the endogenous glandular production and the administered hormone; (2) It is possible to detect transforma- tions that take place in such limited amount as to be undetect- able by ordinary means; (3) It is possible to set precise quantitative limits on the extent to which a particular trans- formation takes place or conversely to demonstrate the absence of such a transformation. These can be well illus- trated from a specific experiment. A normal man received 100 mg. of labelled testosterone intramuscularly in a single injection in solution in sesame oil containing a small amount of benzyl alcohol. Urine was collected for twenty-four hours after the injection. The urinary steroids were isolated by the procedure developed and standardized by Dobriner (Dobriner, Lieberman and Rhoads, 1948; Lieberman, Dobriner, Hill, Fieser and Rhoads, 1948). The a-ketonic fraction was chromatographed and the various fractions from the chromatogram were examined for the presence of deuterium by infra-red spectrometry (Dobriner, Kritchevsky, Fukushima, Lieberman, Gallagher, Hardy, 36 T. F. Gallagher Jones and Cilento, 1949). This permitted the quaUtative detection of the deuterium steroid metaboUtes without loss of material. It was found that five compounds contained appreciable quantities of the isotope; in the order of elution from the column these were androst-2-en-17-one, androst-5- ene-3:17-dione, 8etiocholane-3:17-dione, androsterone and setiocholanolone. It was known from prior studies that androsterone and aetiocholanolone constituted the principal end products from the quantitative standpoint and these substances were therefore examined in detail. Several eluates contained essentially pure androsterone, using the infra-red spectrum as a guide to purity. Several other eluates con- tained essentially pure setiocholanolone by the same criterion. The spectroscopically homogeneous androsterone fractions were combined and freed from a minor non-steroid contami- nant by sublimation in high vacuum. The weight of material obtained in the sublimate (9'0 mg.) was in precise agreement with the colorimetric determination of its 17-ketosteroid content by the method of Callow (Callow, Callow and Emmens, 1938). The amount was therefore taken as the weight of androsterone removed from the total amount excreted in the urine. Although the sample removed was almost analytically pure androsterone by the usual criteria, it was repeatedly recrystallized until its physical constants were those of the purest compound reported in the literature. A sample was then analysed and from the values obtained (2 • 06 atoms per cent excess deuterium in the androsterone, 2-82 atoms per cent excess deuterium in the injected testosterone) it was calculated that 78-4 per cent of the androsterone had been made from the testosterone administered. Similar treatment of the setiocholanolone fraction yielded a pure product con- taining 1 • 95 atoms per cent excess deuterium, from which it was calculated that 74 • 1 per cent of the excreted aitiocholano- lone had its origin in the testosterone injected. It remained therefore to determine the total amount of androsterone in the urine by application of the principle of isotopic dilution. All the fractions from the chromatogram Studies with Deuterium Steroids 37 containing less pure astiocholanolone and androsterone and mixtures of these two were combined, and known quantities of non-isotopic androsterone and aetiocholanolone were added as carriers. A second chromatographing, followed by careful purification of the products, was performed and analytically pure samples of androsterone and aetiocholanolone were obtained. Knowing the amount of carrier added and the deuterium content of the excreted androsterone and the excreted astiocholanolone, it was readily calculated that 4-2 mg. of androsterone and 12-7 mg. of aetiocholanolone had been combined. Adding these values to the weight of pure androsterone removed from the first chromatogram (9-0 mg.) and to the pure aetiocholanolone similarly removed (22-0 mg.) afforded the total excretion of these two compounds. Since the total amount and the dilution were known, both endogenous production and the quantity obtained from the administered hormone were calculated. These results are shown in Table I. Table I Urinary Excretion during 24 Hours after 100 mg. Testosterone Intramuscularly in Man 17KS = 70-5mg. a = 67-8 mg. j3 = 0-6 mg. Wt. in mg. .3itiocholanolone ... "Endogenous" ... Androsterone "Endogenous" ... Androst-2-en-17-one Androstane-3:17-dione ^tiocholane-3:17-dione ... Subtract "endogenous" products Isotopic metabolites recovered ... Total • ■ • ■ • • • • • • 25-7 90 10-3 2-9 0-78 019 Oil 48-98 11-9 37-08 The values for the quantitatively minor ketonic metabolites were obtained by isotopic dilution alone since the amounts were too sm^all to permit direct isolation. In these instances, 38 T. F. Gallagher the endogenous production can be disregarded without error since it is very small in comparison with the amount of carrier added. The values obtained are shown in Table I. Similar experiments have been performed wdth a variety of steroid hormones and their metabolites. Adrenal as well as gonadal hormones have been studied and the general applica- bility of the procedure has been established. It is of course clear that these investigations constitute merely the initial attempts and that they have raised more questions than they have answered. We believe, however, that continued study of these transformations will bring into sharper focus the hazy outlines of the changes the steroid hormones undergo during their passage through the body, and that this know- ledge will in turn lead us to a better understanding of their role in the life process. REFERENCES Block, K. (1945). J. biol. Chem., 157, 661. Bloch, K., and Rittenberg, D. (1943). J. hiol. Chem., 149, 505. Callow, N. H., Callow, R. K., and Emmens, C. W. (1938). Biochem. J., 32, 1312. DoBRiNER, K., Kritchevsky, T. H., Fukushima, D. K., Lieberman, S., Gallagher, T. F., Hardy, J. D., Jones, R. N., and Cilento, G. (1949). Science, 109, 260. Dobriner, K., Lieberman, S., and Rhoads, C. P. (1948). J. hiol. Chem., 172, 241. Dobriner, K., and Lieberman, S. (1950). In Gordon, E. S., A Sym- posium on Steroid Hormones, p. 46. Madison, Wis.: University of Wisconsin Press. Lieberman, S., Dobriner, K., Hill, B. R., Fieser, L. F., and Rhoads, C. P. (1948). J. biol. Chem., 172, 263. DISCUSSION Boscott: Have you studied the metabolism of oestrogens? Have you ooked at the phenolic acids, the steroid phenols? Gallagher: No. Dr. Dobriner has just begun some human experi- ments with the isotopically-labelled oestrogens, but these are still in the exploratory stages. Boscott: Recently I looked up the literature on the metabolism of progesterone in monkeys and chimpanzees, and I was very surprised to find that rhesus monkeys behave very differently from chimpanzees. Chimpanzees excrete pregnanediol in their urine and rhesus monkeys Studies with Deuterium Steroids 39 do not. Do you think that it is possible or Hkely that the metaboUtea occur in the faeces rather than in the urine? Gallagher: In lower animals I think it's rather likely. I am very certain that it doesn't happen with testosterone in man. We had come to that conclusion from the deuterium work, and have recently confirmed with carbon-labelled materials that no radioactivity appears in the faeces. Progesterone in the mouse or in the rat is metabolized in part at least with loss of the side chain. In other words, when you study pregnanediol alone you're looking for only a portion of the end-products of progesterone metabolism. I feel that in most hormones of the type of progesterone, testosterone, and cortisone, some metabolic process will open Ring A, but I have no positive evidence for this view at the moment. Rittenberg: As I understand it, the platinum catalysed exchange in cholesterol resulted in deuterium at the 6, 26 and 27 positions exclusively. Would you care to speculate on the mechanism responsible for this curious reaction. Gallagher: When the double bond of cholesterol w^as reduced there was a loss of deuterium from the compound, i.e. the cholestanol obtained had about • 5 atoms of deuterium per molecule less than the cholesterol. Conversely, when cholesterol is saturated with deuterium in acetic acid-d more than the expected 2 atoms of D appear in the product. This means, we feel, that when carbon atom 6 forms a complex with the catalyst, the hydrogen or deuterium at that position becomes labile and exchanges with the hydrogen or deuterium of the medium. You must have hydrogen gas to achieve this because in the absence of reduction no change or only a very minor one occurs. In the exchange reaction it is therefore readily understandable that this position would undergo exchange. Why only the terminal methyl groups in the side chain exchange hydrogen is at present unknown. Before we found these interesting facts we had thought w^e might make some generalizations about the platinum catalysed exchange reaction, and had studied a great many different compounds. We found that the more unsaturated the compounds, the more functional groups they contained, the more exchange could take place; but as we accumulated more and more evidence, it became clear that each compound behaved uniquely. Apparently, the catalyst-substance complex determines not only the amount but also the distribution of the isotope. Bloch: Do you lose all the ring deuteriums in the conversion of cholesterol to cholestenone? Gallagher: Practically speaking, yes, you lose all of it after equilibra- tion with aqueous alkali. Bloch: I am just wondering how the data which we obtained some years ago on the conversion of deuterium cholesterol to pregnanediol or to the bile acids would appear in terms of these results. We assumed at the time that several hydrogens would be lost by enolization. Such losses would occur if a A'^'^ unsaturated ketone were an intermediate. One would expect a considerable loss since you also lose deuterium from the isopropyl group of the side chain. 40 T. F. Gallagher Gallagher: The pregnanediol conversion comes out to be about 80 per cent of theory, based, of course, on the assumption that blood cholesterol is the precursor. The conversion to cholic acid is higher, of the order of 90 per cent. Bloch: This is assuming that progesterone is an intermediary stage? Gallagher: Yes. And it's also based on the assumption that you do not lose any of the deuterium at C-6. I think the metabolic trans- formation is accomplished before there is any opportunity for extensive exchange with the medium. These calculations confirm the view that deuterium is an excellent tracer element even if it is in a chemically labile position; it is not necessarily labile in the animal body. PART II HEMOGLOBIN AND METABOLIC DERIVATIVES THE BIOSYNTHETIC MECHANISM OF PORPHYRIN FORMATION DAVID SHEMIN and JONATHAN WITTENBERG This discussion will deal with the origin of each of the carbon atoms of protoporphyrin (the porphyrin moiety of haemo- globin) and the conclusions, inferences and hypotheses we have drawn from our data concerning the intermediary steps in porphyrin formation. We will also attempt to demonstrate that in the biological formation of the pyrrole unit two relatively simple molecules are involved. The biological system used for most of the experiments in this study was the nucleated red blood cells of the duck, which can synthesize the porphyrin in vitro (Shemin, London and Rittenberg, 1948, 1950). By merely incubating the blood of a duck with an isotopically labelled precursor of the por- phyrin, labelled hsem is formed. This limited system has several obvious and distinct advantages over work with the whole animal. Since many interconversions among the small molecules do not take place in this in vitro system, the data obtained are readily interpretable. Individual carbon atoms of the porphyrin molecule have been isolated by suitable degradation procedures and the following numbering system is used to designate any particular carbon atom in the porphyrin without ambiguity (Fig. 1). The pyrrole rings are designated A, B, C, D; the methene carbon atoms of the bridges a, p, y, 8 and the ring carbon atoms are 41 42 David Shemin and Jonathan Wittenberg numbered so that similar side chains occur on the same numbered ^-carbon of each pyrrole ring. Since uroporphyrin is the biological porphyrin containing the largest number of carbon atoms, it is used as the parent compound. The carbon atoms of protoporphyrin, the porphyrin found in haemoglobin, would therefore be numbered as shown in Fig. 2. It is now easy to designate a particular carbon atom; e.g. carbon atom ioCOOH 7 COOH 9CH2 r / HC6 W H I D 6 CHp ^ CHo 7C00H 9 CH2 IOCOOH H G- a :G- H 10 QOOH 7.CC0H 9 CH? &QH2 sQHa S I N H 2 N I 2 ! ^h 6 CH< 8 CH2 9 CH2 7 COOH loCOOH UROPORPHYRIN HI Fig. 1. Numbering System for Porphyrins. A4 refers to the /8-carbon atom of pyrrole ring A to which the methyl group is attached, etc. For the purpose of clarity this discussion is divided into two parts: the first part dealing with the role of glycine in porphyrin formation, and the second with the role of a four carbon unsymmetric compound which appears to arise in the tricarboxylic acid cycle. It will be shown below that glycine and this four carbon compound account for all thirty-four carbon atoms of protoporphyrin. Biosynthesis of Porphyrins 43 In 1945 it was found that glycine was the nitrogenous precursor of the porphyrin (Shemin and Rittenberg, 1945, 1946). However, it was not estabhshed from these experi- SCHj 6CH3 3CH 9CH2 6CH38CH HEM!N 5CH3 9CH2 eCH2 5CH3 3CH2 9CH2 'OCOOH lOCOOH PROTOPORPHYRIN IX. CH3 C H, METHYLETHYLMALEIMIDE COOH CH2 CH3 C H2 C*D O^^N^^O HEMATINIG ACID CO2 CC /O, 10) CH3 CHj CH3 c »i CMl__pH2 HC /^U^ H ^ • CH3 CH2 CH2 CH, COOH CCOH MESOPORPHYRIN ¥ CH3 CHj CH2 HOi ; OH .2-(3)-METHYL-3-(2) ETHYLTARTARIMIDE METHYLETHYLMALEIMIDE r CH3-C0-C00H 6 4 3 CH3-CH2-CO— COOH 9 a — 3 — 2 V. CH3-COOH 6 4 + C02 CHj-CHj-COOH 9 8 3 -t- CO2 2 CHjNH2 6 CO2 4 C02 6 CH3-COOH. 9 — a C02 9 Fig. 2. Outline of haemin degradation. The letters and numbers underneath the carbon atoms designate the positions of these atoms in protoporphyrin. *The methylethylmaleimide samples were degraded separately as outlined. 44 David Shemin and Jonathan Wittenberg merits that glycine nitrogen was utilized for the formation of all four pyrrole structures. In protoporphyrin (Fig. 2) two of the pyrrole rings (A and B) contain methyl and vinyl side chains and two (C and D) contain methyl and propionic acid side chains, and it was conceivable that these two different pyrrole structures are synthesized in the organism from different compounds. To elucidate this point ^^N labelled glycine was administered to a human and a duck (Wittenberg and Shemin, 1949), and the resulting labelled h?em was degraded in a manner such that Rings A and B were separated from Rings C and D. It was found that the ^^N concentrations of the porphyrin, pyrrole rings A and B and pyrrole rings C and D were equal (Table I) (Muir and Neuberger, 1949; Table I Distribution of ^^n in Protoporphyrin Synthesized from ^^N Labelled Glycine (Wittenberg and Shemin, 1949) The results are expressed as atom per cent excess ^^N Experiment ^^N concentrations in Porphyrin Pyrroles A and B Pyrroles C and D Duck Human ... 0-293 0113 0-292 0-112 0-295 0-113 Wittenberg and Shemin, 1949). Therefore glycine nitrogen is utilized equally well for both types of pyrrole units in protoporphyrin. This finding suggested that a common precursor pyrrole is first synthesized and that this pyrrole derivative is subsequently utilized for both types of pyrrole units found in the porphyrin. As will be seen, later work will more than support this suggestion and will elucidate the structure of the common precursor pyrrole. In this degradation and the subsequent procedures developed to isolate each carbon atom of each pair of pyrrole rings, the pyrrole units are isolated as pairs. Rings A and B Biosynthesis of Porphyrins 45 and Rings C and D, since pyrrole A is identical with B, and pyrrole C is identical with pyrrole ring D. However, the conclusions which will be drawn from the data are concerned with each of the carbon atoms of each of the pyrrole rings. This appears to be valid since, as shown below, the biosyn- thetic mechanism for the dissimilar pairs of pyrrole rings is the same and it is reasonable to expect that each pyrrole of each pair is made in the same manner. With this assumption we may then conclude that all four nitrogen atoms of the porphyrin are derived equally from the nitrogen atom of glycine. Unless the glycine nitrogen atom were involved in porphyrin formation by some sort of transamination reaction, it could be predicted that the a-carbon atom would also be utilized for porphyrin formation. Indeed it was later demonstrated that the a-carbon atom of glycine is found in the porphyrin (Altman, Casarett, Masters, Noonan and Salomon, 1948; Radin, Rittenberg and Shemin, 1950a). However, carboxyl labelled glycine did not form labelled hsem (Grinstein, Kamen and Moore, 1948; Radin et al., 1950a). Incubation of duck erythrocytes with ^^NHg^^CHgCOOH resulted in hsem containing twice as much ^*C as ^^N (Muir and Neuberger, 1950; Radin et ciL, 1950a), indicating that eight carbon atoms of the porphyrin are derived from the a-carbon atom of glycine since the four nitrogen atoms are known to be derived from this source. In order to locate the positions in the porphyrin of these eight carbon atoms, methods then had to be developed whereby one could systematically degrade the porphyrin molecule. We have utilized methods by which one can isolate unequivocally each carbon atom of each pair of pyrrole rings from known positions in the molecules (Witten- berg and Shemin, 1950; Shemin and Wittenberg, 1951). As can be seen in Fig. 2, hsemin obtained from haemoglobin was converted to protoporphyrin and the latter reduced to meso- porphyrin. The mesoporphyrin was then oxidatively cleaved with chromic acid to yield methylethylmaleimide from Rings 46 David Shemin and Jonathan Wittenberg A and B, and haematinic acid from Rings C and D. The methylethylmaleimide represents fourteen carbon atoms of the porphyrin, since it arises from both Rings A and B, and the haematinic acid represents sixteen carbon atoms since it arises from both Rings C and D, a total of thirty of the original thirty-four carbon atoms. The remaining four carbon atoms were the original four methene bridge carbon atoms which, in this procedure, are oxidized to carbon dioxide. Since side reactions may have taken place during the oxida- tive cleavage, contaminating the carbon dioxide from the methene bridge carbon atoms, the evolved carbon dioxide was not collected. Nevertheless, as will be seen below, the source of the methene bridge carbon atoms can be determined by simple calculations. The haematinic acid was decarboxylated to methylethyl- maleimide and carbon dioxide. This carbon dioxide repre- sented carbon atoms ClO, DlO, i.e. the carboxyl groups of the porphyrin. Both methylethylmaleimide samples were then converted separately with osmium tetroxide to tartari- mide derivatives, and the latter cleaved with periodic acid to yield pyruvic acid and a-ketobutyric acid. The pyruvic acid samples in each case represent the methyl side of the pyrrole rings, i.e. carbon atoms A6, B6; A4, B4; A5, B5 from Rings A and B and carbon atoms C6, D6; C4, D4; C5, Do from Rings C and D. The a-ketobutyric acid samples represent the vinyl side of pyrrole rings A and B and the propionic acid side, minus the carboxyl groups, of pyrrole rings C and D; i.e. carbon atoms A9, B9; A8, B8; A3, B3; A2, B2; from Rings A and B, and carbon atoms C9, D9; C8, D8; C3, D3; C2, D2 from Rings C and D. The keto acids from each maleimide sample were separated on a silica gel column and converted to their respective 2:4-dinitrophenylhydrazones. The hydra- zones were oxidatively decarboxylated to yield acetic acid and carbon dioxide, and propionic acid and carbon dioxide, from the pyruvic and ketobutyric acid hydrazones respective- ly. The carbon dioxide samples from the pyruvic acid hydrazone represented carbon atoms A5, B5 and C5, D5 while Biosynthesis of Porphyrins 47 that from the a-ketobutyric acid hydrazone samples repre- sented carbon atoms A2, B2 and C2, D2. The acetic and propionic acid samples were then degraded stepwise by repeated Schmidt reactions on the free acids, liberating carbon dioxide from positions shown in Fig. 2. In this degradation scheme, not only can the ^*C activities of the individual carbon atoms be determined, but their ^*C activities can be checked by comparing the sum of the activities of the individual carbon atoms to that of the parent compound from which they are derived; e.g. the sum of the individual activi- ties of carbon atoms numbered 6, 4 and 5 can be compared to the activity of the pyruvic acid and likewise the sum of the activities of the keto acids can be compared to the activity of the maleimide samples. These degradation procedures were then utilized for the determination of the position in the porphyrin molecule of the carbon atoms which were derived from the a-carbon atom of glycine. Duck blood was incubated with glycine labelled with ^*C in its a-position and the hsemin isolated and degraded (Wittenberg and Shemin, 1950). It was found that the fraction representing Rings A and B (methylethylmaleimide) had the same activity as the fraction representing Rings C and D (Table II). Further degradation showed that all the activity of these fractions was contained in the a-ketobutyric acid. On degradation of the latter all the activity was found in the carboxyl group. Therefore carbon atoms A2, B2 and Table II Distribution of ^^C Activity in Protoporphyrin Synthesized from i*C Methylene Labelled Glycine,^ (NHa^^CHg-COOH) (Wittenberg and Shemin, 1950) Porphyrin Fragment Total Activity cpm Porphyrin 5236 Pyrrole rings A and B 1290 Pyrrole rings C and D 1324 Pyrrole rings A + B + C + D 2614 4 Methene Bridge carbon atoms 2620 48 David Shemin and Jonathan Wittenberg C2, D2, the alpha carbon atom of the pyrrole ring on the side of the vinyl and propionic acid side chains, were derived from the a-carbon atoms of glycine (Fig. 3). The findings that the activities of Rings A and B is equal to that of Rings C and D and that the carbon atoms in the pyrrole rings derived from glycine are in comparable positions (underneath the longer side chains) supports the earlier suggestion of a common precursor pyrrole. In this degradation we have isolated, as mentioned above, only thirty of the thirty-four carbon atoms of the porphyrin, CH, CH3 CH, CH, ( :h3 ( :h ( A H. j^H-" =C (X HCc5 VK. ?• ^ b — — C ' D H COOH CH- CHp II ^ CH B ■rt ^ N CH^ // CHp CH- CHg COOH PROTOPORPHYRIN 9 CHgNHgCOOH Fig. 3. Positions in protoporphyrin derived from the rx-carbon atom of gly cine. and of these thirty carbon atoms four (carbon atoms numbered 2) are derived from the a-carbon atom of glycine. It can now be easily demonstrated that the four methene bridge carbon atoms are also derived from the a-carbon atom of glycine (Muir and Neuberger, 1950; Wittenberg and Shemin, 1950). The total activity of the porphyrin in this experiment was 5236 cpm (Table II). However, the sum of the total activities of methylethylmaleimide and hsematinic acid, representing Biosynthesis of Porphyrins 49 thirty carbon atoms and containing four tagged carbon atoms, is 2614 cpm. Therefore the four methene bridge carbon atoms must have contained the remaining 2620 cpm. The average activity of the individual radioactive carbon atoms in the rings was 654 cpm, i.e. ( — - — ) while the average activity of the methene bridge carbon atom was 655 cpm, i.e. ( ). This demonstrated the equal utilization of the a-carbon atom of glycine for carbon atoms numbered 2 and for the methene bridge carbon atoms and that eight carbon atoms of protoporphyrin are derived from the a-carbon atom of glycine (Fig. 3). Since glycine accounts for eight carbon atoms of proto- porphyrin, the origin of the remaining twenty-six carbon atoms remained to be determined. It was previously demonstrated that on the administration of deuterioacetic acid (CD3COOH) to a rat, the haemin isolated contained deuterium (Bloch and Rittenberg, 1945). This indicated that some of the side chain carbon atoms of the porphyrin were derived from the methyl group of acetic acid, since these are the only carbon atoms bonded to hydrogen. Subsequent to this finding it was indeed demonstrated that both the methyl group and the carboxyl group of acetate were utilized for hsem formation (Muir and Neuberger, 1950; Radin, Rittenberg and Shemin, 19506). It was found that methyl groups of the porphyrin and the j8-carbon atoms of the pyrrole to which they are attached are derived from the methyl group of acetic acid, whereas the carboxyl group of acetate is utilized for the carboxyl group of the porphyrin (Radin et ah, 1950&). In order to determine the extent of utilization of acetate for porphyrin formation and to locate all the carbon atoms in the porphyrin which may be derived from acetate, duck blood was incubated separately with ^*C methyl labelled and with ^*C carboxyl labelled acetate (Shemin and Wittenberg, 1951), and the resulting labelled hsemin degraded as outlined above. The activities of both samples of acetate were the same. The ISOTOPES 5 50 David Shemin and Jonathan Wittenberg blood of each duck was divided into two parts; half was incubated with the methyl labelled acetate and the other half with the carboxyl labelled acetate. This was done so that the results of the two experiments could be compared, since the dilution of acetate would be the same, and presumably also the rate of synthesis in each of the experiments would be the same. Indeed, that the rate of synthesis was the same was demonstrated experimentally by including ^^N labelled glycine in the incubation mixture. The ^^N concentration of the haemin in the methyl labelled acetate experiment was found to be the same as that of the haemin in the carboxyl labelled acetate experiment. Results of the two experiments are therefore strictly comparable. The ^*C labelled hsemin samples from these experiments were degraded and the activities of the fragments and individual carbon atoms determined. Before discussing the individual activities of the carbon atoms we may come to some conclusions by examining the ^^C activities of some of the fragments. It can be seen from Table III (Shemin and Table III ^*C Activity of Fragments of Protoporphyrin Molecule (Shemin and Wittenberg, 1951) Porphyrin Fragment Total Activity Experiment using "C Methyl labelled acetate Experiment usine i^C Carboxijl labelled acetate Porphyrin (Mesoporphyrin) cpm 23,770 cpm 2,910 Pyrrole Rings A and B (methylethyl- maleimide) 11.620 456 Pvrrole Rings C and D (Ha^matinic \\c\d) 11,840 2,620t P\Trole Rings A + B + C + D* (methyl- ethylmaleimide and haematinic acid) 23,460 3,080 •Addition of activities found for pyrrole rings A + B and C + D. tThe activity of the methylethylnialeimide fragment (haematinic acid minus the carboxyl group) of Rings C and D equals 450 cpm. Biosynthesis of Cholesterol 51 Wittenberg, 1951) that the total activity of the porphyrins produced from ^*C methyl labelled and carboxyl labelled acetate resides in the carbon atoms other than the methene bridge carbon atoms, since the sum of the -total activities of the methylethylmaleimide and haematinic acid samples is equal to the total activities of the porphyrin samples. This is consistent with the previous finding that the a-carbon atom of glycine is the source of the methene bridge carbon atoms. In Table III it can be seen that the total activity of pyrroles A and B is equal to that of pyrroles C and D in haemin made from methyl labelled acetate. This comparison holds for pyrroles A and B and pyrroles C and D in haemin made from carboxyl labelled acetate if one excludes, for the moment, those carboxyl groups in C and D which are not found in A and B. Also, as will be seen later, in both of these experiments the carbon atoms of Rings A and B that occupy similar posi- tions in Rings C and D have the same activities. These findings support the previously suggested hypothesis that in the biosynthesis of protor porphyrin a pyrrole is formed which is the common precursor of both types of pyrrole structures found in protoporphyrin. This was first suggested by Turner (1940-41) but no evidence has heretofore been offered. Another conclusion concerning the mechanism of porphyrin formation can be drawn from the activities of the keto-acid fragments of the pyrrole rings. On cleavage of the pyrrole rings in the methyl labelled acetate experiment the activities of the pyruvic acid and a-ketobutyric acid samples of Rings A and B are equal to those of Rings C and D (Table IV) (Shemin and Wittenberg, 1951). The same relationship was found to exist in the experiment using carboxyl labelled acetate. On further degradation of these keto-acids it was found that, not only did the same number of carbon atoms contain ^*C, but the activities of comparable carbon atoms were similar (Fig. 4). In the porphyrin made from methyl labelled acetate, not only did the methyl group carbon atoms (A6, B6, C6, D6) of each pair of pyrroles have similar activities, 52 David Shemin and Jonathan Wittenberg but their activity was also equal to that of the terminal carbon atoms of the vinyl group (A9, B9) and the correspond- ing carbon atoms of the propionic acid side chains (C9, D9). The methyl-bearing carbon atoms in all the pyrrole rings (A4, B4, C4, D4) had the same activity as the proximal carbon atoms of the vinyl side chains of Rings A and B (A8, B8) and their counterparts in the propionic acid side chains of Rings C and D (C8, D8). Also the carbon atoms numbered 5 in the pyrrole rings had the same activity as all ring carbon atoms to which the longer side chains are attached (A3, B3, Table IV i*C Activities of Fragments of Pyrrole Rings (Shemin and Wittenberg, 1951) Pyrrole Rings Fragments Total Activity Experiment using Methyl labelled acetate Experiment using Carboxyl labelled acetate PjTTole Rings A + B Pyruvic Acid a-ketobutyric Acid Pyruvic Acid + a-ketobutyric acid* . cpm 11,620 5,440 5,560 11,000 cpm 456 206 208 414 Pyrrole Rings C + D (minus carboxyl 'group) Pyruvic Acid a-ketobutyric acid .... Pyruvic Acid -f- a-ketobutyric acid* . 11,840 5,500 5,480 10,980 450 204 190 394 ♦Addition of activities found for pyruvic and a-ketobutyric acids. C3, D3). This diagonal relationship can be seen from Fig. 4. The carboxyl group of the a-ketobutyric acid samples which corresponds to the carbon atoms numbered 2 contained no , ^*C, which is in agreement with the previous finding that the source of these carbon atoms is the a-carbon atom of glycine. Similarly, in the experiment using carboxyl labelled acetate, all carbon atoms numbered 5 and 3 had similar activities (average values given in Fig. 4). The carboxyl group of the porphyrin (ClO, DlO), found only in Rings C and D, is derived Biosynthesis of Porphyrins 53 from the carboxyl group of acetate. These data strongly suggest that not only are the two types of pyrrole unit in protoporphyrin made from the same precursors, but also that in each pyrrole ring the same compound is utilized for the methyl side of the structure and for the vinyl and propionic acid sides of the structure. This conclusion is supported by the .14 C H3COOH Experiment COOH ^^ A>' / \ N H CH3C"^00H Experiment :ooH('i7o) :h2 H3C y c / \ 10 COOH 9 CH2 6 H3C 8 CH2 N H C 5 2C /\;/^ H Fig. 4. Average activities of comparable carbon atoms in all pyrrole units. The activities are given in parentheses. The pyrrole unit represented above contains a carboxyl group which is found only in Rings C and D of protoporphyrin. In the central figure the carbon atoms are designated according to the num- bering system. finding, as pointed out earlier, that the pyruvic acid and a-ketobutyric acid fragments of the pyrrole units have the same activities in each of the experiments using methyl labelled and carboxyl labelled acetate {Table IV). If each side of each pyrrole unit utilizes the same com- pound, the precursor which condenses with glycine to form the pyrrole unit must be either a three or a four carbon atom 54 David Shemin and Jonathan Wittenberg compound. On examination of the structure of protopor- phyrin and noting the quantitative distribution of ^^C among the carbon atoms in the experiments, it can be seen that a three carbon atom compound would satisfy the data as the precursor of the methyl sides of the pyrrole units (carbon atoms 6, 4 and 5) and the same compound would also be consistent with the data as the precursor of the vinyl sides of pyrrole units A and B (carbon atoms 9, 8 and 3), excluding carbon atom 2, which is derived from the a-carbon atom of glycine. However, it would appear that a four carbon atom compound would be necessary as the precursor for the pro- pionic acid sides (carbon atoms 10, 9, 8 and 3) of pyrrole units C and D, again exclusive of carbon atom 2. If a three carbon atom compound were utilized, subsequent carboxyla- tions must have occurred on positions C9 and D9. On the other hand, if a four carbon atom compound Avere utilized, decarboxylations must have occurred on all 6 positions and on positions A9 and B9 subsequent to pyrrole formation. It can be decided which of these two alternative mechanisms operates in the synthesis of protoporphyrin by correlating some of the data obtained in the experiments using methyl labelled and carboxyl labelled acetate. This correlation is valid since the experiments were so carried out that the synthesis of hsem from carboxyl labelled acetate proceeded to the same extent as with methyl labelled acetate. The activities of carbon atoms ClO, DlO (1170 cpm) in the carboxyl groups in protoporphyrin made from carboxyl labelled acetate are equal, with the limits of error, to those of carbon atoms C9, D9 (1128 cpm), adjacent to these groups, in the porphyrin made from methyl labelled acetate (Fig. 4). This equality, i.e. the same degree of dilution, makes it appear that the acetic acid enters as a unit and that the utilization of acetic acid for 2jyrrole formation is via a four carbon atom compound. If ClO and DlO had been introduced by carboxylation the activity of these carbon atoms would have been much lower. Moreover, it has been shown that carboxyl labelled acetate gives rise to labelled carbon dioxide Biosynthesis of Porphyrins 55 in the system used and that radioactive carbon dioxide is not incorporated into hsemin (Radin et at., 19506; Bufton, Bentley and Rimington, 1948). As stated above, in the methyl labelled acetate experiment the activities of the methyl groups (A6, B6, C6 and D6) and the activities of the terminal carbon atoms of the vinyl groups (A9, B9) are equal to that of the corresponding carbon atoms of the propionic acid side chains (C9, D9). Therefore the activities of A6, B6, C6 and D6 and A9, B9 are also equal to that of the carboxyl groups of haem (ClO, DlO) made from carboxyl labelled acetate (Fig. 4). It would appear, therefore, from this distribution and the evidence of the utilization of acetic acid as part of a four carbon atom unit, that in some intermediate stage in the formation of protoporphyrin a pyrrole or porphyrin was formed hearing carboxyl groups attached to the four methyl carbon atoms and to the terminal carbon atoms of the vinyl side chains-, in other words, it would appear that the common precursor pyrrole originally formed contained acetic and propionic acid side chains in its ^-positions (Fig. '7). Although the evidence presented thus far makes it seem highly probable that each of the four pyrrole rings bore two carboxyl groups in some stage of synthesis and that a four carbon atom compound was utilized for pyrrole synthesis, more evidence will be furnished later for these conclusions. The data obtained from these acetate experiments can readily be explained by assuming the participation of the tricarboxylic acid cycle in porphyrin formation. In the light of the relative distribution of the activities among the carbon atoms of the porphyrin derived from acetate, it would appear that acetate was utilized via a member of the tricarboxylic acid cycle. It has been demonstrated that the acetic acid carbon atoms on entering the tricarboxylic acid cycle are the direct source of the y-carbon atom and the y-carboxyl group of a-ketoglutaric acid (Buchanan and Hastings, 1946; Wood, 1946; Lifson, Lorber, Sakami and Wood, 1948; Wood, 1948). The use of a-ketoglutarate in the following argument is not intended as evidence for the participation of this compound 56 David Shemin and Jonathan Wittenberg per se in porphyrin formation. It is merely used to indicate the participation of the tricarboxyhc acid cycle as a whole in this synthesis. For example, if one starts with methyl labelled acetate with a relative activity of 10 in the methyl group after endogenous dilution, the a-ketoglutaric acid formed on the first turn of the cycle would contain ^*C activity only in the y-carbon atom, and the relative activity would be 10 (see Fig. 5 and Table V). When the a-ketoglutarate has been CH2-C00H HOOC-CHp-C-COOH £i I OH CITRATE CH3-COOH ACETATE \| N CH2-COOH I HOOC-CO- C-COOH I H OXALOSUCCINATE [\ CH2-COOH HOOC-CO-CH2 Oi KETOGLUTARATE -7 - V HOOC-CH2-CO-COOH ^ OXALACETATE N OH I HOOC-CH2-C-COOH I H MALATE \[ HOOC-CH=CH-COOH FUMARATE -> HOOC-CH2-CH2-COOH CH3-CO-COOH PYRUVATE SUCCINATE Fig. 5. Abridged scheme of the tricarboxylic acid cycle. converted to the symmetrical succinic acid, the activities of the methylene carbon atoms would (ignoring dilution by endogenous succinic acid) be 5 and 5, those of the oxaloacetate eventually formed would contain half of the activity of the y-carbon atom of the a-ketoglutarate. The recycling of this oxaloacetate with the labelled acetate into the tricarboxylic acid cycle would result in a-ketoglutarate having the relative activities shown in Table V for the second cycle. On the third and after an infinite number of cycles, therefore, the activities of the carbon atoms of the a-ketoglutarate would Biosynthesis of Porphyrins 57 have the relative activities shown in Table V. If a-keto- glutarate or an un symmetric compound derived from a-ketoglutarate were utilized for hsem synthesis, after a finite number of cycles most of the carbon atoms of the hsem would contain ^^C. Moreover, the carbon atoms of the porphyrin derived from the y-carbon atom of the a-ketoglutarate would have the highest activity and the two adjacent carbon atoms in the haem would theoretically have somewhat lower but Table V Relative Distribution of ^^C Activity in the Carbon Atoms of a-KETO- GLUTARic Acid resulting from Utilization of i*C Labelled Acetate in THE Tricarboxylic Acid Cycle From "C Methyl Labelled Acetate {activity of methyl group = 10 cpm) From "C Carboxyl Labelled {activity of carboxyl group = Acetate 10 cpm) a-Keto- glutaric Acid Number of Cycles in the Tricarboxylic Acid Cycle 1st 2nd 3rd 00 1st 2nd 3rd 00 COOH cpm cpm cpm cpm cpm cpm cpm 10 10 10 cpm 10 CH2 10 10 10 10 CH2 5 7-5 10 c=o 5 7-5 10 COOH 2-5 5 5 5 5 equal activities. It can be seen from Fig. 4 that a pattern exists on both sides of each pyrrole unit consisting of three carbon atoms, one with the highest activity and two adjacent carbon atoms having somewhat lower activity. The com- parable carbon atoms numbered 6 and 9 have the highest activities and the carbon atoms numbered 4 and 5 on the methyl side and 8 and 3 on the opposite side have somewhat lower activities. However, the comparable carbon atoms numbered 4 and 8 are slightly more active, on the average, than those numbered 5 and 3. The inequality of activities of these carbon atoms (numbered 4 and 8; 5 and 3) would 58 David Shemin and Jonathan Wittenberg at first suggest that the tricarboxyhc acid cycle is not func- tioning as postulated theoretically above. It was found, however, that carbon atoms numbered 5 and 3 are also in part derived from the carboxyl group of acetate (Fig. 4). Correcting for this dilution, since in the methyl labelled ace- tate experiment the carboxyl group of the acetate contains no ^*C, the average activities of these two pairs of adjacent carbon atoms are equal. The contribution of the carboxyl groups of acetate to positions 5 and 3 is 100 counts on the average. The addition of these 100 counts to these carbon atoms having an average activity of 788 cpm gives a total of 888 cpm, a figure close to the 877 cpm found for the average activity of carbon atoms 4 and 8 in the product from methyl labelled acetate. It would appear therefore that acetate is utilized through the tricarboxvlic acid cvcle, and the relative activities found fit those theoretically predicted on the basis of the distribution of activities in a-ketoglutarate. This also supports the previous conclusion that the same carbon compound is utilized for the methyl side of the pyrrole as for the vinyl and propionic acid sides. On the propionic acid side a compound containing not less than four carbon atoms must have been utilized, since carbon atoms C2 and D2 arise from the a-carbon atom of glycine. If originally a three carbon compound (e.g. pyruvic acid), rather than a four carbon compound were utilized for the methyl side or vinyl side, the methyl group numbered 6 and carbon atom 4 of the pyrrole to which it is attached would have had equal activities, whereas carbon atom 5 of the pyrrole w^ould have been much less active than carbon atom 4. The same relationship would also exist for carbon atoms 3, 8 and 9. The distribution of ^^C in pyruvic acid from methyl labelled acetate can be gathered from Fig. 5 and Table V and has been demonstrated by Wood (1948). The non-utilization of pyruvic acid can also be concluded from the data of the experiment with carboxyl labelled acetate. If pyruvic acid were utilized, carbon atoms 5 and 3 would have been similar in activity to carbon atoms numbered 10. Biosynthesis of Porphyrins 59 Therefore the distribution of activities among the carbon atoms of the porphyrin fits the utihzation of an unsymmetric compound with a minimum of four carbon atoms, as concluded above, rather than a three carbon compound. The finding that in the carboxyl labelled experiment the activities of ClO, DlO are much greater than carbon atoms numbered 5 and 3 eliminates many members of the tricar- boxylic acid cycle, such as all the four carbon atom dicar- boxylic acids, as immediate precursors. Once succinic acid, a symmetrical compound, is formed, the two terminal or carboxyl atoms would have equal activities. If succinate or the fumarate, malate or oxaloacetate subsequently formed were utilized, the carboxyl groups of protoporphyrin (carbon atoms ClO, DlO) and carbon atoms 3 and 5 would have equal activities in hsem made from carboxyl labelled acetate. Also if any of these dicarboxylic acids were utilized, the carboxyl groups of the porphyrin would have had the same activity as carbon atoms C3 and D3 in the experiment in which methyl labelled acetate was used. From all these considerations it would appear that two molecules of an unsymmetrical compound arising from the tricarboxylic acid cycle condense with glycine to form pyrrole units containing acetic and propionic acid side chains (Fig. 7). Lemberg and Legge (1949) have suggested that 2 mols of a-ketoglutaric acid condense with glycine with the elimination of the a-carboxyl group of the keto-acid to form a pyrrole bearing acetic acid and propionic acid side chains. Muir and Neuberger (1950) have adopted the suggestion of Lemberg and Legge with a modification. They believe that the keto-acid condenses with hydroxyaspartic acid, since it has been shown that eight a-carbon atoms of glycine are utilized in porphyrin formation. However, this view is incompatible with the distribution of the a-carbon of glycine in the por-. phyrin (Wittenberg and Shemin, 1950). Our data also appear to eliminate a-ketoglutaric acid as an immediate precursor of the porphyrin. The mode of utiliza- tion of carboxyl labelled acetate in the tricarboxylic acid cycle 60 David Shemin and Jonathan Wittenberg would result in a-ketoglutaric acid labelled only in the y-car- boxyl group in the first turn of the cycle. The a-ketoglutaric acid produced after repeated cycles would contain ^^C in both carboxyl groups but in no other position (Fig. 5 and Table V) (Lifson et al,, 1948; Wood, 1948). If a-ketoglutaric acid were directly utilized, with the elimination of the a-carboxyl group, the protoporphyrin made from carboxyl labelled acetate would contain ^*C only in the carboxyl group. Actually, however, carbon atoms numbered 3 and 5 in the porphyrin contain some ^*C. These atoms correspond to the carbonyl caibon atom of a-ketoglutaric acid, and should contain no ^*C if a-ketoglutaric acid were the immediate precursor, unless the conversion of a-ketoglutaric acid to succinic acid were reversible. From the best evidence to date it would appear that this reaction is irreversible in higher animals, and until this reaction is shown to be reversible another intermediary compound must be postulated. ' The postulated unsymmetric four carbon intermediate must take into account the finding of some activity in carbon atoms numbered 3 and 5 in the experiment using carboxyl labelled acetate. The low activity in carbon atoms 3 and 5 in conjunction with the high activity in the carboxyl groups of haem (ClO, DlO) produced from carboxyl labelled acetate can be explained by presupposing that the compound utilized is derived in greater part from an unsymmetrical compound and in lesser part from a symmetrical compound. A four carbon atom unsymmetric intermediate arising from both a-ketoglutaric and succinic acids would explain the findings (Fig. 6). The detailed mechanism of the conversion of a-ketoglutarate to succinate is at present unknown. It is conceivable that a four carbon atom unsymmetric compound may be an intermediate in this conversion. This compound may be the semialdehyde of succinic acid or more likely a succinyl coenzyme complex. The succinyl coenzyme com- plex may be formed in a manner analogous to the formation of acetyl coenzyme A from both pyruvate and acetate. a-Ketoglutaric acid labelled in either the y-carboxyl group or Biosynthesis of Porphyrins 61 in both carboxyl groups would on decarboxylation yield a succinyl derivative labelled only in the carboxyl group. This compound in turn would be converted to succinic acid, a symmetrical compound. If the latter reaction is reversible (XOC-CHs-CHsCOOH^HOOC-CHa-CHs-COOH) the succinyl derivative arising from the symmetrical succinate would contain equal activity in the carboxyl group and in the other terminal carbon atom. However, since the succinyl derivative 14 C OOH I CHo I CH2 C = I COOH 14 C^OOH C"*OOH C'^OOH I CH? 1 CHp I COX C"^00H CH2 CH2 ' 14 c'^ox. / PYRROLES Fig. 6. Formation of succinyl derivative and distribution of ^*C in this compound in experiment using carboxyl labelled acetate. is presumably formed more extensively from a-ketoglutarate than it is from succinate, the carboxyl group of the pooled intermediate arising from both processes would contain more activity than the other terminal carbon atoms.* Two molecules ♦According to the formulation of the functioning of the tricarboxylic cycle in porphyrin formation, carboxyl labelled succinate cannot give rise to labelled haem unless this reaction occurs. It has recently been found by Shemin and Kumin that carboxyl labelled succinate produced labelled haem in which pyrrole Rings A and B contained 40 per cent of the activity and Rings C and D contained 60 per cent. The carboxyl group of Rings C and D contained one- third of the activity of these rings. This distribution agrees with the theory and therefore further supports the above reaction and the utilization of a four carbon compound for porphyrin formation. 62 David Shemin and Jonathan Wittenberg COOH CHp CH2-- CHo I ^ CH2 ■COX COOH of the succinyl derivative may then condense with glycine to form a pyr- role containing a carboxymethyl and a carboxyethyl group in its ^-positions (see Fig. 7). The formation of protoporphyrin may therefore be visualized as follows (Fig. 8). Four of these mono-pyrroles are condensed, with the loss of the a-carboxyl group of the pyrrole and with the addition of a compound originating from the a-carbon atom of glycine, by a mechanism outlined in a previous paper (Wittenberg and Shemin, 1950), where it was suggested that the compound may be combined with a co- enzyme and be utilized in a manner analogous to the synthesis of porphyrins described by Andrews, Corwin and Sharp (1950). The tetra-pyrrole first formed would be uroporphyrin III, which by decarboxylation of the carboxymethyl side CPX CH2-COOH H2N Fig. 7. Hypothetical scheme for formation of common precursor pjTrole from glycine and the suc- cinyl derivative. COOH COOH CH2 CH2 CH2 H^ N^COOH H 1 COOH CHj CH3 CH2 ^ UROPORPHYRIN ■^ COPROPORPHYRIN PROTOPORPHYRIN H N XOOH H Fig. 8. Hypothetical scheme for formation of protoporphjTin from common precursor pyrrole. Biosynthesis of Porphyrins 63 chain would be converted to coproporphyrin III. The latter by decarboxylation and dehydrogenation of the propionic acid side chains of pyrroles A and B would yield protoporphyrin. An alternative pathway may be suggested in which the first decar- boxylation occurs at the mono-pyrrole stage. Decarboxylation • I I Methyl group of Acetic Acid 9 Corboxyl group of Acetic Acid X Of-Corbon atom of Glycine Fig. 9. Schematic representation of the origin of the carbon atoms of protoporf)hyrin. of the carboxymethyl group and condensation of these derived pyrrole compounds would result in the formation of copropor- phyrin, by-passing uroporphyrin. It is significant that in all of the naturally occurring porphyrins the side chains can theoretic- ally be derived from carboxymethyl and carboxyethyl groups. A summary of the source of each of the carbon atoms of protoporphyrin is given in Fig. 9, where the X is the a-carbon 64 David Shemin and Jonathan Wittenberg atom of glycine, the squares represent those carbon atoms derived from the methyl group of acetic acid, the filled circles represent the carboxyl group of acetic acid, and the carbon atoms which are in main derived from one of the carbon atoms of acetate and partly from the other are appropriately designated. REFERENCES Altman, K. I., Casarett, G. W., Masters, R. E., Noonan, T. R., and Salomon, K. (1948). J. bioL Chem., 176, 319. Andrews, J. S., Corwin, A. H., and Sharp, A. G. (1950). J. Amer. chem. Soc, 72, 491. Bloch, K., and Rittenberg, D. (1945). J. biol. Chem., 159, 45. Buchanan, J. M., and Hastings, A. B. (1946). Physiol. Rev., 26, 120. Button, A. W. J., Bentley, R., and Rimington, C. (1948). Biochem. J., 43, xlix. Grinstein, M., Kamen, M. D., and Moore, C. V. (1948). J. biol. Chem., 174, 767. Lemberg, R., and Legge, J. W. (1949). HemMin Compounds and Bile Pigments. New York: Interscience, LiFSON, H., LoRBER, V., Sakami, W., and Wood, H. G. (1948). J. biol. Chem., 176, 1263. MuiR, H. M., and Neuberger, A. (1949). Biochem. J., 45, 163. MuiR, H. M., and Neuberger, A. (1950). Biochem. J., 47, 97. Radin, N. S., Rittenberg, D., and Shemin, D. (1950a). J. biol. Chem., 184, 745. Radin, N. S., Rittenberg, D., and Shemin, D. (19506). J. biol. Chem., 184, 755. Shemin, D., Lontjon, I. M., and Rittenberg, D. (1948). J. biol. Chem., 173, 799. Shemin, D., London, I. M., and Rittenberg, D. (1950). J. biol. Chem., 183, 757. Shemin, D., and Rittenberg, D. (1945). J. biol. Chem., 159, 567. Shemin, D., and Rittenberg, D. (1946). J. biol. Chem., 166, 621, 627. Shemin, D., and Wittenberg, J. (1951). J. biol. Chem. In press. Turnt:r, W. J. (1940-41). J. Lab. din. Med., 26, 323. Wittenberg, J., and Shemin, D. (1949). J. biol. Chem., 178, 47. Wittenberg, J., and Shemin, D. (1950). J. biol. Chem., 185, 103. W^ooD, H. G. (1946). Physiol. Rev., 26, 198. Wood, H. G. (1948). Cold Spr. Harb. Sym. quant. Biol, 13, 201. DISCUSSION Neuberger: I am wondering whether the activity in the pyrrole rings which is derived from carboxyl-labelled acetate could not be explained without assuming this hypothetical 4-carbon compound. If, for example, the reaction of succinic acid with COj to produce ketoglutarate were Biosynthesis of Porphyrins 65 reversible — I know the free energies would be very much against this equilibrium and would certainly be vastly in favour of the succinate — but if such a reaction occurred in the cell to a small extent, it might explain the findings of Dr. Shemin without assuming the 4-carbon intermediate. He may have some reasons why he excluded that possi- bility. Shemin: I am assuming the 4-carbon intermediate at least until the ketoglutarate-succinate reaction is shown to be reversible. From purely organic chemical lines I can't see how ketoglutarate can go to succinate without intermediates, involving as it does an oxidative decarboxylation; a succinyl derivative, e.g. a succinyl coenzyme complex, may very well be an intermediate. I prefer this 4-carbon compound with reactive groups on a terminal carbon atom. It could very well be that keto- glutaric acid would explain all the findings if ketoglutaric acid were derived from succinate, but so far it has not been demonstrated, as far as I can gather. Wood: I have not seen any convincing evidence for the synthesis of a-ketoglutarate from COg and succinate in animal tissues. In fact, the work that has been done on the reversibility of this reaction with bacteria is not so convincing that it should be accepted as a proved reaction. I think that the chances of an intermediate occurring in this reaction are exceptionally good. Dr. Utter in our laboratory has provided evidence that there is an intermediate C3 compound in the fixation of CO 2 in oxaloacetate. Until now this reaction has usually been considered to be a straightforward reaction between pyruvate and COj. We believe now that it is not, because pyruvate does not always exchange in oxaloacetate at the same rate as COg. Dr. Utter has shown that there are two enzymes involved: with one enzyme present CO2 is fixed in oxaloacetate, but there is no fixation of pyruvate; if then the other enzyme is added, the fixation of pyruvate and CO 2 is approximately equal. Isn't it a misnomer to speak of acetic acid as the "source" of so many carbons? Actually we probably should speak of either a-ketoglutarate or succinate as the source. Have you tried labelled a-ketoglutarate to see whether it will replace acetic acid? Shemin: I agree that we shouldn't refer to acetic acid as the source of the carbon atoms, and as soon as we know the structure of the 4 carbon compound, we will have to call it the precursor. We are at present synthesizing ketoglutaric acid and hope to see whether it gives rise to the same pattern of distribution; we are, in fact, making all members of the citric acid cycle. We can, however, dilute out acetate incorporation by unlabelled members of the citric acid cycle, and in fact can dilute out acetate to zero very nicely with the semialdehyde of succinic acid. But there are objections to this type of experiment. I think one must synthesize the various compounds and do the full degradation. Wood: a-ketoglutarate does dilute it out? Shemin: Yes, and citrate and succinate also dilute the utilization of acetate. These compounds do not damage the cell system; ^^N labelled ISOTOPES 6 66 David Shemin and Jonathan Wittenberg glycine was included in these experiments to check this point. The hsemin samples isolated in these experiments had the same ^^N con- centrations, demonstrating the same rate of synthesis. The members of the citric acid cycle merely dilute the acetate utilization. Bloch: If I understand you correctly, you assume a minimum of three revolutions of the cycle in order to account for the data and for the similarity of the methyl and the carbonyl atoms of pyruvate? Shemin: I am only assuming that it is a finite number. I wouldn't hazard a guess how many cycles it went around. Block: On the other hand, if there is more than one revolution, it would appear that the two carboxyl groups of succinate should become equilibrated, whether you have an unsymmetrical intermediate or not. Would you not then expect a larger incorporation of the carboxyl carbon of acetate into one of the pyrrole rings than you actually find? Shemin: No, because the labelled carboxyl group of succinate would never get into the porphyrin with forward revolutions of the cycle. It would have to go backward for it to get in. I hope we shall be able to demonstrate this by the time I get back. We shall have carboxyl labelled succinate. Theoretically, if we find any labelling in the hsemin it would show the reversibility of this step of the cycle. Neuberger: Mightn't you get a reduction of succinate to your intermediate? Shemin: Yes. That implies a backward step. Neuberger: That would get it into the ring too. But it wouldn't have to go back to ketoglutarate. Shemin: No. To the 4 carbon compound. Rimington: The position of uroporphyrin III is a very important one in the whole scheme of biosynthesis, and it is worth while reporting some information which will help to clear up some of the difficulties which have surrounded this substance in the past. Authentic uropor- phyrin III has not up to the present been isolated from natural sources. The uroporphyrin present in congenital porphyria urine is the series I type. In acute porphyria Waldenstrom and Mertens reported that they had isolated uroporphyrin III, because on decarboxylation they got coproporphyrin III; Watson and his co-workers were able to show, however, that in samples they studied, decarboxylation gave rise to a mixture of isomers. We have approached this problem from two different lines. We have isolated quite unequivocal uroporphyrin III, which we have been able to characterize, from turacin. I showed some years ago that turacin on decarboxylation gave only pure coproporphyrin III. We have gone into the question of the existence of uroporphyrin III in pathological urines and I think the position may be summarized fairly by saying that in some cases the excretion is almost entirely uroporphyrin III, agreeing ^ith Waldenstrom and Mertens, but in other cases there is undoubtedly a mixture of the two isomers. Finally, we have been able to detect uroporphyrin in normal urine, working up large quantities and examining it by paper chromatography. We are now working on larger amounts in order to try and identify the isomeric series. Biosynthesis of Porphyrins 67 Dr. Shemin, has it been possible to demonstrate the tricarboxyhc acid cycle in the avian erythrocyte? Shemin: The activities found in the porphyrin agree with the theo- retical prediction that the citric acid cycle is functioning. Dische, working with fowl erythrocytes, has informed me that the whole cycle exists in the erythrocytes. Rittenberg: Since the more serious part of this discussion is ended, may I raise an objection to the unanimity which exists between Dr. Wood and Dr. Shemin on what the precursor is. In actuality, this is not a scientific but a semantic problem. I would rather go back to the first compound which really works, setting aside theoretical arguments. For that reason, I think acetic is still the best one. Shemin: Purely as a matter of semantics, I still disagree; the four carbon atom compound is the more immediate precursor. The acetic acid was used in this study but the same results should be found with other members of the citric acid cycle. Gurin: I should like to ask what has happened to the one carbon precursors. There was a good deal of work going on with formate and formaldehyde. What is the position with respect to that now? Shemin: The a-carbon of glycine and formate or formaldehyde are equivalent for the 2 and 8 positions of the purines, and it has been th ought by some that they are equivalent in other biological reactions. But in the porphyrins, the a-carbon of glycine is utilized, but formalde- hyde, formate, methyl amine, cyanide and COg are not. Lardy did find a little activity in the porphyrins after formate, but he did not do any degradation. In preliminary studies with i^C labelled methyl alcohol in an experiment in which you can compare the purines and porphyrins, yo u find that while the 2 and 8 positions of guanine have 30,000 counts, the individual methene bridges of the porphyrin have about 100 counts, which is really insignificant; there was more activity in the pyrrole ring from the Ci compounds than there is in the methene bridge. Glycine and the one carbon compounds are just not equivalent for the methene bridge carbon atoms. Neuberger: In the rabbit we had exactly similar experience. We tried a number of one-carbon intermediates, not as many as Dr. Shemin, but got completely negative results. STUDIES ON MAMMALIAN RED CELLS A. NEUBERGER The Life Span of the Mammalian Red Cell Observations in Man In man several methods have been successfully used to determine the life span of the red cell. With all these methods the persistence of a particular component in the circulating red cells over a period of time is determined and in order to do this, the substance under investigation has to be labelled and its concentration must be measurable by a physical, chemical or biological method. If, as is the case, different labelled components are used, it cannot be an a priori assump- tion that the results obtained by the various methods will necessarily be identical. In the technique originally devised by Ashby (1919) group O cells are transfused into recipients belonging to groups A or B and the persistence of cells lack- ing A or B antigens is determined by agglutination. The assumption is made that the transfused group O cells have the same life expectation in the circulation of the recipient as they would have had in their original environment, i.e. in the body of the donor. It is also assumed that the agglut- inogen is not detached before the cells die nor re-utilized in the construction of a new cell. This method of biological labelling has, so far as we are aware, only been applied to man. The other methods to be discussed use the persistence of a label present in the haemoglobin molecule to estimate the life span. Jope (1946) measured the rate of disappearance of sulphgemoglobin from the blood of subjects who had been in contact with trinitrotoluene immediately before the beginning of the experiment. It is assumed that the cell thus chemically labelled has the same life expectation as the normal red cell. 68 Studies on Mammalian Red Cells 69 Shemin and Rittenberg (1946a) found that glycine was a specific precursor of the nitrogen of haem. They then studied the change of isotope concentration with time of the haemo- globin haem of a subject fed ^^N-labelled glycine (Shemin and Rittenberg, 1946^). There is broad agreement between results obtained by the three methods, indicating that the average life span of the human red cell is about 120 days. This conclusion is also supported by the results obtained on the changes of concentration of ^^N in stercobilin after feeding isotopic glycine to a human subject (London, West, Shemin and Rittenberg, 1950; Gray, Neuberger and Sneath, 1950). However, if the data are examined in detail, certain dis- crepancies appear between results obtained by the two methods, i.e. the isotopic labelling of the porphyrin and the agglutination technique. In the Ashby technique the trans- fused cells constitute a mixed population with respect to age, and the mean life expectation is not immediately apparent from the observed data. However, analysis of the survival curves (Callender, Powell and Witts, 1945) indicates (a) that practically no cells survive beyond the 130th day, and (b) that very few cells die before the 90th day. In other words the scatter round the mean value is very small. The main complication in the interpretation of the isotope experiments is the fact that the glycine of the body which is utilized for porphyrin formation will contain significant amounts of ^^N for some time after feeding of the labelled compound; thus the labelled cells are not released into the circulation at one instant, but over a certain period of time. Shemin and Rittenberg (19466) have allowed for this fact in their analysis, and it appears that the release of labelled cells is practically completed by the 20th day. However, the data of the Columbia workers (Shemin and Rittenberg, 19466, London, Shemin, West and Rittenberg, 1949) and also our own obser- vations indicate that the ^^N content of the porphyrin is still about 20 per cent of the maximal value 170 days after feeding and still about 12 to 15 per cent after 220 days. It might be argued that this unexpectedly high isotope content is due 70 A. Neuberger to labelling of red cells formed about 60 to 100 days after the beginning of the experiment. Such an assumption, which implies that the average isotope content of the glycine which is utilized for porphyrin synthesis has only decreased after 80 days to about 15 per cent of its maximum value, is in conflict with the mathematical treatment of Shemin and Rittenberg (19466) and is altogether unlikely. Moreover, if labelled haem should still be incorporated into newly formed cells 50-80 days after the administration of the labelled glycine, the isotope content of the hsem should rise consider- ably between the 50th and 80th day. This is not so. We suggest therefore that a proportion of the hsem or of the haemoglobin of decaying red cells may be utilized in the pro- duction of new cells even in normal man. It is uncertain whether such a re-utilization, which will normally affect only about 15-20 per cent of the total erythrocytes, involves the whole haemoglobin molecule or only the haem. Another, possibly less real, discrepancy concerns the period between the 30th and 100th day. The results of London et al. (1949) and also our own data suggest that there is, at least in some apparently normal subjects, a decrease of about 10-12 per cent of isotope content in the haem between the 40th and 80th day. Moreover this rate of fall increases greatly after 90 days. On the other hand the results obtained by the agglutination method suggest that hardly any normal cells die before the 90th day. It will be necessary to evaluate critically the limits of error of the two methods and also the mathematical analysis employed in the interpretation of the results before any decision can be reached as to whether this discrepancy is real. At the present time the possibility that the haemoglobin in the normal and mature human red cell is metabolically somewhat active cannot be excluded. The Life Span of the Red Cell in the Rabbit In this laboratory we have been particularly interested in the red cell of the rabbit (Neuberger and Niven, 1951); for labelling the cells we have been using both ^^C and ^^N (Muir, Studies on Mammalian Red Cells 71 Neuberger and Perrone, 1951). The curves obtained differ markedly from those observed in man. The isotope content of the porphyrin rose sharply from the 10th-40th hour after injection and a maximum was usually observed between the fourth and the sixth day, but occurred sometimes on the tenth day of the experiment. In some of the rabbits there was a plateau between the 10th and 30th day, but in most animals there was an almost linear decrease from the 10th-12th day -40 60 80 100 120 140 160 DAYS AFTER ADMINISTRATION OF GLYCINE Fig. 1. ^^N content (atom per cent excess) of haemin samples obtained at various times from a rabbit (3-2 kg. body wt.) which had received 3-5 g. of glycine containing 30-5 atom per cent excess ^^N. Broken line corrected for loss of blood. onwards, similar to that found in haemolytic ansemias in man. The isotope content of the haem had generally fallen to half its maximum value by about the 50th day. In most rabbits the rate of decrease after the 50th day slowed down con- siderably, and there was still about 15 per cent of the maxi- mum isotope content present in the haem 160 days after feeding isotopic glycine (Fig. 1). The main difficulty in these experiments is that the withdrawal of blood required for isotope analysis may bring about a secondary anaemia and thus affect the results. In order to minimize this danger, we 72 A. Neuberger used large animals with blood volumes of about 160-180 ml.; moreover, we only took about 3 ml. of blood for each sample and altogether not more than eight specimens per animal. The haemoglobin content did not show any marked changes during the experiments. The interpretation which we adopted at first was that the rabbit red cell has a much shorter life span, mean value about 50-60 days, than that of man, and that the scatter of life spans for individual cells is very much greater. However, more recently, we have become doubtful about the validity of our assumption that the persistence of the label in the haem is necessarily a quantitative measure of the life span, i.e. the period between the release of the cell into the circula- tion and the disintegration of its morphological structure. The possibility that haemoglobin in the mature human erythrocyte is metabolically active was already mentioned above and this may apply to an even greater extent to the red cell of the rabbit. Recently Benard, Gajdos and Gajdos- Torock (1950) have injected folic acid and vitamin B^g into normal rabbits and have found marked increases within 48 hours in the number of red cells without anv significant change in the haemoglobin content. These results are most readily explained by assuming some redistribution of haem or haemoglobin between old and new cells. It was partly in order to get further information on this point that we carried out experiments in which various labelled precursors of the components of the red cells were administered; the haem, the different amino-acids of the globin and of stromatin, and in addition cholesterol were isolated and their radioactivities determined. Cholesterol in the Rabbit Red Cell A rabbit was injected with ^^C-carboxyl-labelled acetate and methylene-labelled glycine (Muir, Perrone and Popjak, 1951). The latter produced activity in the protoporphyrin and glycine residue in the globin, whilst the former gave rise to activity mainly in the cholesterol and, with the dose used. Studies on Mammalian Red Cells 73 produced only slight activity in the porphyrin. The time curves for radioactivities of the cholesterol and porphyrin respectively were very different (Fig. 2) and results show that the cholesterol of the cell membrane is metabolically active to a much greater extent than the porphyrin. The half-life of red cell cholesterol was found to be about 12-14 days, the 300 2 •|200 ♦J u 100 Specific Activity - Time Curves ^"^— . / /^ ' r\ ' "^ \ • ^ \ "^^^ »\ ^ 1 \ ^^ .' \ ^ ' \ ^_ 1 \ ^ \ " ; \ '-1-- \ ^^^<- 1 \ ^ \ >» 1 \ ^ 1 \ ^v 1 X ^ * \ '^ 1 N. "^ \ "^ I \ X \^ ^. 1 ^^"""^^^^S4^i> 1 — —...^^^^ 240 10 20 160 30 TIME (DAYS) 40 50 80 ■r. « T^ J- ^- •^- / counts X molecular weight \ . Fig. 2. Radioactivities I ^^ — I of proto- \ 1000 I porphyrin and cholesterol digitonide obtained from a rabbit which had received 57 fxc ^U^MCU^.COOU and 100 /iC CHg.^^COOH. Right hand ordinate refers to the radioactivity of protoporphyrin, left hand ordinate refers to that of cholesterol digitonide. radioactivity changing in an approximately exponential fashion, whereas the activity of the protoporphyrin fell off linearly. It is also of interest that the maximum activity of the cholesterol was reached after 24 hours, whilst that of the porphyrin was obtained after seven days. Incubation of normal rabbit red cells at 37 °C. with labelled acetate showed that these cells can synthesize cholesterol in vitro. It would 74 A. Neuberger thus seem that part of the cell membrane, i.e. the hpid components, is metabohcally much more active than haemo- globin, even in the circulating erythrocyte. Exchange with the lipids of the plasma has been demonstrated for the red cell phospholipid (Hahn and Hevesy, 1939; Zilversmit, Enteman, Fischler and Chaikoff, 1943). Globin and Porphyrin It is now well established that eight methylene carbon atoms of glycine are utilized for the formation of protoporphyrin (Muir and Neuberger, 1950; Radin, Rittenberg and Shemin, 1950), four for the four methyne carbon atoms and four for one carbon atom in each of the four pyrrolic rings (Wittenberg and Shemin, 1950). Glycine is of course also used in the synthesis of the globin part of the hsemoglobin molecule. If the conversion of glycine into protoporphyrin takes as long as the incorporation of glycine into globin and if both chains of reactions utilize the same intracellular pool of glycine, the specific activity (expressed in molar equivalents) of the proto- porphyrin should be eight times that of the globin glycine. If one component is formed earlier or becomes metabohcally inert sooner, then the activity of that component should at first be lower than that of the other component which is formed later. However, provided the rate of red cell production is constant, the ratio of the molar activities of the protopor- phyrin to globin glycine should, as the activity-time curve flattens out, converge towards eight. We have tested these predictions both in the rabbit and in the rat. In the rabbit experiments fairly large blood samples were obtained from the same animal and the absolute values of the activities are therefore somewhat influenced by the anaemia produced. However, this is unlikely to affect the ratios of activities in which we were primarily interested. In the rat experiments crystalline haemoglobin from two animals was used for each time interval. In these experiments too the absolute values are not strictly comparable owing to Studies on Mammalian Red Cells 75 differences between animals, but again the ratios of activities may be expected to be significant. The results (Figs. 3 and 4, Table I) show that at least in experiments in which samples were taken 24 hours or earlier after the administration of the labelled glycine, the ratios of the molar activities of the porphyrin and of the globin glycine O 10 Specific Activity -Time Curves 10 20 30 TIME (DAYS) 40 50 Fig. 3. Radioactivities of glycine / counts X molecular weight 1000 ) /counts X molecular weight \ ,. . and protop orphynn I — — I obtamed irom \ JL vvl v/ /\ O y a rabbit which had received 57 fic NHg-^^CHg-COOH and 100 IXC CH3.14COOH. were fairly close to eight. The individual values varied between seven and nine, but the average was 7-9. However, during the first 24 hours this ratio was between 5*2 and 6-8. Such low values might suggest that the globin is formed somewhat later in the erythropoiesis than the porphyrin. Such an assumption would not contradict the findings of Thorell (1947) who established by spectrophotometric methods 76 A. Neuberger that formation of protein in the developing red cell precedes that of porphyrin. This protein may, at such an early stage, still be metabolically active, exchanging with the glycine stores of the body, or it may not even be globin. But it is possible that some of the reactions utilizing the methylene P0RPHYR»n/8 V 1-2 ^ GLOBIN GLYCINE 1-0 RATE OF RADIO- • / --^ ..^ -'' X ACTIVITIES 0-8 ^f'^' 0-6 STROMATINGL. ^ 1-2 GLOBIN GLYCINE / \ V 1-0 / \ RATE OF / RADIO- / ^\^ AaivrriES o-8 / ^ 06 ■ / 1 1 1 12 24 72 HOURS 14 22 DAYS 52 Fig. 4. Comparison of radioactivities of porphyrin, globin glycine and stromatin glycine from rat erji:hrocytes obtained at various times after simultaneous administration of the two labelled amino-acids. Broken line refers to results obtained from a rabbit. The porphjTin activities have been divided by eight. carbon atom of glycine for the synthesis of porphyrin are slower than the incorporation of glycine into globin and thus the isotope would appear at a later time in at least some of the eight positions of the porphyrin. Experiments to test this possibility are in progress. Studies on Mammalian Red Cells 77 Table I Radioactivities of Samples of Protoporphyrin Ester, Globin Glycine (DNP) AND StROMATIN GLYCINE (DNP) Time of collection after administration of labelled compound(s) 12 hrs. 72 hrs. 1 week 2 weehs Porphyrin cts./min. cts. Mol. wt. (g.) 162-5 ±2-84 ll-98±0-21 73-5 ±1-76 17-71±0-42 9-3 ±1-18 815±l-03 419-9 ±9-58 30-97±0-71 125-0 ±1-9 3015±0-46 50-92±206 40 08±l-65 297-8 ±8-02 21-97±0-59 104-8 ±1-45 25-26±0-35 30-05±l-28 23-2 ±0-99 375-6 ±1-47 27-7 ±011 136-0 ±3-2 32-78±0-77 24-57±0-84 25-0 ±0-86 8 • 1000 Globin-glycine (DNP) cts./min. cts. Mol. wt. (g.) 1000 Stromatin-glycin (DNP) cts./min. cts. Mol. wt. (g.). Dil. f. 1000 Investigations on Globin Synthesis In the experiments on rats to which reference has aheady been made, DL-vahne labelled with ^*C on the ^^m-methyl groups was injected together with the labelled glycine. The two amino-acids were isolated from crystalline haemoglobin as 2:4 dinitrophenyl (DNP) derivatives (Perrone, 1951). Valine residues supply the terminal amino groups of rat haemo- globin and terminal valine was therefore isolated by treating the intact protein with fluorodinitrobenzene followed by hydrolysis and chromatographic^ separation. The non- terminal valine and also glycine were obtained by similar treatment of the residual hydrolysate. Table II shows the radioactivities of the terminal and non- terminal valine residues in the haemoglobin at various intervals after administration of the labelled amino-acids. It appears that the two types of valine residue have exactly the same activity at any one time, the differences being less than the 78 A. Neuberger standard deviations of the counts. We have also compared the radioactivities of the glycine and valine at each point in time, making allowance for the ^*C dose of each amino-acid Table II Radioactivities of Valine Samples from Rat Globin Vj, = terminal vahne Vjjj=non- terminal valine Results are given in cts./min./cm.^ (infinite thickness) Time of collection Vb Vm 24 hours 72 hours 1 week 2 weeks 331 ±1-9 7419±3-49 58-12±l-78 102-67±6-23 3318±l-58 7605±306 59-45±118 10304±3-63 injected, viz. 5 juc./lOO g. body wt. of glycine and 3-5 /xc./lOO g. body wt. of DL-valine. Fig. 5 shows that the ratio of molar radioactivities of glycine and of valine, corrected for differences 14 12 10 GLYCINE VALINE _i_ \Z 24 HOURS 72 I 2 WETEKS Fig. 5. Comparison of molecular radioactivities of glycine and valine obtained from rat haemoglobin at various times after simultaneous administration of the two amino-acids. Results have been corrected for differences in dosage as discussed in the text. Studies on Mammalian Red Cells 79 of dosage of ^^C, remains reasonably constant at 1-24 ±0-2' A difficulty in interpreting this result arises from the fact that the injected valine was racemic and the exact proportion of the D-isomer which is inverted and becomes available for protein synthesis is therefore unknown. That inversion can take place in the case of leucine has been shown by Ratner, Schoenheimer and Rittenberg (1940) and the same presumably applies to valine. However, the feeding experiments of Rose (1938) show that such inversion may not be extensive. It would thus appear that the average ratio of activities of glycine to that of total available L-valine might be between 0-62 and 1-00. The finding that the ratio of the activity of the terminal valine to that of non-terminal valine is unity at all points examined is of general interest. It suggests that both types of valine residue are built into peptide chains of haemoglobin at approximately the same time. In other words, it appears that the rate of synthesis of valine into terminal and non- terminal positions is fast compared with the rate of change of labelled valine in the body. A comparison of the rate of incorporation of two different amino-acids is more complex, since it involves in addition to the rates of peptide formation, rates of turnover of free amino-acids in the organism and possible permeability differences. The ratio of the molar activity of glycine to that of valine was more variable than the corresponding ratio for terminal and non-terminal valine. However, the differences were relatively small and irregular and it seems therefore that the rates of turnover of these two amino-acids in the rat are very similar. It is also reasonable to conclude from these results that the rates of incorporation of glycine and valine into haemoglobin are similar. Altogether these findings indicate that haemoglobin is formed by simul- taneous condensation of amino-acids or by a process of rapid successive condensations involving intermediates of relatively short life. The results also show that the labelled glycine and valine are diluted in the body to approximately the same extent. 80 A. Neuberger the dilution of glycine probably being slightly larger. This fact seems at first surprising since glycine, unlike valine, is synthesized by the animal. It does therefore follow that the dilution of labelled glycine is largely due to the rapid rate of protein turnover leading to the liberation of both essential and non-essential amino-acids. Stroma Proteins We have also isolated glycine from the washed ghosts of the red cells. The insoluble fraction thus obtained is a mix- ture which still retains some haemoglobin. Results (Table I and Fig. 4) show that there is no clear correlation between the radioactivity of the glycine in this fraction and that of the glycine in haemoglobin. The relatively low count at 12 hours suggests that at least some of the stroma protein is formed at an early stage in erythropoiesis. Summary 1. Measurements of the life span of the human red cell by the agglutination method of Ashby and by the isotopic labelling of haemoglobin have been compared. Certain apparent discrepancies are discussed and the possibility is considered that the haemoglobin in the mature human red cell may not be completely inert, or that there is transfer of haem or haemoglobin between circulating cells of different ages. 2. Experiments reported earlier on the life-span of the rabbit red cell indicate that these erythrocytes have normally a shorter life span of 50 days. As an alternative the possibility is considered that the haemoglobin of the rabbit red cell is particularly labile and that the isotopic labelling of the haemo- globin does not provide a valid method of measuring the life span of the cell. 3. Experiments are discussed which show that the choles- terol of the rabbit red cell is metabolically active, with a half-life of 12-14 days, and can be synthesized in vitro. 4. It was found that both in the rat and the rabbit the ratio of molar radioactivities of the porphyrin and glycine Studies on Mammalian Red Cells 81 isolated between 12 hours and two weeks converge to a value of eight, the earliest samples showing ratios of 5-7. 5. In these experiments radioactive DL-valine was also administered and valine was isolated from terminal and non- terminal positions of the globin. The radioactivities of the two fractions were identical for each point in time. The radioactivities of glycine and valine were also compared and varied by not more than 20 per cent. These results suggest that the haemoglobin molecule is synthesized either by syn- chronized condensation of amino acids, or by a rapid succes- sive formation of peptide bonds, without any appreciable accumulation of intermediates. The experimental work reported was done mainly by my colleagues Dr. Helen Muir and Dr. J. C. Perrone. REFERENCES AsHBY, W. (1919). J. exp. Med., 29, 267. Benard, H., Gajdos, a., and Mme. Gajdos-Torock (1950). C.R. Soc. Biol, Paris, 144, 350. Callender, S. T., Powell, E. O., and Witts, L. J. (1945). J. Path. Bact., 57, 129. Gray, C. H., Neuberger, A., and Sneath, P. H. A. (1950). Biochem. J., 47, 87. Hahn, L., and Hevesy, G. (1939). Nature, Lond., 144, 204. JoPE, E. M. (1946). Brit. J. industr. Med., 3, 136. London, I. M., Shemin, D., West, R., and Rittenberg, D. (1949). J. biol. Chem., 179, 463. London, I. M., West, R., Shemin, D., and Rittenberg, D. (1950). J. biol. Chem., 184, 351. Muir, H. M., and Neuberger, A. (1950). Biochem. J., 47, 97. Muir, H. M., Neuberger, A., and Perrone, J. C. (1951) (unpublished). Muir, H. M., Perront:, J. C, and Popjak, G. (1951). Biochem. J., 48, iv. Neuberger, A., and Niven, J. S. F. (1951). J. Physiol, 112, 292. Perrone, J. C. (1951). Nature, Lond., 167, 513. Radin, N. S., Rittenberg, D., and Shemin, D. (1950). J. biol. Chem., 184, 745. Ratner, S., Schoenheimer, R., and Rittenberg, D. (1940). J. biol. Chem., 134, 653. Rose, W. C. (1938). Physiol. Rev., 18, 109. Shemin, D., and Rittenberg, D. (1946a). J. biol. Chem., 166, 621. Shemin, D., and Rittenberg, D. (19466). J. biol. Chem., 166, 627. ISOTOPES 7 82 A. Neuberger Thorell, B. (1947). Acta med. Scand., 129, Suppl. 200. Wittenberg, J., and Shemin, D. (1950). J. biol. Chem., 185, 103. ZiLVERSMiT, D. B., Enteman, C, Fishler, M. C, and Chaikoff, I. L. (1943). J. gen. Physiol., 26, 333. DISCUSSION Thorell: As a consequence of earlier microspectrographic studies on endocellular haemoglobin formation, we have been interested in the time sequence of the formation of globin and haemin during erythro- % immature in blood e marrow Bon mitosis frequency 00 \<5-10"'2 Hb 20 10 t t tt Hemolysis Hen I hou rs Fig. 1. Haematological changes in hen after injection of phenyl- hydrazine. Abscissa: time in hours. The curve with circles represents the mitosis frequency in the bone marrow. The dotted curve shows the relative amount of erythroblasts in blood, containing ca. 5 per cent haemoglobin as determined micro- spectrophotometrically. The full curve represents erythro- blasts with no haemoglobin. The arrows indicate the injections of labelled glycine. poiesis. I would like to show one of the experiments done in collabora- tion with Dr. Hammarsten and his group. We have used hens, and in order to get a rapid formation of haemo- globin, the hens were made anaemic by haemolysis with phenylhydrazine. After this treatment a rapid regeneration of red cells could be observed (Fig. 1). As shown by cytological analyses, during this condition the erythropoiesis is divided into two parts: the growing, dividing cells are Studies on Mammalian Red Cells 83 in the bone marrow and the haemoglobin synthesis proceeds in the blood stream. We gave two hens the same dose of ^^N-labelled glycine. Hen I was given intraperitoneal injections early in the regeneration and was killed 48 hours later (Fig. 1). At the time of killing the blood cells have thus been under influence of labelled glycine during their growth period in the bone marrow. Hen II was given labelled glycine 12 hours before killing, during the period of maximal amount of immature cells in the blood (Fig. 1). These blood cells thus have been under the isotope influence only a short time during their haemoglobin synthesis. Table I shows some values of ^^N excess in the hsemin and globin from the blood of the two hens. Hen I, which was given labelled glycine during an early stage of regeneration, has a high incorporation of Table I Blood from Hen I and Hen II. Relationship between ^^N Excess in THE Glycine (Chromatographically Isolated from the Globin) and in H^MiN (Crystallized) Atom per cent excess ^^N in glycine isolated from globin Atom per cent excess ^W in hcemin Quotient Hen I . Hen II . 1-61 <1 1-78 4-2 •9 <-2 i^N in the glycine isolated from the globin and also fairly high incorpora- tion in the haemin. The second hen, whose blood cells were under the influence of labelled glycine only a short time during a period of intense haemoglobin synthesis, shows high incorporation of ^^N in the haemin but considerably lower incorporation in the globin. I think that a reasonable conclusion from this experiment might be that during blood cell regeneration in these anaemic hens there is a difference in time between the formation of the globin and the porphyrin during haemoglobin synthesis. Neuberger: One wonders how far these experiments can be compared with experiments done on mammalian red cells without a nucleus. You might get quite a different time relationship of various processes in nucleated cells compared with non-nucleated cells. The other compli- cation is that an anaemia, probably a fairly severe anaemia, w^as pro- duced and that is not strictly comparable to our experiments, in which the haemoglobin content was normal. Wormall: I gather that Dr. Neuberger is anxious to find some other label for the red cell, and that he wasn't quite happy about the labelling of the cholesterol in comparison with the labelling of the haemoglobin. I commend to his notice radioactive zinc, ^^Zn. Tupper, Watts and I have shown in recent times that zinc goes into the red cell fairly quickly and combines with a variety of compounds, though it doesn't enter very 84 A. Xeuberger quickly into the carbonic anhydrase. If the labelHng is done bio- logically over a longer period, however, the radio-zinc will go into the carbonic anhydrase. It may be possible, using ^^Zn, to get something for comparison with haemoglobin; that is, if you feed the zinc to a rabbit and separate the carbonic anhydrase from the red cells you may get something which is very near your haemoglobin value. You will, however, have to exclude other zinc compounds, from which the zinc can be dissociated fairly readily. They may be zinc proteins. I under- stand from Dr. B. L. Vallee that he has recently obtained a new zinc protein from human leucocytes, and my colleagues and I believe there may be certain other zinc compounds besides carbonic anhydrase in the red cell. That is one of the points you will have to consider if you use 6 5Zn for labelling red blood cells. I believe you are still worried about that 15 per cent of activity which remains in the haemoglobin after about 200 days. It might help to solve that problem if you injected ^^N-labelled haemoglobin. Then you might get some idea of how that haemoglobin is used. Neuberger: I think one obvious way of investigating this is to take red cells from one human being, label them and then transfuse them into another experimental subject and see how the isotope content decays. We w'ould have to use a large amount of glycine and it would be rather expensive. Rittenberg: The curve obtained by plotting the ^^N content of the haem against time in normal man, which Dr. Neuberger showed at the beginning of his paper, has attracted our attention for quite a time. There are certain difficulties about this curve. The curve does not approach zero sufficiently rapidly. It is not too clear whether we are dealing with a small re-utilization of the haemoglobin, or w-hether the zero point for the ^^N curve is raised. In other words, you may have changed the "normal abundance" in this subject so that to obtain atom per cent excess you should subtract from the absolute ^^N con- centration a value greater than 0-368 per cent ^^N. I think the plasma protein labelling at this time was somewhat less than the haemoglobin, being about • 060 atom per cent excess. Although of course there may be reutilization, I don't believe it is really very great. Dr. London is now studying the question that Dr. Wormall just raised, the utilization of haemoglobin as such. As far as the middle portion of the curve is concerned, interpretation is difficult because there is nothing you can do about bringing the level of the precursor down rapidly. What you really ought to do is work with a precursor whose concentration went up sharply and then rapidly declined to zero. Since this has not been done you have to go through some complicated arithmetic. The "plateau" certainly goes down, and this downward trend may be due in part to a little extra synthesis or com- pletion of haemoglobin synthesis in the circulating red cells. We feel that this is rather small. The major cause for the downward slope of the plateau is probably a small random destruction of the circulating red cells. WTien we studied sickle cell anaemia and plotted the data semi- exponentially, we got a curve that seemed to be the sum of two straight Studies on Mammalian Red Cells 85 lines. The initial part of the curve falls rather rapidly. It is possible that in the normal system you also have a component in the blood which cannot stand the hurly burly and vicissitudes of circulation and breaks down very rapidly. This may explain the initial part of the plateau. Another difficulty comes when you consider the bilirubin. Its isotope concentration does not fall to zero between the initial peak and the second peak. This is not in accord with the simple theory. The situation is more complicated than the simple mathematical derivation that we presented, but I don't believe that it is worth trying to improve the theory without new and better data than is now available. The curves resulting from transfusion of red cells which Dr. Neuberger showed you have always puzzled me. They shouldn't look that way. As he pointed out, they show a negligible scatter, and we're almost sure, not only from experimental work, but intuitively, that they shouldn't be that way. They should have some sort of curvature at the beginning and the end, and they don't have it. The interpretation of this type of data may be more difficult than the interpretation of the ^^N data. Mollison: There is an apparent discrepancy between the results of experiments with i^N and transfusion experiments. The experiments with i5]s; suggest that a considerable proportion of red cells Uve for less than the average life and that a considerable proportion live on for many weeks after the average life. On the other hand, transfusion experiments suggest that there is very little variation in life-span between one red cell and another. Dornhorst* (1951) has pointed out that the standard deviation of the life-span of red cells can be de- duced from the number of transfused red cells surviving at the end of the average life-span. When the number of cells surviving at different intervals after transfusion is plotted, the points fall on a straight line, and extrapolation of this line to the time axis gives an estimate of the mean cell life. However, the curves often show a small tail at the end so that a small proportion of red cells are still present at the end of the mean cell life and it is this number which is related, mathematically, to the standard deviation of the life-span. Personal observations suggest that the standard deviation may be as small as 5 days. If this is true, the ^^N curves relating to red cell life-span must be re-interpreted, as Dr. Neuberger has said. *Dornhorst, A. C. (1931). Blood, to be published. "Interpretation of red cell survival curves." PRELIMINARY INVESTIGATIONS FOR A STUDY OF ENERGY UTILIZED BY THE SURVIVING FOWL ERYTHROCYTE IN H^M SYNTHESIS C. RIMINGTON From the work of Rittenberg, Shemin, Altman and others in the field a fairly clear picture is emerging of the origin of the various constituent atoms of the porphyrin ring of hgem and this work has made possible speculations as to the actual molecules involved in biosynthesis. One need only refer to the schemes put forward by Lemberg and Legge (1949), Wittenberg and Shemin (1950), and Neuberger, Muir and Gray (1950). That hsem biosynthesis can take place in surviving erythro- cytes in vitro is a discovery of first importance and makes possible an attempt to investigate the system contributing the necessary components and energy for this synthesis. I wish to report some preliminary experiments in this direction carried out with Mr. Bufton some time ago. Heparinized fowl blood was used, incubated with ^^N labelled glycine in presence of streptomycin and penicillin for 24 hours at 37°, the isolated hsemin being eventually analysed for ^^N content. The effect w^as first studied of increasing amounts of added 15^ glycine. Maximal incorporation was achieved at a level of about 2 mg./ml. of blood suspension, in agreement with Shemin's (1948) finding, and this quantity was therefore used throughout subsequent experiments. Since experiments upon erythrocytes, in which inhibitors or nutrilites are added to the system, are complicated by the presence of the cell membranes, the effect of haemolysis was next studied. When sufficient saponin was added to produce 86 Energy Sources for H^m Synthesis 87 nearly complete haemolysis, the incorporation of ^^N glycine was markedly inhibited (Table I). Table I Effect of Hemolysis on Synthesis of Hjem by Fowl Erythrocytes ^^N atom per cent excess in hcemin Control Hcemolysed 101 105 » 0019 030 It was clear therefore that intact red cells would have to be used and that any test in which haemolysis occurred would have to be interpreted with great caution. The most obvious possible source of energy for haem syn- thesis within the erythrocytes is the metabolism of glucose, but when extra glucose was added to the suspending medium, no increase in ^^N incorporation occurred. This would be easily explicable were the quantity of carbohydrate available to the cell already sufficient to supply all energy needs. Cells were therefore washed and resuspended in dialysed fowl plasma. Unfortunately some haemolysis could not be pre- vented during the subsequent incubation but it appeared that the addition of glucose was without any stimulating effect. A similar experiment was carried out with acetate and a negative result obtained in this case also (Table II). Table II Addition of Glucose or Acetate to Washed Fowl Erythrocytes in Dialysed Plasma li 'N atom per cent excess in hcemin - Control Glucose (10 mg.) Na Acetate (10 mg.) 056 054 055 060 0-058 065 Attempts were then made to interfere with the course of carbohydrate metabolism by the use of the well known 88 C. RiMINGTON inhibitors, but in no case was significant interference with ^^N incorporation demonstrable. The substances used were: — Sodium fluoride 10-^ molar final concentration. Sodium iodoacetate 10"* molar final concentration (possibly a slight depression). Sodium malonate 10'^ molar final concentration. 2:4-dinitrophenol (10"^ molar) used to "uncouple" phos- phorylation processes, was similarly ineffective with or with- out added glucose. It would appear therefore either that the inhibitors used did not penetrate in sufficient concentration to the enzyme system providing energy for synthesis from glucose meta- bolism, or that haem synthesis derives energy from sources other than carbohydrate metabolism. That permeability was the limiting factor seems unlikely in view of the fact that these same substances behave as effective inhibitors of the intracorpuscular reduction of methaemoglobin to haemoglobin, a process which is believed on good evidence (Kiese and Schwartzkopf, 1947) to be linked with the metabolism of carbohydrate within the cell. We are doing further work upon this problem. Another observation which I would like to mention is that the biosynthesis of hgem by this isolated red cell system can be markedly inhibited by addition of lead salts (Table III). Table III Inhibitory Action of Pb Acetate on H^m Synthesis by Fowl Erythrocytes ^^N atom per cent excess in hceinin Control Ph Acetate 10-*M (Ringer-Tyrode) 141 134 014 0013 Control Pb Acetate 10*M (Saline) 089 084 042 048 Energy Sources for H^m Synthesis 89 The action of lead might be either an inhibitory one upon some essential enzymic system supplying either energy or a building stone of the porphyrin ring, or alternatively lead ions might be acting by blocking the incorporation of iron into the ring system. There is some evidence from work on lead poisoning to support this view, which I first put forward in 1936 (Rimington, 1936). The isotope technique may offer a means of deciding between these alternative explanations of the inhibitory action of lead and the problem is under further active investigation. REFERENCES KiESE, M., and Schwartzkopf, W. (1937). Arch. exp. Path. Pharmak., 204, 267. Lemberg, R., and Legge, J. W. (1949). Hcematin Compounds and Bile Pigments. New York: Interscience. Neuberger, a., Muir, H. M., and Gray, C. H. (1950). Nature, Lond.y 165, 948. Rimington, C. (1936). Onderstepoort J. vet. Sci., 7, 567. Shemin, D. (1948). Cold Spr. Harb. Sym. quant. Biol., 13, 185. Wittenberg, J., and Shemin, D. (1950). J. hiol. Chem., 185, 103. DISCUSSION Shemin: We did get marked inhibition with malonate and respiration was inhibited about 70 per cent at a concentration as low as 10-^. All the figures have not been published on it. Anaerobic conditions give no synthesis. We haven't tried all the poisons, but most of them inhibit the synthesis. Rimington: In view of w hat Dr. Shemin told us about the mechanism of the synthesis, one would expect an inhibition by malonate. Wood: It seems to me that the extent of inhibition that one would observe would depend to a great extent on the relative rate of synthesis as compared to the rate of respiration. In other words, if the rate of porphyrin synthesis were relatively slow as compared to the rate of formation of C4 acids by respiration, then you would only see the effect of malonate at the time when respiration was reduced to the point where it became the rate limiting reaction. It is rather difficult in this type of experiment to know whether you have reached that point or whether you have not. Sloviter: I wonder if Professor Rimington has any notion why haemolysis inhibits the haem synthesis. Rimington: I'm afraid I haven't. Of course, one could do many things here, for example, add ATP, or inhibit ATPase and so on. I wouldn't like to speculate any further on that point. IRON METABOLISM IN PATHOLOGICAL CONDITIONS A. VANNOTTI In order to studv the intermediate metabolism of iron in normal and pathological erythropoiesis, we have followed the distribution of radioactive iron, ^Te and ^^Fe in the tissues of the rabbit. It was necessary to make an exact computation of radio- active iron, and to study its distribution in the organs half an hour after the injection of a solution of iron lactate (ferric and ferrous lactate). A maior difficult v lav in the method of calcination of the complete animal, which produces a large quantity of ashes which absorb a great proportion of radioactivity: liver ash, 25 per cent; kidney ash, 21 per cent; muscle ash, 17 per cent; bone ash, over 100 per cent. After many trials we have adopted the following method: damp calcination of tissues by Kjeldahl's method, using sulphuric and perchloric acid; precipitation of the excess sul- phuric acid vv'ith titanium trichloride; acidification with concentrated hydrochloric acid (pH=2); re-oxidization of iron chloride with oxvs^enated water; extraction of iron with ether; addition of hvdrochloric acid; distillation of this solu- tion, and precipitation of iron in the form of ferric hydrate; calcination. Iron oxide is measured with a Geiger Miiller counter (very vreak absorption of radioactivity by the small quantities of FcgOg). This method has enabled us to recover 80-90 per cent of injected radioactive iron. A preliminary examination carried out on a normal animal showed that distribution of iron depends partly, during the first hours after injection, on the valence and the quantity of injected iron (200-2,000 /xg.). If large quantities of iron 90 Pathological Iron Metabolism 91 are injected, the organism attempts to react against the invasion of iron by eUminating it through the kidneys and fixing it transitorily in deposit organs hke the hver or in tissues such as skin and muscles, which yield it back gradually as the blood level of iron falls. The liver is the principal organ of iron storage. The spleen does not participate notably in this function, but on the other hand takes up iron after haemolysis. More than 20 per cent of the iron is circu- lating in the blood Ij hours after injection. Skin, muscles, and bone marrow retain iron, whilst kidneys and intestine excrete a certain quantity of the metal. These observations confirm the results reported in numerous papers by American authors on iron metabolism (Hahn, Balfour, Whipple, etc.). We then followed the distribution of iron in pathological erythropoiesis, engendered by anaemia due to repeated bleed- ings, by maintaining the animal in a state of oxygen tension corresponding to an altitude of 18,000 ft., by hyperthyroidism, or by septicaemia. Table I gives the percentage of radioactive iron determined in different organs IJ hours after injection (500 />tg.). In acute ancemia, the quantity of circulating iron is slightly increased, the liver and skin contain strikingly less iron, whilst iron in the bone marrow increases four or fivefold. This experiment emphasizes the intense avidity for iron of the bone marrow when there is increased erythropoiesis. In altitude hypoxemia (18,000 ft.), values are encountered which approach those obtained in acute anaemia. Circulating iron is slightly higher than in normal subjects and a high proportion of the iron is fixed in the bone marrow, and also in muscles, myocardium and nervous system. Apart from the intense erythropoietic stimulation due to hypoxaemia, we find a certain degree of fixation of iron in organs possessing a high respiratory activity. The same phenomenon is also observed in hyperthyroidism, where erythropoietic organs and also heart muscle, and to some extent the liver, fix a significant quantity of iron. 92 A. Vannotti o c -^ c: ?c c; -c JJ l> c: n + c + ■^ cc C: 'f QC r- O d CO c o fe, i^ 01 r- o ■c rl t ^ ecr- — rccrr">i f^ ClMCt-'^wCCOOO 6 ►i5 01 i^ r^ c T-( C C^l O X CO X »-'^ (M Ci O 5<« + + + ccc;s^CiOC5»r:c-H »r: "^ !© (M c c C'^CCf^COCCCwC ►«H CC ri r^ rl c i> c^j -^ -* i> in i-'^ w CO li^ »n + + + b-CCCi— iXi— (CCi— cO c c C0Xi-Hi>i-i;2CCr-iXC o "4^ CO ^ 01 o •r* r— ( s 6i S l>'*'*xeo©ccc:i>t-LO + + •^■^ccCi-^Xr-ccc:!-! c c eOCO:Cl>rHr-CCr-iXC «( M 01 rl c r-i »-O;OXCC:CCC0CC:C + + + XCiwrHXr^XCXwi— c c CCO'^LOt-XCCCt-C »•*« &< 01 T? o c )— 1 g 1 l>l^COCir- > K iO-1, ,-■£, _-»■• — r- 30 T- 60 OOSE(kr) — I— 90 Fig. 3. Survival of E. coli in standard buffer and . 04 M BAL. Burnett, Stapleton, Morse and Hollaender (1951), in press. pletely additive in their protective capacity. The greater survival at optimum concentration for BAL might be explained by the fact that the parent compound of BAL, propanol, is also somewhat protective or that BAL contains two sulphydryl groups (Burnett, Stapleton, Morse, and Hollaender, 1951). 102 HOLLAENDER, StAPLETON AND BuRNETT Doermann (personal communication) in our laboratory found that there is a linear relation between concentration of glycerol and its protective capacity. More complete investigation of the glycols as well as more simple alcohols 1000- PROTECriVE ACTION OF SH COMPOUNDS 8AL CYSTEINE X-RAY DOSE • 60 KILOROENTGENS A.2'C BROTH-GROWN CELLS o.odial 0005 0.0026 O.Ol 0.02 MOLAR CONCENTRATION 004 Fig. 4. Survival of E. coli in buffer suspension of different concentrations of BAL and cysteine. Burnett, Stapleton, Morse and Hollaender (1951). has revealed that all of them have essentially similar charac- teristics both in regard to the concentrations required to give detectable protection as well as the degree of protection afforded by optimum concentrations. Concentrations required to give protection are of the order of a hundred times that required for the sulphydryl compounds, as shown in Fig. 5. X-Ray Sensitivity 103 It should be pointed out that some of the higher concentra- tions of alcohols used here are toxic under conditions other than those employed in these experiments, where the tempera- ture was always held near 0°C. At low concentrations of PROTECTIVE ACTiON OF ALCOHOLS •0--^ I0--4- • METHANOL 9 ETHANOL O PROPANOL-2 60,000 r AIR -v2'C 8R0TH- GROWN CELLS I 2 MOLAR CONCENTRATION Fig. 5. Survival of E. coli in different concentrations of methyl, ethyl and iso propyl alcohols. Burnett, Stapleton, Morse and Hollaender (1^51). alcohols, i.e. less then 1 m, these compounds become pro- tective if the bacteria are incubated in their presence for 30 minutes at 37°C. (Fig. 6). It is of interest that sodium hydrosulphite (NagSaO^), a strong reducing compound, gives protection in low concen- 104 HOLLAENDER, StAPLETON AND BURNETT t rat ions (see Fig. 7). Moreover, sodium hydrosulphite affords some protection in nitrogen-saturated suspensions. This compound appears to protect by depletion of oxygen from the suspensions, possibly by removing bound oxygen from wo- 400- o 3O0- o > w $200- 3 100- ETHANOL 2' -| 1 T- 0,2 03 04 MOLAR CONCENTRATION —r- 06 Fig. 6. Relative survival of E. coli in different concentrations of ethyl alcohol at 2»C. and 37«C. Stapleton, Billen and Hollaender (in preparation). the bacterial cells (Morse, Burnett, Burke, and Hollaender, in preparation). Some of the carboxylic acids which are normal metabolic intermediates or end products have also proved to afford excellent protection to bacterial cells, but again, under rather restricted conditions. A short incubation period for the cells in the presence of the acids produces a striking increase in i X-Ray Sensitivity 105 protection (Fig. 8). The possible mechanism is a depletion of oxygen from the suspension by the bacterial cells in the presence of an oxidizable substrate (Stapleton, Billen, and Hollaender, 1951). < a > •o-i" f^^ t 10 -4 Hi 1 BUFFER CONTROL ^1 -> — r- \0' T 1 1 < 1 1 1 I0-' 10-2 MOLAR CONCENTRATION NAjSjO^ 10-* -T — i 1.0 Fig. 7. Survival of E. coli in different concentrations of sodium hydrosulphite at 2^0,. (60,000 r). Morse, Burnett, Burke and Hollaender (in preparation). The survival curves beginning with 15,000 r are linear over the entire range studied. At lower energy values for aerobi- cally grown bacteria irradiated anaerobically and for bacteria grown in the presence of protecting chemicals the survival curves may appear to be sigmoid in shape. Unfortunately this is a region where it is not simple to obtain highly reliable values. Such data are not available at the present time. In 106 HOLLAENDER, StAPLETON AND BURNETT contrast to this, the sigmoid shape of the anaerobically grown and irradiated organism is quite well established. Tests were conducted to determine if compounds of the different groups were additive in their protective action. We have already mentioned that BAL and cysteine, given at T" 10' 10' lO'' MOLAR CONCENTRATION FORMATE 10 Fig. 8. Survival of E. coli in different concentrations of sodium formate after incubation at 37"C. for 30 minutes. Stapleton, Billen and Hollaender (1951). concentration at plateau level, would not supplement each other in protective ability. In contrast to this, combinations such as alcohol and BAL, sodium hydrosulphite and BAL, and ethanol and hydrosulphite do give additive protection, as indicated in Fig. 9. However, when all three types of compounds were used simultaneously, we did not get signifi- cantly more protection than from two types. It is possible X-Ray Sensitivity 107 that no additional protection can be afforded by the third chemical because the residual killing effect is the direct damage produced by X-rays on essential components of the cell. PER CENT SURViVAL-60 KILOROENTGENS / 41 / PHOSPHATE BUFFER CONTROL \ 40 N II Eton / / 3 \ 9 \ 1 BAL |- Na2S204 * 3 / 18 18 Fig. 9. Compound graph of additivity tests of: Ethyl alcohol plus BAL. Ethyl alcohol plus Na2S204. Na2S204 plus BAL. Ethyl alcohol plus BAL plus Na2S204. The figures immediately below and above the names of com- pounds refer to survival obtained for the individual compound. Discussion It is difficult, on the basis of our meagre knowledge of the mechanism of the effect of ionizing radiation on cell con- stituents, to give a plausible explanation for the protection afforded by this fairly wide variety of compounds against radiation damage. Studies of the effects of X-rays on pure water are of interest in this connection. Many reports indicate the production of radicals like H, OH, HO2, and of HgOg (Allen, 1948; Dale, 1947; Burton, 1947; Stein and Weiss, 1948; 108 HOLLAENDER, StAPLETON AND BuRNETT Bonet-Maury and Lefort, 1948; Weiss, 1944, 1946; Fricke, 1935; Dainton, 1948; Lea, 1947). The influence of X-rays on certain organic chemicals and enzymes is now being studied in many laboratories. Very little is known of the interaction of the products of radiation with macromolecules and struc- tures like chromosomes, mitochondria, etc., in the living cells. One explanation for the effect of sulphydryl compounds is that they compete with or replace the sulphydryl groups of certain enzymes which are very sensitive to X-rays. This has support in Barron's (1949) work, where the special sensitivity of sulphydryl-containing enzymes has been emphasized. There seems to be little doubt that sodium hydrosulphite protects by its ability to tie up oxygen. In addition, it may possibly remove oxygen from inside the cells. The protective ability of alcohol may be partly due to the fact that, at high concentration, it can tie up oxygen. After incubation at low concentration it may entet the metabolic cycle where, in the course of oxidation, a considerable amount of oxygen may be utilized. The possibility exists that metabolic compounds may replace basic cell constituents destroyed by ir- radiation. The utilization of formate in nucleic acid synthesis may explain in part the protective action of this compound. Moreover, the excellent protection afforded by carboxylic acids, w^hen incubated with cells prior to irradiation, suggests also the possibility that these compounds may supply some essential intermediate or overcome a block in the carboxylic acid cycle induced by ionizing radiations. Interference with some step in the Krebs citric acid cycle, brought about by X-rays, is indicated in the report by DuBois, Cochran, and Doull (1951). This suggests the use of some of these com- pounds as protective agents against ionizing radiation damage in mammals. Some aspects of this problem are now being investigated in our laboratory. REFERENCES Allen, A. O. (1948). J. phys. Colloid Chem., 52, 479. Anderson, E. H. (1951). Proc. nat. Acad. Sci., Wash., 37, 340. Bacq, Z. M. (1951). Experientia, 7, 11. X-Ray Sensitivity 109 Baker, W. K., and Sgourakis, E. (1950). Proc. nat. Acad. Sci., Wash., 36, 176. Barron, E. S. G., et al. (1949). J. gen. Physiol., 32, 537. Bonet-Maury, p., and Lefort, M. (1948). Nature, Lond., 162, 381. Burnett, W. T., Jr., Stapleton, G. E., Morse, M. L. and Hollaender, Alexander (1951). Proc. Soc. exp. Biol. Med., in press. Burton, M. (1947). J. phys. Colloid Chem., 51, 611. Crabtree, H. G., and Cramer, W. (1933). Proc. Roy. Soc. B., 113, (1933). 113, 238. Dainton, F. S. (1948). J. phys. Colloid Chem., 52, 490. Dale, W. M. (1947). Brit. J. Radiol. Suppl. 1. Du Bois, K. P., Cochran, K. W., and Doull, J. (1951). Proc. Soc. exp. Biol. Med., 76, 422. Fricke, H. (1935). J. chem. Phys., 3, 364. Gaulden, M. E., and Nix, M. (1950). Genetics, 35, 665. Giles, N. H., Jr., and Riley, H. P. (1949). Proc. nat. Acad. Sci., Wash., 35, 640. Giles, N. H., Jr., and Riley, H. P. (1950). Proc. nat. Acad. Sci., Wash., 36, 337. Hollaender, A., Stapleton, G. E., and Martin, F. L. (1951). Nature, Lond., 167, 103. HoLTHUSEN, H. (1921). Pfliig. Arch., 187, 1. Lea, D. E. (1947). Action of Radiations on Living Cells. New York: The Macmillan Co. Morse, M. L., Burnett, W. T., Jr., Burke, A. W., and Hollaender, Alexander. (Manuscript in preparation.) Stapleton, G. E., Billen, Daniel, and Hollaender, Alexander. (Manuscript in preparation.) Stapleton, G. E., and Hollaender, A. (1951). Bacteriolog. Proc, 62. Stein, G., and Weiss, J. (1948). Nature, Lond., 161, 650. Thoday, J. M., and Read, J. (1947). Nature, Lond., 160, 608. Thoday, J. M., and Read, J. (1949). Nature, Lond., 163, 133. Weiss, J. (1944). Nature, Lond., 153, 748. Weiss, J. (1946). Nature, Lond., 157, 184. WiTKiN, E. M. (1946). Proc. nat. Acad. Sci., Wash., 32, 59. WiTKiN, E. M. (1947). Genetics, 32, 221. DISCUSSION Gray: Clearly, as Dr. Hollaender has suggested, there are two possible protection mechanisms which we must somehow try to separate: chemical protection, which intercepts the radicals as they are formed along the track of the particles; and a mechanism which operates by changing the cell in its reactions to the radicals that are produced. I would like to ask Dr. Hollaender if he has any protection data relating to other types of ionizing radiation, particularly the more densely ionizing particles, such as neutrons and alpha particles, which produce what seem to be identical forms of cytological and other damage to the 110 HOLLAENDER, StAPLETON AND BuRNETT cells, though in quantitatively different amounts. Read and Thoday showed that both as regards gro\\'th inhibition and chromosome damage in root meristems, anaerobiosis affords a roughly 3-fold protection against X-radiation, exactlj^ in line with Dr. Hollaender's results, but no appreciable protection against a-radiation. Mr. Boag and I recently made some experiments with (D-D) neutrons which seem to show little or no effect of anaerobiosis. It would therefore seem that the removal of oxygen in these experiments affects the chemicals formed along the track rather than the condition of the receptor molecules. Unfortunately this isn't absolutely conclusive, because we also know that the chemical effects produced along an alpha particle or a proton track are not identical with those produced along an electron track. Perhaps Dr. Hollaender might be able to say something about this aspect of the problem. Hollaender: The experiments of Read and Thoday on the effects of alpha particles and neutrons were repeated, using different organisms, in our laboratory. We got practically the same results. Giles' work (unpublished) with Tradescantia has shown that neutrons are inter- mediate between alpha and X-rays; oxygen has slightly more effect on neutron than on alpha ray damage, but not enough to be very im- pressive. As yet we have no reportable results on chemical studies. Gordon: I was wondering whether 1,2-glycols have been tried for protective action, in view of the fact that BAL, with two adjacent mercapto groups, has a greater effect than cysteine. Hollaender: No, we haven't tried it. This is on our list. The alcohols and the glycols are about one hundred times less protective than BAL and cysteine. Gordon: Would the protection of a glycol be twice that of a mono- hydric alcohol, or would it be greater than the molar proportions of hydroxyl groups would indicate? Hollaender: As far as we know, propylene and triethylene glycol are about as effective on the basis of molar concentration as the lower alcohols. Holmes: Do you have any idea how much of these compounds, particularly hydrosulphite, we can give to animals? If you can tell me a safe dose to give a rat, I can try it on a nucleic synthesis during the next week or two. I should like to see if it prevents the inhibition by X-rays of nucleic formation. Hollaentder: About 12-5 mg. per mouse, given intraperitoneally or intravenously. A dose of 25 mg. per mouse is toxic. Sloviter: Have you tried glucose? I have seen reports of hyper- glycaemia giving protection against large doses of radiation. Hollaender: If E. coli is incubated in the presence of glucose, the pattern of metabolism is changed and some protection is obtained. If bacteria are grown anaerobically in the presence of glucose, they become very resistant, as we have shown. Stapleton, in our laboratory, has worked out the resistance in aerobically grown organisms throughout the groA\'th cycle. They become very resistant just before they come out of the lag phase; then their sensitivity increases when they go into X-Ray Sensitivity 111 the log phase, with a return to initial sensitivity in the maximum stationary phase. We have also followed organisms grown in glucose, removed samples at certain time intervals, and studied their sensitivity, and they seemed to follow somewhat the same pattern, at least in the early stages of the growth cycle. Parkes: Has this kind of experiment been done on other unicellular organisms? I am thinking of mammalian sperm, for instance. Hollaender: Not as far as I know. Experiments on oxygen tension have been done on a wide variety of organisms. Dr. Kimball in our laboratory has tried Paramecium, but as far as I know it has not yet been done on sperm. Sloviter: Can you grow these cultures in the presence of small quantities of radiation and make them resistant to large doses later? Hollaender: There is no adaptation to radiation as far as we know. However, mutations are produced, and you may get a mutation which is much more resistant. The resistant strain which I discussed, B/r, was obtained by Witkin, irradiating a large bacterial population with ultraviolet; out of the few that survived, one was resistant. Loutit: I have been most impressed by Dr. HoUaender's contribution this afternoon because protection against the generalized effects of radiation is a subject in which at the moment we are all particularly interested. I think Dr. Hollaender has admitted that he has taken the easy path by utilizing bacteria as his biological tool. We, amongst others, have taken the hard road with mice. We were impressed with Dr. Leon Jacobson's contribution to the International Society of Haematology last August, where he showed that it was possible to protect mice, not by measures taken before the irradiation, such as the administration of cysteine or glutathione, but by measures taken after the irradiation, and in his case this was the introduction into the peri- toneal cavities of mice of spleens from infant mice. He showed that splenic grafting, as it were, gave the same degree of protection as shielding the spleen with lead during the irradiation. We have in our laboratory repeated this work of Jacobson's, only on a much smaller scale, and qualitatively we have been able to confirm his results. He was able to show with his strain of mouse, the CF^, that he could alter the median lethal dose from 600 r to something of the order of 1000 r. With our strain of mouse, the CBA, we have not had nearly such drama- tic results. We know that the normal median lethal dose of the unpro- tected mouse is of the order of 800-850 roentgens, and that at 950 roentgens we can get figures like this: in the control mouse no stu'vival out of 5 mice; in the protected mouse something like 3 survivals out of 5. Those are very small figures, it is true, but they do confirm Jacobson's thesis. We have not been able to get any significant protection at higher figures than 950 roentgens. We have also attempted to get protection, not from mouse spleens, but from spleens of other animals. We have tried the introduction of spleens from infant rabbits and from infant guinea pigs, and we find that the hetero-specific test is entirely negative, whereas the homo-specific test did show some positivity. 112 HOLLAENDER, StAPLETON AND BuRNETT Lamerton: Recently, when I was visiting the United States of America, I discussed with Huff at Berkeley some of the work he has been doing with *^Fe. It was suggested to me, in \'iew of the interest of these studies, that I should say something about them. Some of the work is unpublished, but I am sure that Huff would not mind my mentioning it at this meeting. A few months ago Hennessy and Huff published a paper {Proc. Soc. exp. Biol., N.Y., 1950, 73, 436) in which they showed that the uptake of radioactive iron, ^^Fe, by red blood cells was determined critically by a previous exposure of the animal to whole body irradiation. The technique was to expose the animal to whole body irradiation and then at a given time later to inject radioactive iron, afterwards taking a specimen of blood and determining the activity in the red blood cells. It has been found that if the injection is made 24 hours after the whole body irradiation, then quite small doses of radiation will affect the uptake. A dose of 5 roentgens is sufficient to cause an appreciable decrease in the uptake of iron by the red blood cells, and doses of 50 or 100 roentgens make a very great difference indeed. \Vliat is apparently being measured is an effect of the radiation on the erythro- poietic tissue, and the fact that the maximum effect is obtained when the injection is made 24 hours after irradiation is consistent with his- tological evidence of the damage to the marrow. I think these studies have a considerable interest since we are very much in need of criteria of whole body irradiation. By far the greater part of the radiation work being done with animals uses as a criterion of effect median lethal dose or some variant of it. This has for certain investigations considerable disadvantages, as I think many workers agree. In the first place, if you give a dose of the order of mean lethal dose you create widespread intense tissue destruction and the secondary effects of that may very well hide important effects that you would otherwise observe. Secondly, death from acute exposure is a very complex phenomenon indeed, and there may be several more or less independent mechanisms at work here, as has been shown by work with rabbits and also with rats. A very promising line seems to be to use tracers to provide quantita- tive criteria of metabolic processes that could afterwards be used to measure the effect of radiation. Obviously there is no one metabolic process that one should aim to measure; many will be needed. Radio- active iron does, however, offer one possibility there, and one which is very sensitive. Another possibility involves the use of ^*C. The use of radioactive iron has been extremely valuable in other types of investigation. In patients with various blood disorders Huff has been giving radioactive iron in the form of ferric chloride and after these injections making measurements of the counting rate over the surface of the body ,using a highly coUimated scintillation counter, the normal sites chosen being over the sacrum, the liver, the spleen and the heart. It is found that tlie pattern of the rise or fall in counting rate over the first few hours gives a fairly characteristic picture of the particular blood condition. For instance, in the normal male one finds that the X-Ray Sensitivity 113 counts over the sacrum rise fairly steadily over the first 6 hours, indica- ting uptake of iron into the bone marrow; over the spleen there is a slight fall; over the liver the counting rate is fairly constant. But in a patient having an anaemia M'ith a hypoplastic marrow, where there is very little marrow activity, one finds that the counts over the sacrum fall because there is very little uptake of the iron in the marrow. On the other hand, the counts over the liver and over the spleen rise very rapidly, and in the case of the liver, and to a certain extent the spleen, there is a storage of iron. The work is so far unpublished. Summing up, I feel that from the point of view of studies of whole body irradiation on animals, ^^Fe is likely to be very useful indeed. ISOTOPES EFFECT OF X-RAYS ON NUCLEIC ACID AND PROTEIN SYNTHESIS IN THE JENSEN RAT SARCOMA BARBARA E. HOLMES This paper describes attempts to detect any metabolic disturbances in tissues, which might result from treatment with ionizing radiations. This subject became interesting as soon as it was plain that some specificity of action could be expected. Before the action of radiations on substances in solution and the par- ticipation of the water in this action was partially understood, it was regarded as unlikely that any specificity of action on metabolic processes would be found. As soon, however, as attempts were made to show metabolic disturbance as a result of irradiation of tissue with X- and y-rays, it was plain that it was far easier to demonstrate the many systems which were not affected, than to find any which were. The partial inhibition of glycolysis in the retina, found by Crabtree and Gray (1939), is the only case I know of where a clinical dose of irradiation produced an immediate result on the ordinary energy-getting systems of the cell. It is, of course, always possible to show metabolic disturb- ances if the tissue is irradiated in vivo and left in the body long enough for death of some of the cells to occur, and such disturbance is sometimes wrongly described as an action on enzymes, when it is actually a change of cell population. Bone marrow irradiated and left in the body for some days has been described in this way as demonstrating enzyme changes, although much destruction of the dividing cells has occurred. Another example is given by rat gut, irradiated in vivo with 500 r. X-rays and left for four hours before the killing and dissecting of the animal. Sections taken from the gut at this time show destruction of the considerable masses of lymphoid 114 X-Rays and Cell Metabolism 115 tissue and a cessation of mitosis, which usually occurs rapidly, and by which the epithelium is supplied with young cells. These changes must account for part, at least, of the metabolic changes which have been described as directly due to enzyme destruction. Immediate effects, on excised tissue, for instance, cannot usually be demonstrated unless a very large dose is used and even in this case there is selective action and only some enzymes appear to be inactivated (Holmes, 1939). A biological illustration of the lack of effect on the ordinary processes of the cell is given by the difficulty of producing any visible action on a non-dividing tissue, while the biological physicists would agree that the number of ionizations pro- duced by ordinary clinical doses could not be expected to produce any obvious effect on the complex mixture of mole- cules in the cell cytoplasm. Biological studies had shown, on the other hand, that quite small doses do affect the nucleus and the chromosomes and prevent or make abnormal the division of cells. Thus it was plain that the various components of the nucleus had to be the objects of metabolic study. The work of Mitchell (1942), who demonstrated a dis- turbance in nucleic metabolism after X-rays, and also the early work of the Lawrence school in Berkeley, which had demonstrated the uptake of radioactive phosphorus into the nucleic acids of dividing nuclei, immediately suggested that 32p might be used to follow events after irradiation. At that time, just at the end of the war, we did not know that Hevesy (1945) had just completed work along the same lines. The use of tracer phosphorus is- essential for this work, in order that the new formation of nucleic acid may be deter- mined. Taking the average of the cell population, as one does when estimating nucleic acid in the tissues by other methods, the actual amount of deoxyribonucleic per unit wet weight of tissue does not change after irradiation, since it is produced in proportion to the amount of new cell material formed. 116 Barbara E. Holmes By the use of tracer phosphate one can estimate the extent of the new formation of deoxyribonucleic and can study at the same time the turnover of phosphate in the ribonucleic acid and the passage of phosphate through the cell membrane. As measured by ^^P uptake, there is very little synthesis of deoxyribonucleic in non-dividing tissue. Using the Jensen rat sarcoma as a dividing tissue and giving an X-ray dose of 2,000 r. (which gives complete destruction of over 50 per cent of the tumours) it w^as found that the up- take of phosphate into the cell was not affected by this dose, the turnover of phosphate in the ribonucleic acid was not affected, but the deoxyribonucleic synthesis was reduced to about half the normal rate. It should be stated here that other agents may have a profound effect on phosphate uptake and on ribonucleic acid turnover, but, of the systems we have so far investigated, the synthesis of deoxyribonucleic is the only one affected by this reasonable dose of X-rays. We are just beginning to make studies, histological and metabolic together, of later stages after irradiation. In one case, after two days, the deoxyribonucleic synthesis was still proceeding at one half the normal rate, normal mitosis was very hard to find, but many of the enlarged cells so com- monly found after irradiation were present and it was tempt- ing to suppose that some newly formed nucleic acid was being laid down in these enlarging cells instead of forming nucleic in new cells. There is no explanation as to why this should happen, unless, perhaps, one can imagine that other cell processes are proceeding at speed while the partial lack of this one substance prevents the actual normal division and that, thus, freaks may develop. To carry supposition further, it may be pictured that some drugs and chemical compounds, having a more generalized effect on the cell, might slow up cell division and many other processes for a time and, when the drug was eventually removed the inhibited cell might simply resume its normal existence. In this case the cell would behave in a sense, like an organism in which growth has been temporarily stopped X-Rays and Cell Metabolism 117 by starvation. The inhibition of mitosis by shock, which will be mentioned later, seems to be a more generalized process, which passes off altogether unless the shock is fatal. It was plain that the synthesis of protein, both cytoplasmic and nuclear, in the dividing cells should next be considered, and the effect of irradiation upon this should be determined. The protein fractions we found most easy to separate were the histone from the sedimented nuclei, the ribonucleoprotein (precipitated by calcium) remaining in the supernatant fluid, and the heat coagulable protein remaining after the removal of the ribonucleoprotein. Sulphur 35 was injected as sulphide or as methionine into rats bearing the Jensen sarcoma, which were killed 1^ hours later. Some of the tumours received 2,000 r. X-rays immediately after the injection and the uptake of ^^S was measured in the different protein fractions of the irradiated and control tumours. The histone fraction showed a variable uptake of radio- active sulphur, but it was not possible, in this or either of the other fractions, to show a constant effect of irradiation on the protein turnover. If the histone is actually attached to the deoxyribonucleic acid, the lack of correspondence in X-ray effect on the two substances is remarkable. Some final experiments were carried out in which ^'^P and methionine ^^S were injected simultaneously, in order that the protein turnover and the nucleic acid formation could be traced in the same animals. Table I gives the results and shows the complete lack of any relation between the effects of irradiation on the uptake of the two substances. These results will be more fully discussed in a paper now in pre- paration. So far, the work described suggests a somewhat specific effect of irradiation. I should now like to describe the action of very non-specific mitotic inhibiters, that is, the "shock" substance or substances formed in a ligatured limb. In our earlier work we used to fix the animal for irradiation partly by threads passing round the wrists and ankles ; these were not tight but were sometimes pulled tight if the animal 118 Barbara E. Holmes O < J/2 Q W H o o s iX! Q < H a o » o S o o en a, CO Deoxy- ribonucleic X + "^ CO c + rH + (>1 X CO + (N CO X ^ 4- -* c + s e a: Oi Tjl CO r— I i-H + r-l >-H + + rH C'l (N X + CO CO CO + C l> + S J Q. T3 fr- a; a; .^ rS "1^ _ T3 fc. S3 CK ^3 rH a; — HH I-H 'O rH a; • Si— I-H h^ c -M T3 rH c3 a> ;-- 4J r^ hH hH H^ s * * Tf CO + * * r-l 05 CO ^ + * • * r-l r-l ^ + * * X to CO CO + * * ^ «0 CO >o rH '^ -i- ■0 rH X CO r^ CO -* + 00 in CO C + X X (N rH + * * X b- CO CO (M :^ j- sO i-O rH CO CO CO + Jfl rH r^ rH '^ 4- 03 2 s -1 CO ■^ + X CO CO -* + X CO + CO C2 CO c + CO C5 + CO r-l rH (>J + so s .2 ■■ CO '>i CO ^ + CO I- CO X CO iM CO CO r-l rH X "* CO 2 2 rH X CO (N + X X ri CO + CO CO CO 10 CO ^ rH c + X + ^ 4- .2 hi u cfl 0/ 1 a 0. 3 a> C o c o a o ti 0. CD 4; .S .t: c a o 4) "O cC d .£ = '^ 2 ,^ ■ ^ o £ = £ c - c «< S O 03 — O 3 S S - c c 5 3 73 O .2 o .= ^ <" s c =s 5 - o O to y T3 O _ " i' «i 3 ''J ■" S ^ C3 f ^ C >< ^ p ^ di ^ " -^ '^ ,CJ 3 C5 !< » . — O -S — O 3 2 7-^ 3 i I' > 01 > In CA 3 c« to <1> to to en en 'c 'E t£ M o + o o -E o S ? t" ? W 1; 2 4* O o '-2 o O o ^ o CU X > 72 X-Rays and Cell Metabolism 119 struggled. In these earlier experiments we always appeared to show a temporary effect of X-rays on ribonucleic acid synthesis and sometimes on the uptake of phosphate into the cell, as well as on deoxyribonucleic synthesis. Hevesy also had found that, if a long, slow irradiation was given, the fastening down of the animal caused a diminution in the amount of phosphate with ^^P arriving in the cell. The contrast of such results with our later specific ones caused a suspicion that they may have been due partly to "shock", and Professor Green's observations (Bullough and Green, 1949) that mitotic inhibition occurred in the skin of the ear after release of a ligatured limb made it seem much more probable. Professor Green very kindly carried out one of his experiments on rats bearing the Jensen sarcoma and gave me a chance of following nucleic synthesis with ^^P at the same time. The chief effect visible in the first hour of shock, was inhibition of ribonucleic turnover. Later, the inhibition of uptake of ^^p into the cell [a mechanism which is now being shown by Popjak (1950), Sacks (1948) and others to be a phosphorylation process] became so marked that the radio- active counts in the nucleic acid fractions were reduced to a very low level. This effect is not, apparently, due to cir- culatory collapse, since we have not seen it in animals at the point of death from nembutal or chloroform anaesthesia. Inhibition of deoxyribonucleic synthesis did occur, but not to a very marked extent. This more generalized type of phosphorylation inhibition is being examined further, but I should like to return for a minute to the specialized effect of X-rays. When such a definite effect on the life of the cell is caused by the small amount of ionization produced by a clinical dose, one is forced to suppose that the destruction or inhibition of an enzyme is involved. In the synthesis of a substance like deoxyribonucleic acid, necessitating phosphorylation and many other processes, a very large number of enzymes must be concerned, the loss of any one of which would put the system 120 Barbara E. Holmes out of gear. This, perhaps, would provide the large or diffuse target which is demanded by biophysicists, Dr. L. H. Gray and Dr. S. Pele among them. However, the particular vulnerability of this system to irradiation may be due to its position rather than to any particular sensitivity. Enzymes placed on a structure such as a chromosome or chromatin thread and perhaps out of reach of the protecting substances in the cytoplasm might show great vulnerability. There are many obvious fallacies in the comparison of inhibitions caused by chemical agents and by ionizing radiations, but, with due regard to these, it is probably worth attempting to imitate these metabolic effects of X-rays by the use of inhibitors of known enzyme systems, together with the tracer materials. REFERENCES BuLLOUGH, W. S., and Green, H. V. (1949). Nature, Lond., 164, 795. Crabtree, H. G., and Gray, L. H. (1939). Brit. J. Radiol., 12, 39. Hevesy, G. (1945). Rev. mod. Phys., 17, 102. Holmes, B. E. (1939). Proc. Roy. Soc. B., 127, 233. Mitchell, J. S. (1942). Brit. J. exp. Path., 23, 296. PopjAK, G. (1950). Nature, Lond., 166, 184. Sacks, J. (1948). Cold Spr. Harh. Sym. quant. Biol., 13, 180. DISCUSSION Boscott: a paper by Butler and Smith (Butler, G. C, and Smith D. B., 1951, J. Amer. chem. Soc., 73, 258) states that X- and gamma-rays degrade sodium thymonucleate. I wondered if that would be of any interest to you. Holmes: I think it probably is, only as far as we know, it does usually take a rather enormous dose. I was thinking of it when Dr. HoUaender was saying that perhaps you have got two things to consider in irradia- tion damage. It does look very like it from the biochemical data. You may have to think in one case of the enzyme inhibition, which you may be able to stop with something like hydrosulphite, and you may at the same time have to think of some direct effect like the effect of a hydroxyl radical on your sodium thymonucleate. Boscott: In the biochemistry of mitosis perhaps purely analytical biochemistry may not give results that one would desire, because one may be up against such things as the degree of polymerization and depolymerization of the nucleic acid. X-Rays and Cell Metabolism 121 Holmes: That is why we try to follow it in the body with tracers. As soon as you take the tissue out I suppose mitosis stops, and I don't think we should have much hope of seeing mitosis effects on tissue slices. Brown: With regard to the problem of interference with nucleic acid synthesis, you probably are familiar with the observations by Skipper and co-workers at Southern Research Institute that an anti-folic-acid will decrease the incorporation of formate into the purines of mixed nucleic acids, PNA and DNA. We have done a similar experiment in which, rather than a chronic toxic dose, a single lethal dose of Amethopterin was given. In this case the incorporation of formate into the PNA is decreased on the first day, but if the formate is given 24 or 48 hours afterwards, there is a much greater incorporation of that formate. The animal still dies on about the fourth day, but on the third day it has largely recovered the ability to elaborate purines from formate. Another experiment has been done in which the formate was labelled with i*C and the adenine, which was given simultaneously, was labelled with ^^N. This experiment indicates that the anti-folic-acid has apparently greatly decreased the synthesis of purine from formate, but has not correspondingly decreased the synthesis of nucleic acids from preformed adenine. The fact that two different precursors of the same final product can yield such different results adds just one more complication to any studies of the effects of a drug or of radiation. RADIATION DOSE IN TRACER EXPERIMENTS INVOLVING AUTORADIOGRAPHY S, R. PELC In many tracer experiments the radiation dose due to the radioactive element hmits the scope of the investigation. The radiation dose should, in all cases, be below a value which would affect the process under investigation, and in many applications, especially those involving humans, radiation damage to organs other than those investigated must also be avoided. To obtain useful results a certain minimal concentration of radioactive material in the final sample is required and this concentration will be one of the determining- factors for the lowest possible radiation dose. Other factors are the time which elapses between application of the radio- active material and removal of the sample, the uptake and excretion of the material and the radiation characteristics and half-life of the radioactive tracer. This paper is an attempt to co-ordinate these factors in a mathematical form which will enable the investigator to estimate the quantity of radioactive material necessary to give the desired result, as well as the radiation dose to the tissues. In some cases calcula- tion might show that an experiment is not feasible, in others the extent of radiation damage might be evaluated. It should be realized that such calculations will normally be based on very insufficient data and the result should be regarded only as a guide in these cases. Calculations relevant to this problem have been published by Morgan (1947) and Marinelli (1949), but no comprehensive equations to describe the processes in question seem to have been developed so far. The equations given in this paper are designed to meet the needs of the experimenter desiring to produce autoradio- 122 Radiation Dose in Autoradiography 123 graphs, but the general form should be equally suitable for work with counters. In general the concentration in an organ of an element applied at time will follow a curve similar to Fig. 1. The dose will be due to the total concentration of tracer, only a fraction of which is utilized in producing the autoradiograph. Figs. 2, 3 and 4 show photomicrographs of cells and auto- radiographs of tissue cultures of avian fibroblasts (Felc and Spear, 1950a) grown in plasma containing ^^p (Pelc and fC-c) Fig. 1. Uptake and excretion of tracer (assumed to be stable) as a function of time (t). Spear, 1950b). The distribution of ^^P in the cells after fixation in 80 per cent alcohol and washing (Fig. 2) is fairly general, and it is safe to assume that a more even distribution would have been obtained if the water soluble phosphorus compounds had been retained. After extraction with 5 per cent trichloracetic acid at 15°C. a more detailed autoradio- graph appears (Fig. 3). This trend is even more marked after extraction with normal hydrochloric acid for 10 minutes at 60 °C. when mainly phosphorus compounds situated in the cell nuclei are retained (Fig. 4). The radiation dose given 124 S. R. Pelc to the cells will be due to the total concentration of ^^P before fixation whilst only ^^P insoluble in hydrochloric acid will be of value in an investigation concerning deoxyribonucleic acid in this material.* The fraction of the tracer utilized in producing the autoradiograph depends on (a) tracer lost involuntarily during preparation of the specimen, {b) tracer extracted during the preparation in order to retain only the radioactive element in those compounds which are of interest. Let C be the concentration in /xC per ml. in the latter com- pound or compounds at time t, a, the total concentration of tracer and a/C=k(T) their ratio. If an amount, A. ixC per ml. is applied at time the concentration C at time r will be: — C=Af(T)e-^- /xC/ml. ^ (1) If a minimum concentration Cq is needed at the time (t) of removal of a specimen to obtain an autoradiograph, either after an exposure of two half-lives for short-lived isotopes or after d days for long lived ones, the minimum concentration Aq to be applied will be: — Ao=-^ (2) Electronic equilibrium in the tissue in question has been assumed for the following calculations. This will be so when the energy of the emitted electrons is dissipated in the material. It is realized that this condition will not always be fulfilled, but the mathematical difficulties of the more general case would make the final equations unwieldy. The dose rate will be: — d.r. = 61aEy3==61 Ck(T)E^ reps per day (3) where E^ is the average electron energy in ]\IeV. *More detailed information on this work can be found in a paper by A. Howard and S. R. Pelc in this volume. Figs. 2, 3 and 4. Autoradiographs of avian fibroblasts grown in plasma containing ^-P. Left: phase contrast ; right: same position without phase contrast. [To face page 124 tf. O o a; • :^- O o X ^ « bC CO C c3 :3 o 0» J. # I &^. #**'^ y^ J iS^ -* *•/ O o o o c3 * ' — CO d C "o? C u 0/ C £ o O o C 6 i-i Radiation Dose in Autoradiography 125 If Aq has been applied, combination of equations (1), (2) and (3) gives the general expression for the dose rate at time t: — d.T. = 6l^^^^^^^ e^t f(T)e-^^T reps per day (4) The total dose for the time T the tracer remains in the organism will be: — ^^ei^Co ^^, J k(^)f(^)e-Ar dr reps (5) o For practical use, knowledge of the minimum concentra- tion Cq is essential, and solutions for the integral have to be found. The factor k itself can be a function of time and in many cases the function k(r)f (r) will be simpler than f(T) itself. In most cases two possibilities are of interest: (a) when the radioactivity present in the organism after removal of the specimen at time t is not important, in which case the integral has to be evaluated for T=t; (b) when the dose after removal of the specimen is important for a longer or shorter period, i.e. when T>t. This applies especially to humans, where T=oo. Minimum Concentration of Radioactive Material to Obtain Autoradiographs Using short-lived isotopes the autoradiographs (ARG) can be exposed for times of up to two half-lives. After this time 75 per cent of the possible radiation has been utilized and longer exposures would not increase the strength of the ARG to any appreciable extent. Only one half of the particles emitted will hit the film and thus the overall efficiency will be 37-5 per cent, or less if developed grains in the penumbra are not counted. The grain yield in terms of the number of photographic grains rendered developable per incident electron has been found by Berriman et al. (1950) to be of the order one for 126 S. R. Pelc X-ray emulsions as well as for special stripping film. Lamerton (1950) fomid values of 0-5 to 0-7 for ^^F and 1-5 for ^^S on special stripping film for autoradiography. These discrepan- cies are not surprising in view of the difficulties of absolute measurement of radioactive substances, reproducible develop- ment and grain counts under the microscope. For the purpose of preliminary calculations, a value of one grain per incident electron can be assumed. To observe the ARG we require a minimum grain density above background which is dependent on the magnification. At one extreme are autoradiographs of large objects, such as leaves, etc., which may be viewed directly or with low magni- fications. Resolving powers of 50/x-lOO/x are sufficient for this purpose, and the work can conveniently be done with X-ray film pressed against the specimen. At the other extreme are autoradiographs of high resolving power viewed at high magnifications for which we use fine grained emulsions and more specialized techniques. When only macroscopic detail is wanted we observe the optical density of the autoradiograph, while with high magni- fications the number of grains per unit area is of importance. Various estimates of the number of electrons necessary to give a visible autoradiograph on X-ray film range from 2 x 10^ to 10' electrons per sq. cm. It w^as found in recent experi- ments (Howard and Pelc, 1950) that concentrations of 10 grains per lOO/x^ (above background) can be observed using Kodak special stripping film for autoradiography, for low background. This is equal to 10' grains per sq. cm. It should be stressed that 10 grains per lOO/x^ will not result in a visible blackening of the film and that they are observable onlv under favourable conditions. For a general calculation, it will be prudent to assume a minimum of 2 x 10^ grains per sq. cm. for X-ray film and 10' grains per sq. cm. for special stripping film. The thickness of the specimen determines the volume of material available, and the necessary concentration of radio- active material will be inversely proportional to the thickness. Radiation Dose in Autoradiography 127 For work with X-ray film, sections up to 20/z thickness can be used, whilst for high resolution not more than 5/^ thickness can be tolerated. Special conditions are presented in smears and squashes where the thickness of the specimen can usually not be measured. The area taken up by one cell nucleus in the final preparation can be measured and from this the actual volume per unit area calculated. Autoradiographs are usually used for the determination of the concentration in certain parts of the material, under investigation, and consequently the minimum concentration per ml. of tissue will be proportional to the ratio of the radio- active volume to the total volume of tissue, e.g. for a study of deoxyribonucleic acid labelled with ^^P in cell nuclei the ratio of the volume taken up by the nuclei to the total volume. (See Fig. 1 in the paper by A. Howard and S. R. Pelc in this volume.) This ratio has to be determined with great caution when the size of the object is of the same order as the resolving powxr. Thus, sparsely distributed cell nuclei of 10/x diameter in autoradiographs of 100/x resolving power have to make developable sufficient grains in an area of approximately 7,500/x2. jj^ high resolution work a similar limitation appears when the area in question is smaller than approximately 100^^. While 10 grains per lOO/x^ can be regarded as an autoradiograph, one grain per lOft^ can not; we may regard an accumulation of 10 grains above back- ground per single object as the minimum when this object covers an area of less than lOO/x^. Thus if we need n grains per sq. cm. to obtain an observable autoradiograph, the specimen must contain nf/0 • 375 radio- active atoms per sq. cm., where f is the ratio of radioactive volume to non-radioactive volume. This concentration must, of course, be present after extraction of all the labelled com- pounds which are not under investigation. The volume per sq. cm. of specimen will be equal to the thickness (8), and the number of radioactive atoms per ml. of tissue must therefore be nf/0 -3758. One fxC of radioactive material contains 4 • 64 x 10^ X H atoms, where H is its half-life 128 S. R. Pelc in days. The concentration necessary for obtaining an auto- radiograph after exposure for two half-hves will therefore be: — ^ = 0.3758x4"6xlO»xH =mX^-^X^""VC/°^'- (^> For work with special stripping film, taking n = 10^ grains per sq. cm. and 8 = 5 xlO-* cm. we get: — Cs=iij^^ (7) Similarly for X-ray film, where n— 2 xlO^ 8=2 x 10"^. Cx-0-58g (8) For elements with long half-life, exposure for two half-lives will be out of the question and the minimum concentration will then depend on the exposure time chosen. For inter- mediate half-lives, such as ^^S, where the full exposure time might be practicable, but tedious, allowance for the shorter exposure time can be made, taking the decay into account. For elements with very long half-life, the decay during exposure can be neglected. We need a total of nf/0 • 5 electrons emitted per ml. of tissue. 1 fiC gives 3-2 xlO® disintegrations per day and the necessary concentration, w^here d is the exposure time in days, is: — For stripping film and X-ray film we obtain: — C's=i^V/ml. (10) 0-625f Cx= — T — /xC/ml. (11) Radiation Dose in Autoradiography 129 Minimum Dose for Short-lived Elements Inserting the expression for the minimum concentration (equation 6) in the general equation for the dose (equation 5) we obtain: — T a-d 5X1U ^^^^^^ e k(T)f(T)e-^^ dr reps (12) o Using the same assumptions which led to equations (7), (8), (9) and (10), equation (12) can be simplified. T A ^At k(T)f(T)e-^'' dr reps (13) o f(t) where B=6lE^Co. Bs== — „ for stripping film (14a) 11 Bx=36 ^ for X-ray film (14b) For work with long-lived isotopes: — Bs=800 ^ (15a) B'x=38 5|! (15b) The following solutions of equation (13) were computed under the assumption that k is independent of time. It should be noted that for computations, the values for the time should be in days, and the value for A based on time of decay in days. (a) Linear increase or decrease of concentration f(T)=aa-|-br ISOTOPES 10 130 S. R. Pelc ic"RpAt r d=. , ^,^.: .., e(a^tA+b)_e-^T(aA+b+bAT) reps (16) ^a-j-Dt jA L J When the organism is killed for preparation, i.e. T=t, equation 16 can be shortened to : — kB d= (a+bt)A2 _ e^t(aA+b)-(aA+bAt+b) reps (17) When the element can be assumed to have decayed com- pletely in the organism, i.e. T=oo, e"^^=0. kBe^* These equations will at once give the appropriate expres- sions for constant concentration when b=0, and for concen- tration proportional to time when a=0. (b) Exponential increase or decrease of concentration f(T)=be'^^ ^^K^e — (g(a_A)T_i) reps (19) a— A kR forT=t:d=— ^ (l-e^^-'^H) reps (20) a — A for T=oo, a solution which is valid only where (a — A) is negative: — kBe(^-«)t d= r — reps (21) a— A The solution for concentration proportional to l-e""'' can easily be computed by combining the solutions given under (a) and (b). (c) Concentration increasing at first and then decreasing through excretion, etc., similar to Fig. 1. d= Radiation Dose in Autoradiography f(T)==bTe-^ kBe(«+^)t " t(a+A)2 [ l_Te-T(-+A)(a+A)-e-T(«+ A) reps 131 (22) for T=t: d: kB t(a+A)2 forT=oo: d = kB t(a+A)2 et(a+A)_(^_|_;^)t_i et(a + A) reps reps (23) (24) Minimum Dose for Long-lived Elements For elements with very long half-life or when the time taken up by the experiment is much shorter than the half-life, the radioactive decay during the experiment can be neglected and simpler equations are obtained. Applying AjuC per ml. of tracer will result in a concentra- tion of Af(T)jLtC per ml. at time r. Thus to obtain a concen- tration Cq at time t we have to apply: — A=;p— - fxC per ml. i(t) (25) The dose rate at time t will be: — d.r.=?ig^k(.) f(.) reps per day (26) Substituting as before, B=61 E^ Cq which will have the values given in expressions (15a) and (15b) we obtain for the dose: — B f(t) k(T) f(T) dr reps (27) o The following are solutions of equation (27) for a number of functions for f(T), assuming k to be independent of time. 132 S. R. Pelc (d) Linear increase or decrease of concentration with time: — f(T) = a+bT Bk fort=T:d = ,^^^-^3^j (2at+bt2) reps (29) (e) Exponential increase or decrease with time: — f(T):=be" Rk d~, (e^-'-l) reps (30) Rk forT=t:d= — (l-e-*^*) reps (31) a (/) f(r)=b(l-e-") T?k ^== a(l-e-°t) (aT+e-°T-i) reps (32) for t = T: d=Bk ( _^ _^^ - i) reps (33) (^) Concentration increasing at first and then decreasing through excretion, etc., similar to Fig. 1. f(t)=bTe- 1_ n — p-aT — p-aT a^te" Bk d=-S7lS (l-e-T-e-TaT) reps (34) Bk for T=t: d=:-^ (C^t.j.^t) reps (35) a^t forT=oo:d=:^^ reps (36) a-^t Radiation Dose in Autoradiography 133 Discussion The varying conditions under which tracer experiments are undertaken make it impossible to draw any conclusions from these calculations which might be applicable to all cases. A few cases may, however, shed some light on the order of magnitude of the radiation dose involved even if they are not applicable to a specific experiment. The following computations were based on the data for stripping film; for X-ray film the doses would be about one twentieth. In many applications the tracer element will be distributed through an organ and be subsequently incorporated in various compounds. Mathematically this can be regarded as immedi- ate uptake with a constant concentration for a longer or shorter period. Using equation (17) with b=0 and assuming k(T)=k=l we obtain for stripping film in the case of the organism being killed for the preparation (HA =0-693): — d= '^^^fi^ k(e^^-l) = 10Q0 E/k(e^t_i) ^eps (37) MA Computations of equation (37) shown in Fig. 5, show that where these assumptions are fulfilled autoradiographs can be produced with a fair prospect of escaping radiation damage for the isotopes ^^^I, ^^P and ^^S. For medical applications the tracer will usually remain in the body until it is either excreted or decayed. This case can be calculated, again under the assumption that the con- centration reaches a permanent value shortly after application of the tracer. Taking equation (18) and inserting b=0, b from equation (14a) and k(T)=k = l:— d=-^e^*=1000kE/e^t reps (38) A Calculated values for various isotopes are shown in Fig. 6. It may be of interest that the minimum exposures are high 134 S. R. Pelc and the radiation dose presents a very severe limitation for elements other than ^^S unless a high proportion of the element is excreted or the values for the factors k and f are very favourable. It is obvious that the factor k will remain minimum dose to tissue for a.r.g. (organism killed) '^-•) REPS. f CO -CONST 1 10^ E^ (e IQOOO lOOO 1 (REPS) lOO lO J A^^ p'/y I b^ / / / / 1 2 3 i 8 I'i Id 2 4 HOURS aoi o-i 10 lOO lOOO DAYS Fig. 5. small if labelled precursors of compounds of interest rather than inorganic compounds are used in the first instance, and better knowledge of precursors for various substances will be of great advantage for further work. As an example of work with long-lived isotopes, radioactive carbon ^*C may be chosen. For immediate uptake and sub- sequent constant concentration equation (29) can be used Radiation Dose in Autoradiography 135 when the organism is killed at the time the specimen is removed. d=!^^ tE, reps (39) If we assume an exposure time of 40 days we can deduce the simple rule that autoradiographs using ^^C can be made MINIMUM DOSE TO TISSUE FOR A.R.G. (TRACER REMAINS IN ORGANISM) f(r)=CONST jI-Io'e^ e^^ 10,000 lOOO REPS 100 IQ La'* _p!L. l^ ^ --1^ s'^ , lO DAYS Fig. 6. 15 20 2S with a radiation dose of 1 rep per day to the organism. The same figure can, without undue error, be used for work with ^^S. If X-ray film is used autoradiographs can be produced with a radiation dose not exceeding the tolerance dose under favourable conditions. These calculations assume that fine grained stripping film is used for making the autoradiographs. For X-ray film the 136 S. R. Pelc doses will be one twentieth of the calculated values. Also the dose will in general be lower for gradual uptake and subsequent excretion. These calculated values can therefore be taken as maximum values, unless the factors k and f be very unfavourable. REFERENCES Berriman, R. W., Herz, R. H., and Stevens, G. W. W. (1950). Brit. J. Radiol., 23, 474. Howard, A., and Pelc, S. R. (1950). Brit. J. Radiol., 23, 634. Lamerton, L. F. (1950). Private communication of preliminary results. Marinelli, L. D. (1949). J. din. Invest., 28, 1271. Morgan, K. F. (1947). J. phys. colloid Chem., 51, 984. Pelc, S. R., and Spear, F. G. (1950a). Unpublished. Pelc, S. R., and Spear, F. G. (19506). Brit. J. Radiol., 23, 287. DISCUSSION Lamerton: I would like to ask Dr. Pelc how he calculated the value 2-5 roentgens for the dose given to the thyroid — whether he did in fact assume a uniform distribution of iodine? Pelc: That figure was based on actual experiments. We gave the rats varying amounts of radioactive iodine and found the smallest amount which will still give an autoradiograph. One lobe was taken out, the total amount of iodine in this lobe was determined and the radiation dose calculated, assuming that the radioactive material is evenly distributed throughout the thjToid. Lamerton: Isn't there a fairl^'- considerable error there, particularly with the variation in concentration you get in the thyroid? Pelc: It depends on the time of removal. In the beginning you have a fair amount of iodine present, and therefore the difference will not be large. Later on, 24 or 48 hours, the difference might be appreciable. It is unlikely to be tremendous, however. Gray: I think it is a matter of the greatest importance to radio- biological research that it should be possible, as Dr. Pelc has demon- strated, to prepare autoradiographs showing the localization of isotopes such as 3 2p and ^^S in living cells without appreciable damage being caused to those cells by the radiation emitted from the isotope. It is characteristic of certain aspects of the damage induced by ionizing radiation that it appears only in a proportion of the cells. One cell may be heavily damaged and its neighbour apparently unaffected. Bio- chemical analysis of tissues is too coarse for the study of such phenomena, because the results obtained represent the average response of a cell population in which the proportion of damaged cells may be very small. Cell-autoradiography, on the other hand, reveals the metabolic condition Radiation Dose in Autoradiography 137 of individual cells, and I think we may look for great advances to follow upon the application of this method in radiobiology. Hollaender: Isn't 2-5 roentgens sufficient to upset the rate of mitosis? Gray: In exceptionally sensitive tissues, such as grasshopper embryos, it is, but it would not appreciably interfere with the growth of plant meristem or cause chromosome damage in more than one or two out of every 100 cells. Holmes: If 2-5 roentgens were delivered to a whole animal the radiation effects might not be negligible. If you were interested in the blood forming organs, the dose of 2-5 roentgens to the whole body due to the 3 2p incorporated into the tissues might not be without effect on the biological change which you were studying. SYNTHESIS OF DEOXYRIBOSE NUCLEIC ACID AND NUCLEAR INCORPORATION OF ^^S AS SHOWN BY AUTORADIOGRAPHS ALMA HOWARD and S, R. PELC AuTORADioGRAPHs Can be used as a cytochemical method if their resolution is sufficient to distinguish radioactivity in individual cells and cell parts, and if, of those compounds which contain the administered isotope, all can be removed from the cell after fixation except the one under consideration. When this is done, the method makes it possible to measure the time of incorporation of the isotope into specific com- pounds. This paper extends work already reported (Howard and Pelc, 1951a) and describes experiments designed to examine the time of synthesis of deoxyribonucleic acid (DNA) in relation to the mitotic cycle, nuclear uptake of ^^S, and the effect of X-rays on these processes. The material used was the growing main root of Viciafaba seedlings, cultured in aerated water at 19°C. Gray and Scholes (1951) have shown that under similar conditions of growth the meristematic cells, which occupy a region 2 to 3 mm. long at the tip of the root, have a mean intermitotic time of 19 hours, while division, from prophase to the end of telophase, occupies 2^ hours. Thus every 21 J hours all the meristematic cells divide, and of their daughter cells, one half remain meristematic and the other half differentiate. At the upper limit of the meristem the mitotic rate falls sharply, reaching virtually zero in the cortex at about 3 mm. from the tip. In the developing vascular region mitotic activity is present at more proximal levels in the root, and at a distance of 1 to 3 cm. the initials of the lateral root meristems can be seen in these tissues. Thus some at least of the cells in this region must be considered as premitotic. 138 r- o S ^ ^ ^ 03 "> -M K* Td (— r— 1 4> "B. CC a; bJj O ■'5 t/; C5 ^ -M O ■M 3 tf I 9 - o <1> c ^ ^^ ^-* 'c, ^ ■s ^^ i^^ <^ K o.:: . O'C 6 *^ c > |2 ^ G bD o OJ 0; i^-^^ 4i 11 ■*- ^ C P X e bjj O « ■-s • cS m ^ u c w*« -M w C3 «\ .-^ < ¥^ "a: c S ^ "x ^ X O -M r* r" +J C X ■M rt X ^ -M 'C 0; --- c o -M 0^ o VI c 2 PC r^ ti ^■H ^ o fe ^ To /ace page 139] DNA Synthesis 139 I. Synthesis of DNA Method Roots were grown for periods of time from 2 to 48 hours in water containing ^^p as NaH2P04, with 16 mg. per htre of added carrier phosphate. The activity was such that the product concentration X time was constant at 4-8: thus roots grown for 24 hours were put into a solution containing 0-2 /xC. per cc, while roots treated for shorter periods received relatively higher activity. An exception w^as made in the case of roots treated for 36 and 48 hours: these received an activity equal to that of the 2 4 -hour sample. After treatment the roots were fixed in alcohol-acetic acid (3:1), washed in water, hydrolysed in N. HCl at 60°C. for 10 minutes, washed in water, and portions squashed in 45 per cent acetic acid. After removal of the cover slip and further water washing, the slides were autoradiographed by the stripping-film method (Doniach and Pelc, 1950). Exposure was from 2 to 28 days. After development and fixation, slides were mounted in chrom- jelly and observed with phase contrast. Controls 1. Autoradiograph. Material treated as above showed definite autoradiograph over some nuclei (Fig. 1). Material treated in an exactly similar manner but without ^^P showed no autoradiograph. 2. Growth Rate. Roots were treated for 48 hours with 0-2 jLtC. per cc. of ^^P, and then grown in carrier without further ^^P for 10 days. These roots grew as fast throughout the 12-day period as did inactive controls. We may therefore conclude that ^^P at the activities used does not cause radia- tion damage sufficient to aff'ect the mitotic rate of the meristem. 3. Interpretation of Nuclear Autoradiograph. The phos- phorus compounds which may be expected to remain in the cell nuclei after fixation and hydrolysis as described above 140 Alma Howard and S. R. Pelc are nucleic acid, phosphoprotein, and some phospholipid. The one most probably present in quantity is DNA. Squashes were subjected to a solution of deoxyribonuclease.* These squashes showed no nuclear autoradiograph, while an auto- radiograph could be obtained over the nuclei of cells subjected in the same way to a boiled solution of the enzyme. We therefore conclude that the fixation, hydrolysis and washing of the tissue has removed from the cell nuclei all phosphorus compounds present in sufficient amounts to give an autoradio- graph except the DNA, and that the presence of an autoradio- graph over a nucleus means that ^^P is present in the nucleus in the form of DNA. This could occur either throufifh the synthesis of new DNA since the beginning of ^^P treatment, or through the exchange of stable for radioactive P. As is discussed below, the turnover of P in non-dividing nuclei is very low indeed, and we therefore consider that a nucleus which shows an autoradiograph has synthesized DNA during the period of treatment. Observations ^^P was given to roots over time periods of 2, 4, 6, 9, 12, 16, 20, 24, 36 and 48 hours. Squashes were then examined and the percentage of dividing and non-dividing meristematic cells which show autoradiograph was determined. The results are expressed in Fig. 2. The number of non-dividing nuclei which show autoradiograph rises from 20 per cent at 2 hours to above 90 per cent at 48 hours. Dividing cells show no autoradiograph at 2 or 4 hours. At 6 and 9 hours the figure is about 10 per cent, and this value rises steeply to 100 per cent at about 16 hours. Thus there is a lag of about 6 hours between the appearance of positive non-dividing and positive dividing cells (Fig. 3). If the dividing cell class is broken down into stages of division, we see a few positive prophases appearing first, and at later times, metaphases and finally anaphases and telophases (Fig. 4) which are positive. ♦We are indebted to Dr. Barbara Holmes for a preparation of the enzyme. »i "» O a; °<^ o o +-> !^ S^ M cS O +-■0 1^ o '^ S a ct ■— ^ a, a CO ^ ..^ife.. --.« o be [To /ace page 140 Zi. (^S C W u w^^ ^/ ^/ c^ S-^ C o o -M ^ -M ^-' 0) o < •— "^ 4-- O x w^ <^ 4J -^ £1 M tt +J ^^ ^W >w o X o X «w o o •^ •/: MBi- ^f O ij :3 s o O HH .-rt "iC "S , ^^ b a; X +j |MH 55 Ip X C 2 +^ *J n HM ■0 « ^*-) o o t» (L> cc ^ X ^s .^ ^■^ — 3 »-^ ^f o JJ 00 ^ S > o X S3 ^> ^^ ;^ o '-\ +-> CS s C o C X -t ■kJ ^ T;. DNA Synthesis 141 These observations are interpreted to mean that DNA is not synthesized during cell division but only during the inter- phase, and that there is a lag between the end of synthesis and the beginning of visible prophase of about 6 hours. The relationship between DNA synthesis and the later behaviour of the cells can be discovered by making use of either of two facts. The first is that differentiating cells elongate, while meristematic ones are rarely more than 50 /x long. By taking squashes from the final 3 mm. of the tip of roots which had been given ^^P for 20 and 40 hours, a cell 100 80- o, 60- X X— 7 /o ARG 40- 20- Q-t^nm — r O lO 20 30 40 TIME HRS. 60 Fig. 2. Percentage of cells showing autoradiograph in relation to time of ^^p treatment. Solid line: non-dividing meristematic cells. Broken line: dividing cells. population is obtained which includes elongating as well as meristematic cells. The ratio of positive to negative nuclei from such a squash, in relation- to cell length in microns, is shown in Fig. 5. It is clear that the proportion of positive cells falls sharply with increasing cell length. The maximum value is higher for the 40-hour root than for the 20-hour, and some of the elongated cells show autoradiograph after 40 hours. The second fact which may be used in this connection is the distribution of mitotic cells in the length of the root. The 142 Alma Howaed and S. R. Pelc mitotic rate is very high in the final 2 mm. of the tip and decreases with distance from it. In order to discover the percentage of positive cells in relation to their distance from the meristem, longitudinal sections were made of roots grown for 24 and 48 hours in 0-2 /xC. per cc. of ^^P. The resolving power in sections is not as good as that in squashes due to greater thickness, overlapping of cells, and absence of the spreading which is achieved in squashes. However, it is 40n 3-0- RATIO 20- l-O- I 1 I I I O 20 40 60 80 lOO I20-200 LENGTH OF CELLS p 20 HRS. P^^ -t— O 20 40 60 80 lOO I20-200 LENGTH OF CELLS >j 40 HRS P^^ Fig. 5. Percentage of cells showing autoradiograph in relation to cell length. The indicated standard deviations are based on the number of observations only. possible adequately to resolve individual nuclei in the elon- gating cells of these roots. The results are summarized in Fig. 6, which gives readings for two roots, one of which had 24 hours in ^^p and during this time grew 3-0 cm., while the other had 48 hours treatment and grew 4-2 cm. It will be seen that in the cortical cells the percentage of positive nuclei falls towards zero in both roots, from a level similar to that found in the meristem in squashes. In the developing vascular region, where mitotic activity is appreciable at more proximal levels, the meristematic value is maintained for DNA Synthesis 143 some distance above the meristem, and then falls away (in the 48-hour root) with a slope similar to that seen in the cortical cells. (In the 24-hour root, sections were not cut above 3 cm. from the tip.) 5 loor 4. O o Cortical Cells < i o z % o I 80 60- 40- -I ui O 20 u O X 24Hrs.^^P O 48 Hrs. " P O o X X ->e -x^ -L A. 2 '^ '" 3 4- 5 Distance from Tip (cms) 6 'Vascular*' Cells I a < o o < o »- D 4 lOOr 80 60- X 24 Hrs ^^ P 22, o 4.8 HRS "^ p O 5 I 40|- Ui U 20 o X X J. -L 2 3 Distance 4 5 FROM Tip (cms.) Fig. 6. Percentage of cells shoAving autoradiograph in relation to distance from the tip of the root. (Permission has been obtained by the authors from the Jotirnal for Experimental Cell Research for this figure to be reproduced). 144 Alma Howard and S. R. Pelc These observations are interpreted to mean that DNA is synthesized by those cells which are preparing for a further division, but is not synthesized by cells which will differentiate without further division. A cell which synthesizes DNA transmits this to its daughter nuclei, which then retain it. The fact that most cells in the non-mitotic upper parts of the root do not show autoradiograph after 48 hours in ^-P means that the exchange rate of phosphorus between DNA in cell nuclei and other phosphorus compounds in the cell is very low indeed. The autoradiographs used in this work are of such a sensitivity that a nucleus containing as few as 50 atoms of tracer which disintegrated during exposure would be counted as positive. II. Nuclear Uptake of ^^S The methods of culture of roots, and the fixation, hydrolysis and autoradiographing of cells, were the same as those described for the previous experiments. ^^S was given as sodium sulphate, at 1 fiC. per cc, without carrier, for 20 hours. Autoradiographs appeared over a proportion of non-dividing nuclei (Fig. 7) and over dividing cells. In the latter, the auto- radiograph was associated with the chromosomes (Fig. 8). The compound in which ^^S exists in these cells is not known to us. The treatment of the tissue before autoradiography can be expected to have removed all the inorganic and acid- soluble sulphur compounds, including the histones. Since the autoradiograph is seen over some nuclei only, and not over all (Fig. 7), it is unlikely to be due to ^^S which has been exchanged for stable sulphur in compounds present in all cells. For the same reason, it is unlikely to be due to proteins which have been precipitated on to nuclear structures. A reasonable possibility thus exists that the autoradiographs are due to ^^S in the form of nucleoprotein which has been synthesized during the 20 hours of treatment. The likelihood that this is true is increased by the fact that the percentage of cells of different classes which show nuclear autoradiograph after 20 hours in ^^S is similar to the percentage which show 'V ,^3 %* • .». ■ . "If <» • < ¥ • . O o "? m c t» o cS c f^ "a ii 0; ^ M J 0) • O — XI r^ ^^ ^ cS 5C' ^ 5^ bX) -w^ O I— 1 \ ^ -5 A c3 O CO +-' 3 • 03 ;h ^ O O ^ (M xn •\ O ^^ +J in c3 cr ;h C/3 +J fl a O o a; -M OJ .^ t» 03 "E. »— 1 4-> ■tJ 3 o O o • ^ i> C ■ a> o^ 03 [To /ace pa^e 144 'Ss \ G ■ ^*». c . 1 .,, c3 ^ ^ c 3t o o •^ ' o ' ■5 S3 O A-i +J i: ^ & c« u +-> > X C ^ +-' o O ^ o ^^ 0) "aD & IX! O r— ^ 2 ■^ a> ^^M o 4^ 4-> 0; Cm ^ 'V +-< +- ^ C O o 1— o c • o ^ 0; ^ c X "H, in -M 1 in ^ ■M n o ^ ^ o ~ 0; O 2 -p o ^^ u ^ ^^ X % o s cC Is a; s- +j lU a> > • W o: 7^ ^ ^H , . +i -M o F— H ) X ^ ^ ox DNA Synthesis 145 it after 20 hours in ^sp (Howard and Pelc, 19516). This would be expected if the ^^S is present in protein attached to DNA and if the two parts of the nucleoprotein entity are syn- thesized at the same time in the nuclear cycle. lOO 80 7 /o ARG 60- 40- 20- III. Effects of X-Radiation Roots which w^ere treated as described above were given 140 r. of X-rays (190 kV., | mm. Cu +1 mm. Al) immediately before a 24-hour period in ^^P. These roots showed a marked reduction in DNA synthesis when compared with unirradi- ated controls (Fig. 9). The number of positive cells was re- duced in all classes, and also the strength of autoradiograph which appeared over positive nuclei (about 24 grains per nuc- leus in irradiated roots compared with about 65 in unirradiated). In contrast with this obser- vation, we have not, in pre- liminary experiments, been able to detect any effect of a similar dose of X-rays on the nuclear uptake of ^^S during 20 hours. If nuclear uptake of ^^S is due to synthesis of nucleoprotein, this means that this protein can be synthesized in the absence of DNA synthesis. n / / / ; / / / / / / / / / / / / / / / / i / / / /-, / / AC MR D Fig. 9. Percentage of cells showing autoradiograph after 140 r. X-rays followed by 24 hrs. ^^p. Hatched columns: unirradiated controls. Empty columns: irradiated. AC. All cells in the final 2 mm. of the root. Meristematic resting nuclei. Dividing cells. MR D. Discussion 1. Time of Synthesis of DNA While earlier work (Caspersson, 1939; Ris, 1947) suggested that DNA is synthesized by cell nuclei during prophase, three authors have recently put forward evidence to the contrary. All have used photometric analysis of Feulgen stained material to provide relative estimates of the amount of DNA ISOTOPES 11 146 Alma Howard and S. R. Pelc present in nuclei. Seshachar (1950) observed that in the micronucleus of Chilodonella uncinatus the DNA content of daughter nuclei soon after division was about one half that of interphase nuclei. This would mean that the synthesis of DNA Avas already complete before prophase began. Swift (1950a) examined the pronephros of Ambystoma larvae, where there is considerable mitotic activity, and found that the earliest recognizable prophase stages had twice as much DNA as most of the resting cells. He concluded that DNA is built up in the interphase, and that prophase is initiated when the doubled amount has been reached. In plant tissues having a high mitotic rate, Swift (19506) found essentially the same picture. Early prophases had twice the telophase amount of DNA, and the interphase values fell between the two. In the root of Zea mays most interphase nuclei showed the double amount, a fact which suggests that synthesis takes place soon after telophase and that there may be a lag between the end of synthesis and the beginning of prophase. Lison and Pasteels (1951) have recently put forward evidence that in the segmenting egg of Paracentrotus lividus synthesis of DNA occurs during the reconstruction of the daughter nuclei at telophase. In an earlier paper (1950) they had concluded that in certain vertebrate embryonic and adult tissues the DNA content of the daughter nuclei is doubled between anaphase and the end of telophase, and they suggest that this observation may have general validity. From the observations described here and summarized in Fig. 2, it is plain that in the meristem of Vicia faha the synthesis of DNA occurs during interphase, and that a time interval of about 6 hours exists between the end of svnthesis and the beginning of visible prophase. It follows that the morphological appearance of chromosomes is not an imme- diate expression of the synthesis of new DNA molecules, although the arrangement of these molecules may alter as the chromosomes become visible. In considering the significance of the conclusions of Seshachar, Swift and ourselves, it should DNA Synthesis 147 be borne in mind that "prophase" is a descriptive term relating to a certain morphological condition of the chromosomes. The time of synthesis of DNA may well be a more significant point in the cell cycle with regard both to the metabolism and the reproduction of the nucleus, and it is possible that in different tissues this synthesis and the onset of visible prophase may occur at relatively different times in the cycle. 2. The Effect of X-Rays In 1944 and later papers Hevesy (1948) and his colleagues reported that X-rays decrease the rate at which ^^P is incor- porated into DNA. This is true for growing rats, for several normal organs of adult rats, and for Jensen rat sarcoma: the percentage formation of new DNA is similarly depressed in all tissues studied. This effect of X-radiation on the synthesis of DNA was also found by Holmes (1947) in Jensen rat sarcoma and in rat gut. Our results (Fig. 9) show that X-rays decrease both the amount of DNA synthesized in cell nuclei, and the number of cells which synthesize it in observable amounts, during a period of 24 hours after irradiation. Since only cells preparing for division synthesize DNA in this tissue, there can be little doubt that the depression of this synthesis is one of the factors responsible for the known effect of irradiation in delaying the entry of cells into mitosis. The apparent absence of X-ray effect on nuclear uptake of ^^S brings up the possibility that some of the mitotic and chromo- somal abnormality to be seen after irradiation may be due to a disturbance in the relative rates of synthesis of DNA and protein. This possibility forms the basis of further experi- mental work now in progress. REFERENCES Caspersson, T. (1939). Chromosoma, 1, 147. DoNiACH, I., and Pelc, S. R. (1950). Brit. J. Radiol, 23, 184. Gray, L. H., and Scholes, M. E. (1951). Brit. J. Radiol, 24, 82. Hevesy, G. (1948). Radioactive Indicators. New York: Interscience. Holmes, B. E. (1947). Bril J. Radiol, 20, 450. Howard, A., and Pelc, S. R. (1951a)- J- exp. Cell Res., 2, 178. 148 Alma Howard and S. R. Pelc Howard, A., and Pelc, S. R. (19516). Nature, Loud, (in press). LisoN, L., and Pasteels, J. (1950). Arch. Biol., 61, 445. LisoN, L., and Pasteels, J. (1951). Arch. Biol., 62, 2. Ris, H. (1947). ColdSpr. Harb. Sym. quant. Biol, 12, 158. Seshachar, B. R. (1950). Nature, Lond., 165, 848. Swift, H. (1950a). Physiol. Zool., 23, 169. Swift, H. (19506). Proc. nat. Acad. Sci., Wash., 36, 643. DISCUSSION Heidelberger: At the McArdle Laboratory at the University of Wisconsin, Price and Laird have obtained results in regenerating rat liver almost exactly parallel to yours. They showed by cell fractiona- tions and chemical analyses that the amount of DNA doubles just before the maximum rate of mitosis, so that the plant and animal pictures seem to be very similar. Howard: Yes, this comes out also in recent work of Swift using photometric measurement of Feulgen-positive material in the nucleus both in plant and animal cells. In this work the conclusion is inescapable that the synthesis must be during the resting stage, and possibly also during the first part of the resting stage in the case of the plant cells, with an interval between the end of the synthesis and the beginning of division. Leblond: This work is very important from both the practical and theoretical points of view. The results raise some doubt regarding contemporary theories of nucleic acid synthesis (Caspersson and others). Even the interpretation of theories of the origin of cancer may have to be revised. I would like to ask a question similar to that of Dr. Heidelberger, since in 1948 I published with Dr. Stevens a paper in which we showed the entry of phosphorus into animal cell nuclei. The reaction of individual nuclei was observed in the intestinal epithelium. By sacrificing the animals at various times after injection we could see the nuclei climbing along the villi of the intestine and gradually moving from the crypts of Lieberkiihn up to the villi of the intestine. We calculated in three different ways the time taken for the nuclei to ascend the villi. First, on the basis of mitotic counts after colchicine, it was estimated that in the rat the rate of renewal of the intestinal epithelium was of the order of 1 • 5 days. The second method was similar to that of Hevesy, estimating the entry of radio-phosphorus in the deoxyribo- nucleoprotein. By that method we obtained a figure that was a little over 2 days. The third method consisted in estimating on ^^p auto- radiographs the speed at which nuclei climbed to reach the tip of the villi, and the figure that was obtained was also over 2 days. We thought that the methods were somewhat rough and maybe the dis- crepancy between mitotic counts (1-5 day) and ^ap methods (over 2 days) was due to error variations of one kind or another, but the work of Dr. Howard would suggest that possibly it may be a true discrepancy. Perhaps the cells moving out of the villi soon after the injection are not DNA Synthesis 149 labelled. The only labelled cells would be those that move out of the villi after a lapse of time of about 6 to 8 hours. Then the level at which we get the autoradiograph reaction would be about 6 or 8 hours behind what it should be. And that would explain the discrepancy between the figure we obtained with the autoradiographic method and with the other two methods. Popjak: Do you know what happens to the other phosphate com- pounds during mitosis? Does, for example, inorganic phosphate penetrate the cell or cell nucleus during mitosis? Howard: I don't know. Our work hasn't covered any phosphorus compounds other than DNA, and we don't know at all from our own work in what form the phosphorus goes into the nucleus. Popjak: There is a possibility that during mitosis the cell nucleus maintains a peculiar autonomy; i.e. that it metabolizes only those substances which it has accumulated before, and that there is actually no penetration of phosphorus into the nucleus during mitosis. Howard: We don't have any uptake in the DNA fraction during the division. Pelc: There is some indication about that point in our tissue culture experiments. After extraction with 80 per cent alcohol only (which takes out most of the inorganic phosphorus but still leaves quite a lot of other compounds, such as the nucleic acids) we can see that some of the nuclei do not contain phosphorus, while others do. It seems from our experiments that this autonomy of the nucleus would have to start some hours before visible prophase. Leblond: This seems to be an extremely important point. Since the earliest work of Hevesy there has been a discrepancy between the percentage of DNA present in the nucleus and the rate of mitosis. In organs like the intestinal mucosa and the spleen the discrepancy is small, but in the case of the liver the discrepancy is large. To give you the precise figure, in the adult rat liver the percentage of cells that go into mitosis per day has been calculated in our laboratory to be about 0- 1 per cent of the cells, while the renewal rate, calculated on the basis of DNA, is of the order of 1 per cent, and in some authors' work, up to 2 per cent and more. This discrepancy has not been explained, and the situation has been further complicated recently by the study of the incorporation of substances other than phosphorus into the DNA. Thus, the work of Brown and associates indicated that the entry of adenine into the nucleoprotein corresponds far more accurately to the true rate of mitosis (about • 1 per" cent renewal per day). With [i^N] glycine Reichard found values of about 0-7 per cent per day, similar to those obtained with radiophosphorus. Finally, with [^^C] glycine (LePage and Heidelberger) the incorporation is so great that calculations of renewal rate make no sense. The type of explanation that has been proposed is that some parts of the molecule may be renewed either by exchange or true chemical processes, but that the bulk of the molecule is not. Brown: In regard to the point about adenine, in our original work with adenine it was shown that there is little incorporation of adenine 150 Alma Howard and S. R. Pelc into the DNA of non-growing liver, but that there is extensive incor- poration into the DNA of regenerating liver, and in long-term survival experiments there is extensive retention of that adenine once incor- porated. That has been rather hard to correlate with the original observations of Hammarsten with [^^N] glycine and the subsequent observations of LePage and Heidelberger with [^*C] glycine, which enters the DNA of non-gTowing tissues. Currently we have carried out an experiment in which [^^N] glycine and [^^C] adenine have been fed simultaneously, both to normal rats and to rats that had been partially hepatectomized. The combined results confirm all the previous experiments, that is, a certain amount of glycine will go into the DNA when adenine does not; and that the relative incorporation of the two compounds into the two types of nucleic acids in rest and into the two nucleic acids in regeneration are not parallel to one another. I think the simplest thing is to say that there are two mechanisms for formation of DNA. I would hesitate to call it two DNA's, but it might be "core" versus side-chains. Certainly there is an incorporation of glycine, formate, and other purine precursors which is not parallel to the incorporation of preformed purine given to the animal. We have also extended the adenine retention data to a 96-day experiment, at which time the DNA adenine, incorporated during regeneration, still contained 1100 counts, and by that time the isotopically labelled PNA adenine had been lost by dynamic equilibrium down to 7 counts, thus confirming a long term stability for at least some DNA. Leblond: Does your work suggest an explanation for what has been observed with phosphorus? Brown: I would say that the phosphorus and the adenine are fairly well correlated, as can be the results with glycine and formate. We have also given formate to normal animals, with non-growing liver, and the formate which is incorporated into the 2 and 8 positions of the DNA purines also disappears as though it were in dynamic equilibrium. At present the simplest thing is to say that there is some dynamic DNA, some static DNA. The adenine and the phosphorus measure chiefly the static DNA, which Dr. Howard has indicated is only formed prior to mitosis. I hesitate to refer to two molecules; it can be two portions of the same molecule. Holmes: If you use phosphorus in brain tissue, w^here there is prac- tically no mitosis, you get very little incorporation in the DNA. You would agree with that? Brown: We haven't done anything on brain tissue, but since there is very little cell division, that would fit the picture that phosphorus and the preformed purine are giving one type of result and the purine pre- cursors are giving another type. Heidelberger: I have a couple of slides that I hope to show to- morrow on this glycine incorporation that Dr. Brown just mentioned. We get very extensive incorporation of glycine into the purines of the deoxyribonucleic acid of resting liver. Howard: I think that the question of the discrepancy between the amount of DNA and the mitotic rate is due in part to the fact that, DNA Synthesis 151 particularly in liver, there are different classes of nuclei, some of which contain twice as much DNA as the usual amount, others which contain four times as much, and others which contain eight times as much. This has come out of recent work by Swift, and by Leuchtenberger and the Vendrelys. Therefore the synthesis of DNA must be related not only to mitosis but also to endomitosis, or whatever the process is which gives rise to this doubling of DNA content, presumably associated with doubling of the chromosome number. Other organs such as the kidney show far less markedly this appearance of high DNA content. This might explain some of the difficulties associated with work on DNA synthesis in so-called non-mitotic tissue or tissues in which the visible mitotic rate is very low. McFarlane: I am particularly interested to learn that both investi- gators find that DNA synthesis is depressed after radiation, but can find no similar effect on the synthesis of proteins associated with the nucleic acid. I do not know of any figures for the histones of the nucleus either in resting or dividing cells, but certainly in normal tissues the turnover period of the DNA is of the order of days, while that of the cytoplasmic proteins is of the order of hours. May this not explain why you find no radiation effect on intracellular protein metabolism — namely that the effect is evident only in a much shorter time interval than your experiments cover? Holmes: I think that is so. I think it is obliterated by the tremen- douse amount of turnover that goes on, quite apart from any new synthesis. Professor Hevesy suggested that when I first told him about it. On the other hand, if this turnover is going on just as fast as ever after X-rays, I should think it must produce curious results in the cell. I have been dealing with the histones, having been forced to discard the other chromosome proteins, because you can't do both, and I think almost certainly Miss Howard has been dealing with the other proteins of the chromosome and washing out the histones; thus we seem to have both proteins represented in the work. But I do think it is possible that it is just such a big general turnover that any extra bit due to the formation of a new nucleus wouldn't necessarily show, but that the X-rays don't slow up that rate of turnover. Pelc: But there are differences in our results which I don't think can be explained by turnover. After 20 or 24 hours many of the elongated cells show no autoradiographs of the nucleus. The amount of ^^S in the protein of the differentiated cells should indicate the turnover rate, since ^ss was available to these cells. The only thing which might make any difference there is the change of permeability of the membrane, a point about which we know very little. PART IV NUCLEIC ACIDS THE BIOSYNTHESIS OF PYRIMIDINES IN VITRO* D. WRIGHT WILSON The metabolism of the pyrimidines seems to be quite different from that of the purines. For many years it has been known that while purines are excreted as uric acid or allantoin, the pyrimidines form neither of these but are excreted as urea and CO2. With the introduction of the use of the rare isotopes as tracers it became possible to study the precursors of the pyrimidines. It was early shown that ammonium salts labelled with ^^N could become incorporated in pyrimidines of nucleic acid (Barnes and Schoenheimer, 1943). If, however, a labelled pyrimidine, uracil or thymine or cytosine, is fed to rats, it does not appear in the nucleic acids (Plentl and Schoenheimer, 1944; Bendich et al., 1949). Purines are not the source of pyrimidines nor are pyrimidines the source of purines. Although pyrimidines are catabolized to urea, urea is not one of the building stones of pyrimidines. Heinrich (Heinrich and Wilson, 1950) has shown that CO2 is incorporated into the pyrimidine molecule in rats when bicarbonate is given. Most of it enters position 2 (the ureide carbon) although some may enter position 4 ( Lager kvist, 1950). Carbon of radioactive formate was not found in thymine. Although formate is known to be a good methyl ♦Aided by a grant from the American Cancer Society administered by the Committee on Growi:h of the National Research Council. The i*C used in this investigation was obtained on allocation by the United States Atomic Energy Commission. 152 Biosynthesis of Pyrimidines 153 donor and in our experiment was incorporated in purines to give high specific activities, not sufficient radioactivity was found in the thymine to be detected. This is not in agreement with the results of Elwyn and Sprinson (1950) who used [)S-i*C]-L-serine or [a-i*C]-glycine for methylation or those of Totter, Volkin, and Carter (1950) who used formate. There is agreement with LePage and Heidelberger (1951) who used [a-^*C]-glycine. Hammarsten and his group (Arvidson et aL, 1949) have demonstrated that orotic acid labelled with ^^N enters the pyrimidines when it is fed to rats. Mary Edmonds (1950) has obtained a similar result when yeast is grown in a medium containing orotic acid labelled with ^^C in carbon 4. Glycine carbon is not incorporated in pyrimidines. Lactate labelled in 2 and 3 positions with ^^C is incorporated probably in positions other than in the ureide group. Recently Dr. Lawrence L. Weed in my laboratory has prepared orotic acid with ^^C in position 2, and has been studying pyrimidine formation in vitro (Weed, Edmonds, and Wilson, 1950). He has found that slices of tissues, when incubated with the radioactive orotic acid will incorporate the radioactive carbon in the pyrimidines. Cy tidy lie and uridylic acids of ribonucleic acid (RNA) have been studied mainly, but thymidylic and cytidylic acids of deoxyribo- nucleic acid (DNA) have been examined to some extent. We have carried out experiments as follows: Slices of tissue suspended in Krebs saline with added phosphate at pH 7 • 4 were digested with 1-10 mg. of radioactive orotic acid at 38° for 4 hours. After digestion the sUces were washed several times, homogenized and precipitated with trichloracetic acid. The precipitate was freed from lipids by heating with alcohol-ether mixture. Nucleic acid was extracted from the precipitate with hot 10 per cent NaCl solution for 24 hours with continuous stirring. The nucleic acid was precipitated from the filtered solution with 2.5 volumes of alcohol. Sometimes DNA and RNA were separated by means of the Schmidt-Thannhauser procedure (1945). We feel that the 154 D. Wright Wilson Schmidt-Thannhauser procedure may not separate completely the RNA from the DNA. We have usually separated the DNA constituents from those of the RNA by the very effective combination of paper chromatography and ion exchange. The nucleic acids were hydrolysed by heating at 100° with 1 N HCl for 1 hour. This procedure liberated purine bases and the pyrimidine nucleotides. The separation of these compounds by means of paper chromatography was being studied when Smith and Markham (1950) reported that they could be separated by means of a 70 per cent solution of tertiary butyl alcohol in water, the w hole solution made • 8 N with HCl. The com- pounds are now separated by using a combination of paper chromatography and ion exchange and identified by means of the Beckman ultraviolet spectrophotometer. We are now able to identify the four pyrimidine nucleotides, two from RNA and two from DNA. About • 1 ml. of the hydrolysate is placed as a band near the bottom of Whatman No. 1 filter paper, 12 cm. wide and 40 cm. long and allowed to dry. The paper is set up as an ascending column, using Smith and Markham' s tertiary butyl alcohol. After 48 hours the paper is dried at room temperature and examined by the Mineralight lamp. The dark regions are circled with a pencil. From the solvent front down, the bands contain thymidylic acid, uridylic acid, cytidylic acid, adenine, and guanine. Each band is cut out and extracted with water. Each solution is made faintly alkaline and poured on a column containing Dowex-1, 300-400 mesh, which had been previously washed with alkali, acid and water. After thorough washing, the columns are eluted with varying strengths of HCl. Thymidylic acid comes off the column with • 1 N HCl, uridylic acid with the 0-01 N HCl. RNA cytidylic is eluted with -002 N HCl and after it is removed from the column 0-01 N HCl will remove DNA cytidylic acid. The adenine and guanine are not absorbed by the column. It has been demonstrated that the added orotic acid does not contaminate the isolated products (Weed and Wilson, Biosynthesis of Pyrimidines 155 1951). We calculate that the washings alone in our procedure will leave less than 1 • 10* per cent of the original orotic acid. We have demonstrated that the paper will separate orotic from uridylic acids (Table I) and the column will separate orotic and cytidylic acids. Table I Separation of Uridylic and Cytidylic Acids from Radioactive Orotic Acid Exp. Substance Actual count cts.lmin. 1 2 3* Uridylic acid Orotic acid Cytidylic acid Uridylic acid Orotic acid Uridylic acid Orotic acid Cytidylic acid 17 31,400 26,700 5 656 23 618 4 *The cytidylic acid was removed from paper and run through a resin column before the radioactivity was determined. The isolations are moderately quantitative. Table II shows recoveries of pyrimidines and pyrimidine nucleotides as determined by ultraviolet absorptions. They were from 90 per cent to 100 per cent. The combination of paper chromatography and ion ex- change is important for the separation of certain constituents and for obtaining good planchejts for radioactivity determina- tion. In our experiments with radioactive orotic acid no radioactivity was found in the purines and all the activity of the uridylic and cytidylic acids was found in the pyrimidine bases. It is, therefore, reasonable to conclude that the orotic acid enters the pyrimidines as a whole and does not break up to give appreciable amounts of radioactive CO2 in our experiments. 156 D. Wright Wilson Table II Recovery of Pyrimidines put on Paper Exp. Substance Wt. used Wt. recovered tig- Recovery % 1* 2* 3* 4* Uracil Thymine Uracil Thymine 20 20 20 11 19-6 180 200 10-8 98 90 100 98 Recovery of Pyrimidine Nucleotides Put on Paper Exp. Substance Wt. used fig- Wt. recovered tig- Recovery % Ratio 1* 2* 3t Uridvlic acid . Cytidylic acid Uridylic acid . Cj'tidylic acid 79 82 79 82 76 80 78 82 92 97 98 100 0-46 1-64 0-46 1-63 ♦The substance run on paper separately. •fA mixture of the substances run on single paper, producing two separate bands which were eluted separately. Ratio = Reading at 278 m^/Reading at 262 m/x. Table III shows the type of data which we have obtained. The specific activities are accurate to within 5 per cent except where the counts were small, as shown in later tables. Within each part of our experiment, the relative weights Table III Normal Rat Liver Incubated with 4 mg. of [2-^*C] Orotic Acid Part Substance Amount lig- Actual count cts.jmin. Cts.jmin. Img. Base A* Bt Uridylic acid .... RNA cytidylic acid Adenine Guanine Uridylic acid .... RNA cytidylic acid 518 978 232 284 1742 1975 112 57 312 112 629 170 522 166 ♦Separation with paper followed by resin purification of each band. fSeparation on resin alone. Biosynthesis of Pyrimidines 157 recovered may be of some value. In this experiment, one may note that the specific radioactivity calculated per milli- gram of base was much higher for uridylic than for RNA cytidylic acid, the former being 3-4 times higher than the latter. There was isolated about twice as much cytidylic acid as uridylic acid. This was true for the separation by resin alone as well as for the separation by paper and resin. We have come to believe, however, that the combination of paper and resin is the most satisfactory procedure. Some results of our studies on tumour tissue from rats are given in Table IV. The specific activity of uridylic acid was Table IV Rat Walker Carcinoma 256 Incubated with 3 mg. of [2-^^C] Orotic Acid Substance Amount Actual count cts.jmin. Cts.jmin. (mg. Base Part I, Control Thymidylic acid Uridylic acid RNA cytidylic acid 224 622 1410 3-2 195 48 37 920 99 Part II. Methopterin Thymidylic acid Uridylic acid RNA cytidylic acid 186 750 1150 4.0 233 42 56 905 107 about 9 times that of the cytidylic acid, although, again, there was twice as much cytidylic as uridylic acid. A small amount of thymidylic acid was obtained which had a very small specific activity. In part II of the experiment, alternate slices of tissue were taken, a small amount of methopterin added, and a parallel digestion and determination carried out. The results indicate, we believe, that methopterin had no effect and the data therefore show the accuracy of our duplicate determinations. 158 D. Wright Wilson Table V shows the specific activities of compounds isolated after incubating some human tumours with radioactive orotic acid. The specific activity of uridylic acid in the first experiment was the highest we have obtained but, as different Table V Human Tumours Incubated with 3 mg. of [2-i*C] Orotic Acid Fibro- sarcoma Gastric Carcinoma C cecal Carcinoma Teratoma Specific activity — cts.jmin.jmg. base Thymidylic acid Uridylic acid RNA cytidylic acid Ratio u/c .... 124 2300 348 7 1130 122 9 1033 123 8 32 1070 94 11 48 1346 187 7 tissues vary considerably, we do not feel that it is necessarily characteristic of this tissue. However, the ratios of specific activities of uridylic and RNA cytidylic acids seem to be significant for tumours. The ratios for tumours range from 7 to 11, while for normal tissues the ratios range from 2 to 4, although normal spleen gave ratios as high as 7. The specific activities of thymidylic acid were low. The specific activities of the bases from the nucleic acids of regenerating liver are given in Table VI. Regenerating liver has been studied because it is a rapidly growing tissue, as are Table VI Regenerating Liver Incubated with 3 mg. of [2-i*C] Orotic Acid Substance Amount Actual count cts.lmin. Cts.lmin. Img. Base Thymidylic acid Uridylic acid RNA cytidylic acid . 136* 63t 2680 4770 7-8 60 578 352 149 246 630 216 ♦Eluted from Resin column with 01 N HCl. tEluted from Resin column with 01 N HCl. Biosynthesis of Pyrimidines 159 tumours. The ratio of activities of uridylic to cytidylic acids is 3, the same that we have obtained for normal hver and much less than for the tumours. On account of the larger amount of thymidylic acid obtained from this growing tissue, the thymidylic acid was collected in two fractions. Although only small counts were obtained the specific activities were higher than usual. Furst, Roll and Brown (1950) have found a similar result with purines. Data from another normal tissue, a cat spleen, are shown in Table VII. Here the ratio of specific activities of uridylic and RNA cytidylic acids was 6. The specific activities of the DNA cytidylic acid and thymidylic acid were very low. Table VII Normal Cat Spleen Incubated with [2-i^C] Orotic Acid Substance Amount Y Actual count cts.jmin. Cts.jmin. jmg. Base Thymidylic acid Uridylic acid RNA cytidylic acid . DNA cytidylic acid . 243 461 812 176 1 183 55 2 11 1160 199 31 One may not draw any startling conclusions from these data. However, the following comments may be made: We have developed a method for determining readily the incorporation of radioactivity into pyrimidines of the nucleic acids after incubation of tissue slices with radioactive orotic acid. The orotic acid goes into the pyrimidines without being degraded to COg. The specific a^ctivity of resulting uracil is from 3 to 11 times greater than that of the RNA cytidylic acid or cytosine. This ratio is associated with a concentration of cytidylic twice as high as that of uridylic acid. The differences in incorporation of orotic acid in uridylic and cytidylic acid emphasized the uncertainty of studies where incorporation in total DNA is compared with incorpora- tion in total RNA. These considerations lead to the question 160 D. Wright Wilson whether the nucleic acids are built up from small units de novo or whether the bases may go in and out of the nucleic acid without its complete disruption. The pyrimidine bases of the DNA have always shown far less incorporation of radio- activity than those of RNA. However, in the case of regener- ating liver the incorporations in the DNA pyrimidines were unusually high. We have studied other intermediates besides orotic acid (Table VIII). I wish to report here an experiment carried out in collaboration with a team under the direction of Lemuel Wright (1951). Ureidosuccinic acid was fed to a strain of Lactobacillus bulgaricus which required orotic acid, Table VIII Growth Factor: Orotic Acid ' Cts.jmin.l m. mole Uridylic acid Cj'tidylic acid 7500 7950 Growth Factor: Ureidosuccinic Acid Uridylic acid Cvtidine 7600 7000 in amounts sufficient to obtain a rate of growth similar to that brought about by orotic acid. The ureidosuccinic acid had the same specific activity as did the orotic acid. Bacteria which were grown in such a medium contained uracil and cytosine having the same specific activities as they had when the bacteria were grown with radioactive orotic acid. It would appear that ureidosuccinic acid is a precursor of pyrimidines in these bacteria. With S. S. Cohen (Weed and Cohen, 1951) we have studied the' incorporation of radioactivity into bacteria and viruses. See Table IX. When E. coli is grown in a medium containing radioactive orotic acid, radioactivity is incorporated in the Biosynthesis of Pyrimidines 161 Table IX DNA Components of Bacteria Substance Cts.jmin.lmg. base Thvmidvlic acid 2780 3420 DNA Cvtidvlic acid DNA Components of Virus Substance TQR^ Natural Lysis — 90 min. Cts.lmin.jmg. base T6R"^ Artificial Lysis — 35 min. Cts.jmin.lmg. base TQR Natural Lysis — .35 min. Cts.jmin.lmg. base Thymidylic DNA cytidylic .... Relative DNA increments mg./5 ml 217 401 0-208 894 1490 048 766 1252 075 nucleic acid. We report the incorporation in DNA pyrimi- dines. The RNA has been shown to be completely inert in the infected cell. DNA cytidylic acquired slightly more specific radioactivity than did thymidylic acid. These bacteria were then washed, suspended in a non- radioactive medium and infected with virus. The problem we wished to study was this: Do the virus particles which are synthesized early receive a greater proportion of host DNA than those formed later, or do all particles receive a constant proportion of material from host DNA and medium? Three portions of virus were prepared. Some bacteria were infected with T6r virus which caused lysis of bacterial cells after 35 minutes. Two other portions of radioactive bacteria were infected with T6r+ virus. One portion was allowed to lyse the cells naturally (this started in 90 minutes) and the other portion was treated with cyanide after 35 minutes. This caused cessation of metabolism and lysis of bacterial cells. The amount of newly formed DNA in the virus was deter- mined in each case. The values are found at the bottom of the table. The smallest amount of DNA formed is associated ISOTOPES 12 162 D. Wright Wilson with the maximum incorporation of radioactivity in the DNA pyrimidines of the virus. Thus as the DNA of the virus increased 4 times (from 0-048 to 0-208) the specific activity of each of the DNA pyrimidines was reduced to one-quarter. The experiments are not in agreement with the idea that the host furnishes a definite amount of its preformed DNA to each virus particle. They may be explained by saying that the virus DNA is formed from the total available pool of material, which diminishes in radioactivity as more DNA is synthesized. REFERENCES Arvidson, H., Eliasson, N. A., Hammarsten, E., Reichard, P., von Ubisch, H., and Bergstrom, S. (1949). J. biol. Chem., 179, 169. Barnes, F. W., Jr., and Schoenheimer, R. (1943). J. hiol. Chem.y 151, 123. Bendich, a., Getler, H., and Brown, G. B. (1949). J. biol. Chem.^ 177, 565. Edmonds, M. P., Delluva, A. M., and Wilson, D. W. (1950). Fed, Proc, 9, 167. Elwyn, D., and Sprinson, D. B. (1950). J. Amer. Chem. Soc., 72, 3317. FuRST, S. S., Roll, P. M., and Brown, G. B. (1950). J. biol. Chem., 183, 251. Heinrich, M. R., and Wilson, D. W. (1950). J. biol. Chem., 186, 447. Lagerkvist, U. (1950). Acta chem. Scand., 7, 1151. LePage, G. a., and Heidelberger, C. (1951). J. biol. Chem., 188, 593. Plentl, a. a., and Schoenheimer, R. (1944). J. biol. Chem., 153, 203. Schmidt, G., and Thannhauser, S. J. (1945). J. biol. Chem., 161, 83. Smith, J. D., and Markham, R. (1950). Biochem. J., 46, 509. Totter, J. R., Volkin, E., and Carter, C. E. (1950). Abstracts, Division of Biological Chemistry, American Chemical Society, 118th Meeting, Chicago, 55C. Weed, L. L., and Cohen, S. S. (1951). Unpublished data. Weed, L. L., Edmonds, M. P., and Wilson, D. W. (1950). Proc. Soc. exp. Biol., N.Y., 75, 192. Weed, L. L., and Wilson, D. W. (1951). J. biol. Chem., In press. Wright, L. D., Miller, C. S., Skeggs, H. R., Huff, J. W., Weed, L. L., and Wilson, D. W. (1951). J. Amer. Chem. Soc. In press. DISCUSSION Hammarsten: We recently tried orotic acid on liver slices and found that part of it was incorporated in the pyrimidines. Of course, a large part was combusted, and the part escaping combustion and incorporation Biosynthesis of Pyrimidines 163 was diluted. That might indicate that orotic acid is a normal inter- mediate. It has to be confirmed in other ways. We also tried carbon- labelled aspartic acid (glutamic M^as not used) and carbon from this also was incorporated into the pyrimidines. This of course doesn't mean that aspartic acid is a precursor to orotic acid. We don't know anything about that. At the moment we are trying pyruvic acid. Aterman: I would like to point out that the regenerating liver may not be a normal organ. I have slides which show that within 5 minutes after partial hepatectomy in the rat there is in the liver remnant a type of vacuolation which is exactly identical with the vacuolation described in hypoxic states, and I should think that functional changes must occur as well. Neuberger: How are the samples to be counted prepared, and what amount of material is counted? Wilson: They are obtained in solution, of course, from the columns. The solution is evaporated down practically to dryness. Then single drops of water are added to the little evaporating dish, the liquid is taken off, placed on an aluminium planchet and evaporated again to dryness. The great advantage of using the column over the paper alone is that the solution is much cleaner when the column is used. If we evaporate the solution from the paper, we have never been able to purify the tertiary butyl alcohol enough so that there is no residue at all; there is usually some residue from the alcohol and a little residue from the filter paper. The amounts of pyrimidine nucleotides placed on the planchets were from 100 to 1500 /ng., as indicated by the spectrophotometric absorptions. STUDIES WITH ORGANIC- AND BIO-SYNTHETIC NUCLEOSIDES AND NUCLEOTIDES G, B. BROWN* Of nine labelled purines thus far studied only two have been found to be utilized by the rat for the synthesis of nucleic acids: adenine (Brown, Roll, Plentland Cavalieri, 1948) which may serve as a precursor of both adenine and guanine of the polynucleotides, and 2,6-diaminopurine (Bendich and Brown, 1948) which leads only to polynucleotide guanine. The earlier studies of Plentl and Schoenheimer (1944) with [2-amino-l,3-^^N3] guanine showed that it was not utilized by the rat to a significant extent and this conclusion was later affirmed in our laboratory. It was subsequently demonstrated, however, that this guanine could be utilized by the C57 black mouse to an appreciable degree (Brown, Bendich, Roll and Sugiura, 1949). In view of this utilization of guanine by the mouse, and also its extensive utilization by yeast (Kerr, Seraidarian and Brown, 1951) and Lactobacillus casei (Balis, Brown, Elion, Hitchings and Vanderwerff, 1951), the fate of guanine in the Sherman rat was reinvestigated with highly active [8-^*C] guanine. This more sensitive tracer not only permitted detection at a greater dilution but also avoided the general contribution of ^^NHg to body pool ammonia which obscured any small specific incorporation of the intact purine. This ^*C labelled guanine was incorporated into the pentose poly- nucleotide guanine of the rat at a 1:1,000 dilution (Balis, Marrian and Brown, 1951), which is about one per cent of the ♦The author wishes to express appreciation to the several collaborators responsible for the studies summarized here and for the support of the National Cancer Institute of the United States Public Health Service, Tlie United States Atomic Energy Commission, Contract AT (30-l)-910, The Nutrition Foundation, Inc., and the James Foundation of New York, Inc. 164 Synthesis of Nucleosides and Nucleotides 165 extent to which adenine is incorporated under similar con- ditions. This trace utihzation of guanine now makes the difference between the rat and the mouse a large quantitative one rather than the heretofore apparent qualitative one. These species differences also indicate that caution must be exercised in drawing generalizations in this field. In the survey of the utilization of various purines by the rat the individual purines were isolated from the tissue nucleic acids. However, upon the recognition by Carter and Cohn (1949) that two isomeric adenylic acids and two guanylic acids could be obtained upon alkaline hydrolysis of pentose nucleic acids, it became of interest to know whether these isomers could be differentiated on any metabolic basis. The incorporation of [1,3-^^N2;8-^*C] adenine into the indi- vidual nucleotides (Marrian, Spicer, Balis and Brown, 1951) was studied and the results did not indicate any preferential incorporation into either of the isomers of adenylic acid or of guanylic acid. In those experiments it was also demonstrated that there was a parallel incorporation of the two labelling isotopes into each of the purines, and this further sub- stantiated an earlier conclusion that the purine ring is retained intact during the process of transformation of adenine into polynucleotide guanine. Approaches to the Biosynthesis and Synthesis of Nucleosides and Nucleotides Our programme is aimed at the elucidation of the immedi- ate precursors of nucleic acids in the hope of providing a rational approach to the design of antimetabolites of possible chemotherapeutic import. We have therefore wished to investigate the utilization of labelled nucleosides and nucleo- tides, compounds which are intermediate in size between the bases and the polynucleotides. The first approach to these compounds involved the growth of yeast in a medium containing isotopic ammonia. The yeast (pentose) nucleic acid was then isolated and four nucleotides have now been isolated by chemical procedures 166 ''* G. B. Brown from a single sample of this ^^N-labelled nucleic acid. It should be noted that these are mixtures of the isomeric nucleotides demonstrated by Cohn and this fact must be kept in mind when assaying the biological results obtained. From the mode of production these nucleotides are labelled in all of their nitrogens, including the substituent amino groups. The group at Karolinska Institute has also utilized this method of biosynthesis and has prepared the ribonucleosides (Hammarsten, Reichard and Saluste, 1950), and more recently they have extended this to the preparation of deoxy- ribonucleosides from the DNA of Escherichia coli (Reichard and Estborn, 1951). For certain purposes the homogeneously labelled derivatives will be of limited usefulness, and we have also investigated another biosynthetic approach which permits the production of specifically labelled nucleosides or nucleotides. When yeast (Torulopsis utilis) was grown in the presence of ^*C labelled purines, either adenine or guanine was efficiently utilized by the yeast and essentially all of the administered radioactive purine could be recovered from the nucleic acids. By the application of the ion-exchange method for the separation of the nucleotides (Cohn, 1950), each of the individual nucleotides may be recovered. The corresponding nucleosides may be prepared from them by enzymatic hydro- lysis and ion-exchange purification (Kerr, Seraidarian and Brown, 1951). The specificity of the labelling of derivatives prepared by this type of biosynthesis is dependent only upon the position of the label in the purine furnished, and there is little waste of isotope. Small samples of labelled adenosine, adenylic acid, guanosine and guanylic acid, each labelled in the 8-position of the purine ring, have been so prepared. The biosynthetic approaches are cumbersome for the production of large samples, and offer no possibility of obtain- ing the derivatives of the 2,6-diaminopurine in which we are interested. We were therefore interested in total synthesis Synthesis of Nucleosides and Nucleotides- • 167 and Dr. Davoll has been investigating the possibiUties of practical syntheses of some of the desired compounds. The pubhshed routes (Davoll, Lythgoe and Todd, 1948) for the synthesis of adenosine and guanosine are via uric acid, and they can be expected to result in over-all yields from inorganic carbon of about 2 • 5 and • 5 per cent, respectively. For this reason it was deemed desirable to study possible direct syntheses from the parent purines. Briefly this study has culminated in the use of the chloromercuri derivatives of acyl-substituted aminopurines and acetochloro-D-ribofuranose (Davoll and Lowy, 1951). By this method adenosine has been obtained in a yield of 20 per cent from adenine. The application of this procedure to 2,6-diacetamidopurine has resulted in a synthesis of 2,6-diamino-9-j3-D-ribofuranosyl- purine, the nucleoside which corresponds in configuration to natural adenosine. This nucleoside could be partially deaminated by the action of nitrous acid, in analogy to the chemical behaviour of 2,6-diaminopurine, to yield 6-amino-2- hydroxy-9-/8-D-ribofuranosylpurine, or crotonoside. It has also proven possible to preferentially deacetylate the sugar and the 6-amino group of pentacetyl-2,6-diamino-9-/8-D- ribofuranosylpurine and, by subsequent treatment with nitrous acid and further deacetylation, to obtain guanosine. Metabolic Studies The initial studies with labelled materials larger than the purines or pyrimidines involved the feeding of the intact ^^N-labelled yeast nucleic acid. The mixture of mono- nucleotides derived from it was also administered by intra- peritoneal injection (Roll, Brown, DiCarlo and Schultz, 1949). There was essentially no incorporation of the purines of the orally administered nucleic acid into the tissue polynucleo- tides, and although there was a definite incorporation of the purines of the intraperitoneally administered mixture of nucleotides, it was obvious that these were much less effective- ly utilized than was free adenine. It was significant that either the intact nucleic acid or the mixture of nucleotides 168 G. B. Brown did furnish a precursor of the polynucleotide pyrimidines, since none of the individual pyrimidines occurring in the nucleic acids has been found to be utilized by the rat. Almost simultaneously Hammarsten, Reichard and Saluste (1949, 1950) studied cytidine and uridine, also prepared from yeast grown with isotopic ammonia, and found that it was chiefly the cytidine which led to all of the pyrimidines of both the pentose and the deoxypentose polynucleotides. Dr. Roll's results with cytidylic acid are quantitatively almost a perfect parallel to the earlier results with cytidine as far as the pyrimidines of both types of nucleic acids are concerned. There is one interesting and perhaps very signi- ficant point of difference, in that Hammarsten and co-workers found that cytidine nitrogen was also incorporated to a very considerable extent in both the adenine and the guanine of the deoxypentose nucleic acids, while with cytidylic acid we have not obtained any evidence of such a transformation. Our first studies of the metabolism of individual purine nucleosides and nucleotides were carried out with the [8-^*C] derivatives obtained by biosynthesis in yeast. The fate of these has so far been studied only in this same yeast. Rela- tively little (3 to 10 per cent) incorporation of any of these ribose derivatives was observed, and since these small incor- porations could have been due to incorporation of degradation products, the only conclusion possible was that attachment of the ribose or ribose phosphate did not facilitate the incor- poration of the purines into the polynucleotides. In the light of subsequent findings in the rat the small utilization by yeast could perhaps be assigned greater significance. Dr. Roll has now completed studies in the rat of each of the ^^N-labelled purine nucleotides prepared from yeast nucleic acid. These were administered intraperitoneally and it was found that both the adenylic acid and the guanylic acid are utilized as precursors of the pentose polynucleotides (Roll and Weliky, 1951). The behaviour of guanylic acid, which leads only to the guanine of the polynucleotides, makes it unique among guanine derivatives since neither guanine nor Synthesis of Nucleosides and Nucleotides 169 guanosine (Hammarsten and Reichard, 1950) is utilized by the rat. Thus it would appear that guanylic acid can be incorporated directly into the polynucleotide, since prior degradation to either guanosine or guanine cannot be involved. The incorporation of adenylic acid, which leads to both the adenine and the guanine of the polynucleotide, could not in itself be of much significance since degradation to the purine, and incorporation thereof, could explain it. However, by analogy to the result with guanylic acid, it may be that both adenylic and cytidylic acids may be incorporated directly and it seems possible that nucleotides may be the immediate precursors in the biosynthesis of polynucleotides. The fact that guanylic acid is the only potential intermediate between adenine and polynucleotide guanine also suggests that the conversion of an adenine, or a 2,6-diamino-purine, into a guanine derivative may occur at the nucleotide stage. It is worthy of note that the sum of the renewals of poly- nucleotide guanine from the two sources, that is, directly from guanylic acid and indirectly from adenylic acid, is approxi- mately equal to the renewal of polynucleotide adenine from the adenylic acid alone, and that these values are in close agreement with the approximately equal renewal of the two purines observed in the original experiment with a mixture of nucleotides. The relatively low utilization of the adenylic or guanylic acids when compared to that of free adenine, or to that of cytidylic acid, may be attributed to the greater abundance of tissue enzymes capable of degrading the purine nucleotides. However, the possibility that only one of the isomers present is utilized for anabolic purposes could also help to account for these results and must be considered. These purine nucleotides led to the incorporation of the isotope into the purines of the ribose nucleic acids but there was not more than a trace of incorporation into the deoxyribo- nucleic acids. In the case of a pyrimidine ribose derivative, either cytidine or cytidylic acid, there was a considerable conversion into the deoxyribose polynucleotide pyrimidines, 170 G. B. Brown with the relative incorporation into the PNA and into the DNA cytosines being in the ratio of about 2 to 1. The con- trast between that extent of the conversion of the pyrimidine derivative and a corresponding ratio of about 8 or 10 to 1 for the purine derivatives may have a bearing on another of the currently perplexing problems in nucleic acid metabolism. The above observations, in conjunction with the facts that DNA purines are not derived from adenine (Furst, Roll and Brown, 1950) to the extent that they are derived from glycine (Bergstrand et al., 1948; Elwyn and Sprinson, 1950; LePage and Heidelberger, 1951), formate (Totter et al., 1950) and cytidine (Hammarsten and Reichard, 1950) may be related in some as yet obscure manner. The implication that the deoxyribosenucleic acid purines arise by pathways perhaps quite different from those for the ribosenucleic acid purines suggests that studies with purine deoxyribonucleotides are urgently needed. Another instance of marked dissimilaritv of results obtained with different precursors is found in the relative incorporation of [1,8-^^X2] adenine and ^^C-formate into various organs. Orally administered adenine is incorporated into liver, kidney, intestine, spleen, testis in that decreasing order (Furst et aL, 1950), with the ratio of the incorporation into liver and intes- tine being somewhat less than 2 to 1. The incorporation of intraperitoneally administered formate into the adenine of the pentose nucleic acids of various organs was in the order : intestine, kidney, spleen, liver, pancreas, testis; with a liver to intestine ratio of • 16 to 1 (Marrian, Drochmans, unpub- lished). In connection with another experiment, [1,8-^^X2] adenine and ^*C-formate were administered simultaneously by intraperitoneal injection (GoldthAvait, Bendich, unpublished) and, in confirmation of the separate experiments mentioned above, the liver to intestine ratios were 1 • 6 to 1 for the incor- poration of adenine and 0-25 to 1 for the incorporation of formate. This preferential incorporation of a purine into liver nucleic acids and of a purine precursor into those of intestine adds to the complications and points up the necessity Synthesis of Nucleosides and Nucleotides 171 for studies on individual tissues and also the possible pitfalls to be encountered in comparing results obtained with different precursors. REFERENCES Balis, M. E., Brown, G. B., Elion, G. B., Hitchings, G. H., and Vanderwerff, H. (1951). J. biol. Chem., 188, 217. Balis, M. E., Marrian, D. H., and Brown, G. B. (1951). J. Amer. chem. Soc, in press. Bendich, a., and Brown, G. B. (1948). J. biol. Chem., 176, 1471. Bergstrand, a., Eliasson, N. A., Hammarsten, E., Norberg, B., Reichard, p., and von Ubisch, H. (1948). Cold. Spr. Harb. Sym. quant. Biol., 13, 22. Brown, G. B. (1948). Cold Spr. Harb. Sym. quant. Biol., 13, 43. Brown, G. B., Bendich, A., Roll, P. M., Sugiura, K. (1949). Proc. Soc. exp. Biol., N.Y., 72, 501. Brown, G. B., Roll, P. M., Plentl, A. A., and Cavalieri, L. F. (1948). J. biol. Chem., 172, 469. Carter, C. E., and Cohn, W. E. (1949). Fed. Proc, 8, 190. Cohn, W. E. (1950). J. Amer. chem. Soc, 72, 1471. Davoll, J., Lythgoe, B., and Todd, A. R. (1948). J. chem. Soc, 967 and 1685. Davoll, J., and Lowy, B. A. (1951). J. Amer. chem. Soc, in press. Elwyn, D., and Sprinson, D. B. (1950). J. Amer. chem. Soc, 72, 3317. FuRST, S. S., Roll, P. M., and Brown, G. B. (1950). J. biol. Chem., 183, 251. Hammarsten, E., Reichard, P., and Saluste, E. (1949). Acta chem. Scand., 3, 433. Hammarsten, E., and Reichard, P. (1950). Acta chem. Scand., 4, 711. Hammarsten, E., Reichard, P., and Saluste, E. (1950). J. biol. Chem., 183, 105. Kerr, S. E., Seraidarian, K., and Brown, G. B. (1951). J. biol. Chem., 188, 207. LePage, G. a., and Heidelberger, C. (1951). J. biol. Chem.. 188, 593. Marrian, D. H., Spicer, V. E., Balis, M. E., and Brown, G. B. (1951). J. biol. Chem., 189, 533. Plentl, A., and Schoenheimer, R. (1944). J. biol. Chem., 153, 203. Reichard, P., and Estborn, B. (19^51). J. biol. Chem., 188, 839. Roll, P. M., Brown, G. B., DiCarlo, F., and Schultz, A. A. (1949). J. biol. Chem., 180, 333. Roll, P. M., and Weliky, I. (1951). Fed. Proc, in press. Totter, J. R., Volkin. E., and Carter, C. E. (1950). Abstracts, 118th Meeting, Amer. Chem. Soc, 55c. DISCUSSION Heidelberger: While we are on the subject of these different precursors of purines I would like to mention some of our experiments. 172 G. B. Brown This is work which was done in collaboration with Dr. LePage and begun about two years ago. We have been studying the incorporation of [2-^*C] glycine into the nucleic acids of rats. We gave these injections of glycine primarily to study the incorporation by tumours, and we used liver as a contrast. In addition, we separated the nucleic acids into the DNA and the RNA, and we accidentally came across a rather remarkable result, which has been obtained independently by Eh\yn and Sprinson at Columbia, and by Totter, Volkins and Carter, and also by Dr. Brown. We found an extremely high incorporation of i*C from glycine into the purines of DNA, almost identical with the incorporation into the purines of the RNA. These data were published as Table IV in the February issue of the Journal of Biological Chemistry (J. hiol. Cheryl., 1951, 188, 593). The isolation procedure was a modification of the Schmidt-Thann- hauser technique. The RNA samples according to analysis contained no DNA, and the DNA samples were contaminated with not more than 10 per cent of RNA. As Dr. Wilson mentioned, we found practically no incorporation from glycine into thymine. The specific activities in counts per minute per mg. of the RNA, adenine and guanine are compared with the DNA. The specific activities for DNA are usually somewhat lower than those for RNA, but they are all of the same order of magnitude, whereas the data Dr. Brown had previously obtained with adenine indicated a much lower incorporation, and of course the work with phosphorus indicates very little turnover of ^ap in the DNA. There is considerable variation from animal to animal, but our results are for normal livers of so-called adult rats, which showed ample incor- poration into the purines of the RNA. We also studied regenerating liver, in which the incorporation was higher, and tumour, in which the incorporation was generally somewhat higher than we found in normal liver. This, of course was rather disturbing because of the previously assumed biochemical stability of DNA and its relationship to the gene. Since it has been shown by Dr. Wilson and others that the 2-carbon of glycine is a precursor of the 5-carbon of the purines, and since it has also been shown that glycine is converted into formate, which is a precursor of the carbon in the 2 and 8 positions (the ureide positions) of the purines, it seemed of some importance to degrade the purines that had been obtained, since we might be observing simple exchange of the ureide carbons which would result from the conversion of glycine into formate. When the degradation is carried out by hydrolysis of the purines to produce glycine, the glycine is made of the same carbons which it supplied synthetically. In other words, the 5-carbon of the purine gives the 2-carbon of glycine. The glycine was isolated and treated with nin- hydrin to liberate formaldehyde, which was distilled, precipitated as dimedon, and counted. A separate aliquot of the formaldehyde was determined colorimetrically. We determined the specific activity of the purines, the specific activity of the formaldehyde obtained on degradation, and calculated the specific activity that the formaldehyde would show if all the ^^C were present in the 5-carbon atom. These data Synthesis of Nucleosides and Nucleotides 173 are reported in Proc. Soc. exp. Biol., N.Y., 76, 464-5 (1951). We found in all eases except one (in which we had an unexpectedly high value for the purines) that 35-75 per cent of the activity in the purine is actually contained in the 5-carbon atom, and in most cases the value from the DNA is somewhat higher than the value from the PNA. This proves, then, that to the extent of 35-75 per cent this glycine is being incor- porated into the purine as glycine, and the work of Dr. Wilson and Heinrich and also the early work of Carlson and Barker indicate that presumably the remainder of the purine carbon is derived from formate. This, I think, necessitates the postulation that there are two separate pathways of DNA purine synthesis, one involving the incorporation of adenine, which essentially does not exchange; the other involves these smaller precursors, which apparently enter and leave the DNA, which by chemical analysis by phosphorus incorporation is known not to be turning over. Thus there is considerably more lability of the purine of the DNA than there is of the phosphorus, and of course there is no information yet on the ribose. At the present time we are doing a rather extensive turnover study to try to find the time sequence of these reactions. Hammarsten: I would like to ask Dr. Heidelberger if he extracted all of the DNA from his tissue or only part of it. It might not be quite homogeneous. Heidelberger: We feel that we have extracted in these procedures about 75 per cent of the DNA, and in a separate experiment we found identical specific activities in the purines of this extract and in com- pletely extracted aliquots of the same sample. Of course, the work of Chargaff and others has shown that DNA is probably not a homogeneous compound, and we hope to try to study the differences in specific activities in different parts of the DNA molecule. We do have con- tamination of up to 10 per cent of RNA in our DNA sample, but in order to invalidate these results we would have to have about 200 per cent contamination. The contamination alone cannot be the explanation. Brown: I agree with Dr. Heidelberger. Contamination alone cannot be the explanation. I would like to stand up for the Schmidt-Thannhauser method, a modification of which we are using now, involving a salt extraction of nucleic acid, precipitation with alcohol to get it away from protein, then an alkaline digestion of the mixture of nucleic acids. You can repeat the salt extraction until you are sure you have all the nucleic acids out of the tissue. Dr. Roll is now using an alkahne Dowex column which will separate five bases: adenine and guanine, as well as cytosine, uracil and thymine. The nucleotide experiments which I described were worked up by that method. There uracil and thymine are the cross-contamination criteria and 1 per cent of either can be readily picked up on the column, and each nucleic acid is free from the other to within 0-4 per cent. We do not have that data on the nucleic acids that were used for our PNA-DNA turnover studies, because at that time we were depending on either a modified isotope dilution experiment or on colorimetric methods, neither one of which were perfect. 174 G. B. Brown We originally approached the problem of the differences between glycine and adenine because Dr. Hammarsten reported in 1948 that glycine went into DNA to a certain extent. When we did our adenine experiments, we found that adenine wasn't incorporated and we were curious for some time. I think about five laboratories have recently obtained data on the incorporation of purine precursors into DNA. Our experimental approach was to quit worrying about differences in experimental conditions between laboratories and to give two tracers in the same set of animals, ^^C adenine and i^N glycine. While we had essentially no incorporation of intact adenine into the DNA, we had an appreciable incorporation of glycine, as mentioned after Dr. Howard's paper yesterday (p. 150). Hammarsten: I didn't mean to say anything about contamination of your material, Dr. Heidelberger; I meant another thing, the extent to W'hich this DNA was extracted from the cell nuclei. Cell nuclei are very difficult to extract when not mechanically completely disintegrated, and a lot of extraction work is done with incompletely disintegrated cell nuclei; and even with weak alkali, extraction of DNA is not complete unless the nuclei are disintegrated. We always shake the nuclei with glass beads at 100 vibrations per second; then they are completely disintegrated in 24 hours and can be completely extracted. If you don't do that we might get different fractions of DNA. We don't know, but it is a possibility. Heidelberger: I don't wish to take up time with our experimental procedures, which are published, but I would like to add that in every case, in checking the purity of our DNA samples, whatever they might be, for contamination with PNA, in addition to the colorimetric tests, we also test for uracil on the chromatograms. In every case there was considerably less uracil in the DNA samples than the sugar analyses indicated. We have thus used the amount of uracil in the DNA as an additional measure of RNA contamination. Brown: If there are two DNA's the chemical separation and charac- terization is going to be an extremely difficult problem. Dr. Bendich* in our laboratory has an experiment in progress in which labelled formate has been administered in quantity to rats. The DNA is isolated, and then an attempt is made to fractionate it into components with different isotope values; in other words, a fractionation based upon differential metabolism of the sample before it is isolated for study. I don't know any better method to work on a small sample of DNA and try to show that it is inhomogeneous. At the present moment he has two fractions of DNA, obtained by salt extraction, in which the purine label is essentially identical in the two. The thymine label is about 30 per cent less in one than in the other, and that seems hopeful. I don't say it is evidence for the presence of two kinds of DNA. ♦Bendich, A.. Abstracts, American Chemical Society, Boston; April, 1951, p. lie. THE USE OF RADIOPHOSPHORUS IN THE STUDY OF THE NUCLEIC ACIDS* J. N. DAVIDSON This paper is an account of recent work carried out by a group of biochemists in Glasgow using ^^P for nucleic acid work. When radioactive phosphorus is used as a tracer in experi- ments on the nucleic acids and nucleoproteins, it is possible when dealing with moderately large amounts of tissue to isolate the nucleic acids in a purified state before determining specific activities; but when only minute amounts of tissue are available attempts have been made to determine specific activities on tissue fractions rather than on isolated sub- stances. Thus when a sample of liver tissue, from which acid soluble phosphorus and lipid phosphorus have been extracted, is incubated in alkali in accordance with the usual Schmidt- Thannhauser (1945) procedure, it has been assumed by some authors that the precipitate obtained on acidifying the alkaline digest will give the activity of the deoxyribonucleic acid (DNA), while the supernatant fluid containing the nucleo- tides derived from ribonucleic acid (RNA) will give the activity of the RNA. That this is by no means the case is shown from Tables I and II. In the first place the acid supernatant fluid in the Schmidt- Thannhauser separation contains, in addition to ribonucleo- tides, a minute amount of inorganic phosphate usually supposed to be derived from "phosphoprotein." This "phos- phoprotein phosphorus" has a much higher activity than has the ribonucleotide phosphorus (Davidson et al., 1949) but its presence is frequently disregarded. Moreover, comparisons ♦This work was supported by grants from the Medical Research Council and the British Empire Cancer Campaign, to whom grateful thanks are due. 175 176 J. N. Davidson of pentose and phosphorus estimations in the ribonucleo- tide fraction suggest the presence of a phosphate ester (or esters) other than the ribonucleotides (Davidson, Frazer and Hutchison, 1951). Table I Concentrations and Activities of Phosphorus Compounds Obtained BY the Schmidt-Thannhauser Procedure from Rabbit Liver Two Hours after Injection of lOfxc. ^-P per 100 g. body weight Phosphorus fraction Mg. P/100 g. fresh tissue Sp. activity cts.lmin. 1 100 ng. P. Inorganic, acid soluble Organic, acid soluble .... Lipid "Ribonucleotide" "Phosphoprotein" "DNA" 320 109-4 1161 65-7 0-9 20-4 4933 1572 127 225 818 24 Table II A Comparison of the Specific Activities of Phosphorus Fractions Obtained from Rabbit Liver Homogenates by the Schmidt-Thann- hauser Procedure with those of Nucleic Acids Isolated from the Same Material 10 /iC. per 100 g. body weight were injected two hours before killing. Specific activities in cts./min./lOO /xg.P. S. and T. fractions Isolated N.A. '^'Ribonucleotide'' "DNA'' RNA DNA 281 173 487 311 118 322 237 15 32 61 16 30 21 16 57 54 105 45 52 59 91 7 3 11 5 15 13 12 In the second place, the specific activity of the phosphorus in the ribonucleotide fraction is invariably much higher than that of the phosphorus in isolated RNA, while the activity of the DNA fraction is greater than that in isolated DNA (Table II). Radiophosphorus and Nucleic Acids 177 Experiments along these lines by W. C. Hutchison and S. C. Frazer have also been extended to an investigation of the Schneider (1945) procedure in which the extracted tissue powder is treated with hot 10 per cent trichloracetic acid (TCA) to remove nucleic acids as acid soluble fragments on which pentose and deoxypentose are estimated. It might be supposed that nucleic acid phosphorus would be removed in the hot acid extract, leaving "phosphoprotein" phosphorus in the insoluble residue, but Table III shows that the residue contains much more phosphorus than would be expected and that its activity, due in part to highly active "phospho- protein," is much greater than that of the phosphorus in the Table III Concentrations and Activities of Phosphorus Fractions Obtained FROM HoMOGENATES OF RaBBIT LiVER BY THE SCHMIDT AND ThANNHAUSEB Procedure and by the Schneider Procedure 10 fxc. ^^P per 100 g. body weight were injected two hours before killing. Schmidt-Thannhauser Schneider Ribonu- cleotide Phospho- protein DNA Extract Residue mg. P/100 g. fresh tissue cts./min./lOO |ug.P 74-2 311 0-8 2585 17-9 16 600 203 38-5 525 acid extract which is derived from the mixed RNA and DNA. When the Schneider residue is submitted to a Schmidt- Thannhauser separation, the inorganic phosphate released on alkaline incubation corresponds both in amount and in activity to the "phosphoprotein" phosphorus found by the original Schmidt-Thannhauser method, but it is accompanied by phosphate esters of surprisingly high activity (Table IV). To examine the possibility that the highly active "phos- phoprotein" phosphorus found by the Schmidt-Thannhauser procedure might consist mainly of inorganic phosphorus which had failed to be removed in the preliminary extraction of the tissue with TCA, Hutchison and Frazer compared the effects of extracting fresh liver homogenates 3, 6 and 20 ISOTOPES 13 178 J. N. Davidson Table IV Concentrations and Activities of Phosphorus Compounds in Fractions Obtained from Rabbit Liver Homogenates by the Schmidt-Thann- hauser Procedure and by the Schneider Procedure with the Schmidt- Thannhauser Procedure Applied to the Schneider Residue 10 fic. 32p per 100 g. body weight were injected two hours before kilHng. Fraction Mg. P/100 g. fresh tissue Sp. activity cts.lmin.l 100 ng. P Schmidt-Thannhauser "Ribonucleotide" .... "Phosphoprotein" .... "DNA" Schneider Extract Residue Residue divided into "Ribonucleotide fraction" . "Phosphoprotein fraction" . "DNA fraction" .... 470 0-4 11-7 44-5 11-6 9-9 0-2 1-5 156 1641 19 52 359 369 1850 235 times with ice cold 10 per cent TCA in a refrigerated centri- fuge. They showed that even 20 extractions reduced neither the amount nor the activity (Table V) of the "phosphopro- tein" phosphorus, which would therefore appear to be a Table V The Extraction of a Homogenate of Rat Liver Tissue 3, 6 and 20 Times with 10 per cent Trichloracetic Acid, followed by Application OF the Schmidt-Thannhauser Procedure 10 fic. 32p per 100 g. body weight were injected two hours before killing. Number of TCA extractions 3 6 20 Phosphorus fraction Specific activ ity in cts./min./lOO ixg. P. Inorganic, acid soluble Organic, acid soluble Acid sol. (20th extn.) Lipid "Ribonucleotide" . "Phosphoprotein" "DNA" 7450 3415 871 470 1340 32 7635 3490 891 403 1195 30 7820* 3910* 189 874 383 1221 28 •1st to 10th extractions only. Radiophosphorus and Nucleic Acids 179 truly protein-bound tissue constituent and not merely a residue of acid soluble inorganic phosphate which has failed to be extracted with TCA. Investigations of this type have, however, raised the question of the number of washings which are necessary to remove all traces of acid-soluble phosphate from a tissue homogenate. It is generally considered that three or four extractions are adequate, but in dealing with ^^p it should be emphasized that the exceedingly minute amounts of phos- phorus removed in the later stages of repeated washing with TCA may nevertheless be significantly radioactive, since the acid soluble inorganic phosphate in a tissue has a much greater activity than have any of the protein-bound phos- phorus fractions. Frazer and Hutchison have investigated this matter by adding ^^p along with TCA to a homogenate of liver tissue from a non-radioactive animal. Such artificially added inorganic radiophosphate might be expected to be completely removed by subsequent extractions with TCA, leaving the protein-bound phosphorus fractions inactive, but this has not proved to be the case. In one experiment as many as 57 washings with 10 per cent TCA in a refrigerated centrifuge failed to remove all traces of radioactivity from the protein, and the protein-bound phosphorus fractions showed significant activity. Repeated washings with TCA containing 15 per cent non-radioactive NaH2P04.2H20 also failed to remove contaminating radioactivity (Davidson, Frazer and Hutchison, 1951). It should be emphasized that in our experience the activity of the protein-bound phosphorus fractions obtained in this type of experiment show a much greater variation than those from animals which had received the ^^P by injection. These experiments serve to emphasize the need for rigid purification of a tissue constituent before its specific activity can be determined. It is clear from our results that before any reliable informa- tion can be forthcoming about the specific activities of the ribo- nucleotides obtained in a Schmidt-Thannhauser separation, 180 J. N. Davidson it is necessary first to isolate the various components, known and unknown, of the ribonucleotide fraction. Existing methods for the separation of ribonucleotides by ion-exchange chromatography or by paper chromatography all suffer from certain disadvantages, and an attempt has therefore been made to utilize the mobility of nucleotides in an electric field. R. M. S. Smellie has succeeded in developing a method for the separation of ribonucleotides by ionophoresis either on agar gel or preferably on paper, using a technique similar to that described by Durrum (1950) for amino-acids. The nucleotides separate in the order: uridylic acid, guanylic acid, adenylic acid and cytidylic acid, in decreasing order of mobility, and their positions after separation can easily be located in ultraviolet light by the procedure of Holiday and Johnson (1949). If a permanent record is desired, a photo- graph may be made by the method of Markham and Smith (1949). The nucleotides can be removed quantitatively on elution by the technique of Consden, Gordon and Martin (1947) and can be identified by their ultraviolet absorption spectra and estimated by determination of phosphorus. When the method is applied to a mixture of nucleotides to which radioactive inorganic phosphate has been added, the nucleotides on isolation show no contamination with radio- active material. The ionophoresis method can be applied to the analysis of a purified ribonucleic acid, but it can also be used with the mixture of ribonucleotides obtained in the Schmidt- Thannhauser separation procedure (Smellie and Davidson, 1951). In this case it is preferable to acidify the KOH digest with perchloric acid rather than with trichloracetic acid. When the ribonucleotide fraction is submitted to ionophoresis, uridylic acid is found to be preceded by a fast moving spot showing up faintly in ultraviolet light. This unknown component accounts for 10-20 per cent of the total phosphorus in the fraction (Table VI) when liver tissue is employed, and consists of a small amount of inorganic phosphate (derived from "phosphoprotein") and a larger Radiophosphorus and Nucleic Acids 181 proportion of organic phosphate, which is responsible at least in part for the fact that the amount of reactive pentose in the ribonucleotide fraction is invariably less than is expected on the basis of the phosphorus content. Table VI Concentrations and Activities of Phosphorus Fractions Separated by lonophoresis from the ribonucleotide fraction in a schmidt- Thannhauser Separation of Rat Liver Tissue 10 /iC. ^^P per 100 g. body weight were injected two hours before kilhng. Components of ribonucleotide fraction mg. P/100 mg. fresh tissue Sp. activity in cts.jmin.llOO fig. P. Cytidyhc acid Adenyhc acid Guanyhc acid Uridyl ic acid Inorganic phosphate (from "phos- phoprotein") Organic phosphate 11-9 17-9 240 120 1-8 2-7 300 285 121 482 3050 660 Smellie has also applied the ionophoresis procedure for ribonucleotide separation to tissue fractions from the livers of rats which had received radiophosphate by injection two hours previously, in order to determine the individual specific activities of the four nucleotides derived from RNA. The results of such an experiment are shown in Table VI, from which it will be seen that in the ribonucleotide fraction ob- tained from the Schmidt-Thannhauser separation of rat liver, the uridylic acid has a higher activity than the purine nucleo- tides, and all nucleotides have a lower activity than the inorganic phosphate from "phosphoprotein" and the organic phosphate which accompanies it. Since ribonucleic acid occurs in several different cellular components it is obviously desirable to compare the activities of the RNA phosphorus of different origins. This has been done by W. M. Mclndoe and R. M. S. Smellie using rat liver tissue for the isolation of nuclei by the citric acid method of Mirsky and Pollister (1946) and for the isolation of cytoplas- mic particles from a homogenate in saline by the method of 182 J. N. Davidson Claude (1946). From the figures shown in Table VII, it is clear that in agreement with the results of Marshak and Calvet (1949) and of Barnum and Huseby (1950), the nuclear RNA has a much higher activity than the cytoplasmic RNA. It also appears that of the cytoplasmic fractions the ribonucleotide phosphorus in the small granules has a lower activity than that in the large granules or the supernatant fluid. This is in general agreement with the results of Jeener (1949) and of Barnum and Huseby (1950) who used ^^p. Table VII Activities of Phosphorus Fractions Separated by Ionophoresis from THE Ribonucleotide Fractions in a Schmidt-Thannhauser Separation OF Isolated Nuclei and Cytoplasmic Components Obtained from Rat Liver Tissue 10 fxc. ^-P per 100 g. body weight were injected two hours before kilhng. • Specific activity of components of ribonucleotide fractions cts.lmin.llOO ng. P. Nuclei Large Granules Small Granules Supernatant Cytidyhc acid . Adenylic acid GuanyUc acid . Uridyhc acid Inorganic phosphate (from "phosphopro- tein") .... Organic phosphate . . 2900 3300 1925 2660 6020 204 216 188 325 3600 500 65 128 33 120 4000 50 375 260 135 314 4660 180 Reichard (1950), however, using ^^N in regenerating rat liver, found no difference in the incorporation of the isotope into the purines of the RNA in mitochondria, microsomes, and "cell sap," i.e. the supernatant fluid in the differential centrifugation procedure. The reason for this discrepancy is not yet clear. Neither Jeener nor Barnum and Huseby isolated the separate nucleotides in their fractionation procedures. This has now been done by Mclndoe and Smellie using the iono- phoresis technique, and the results are shown in Table VII, Radiophosphorus and Nucleic Acids 183 which indicates that, whereas in most cytoplasmic fractions the phosphorus has a higher activity in the pyrimidine nucleotides than in the purine nucleotides, this is not so for the nuclear ribonucleotides. REFERENCES Barnum, C. p., and Huseby, R. A. (1950). Arch. Biochem., 29, 7. Claude, A. (1946). J. exp. Med., 84, 51. CoNSDEN, R., Gordon, A. H., and Martin, A. J. P. (1947). Biochem. J., 41, 590. Davidson, J. N., Frazer, S. C, and Hutchison, W. C. (1951). Bio- chem. J., 49, 311. Davidson, J. N., Gardner, M., Hutchison, W. C, McIndoe, W. M., Raymond, W. H. A., and Shaw, J. F. (1949). Biochem, J., 44, Proc. XX. DuRRUM, E. L. (1950). J. Amer. Chem. Soc, 72, 2943. Holiday, E. R., and Johnson, E. A. (1949). Nature, Lond.y 163, 216. Jeener, R. (1949). Nature, Lond., 163, 837. Markham, R., and Smith, J. D. (1949). Biochem. J., 45, 294. Marshak, a., and Calvet, F. (1949). J. cell. comp. Physiol., 34, 451. MiRSKY, A. E., and Pollister, A. W. (1946). J. gen. Physiol., 30, 117. Reichard, p. (1950). Acta chem., Scand. Schmidt, G., and Thannhauser, S. J. (1945). J. hiol. Chem., 161, 83. Schneider, W. C. (1945). J. hiol. Chem., 161, 293. Smellie, R. M. S., and Davidson, J. N. (1951). Biochem. J., 49, Proc. XV. RATE OF SYNTHESIS AND QUANTITATIVE VARIATIONS OF THE RIBONUCLEIC ACID DURING THE GROWTH OF A CULTURE OF POLYTOMELLA COECA* R. JEENER The cytoplasm of the cells of a colourless flagellate, Polyto- mella coeca, has 50 to 80 per cent of its total proteins bound to ribonucleoprotein particles, resembling those of mammalian liver cells in their speed of sedimentation, their ribonucleic acid (RNA), phospholipid and enzyme content and their extreme heterogeneity (Claude, 1946; Brachet and Jeener, 1944; Chantrenne, 1947). The study, by means of labelled phosphate, of the RNA increase in a culture oiPolytomella showed apparently that the quantity of RNA synthesized at any instant is proportional to the quantity of RNA present. The only interpretation of this fact which we believe to be pertinent is that the cyto- plasmic particles, to which all the RNA is bound, multiply by autoduplication, as Brachet (1944, 1949a, 1949&, 1951) suggested several times. We present here a short survey of some results, which, we believe, bring arguments in favour of this assumption and which will be published with more details elsewhere. We study a culture of Polytomella, the growth of which is limited by the complete utilization of the phosphate of the medium, labelled with ^^P at the moment of seeding; to this culture we add a large excess of non-labelled phosphate. The growth of the culture resumes and a particularly rapid protein synthesis begins. The evolution of the RNA (per cell or per unit volume of culture) shows three characteristic phases: a phase of decrease for 2 to 5 hours, followed by a return to the ♦Part of the communication presented by D. Szafarz. 184 RNA Synthesis in Polytomella 185 initial content, a phase of synthesis during which the RNA increase follows approximately the cellular multiplication, and, occasionally, a phase of diminution accompanying the "i o 1000 - o o w *> 5oo - Koars Fig. 1 . Modifications undergone by the quantity of RNA present in 100 ml. of culture, and the specific radioactivity of RNA P following addition of phosphate at time 0. slowing down of the growth, which is caused at the end of the experiment by the fall in oxygen concentration of the medium. From the very moment of the addition of the non-labelled phosphate to the culture, the specific radioactivity of RNA P (counts per minute per 100 fig. P) starts to decrease, without 186 R. Jeener any ^^p appearing in the culture medium. The fact which we beheve important to emphasize is that this decrease takes place at a rate (defined as 1/r. dr/dt) which is almost constant whether the quantity of RNA decreases, remains constant or rapidly increases. We would have expected that the specific radioactivity of the RNA would depend at any instant on the relation existing between the number of newly formed mole- cules and the number of old molecules, i.e. that it would diminish as rapidly as the quantity of RNA would become smaller or as the rate of increase of RNA goes up. Two hypotheses seem to be able to account for the fact, at first difficult to interpret, that the decrease of the specific radioactivity of the RNA is independent of quantitative variations of the RNA. We could consider that the rate of renewal of the RNA molecules or of the phosphate groups of their constituent nucleotides is sufficiently high so that the specific radioactivity of the RNA remains at any instant independent of the quantitative variations that it undergoes. But, in this case, the specific radioactivity of the RNA phosphorus should be very close to that of the intracellular inorganic phosphate. This is by no means the case in our experiments. Moreover, the turnover of the nucleic P should be very high, whereas the maximum value ascribable to it is 3 to 6 per cent per hour. This fiirst hypothesis being discarded, we do not see any possibility of avoiding the other one, i.e. to assume that the quantity of RNA synthesized at any instant is proportional to the quantity of RNA present, which involves necessarily the consequence that the evolution of the specific radioactivity of RNA is independent of its concen- tration. The conclusion to which we are thus led only becomes significant if we remember that the RNA is part of ribonucleo- protein particles. We may consider their autocatalytic multi- plication as possible, if we remember the analogy of the smallest particles and the viruses, and the arguments in favour of the "genetic continuity" of many types of larger and more complex particles, such as mitochondria, plasts, kinetosomes, RNA Synthesis in Polytomella 187 kappa particles, etc. If these particles, partly composed of RNA, multiply autocatalytically, the quantity of RNA syn- thesized at any instant would depend on the quantity present, as the study of the evolution of the specific radioactivity of the RNA led us to assume. The hypothesis which we present is valid whatever the intimate mechanism of the multiplication of the particles may be. This multiplication could take place by an increase in size of the particle, followed by its fragmentation into two parts, as happens in the mitochondria and the kinetosomes; or it might be that we are dealing with a more complex mechanism, of which the phage multiplication gives a model, which would involve the rupture of the particle into smaller units, the autocatalytic multiplication of the latter, and their final recombination (Luria, 1950). The facts observed till now are not in favour of the latter representation. If, after adding the non-labelled phosphate to the culture which has stopped growing, we study the evolu- tion of the specific radioactivity of the RNA of the smaller particles and at the same time that of the larger particles separated from the former by 10 minutes centrifugation at 60,000 g, we find that the radioactivity decreases just as rapidly in both fractions from the moment of addition of the excess phosphate. If the smaller particles were the only site of synthesis of the RNA, the specific radioactivity of the RNA would rapidly decrease in this fraction and would start to diminish in the larger ones only after an appreciable delay. In conclusion, we think it possible to interpret all the facts known at present by assuming that the Polytomella cells contain a mixture of various categories of specific particles, each of them being characterized by its size, its RNA content, a collection of enzymes, etc. In each of these categories there takes place autoduplication, the rate of which may be different from the rate of cellular division, and may differ from one category of particles to another. The normal ribonucleo- protein particles would thus possess the characteristics of Sonneborn's plasmagenes (Sonneborn, 1949, 1950) but would 188 R. Jeener have a quantitative importance in the cell, since in Polytomella the ribonucleoprotein particles contain more than 50 per cent of the cellular proteins. It is obvious that we consider these views as a mere working hypothesis, susceptible of experimental verification both by genetical and cytochemical methods. At present, their princi- pal weakness is that they are exclusively based on data obtained from studies with ^^P on RNA synthesis. It is evident that the simultaneous use of ^^N or ^*C is unavoidable, since only this would enable us to study simultaneously the protein constituents of the particles too. REFERENCES Bracket, J. (1944). Embryologie chimique. Liege: Editions Desoer. Bracket, J. (1949a). Experientia, 5, 204. Bracket, J. (19496). Puhbl. Staz. Zool. Napoli, suppl. 21, 77. Bracket, J. (1951). Ann. Soc. Zool. Belg. In press. Bracket, J., and Jeener, R. (1944). Enzymologia, 11, 196. Chantrenne, H. (1947). Biochim. Biophys. Acta, 1, 437. Claude, A. (1946). J. exp. Med., 84, 51, 61. LuRiA, S. E. (1950). Science, 111, 507. SoNNEBORN, T. M. (1949). Amer. Scientist, 37, 33. SoNNEBORN, T. M. (1950). Heredity, 4, 11. DISCUSSION Brown: With regard to Dr. Davidson's paper, I think it is a very fine approach to get a characterizable compound representing the ^-P rather than total nucleic acid. In some of our Schmidt-Thannhauser supernatants, prepared after extractions with trichloracetic acid, the total phosphorus recovery from rat tissue was 50-70 per cent, whereas the optical density recovery over columns and adding up all of the six fractions was 80-90 per cent. In one trial experiment, Dr. Marrian used some glucose uniformly labelled with i*C. The total nucleic acid fraction, as precipitated with salt, had quite a bit of radioactivity. The nucleotides, as they came off the column, had essentially none. Do you suppose that there is some sugar phosphate coming along with the nucleic acid fraction, and is there any derivative of sugar phosphate in your fast-running phos- phorus fraction in the ionophoresis? Davidson: We are still working on that fast-running fraction. We are separating it out by chromatographic methods and we haven't found any trace of any sugar phosphate yet. There is evidence of a small amount of what is probably ethanolamine phosphate. I under- stand from what Dr. Brown said earlier that he did not apply the Discussion 189 Schmidt-Thannhauser method to a whole tissue but to extracted nucleic acid, which is a very different thing. Brown: The experiment I am now quoting was done by an older technique, which was to use desiccated tissue which had been extracted with fat solvent and trichloracetic acid. Davidson: In such an experiment there might well be the additional phosphate ester. Leblond: I wanted to clarify the meaning of the curve of Dr. Szafarz. In your interpretation of that slope, do you mean that all the radio- activity comes in at the beginning, and then when particles duplicate themselves they retain the same activity, and since they increase in quantity, the radioactivity per particle decreases? Szafarz: In the beginning all the fractions of the ribonucleic acid are marked equally and have the same specific radioactivity. Then at the beginning of this graph we add non-labelled phosphate. At that time the culture begins to grow again and the cells multiply exponentially, so the specific activity of the RNA begins to decrease. If we try to establish a relation between the specific radioactivity and the quantity of RNA we see that this regularity in the slope of the specific radioactivity doesn't correspond to the evolution of the quantity of RNA. The quantity of RNA doesn't begin to increase just at the moment when the specific radioactivity begins to decrease. If we plot the quantity of protein synthesized, the proteins begin to in- crease at the same time. We have found this behaviour of RNA in all our experiments. In the beginning we thought it was a mistake, but we regularly find the same thing. So the specific radioactivity is inde- pendent of the quantity of RNA present, and this can be explained by assuming that the quantity of RNA synthesized is at any instant proportional to the quantity of RNA present. It is an autocatalytic multiplication. Popjak: In connection with the differences of the specific activities of the nucleic acid fractions found in the large and small granules, it might be worth mentioning that Mr. Ada, who worked in Dr. McFarlane's laboratory some years ago, studied with ^^p the phospho- lipid turnover in the mitochondria and microsomes of the liver cell. He found that the specific activity of the phospholipids of the micro- somes was always higher than that of the mitochondria. The two types of particles seem to have been quite independent of each other as regards phospholipid metabolism. PART V PROTEINS AND AMINO-ACIDS A METHOD FOR THE EVALUATION OF THE RATE OF PROTEIN SYNTHESIS IN MAN D. RITTENBERG Until now the problems to which this conference has devoted its attention have in the main been of such character that, complex though they were, they could be clearly formu- lated. Such unfortunately is not the case for the question concerning which I am expected to enlighten you. The syn- thesis of proteins is one of the characteristic functions of the living cell. Unfortunately we do not clearly understand the nature of a protein and would surely have difficulty in agree- ing on definitions of the words protein or synthesis and almost certainly on the meaning of the phrase "protein synthesis." Rather than attempt to formulate definitions which exactly define these terms I shall depend on one patterned after that used by Eddington (1949) to define physical knowledge. Protein synthesis will consist of that which any right thinking person would accept as protein synthesis. The difficulty in defining the subject matter of this lecture arises from ignorance. Any more exactly formulated de finition would almost surely arise to plague me in the future . Recognizing that I am assuming that we int uitively under- stand the meaning of the phrase "protein synthesis," let us consider some kinetic aspects of this problem. We now know, as a result of recent studies, that the tiss ue proteins are rapidly being formed and degraded. Indee d, the proteins 190 Rate of Protein Synthesis in Man 191 of the living cell owe their secular stability not to an absence of reactions nor to existence of a time independent equilibrium state, but because they are a part of a kinetically stationary state. It is important that we clearly differentiate between an equilibrium and a steady state (Burton, 1939; Bertalanffy, 1950; Denbigh, Hicks and Page, 1948). A system at equili- brium is a closed one; material neither enters nor leaves and the state is determined by thermodynamic quantities. A system in a stationary state is an open system ; materials enter and leave and the state is determined by kinetic factors. I have previously indicated the advantages which accrue to a living cell from this form of organization (Rittenberg, 1949). These general considerations indicate that kinetic data are required if we are ever to understand the problem of the living cell. Some years ago Shemin and I (Shemin and Rittenberg, 1944) attempted to measure the rate of protein formation in the rat. We found that the liver proteins of this animal were being regenerated with a half time no greater than 8 days. Even faster was the regeneration rate of glutathione ( Waelsch and Rittenberg, 1942). The half time of this tripeptide is less than 4 hours. Studies in our laboratory on the rate of formation of the human plasma proteins indicate that they have a half time of about 10 days. In none of the experiments reported above was it possible to determine the metabolic activity of the great bulk of the tissue proteins, the muscle. Recently a method has been developed (Sprinson and Rittenberg, 1949) by which the absolute rate of protein synthesis in an intact animal could be estimated. The assumption underlying this method is that a dietary amino-acid is either used for protein synthesis or is oxidized and its nitrogen excreted. This is not true in all details, for we know that dietary nitrogen can be used for the synthesis of nitrogenous compounds other than protein. Quantitatively these are but minor side reactions, and we have ignored them. The excretion of dietary nitrogen will thus reflect the rate of protein synthesis. The more rapid the synthesis, the less dietary nitrogen will 192 D. RiTTENBERG be excreted. We measured this partition of dietary nitrogen by adding to a normal diet a small amount of ^^N labelled amino-acid. These experiments using ^^N labelled glycine indicated that in a normal adult human about 1 • 3 grams of protein per kilo weight were daily being synthesized. At the same time the magnitude of the metabolic pool was calculated. This pool was defined as being all those nitro- genous substances which arise either from the diet or from the degradation of tissue proteins which can be used for the synthesis of tissue proteins. The size of the metabolic pool was found to be 0-5 g. N/kilo weight. This metabolic pool is really a mathematical construct, since we are unable to define exactly its chemical composition or even its location in the organism. It appears to be large, for it would require that the metabolic pool of the normal human adult contain about 35 g. of nitrogen. This value is more than 10 times the total free amino-acid nitrogen ('^2-5 g. N). While it is to be expected that the metabolic pool should be greater than the total free amino-acid nitrogen, such a large difference is surprising. Data similar to that which we have obtained have been reported by White and Parson (1950), who fed ^^N labelled glycine and ^^N labelled yeast, and by Wu and Snyderman (1950) who fed labelled L-aspartic acid. The most obvious defect of the theoretical system elaborated by Sprinson and Rittenberg is their tacit assumption that the rate of urea excretion is not the rate determining step. Dr. San Pietro and I have re-analysed the problem without making this assumption. The system we assume is graphically indicated in Fig. 1. The metabolic pool consists of organic compounds, previously defined, containing P mg. of nitrogen. Dietary nitrogen enters the metabolic pool at the rate of D mg. per day. Some of the components of the metabolic pool are used for protein synthesis at the rate of S mg. N per day; another part, £„ mg. N per day, is converted to urea and mixes with urea already present. The remainder of the nitrogen excreted is denoted by Ejj. We assume E, to be negligible as compared to E^. Rate of Protein Synthesis in Man 193 The total urea in the organism, U mg. N, is the urea pool. Urea enters it at the rate of E^ mg. N per day and leaves it at the same rate via the urine. Since we assume the animal is in the stationary state, the rate of protein breakdown, R, is equal to S. Protein Pool F Grams N- R Metabolic Pool P Grams N- Y Urea Pool U Grains M- ■>Ku For the Stationary State : S=R, D=Et. where : — D. Rate at which Dietary Nitrogen enters the Metabolic Pool as Grams/day. R. Rate at which Nitrogen from Protein breakdown enters the Metabolic Pool as Grams/day. S. Grams of Nitrogen/day used for Protein Synthesis. Et. Rate of Total Urinary Nitrogen Excretion as Grams/day, and Et = Eu+Ex. Fig. 1. If we define the total excretion of N in the urine to be Et, then:- Et = Eu -|- Ex R-S D=Et It is clear that if we are to be able to understand the factors which determine the rate of excretion of a sample of ISOTOPES 14 194 D. RiTTENBERG a labelled amino-acid added to the diet, we must evaluate the parameters which govern the kinetic aspects of the urea pool. We have done this by direct measurements. A small amount of ^^N labelled urea (~50 mg.) was injected intra- venously into a normal adult. It is possible to predict the fate of this material. Let y=mg. ^^N injected as urea. Au=mg. ^^N in urea pool at any time ^^Nu = ^^N concentration (atom per cent excess) in urea pool at any time. It is relatively easy to show that and that E _100Au _ lOOye-u* (2) if we assume mixing of the exogenous and endogenous urea to be fast. When t=0 ("N„) t-o=^ (3) °^ ^=pNiro (*) In Fig. 2 is shown the results of three experiments. In each case you observe a rapidly declining curve which then becomes almost linear. Extrapolation of the linear portion of the curve to zero time gives the isotope concentration which would have existed had mixing of the injected urea been instantaneous. These values shown on the left hand side of the figure are (^5Nu)t=o- Inserting this value in equation 4 gives U. In normal adults U is approximately 6 g. nitrogen. Since the rate of excretion of urea in urine is easily measured, all the constants in equations 1 and 2 are known. Rate of Protein Synthesis in Man 195 The rate of clearance of urea from the urea pool in equation 1 is determined by the size of 'EJV. The half time, t^yg) is given by ^ti/2=In 2=0-693 (5) Since Eu/U is approximately 2, ti/2 is close to 1/3 day or eight hours. In this period half of the urea present in the urea pool will be excreted. o S z < UJ a: o o o -J Ol lu X u. 0.1 O z i= 0.067 ABS JI _, * 50 100 TIME MINUTES 150 Fig. 2. Isotope concentration of the urea pool following intra- venous injection of ^^N urea. The same results can be obtained in determining the rate of excretion in the urine of the labelled urea. If A^ is the total urea ^^N excreted in the urine, then a plot of the logarithm jl — ^1 vs. time yields a straight line (see Fig. 3). The slope of this line is equal to E^/U. of 1 196 D. RiTTENBERG Having determined the total urea in the human we next calculated the volume of solution (in litres) that this amount of urea would occupy were it uniformly distributed at a concentration equal to that found in the blood. This volume we call the urea space. It is calculated from equation 6. Total Urea Nitrogen=10 x B.U.N, x Urea space (6) in which B.U.N, represents the blood urea nitrogen in mg. per cent. Some values for the urea space are shown in Table TIME DAYS 0.5 -0.5 • Fig. 3. Calculation of the size of the urea pool. I. The urea space appears in these three subjects to be about 50 per cent of the body weight. These values are approxi- mately equal to the total body water values which exist in the literature. We have directly measured the total body water of a subject by administering DgO as well as ^^N labelled urea. The DgO and urea values are in excellent agreement, indicating that urea is uniformly distributed through all the water of the body. Knowing the rate of excretion of urea in the human, we can set up a system in which we need not assume that the Rate of Protein Synthesis in Man 197 excretion of urea is not the rate determining step. If we set up the equations governing the system shown in Fig. 1 it is a relatively simple matter to show that -Bt Ap = AqC in which Ap=mg. ^^N in the metabolic pool at any time t Ao=mg. ^^N administered as amino-acid ET"i"S (7) B=- Table I Measurement of the Urea Space and its Relation to Body Weight in Normal Male Subjects Expt. Body Wt. Kg. y Mg. "iV "A^M at zero time U g.N 4 BUN Mg. per cent Urea space Atom per cent excess Litres Per cent body wt. DR AGS JI 68 72 63 8-38 8-38 6-53 146 0132 067 5-74 6-35 9-75 17 18 32-5 33-8 35-3 300 50 49 48 Equation 7 represents the amount of isotope remaining in the metabolic pool at any time t. The larger the value of B the more rapidly will the isotope content of the metabolic pool decline. Part of the ^^N leaving the metabolic pool will do so by conversion to urea and transfer to the urea pool. The amount of ^^N in the urea pool, A^, will be Au= AoEuU P(Eu-BU) /e-Bt_e-u* j (8) and the amount of ^^N excreted as urea in the urine, Ag, will be A.= AqEu u P(Eu-BU) LB^^ ^ ^ E ^ ^ ^ (9) 198 D. RiTTENBERG In Fig. 4 are shown a plot of equations 7, 8 and 9 with arbitrary values of the constants. You notice that the isotope concentration of the metabolic pool drops rapidly. Indeed, after 3-4 hours it drops almost to zero. The isotope concen- Curve b Simmation of Fraction Excreted as Urea Nitrogen Curve c Fraction Remaining in Urea Pool lax TIME Fig. 4. Representative curves showing distribution of isotope following intravenous injection of labelled amino-acid. Curve a. Curve b. Curve c. Ao Ae Eu2 1/, _-Btx U Eu 1 -Ut) Ao P[Eu-BU] Au EuU b(1 e ) -Bt ^-TT t Eu Ao P[Eu-BU] 100 Ao Eu " P(Eu-BU ) En e-Bt_e-u t tration in the urea pool rapidly reaches a maximum value and then slowly declines. The total excretion steadily increases along an S-shaped curve. If sufficient data are available, equations 7, 8 and 9 can be solved and used to determine S and P. Rate of Protein Synthesis in Man 199 The data from a typical experiment are given in Table II. With these data we determine the time at which the urea Table II Rate of Utilization of Glycine Nitrogen by Human Subject (DR-1) Time Urea nitrogen excreted Mg. ^W concentration 2x100 Days Urea Ammonia Calcu- lated* Observed Corrected'^ Atom per cent excess -0-017 100 0-094 3-940 0-21 0-21 0-15 0-017-0 038 246 107 1 354 73 80 66 0-038-0 059 257 129 558 1 25 1 54 1 32 - 059-0 081 247 117 336 1 83 2 18 1 89 0-081-0 101 196 122 233 2 32 2 72 2 34 0-101-0 127 310 112 185 2 92 3 49 3 02 0-127-0 156 300 103 146 3 58 4 18 3 60 0-156-0 229 647 101 125 5 11 5 64 4 79 0-229-0 485 2500 100 091 9 26 11 2 9 61 0-485-0 827 2620 071 050 12 7 15 4 12 3 • 827-1 23 3330 047 054 15 18 9 14 3 1-23 -1 54 2945 040 034 16 21 5 15 8 1-54 -1 83 2540 034 036 16 6 23 4 16 7 1-83 -2-13 2840 0-023 0-040 170 24-9 17-0 *Eu = 8-95g. N/day U=5-65g. N fCorrection: B = 83-2/day Et = 11 - 5 g. N/day P = 0-612g. N. Ao=44-8mg. »N 17-24-9 213 — 3-70 per cent/day. concentration attains its maximum value. At this time Eu U E u U*^max.__-g g-Bt ''max. (10) B is calculated from this equation and used in equation (8) to calculate P. S is determined by equation (11). S=BP-Et (11) There is certain evidence that these calculations approxi- mate the real situation. From equation 7 we deduce that the administered amino-acid is rapidly cleared from the metabolic pool. The rapid decrease in the ^^N concentration of the urinary ammonia is consistent with this, for the urinary 200 D. RiTTENBERG ammonia is known to arise from deamination of the free amino-acids of the blood. Further, it is known from experi- ments in which rats are kept on a diet deficient in an essential amino-acid that to produce growth the essential amino-acid must be fed at the same time as the deficient protein. If any considerable time elapses between the feeding of the amino- acid and the deficient protein, the supplement will all be destroyed before the incomplete protein enters the system. Protein can only be formed when all the required amino-acids are present. The values we obtain for S have been corrected for the feed- back of isotope from the protein pool to the metabolic pool. The values for S are approximately 0-4 g. nitrogen per day per kilo weight. This does not differ greatly from the values calculated by Sprinson and Rittenberg. Indeed, it is possible to show that the value of S does not depend on the system assumed. The value of P calculated here is much lower than pre- viously. It is about 2 grams nitrogen rather than 35 previously found. The reason for this large discrepancy is quite obvious, and in defence I would like to point out that Sprinson and Rittenberg stated that their metabolic pool was a mathematical construct. I believe that the present pool is a reality. The conclusions which we previously drew concerning the metabolic heterogeneity of human muscle are still valid. The further analysis of this question may clarify many obscure corners of nutrition. REFERENCES Bertalanffy, L. von (1950). Science, 111, 23. Burton, A. C. (1939). J. cell. comp. Physiol., 14, 327. Denbigh, K. G., Hicks, M., and Page, F. M. (1948). Trans. Faraday Soc, 44, 479. Eddington, a. (1949). Philosophy of Physical Science. Cambridge University Press. Rittenberg, D. (1948-49). Harvey Leet., Series XLIV. Springfield^ 111.: Charles C. Thomas. Shemin, D., and Rittenberg, D. (1944). J. biol. Chem., 153, 401. Rate of Protein Synthesis in Man 201 Sprinson, D. B., and Rittenberg, D. (1949). J. biol. Chem., 180, 715. Waelsch, H., and Rittenberg, D. (1942). J. biol. Chem., 144, 53. White, A. G. C, and Parson, W. (1950). Arch. Biochem., 26, 205. Wu, H., and Snyderman, S. E. (1950). J. gen. Physiol, 34, 339. DISCUSSION Neuberger: We have been worried for the last few years by problems similar to those Dr. Rittenberg has indicated about assmnptions made in the interpretation of experiments on nitrogen metabolism and turn- over rates, in this case not in man but in the rat. We, of course, realized that one has to make certain simplifying assumptions to get a system which is susceptible to mathematical treatment. However, there were two points in the treatment of Dr. Rittenberg and Dr. Sprinson which worried us slightly and I would just like to put them forward for discussion. The first assumption made is that the metabolic pool is physically essentially homogeneous. For example, if glycine is used for synthesis of protein in the liver and in muscle, it is assumed that this glycine is rapidly equilibrated between muscle and liver. I think that this is unlikely to be quantitatively correct because the penetration of amino acids into certain tissues like muscle and to a much greater extent into the cells of the central nervous system, is relatively slow. If the rate of protein synthesis in the liver is of the same order of magnitude as the rate of penetration of glycine from the intracellular space of the liver to the intracellular space of the muscle, you may get a steady state, in which the isotope concentrations are different, and the assump- tion of a homogeneous compartment falls to the ground. Further, if, for instance, the degradation of glycine concerns mainly the glycine in the liver, and if this liver glycine is not in equilibrium with muscle glycine, you again may get a considerable error by assuming that the urinary nitrogen is derived from a homogeneous pool in the body. The other point is that, if you give labelled nitrogen in any form, either in the form of ammonia or of glycine, one makes the assumption that this labelled nitrogen gets quickly distributed over other nitro- geneous compounds. I am not quite sure whether this is an assumption which has to be made, but if it is so, then how far is this assumption correct? There are these two points where lack of homogeneity might be quite serious and I am just wondering whether it wouldn't be possible to assume that you have a series of compartments; this of course makes the mathematical treatment extremely complicated, but one might get somewhat nearer the true state of affairs. I'm just putting this forward with some hesitation as points for discussion. Rittenberg: The point which Dr. Neuberger has raised of course has troubled us greatly. I trust we are in agreement that at the present stage of matters it is not worth while to be too fussy about little points, because you first want to get the broad outline of the matter straightened out. The two points you have raised are indeed very important ones. 202 D. RiTTENBERG and I am afraid that I cannot answer them completely, but I should like to present certain notions to you. Firstly, concerning the lack of homogeneity of the metabolic pool. The mathematical treatment of this is extremely difficult, and I am certain that at the present time it would not be worth the time it would take to solve it, but intuitively you can discover some things which at least are comforting. Suppose the pool contains Ni compartments instead of one, and let's call compartment Ni, the liver pool. Let the i compartments be connected with each other and let the rate of transfer of constituents from one pool to the other be denoted by connection constants kNiN,- and kNjNi* Let us consider the situation which would arise where any set of k's (e.g. kNjNa and kNaNi) are very large or infinite. Then whichever pools are connected by these infinite constants are essentially identical pools. On the other hand, if any set of k's are zero or close to zero then you can neglect that pool; it just doesn't enter the problem. So at either end, at the boundaries, the situation is satisfactory. But the question that you wished to have answered is, "What happens in the middle?" In the middle, I don't know, but I suspect that the answer is that in the middle the actual pool can be replaced by another pool which is smaller. In other words, if pool Na contains 4 grams of nitrogen, you replace it by a pool, say of 2 grams of nitrogen, and treat it mathematically as if it has an infinite connection constant. At some time in the not too distant future this problem will have to be considered. The other question concerning the inhomogeneity of the labelled materials is also a question which I find difficult to answer, but again intuitively I feel that it is not a question which will create too great difficulties. When Dr. Hsien Wu was at our laboratory he investigated this excretion problem with L-Aspartic acid. L-Aspartic acid is quite different from glycine, and its transfer rate of nitrogen to other amino acids is quite different from that of glycine. Nevertheless, he gets almost identical values with those we obtained. I think the answer to this is that there are other things which are of more importance than the question of transfer from one amino acid to the other. There are other holes in this argument which unfortunately I cannot see at the present time, but which I trust future work will bring to the surface. TURNOVER RATES DURING FORMATION OF PROTEINS AND POLYNUCLEOTIDES IN REGENERATING TISSUES E, HAMMARSTEN The occurrence of a functional interrelation between nucleic acids and proteins during synthesis of these substances in cells has been investigated by many on the basis of Miescher's original discovery in 1872. A group at our laboratory at present engaged with this problem consists of E. Anderson, N. A. Eliasson, E. Hammarsten, U. Lagerkvist, B. Low, P. Reichard, A. Wretlind and S. Aqvist, working in collaboration with B. Thorell and G. Heden at the Institute for Cell Research and G. Ehrensvard at Wenner-Grens Institute. We are investigating a possible correlation between the turnover and synthesis of polynucleotides and that of proteins during growth. Glycine marked with ^^N, and in some experiments with ^^C in the methyl and ^*C in the carboxyl groups, was used. The rate of protein turnover was measured by the concentration of isotope in the glycine incorporated in the proteins at different stages of regeneration in rat liver after partial hepatectomy. In the same experiments the polynucleotides were degraded and the rate of their turnover measured by the isotope content in the nitrogenous bases. In other experiments hen bone, marrow was used in order to study metabolic changes during growth and differentiation of the blood cells. Results of these experiments will be mentioned here only in so far as they disclose a certain relationship between the isotope content in amino-acids and purine bases. The investigation of the relationship between the turnover rates of two different cellular substances will naturally be 203 204 E. Hammarsten dependent to a large extent on the precursor used. The ideal precursor should be directly incorporated into both of the actual substances without any intermediates. In order to investigate to what degree glycine has been incorporated as an intact molecule, thrice marked glycine was used, and the excess of isotopes was determined in a sample of glycine isolated from the proteins (Table I). The Table I Incorporation of ^^NHg^^CHg^^COOH into Proteins of Regenerating Rat Liver Isotope Atom per cent excess Counts /min. Atom per cent excess I. Administered Glycine 13-43 30336 31-0 II. Isolated Glycine 0-945 1692 2-38 Excess in isolated Glycine as per cent of administered Glycine 7-0 5-6 7-7 figures are the direct analytical values measured on the basis of total carbon. No attempt has been made to degrade the glycine and determine the isotope figures in each carbon atom. The values show that the ratio of ^^C and ^^N excesses in glycine isolated from the proteins is nearly the same as in the administered compound (7- 7/7-0 and 100/100 respec- tively). The value for ^*C (5-6) shows that the glycine incorporated into the protein molecule has undergone slight rearrangement involving rupture of the bond to the carboxyl group. According to this result the administered glycine has been directly incorporated into the proteins with only some intermediate turnover of the carboxyl group. The glycine is probably fairly well distributed in the protein molecule, and we think that the isotope content of the glycine isolated from the proteins can be used as a measure of the protein turnover. In the polynucleotides the incorporated glycine, according to earlier measurements, is probably situated in positions 4, 5 Protein and Polynucleotide Turnover 205 and 7 of the purines, and the ^^N content in adenine and guanine are therefore used as a measure of nucleic acid turn- over. On the incorporation into pyrimidines, on, the other hand, httle is known. Concerning experimental conditions I will only mention some facts that are necessary for the interpretation of the results. Partial hepatectomy was performed on 20 to 40 rats in each group. At different times after this operation isotopic glycine was injected subcutaneously during the last eight hours before the animals were sacrificed. The regeneration times, viz. lifetime in hours after hepatectomy, were in our experiments 11, 26, 32, 56 and 170. The only intentional difference between these different groups is their stage of regeneration. Each group received the same amount of isotope during the same period before death. The regeneration was calculated from weighing the excised and remaining parts of the liver in the usual way. The pooled livers from each group were fractionated with a view to separa- ting cell nuclei and cytoplasm. These tissue fractions were mechanically disintegrated by vibrating with glass beads, and were then extracted from lipids and after that from poly- nucleotides. The remaining substance was treated with trichloroacetic acid according to Schneider and after hydro- lysis fractionated for amino-acids on starch and Dowex 50 columns. The polynucleotides were also degraded and the nitrogenous bases separated on starch columns as adenine, guanine, cytidine, uridine, thymine and cytosine. All substances were analysed for purity. As to the conditions in the experiments on hens it may be sufficient to say that in some cases the regeneration in the bone marrow was stepped up considerably by destroying about 50 per cent of the red blood corpuscles. [^^N] glycine was used as precursor. The results will be given in the follow- ing figures. When [^^N] glycine was used as a tracer precursor in more or less rapidly regenerating bone marrow and non-regenerating liver of the same hen, a correlation could be observed between 206 E. Hammarsten the isotope contents in guanine and adenine from the poly- nucleotides and the isotope contents in glycine from the proteins (Fig. 1). Certainly the previously mentioned direct incorporation of glycine into both the purines and the proteins plays an important role in this correlation. In spite of the wide deviations from a mean value we think that this correlation is not accidental but indicates a general tendency to simultaneous metabolic activity of polynucleo- 10 Glycine/ /Adenine Glycine/ /Guanine 10 20 30 Fig. 1. Relation of isotope contents in glycine from proteins (abscissa) and in guanine and adenine from PNA-polynucleo- tides (ordinate) in bone marrow and liver from hen. The values are calculated on the basis of 100 per cent in administered glycine. tides and proteins. This connection might possibly signify a functional interaction of polynucleotide and protein turn- over, and should be considered along \vith the turnover rates found in the experiments on regenerating rat liver (Figs. 2, 3 and 4). These experiments show that the maximum turnover rates for all nitrogenous compounds in PNA in cell nuclei and cyto- plasm and in DNA appeared at about 30 hours after partial hepatectomy. The greatest increase in the amount of PNA Protein and Polynucleotide Turnover 207 70 o w> CI o O C 6.0 ■°T> c >> O CT> 41 "o 5.0 2 fe ?, ^ T^.S 4.0 o cr •a T> 5.0 4) 4) o D in is! o c 4.0 o F in •o z o c _c 30 a; *^ o u ex 41 ?0 r- Q. c o O < O 1.0 Glycine •—Adenine Pna cu Ade nine PNA Cy tnpinc^ 12 24 36 48 60 72 84 96 108 120 132 144 156 168 Time in hours Fig. 3. 208 E. Hammarsten 7.0 o c o I 6.0 "o >. cr -O T3 5.0 ti .— - B e 4.0 « e "} -o z o c E 3.0 o c °- fc 2.0 E "• 9 ° <2 1.0 Glycine UPC = Uridine PNA Cytoplasm 12 24 36 48 60 72 84 96 108 120 132 144 156 168 • Time in hours Fig. 4. and DNA per dry weight liver tissue appeared at the same time, demonstrating the coincidence of high turnover rates with an increase of polynucleotides per dry weight of tissue (Fig. 5). Mq .2 ..Protein NUrpqen/O.!, Regeneration 160 140 120 100 80 60 40 UO 12 24 35 48 60 72 84 95 108 120 132 144 156 168 Time in hours after partial hepatectomy Fig. 5. Mg. per 100 mg. dry weight of liver. PNA and DNA as phosphorus. Protein as nitrogen. Regeneration as increase in per cent of the weight of the remaining lobe after partial hepatectomy. Protein and Polynucleotide Turnover 209 The maximum turnover in glycine in the proteins appeared approximately 30 hours later (60 hours after partial hepatec- tomy). All of the other amino acids also had a maximum at this point, indicating a common maximum of transamination. The cytological changes during regeneration (Fig. 6) are in accordance with earlier similar findings, and bear out the conclusions from tracer analyses that nucleic acid turnover and synthesis play the primary role in nucleic acid-protein Nudeus-CytopjQsmic Ratio 168 Hours after the operation. Nucleic acid Protein maxima. maximum. Fig. 6. Constituents of regenerating liver cell. In order to facilitate comparison, the curves have been separated by multi- plying each value by an arbitrary factor and adding a constant. The maximum values, expressed as percentage increase of normal liver cell constituents, are: nucleolus, 260; nucleus, 160; cyto- plasm, 145; and nucleus-cytoplasm ratio, 130. The mitosis frequency increased from less than 0.05 per cent in normal liver to 10ii:0-32 per cent at 24 hours regeneration. formation. One notes the coincidence of the maximum rate of polynucleotide synthesis with the maximum of mitosis frequency and with the size of nuclei and nucleoli. Also it can be seen that during the period of maximum protein syn- thesis, at regeneration times 48-56 hours, these cytological quantities have already diminished. In this connection I would like to mention some measure- ments that have been made recently. The rats were fed by ISOTOPES 15 210 E. Hammarsten stomach tube a diet containing amino-acids as nitrogen source. The feeding began four hours after the operation and an adequate amount was given in portions six times per 24 hours. Fig. 7 shows that the PNA synthesis starts earher when the hver is forced into activity and that the protein synthesis lags somewhat behind, as in the earher experiments. Some very important experiments by A. L. Schade and co-workers on B. jjroteus have shown that growth may be in o in *«/) a r> c o ■o a; o ■D O o o c u o. E o < . 1.5 c o ■o i_ ^ 1.0 'c E o §0.5 o o 'I PNA Guanine TCA Protein 20 40 60 80 Time in hours Fig. 7. Isotope contents at different stages of regeneration in rat liver. inhibited by cobalt without any influence on the synthesis of PNA. This experiment emphasizes in a convincing way the secondary place of protein synthesis to that of PXA synthesis in the total picture. From all this one would expect that an inhibition of nucleic acid synthesis and turnover would decidedly inhibit protein synthesis. Whether the experiments on the purine isomers (for example those by Hitchings and others) have anything Protein and Polynucleotide Turnover 211 to do with inhibition of nucleic acid synthesis seems however to be doubtful, according to information from G. B. Brown. We intend to study these inhibitions with the help of tracer methods. We regard our results as definite in one respect, namely the correlation and time sequence in the general pattern of nucleic acid-protein synthesis. Our experiments cannot elucidate the underlying mechanism because they are preliminary and incomplete, and the enzyme systems were not studied at all. There are indications that it would be advisable to study a third group of substances, the phosphatides, which are apparently closely connected with nucleic acids and proteins in the cell, and we are intending to extend our investigations to these substances. Our reasons for this are mainly specula- tive. One of us, S. Aqvist, recently found that Avith [^^N] glycine as a precursor the amino-acids or peptides to be found in the phosphatide fractions of the liver are highly marked with ^^N. At different regeneration times the values are about the same as those which would be expected if this glycine represented the precursor at the site of incorporation. DISCUSSION Rittenberg: Do I understand you to state that the synthesis of the nucleoproteins not only precedes the synthesis of the proteins but it is necessary for their synthesis? Hammarsten: That is what we are trying to show, and I think that these experiments go some way toward showing it. Maybe the Hitchings experiment with different purines would also show that you can inhibit protein synthesis without inhibiting nucleic acid synthesis, but that you cannot inhibit nucleic acid synthesis without inhibiting all the protein synthesis. It might be so. Brown: I think that a lot of the confusing points about the ratio of incorporation of isotope into adenine and guanine are now solved. It seems from your work that much depends on the time interval involved. You referred to diaminopurine and other purines inhibiting the growth of Lactobacillus casei. We have been working with Dr. Hitchings lately on the purine metabolism pattern of Lactobacillus casei, an organism which very readily converts guanine into adenine or vice versa; if you give the organism both, it makes about half of its guanine into adenine and half of its adenine into guanine. If it is offered 2:6-diaminopurine at a non-inhibitory level, that is, about 2 /xg. of adenine per ml. (or 212 E. Hammarsten 2 ng. of guanine) plus 5 ^g. of diaminopurine, the additional purine stimulates the gro\^'th; and if [i*C]-dianiinopurine is administered, it is converted into both adenine and guanine in about the proportion that it is present in the medium. If the diaminopurine is as high as 50 ^g. per ml., which is an inhibitory level, you still get about half as much grow-th. This still gives something to work up for the tracer experiment, and even here diaminopurine is still very efficiently utilized for both adenine and guanine synthesis. Recently Dr. Hitchings has been doing some growth studies which we are now following up with tracer studies. If you give the micro-organism an induction period w ith about 0-2 ixg. of adenine for a day, during which time there is very little growi:h, then put in plenty of diaminopurine, the organism grows perfectly well with diaminopurine as a precursor of both of the purines. We are trying to find out how diaminopurine can behave both as a metabolite and as an anti-metabolite, but as yet cannot find any great difference in its metabolism when it is acting as a growth inhibitor. Unfortunately we don't have parallel studies on the metabolism of the protein, but we'll leave that to you. SYNTHESIS OF PHENYLALANINE AND TYROSINE IN YEAST KONRAD BLOCK The advances of biochemical knowledge which are attribu- table to the tracer technique are nowhere as evident as in the area which deals with the origin of organic molecules in biological systems. The use of labelled molecules must surely be credited for whatever limited information is at hand on subjects such as the biosynthesis of steroids, por- phyrins, purines, etc. The success which has been attained should however not obscure the fact that most of the findings in this field were due to accidental observations and not to rationally designed experiments. The experiments which I would like to discuss today illus- trate the power of the tracer technique, but also underscore the highly empirical nature of present-day biochemical research. During recent years the role of acetic acid as a building block for heterocyclic and alicyclic systems has been well established (Bloch, 1947). In particular, good evidence exists that the cell uses acetate as the principal carbon source for the polynuclear as well as for the aliphatic moiety of the sterols (Bloch and Rittenberg, 1942). Working on the assumption that the study of simple monocyclic compounds might aid in clarifying the mechanism of synthesis for the more complex polycyclic steroids. Dr. Gilvarg in our labora- tory has investigated the biosynthesis of the amino-acids phenylalanine and tyrosine, two readily accessible benzene derivatives. I may anticipate here the final conclusions of this research by saying that the biosynthesis of benzene rings proved to be totally unrelated to that of the ci/c/o-pentano- phenanthrene derivatives. In this sense these experiments 213 214 KoNRAD Block failed to serve the specific purpose for which they were designed. Time does not permit to discuss the hmited and on the whole inconclusive information which has previously been obtained on the biosynthesis of phenylalanine and tyrosine. The studies of Ehrensvard and his associates (Baddiley, Ehrensvard, Klein, Reco and Saluste, 1950) which resemble ours in many respects and which have greatly aided in the interpretation of our results, will be referred to later. Amino-acid synthesis was studied by Dr. Gilvarg under the following experimental conditions. Yeast (S. ccrevisice) was grown from a small inoculum in a medium containing glucose and acetate as carbon sources, citrate buffer, inorganic salts, and 7?z^50-inositol as growth factor. Acetate was the labelled substrate in one series of experiments, while in another [1-^*C] glucose was the source of isotopic carbon (Koshland and Westheimer, 1950). After a suitable period of growth, the cells were harvested. The yeast protein was hydrolysed and the resulting amino-acid mixture was resolved with the aid of ion-exchange resins, according to Stein and Moore (1950). A modification found to be useful for our purpose was a preliminary separation of phenylalanine and tyrosine from the hydrolysate by adsorption on charcoal as described by Partridge (1949). Prior to isotope analysis the amino-acids were crystallized either as such or in the form of suitable derivatives. The identity of the amino-acids was in all cases confirmed by paper chromatography. The isotope concentrations of phenylalanine, tyrosine and a few other amino-acids from the proteins of yeast grown with labelled acetate as substrate, are listed in Tables I and II. The mean of the isotope concentrations of the acetate added initially and of the recovered acetate serves as a measure of the average level of isotope in the acetate available to the yeast cell during the growth period. The yeast fatty acids, which are formed by the condensation of acetyl units, contain isotopic carbon in a concentration which is about two-thirds of this average. Phenylalanine and Tyrosine Synthesis 215 From the data in Tables I and II two general conclusions may be drawn. (1) Yeast growing on a medium containing glucose as the principal nutrient derives about 15-20 per Table I Isotope Concentrations of Amino-acids Obtained from Yeast Protein AFTER Incubation with Labelled Acetate Yeast grown in the presence of 10 parts of glucose and 1 part of labelled acetate. Initial Acetate . Final Acetate Protein . Glutamic Acid . Aspartic Acid Alanine (Ca-(-C)3) Tyrosine . Benzoic Acid* . Serine Glycine . Precursor CH^'^COOH ^^CHsCOOH Exp. I Exp. II counts I min. counts Imin. 13200 19500 2300 7200 1040 2700 2040 7150 740 1120 560 35 8 2 200 75 *Froni phenylalanine. Table II Isotope Concentrations of Amino-acids Obtained from Yeast Protein AFTER Incubation with Acetate Labelled with ^^C and ^*C Yeast grown in medium containing 10 parts of glucose and 1 part of acetate "CHg^^COOH. RIC is the ratio of the isotope concentration of the acetate and of the compound isolated. (countsjmin.) Initial acetate 22800 Final acetate 1270 Protein 1980 Fatty acids 7800 Glutamic acid 3700 Aspartic acid 730 Phenylalanine 20 Tyrosine 19 Leucine 3300 Valine 720 Lysine 6150 Glycine 100 "C {Atom % excess) 10 3 1 1 2 •4 •70 •91 •66 •83 •25 •02 •03 •27 •15 •97 •06 RIC "C RIC "C 100 0-83 0-99 98 0-92 1-34 119 2 20 0-97 0-76 216 - KoNRAD Block cent of the protein carbon from acetate and from the Cg units formed from glucose. (2) The isotope concentrations in different protein constituents vary over a wide range from maximal values in lysine, which approach those of the higher fatty acids, to very low levels in serine and glycine and insig- nificant concentrations in phenylalanine and tyrosine. Evidently carbon from glucose and from acetate is used in amino-acid synthesis in widely varying proportions. For the purposes of our discussion the isotope concentrations in the dicarboxylic amino-acids and in alanine are pertinent because they reflect the ^^C and ^*C levels in the corresponding a-keto acids, which are members of the citric acid cycle. The extensive labelling of glutamate, aspartate and alanine is taken to indicate that this cycle is operating under the conditions of our experiment, i.e. in growing yeast.* This has previously been demonstrated for resting yeast (White and Werkman, 1947; Weinhouse, Millington and Lewis, 1948). By the same token the failure of acetate to provide carbon atoms for phenylalanine and tyrosine excludes as precursors not only acetate itself but also pyruvate and the intermediates of the tricarboxylic acid cycle. This restriction applies to both the benzene rings and to the side chains of the aromatic amino-acids. The amino-acids which have the same carbon chains as the cvcle intermediates do not all have the same isotope concentrations. The values are highest in glutamic acid, considerably lower in aspartic acid and still lower in alanine. Most likely, these differences are the result of the following events. The condensation of acetate with oxalo- acetate produces highly active a-ketoglutarate and hence glutamate with a high isotope concentration. The dicar- boxylic acids and oxaloacetate formed in the course of the citric acid cycle should contain isotopic carbon at the same level, but are actually diluted by oxaloacetate resulting from the carboxylation of unlabelled pyruvate, which is continually produced from glucose. The isotope concentration in aspartic ♦This conclusion is supported by some preliminary studies of the isotope distribution in these amino-acids. Phenylalanine and Tyrosine Synthesis 217 acid will be lower to an extent which is determined by the relative rates of the cycle reactions and of CO2 fixation. Some labelled pyruvate will be formed by decarboxylation of oxaloacetate, but its quantity is probably small compared to that of non-isotopic pyruvate originating directly from glucose. The isotope concentrations in alanine will therefore be reduced still further. These relationships are illustrated in Fig. 1. PYRUVI^C ACID °CH. GLUCOSE = ALANINE °CH, ol ^ =^ HCNH, COOH ASPARTIC ACID tOCH "«-KETOGLUTARICAClD GLUTAMIC ACID ,CH,COOH CH-COOH I ^ CH, I ^ HCNH, I ^ COOH HO-C-COOH ^CHo *COOH CITRIC ACID COOH *COOH ACETIC ACID Carbon atoms derived from glucose marked ° Carbon atoms derived from acetate marked Fig. 1. Probable pathway of incorporation of acetic acid into benzenoid amino acids. The non-participation of acetate, pyruvate and inter- mediates of the citric acid cycle in the synthesis of phenylala- nine and tyrosine limits the possible carbon sources for these amino-acids to glucose itself, or to intermediates of glycolysis. For the identification of the precursors, [1-^^C] glucose seemed a suitable and relatively accessible substrate. The yeast growth experiment carried out with this labelled glucose was in all respects identical with earlier ones, except that the acetic acid in the medium was labelled by ^^C only (Table III). In this experiment the isotope distribution in pheny- lalanine and tyrosine was of primary interest. 218 KoNR.AD Block Table III Isotope Concentrations in Amino-acids Obtained from Yeast Protein AFTER Incubation with Labelled Glucose and Acetate Yeast grown on a medium containing [1-"C] glucose and [carboxy-^^C] acetate. "C "C (counts/min.) (Atom % excess) Glucose 810 Initial acetate 10-4 Final acetate 800 315 Protein 690 1-20 Tyrosine 530 005 Phenylalanine 540 007 Serine 650 Glycine 240 Table IV contains the ^*C concentrations of various carbon atoms in phenylalanine and tyrosine. The ingenious degra- dation procedures of Baddiley, Ehrensvard et al. (1950) were employed to establish the isotope distribution in the benzene rings. These methods take advantage of the fact that carbon atoms of the benzene ring which carry a nitro substituent will yield bromopicrin {CBr3N02) on oxidation with hypo- bromite, while ring carbons not so substituted will be oxidized Table IV Distribution of ^*C in Phenylalanine and Tyrosine from Yeast Grown ON [1-"C] Glucose Total COOH Qa (calc.) . . . C3 Benzene ring: (a) calc {b) Picric acid (c) 2:4-Dinitroaniline ^1,3,5 ^2,6,4 ^2,6 C4 (calc.) .... Tyrosine {counts Imin.) Phenylalanine {counts jmin.) 26 22 8 1 14 3 77 97 23 16 23 17 2 22 37 7 Phenylalanine and Tyrosine Synthesis 219 to CO2. Suitable nitro-derivatives must therefore be prepared (Fig. 2). Phenylalanine and tyrosine isolated from the experiment with labelled glucose show no significant radioactivity at ring carbon atoms 1, 3, 4 and 5. On the other hand, C2 and ( 3i|) HO^/ \_CH2CJcppJH ^ CO NH- [ 2^-= ' H0/\ cop: H N02' NO2 I 3 CBr3 NO2 C,.3.5 (2tl) NO2 (2314) (a) Tyrosine. (77+8) CO2 3 2 / ( 1*1) » H CHg cIcOO^'h— COs NH- ^^ _-y / 22t 2 I I ICOOlH (97*10) ■* CO2 B ^«H. Br/^''"'">">f\„„. 1 NO2 HO 2^^ /| NO. CBfjNOg (37±4) 2 CBt J NOg (Cg .OR Cg.C^) (22±2) (b) Phenylalanine. Fig. 2. Degradation of Benzenoid amino-acids. Cq in the phenylalanine ring account for a minimum of three 370 200 75 5C 5 460 7. Y 75 t o II. After addition of [l-i*C] glucose Side chain phenylalanine Side chain tyrosine . Serine Glycine 690^50 650 — 50 650 240 1900^100 1600- 80 1500 00-^20 290-60 24 20-10 60-20 Y particular, pyruvate and alanine, which become labelled under these conditions, cannot be major precursors. On the other hand, carbon from [1-^*C] glucose is incorporated into the alanyl side chains, and, as shown in Table V, is located primarily in the ^ positions. The specific activity found at the ^ positions is about half of that at C^ in ^*C glucose, a value in good agreement with the concept that a three-carbon intermediate of glycolysis had furnished the skeleton for the side chains. Phenylalanine and Tyrosine Synthesis 223 Two other amino-acids isolated from yeast proteins, serine and glycine, also do not appear to be closely related meta- bolically to acetate and intermediates of the citric acid cycle. Serine and glycine contained relatively low isotope concentra- tions when labelled acetate was the precursor (Table V), and therefore the major source of carbon for these two amino- acids must be confined to the same group of compounds which were considered for the side chains of the aromatic amino- acids. Moreover, in the experiments with [1-^^C] glucose the isotope distribution and the level of ^*C in the ^ positions are similar for serine and for the side chains of phenylalanine and tyrosine, pointing to a com.mon precursor. The very low concentrations of isotopic carbon in glycine, both in the experiments with labelled acetate and with [1-^*C] glucose, agree well with what is currently known about the metabolic relationship of serine and glycine. The breakdown of serine to glycine by removal of the ^ carbon atom, as demonstrated by Shemin (1946) will, under the conditions of our experiments with labelled acetate, involve a precursor which already contains very little isotope in the carboxyl and a positions. When glycine is formed from serine which had been derived from [1-^^C] glucose, the highly labelled ^ carbon will be eliminated and a product with a low ^*C concentration will result. REFERENCES Baddiley, F., Ehrensvard, G., Klein, E., Reco, L., and Saluste, E. (1950). J. biol. Chem., 183, 777. Block, K. (1947). Physiol. Rev., 27, 574. Block, K., and Rittenberg, D. (1942). J. hiol. Chem., 145, 625. KosKLAND, D. E., and Westheimer, P. H. (1950). J. Amer. chem. Soc, 72, 3383. Partridge, S. M. (1949). Biochem. J., 44, 521. Skemin, D. (1946). J. biol. Chem., 162, 297. Stein, W. H., and Moore, S. (1950). ColdSpr. Harb. Sym. quant. Biol., 14, 179. Weinhouse, S., Millington, R. H., and Lewis, K. F. (1948). J. Amer. chem. Soc, 70, 3680. Wkite, a. G. C, and Werkmann, C. H. (1947). Arch. Biochem., 13, 27. 224 KoNRAD Block DISCUSSION Bentley: I would like to describe very briefly another direct con- version of glucose into a 6-membered ring compound, which Dr. Arnstein and myself have observed, apparently without the inter- mediates of the glycolysis cycle being involved. We have been investi- gating the biosynthesis of a mould product, kojic acid, which has this structure: — H0-C2 4CH -CHl 5C-CH,0H \ / This compound is a y-pyrone, and contains the same number of carbon atoms as glucose. It is formed by various Aspergilli, and we have, rather similarly to Dr. Bloch, used [l-^^CJglucose. When we allowed the organism to metabolize this glucose, isolated the kojic acid and degraded it, we found that at least 80 per cent of the ^^C was located in carbon atom 1. The degradation was an alkaline hydrolysis of the dimethyl compound, which results in the formation of formic acid from carbon atom 1, with the production of methoxyacetone and methoxyacetic acid. We have also shown that this hydrolysis goes through a symmetrical intermediate, and at the moment the only carbon atoms on which we can really argue are 1 and 4. (The latter is obtained as iodoform after treating the methoxyacetone with alkaline hj^oiodite.) However, since 80 per cent of the activity is in carbon atom 1, we believe that this represents a direct conversion of glucose to this 6-membered carbon compound without the intermediary for- mation of triose phosphate. It is interesting that this compound can also be formed using dihy- droxyacetone as a sole carbon source, and we have been very interested to know how it jarises in this case; whether the dihydroxyacetone is first synthesized into glucose or whether there is a direct utilization. For this purpose we have synthesized dihydroxyacetone containing i*C. The synthesis involves the condensation of formaldehyde with ["CJnitromethane, yielding "CN02(CH20H)3; this compound is treated with sodium in methanol, when one hydroxymethyl group is lost with formation of the sodium salt of nitropropanediol, (HOCH2)2^*C = N02Xa. On treatment with acid the nitro group is split off and we get the ketone formed. It is not possible to crystallize the dihydroxyacetone at this stage, and we have to add carrier material in order to get it out. But the activity is certainly there, as shown by constant activity on recry- stallization and the preparation of derivatives. We are currently studying the formation of kojic acid from this compound labelled in the 2 position. It may be that the 80 per cent of activity that we find in Phenylalanine and Tyrosine Synthesis 225 r carbon atom 1 arises by the C5-C1 mechanism that Dr. Bloch has been discussing; we have httle information about that at the moment, except that carbon dioxide and formic acid are not incorporated into the kojic acid to any extent. Arnstein: I would hke to add one experiment which Dr. Bentley did not mention, and that is that in these fermentation reactions we have collected the CO 2; the specific radioactivity of the CO 2 during the early part of the fermentation is considerably higher than that from the later periods. That would indicate a preferential oxidation of the 1 -carbon of glucose and not an equal oxidation of all carbon atoms. Pochin: As most of the glucose in themedium is unlabelled, Dr. Block, you presumably would get only a very small proportion of your recombined trioses in the phenyl ring with two labellings, and does not your argu- ment depend upon such double labelling? Rittenberg: I think Dr. Rloch's line of reasoning is perfectly correct, because he does not insist that in his single phenylalanine two positions he labelled, but merely that in the mixture he finds both positions labelled. Dr. Bloch, as I remember your tables, the j8-position is much more active than the 2,6-position, and offhand I should think they should be the same. Bloch: This is correct, The ^-position has a specific activity very close to the value which one would expect from a 3>-carbon unit formed in glycolysis (1900 as compared to a calculated value of 2400). The activity in the ring, when recalculated for undiluted material, is very similar to this value, although it should be twice as high if the benzene ring were formed directly from glucose. It should have one labelled carbon atom with the same specific activity as that of glucose. The count in the j3 carbon of the side chain indicated that the ^ carbon is derived from a 3-carbon unit, which has half the isotope concentration of Ci in glucose. This is in agreement with the glycolytic scheme. On the other hand we cannot explain the finding that the activity of the benzene ring is only 50-60 per cent of what it should be on the basis of the cyclization scheme. Rittenberg: I would like to ask Dr. Krebs what his reaction is to the manner in which his citric acid cycle is invoked to explain much of current biochemical reactions. Krebs: The reactions which make up the tricarboxylic acid cycle can certainly occur in yeast cells and in many types of biological material, but I doubt whether in yeast cells they are the main mechanism of the oxidation of acetate or carbohydrate. In cells like yeast which grow rapidly there are two types of major reactions going on, those which build up the cell material and those which supply energy. If we find that reactions take place at a fairly fast rate it remains to be investigated whether they are required for one or the other type of reaction. Many facts are in favour of the assumption that the synthesis of citric acid in yeast cells is a link in synthetic processes, being con- cerned, for example, with the supply of a-ketoglutaric acid for the synthesis of glutamic acid. One reason in favour of this view is the ISOTOPES 16 226 KoNRAD BloCh fact that at least one step of the tricarboxyHc acid cycle is extremely slow in yeast, namely the conversion of malate into oxaloacetate. A recent paper by Foulkes (1951, Biochem. J., 48, 378) shows that the reactions of citric acid are also slow. Thus I regard it as fair to invoke the reactions of the tricarboxylic acid cycle to explain a number of synthetic processes, because the reactions can occur, though I doubt the validity of the assumption that the tricarboxylic acid cycle is the major pathway of respiration in micro-organisms. Wood: I would like to raise the question of whether we are on solid ground when we use succinate oxidation as a measure of the occurrence of the Krebs cycle. Do we know the actual intermediates of the Krebs cycle? Perhaps we know the carbon skeleton of each intermediate, but I do not believe we know the structure of the actual intermediates. There have been several experiments, especially in the isotope field, where unlabelled a-ketoglutarate has been added as a pool to trap isotopically labelled a-ketoglutarate that presumably would be formed during the oxidation of a labelled substrate. I question very much whether any definite conclusion can be drawn from the fact that the pool of a-ketoglutarate is not labelled as much as expected, because I am not at all certain that a-ketoglutarate as such is the real intermediate, or in the case of succinate, that succinate as such is the intermediate. If one considers the numerous intermediates of anaerobic glycolysis and how little we knew a few years ago of the true structure of the intermediate compounds of this transformation, one is certainly justified in thinking that the intermediates in the Krebs cycle may be different than we now believe, and that we may not detect or recognize them by the usual methods. Krebs: In general terms I agree with what Dr. Wood has said, but in animal tissues the position is somewhat different because the reactivity of all the intermediates of the tricarboxylic cycle can be easily demonstrated in this material. In yeast the evidence is still largely missing, and it is not impossible that the intermediates are not succinate or malate but some closely related substances. The main point is that at present the evidence in support of the view that the tricarboxylic acid cycle is a major respiratory mechanism in yeast should not be regarded as complete. PART VI CARBOHYDRATE AND FATTY ACID METABOLISM A STUDY OF ACETONE METABOLISM USING GLYCOGEN AND SERINE AS INDICATORS, AND THE ROLE OF Ci-COMPOUNDS IN METABOLISM HARLJND G. WOOD The first experiments concerning the distribution of tracer carbon in the ghicose unit of hver glycogen were reported in 1945 and showed that the normal rat fixes CO 2 in the 3 and 4 positions of the glucose (Wood, Lifson and Lorber, 1945). The results were of special interest because they demonstrated that it was possible to confirm in the intact aniinal concepts of carbohydrate metabolism that had been based largely on studies with enzyme preparations from a wide variety of tissues. These concepts included the tricarboxylic acid cycle, fixation of CO 2 in oxaloacetate, and the synthesis of glycogen by the reversal of the reactions of anaerobic glycolysis. The finding that CO 2 carbon occurred only in the 3 and 4 carbons was most gratifying not only because it verified the predic- tions arrived at by in vitro studies but because it showed the possibilities for other similar in vivo studies. These experi- ments marked the beginning of the use of the distribution patterns of the tracer in glycogen as an indicator of metabolism studied in vivo. Numerous experiments from several labora- tories have yielded a body of information which permits rather reliable deductions about mechanisms of intermediary metabolism. 227 228 Harland G. Wood In animals it is now well established that a labelled sub- strate which gives rise to carboxyl-labelled "acetate"* or pyruvate gives rise in turn to a 3, 4-labelled glucose, and a substrate that yields methyl-labelled "acetate" leads to the formation of a sugar labelled equally in the 1, 2, 5 and 6 positions. If pyruvate or lactate with the highest labelling in the beta position is fed or formed, the glucose is labelled highest in the 1, 6 positions; on the other hand if the labelling is highest in the alpha position the 2, 5 positions of the sugar contain the most tracer. Thus far the distributions found in animals have been uniformly in accord with present concepts of glycolysis and oxidation by the tricarboxylic acid cycle. f ♦"Acetate," "formate" and "pyruvate" denote the compound itself or a metabolite thereof which is the active form of the compound. fThere most likely are other mechanisms of breakdown but apparently they either are not of major quantitative significance or the present method with glycogen is not suitable to detect them. For example, there is considerable evidence of a terminal oxidation of glucose to phosphogluconic acid and its breakdown by subsequent decarboxylation (Dickens, 1938; Scott and Cohen, 1951; Horecker and Smyrniotis, 1950). Of interest in this respect are the observations of Gibbs and Gunsalus (private communication) with Leiiconostoc mesenteroides. They have shown that these bacteria ferment [l-^^C]-glucose to the folloAving products: — i4C-C-C-C-C-C->i4C02 + CHg CH2OH + COOH CHOH CHj 12 3 4 5 6 Contrary to the usual alcoholic type of fermentation the CO2 in this case arises from the 1 position. AVith [3,4-^^C]-glucose the labelling was: — C-C-i*C-"C-C-C^C02 + CH3 i^CHoOH + i^COOH CHOH CH3 12 3 4 5 6 If the above conversions were reversible it would be expected that CO, would enter position 1 of the glucose. Apparently such a reversible conversion does not occur in rats, or if it does so it must be slow in the liver. Shreeve, Feil, Lorber and Wood (1949) considered that there was some activity in the 1, 2, 5, 6 positions of liver glycogen but Gibbs, Dumrose, Bennet and Bubeck (1950) have shown that the bacterial degradation to lactate is not completely specific and this non-specificity may account for the findings of Shreeve et al. (1949). Gibbs et al. found about 3 per cent of the 3, 4 position of glucose to be converted to the methyl group of lactate during the fermentation. It is not yet clear how the small amount of tracer gets into the 2, 5 position of the glucose or the a-carbon of lactate. This in itself is an interesting problem. In support of the view that COg is only fixed in the 3, 4 position. Topper and Hastings (1949) have found, using liver slices, that there is no detectable fixation of COj in the 1 and 2 positions of the glucose of glycogen. A chemical degradation was employed in this case and the activity in the sugar was high Acetone Metabolism 229 The results with dififerent types of labelled fatty acids are also in accord with the predictions of j8-oxidation and thus lend support to the occurrence of these reactions in the intact animal (Lifson, Lorber, Sakami and Wood, 1948). The recent investigations by Sakami best illustrate the usefulness of glycogen as an indicator of metabolism. His first studies on formate metabolism (Sakami, 1948) were based on this technique and led to the discovery that formate is incorporated in the j8-carbon of serine. It was known that glycine was glycogenic but there was no known mechanism for a net synthesis of glycogen from glycine. The most attractive explanation appeared to be the one shown below which included reversal of the reaction which Shemin (1946) had postulated for the formation of glycine from serine. IP^COOH+CHsNHgi^COOH -> i^CHgOH • CHNHg • i=^COOH i^CHsOH • CHNH2 • 13C00H -> 14CH3 • CO • i^COOH 214CH3 • CO • 13C00H -> 14C . 14C . i4,i3C . 14,13C . 14C . 14C ^*C-Formate and CHgNHa^^COGH wxre therefore given to rats and the glycogen w^hich was formed w^as isolated and degraded. The ^^C was found only in the 3, 4 carbons and the highest ^*C labelling was in the 1, 6 carbons. These results, therefore, fit the predicted distribution pattern of the glycogen as made from the above equations. To verify further the scheme, Sakami (1948) isolated serine from the liver pro- tein of the above experiments and degraded it by periodic acid oxidation to obtain formaldehyde (/8-C), formate (a-C) and CO 2 (carboxyl-C). The labelling of the serine was found to agree with the above equations, i.e. the ^*C was in the in the 3, 4 positions so even a small fixation in other positions should have been detectable. Another interesting example of a new type of fermentation is that studied by Lampen, Gest and Sowden (1951). [l-^^CJ-Xylose was fermented with Lactobacillus pentosus and gave the following labelling of products: — "C-C-C-C-C-^^CHj • COOH + COOH CHOH CH3 12 3 4 5 The aldehyde of the xylose thus becomes the methyl group of acetate. What part the reversal of such a cleavage may play in the synthesis of pentose is a question of great interest. 230 Harland G. Wood j8-position and the ^^C was in the carboxyl-position. Following this discovery it was possible to use serine as an indicator of metabolism, i.e. it could be used to detect the formation of "formate" from other labelled substrates. With this method much information was obtained relative to the role of 1 -carbon compounds in metabolism. Evidence has been obtained with serine as an indicator, that "formate" is formed from the a carbon of glycine (Sakami, 1949a), from the methyl groups of choline (Sakami, 19496), and from the methyl groups of acetone (Sakami, 1950). The demonstration that methyl groups are converted to "formate" made it seem likely that the methyl groups of methionine and choline might be synthesized from formate by a reverse reaction. When this idea was tested by Sakami and Welch (1950), it w^as found that the synthesis occurred. In the present report the use of glycogen and serine as indicators of intermediary metabolism of acetone will be considered in some detail. In addition some recent studies with the propionic acid bacteria will be reported since they reinforce the idea that the C^-compounds have an important role in metabolism. ^ » Acetone Metabolism Although it has been recognized that acetone is formed by animals, especially in diabetes mellitus, it has generally been considered that acetone is formed by spontaneous decarboxy- lation of acetoacetate and that once formed it plays no part in metabolism but is excreted as such. This idea has been held in spite of the fact that from time to time reports have appeared which indicated that acetone is utilized (Schwarz, 1898; Rothkopf, 1936). The recent tracer experiments of the Columbia group (Borek and Rittenberg, 1949; Price and Rittenberg, 1950) served to refocus attention on this problem. These workers demonstrated that rats utilize small amounts of acetone efficiently and that it is oxidized in part to COo and in part to "acetate." The latter was indicated by the observed conversion of acetone to cholesterol and by the Acetone Metabolism 231 acetylation of phenyl- y-aminobutyric acid with acetone. Here we see the use of another indicator of metabohsm. In this case, the acetylation of a foreign amine is used as evidence of the occurrence of "acetate" in the intermediary metabolism of a compound (Bloch and Rittenberg, 1944). The above results could be explained by a pathway of utilization of acetone via carboxylation to acetoacetate, with subsequent oxidation of the acetoacetate. C*H3COC*H3 + C02 -> C^HgCOC^Ha-COOH Evidence that the carboxylation reaction occurs has been obtained by Plant and Lardy (1950) and Coon (1950). It would be expected that the acetoacetic acid from the above reaction would be metabolized by conversion to methyl labelled acetate and then be oxidized via the tricarboxylic acid cycle: — C*H3 • CO • C*H3+C02 -> C*H3 • CO • C*H2 • COOH -> "C^HsCOOH" via Krebs cycle -> C*-C*-CO-CO-C*-C* (glucose) and glycolysis Accordingly, the distribution of the tracer in the glucose would be expected to be like that obtained with methyl labelled acetate (Lifson et al., 1948). However when acetone was tested (Sakami, 1950) the results (shown below) were found to be quite different from that found with methyl labelled acetate.*)* ♦Indicates highest activity, o indicates lower activity. I In these and subsequent experiments a comparison will be made between the activities of glycogen obtained under conditions which are not always identical. Although the absolute values are not comparable between the experiments the relative distribution pattern probably is comparable. The method of conducting the glycogen experiments was to fast the animal for 24 to 48 hours and then administer the isotopic compound along with un- labelled glucose or glycine. After 3 to 10 hours the animals were sacrificed and the liver was worked up. When glycine was being used a longer time was allowed because of the delayed glycogen deposition with this substrate (MacKay, Wick and Carne, 1940). Usually 2-5 mM. of glycine or 2-5 mM. of glucose were given by stomach tube per 100 g. of rat and 0-5 to 2 mM. of the labelled compound containing 1-5 to 0-5x10* counts/min. Sometimes the labelled compounds were given intraperitoneally, sometimes subcutan- eously. The glucose from the glycogen was usually degraded by bacterial fermentation with Lactobacillus casei (Wood et al., 1945). 232 Harland G. Wood Ci - - c, - - C3 - - c, - - c, - - c. i^CHg • COOH 176 177 157 157 177 176 i^CHgCO-i^CHa 447 340 399 399 340 447 (The values in this and subsequent diagrams are in counts/min./mg. C.) With methyl labelled acetate the tracer occurred equally in the 1, 2, 5, 6 while with the acetone the tracer was higher in the 1, 6 than the 2, 5 positions and in addition the 3, 4 positions contained a high concentration of ^*C. It therefore appeared that the major pathway of acetone metabolism was not via acetoacetate and "acetate." In evaluating these and the subsequent results it should be borne in mind that the 1, 6 and 2, 5 labelling is indicative of the relative isotope concentration of the methyl and carbonyl carbon of the precursor "pyruvate" that took part in the synthesis of glycogen. On the other hand, the 3, 4 position is indicative of the labelling of the carboxyl carbon of the "pyruvate." Apparently, therefore, the composite "pyru- vate" as formed from the acetone was labelled highest in the methyl and carboxyl groups. Disregarding for the moment the high labelling of the 3, 4 positions and considering only the 1, 2, 5, 6 positions, it is seen that the distribution found with acetone was comparable to that obtained with methyl labelled lactate or with labelled formate as illustrated below: — C -C -C -C -C -C i^CHgCHOHCOOH 826 679 204 204 679 826 (Lorber et al., 1950fl) Hi^COOH 512 321 212 212 321 512 (Sakami, 1948) The idea that "formate" might arise from acetone by a C2 and Cj split was therefore considered a possibility: — C*H3 • CO • C*H3-^"C*H3COOH" + "HC*00ir' / \ Q*_Q*_QO_QO_Q*_Q* Q*_QO_QO_QO_QO_Q* and H0C*H2 • CH(NH2) • COOH Acetone Metabolism 233 The resulting labelled "formate" would then yield 1, 6 labelled glucose as previously discussed and would be mixed with glucose formed simultaneously from the "acetate" but this part of the glucose would be equally labelled in the 1, 2, 5, 6 positions and would not mask the detection of "formate." However, the formation of labelled "formate" from acetone would not account for the high labelling in the 3, 4 positions and it therefore was considered that there might be a direct conversion of the acetone to "pyruvate." C*H3-COC*H3 -> ''C*H3-C0C*00H" -> C*-CO-C*-C*-CP-C* In fact this type of labelled pyruvate would in itself yield a glucose that would fit the results since from it a glucose would presumably be formed which would be labelled high in both the 1, 6 and 3, 4 positions. Sakami's first tests were concerned with the hypothesis that the acetone was split to "formate" and "acetate." For this test he used serine as the indicator. The serine was isolated from the liver protein (Sakami, 1950) and was degraded. The results of this experiment, as well as one in which labelled formate was given (Sakami, 1948), are shown below: — HOCH2 - CH(NH2) - COOH 420 15 18 Hi^COOH 377 20 4 14CH3-C0 14CH3 420 15 18 Since the labelling of the ^-carbon of serine in the two experi- ments was of the same order of magnitude and the concen- tration and activity of the administered formate and acetone were roughly the same, the indications are that the Cg and C^ pathway of acetone metabolism may be of major quantitative significance. When C*H2-CO*C*H2-COOH was administered in similar concentration, the labelling in the serine was much lower (Sakami, 1950):— HOCH2 - CH(NH2) - COOH 14CH3CO14CH2COOH 15 7 4 234 Harland G. Wood These latter results are a good indication that the acetoacetate was not metabolized via acetone, and the data as a whole indicate that there was not an extensive interconversion of acetone and acetoacetate. As noted previously Cg and C^ cleavage of methyl-labelled acetone could account for the 1, 6-labelled glucose but not for the high labelling in the 3, 4 positions of the glucose. It appeared possible to test for a direct oxidation of acetone if carbonvl labelled acetone were used. This is illustrated below: — 1 . Carbox vlation of acetone CH3 • C*0 • CH3 + COo ->CH3 • C*0 • CHg • COOH ->"CH3 • C*OOH"^ C-C-C*-C*-C-C 2. C2 and C^ split CH3C*OCH3 -^ CH3C*00H+HC00H -> C-C-C*-C*-C-C 3. Direct oxidation CH3 • C*0 • CHg-^'THg • C*0 • C00H"^CO-C*-CO-CO-C*-CO Only in case of direct oxidation should the tracer occur in all positions of the glucose unit, and in addition it should be highest in the 2, 5 positions. The results of this experiment (Sakami and Lafaye, 1951) are shown below, as well as an experiment with a-labelled lactate (Lorber et al., 1950a): — CH3ICOCH3 180 283 278 278 283 180 CH3 • 13CHOH • COOHt 18 -27 05 05 .27 -18 It is seen that the carbonyl labelled acetone gave a tracer distribution pattern in the glycogen quite similar to a-labelled lactate, i.e. the labelling was high in the 2, 5 positions. Thus direct oxidation is indicated to occur. The only difference in fThe experimental values are given as atoms per cent excess ^^C. Acetone Metabolism 235 distribution with the acetone and lactate is that the 3, 4 positions are labelled higher in the case of the acetone. The high labelling in the 3, 4 positions, is considered to reflect the simultaneous occurrence of the C 2 and C\ split of acetone, which yields carboxyl labelled "acetate" and thus 3, 4 labelled glucose. The results are thus in complete agreement with the idea that acetone may be oxidized in part to yield a 3-carbon carbohydrate precursor directly and may in part be cleaved to Cg and C^ compounds. Although the results with 2, 4 labelled acetoacetate made it fairly certain that acetone Avas not mxCtabolized via aceto- acetic acid, or acetoacetic acid via acetone, Sakami and Lafaye (1951) repeated the experiment with 1, 3-labelled acetoacetate. The results were as follows: — c. - - c. - C3 - c, - - c, - - Ce 22 26 1264 1264 26 22 CH3 • 14C0 • CH2 • i^COOH It is seen that in contrast to the results with carbonyl labelled acetone there was little labelling of the 1, 2, 5, 6 positions and thus no evidence of significant conversion of acetoacetate to acetone. Moreover, since the 3, 4 positions are highly labelled, there is no indication that dilution of acetoacetate was such as to mask detection of the conversion of acetoacetate to acetone. The significance of these two pathways of acetone meta- bolism is not apparent at present. It is well known that liver does not utilize acetoacetate as rapidly as does skeletal or heart muscle. Therefore, liver may not be the best organ for study of acetoacetate metabolism. The conversion of acetoacetate to acetone and of acetone to a carbohydrate precursor offers an attractive mechanism for the net con- version of fatty acid to carbohydrate. There has long been a controversy about such conversion but until now there has been no conclusive evidence and indeed no satisfactory mechanism for such net conversion. In heart muscle there is some indication that there may be glycogen formation from ketone bodies (Lackey, Bunde and Harris, 1947). It would 286 Harland G. Wood be very interesting to have data for heart glycogen similar to those presented here for liver. Practically all the experi- ments on the conversion of labelled fatty acids to glycogen in which isotope distribution patterns have been determined have dealt with liver glycogen. However, it is more than likely in experiments of short duration, even though done with whole animals, that the liver glycogen may in a large part reflect only liver metabolism. The significance of acetone metabolism as a source of metabolites such as "formate" is likewise an interesting problem. Propanediol or Propanediol Phosphate as a Possible Intermediate of Acetone Metabolism Propanediol phosphate was found by Lindberg (1943) to occur in sea urchin eggs, and he later isolated the ester from mammalian liver, kidney and brain (Lindberg, 1946). The highest concentration was in brain, in which the ester made up about 5 per cent of the acid soluble phosphate. The metabolic significance of this compound is relatively unknown. Lindberg (1946) has made ^^p-labelled propanediol phosphate and has found that it is fairly rapidly metabolized and that the phosphate group is transformed into an easily hydro- lysable ester. The unphosphorylated compound is well metabolized and is glycogenic (Hanzlik, Newman, Van Winkle, Lehman and Kennedy, 1939). The possibility that propanediol is an intermediate of acetone metabolism is an attractive hypothesis and is being investigated in our laboratories by Rudney (1950). The following reaction is suggested: — OH H II O HOH O CH3CCH3^CH3C:CH2+ or -^CU^C- H3PO, CH2OH -> "acetate" + "formate" H This reaction is analogous to the conversion of phosphoenol pyruvate to phosphogly eerie acid: — Acetone Metabolism 237 CH2:C(OP03H2)COOH+HOH -> CH20HCH(OP03H2)COOH In addition to C2 and C^ cleavage the propanediol might be oxidized to "pyruvate" or "lactate" and thus account for the direct conversion of acetone to glycogen. Rudney prepared CHg-CHOH-^^CHgOH by reducing car- boxyl labelled lactic acid. He tested this compound in rats, using liver serine and liver glycogen as indicators (Rudney, 1950). The distribution patterns were as follows: — {^ f^ f^ (^ f^ f^ CH3 • CHOH • 14CH2OH 66 48 448 448 48 6% HOCH2 - CH(NH2) - COOH CH3 • CHOH • i^HgOH 166 3 14 In addition, the methvl carbons of the choline of the viscera and liver were isolated and found to have 66 counts/min./mg. C. The maximum activity in the respiratory CO 2 occurred 1 hour after the administration of the ^*C propanediol and contained 500 counts/min./mg. C. It appears from these results that the propanediol is meta- bolized very much as is acetone. It labels serine and choline, and thus the indications are that a 1 -carbon compound is formed from the hydroxymethyl carbon of propanediol. The distribution of tracer in the glycogen is similar to that obtained with acetone, in that the 1, 6 carbons are labelled higher than the 2, 5 carbons, but differs in that the concen- tration in the 3, 4 positions is higher relative to 1, 2, 5, 6 than was found with acetone. The high activity in the 3, 4 positions is probably indicative of direct conversion of the propanediol as a 3-carbon unit to a carbohydrate precursor. The activity of the respiratory CO 2 does not appear to have been high enough to account for the activity in these positions by CO 2 fixation. The relatively low activity in the 1, 2, 5, 6 positions as compared to that observed with ^^CHg • CO • ^^CHg can be accounted for if it is recalled that i*CH3-CO-^*CH3 yields methyl labelled "pyruvate" by direct conversion and methyl labelled "acetate" by C2 and Cj cleavage. Both of these 238 Harland G. Wood labelled products would contribute activity to the 1, 2, 5, 6 positions. On the other hand with CH3 • CHOH • I'^CHaOH the situation would differ in that the "pyruvate" from the direct conversion would only be labelled in the carboxyl group and the "acetate" from the Cg and C^ split would be unlabelled. These products would not label the 1, 2, 5, 6 positions. Thus, contrary to the situation wdth labelled acetone, the only source of tracer for the 1, 2, 5, 6 positions from propanediol is the "formate" of the Cg and C^ cleavage. Thus a lower activity would be expected than was observed with the acetone. The results are sufficiently encouraging to warrant further study of the relation of acetone metabolism to propanediol phosphate. Rudney has now developed methods for the isolation of propanediol phosphate in pure form from tissues, and will shortly feed labelled acetone and isolate the propane- diol phosphate to see if there is an indication of a direct conversion of the acetone to propanediol phosphate. The Role of Formaldehyde in the Propionic Acid Fermentation It is interesting that the propionic acid fermentation, which played such an important part in the discovery of the role of CO 2 in metabolism, has now assumed a somewhat similar part in the demonstration of the importance of the C\- compound, formaldehyde, in metabolism. Leaver (1950) has shown in our laboratories that [^*C] formaldehyde is utilized by the propionic acid bacteria during the fermentation of a number of substrates, and that the tracer is fixed in every position of the resulting propionic acid. Typical results are as follows: — H14CH0 Hi^CHO Hi^CHG The formaldehyde was fixed in approximately equal amount in the a and ^ positions of propionate and in still greater CH3 - CH2 -COOH Substrate 132 137 304 erythritol 7G 71 131 glycerol 173 187 244 pyruvate Acetone Metabolism 239 amount in the carboxyl groups. Since CO 2 is fixed only in the carboxyl group it is evident that the formaldehyde was not used in the form of CO 2. The formaldehyde was added to a resting cell fermentation in -001 m. concentration and contained 142,000 counts/min./mg. C. Almost one-half of the added activity was recovered in the propionic acid. Additional activity was found in the succinate and CO2, and a small amount in the acetate. The succinate was labelled in both the carboxyl and methylene positions, but did not always have a distribution similar to that in the propionate. The reactions by which the fixation of formaldehyde occurs are obscure, but this discovery certainly opens up new possi- bilities of revealing the mechanism and of studying the propionic acid fermentation. It has never been completely certain how propionate is formed in this fermentation and the interrelation of succinate and propionate has remained an interesting problem. Although there is evidence that propionate may be formed by decarboxylation of a C4 dicar- boxylic acid, it is not at all certain that this is the major pathway of its formation (Wood, 1942; Delwiche, 1948, 1950). In this connection it is of interest that either a or ^8 labelled propionate yields equal labelling in the 1, 2, 5, 6 positions of rat liver glycogen (Lorber, Lifson, Sakami and Wood, 19506). These results are quite different from those with a or j3 labelled lactate, with which the labelling is always unequal in the 1, 2, 5, 6 positions, apparently because of direct conversion of the lactate to glycogen (Lorber, Lifson, Wood, Sakami and Shreeve, 1950a). Apparently with propionate there is little or no direct conversion, and the conversion to glycogen involves complete randomization of the a or jS carbon of propionate. Perhaps the reason this occurs will be revealed when we understand more about the reactions whereby propionate is formed from glucose in the bacterial fermentation. The most recent investigations by Leaver (unpublished) are of further interest because they appear to show that free formaldehyde is formed in the propionic acid fermentation. 240 Harland G. Wood This is a most important observation because it indicates the possibihty that formaldehyde per se may be an intermediate. The conversion of added formaldehyde or formate* to a fermentation product is of course not proof that the formal- dehyde or formate is an intermediate. These added com- pounds may be converted to the true intermediate or possibly they may never occur in a normal fermentation. Although formaldehyde has been detected previously in the propionic acid fermentation by use of the fixative dimedon (Wood and Werkman, 1935), in the present case Leaver detected it without the use of a fixative. [1, 3-i*C]-glycerol| was fermented by resting cells in the presence of a pool of 8 x 10* m. unlabelled formaldehyde. After 18 hours the fermentation was stopped and the residual formaldehyde was distilled and determined in an aliquot colorimetrically (Alexander, Landweler and Seligman, 1945). To the 28 /lim. of formaldehyde that remained, unlabelled formaldehvde was added as a carrier and the formaldehyde was isolated as the dimedon derivative. This derivative was recrystallized from alcohol and water to a constant activity of 119 counts/min./mg. C. It was then recrystallized from acetone and water and contained 117 counts/min./mg. C. The melting point and mixed melting points were 190°. When corrected for dilution by the carrier formaldehyde, the formaldehyde recovered from the fermen- tation was calculated to contain 720 counts/min./mg. C. The original glycerol contained 808 counts/min./mg. of labelled carbon. It would thus appear that the pool of added for- maldehyde was largely replaced by ^*C -formaldehyde formed ♦Leaver (1950) found that formate is not fixed in the propionic acid fer- mentation. In the synthesis of serine, Siegel and Lafaye (1950) found that formaldehyde was utilized better than formate by rat liver homogenates and they suggested that formaldehyde is a more direct precursor of the jS-carbon of serine than is formate. Siekevitz and Greenberg (1950) found that rat liver slices do not convert formate to formaldehyde although they do produce formate from the a carbon of glycine. In the synthesis of methyl groups of choline (du Vigneaud, Verly and Wilson, 1950) both formate and formaldehyde are utilized, but Berg (1951) has found that formaldehyde per se probably is not an intermediate in the formation of methyl groups of methionine from formate. •{•The [1, 3-^*C]-glycerol was a generous gift from Dr. M. L. Karnovsky. Acetone Metabolism 241 from the labelled positions of the glycerol. Control deter- minations on the unfermented glycerol were found not to yield formaldehyde. Formaldehyde has also been isolated from resting cell fermentations to which no pool of unlabelled formaldehyde was added. A similar attempt to demonstrate the formation of formaldehyde from uniformly labelled glucose gave no activity in the residual formaldehyde pool. When [1, 3-^*C]-glycerol was fermented and the propionate was isolated, the distribution pattern in the propionate was as follows: — CH3 - CH2 - COOH HQi^CHg • CHOH • i^CHgOH 397 403 675 The succinate contained 525 counts/min./mg. C in the car- boxyl group and 425 counts/min./mg. C in the methylene carbons. The distribution is strikingly similar to that observed when labelled formaldehyde was added to the fermentation. This again suggests that glycerol may be metabolized via formaldehyde. Although no attempt will be made at present to consider the details of the mechanism of the propionic acid fermen- tation, it is tempting to suggest that formaldehyde may play an important role in the fermentation of 4 and 5 carbon polyalcohols. In these fermentations propionate and some succinate are the major products, and it seems obvious that there must be extensive cleavage and resynthesis. Isotope dilution experiments done by Leaver indicate that CO 2 is not turned over in sufficient quantity to suggest that it has a major role as a synthetic unit in these fermentations. Perhaps formaldehyde assumes this role. - It is interesting that serine and propanediol with animals and glycerol with bacteria yield "formate" or formaldehyde. One is reminded of the fact that oxidation of these com- pounds with periodic acid gives formaldehyde from the primary carbinol of each. One wonders whether there may not be an analogous biological oxidation, so that compounds which are activated and contain two adjacent hydroxy Is or ISOTOPES 17 242 Harland G. Wood an adjacent hydroxy! and amino group yield formaldehyde or "formate." A reversal of these reactions would constitute a synthesis with formaldehyde. It is rather a strange coincidence that formaldehj^de was postulated as the major intermediate of photosynthesis for many years but then lost favour. It is now almost certain that it will once more receive serious consideration in photo- synthesis, though probably not in terms of the old proposal of a carbon-bj^-carbon build up of sugars by aldol condensa- tion of formaldehyde. Formaldehvde is also formed in the oxidation of N-methvl groups, as first observed by Handler, Bernheim and Klein (1941) and recently confirmed by Mackenzie (1950). One looks with new interest on the experiments with Fusarium (Goepfert, 1941), in which formaldehyde was obtained during the oxidation of acetone, isopropyl alcohol and propanediol when dimedon was used as a fixative. Summary* Glycogen and serine have proved to be very useful indicators of the mechanism of metabohsm of acetone by animals. Evidence is reviewed which apparently shows that there is a Cg and Ci cleavage of acetone, and in addition a conversion of the intact 3-carbon chain to a carbohydrate precursor. No evidence has been obtained by the glycogen and serine indicator method that a major pathway of metabolism of acetone occurs via its carboxylation to acetoacetic acid, nor was there evidence that acetoacetic acid was metabolized to any significant extent via acetone. These results are con- *The bibliography is by no means complete because the discussion is mostly intended to illustrate how glycogen and serine may be used as indicators of metabolism. It should be noted that Sonne, Buchanan and Delluva (1948) demonstrated that formate was converted to the 2 and 8 positions of uric acid, and thus drew attention to the fact that formate might play a role in metabolism by animals. Greenberg's group in California and Neuberger's in England have been very active in studies of serine synthesis, and Ehrensvard and co-workers in Sweden found in 1947 that glycine was converted to serine by Torula. Hastings' group at Harvard has long been active in glycogen studies and has done excellent work in determining the isotope distribution pattern of glycogen. Acetone Metabolism 243 sidered to be more typical of liver than other organs and may not necessarily reflect, the metabolism of the latter. Evidence indicating that propanediol may be an intermediate of acetone metabolism is reviewed. Further indication of the role of formaldehyde in meta- bolism is drawn from the experiments with propionic acid bacteria. These bacteria apparently convert the terminal carbon of glycerol to free formaldehyde. Furthermore, they appear to utilize formaldehyde very extensively in the syn- thesis of propionic acid, the formaldehyde carbon being incorporated in every position of the molecule. It seems quite certain that, in addition to utilization of COo, there is exten- sive utilization of other C^-compounds in metabolism. The understanding of how these Ci-compounds are utilized will no doubt lead to new important concepts of the mechanism of metabolism. The author wishes to express his sincere thanks to Dr. Warwick Sakami, to Mr. Harry Rudney and Mr. Fred Leaver for permitting him to use unpublished data that they have obtained and for their many valuable comments. REFERENCES Alexander, B., Landweler, G., and Seligman, A. M. (1945). J. biol. Chem., 160, 51. Berg, P. (1951). J. bioL Chem., 190, 31. Block, K., and Rittenberg, D. (1944). J. biol. Chem., 155, 243. BoREK, E., and Rittenberg, D. (1949). J. biol. Chem., 179, 843. Coon, M. J. (1950). J. biol. Chem., 187, 71. Delwiche, E. a. (1948). J. BacL, 56, 811. Delwiche, E. a. (1950). J. Bad., 59, 439. Dickens, F. (1938). Biochem. J., 32,-1626, 1645. DU ViGNEAUD, V., Verly, W. G., and Wilson, J. E. (1950). J. Amer. chem. Soc, 72, 2819. GiBBS, M., DuMROSE, R., Bennet, F. a., and Bubeck, M. R. (1950). J. biol. Chem., 184, 545. Goepfert, G. F. (1941). J. biol. Chem., 140, 525. Handler, P., Bernheim, M. L. C, and Klein, J. R. (1941). J. biol. Chem., 138, 211. Hanzlik, p. J., Newman, H. W., Van Winkle, W., Jr., Lehman, A. J., and Kennedy, N. K. (1939). J. Pharmacol., 67, 101. 244 Harland G. Wood HoRECKER, B. L., and Smyrniotis, P. Z. (1950). Arch. Biochem., 29, 232. Lackey, R. W., Bunde, C. A., and Harris, L. C. (1947). Proc. Soc. exp. Biol., N.Y., 66, 433. Lampen, O. J., Gest, H., and Sowden, J. C. (1951). J. Bad., 61, 97. Leaver, F. W. (1950). J. Amer. Chem. Soc, 72, 5326. LiFSON, N., LoRBER, V., Sakami, W., and Wood, H. G. (1948). J. biol. Chem., 176, 1263. LiNDBERG, O. (1943). Ark. Kemi., 16A, No. 15. LiNDBERG, O. (1946). Ark. Kemi., 23A, No. 2. LoRBER, v., LiFSON, N., WooD, H. G., Sakami, W., and Shreeve, W. W. (1950a). J. biol. Chem., 183, 517. ' Lorber, v., Lifson, N., Sakami, W., and Wood, H. G. (19506). J. biol. Chem., 183, 531. MacKay, E. M., Wick, A. N., and Carne, H. O. (1940). J. biol. Chem., 132, 613. ^Mackenzie, C. G. (1950). J. biol. Chem., 186, 351. Plaut, G. W. E., and Lardy, H. A. (1950). J. biol. Chem., 186, 705. Price, T. D., and Rittenberg, D. (1950). J. biol. Chem., 185, 449. RoTHKOPF, H. (1936). Z. ges. exp. Med., 99, 464. RuDNEY, H. (1950). Arch. Biochem., 29, 231. Sakami, W. (1948). J. biol. Chem., 176, 995. Sakami, W. (1949a). J. biol. Chem., 178, 519. Sakami, W. (1949&). J. biol. Chem., 179, 495. Sakami, W. (1950). J. biol. Chem., 187, 369. Sakami, W., and Lafaye, J. M. (1951). J. biol. Chem., in press. Sakami, W., and Welch, A. D. (1950). J. biol. Chem., 187, 379. ScHWARz, L. (1898). Arch. exp. Path. Pharmak., 40, 168. Scott, D. B. M., and Cohen, S. S. (1951). J. biol. Chem., 188, 509. Shemin, D. (1946). J. biol. Chem., 162, 297. Shreeve, W. W., Feil, G. H., Lorber, V., and Wood, H. G. (1949). J. biol. Chem., 177, 679. SiEGEL, I., and Lafaye, J. (1950). Proc. Soc. exp. Biol., N.Y., 74, 620. SiEKEViTZ, P., and Greenberg, D. M. (1950). J. biol. Chem., 186, 275. Sonne, J. C, Buchanan, J. M., and Delluva, A. M. (1948). J. biol. Chem., 173, 69. Topper, Y. J., and Hastings, B. (1949). J. biol. Chem., 179, 1255. Werkman, C. H., and Wood, H. G. (1942). Advances in Enzymology, 2, 160. Wood, H. G., Lifson, N., and Lorber, V. (1945). J. biol. Chem., 159, 475. Wood, H. G., and Werkman, C. H. (1935). J. Bad., 30, 652. DISCUSSION Krebs: Can formic acid be replaced in the animal experiments by methyl alcohol or formaldehyde? Wood: Siegel and Lafaye in our laboratory, using rat liver homo- genates, have shown that formaldehyde enters serine even more rapidly Acetone Metabolism 245 than does formate. The Neuberger group has shown that other 1 -carbon compounds are also incorporated into serine. There is no doubt that formaldehyde has to be considered. I think that the results of pro- pionic acid fermentation are especially significant, because in this case we know that formaldehyde itself is actually formed. Rittenberg: Dr. Price in my laboratory has been interested in the oxidative route taken by acetone, especially with regard to its utiliza- tion for cholesterol synthesis. In tissue slice experiments he has found by use of the washing-out procedure that pyruvic acid does not diminish the utilization of methyl-labelled acetone for cholesterol formation. This presumably means that pyruvate is not on the pathway from acetone to acetate. Curiously enough, methyl glyoxal does diminish the utilization of acetone for cholesterol synthesis. We are not sure now whether we are dealing with a true intermediate or whether methyl glyoxal by some indirect method interferes with this conversion. Wood: When I put pyruvate in quotes, that implies any Cg precursor of glycogen: methyl glyoxal, lactic acid. I think there is a fair chance that methyl glyoxal might be formed. Block: If I remember correctly, the isotope concentrations in carbons 1 and 6 were about 20 per cent higher than in carbons 2 and 5 after methyl-labelled pyruvate or lactate. Would you interpret this to indicate that the direct conversion of pyruvate to glucose is a rela- tively minor pathway, and that the major part of pyruvate is meta- bolized by way of the Krebs cycle, at least in the liver? Wood: Yes. I would interpret it that way, except I don't know whether the pyruvate is metabolized via the entire cycle or just shuttles back and forth very rapidly to C4 carboxylic acids. The fixation reaction is much faster than our previous results indicated. Dr. Utter now has found in the carboxylation reaction that oxaloacetate is almost com- pletely equilibrated with CO 2 in 10 minutes. Thus the conversion of CO2 and pyruvate to oxaloacetate is not a slow reaction. ASYMMETRIC CITRIC ACID CHARLES HEIDELBERGER and VAN R. POTTER The classical investigations of Krebs on the oxidation of carbohydrates and their metabolites led him to formulate the now widely accepted citric acid cycle that bears his name. The subsequent use of isotopic tracers has in many cases COg + CH^CCXIOOH HOOCCHoCO + (-CHoCOX) COOH ^ HOOC-CHgCOHCHgCOOH COOH (Not isolated) HOOCCOCHoCHoCOOH I Pigeon liver mince COg + HOOCCH2CH2COOH 100^ 0% Fig. 1. Distribution of COg fixed into a-ketoglutarate (Evans and Slotin, 1941; Wood et ah, 1942). The percentage values showTi on this and the following figures denote the proportion of the total isotope content of a substance found in individual chemical groupings. strengthened the evidence for this cycle. The important demonstration by Evans and Slotin (1941) and by Wood and his group that animal tissues are able to incorporate carbon dioxide into intermediary metabolites marked the beginning of the use of carbon isotopes in animal metabolism. When ^^COg or ^^C02 was incubated with pigeon liver minces the isotope was present in the a-ketoglutarate, but when this 246 Asymmetric Citric Acid 247 compound was oxidized the isotope was found to be present only in the carboxyl group adjacent to the carbonyl carbon as shown in Fig. 1 (Evans and Slotin, 1941; Wood, Werkman, Hemingway and Nier, 1942). It was universally agreed that citric acid, a symmetrical molecule, could not have been an intermediate in this process, and hence it was thought not to be a direct intermediate in the cycle. The brilliant proposal of Ogston (1948) that isotopic citric acid would probably behave in an asymmetric fashion in enzymatic reactions has led to a resolution of the difficulties. According to this idea, labelled citric acid should exist in two isotopically antipodal forms as illustrated in Fig. 2. This proposition was peculiarly easy to test experimentally. It was only necessary to repeat the earlier experiments, this HOOCv^ ^CHr>COOH HOOCCHo- ^COOH H00CCH2'^^ ^OH HO-^ "^CHgCOOH Fig. 2. Isotopically antipodal forms of citric acid. time sub-divided into two steps and with isolation of the citric acid itself. When ^^COg was fixed in a rat kidney homogenate, the citric acid isolated was radioactive, and when it was used as the labelled substrate in a rat liver homogenate, the a-ketoglutarate produced was shown by chemical oxida- tion to have a completely asymmetric distribution of ^*C (see Fig. 3, Potter and Heidelberger, 1949). This result confirmed Ogston's proposal and together with the work of Stern and Ochoa (1949) restored citric acid to its rightful place as a true intermediate in the Krebs cycle. In our experiments the labelled substrate is added to a rat kidney homogenate maintaining active phosphorylation in the presence of pools of fumarate, citrate, and with malonate to prevent recycling. The citric acid is isolated and purified by silica-gel partition chromatography (Isherwood, 1946), and its specific activity determined. This citrate is added to a second rat liver homogenate in the presence of arsenite to 248 Charles Heidelberger and Van R. Potter prevent further oxidation of the a-ketoglutarate, which is isolated and purified as the dinitrophenylhydrazone. Malo- nate is also added to the homogenate to eliminate the possi- bility of any recycling. After determination of the specific activity, the compound is oxidized with acid permanganate COg + CH^COCOOH HOoScHpCO + (-CH9COX) COOH '^ I 65^ j BOOCCH2COHCH2COOH Rat Kidney Homogenate I. COOH Rat Kidney Homogenate II. HOC5COCH2CH2COOH KMnOi^ CO2 + HOOCCH2CH2COOH lOOJg <.01Jg Fig. 3. Distribution of CO2 fixed into citrate and a-keto- glutarate (Potter and Heidelberger, 1949). to CO2 and succinic acid, and the radioactivity in each com- pound is measured. Our experiment was confirmed and extended by the elegant work of Lorber, Utter, Rudney and Cook (1950) as indicated in Fig. 4. They used oxaloacetate as the labelled substrate, partially purified enzyme preparations, and degraded the citric acid chemically to distinguish between the radioactivity in the primary and the tertiary carboxyl groups. We have adopted the latter practice in our more recent experiments. In order to illustrate more fully the stereochemical implica- tions of this work a consideration of Fig. 5 will be helpful. Ogston pointed out the necessity of assuming that there is at Asymmetric Citric Acid 249 least a three point interaction between the substrate and the enzyme. In order to illustrate the formation of citric acid we shall assume for convenience that the oxaloacetate inter- acts with the enzyme through the two carboxyl groups, which are in the plane of the paper. Furthermore, we shall assume that the "activated acetate" interacts with the enzyme, which has a configuration such that it can only react when it approaches the carbonyl group of the oxaloacetate from HOo5cH2CO + (-CH2COX) COOH HOOCCHpCOHCH^COOH Pigeon Liver Extract I. COOH HOOCCOCH2CH2COOH 1 Pigeon Liver Extract II. KltoOi^ CO2 + HOOCCH2CH2COOH 100^ 05? Fig. 4. Distribution of label from oxaloacetate in citrate and a-ketoglutarate (Lorber et al., 1950). the rear. In this way, the citric acid derived from labelled CO2 would have the configuration shown in the diagram. If it were derived from a labelled Cg fragment, it would have the antipodal configuration, since it would be a non-superim- posable mirror image of the first. This can be clearly seen by examination of molecular models. In order to consider the steps leading from citrate to a- ketoglutarate let us suppose that the asymmetric interaction of enzyme and substrate occurs at three points: with the tertiary carboxyl group at a, a primary carboxyl group at 250 Charles Heidelberger and Van R. Potter b, and with the hydroxyl group at c above the plane of the paper. When dehydration occurs specifically between the hydroxyl group and the hydrogens of the acetate group at point b, the m-aconitic acid w^ill have its double bond CH3COCOOH Pyruvic acid Step l||c*02 HOOC\ >CO hooc*ch/ Oxalacetic acid 10 CH3CO— X /-Malic acid Furaaric acid ati Succinic acid 7 aHOOC bHOOC OOCv ,.CH Citric acid .CHjCOOH c HOOC*— CO / t-co, /CHjCOOH CH,/ a-Ketoglutaric ejj— CO, HOOCv yy W -UiO HOOC^/ HOOC*C^ H c»5-Aconitic acid CH2COOH -^ CH2COOH HOOC*— CO/ Oxalosuccinic acid 5J|-2H HOOC*— CH ^ \ OH Isocitric acid Tj r\ yCH HOOCv yCHjCOOH Fig. 5. The Krebs citric acid cycle. situated towards the labelled primary carboxyl group. Once the double bond has been formed, the other reactions leading to a-ketoglutarate are straightforward. In Fig. 6 is formulated the enzymatic dehydration of the antipodally labelled citric acid, so that the double bond is situated away from the Asymmetric Citric Acid 251 isotope. It must be remembered that although this formulation is purely arbitrary, the general proposition must hold true. Since it was evident from these considerations that the determining factor is the asymmetry of the central carbon a KOOC\ ,.CH2C«00H b KOOCCHo i>K OK c it HOOCv . ^C— CHpCOOH HOOCCK^ '^ Fig. 6. Enzymatic dehydration of a labelled antipode of citric acid. atom, it followed that a chemical preparation of the asym- metrically labelled citric acid was entirely feasible. Therefore, its preparation was carried out as illustrated in Fig. 7 (Wilcox, Heidelberger, and Potter, 1950). By this means a compound closely related to citric acid and having a true asymmetric Mg' HCN ClCHpCOOEt -rrT-^ ClCHoCOCSoCOOEt T^ :* ?^ H^ ?^ ClCH^CCH^COOEt ^ » ClCH«CCHoCOOH Fig. 7. Synthesis of citric acid precursor. carbon atom was synthesized. This substance was resolved into optically active form with brucine, and was then con- verted into labelled citric acid bv means of the reactions shown in Fig. 8. Samples of this labelled citrate in various stages of pH * ^ OH ClCHgC-CHgCOOH ^*g^^^> HbOCCHgC-CHgCOOH COOH ' ^ COOH *• •'D / Fig. 8. Conversion of optically active precursor into labelled citric acid. 252 Charles Heidelberger and Van R. Potter optical purity of its precursor were subjected to enzymatic conversion into a-ketoglutarate as illustrated in Fig. 9. The success of this experiment adds further support to the stereo- chemical interpretation of this phenomenon. Recently Martins and Schorre (1950) have reported experiments, similar to these in scope and results, with the use of deuterated citric acids. 85^' r 1 2| 2 Chemically Prepared ' COOH 15^ Rat Kidney ,, Homogenate II. HOOCCOCH^CH^COOH KMnO, CO^ + HOOCCH^CH^COOH M 25 = D - 9-3 39^ 615^ (*^) 25 = D -25-9 20^ 80^ w §^ = -¥f 9 <2% >^8% Fig. 9. Distribution of label in citrate and a-ketoglutarate from chemically prepared asymmetric citric acid (Wilcox, Heidelberger and Potter, 1950). On the basis of the experiments with isotopic CO2 and oxaloacetate it would be predicted that citric acid derived from labelled acetate or pyruvate should be asymmetrically labelled, and that the ^*C should be located in the primary carboxyl group other than the one labelled by COg. Further- more, if any recycling should occur one would anticipate a balance between the radioactivity of the "a" and the tertiary carboxyl group. That such is the case was strongly Asymmetric Citric Acid 253 indicated by the results obtained previously by Buchanan, Sakami, Gurin and Wilson (1945) and Weinhouse, Medes, Floyd and Noda (1945) (Fig. 10) although they did not isolate the citric acid. We have repeated the experiments in our usual way with the use of [carboccy-^^Cyacetate and [carbonyl-'^^C]-py ruY ate (Potter and Heidelberger, 1951). The CI- I COCH^COOH HOOCCH2CO + (-CH^COX) COOH HOOCCH^COHCH^COOH COOH (Not isolated) (B) CH^COOH (W) Slices and Homogenates HOOCCOCH CH COOH COo + HOOCCH CH COOH 2 2 2 KMnOj^ (B) 10$^ 90^ (w) 17% 83^ Fig. 10. Distribution of label in a-ketoglutarate from labelled acetate and acetoacetate (Buchanan et al., 1945; Weinhouse et al., 1945). results indicated in Fig. 11 are somewhat surprising in that the citric acid was not completely asymmetric, while the tertiary carboxyl of the citric acid contained no isotope. Some additional evidence bearing on this result was obtained by studying the asymmetry of citric acid produced in vivo (Potter, Heidelberger and Busch, 1951). When sodium fluoroacetate is administered to rats there is an enormous increase in the citric acid content of many tissues, and when [carbOvCy-'^'^C]- acetate was administered to such an animal, enough labelled citric acid could be isolated from the kidneys of the rat to be degraded by the usual procedure. The very 254 Charles Heidelberger and Van R. Potter surprising result shown in Fig. 12 is that although the citric acid is indeed asymmetric, it is in the opposite direction from that which might have been predicted. One common finding in the citric acids derived from acetate both in vitro and in vivo, is that there is a much higher isotopic content in the "a" carboxyl than in the tertiary carboxyl CH^COCOOE ' i HOOCCH^CO + (-CH^50X) <— CH^SoOH 2i COOH EOOCCH^COHCEoCOOH 2| 2 COOH 0-5% HOOCCOCH^CH^COOH 2^^"^' ^^ ^^2 Rat Kidney Eomogenate I, 1. Rat Kidney Homogenate II. 2. Pigeon Liver Extract I KMnOij. CO^ + EOOCCH^CH^COOH 15-23^ 85-775^ Fig. 11. Distribution of label in citrate and a-ketoglutarate from isotopic acetate and pyruvate (Potter and Heidelberger, 1951, unpublished). carbon and suggests that there may be a common mechanism for the process. Several explanations may be advanced for the finding of an isotopic content 20 per cent in the "a" carboxyl of citrate in the acetate and pyruvate experiments. An artifact in the permanganate oxidation of the a-ketoglutarate dinitro- phenylhydrazone has been ruled out by the demonstration that under very drastic conditions labelled succinic acid is converted to CO 2 to the extent of only 3 per cent. Recycling Asymmetric Citric Acid 255 in the degradation by the homogenate has also been ehminated because identical results were obtained when aliquots of the same sample of citric acid were degraded by the homogenate and pigeon liver extract, in which cycling is not possible. Another plausible explanation, that of COg fixation, has also been ruled out by an experiment with [car&oa??/-^*C] -pyruvate, in which no ^^C could be detected in the citric acid. Two other possibilities are now under investigation. One is that some succinate can break through the malonate block, HOOCCH CO + (-CH COX) *- CH 5oOH COOH Intact Rat (Kidney) Fluoroacetate Poisoned HOOCCH^COHCH^COOH 2| 2 COOH 20^ EOOCCOCH2CE2COOH 1 1 Pigeon Liver Extract KMnOy^ CO^ + HOOCCH^CH^COOH 80% 20% Fig. 12. Distribution of label in citrate and c -ketoglutarate from labelled acetate in the intact rat (Potter, Heidelberger and Busch, 1951, unpublished). be preferentially oxidized and cajry through to oxaloacetate in an asymmetric fashion, possibly as phosphorylated inter- mediates. The other hypothesis currently under investigation is that oxaloacetate may exchange with acetate by the splitting off of oxalic acid or a derivative. The results of these experiments are not yet available. In conclusion, it may be seen that a study of the asymmetry of labelled citric acid has substantiated Ogston's proposal, 256 Charles Heidelberger and Van R. Potter has helped to remove the objections to the inclusion of citric acid in the Krebs cycle, has given further information on the stereochemistry of substrate-enzyme interaction, and may be used as a tool for the investigation of new phenomena in intermediary metabolism both in vitro and in vivo. REFERENCES Buchanan, J. M., Sakami, W., Gurin, S., and Wilson, D. W. (1945). J. biol. Chem.y 159, 695. Evans, E. A., Jun., and Slotin, L. (1941). J. biol. Chem., 141, 439. IsHERWOOD, F. A. (1946). Biochem. J., 32, 108. LoRBER, v.. Utter, M. F., Rudney, H., and Cook, M. (1950). J. biol. Chem., 185, 689. Martius, C, and Schorre, G. (1950). Liebig's Ann., 570, 140, 143. Ogston, a. G. (1948). Nature, London, 162, 963, Potter, V. R., and Heidelberger, C. (1949). Nature, London, 164, 180. Potter, V. R., and Heidelberger, C. (1951). Unpublished. Potter, V. R., Heidelberger, C, and Busch, H. (1951). Unpublished. Stern, J. R., and Ochoa, S. (1949). J. biol. Chem., 179, 491. Weinhouse, S., Medes, G., Floyd, N. F., and Noda, L. (1945). J. biol. Chem., 161, 745. Wilcox, P. E., Heidelberger, C, and Potter, V. R. (1950). J. Amer. Chem. Soc., 72, 5019. Wood, H. G., Werkman, C. H., Hemingway, A., and Nier, A. O. (1942). J. biol. Chem., 142, 31. DISCUSSION Arnstein: In the synthesis of citric acid (Fig. 7) some of the radio- activity was incorporated into the tertiary carboxyl group. Have you any idea of the mechanism of this reaction? Heidelberger: We added HCN to ethyl y-chloroacetoacetate and got this compound ^ OH CI-CH2-C-CH2COOH CN The cyanide group was then hydrolysed to a carboxyl group but we could not recrystallize the acids. There is some possibility that there may have been some nitrile left that was not completely hydrolysed under these conditions; if there were, this might have exchanged with the labelled cyanide we added in the next step, and given us the result that was obtained. This is a possible explanation, but fortunately it did not affect the results of the a-keto story. Asymmetric Citric Acid 257 Bentley: In the synthesis you reacted chloroacetoaeetie ester with cyanide. How do you get the preferential condensation on the ketone instead of reaction with the halogen? Heidelberger; It was an ordinary hydrocyanic acid addition which was catalysed by cyanide ion. The experimental details are published (1950, J. Amer. cheni. Soc, 72, 5019). Incidentally, the synthetic citric acid was characterized by a number of methods, including a chromatographic one, in which we took an equal amount of the [i*C]citric acid and of authentic non-labelled citric acid and chromatographed them on silica gel. Only one peak was found and again the titration paralleled the count. We feel that there is no doubt about the citric acid being authentic. I wish some- body would give us some better ideas about the peculiar result of finding i*C in the wrong end of the citric acid in our biosynthetic experiments using pyruvate and acetate. Krebs: Would it help in any way to assume that the reaction forming citric acid from acetate and oxaloacetate was reversible and that the reverse reaction did not involve a three-point contact? The reverse reaction has not been shown in animal tissues but there are indications that that type of reaction occurs in micro-organisms, citrate being split into oxaloacetate and acetate. Heidelberger: I think that if such an explanation were true it would help, but Professor Wood's laboratory has good evidence, I think, that the split only takes place to a small extent. Furthermore, I would hate to think of it taking place by a non-3-point contact. I think the same enzyme that splits the citric acid would have to make it. •ISOTOPES 1 8 MODE OF FORMATION OF FATTY ACIDS FROM ACETATE AND GLUCOSE AS STUDIED IN THE MAMMARY GLAND G. POPJAK Soon after the introduction of deuterium as a tracer element by Schoenheimer and Rittenberg (1935), it was discovered that when the body water of animals is enriched with DgO, the fatty acids, among other body constituents, become labelled with deuterium. Schoenheimer and Ritten- berg found that when the animals were given DgO to drink for many days the concentration of D in the fatty acids reached about 50 per cent of that in the body water. They adduced convincing proof that the incorporation of the isotope into fatty acids in stably bound form was due to the synthesis of the acids in the body, and with remarkable intuition they also inferred that in this synthesis small molecules must have been involved (see Schoenheimer, 1941). A few years later Bloch and Rittenberg (1942a, h) made the further important discovery that the small molecules from which fatty acids and cholesterol are synthesized are acetate units or some more reactive derivatives of acetate. The precise mode of the condensation of these Cg units to form the long-chain fatty acids and the nature of the intermediate compounds have, however, eluded discovery to a large extent. Recently I have been fortunate in having been associated with some work on the metabolism of the mammary gland, the results of which seem to offer at least a partial solution of this problem. The work to be described was done in collabo- ration with a number of colleagues. Dr. S. J. Folley and Dr. T. H. French of the National Institute for Research in Dairying, L^niversity of Reading, and Dr. G. D. Hunter and 258 Formation of Fatty Acids 259 Dr. A. J. P. Martin, F.R.S., of the National Institute for Medical Research. Most of the work to be described has been published in the Biochemical Journal. References to the individual papers will be found in the text and in the biblio- graphy at the end of this article. The experiments have shown that the mammary gland is singularly suited for the investigation of biosynthetic pro- cesses. It is not intended, however, to present the results from the particular point of view of the metabolism of the mammary gland, but to discuss their bearing on the more general problem of the biosynthesis of fatty acids. Neverthe- less, reference has to be made to certain salient features of mammary gland secretion reviewed recently by Folley (1949). Milk is a remarkable biological product in that it provides all the nutrients necessary for the growing young. In essence it is a store of energy. A lactating rabbit may provide in 3-4 weeks an amount of milk equivalent to her own body weight! We are convinced now that this remarkable output of stored energy in milk is the result primarily of the metabolic activities of the mammary gland cells. So far we have studied mainly the synthesis of milk fat which, of all animal fats, is the only one that contains short- chain fatty acids. The milk of ruminants is particularly rich in these acids, containing every even numbered member of the saturated series from butyric to stearic acid. The origin of milk fat and particularly that of the short-chain acids has been the subject of divergent opinions (for review see Folley, 1949). Graham, Jones and Kay (1936) concluded that the amount of glyceride fat which disappeared from the blood during passage through the udder of lafctating cows was sufficient to account for the production of milk fat. Voris, Ellis and Maynard (1940), using a more specific method for the estima- tion of blood glycerides, confirmed the uptake of blood fat by the udder of lactating cows. On the other hand, the finding of a high respiratory quotient for the active udder in vivo led to the suggestion that part of the milk fat might 260 G. PopjAK be synthesized in the gland from carbohydrate sources (Graham, Houchin, Peterson and Turner, 1938). To explain the occurrence of the short -chain acids several speculative views have been put forward. It was suggested for instance by Reineke, Stoneciphcr and Turner (1941) that if part of the milk fat is synthesized in the gland from carbohydrate, the short-chain acids might be the part so formed. The most fully developed theory, however, that of Hilditch, assigns no synthetic role to the mammary gland, but insists that the volatile acids are derived from blood glycerides, especially their oleic acid component (Hilditch, 1947) by an a>-type of oxidation (Achaya and Hilditch, 1950). In view of the central role of acetate in fat metabolism as revealed by the work of Bloch and Rittenberg (Bloch, 1947, 1948), Folley suggested in a chapter written in 1945 for the new edition of Marshall's Physiology of Reproduction (which unfortunately is still not published) that the short-chain acids in milk might also originate by synthesis from acetate. Folley and French (1948, 1950) have found that tissue slices made from the lactating mammae of ruminants utilized acetate (but not glucose) with a respiratory quotient (R.Q.) greater than one. On the other hand, slices from the glands of non-ruminants (rat, rabbit) showed a R.Q. greater than one only in the presence of glucose, or glucose +acetate, but not with acetate alone as substrate. They interpreted their results as indicating the preferential use of acetate rather than glucose for milk fat synthesis in ruminants. In the course of studies on foetal fat metabolism Miss Beeckmans and I investigated the fat in the mammary gland of non-lactating pregnant rabbits. We were able to show after giving DgO and ^^C-labelled acetate to the animals that active fat synthesis occurred in the gland (Popjak and Beeckmans, 1949, 1950). There was a particularly high incorporation of ^^C from acetate into the glyceride fatty acids and it was subsequently shown that this was largely due to the uptake of acetate into the short chain volatile fatty acids (Popjak, Folley and French, 1949; Popjak and Formation of Fatty Acids 261 Beeckmans, 1950). The data of Table I show that the short- chain acids could not have originated from the degradation of oleic acid as postulated by Hilditch, but must have been synthesized from acetate. Table I "C Content (1 x 10-^ /xc./mg. C) of Glyceride Fatty Acids in Mammary Glands of Non-lactating 28 Days-pregnant Rabbits Injected Intra- venously with CHg^^COONa Volatile fattv acids: — Water-soluble (Ci-Cg) Water-insoluble (Cg-Cij) Non-volatile fattv acids: — Solid . . . . . Liquid Exp. 1* Exp, 2* 38 1 184-5 52-6 136 3-4 7-2 5-9 10-2 *In Exp. 1 two equal injections of the labelled acetate were given daily for 4 days; the total **C dose was 50/uc. (4-3 nig. of acetate). In Exp. 2, two injections, 50 ^c. each, were given 6 hours apart and the animal was killed 20 hours after the first injection. In these experiments on pregnant rabbits, the fat extracted from the mammae contained volatile acids and in this respect was comparable to milk fat, but the extent of its admixture with tissue fat could not be ascertained. In order to continue these observations it was desirable to obtain adequate quan- tities of true milk fat for detailed study over a period which also allowed the time course of the synthetic process to be followed. For this purpose a lactating goat of 117 lb. body weight was injected intravenously with 5 mc. of CHg^^COONa (430 mg.) and was then milked at frequent intervals up to 48 hours after the injection (Popjak, French and Folley, 1950, 1951). It is of particular interest to mention here (the goat being a ruminant) that it has b.een shown by Barcroft and his collaborators that ruminants assimilate their carbo- hydrate food as acetic acid, by fermentation of cellulose and other polysaccharides in the rumen (see review by Elsden and Phillipson, 1948). Thus acetate is of particular impor- tance in ruminant metabolism. 262 G. POPJAK In order to obtain evidence on the question whether or not milk fatty acids are derived primarily from the plasma lipids or are synthesized in significant amounts in the udder, the i*C contents of plasma fatty acids were compared with those of four crude fatty acid fractions of the milk glycerides. The results are shown in Figs. 1 and 2. It should be pointed out that the scale of the ordinate in Fig. 1 is one hundred u (J I o X >^ *> u rJ to 0-5 0-4 0-3 ^.■^--^ •-£i 10 20 30 40 Time after injection (hr.) Fig. 1. Change of specific activities of plasma fatty acids and of plasma and milk cholesterol after the injection of 5 mc. CHgi^COONa into a lactating goat. Plasma total fatty acid. © Plasma phospholipid fatty acids. Q Plasma non-phospholipid fatty acids. • Milk cholesterol. A Plasma cholesterol. (Popjak, Folley and French, 1951.) times larger than in Fig. 2. The comparison shows that the milk fatty acids behaved quite differently from the plasma fatty acids. The specific activity of the latter reached a pla- teau 24 hr. after the injection but continued to rise slightly up to the end of the experiment, whereas the milk fatty acids reached their maximum ^^C content 3-4 hr. after the injection, when they contained several hundred times more isotope than the plasma fatty acids did at any time. Formation of Fatty Acids 263 This gross comparison between plasma and milk fatty acids indicates that synthesis of fatty acids in the udder from small molecules (e.g. acetate or some other C2 unit) is of primary importance in the origin of milk fat. It has been possible 40 r 10 20 30 Time after injection (hr.) Fig. 2. Specific activity: time curves of 4 crude fatty acid fractions obtained from milk glycerides of a goat after tlie injec- tion of 5 mc. CHa^COONa. Steam volatile water-soluble fatty acids. ^ Steam volatile water-insoluble fatty acids. • Non-volatile solid acids. A Non-volatile liquid acids. (Popjak, Folley and French, 1950, 1951.) to calculate the amount of acetate used for milk fat synthesis during the first six hours of the experiment. Eighty per cent of the injected acetate appeared as respiratory COg, and 10 per cent of the total dose, or 50 per cent of the acetate retained in the body, was converted into milk fatty acids. If we 264 G. PopjAK assume that the tracer acetate was mixed completely with the body acetate, the calculation implies that in the lactating goat 50 per cent of the body acetate which is not oxidized is converted into milk fat in six hours. Another important point emerged from the relationship between the specific activities of the short-chain volatile acids (water-soluble and water-insoluble steam volatile acids) and those of the non-volatile (solid and liquid) acids. At the time of their maximum isotope content the former con- tained several times more ^C than the latter, and therefore the short-chain acids must have originated by synthesis from acetate and not by the degradation of oleogiycerides, con- firming our findings on non-lactating pregnant rabbits (Popjak, Folley and French, 1949; Popjak and Beeckmans, 1950). In casting for an explanation for the differences in the iso- tope contents of the milk fatty acid fractions we were im- pressed by two points: (1) although the fractions were not homogeneous, their specific activities decayed with almost identical half- lives of about four hours and (2) that the specific activities did not bear a simple relationship to the average chain length of the fractions. The steam-volatile fraction insoluble in water, and which had the highest ^^C content, consists mainly of Cg to C12 acids and the water- soluble fraction of C4 to Cg acids. It seemed to us probable therefore that the differences were connected in some way with the biochemical mechanisms of fatty acid synthesis rather than with any other factor, e.g. differences in turn-over rates or dilutions by blood fat of low isotope content. In order to interpret more fully the role of acetate in milk fat synthesis, we have resolved the crude fractions into individual acids (Popjak, French, Hunter and Martin, 1951). Since it would have been an enormous task to carry out this type of work on all the 22 milk samples separately, we decided to pool the material from various parts of the experiment. If the specific activities of the four crude fatty acid fractions had maintained the same order relative to one another throughout the experiment, it would have been justifiable to Formation of Fatty Acids 265 pool all the samples for the isolation of the individual acids. However, as can be seen from Fig. 2, the relative position of the specific-activity time curves changed at 12 hr. after the injection. For this reason it was decided to divide the acids, into two main groups: (a) those obtained up to 12 hr. and (b) those obtained between 12 and 48 hr. of the experiment. Thus, during the first period the fatty acids had their highest specific activities, and during the second period the specific activities of all the fractions were already declining. In addition to these two main parts of the experiment some observations were also made on two small samples of the water-soluble volatile acids, one representing 0-7 hr. and the other approximately the entire experimental period. Fig. 3 shows graphically the specific activities of the individual fatty acids which, with the exception of acetic acid, were chromatographically pure (for technical details of isolation of the individual acids see Popjak, French, Hunter and Martin, 1951). The acetic acid was contaminated with butyric acid to the extent of about 10 per cent. But as only small quantities (5--10 mg.) of this acid were obtained, no further purification was attempted. It is seen that there is a relationship between isotope content and chain length of the fatty acids. During the first 12 hr. of the experiment the specific activities increased from butyric acid up to and including capric (decanoic) acid, whereas during the second part of the experiment the specific activities increased up to myristic (tetradecanoic) and palmitic (hexadecanoic) acids. The finding of acetic acid in milk fat requires some comment. Its presence was quite unexpected, but we have demonstrated it quite consistently among the water-soluble steam-volatile fatty acids. We are uncertaih whether it is present as glyceride or is esterified with some other substance. We have now found it also in the milk fat of rabbits and there can be no doubt as to its genuine occurrence. The meaning of the results shown in Fig. 3 became apparent from the information obtained from the chemical degradation of acetic, butyric and caproic acids shown in Table II. 266 G. POPJAK SPECIFIC ACTIVITY OF MILK FATTY ACIDS AFTER INJECTION OF 5 MC OF CH^'^CO^Na O- 12 HR. 2-48 HR. 8 lO 12 t4 16 18 18 LENGTH OF 4 6 8 lO 12 14 16 IB 18 CARBON CHAIN Fig. 3. Specific activities of pure fatty acids obtained from milk glycerides of a goat after the injection of 5 mc. CHg^^COONa. (Popjak, French, Hunter and Martin, 1951.) Table II Specific Activity (1 x 10"^ /xc. ^^C/mg. C) of Individual Carbon Atoms IN Acetic, Butyric, and Caproic Acids from Milk Glycerides after Intravenous Injection of 5 mc. of CHg^^COgNa Period of experiment ihr.) Fatty acid C atom numbers Average of all C atoms 6 5 4 3 2 1 {COOH) 0- 7 0-48 12-48 Butvric Caproic Acetic Butyric Caproic Butyric Caproic 000 0-00 00 36-20 13-20 0-556 0-00 000 0-00 0-00 000 0-00 36-00 36-20 15-26 13-20 0-700 0-556 00 0-00 000 00 0-00 0-00 000 36-00 96-70 23-60 15-26 32-60 0-700 1-504 18 00 28-75 11-80 7-63 9-86 0-350 0-424 i Formation of Fatty Acids 267 Whatever the origin of the acetic acid its occurrence never- theless served as a useful guide showing that redistribution of the label from the injected acetate did not occur in the animal since ^*C was found only in the carboxyl carbon. The degradation of butyric acid revealed the presence of equal amounts of ^*C in carbon atoms 1 (COOH) and 3, carbons 2 and 4 being totally inactive. This finding indicates that butyric acid has been synthesized in the udder by con- densation of two acetate units of equal isotope content, the methyl carbon of one having been linked with the carboxyl carbon of the other. However, we do not consider acetate to be the only source of butyric acid in the milk of the goat, but it is thought that another precursor, probably a C4 compound, is also involved. The necessity for this assumption should be clear from the information obtained by the degra- dation of caproic acid. The results of the degradation of caproic acid from the early part of the experiment (cf. Table II) should be considered first. Only carbon atoms 1 (COOH), 3 and 5 contained ^^C. The specific activity of the COOH carbon was 2 • 5 times that of carbons 3 and 5, whose specific activities were equal, just as in carbons 1 and 3 of butyric acid. Further, the specific activities of carbons 3-6 of caproic