tively recent reviews3 and the increasing detail of

Transcription

1 THE BIOSYNTHESIS AND INTERCONVERSION OF PURINES AND THEIR DERIVATIVES' ALBERT G. MOAT AND HERMAN FRIEDMAN2 Department of Microbiology, Hahnemann Medical College, Philadelphia, Pennsylvania I. Introduction II. The Precursors of the Purine Molecule A. Studies with Avian Systems B. Studies with Mammalian Systems C. Studies with Microbial Systems D. Inconsistencies in Precursor Studies III. Intermediary Reactions in Purine Biosynthesis A. The Pigeon Liver Pathway B. The Microbial Pathway to Purines Aminoimidazolecarboxamide and the role of folic acid Formamidoimidazolecarboxamide Sources of the one-carbon unit for positions 2 and Aminoimidazolecarboxamidine Formylglycinamide ribotide Aminoimidazole ribonucleotide and the role of biotin Mechanism of aminoimidazolecarboxamide ribonucleotide and adenosine monophosphate synthesis Synthesis of guanosine 5'-phosphate Evidence for other intermediates C. Metabolic Interrelationships and Feedback Mechanisms Purine-histidine interrelationships Feedback control mechanisms D. Purine Biosynthesis in Mammalian and Cancer Systems IV. Interconversion of Purines A. Microbial Interconversions Specific enzymatic mechanisms Other evidence for interconversion mechanisms B. Interconversions in Mammalian, Avian, and Cancer Tissues Specific enzymatic mechanisms Metabolism of unnatural purine derivatives Other evidence for interconversion mechanisms Reutilization of host purines by tumors V. General Discussion and Summary VI. Acknowledgments VII. References I. INTRODUCTION Our present knowledge of purine biosynthesis represents a compendium of information obtained from experimental observations on avian, mammalian, neoplastic, and microbial systems. Major advances have been made in this field within the past few years as indicated by the number of rela- 1 Supported in part by grants from the National Science Foundation (G-6398) and the Pender Fund. 2 Present address: Department of Microbiology, Albert Einstein Medical Center, Northern Division, Philadelphia, Pennsylvania. tively recent reviews3 and the increasing detail of information which has appeared therein with each succeeding year. Most of the previous reviewers have chosen to weave the information from various sources into a single composite pic- 3 The following reviews have been used in referring to many of the original contributions to this subject: 25, 26, 28, 29, 33, 38, 66, 76, 77, 83, 86, 88, 128, 150, 165, 172, 184, 187, 220. It is regrettable that the volume of published work has precluded direct reference to the many authors whose works have bearing on the biosynthesis and interconversion of purines. 309

2 310 MOAT AND FRIEDMAN [VOL. 24 ture for all biological systems. Since present indications are that there is, indeed, a remarkable similarity in the mechanism of purine biosynthesis and interconversion by rather widely divergent biological forms, presentation of the subject in this manner is, perhaps, well justified. However, this could lead to a false sense of security regarding the degree to which any given organism may conform to this general pattern. This is especially true of microorganisms. For example, if we compare purine formation with other metabolic pathways, it would seem that microorganisms should exhibit a comparable diversity of reactions in this system. Although there are indications of alternate reactions in some organisms, the presence of a number of the intermediary steps in the synthesis of purines by microorganisms as well as in many normal and neoplastic tissues of other biological forms has been inferred on the basis of over-all similarity to the rather complete description of the mechanism of purine synthesis in pigeon liver systems (28, 29, 33, 66, 76, 77, 82, 83, 165). It is our purpose, therefore, to compare the extent of our knowledge of the intermediary reactions in purine biosynthesis and interconversion by microbial species with what is known of these activities in avian, mammalian, and neoplastic systems. II. THE PRECURSORS OF THE PURINE IIOLECULE A. Studies with Avian Systems Early studies on precursors to the purines employed birds because of the high percentage of nitrogen excreted in the form of uric acid (greater than 80 per cent). Feeding of potential precursors to birds contributed important preliminary information (see reviews by Rose (172) and Christman (38)). Many of these experiments were difficult to interpret, however, because such a great variety of metabolites were found to be capable of increasing the amount of uric acid excreted (e.g. arginine, histidine, glycine, alanine, glutamic acid, aspartic acid, asparagine, and ammonia). Studies on the incorporation of compounds labeled with C'4 or N" into nucleic acids and subsequent analysis of the nucleic acid purines to determine the fate of these compounds substantiated earlier evidence that birds and mammals were capable of synthesizing purines from simple precursors. Pigeons incorporated N"5-labeled ammonia into the nucleic acids of various organs and also into the excreted uric acid and allantoin. Buchanan et al. (28, 29) showed that glycine was incorporated into the 4, 5, and 7 positions, formate into the 2 and 8 positions, and carbon dioxide into position 6 of the purine ring. The demonstration' that pigeon liver slices could synthesize hypoxanthine in the presence of glutamine or oxalacetate and ammoniumn salts paved the way for later work with pigeon liver homogenates. Greenberg (66, 76, 77) indicated that inosinic acid (inosine 5'-phosphate) was the first form of the completed purine ring to be synthesized. He presented evidence that the precursors were incorporated into hypoxanthine via the following general mechanism:' CO2 + 3 NH, + glycine + "formate" + R-1-P -H3P04-3P4 riboside intermediates Riboside intermediates + "formate" + H3PO4 inosine-5'-p Inosine-5'-P -+ inosine + H3PO4 Inosine + H3PO4 -* hypoxanthine + R-1-P 5-Amino-4-imidazolecarboxamide did not act as an intermediate in the system, a fact which delayed the confirmation of the participation of this compound for some time. Nevertheless, these studies provided the experimental system which later proved useful in identification of the inter- 4The following abbreviations are used: R-1-P, ribose 1-phosphate; R-5-P, ribose 5-phosphate; PRPP, 5-phosphoribosylpyrophosphate; PRA, 5-phosphoribosylamine; GAR, glycinamide ribonucleotide; FGAR, formyl-glycinamide ribonucleotide; FGAM, formyl-glycinamidine ribonucleotide; AIR, 5-aminoimidazole ribonueleotide; C-AIR, 5-amino-4-carboxyimidazole ribonucleotide; SAICAR, 5-amino-4-imidazole- (N-suceinylo)-carboxamide ribonucleotide; AICA or AICAR, 5-amino-4-imidazolecarboxamide or its ribonucleotide; FAICAR, 5-formamidoimidazole-4-carboxamide ribonucleotide; IMP or I-5'-P, inosine 5'-phosphate; AMP-S, adenylosuccinate; AMP or 5'-AMP, adenosine 5'-phosphate; XMP, xanthosine 5'-phosphate; GMP, guanosine 5'-phosphate; IGP, imidazoleglycerol phosphate; FADG, a-l-formamidinoglutarate; THF, N5-formyl-tetrahydrofolic acid; THF-CHO, N5,AIO-anhydroformyltetrahydrofolic acid; 2,6-DAP, 2,6-diaminopurine; 6-MP, 6-mercaptopurine; RNA, ribonucleic acid; 6-TG, 6-thioguanine; DON, 6-diazo-5- oxo-l-norleucine; PNA, pentose nucleic acid; DNA, deoxyribonucleic acid; Pi, inorganic phosphate; P-P, inorganic pyrophosphate; ATP, adenosine triphosphate; GTP, guanosine triphosphate; PAB, p-aminobenzoic acid.

3 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES 311 ASPARTATE - HRe-0QO-COo FORMATE - OC' 3 HN Cs-;NH -GLYCINE 1 O-FORMATE NH AMIDE-N OF GLUTAMINE Figure 1. Precursors of the purine molecule. In avian systems, the contribution of the amino nitrogen of aspartate to N1 and the amide nitrogen of glutamine to N3 and Ng has been verified with N'5-labeled compounds. Comparable isotopelabeling studies have not been performed with microorganisms. However, other evidence indicates that these same donors provide the nitrogen for positions 1, 3, and 9 in the purines of microbial svstems. mediary steps involved. During this same period it was confirmed that the methylene carbon of glycine was incorporated into position 5 in the purine ring and that formate was incorporated almost exclusively into positions 2 and 8 (see references, footnote 3). Continued investigations in several laboratories (26, 28, 33, 77, 184) succeeded in determining the immediate precursors to the purine molecule (figure 1). In particular, the question as to the immediate precursors of the nitrogen atoms of purine was clarified (28). Aspartic acid contributed the nitrogen of position 1 and glutamine (from its amide group) the nitrogens of positions 3 and 9. It is interesting that folic acid deficiency failed to depress the incorporation of formate into purines of the chick whereas glycine incorporation was depressed under these same conditions (200, 201). Folic acid-deficient rats exhibited a decreased ability to incorporate formate into purines (45). Such variations indicate that the role of a vitamin cofactor may not be evident under all experimental conditions, even when an obvious deficiency state has been achieved. It may also be that formate is not the sole precursor of the C2 and C8 positions, as discussed in a later section on sources of the Cl unit. B. Studies with Mammalian Systems Concurrently with the elucidation of the precursors of purines in avian systems, a variety of mammalian systems, including man, were also employed (in many instances by the same investigators). N'5-labeled ammonia was incorporated into the purines and pyrimidines of the internal organs as well as the excretory purines and pyrimidines by the rat. Arginine, previously held to be a precursor of purines, was eliminated as a direct precursor. Histidine nitrogen did not serve as a direct precursor of purines. The nitrogen in position 7 of purines was shown to arise specifically from glycine in human males. Rats incorporated carbon dioxide into C6, the a- and f3- carbon atoms of glycine into C4 and C5, and formate carbon into C2 and C8 of purines. The presence of exogenous purines suppressed the incorporation of precursors (see references, footnote 3). Although most of the investigations in mammalian systems were not as detailed as those made with pigeons or pigeon liver systems, the bulk of evidence suggested a similar, if not identical, list of precursors for the synthesis of purines by mammals. Drysdale et al. (45) were able to exhibit a decreased ability to incorporate C'4-labeled formate into purines on the part of folic acid-deficient rats. Skipper et al. (190) obtained similar results with mice treated with folic acid analogues. The utilization of formate and carbon dioxide for nucleic acid synthesis was impaired, with the effect on formate utilization being more marked. The effect of biotin deficiency on the incorporation of carbon dioxide into nucleic acid purines reported by Lardy et al. (106, 124) is of interest in view of the limited number of observations regarding the role of biotin in mammalian metabolism (114). Biotindeficient rats incorporated much lower quantities of C14 from the radioactive carbon dioxide administered into adenine and guanine than did normal rats. As compared with experimental observations on the precursors in avian and mammalian systems, the role of vitamins in purine biosynthesis has been investigated more extensively by means of avian and microbial systems. The effect of these factors on the synthesis of purines by microorganisms and studies in vitro with avian and mammalian systems will be discussed in a subsequent section. C. Studies with Microbial Systems Abrams et al. (7) found that Torulopsis utilis utilized glycine-n15 equally for the synthesis of adenine and guanine. N15-Labeled ammonia and glycine, but not histidine or lysine, were found by

4 3MOAT AND FRIEDMXIAN 312 [VO L. 24 Di Carlo et al. (44) to be utilized for purine and pyrimidine synthesis by T. utilis and Saccharomiyces cerevisiae, Hansen strain. Using H2N'5CH..- C1"OOH, Sutton et al. (194) concluded that Aero bacter aerogenes incorporated glycine in toto into both adenine and guanine as indicated by equal labeling of the C4 and N7 positions. Roberts et al. (171) and Koch et al. (97) found that labeled carbon dioxide was incorporated into the C6 of all purines of Escherichia coli and that added purines suppressed this incorporation. Edmonds et al. (46) investigated the formation of nucleic acid purines and pyrimidines of growing yeast with the aid of C'4-labeled formate and glycine-2- C'4. Formate was incorporated into positions 2 and 8 of both guanine and adenine whereas the methylene carbon of glycine was incorporated into position 5. Kerr and Chernigoy (94) found that T. utilis incorporated C14-labeled formate equally into adenine and guanine and that the distribution was approximately the same in RNA as in DNA. Adenine repressed the incorporation of formate into either adenine or guanine. Guanine or 2,6-diaminopurine repressed formate incorporation into guanine while enhancing formate utilization for the synthesis of adenine. Roberts et al. (171), in their extensive investigations on the metabolism of E. coli, summarized their data regarding purine synthesis from labeled precursors as follows: Purine C atom Source C6 Carbon dioxide C4, Cs Glycine, serine, or threonine C2, C8 "C," residue arising from formate, formaldehyde, C, of glucose, glycine, serine, and threonine They concluded from their findings that the pathway of purine synthesis in E. coli was consistent with that which had already been demonstrated in pigeon liver. It is evident that the investigations with use of isotopically labeled compounds to determine the precursors of purines in microorganisms have not been as extensive as those conducted on avian or mammalian systems. There has been a lack of sufficient evidence for the immediate precursors of the nitrogen atoms of positions 1, 3, and 9 of the purine molecule. Direct isotopic labeling experiments of the type used to indicate the role of aspartate and glutamine as contributors of the nitrogen atoms in the avian pathway have not been performed with microorganisms. Other types of evidence (see section III) leave little doubt, however, that the nitrogen of position 1 arises from aspartate and that the nitrogens of positions 3 and 9 arise from the amide nitrogen of glutamine, in microbial systems as well as in avian and mammalian tissues. D. Inconsistencies in Precursor Studies Not all of the experiments relating to purine precursors in birds, mammals, and microorganisms were as clear-cut as might be envisioned from the evidence just presented. In the original attempts to find the immediate precursors of the purine molecule, contributions from a variety of precursors, usually at low levels, were dismissed as representing metabolic conversion of some portion of these compounds to the immediate precursors. Under certain conditions, considerable breakdown of endogenous nucleic acids may occur, miaking it difficult to interpret an increased concentration of purines as an indication of synthesis of new purine (170). However, it may be cons;idered that the application of isotopic methodology proved to be the key which led to the determination of the precursors of the purine molecule. Incorporation of labeled compounds as a measure of rea(- tion in the direction of synthesis macde it possible to delineate the direct precursors of the purine molecule and provided a firm basis for subsequent inquiries as to the intermediary steps involved. Most of the evidence derived from studies on the incorporation of labeled compounds indicated that there was a general similarity of precursors and that birds, mammals, and microorganisms possessed pathways for the biosynthesis of purines that were not markedly different. However, some evidence is available which suggests that metabolic interrelationships exist in a variety of species and that these may affect purine synthesis quite markedly. These divergent pathways, which may alter purine synthesis quite differently in one organism as compared to another, may not truly represent alternate routes of purine synthesis. However, their importance is obvious, and they will be discussed in a subsequent section after presentation of the direct, pathway to purines in various forms. III. INTERMEDIARY REACTIONS IN PURINE BIOSYNTHESIS A. The Pigeon Liver Pathwvay Using pigeon liver homogenates, Greenberg (64, 66, 76, 77) demonstrated the synthesis of

5 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES TABLE 1 Intermediary reactions in purine biosynthesis 313 Intermediates Presence of Intermediate or Reaction Birds Mammals Neoplasms Microbes HO // C OH HO C CH2OPOsH2 66, 76, 77 H H Ribose 5-phosphate OH OH Mg l +ATP =P-O-P-O / \ CH20PO3H2 1H OH C OH HO\C H7 b b H + AMP H /N 5-Phosphoribosylpyrophosphate (PRPP) Glutamine + Mg Amidotransferase /0 CH20PO3H2 C OH HO C H7 b b H + P-P H H + Glutamate 5-Phosphoribosylamine (PRA) Glycine + ATP + Mg- HtC-NH2 O=^-NH Ribose-P Kinosynthase + ADP + Pi Glycinamide ribotide (GAR) + THF_4~---* Transformylase H2C-NH CHO + THF O= -NH ILibose-P 28, 29, 99, , , 30, 63, 64, , 81, , 64, 66 28, 29, 38, 66, 82 28, 29, , 77, 91, , 79 82, 84, 208, , 109, 110, 144, 177, , 109, 110, , , , 36, 37, 73, 141, 147 7, 44, 46, 171, 187, , , 94, 97, 171, , 197, 198 a-n-formylglycinamide ribotide (FGAR) H2C-NH CHO + Mg 4 + ATP + Glutamine + Glutamate + ADP + Pi HNd -NH Libose-P a-n-formylglycinamide ribotide (FGAM) 81, 82, 84, 112, 17, 113, 119, , 109, 110, , 73, 197, 198

6 314 MOAT AND FRIEDMAN [VOL. 24 TABLE 1-Continued Intermediates Presence of Intermediate or Reaction Birds Mammals lneoplasms Microbes a-n-formylglycinamide ribotide (FGAM) HC-N + Mg + ATP 111, , 37, 57, 117, 118, 141, 160, 224 H2N- -N Ribose-P + ADP + Pi 111, Aminoimidazole ribotide (AIR) + HCOS- + Mg++-, (+ Biotin?) COOH CH H2N -,/f Ribose-P 5-Amino-4-carboxyimidazole ribotide (C-AIR) + ATP + Mg++ + Aspartate HOOC-GCH2 HOOC-A--N-O=O H H _:N O\CH + ADP + Pi H2N-C-N Libose-P , 29, 53, , , 106, 124, ,114, 142, 143, 167, 205, , 37, 57, 117, 142, 143, 205, , 74, 137 Downloaded from 5-Amino-4-imidazole- (N-succinylo-)-carboxamide ribotide (SAICAR) H2N-C=O b_n H2N O\ H l -N-R-P Adenylosuccinase + Fumarate 5-Aminoimidazole-4-carboxamide ribotide (AICAR) H + K+ + THF-C=O-* Transformylase 120, , 66, 76, 77, 138, 139, , 14-16, 60, 69-72, , 94, 103, 137, 147, 159, 171, 187, 188, 191, 193, 196, 205, 206, 210, 211, 220, 221, 224 on June 8, 2018 by guest H2N-C=O O=CH 6-N O\CH + THF H2N-O-N-R-P 5-Formamidoimidazole-4-carboxamideribotide (FAICAR) 66, 76, 77, , 46, 53, 65, 76, 77, 138, 139, 190, 196, 199, , 139, 196, 217

7 TABLE 1-Concluded Intermediates Presence of Intermediate or Reaction Birds Mammals Neoplasms Microbes 5-Formamidoimidazole-4-carboxamide ribotide (FAICAR) 28, 30, 88, 208, 14-16, 88 Inosinicase 209 N==C-OH Hb d-n CH / + H20 28, 208 NC-N-R-P Inosine 5'-phosphate (IMP) + GTP + Aspartate + Synthase 64, 66, 76, 77, , 57, 117, 143, 154, 159, mg H N=C-N--CH-COOH HG I 1 C-N H2COOH CH + GDP + Pi , 35,115 N-C-N-R-P Adenylosuccinate (AMP-S) Adenylosuccinae 92 13, 33-35, 61, 68, 74, 115, 154, 212 N=C-NH2 C-N + Fumarate 4, 5 115,154,159, 224 ~CH NCN-R-P Adenosine 5'-phosphate (AMP) N=C-OH C6k'-N ~CH N-N-R-P Inosine 5'-phosphate (IMP) + DPN+ + K+ j Dehydrogenase , 130, 148 N=C-OH HO- C-N CII 101, , 128, 130, 148 N--C-N-R-P Xanthosine 5'-phosphate (XMP) + ATP + HO + Gluta- Amidotransferase 101, , 130, 148 mine A N=C-OH H2N-C C-N CH + Glutamate + AMP + PP N-C-N-R-P Guanosine 5'-phosphate (GMP) * The true donor here is NiNio-anhydroformyl tetrahydrofolic acid. With this donor, ATP is not required. ATP is required to convert Ns-formyltetrahydrofolic acid to the cyclic form. 315

8 316 MOAT AND FRIEDMAN [VOL. 24 purines from elemental precursors. He concluded that inosine 5'-phosphate was the first complete purine molecule to be synthesized and was the precursor of other purine derivatives. As previously mentioned, he was unable to implicate AICA as an intermediate in this system; but others eventually found that CG4-labeled AICA could be incorporated into the nucleic acid adenine and guanine of rats, into the excretory uric acid of pigeons, and into hypoxanthine in pigeon liver homogenates or extracts (28, 29, ). Following these important discoveries, there ensued a concerted effort on the part of Greenberg, Buchanan, their colleagues, and a number of other investigators, to elucidate the intermediates and the enzymes involved in the pathway of biosynthesis in the pigeon liver system. The cumulative results of these efforts are depicted in table 1. Since the chronological development of this pathway has been quite adequately summarized in several recent reviews,3 no attempt will be made to recreate the details of these experiments here. Suffice it to say that this appears to be the major route of synthesis of purines in avian liver. From a comparison of the number of intermediates or enzymes mediating their utilization which have been demonstrated in microbial species or mammalian or cancer tissues, it can be seen that our knowledge of the pathway in these other systems is somewhat limited. It should be emphasized, however, that many of the studies conducted in other systems were instrumental in aiding the clarification of some of the key reactions in the avian pathway. In fact, early work with microorganisms provided the first indication of possible intermediates. This, perhaps more than any other factor, is responsible for the assumption that the intermediary pathway of purine synthesis by microorganisms is identical, or very nearly identical, with that in avian metabolism, despite the lack of detailed evidence for the presence of several important steps in microbial systems. B. The Microbial Pathway to Purines 1. Aminoimidazolecarboxamide and the role of folic acid. Very early in the work on the function of folic acid in metabolic reactions it was recognized that this vitamin was in some way related to the synthesis of purines in microorganisms. Particularly with lactobacilli and streptococci, purines or their derivatives reversed the inhibition of growth by folic acid analogues or exhibited a sparing action for folic acid or PAB in their nutrition (76, 77, 187, 220). However, the first indication regarding possible intermediates in the pathway of purine biosynthesis in living systems arose from the studies of Fox (56) and Stetten and Fox (193). Under the conditions of sulfonamide bacteriostasis it was shown that a diazotizable substance accumulated in cultures of E. coli. Shive et al. (187) correctly identified the accumulated compound as 5-amino-4-imidazolecarboxamide (AICA). It was also found that glycine increased the accumulation of AICA in sulfonamide-inhibited cultures of E. coli, indicating glycine as a precursor of this intermediate. The accumulation of AICA in E. coli cultures inhibited by 4-aminopteroylglutamic acid (221) or by PAB-requiring mutants (72) related folic acid directly to the further utilization of AICA. The significance of AICA as an intermediate in purine biosynthesis was not immediately accepted, as many organisms could not utilize it in place of purines for growth. For example, Pomper (159) found that an adenine- or hypoxanthine-requiring strain of S. cerevisiae could not utilize AICA for growth. Fries (60) reported that a purinerequiring strain of Ophiostoma utilized AICA for growth only if small quantities of purine were added to "trigger" AICA utilization. Skeggs et al. (188) found that Lactobacillus bifidus could not utilize AICA in place of adenine or hypoxanthine. However, Lactobacillus arabinosus (210) or purine-requiring E. coli mutants (69) utilized AICA or its ribonucleoside for purine synthesis and aminopterin inhibited its utilization. Shive and his colleagues (187) developed further relationships of folic acid to the formation of purines and other metabolic reactions in microorganisms. During the same period Greenberg and his colleagues (66, 76, 77) established specific requirements for folic acid in the pigeon liver system. 2. Formamidoimidazolecarboxamide. Ben-Ishai et al. (15, 16) suggested the participation of the formyl derivative of AICA (5-formamido-4- imidazolecarboxamide as an intermediate in the conversion of AICA to hypoxanthine. Methionine, in the presence of PAB, decreased the amount of AICA accumulating in sulfonamideinhibited cultures of E. coli, implicating methionine as a possible donor of carbon 2 of the purine ring. In avian systems, formyl-aicar is converted to inosinic acid (209), although the evidence for the two separate steps shown in table 1

9 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES 317 is by no means complete. The formation of the formyl derivative from AICA has not been directly demonstrated in any system (88). The participation of folic acid in the conversion of AICAR to purines in both microbial and avian systems has been well demonstrated, making a formyl derivative the most likely intermediate in this conversion. It is likely that the formyl derivative of AICAR represents an intermediate which is stable only in the presence of the enzyme (inosinic acid transformylase) and cyclizes to form IMP upon release. 3. Sources of the one-carbon unit for positions 2 and 8. The question of the participation of methionine in purine synthesis remains to be fully clarified. Whereas formyltetrahydrofolic acid represents the obvious contributor of the Cl unit, there have been a number of implications that methionine may contribute Ci units, particularly in microorganisms. Few, if any, direct references as to the contribution of the methyl carbon of methionine are available. However, several indirect implications may be found. In E. coli and other organisms possessing an active formic hydrogenlyase, formate is rapidly converted to carbon dioxide and a considerable portion of the total formate may be recovered in the C6 of purines (97, 171). Thus, C, donors such as methionine or serine, which may contribute to the formation of formyltetrahydrofolic acid without going through a free formate stage, would represent more efficient Cl donors than formate itself. In systems which do not readily degrade formate, conversion of free formate to formyltetrahydrofolic acid represents the most direct source of C, units for purine formation. Gots (69) reported the accumulation of AICA by a purine-requiring mutant of E. coli when grown at limiting concentrations of purines. Gots (70), Greenberg and Spilman (78), and Ben-Ishai et al. (14) independently provided evidence that the riboside of aminoimidazolecarboxamide was produced by sulfonamide-inhibited cultures of E. coli or purine-requiring E. coli mutants, supporting the concept that ribosidation occurred prior to ring closure. This coincides with the consistent findings that purine synthesis in avian liver proceeds via the ribotide derivatives, as first demonstrated by Greenberg (66, 76, 77, 78). Gots and Chu (72) found that a PAB-requiring mutant of E. coli accumulated AICA under the conditions of PAB deficiency or in the presence of competitive inhibitors of PAB. Methionine was found to depress the accumulation of AICA, again suggesting that methionine may serve as the donor of the C2 carbon. In more detailed investigations, Gots and Love (75) established that serine, glutamic acid, and aspartic acid serve as nitrogen donors for the formation of AICA. Histidine, threonine, glycine, and aspartic acid stimulated AICA production in the presence of inorganic nitrogen. Methionine, as well as a-aminobutyric acid, valine, and homoserine depressed AICA accumulation. Depression of AICA accumulation by methionine in this instance may reflect inhibition of AICA synthesis. Contribution to the C2 position would be unlikely here since the organism is genetically deficient in transformylase. Wild strains of E. coli have been found to accumulate AICA when grown in purine-free medium (191) suggesting that closure of the purine ring may be a rate-limiting step in purine biosynthesis even under more or less normal conditions. 4. Aminoimidazotecarboxamidine. Aaronson et al. (1-3) have reported utilization of AICA or its amidine derivative (4-amino-5-imidazolecarboxamidine) for growth in lieu of purines. Staphylococcus flavocyaneus, Crithidia fasciculata, and Gaffkya homari utilized the amidine more effectively than the amide, although purines elicited a greater growth response. Utilization of the amidine represents a possible alternate pathway to adenine. If the amidine can be utilized in toto for purine synthesis, adenine would arise directly as a result of incorporation of a C, unit. Closure of the ring would be similar for either AICA or the amidine: H2N-C==NH b N + "Cl" CH H2N-C-NH 4-Amino-5-imidazolecarboxamidine N==C-NH2 -OH~ (-N i-xnch Lei-NH Adenine It has not been established whether the amidine can be synthesized from earlier precursors. Only under such circumstances could this represent an alternate pathway to adenine. This, and the

10 318 MOAT AND FRIEDMAN (VOL. 24 possibility of an intermediate formyl derivative of AICA, represent interesting aspects which have yet to be fully clarified. 5. Formylglycinamide ribotide. Following the demonstration of the antitumor activity of azaserine and the inhibition by it of purine biosynthesis in mice and pigeon liver systems (27, 30), several groups investigated the activity of this agent on purine synthesis in microorganisms. Azaserine inhibited the accumulation of AICA by purine-requiring mutants of E. coli. Tryptophan, phenylalanine, tyrosine, and, to a lesser extent, glutamic acid, leucine, isoleucine, and norleucine prevented the inhibition of AICA formation by azaserine (73). Bennett et al. (17) found that the inhibition of the growth of E. coli by azaserine was prevented by AICA, adenine, guanine, hypoxanthine, xanthine, methionine, glutamine, phenylalanine, tyrosine, and tryptophan. Tomisek et al. (197, 198) found that the riboside and ribotide derivatives of formylglycinamide accumulated in azaserine-inhibited cultures of E. coli. Incorporation of labeled formate, glycine, and serine into purines was prevented, but assimilation of labeled adenine, hypoxanthine, and guanine was unaffected. These findings provided convincing evidence that a major site of action of azaserine in E. coli is at a point subsequent to the formation of FGAR. Even more important, in view of the limited number of demonstrated steps in purine synthesis in microorganisms, is the demonstration of FGAR as an intermediate. In pigeon liver systems, azaserine inhibition has been shown to affect the amidination of FGAR to FGAM in the presence of glutamine and ATP (81, 82, 84), indicating that azaserine acts as a specific inhibitor of glutamine (30). Whereas glutamic acid or glutamine reversed the inhibitory action of azaserine on E. coli, a number of other compounds were more effective. Reversal by purines or AICA can be readily explained on the basis of supplying preformed products of the inhibited reaction. However, reversal of azaserine inhibition by aromatic compounds such as phenylalanine, tyrosine, and tryptophan was not readily explainable. Tomisek et al. (197) could find no evidence for an alternative metabolic pathway. Destruction or complexing of azaserine by these compounds was suggested as a possible mechanism. Participation of glutamine in the formation of tryptophan would explain the reversal of azaserine inhibition with this compound. Another possibility may be competition amongst a variety of substrates in a complex system as suggested by several recent reports (see section III C on metabolic interrelationships and feedback mechanisms). 6. Aminoimidazole ribonucleotide and the role of biotin. Love and Gots (118) found that a purinerequiring mutant of E. coli accumulated an arylamine which differed from previously described intermediates. Another purine-requiring mutant converted the new compound to AICA and chemical and physical characterizations suggested that the accumulated product was aminoimidazole or its riboside. Description of aminoimidazole and aminoimidazolecarboxylic acid as products of purine degradation by Clostridium acidi-urici (162) suggested these compounds or their derivatives as possible intermediates in both dissimilatory and assimilatory pathways. Moat et al. (143) provided evidence that an arylamine produced by S. cerevisiae under the conditions of biotin deficiency was aminoimidazole or a derivative. They noted its similarity to an arylamine described by Chamberlain et al. (36, 37), which was produced by adenine- or biotin-deficient yeast. Subsequent studies by Levenberg and Buchanan (111, 112) demonstrated the formation of AIR from FGAM in the pigeon liver pathway. Carbon dioxide, ATP, aspartic acid, and tetrahydrofolic acid were required to convert AIR to IMP, the reaction proceeding via C-AIR, SAICAR, and AICAR. Earlier studies with microorganisms (205, 206) had suggested a role for biotin in purine biosynthesis in that biotin and carbon dioxide were found to replace the aspartic acid requirement for purine synthesis by L. arabinosus. In rats (106, 124) biotin deficiency was found to depress the fixation of carbon dioxide into purines. Demonstration of a relationship between biotin deficiency and the accumulation of purine precursors by Chamberlain et al. (36, 37) and Moat et al. (143) stimulated further investigation in their laboratories (57, 117). Both groups provided evidence that the arylamine accumulated by biotin-deficient yeast was AIR rather than the free aminoimidazole. Lones et al. (117) identified a purine which accumulated simultaneously with AIR as hypoxanthine. Aspartic acid depressed accumulation of both compounds, leading them to the conclusion that the main action of biotin was in providing an adequate supply of aspartic acid for the formation of succinyl intermediates. The yeast studied by

11 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES 319 Friedman and Moat (57) was found to accumulate inosine, rather than hypoxanthine, along with AIR. They also concluded that a major role of biotin was in the synthesis of aspartic acid. However, the inability of aspartic acid to suppress AIR accumulation completely and the earlier evidence relating biotin to the fixation of carbon dioxide into purines (106, 114, 205, 206) suggested a possible role for biotin in the carboxylation of AIR. In more recent experiments (142) the conversion of AIR to AICAR in biotindeficient yeast has been shown to require biotin, aspartic acid, and bicarbonate. Upon addition of these compounds, AIR was utilized, and, in the presence of amethopterin, accumulated as AICAR. Bicarbonate and aspartic acid exerted little or no effect upon the conversion of AIR to AICAR in the absence of biotin but enhanced the rate of conversion in the presence of biotin. The relationships of biotin to the pathway of purine synthesis in microorganisms may be tentatively depicted as shown in diagram 1. IMP + L-aspartic acid + GTP AMP-S + GDP + Pi (1) AMP-S -- AMP + fumaric acid C-AIR + aspartic acid SAICAR SAICAR -* AICAR + fumaric acid (2) Citrulline + aspartic acid -+ arginosuccinic acid Arginosuccinic acid -- arginine + fumaric acid (3) Aspartic acid -f fumaric acid + NH3 (4) It is of significance that demonstration of the mechanism of AMP synthesis led to the discovery of a comparable reaction leading to AICAR formation. Miller et al. (137) reported that an enzyme which could be isolated from pigeon and chicken liver and also from yeast, E. coli, Salmonella typhimurium, and Neurospora crassa, catalyzed the cleavage of SAICAR to AICAR and fumaric acid. Gots and Gollub (74) investigated purine-requiring mutants of E. coli and S. typhimurium that exhibited a specific requirement for adenine or its ribonucleoside. The +C002 AIR - C-AIR - SAICAR - > AICAR (Biotin) I +NH3 I Aspartate ( - Fumarate I (Biotin) I AICAR - IMP > AMP-S * AMP Diagram 1 Biotin appears to play a dual role in the biosynthesis of purines, controlling two reactions through its relationship to aspartic acid synthesis and a third reaction through its role in carbon dioxide fixation. Such a complex and intimate relationship with purine biosynthesis would serve to explain the marked effect of biotin deficiency on many biological systems. In view of the evidence that biotin is related to the formation of aspartic acid (114) and the increasing evidence for the role of biotin in carbon dioxide fixation (106, 114, 123, 218), the relationship of biotin to the fixation of carbon dioxide and ammonia into purines merits more detailed investigations in purified enzyme systems. 7. Mechanism of AICAR and AMP synthesis. Carter and Cohen (34, 35) and Lieberman (115) provided evidence for the mechanism of AMP synthesis for IMP. These authors found that yeast and E. coli convert IMP to AMP via reaction 1: specific requirement for adenine reflected the loss of an enzyme responsible for the conversion of IMP to AMP-S or AMP-S to AMP. It was also observed that these organisms accumulated SAICAR, indicating the absence of an enzyme capable of splitting SAICAR to AICAR. Their findings represent convincing evidence that the enzyme which cleaves AMP-S to AMP is bifunctional and is also responsible for the cleavage of SAICAR to AICAR (reaction 2). Accumulation of AMP-S or its aglycone by adenine-requiring mutants of N. crassa (13, 61, 154, 212) lends additional support to the concept that amination by the addition of aspartic acid and subsequent cleavage to fumaric acid and the aminated derivative is catalyzed by enzymes which function equally well with the complete purine or a precursor of similar configuration. A recent report by Gollub and Gots (68) reveals that a number of auxotrophs of E. coli and S. typhimurium which lack adenylosuccinase accumulate

12 320 MOAT AND FRIEDMAN [VOL. 24 SAICAR. It has also been suggested that this may represent a more common mechanism of amination which is not necessarily limited to purines or heterocyclic compounds. The synthesis of arginine also occurs via the formation of a succinyl derivative from citrulline and aspartic acid (164). Aspartic acid deaminase, a biotinrequiring enzyme found in a number of microorganisms (114), may be similar to adenylosuccinase with regard to mechanism in that fumarate is an end product in each case. Arginosuccinase and adenylosuccinase would appear to be even more closely analogous reactions, however, they do not appear to be the same enzyme (137). Furthermore, mutants lacking the enzyme which cleaves SAICAR to AICAR possess arginosuccinase activity (68). 8. Synthesis of guanosine 5'-phosphate. From the preceding discussion, it has already been established that inosine 5'-phosphate (IMP) is the first complete purine to be synthesized and that adenosine 5'-phosphate (AMP) is synthesized via the intermediary formation of adenylosuccinic acid (AMP-S). The intermediary steps to guanosine 5'-phosphate (GMP) have been described for several systems. Magasanik and Brook (128) first observed that a guanineless mutant of A. aerogenes accumulated xanthosine, suggesting that the formation of nucleic acid guanine involved the formation of xanthosine or a derivative (e.g. xanthosine 5'-phosphate). Gehring and Magasanik (128) found that inosinic acid dehydrogenase was a DPN-requiring enzyme and that XMP was the product of this reaction. Moyed and Magasanik (148) subsequently found that the over-all reaction from XMP to GMP required the presence of ATP, ammonia, and Mg+. AMP and inorganic pyrophosphate were products of the reaction along with GMP. Magasanik et al. (130) demonstrated further that the intermediary formation of XMP was an obligatory step in the formation of nucleic acid guanine and that the reaction was apparently irreversible. A comparable reaction was shown to occur in bone marrow extracts by Abrams and Bentley (4-6). IMP dehydrogenase was shown to be a DPN-linked enzyme and XMP the product of the reaction. XMP aminase required the presence of L-glutamine, ATP, and Mg++. Lagerkvist (101, 102) demonstrated a similar system in pigeon liver extracts. The mechanism of formation of GMP from IMP is shown in table Evidence for other intermediates. Yura (224) has provided genetic and chemical evidence for a series of genetic blocks in purine synthesis by purine-requiring mutants of salmonellae. Mutants were obtained which accumulated AIR, AICAR, IMP, and another compound believed to be an imidazolone. Although the evidence for the identity of this compound was scant, the possibility of an imidazolone as a precursor to AIR would represent an interesting variation in the pathway as presented in table 1. Ring closure with FGAR as the substrate would yield an imidazolone. Amination at the 5-position could occur after ring closure as well as before. This possibility bears further exploration. It has been assumed that adenine and guanine are the only purine bases which occur in nucleic acid. However, Littlefield and Dunn (116) have recently reported the isolation of methylated derivatives of adenine from the RNA of several microorganisms. 2-Methyladenine, N-methyladenine, and N, N-dimethyladenine were demonstrated in E. coli, A. aerogenes, Staphylococcus aureus, and a yeast. Plant tissues (wheat germ) also contained measurable quantities of the methyl purines, but two strains of plant viruses were devoid of these derivatives. 6-Methylpurine was found in rabbit liver and rat liver microsomes, however. C. Metabolic Interrelationships and Feedback Control Mechanisms 1. Purine-histidine interrelationships. Feeding experiments in birds and mammals frequently produced evidence of a significant contribution of histidine to the formation of purines (38, 172). With labeled isotopes, contributions from a variety of compounds, usually at lower levels, were eventually dismissed as representing metabolic cycling of some portion of these compounds to the direct precursors. Although this conclusion is essentially correct, the ability of a given organism to carry out interactions of this type may represent important alternate sources of purine precursors. Tesar and Rittenberg (38) eliminated histidine as a major contributor of nitrogen to the purines of the rat and Di Carlo et al. (44) produced similar evidence in S. cerevisiae. On the other hand, several authors (28, 29, 127, 128, 166, 167) reported significant incorporation of the C2 of the imidazole ring of histidine into the C2 and C8 position in the purines of pigeons and rats.

13 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES Purine-histidine relationships were described for NH3+GLUTAMATE Lactobacillus casei and L. arabinosus by Broquist GAR and Snell (24) and for an E. coli mutant by h-thf-cho DEGR FGAR PATH Luzzati and Guthrie (122). This E. coli mutant i,thf was found to incorporate histidine carbon into ATP FADG F nucleic acid purines (11). GLUTAINE SYNTH R-5-P-AMP----TAMI- AICAR*IGPPAH Elucidation of the mechanism of histidine biosynthesis and degradation in mammalian tissues FUMARATE FADG ISTIDINE and microorganisms by several I *--THF-CHO0 DEGR groups provided a ASPARTATE THF PATH basis for the explanation of the relationship of AMP-S- IMP NH3+GLUTAMATE histidine to purine synthesis (128, 169, 204). Due credit is given to these authors by Magasanik Figure 2. Metabolic cycling of purine and histidine metabolites. and Bowser (129) and Revel and Magasanik (169), who have compared the degradation of histidine by liver with that occurring in microorganisms. An organism which can degrade enzymes capable of the further conversion of which accumulated AIR presumably contained histidine to formate (e.g. Pseudomonas fluorescens) can utilize the 2 carbon of histidine for being formed from the added ATP via the reac- AIR to AICAR. Indications are that AICAR was the synthesis of purines. One which forms formamide (e.g. A. aerogenes) cannot. The unequal pathway. tion above rather than through the biosynthetic utilization of histidine carbon for formation of If histidine contributes to the C8 position of the Cs position of purines was defended on the purines (and possibly to C2) and purines contribute to the N1-C2 of the imidazole ring of histi- basis of similar findings by Koch (95) and Paterson and Zbarsky (156), who reported that the dine, then it is possible for metabolic cycling to distribution of formate and glycine depended occur in both directions under certain conditions upon the initial concentration of the precursors. (figure 2). The contribution of purine nitrogen and 2. Feedback control mechanisms. Gots (71) carbon to histidine has also drawn recent attention. Utilization of the C2 of guanine for the requiring mutant which accumulates AICAR investigated the mechanism whereby a purine- formation of the C2 of the imidazole ring of ceased to accumulate this intermediate when histidine has been reported by Mitoma and purines were added. Assuming that the genetic Snell (141) for L. casei and by Magasanik et al. block is complete, this action cannot be explained (127, 131) for A. aerogenes and E. coli. Neidle simply on the basis of product inhibition. In a and Waelsch (149) found that the amino group of proliferating system, prevention of the accumulation of an intermediate could be explained on the guanine did not enter the imidazole ring of histidine. Moyed and Magasanik (147) determined that an adenine derivative, rather than of other products. In a nonproliferating system, basis of precursors being utilized for the synthesis guanine, appeared to be the actual donor of the obtained by imposing a secondary requirement N1-C2 fragment which entered histidine. IGP for tryptophan, AICAR still accumulated and (imidazoleglycerolphosphate) and AICAR were purines prevented this accumulation. The inhibition by purines was immediate and proportional the immediate products of the reaction in extracts of a histidineless S. typhimurium: to the logarithm of the purine concentration. Inhibition of enzyme synthesis was eliminated as + Glutamine R-5-P + AMP AICAR + IGP a possible explanation by demonstration of the + ATP presence of the enzymes required for AICAR An intermediate formed from the coupling of synthesis under the conditions of the experiment. R-5-P at the N1 position of AMP was postulated Inhibition of the action of the enzymes involved, and partially characterized. This reaction appears or the limitation of availability of an earlier to be a major route of histidine synthesis in S. substrate were suggested as possible explanations. Phosphoribosylpyrophosphate (PRPP) typhimurium in that mutants blocked prior to IGP were devoid of this activity. These authors represents such a compound since it is required indicated that their findings explained the observation of Love (119) that an E. coli mutant reactions leading to the synthesis of for nucleotide formation as well as for initial phosphoribo- 321

14 322 MOAT AND FRIEDMAN [VOL. 24 sylamine in avian systems. Wyngaarden and Ashton (223) provided evidence that competition for PRPP is the explanation for similar feedback mechanisms in the avian liver system. Only those compounds which could react with PRPP could be shown to inhibit intermediary reactions in the synthetic pathway. Thus competition for available precursors emerges as the logical explanation for the inhibitory action of purines on earlier reactions in the system. Recent reports by Goldstein and Gots (62) and Moyed and Friedman (146) indicate that feedback control may be dependent upon purine-histidine interrelationships. Such interrelations provide a fortuitous and economical metabolic control mechanism, particularly for rudimentary systems not subject to hormonal or other control mechanisms. They also serve as potential explanations for seemingly nonspecific stimulatory or inhibitory actions by other metabolites such as methionine, phenylalanine, and tryptophan (57, 72, 73, 75, 143, 197, 198). A very recent report by Talman et al. (195) concerning experimental porphyria in chick embryos induced by allylisopropylacetylcarbamide or allylisopropylacetamide (AIA) seems referable to interactions of the type discussed here. Under the conditions of experimental porphyria induced by AIA, uric acid synthesis is depressed. Exogenous adenine was excreted as uric acid by porphyric embryos. The incorporation of labeled glycine into uric acid was depleted. Adenine reduced the amount of porphyrin excreted and improved the growth of AIA-treated embryos. Since AIA impairs purine synthesis while enhancing porphyrin synthesis, it appears that some common precursor is blocked from entering the purine pathway but continues into porphyrin. The authors suggested that a block in the succinate-glycine cycle (186) which retarded the entry of the 5-carbon of 6-aminolevulinic acid into the C1 pool and from there to purines would increase the amount of 3-aminolevulinic acid excreted via porphyrin. Such a mechanism is comparable to those proposed in explanation of the observations on the interrelationships of purine and histidine. Metabolic interactions of this nature may have little to do with the synthesis of purines from elementary precursors under one set of conditions. Under altered conditions they may well play a THF-CHO R-5-P+GLYCINE+GLUTAMINE-GAR-- FGAR GLUTAMINEI FUMARATE- ASPARTATE FGAM 2 (BIOTIN) I AICAR SAICAR C-AIR' Go, IR jthf-cho QIOTW IMP AMP-S -AMP 1 DPN XMP ASPARTATENFUMARATE ATPJGLUTAMINE IOTIN) GMP Figure S. Pathways of purine biosynthesis in microorganisms. Intermediates or cofactors whose role in microbial systems require further verification are indicated by brackets. major role in affecting the manner in which this pathway operates, resulting in many of the seemingly unexplainable findings that have been observed under the conditions of analogue inhibition (211), genetic blocks (128), or experimentally altered concentrations of the various reactants (95). In summary of the information which has been presented, a general scheme may be formulated for the synthesis of purines by microorganisms (figure 3). From this abbreviated scheme it is apparent that some important reactions known to be present in the avian system have not been adequately described in microbial systems. Hartman and Buchanan (83) refer to unpublished observations in which these steps have now been demonstrated. Since the over-all pathway is suggestive of a distinct similarity with the avian system, there is little doubt that the intervening steps will eventually be demonstrated in detail. Thus far, no completely alternative pathways leading from simple precursors to the completed purine ring have been found. Metabolic cycling, which has been shown to play an important part in providing precursors to microbial purines, may be considered the equivalent of alternative pathways under circumstances where the organism is confronted with an excess of a metabolite which it can readily utilize for purine formation. Under conditions where the rate of synthesis from these alternate sources of precursors equals or exceeds the main pathway, they may play an important part in purine synthesis. Such a situation might arise where the main pathway is inhibited by a chemotherapeutic agent or metabolic antagonist.

15 19601 BIOSYNTHESIS AND INTERCONVERSION OF PURINES 323 D. Purine Biosynthesis in Mammalian and Cancer Systems The pathway to purines in mammalian and neoplastic tissues has not been fully clarified by studies of a specific nature. The precursor experiments cited earlier provided evidence that the direct contributors to the purines in a variety of mammals and tumors are similar to those found in avian and microbial systems. Otherwise, one finds a relative paucity of reports delineating specific intermediates. In the few studies where intermediates have been isolated, they are identical Ṁost studies in mammalian and cancer systems have followed a similar design. The incorporation of known purine precursors such as formate, carbon dioxide, or glycine into the nucleic acid purines is taken as evidence of purine synthesis from these elementary precursors (the so-called pathway de novo). Prevention of the incorporation of these precursors by antimetabolites is, in turn, taken as evidence that the pathway de novo is a major site of inhibition. If the site and mechanism of action of an inhibitor has previously been determined to be a specific step in the biosynthesis pathway, then inhibition of precursor incorporation in a less well defined system presumably indicates the presence of a comparable reaction or reaction sequence in the system under study ( ). For example, Goldthwait and Bendich (65) found that aminopterin depressed formate incorporation into nucleic acid purines and pyrimidines to a greater extent than utilization of preformed purines in a variety of tissues of the rat. Totter and Best (199) reported similar results in rabbit bone marrow. Tetrahydrofolic acid, but not folic acid, partially reversed the action of aminopterin. Tomisek et al. (196) found that amethopterin decreased formate incorporation into the purines of the acidsoluble fraction of L-1210 leukemia and normal mouse intestine. It was further demonstrated that amethopterin diverted the administered formate from purines to a variety of amino acids and hydroxy acids. AICAR accumulated and was highly labeled, indicating that final closure of the purine ring was the site of inhibition. In an amethopterin-dependent line of L-1210 leukemia, formate incorporation into purines was increased in the presence of the inhibitor. Attempts to demonstrate an "alternative pathway" of synthesis in the dependent cell line were unsuccessful. Miller and Warren (139) had previously demonstrated the utilization of AICA by rat liver homogenates and, in a later report (138), by embryonic tissues, regenerating liver, and several rapidly growing malignant tumors. The fate of AICA was not determined in these studies, however. In summarizing the findings of these various groups, one may conclude that AICA or AICAR is an intermediate in both mammalian and cancer cells and, on the basis of the inhibitory action of known antifolic agents, that a functional form of folic acid is concerned with final closure of the purine ring. Azaserine was shown by Bennett et al. (17) to decrease the incorporation of C14-labeled formate and glycine, but not the utilization of C14- labeled adenine, AICA, or hypoxanthine for the synthesis of nucleic acid purines. These results suggested a site of action prior to the formation of AICA in normal and neoplastic systems in the mouse and in microorganisms. Greenlees and LePage (79) and LePage and Sartorelli (110) found glycine incorporation into tumor nucleic acid purines to be inhibited by azaserine. A compound resembling glycinamide ribotide (GAR) accumulated and was found to contain relatively large amounts of the radioactive C, from glycine. Moore and LePage (144) later found that FGAR accumulated in the presence of azaserine and DON (6-diazo-5-oxy-L-norleucine) along with a second compound which was not identified. Glutamine was found to increase the amount of FGAR accumulating. These findings may be compared with the demonstration that azaserine and DON appear to be glutamine antagonists and specifically prevent the conversion of FGAR to FGAM (27, 30, 113). The similarity of effects in normal mammalian tissues and a variety of transplantable tumors suggests that purine biosynthesis proceeds via FGAR in these tissues. More recent observations by Sartorelli and LePage (176), concerning the development of azaserine resistance in TA3 ascites tumors is of interest in that cross-resistance with DON and N-methylformamide was reported. This suggests a similar mode of action of these three agents. Furthermore, demonstration of an increased ability on the part of the resistant cell

16 324 MOAT AND FRIEDMAN [VOL. 24 line to utilize preformed purines for nucleic acid synthesis provides an explanation of the resistance mechanism. The occurrence of AMP-S and its aglycone in mammalian liver (92) appears to be the only indication, to date, for the presence of these compounds as intermediates in mammalian systems. However, a report by Tschudy et al. (202) suggests that aspartic acid contributes nitrogen to purines in humans. Increased incorporation of aspartic acid nitrogen into urinary uric acid following surgery was observed. No differences were found in the distribution of aspartic nitrogen between normal individuals and patients with a variety of cancers, however. Although the information is relatively sparse, these reports suggest that the formation of succinylo-derivatives is involved in the formation of purines by mammalian tissues via a mechanism comparable to that shown in avian and microbial pathways. The biosynthesis of GMP from IMP via XMP has been demonstrated in calf thymus and bone marrow extracts (4-6). In microorganisms this reaction requires ammonia rather than glutamine (148), indicating that this step represents ammonia fixation rather than an amido transferase as found in mammalian and avian tissues (4-6, 101, 102). Further evidence is required before it may be stated unequivocally that the biosynthetic pathway in mammalian and cancer tissues is identical with that shown in pigeon liver and microorganisms. IV. INTERCONVERSION OF PURINES In the previous section, the conclusion was drawn that the first purine derivative to be synthesized is IMP. Organisms supplying their purine requirements through synthesis must convert IMP to AMP and GMP in order to provide nucleic acid adenine and guanine. These two pathways were, therefore, considered a part of the biosynthetic pathway. They are also concerned with the utilization of preformed purines for nucleic acid synthesis, and reactions which are of greatest importance in synthesis are those which channel other purines into these two pathways. Unlike purine biosynthesis, the metabolic interactions of purines have been known and studied over an extended period of time. A number of reviews have appeared (25, 26, 33, 38, 86, URIC ACID Figure 4. Interconversion of purines and their derivatives. 88, 165, 179, 180, 184), which summarize earlier work in considerable detail. The activity of a number of potential anti-cancer agents against interconversion mechanisms (9, 23, 25, 26, 39, 48, 50, 51, 135, 177, 178) has stimulated a renewed interest in these reactions. As a result, many important contributions have been made since the most recent reviews were published. A much clearer understanding of interconversion reactions and their relationship to nucleic acid synthesis and tumor development may now be presented. The known metabolic interactions of purines which may be summarized from recent publications and earlier reviews is shown in composite form in figure 4. In comparison with the rather singular pathway of biosynthesis thus far observed in most biological systems, the potential reactions available for the interconversion of purines would seem to represent a veritable maze leading eventually to nucleotide adenine and guanine. It is with this thought in mind that we compare recent findings concerning the metabolic utilization of purines for nucleic acid synthesis by microorganisms with comparable reactions in mammalian, avian, and cancer systems.

17 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES 325 A. Microbial Interconversions 1. Specific enzymatic mechanisms. Unequivocal demonstration of the reactions depicted in figure 2 is limited to a relatively small sampling of microbial species. No completely defined pattern has been established for any given organism, although L. casei, E. coli, and a few others have been investigated in considerable detail. Deamination of adenylic acid and adenosine has been demonstrated in E. coli (192), Aspergillus oryzae (140), and Tetrahymena pyriformis (47). T. pyriformis deaminated deoxyadenosine, yeast adenylic acid, and 5'-adenylic acid. Adenase and guanase could not be demonstrated in cell free homogenates. Variations in the ratio of deamination of adenosine and deoxyadenosine with the age of the organism suggested that two different enzymes may be involved in the metabolism of the two nucleosides (47). Friedman and Gots (58, 59) found that E. coli deaminated adenine, adenosine, adenosine 3-phosphate, isoguanine, and 2,6-DAP. Adenase (adenine deaminase) has been shown to be present in Azotobacter vinelandii (87). Nucleosidases which split adenosine, xanthosine, guanosine, and inosine as well as pyrimidine nucleosides have been demonstrated in extracts of Lactobacillus pentosus (105). Nucleoside phosphorylases and trans-n-glycosidases were not observed. Deoxyribose nucleosides were found to be cleaved by reversible nucleoside phosphorylases in cell free extracts of E. coli (134). Phosphorolytic cleavage of inosine and guanosine has been observed in baker's yeast (86). Hydrolytic cleavage of nucleosides was also demonstrated, the latter reaction being active against adenosine as well. Nucleosidase activity for inosine, guanosine, and hypoxanthine deoxyriboside has been observed in E. coli (59). Nucleoside phosphorylases have been found in E. coli (95-97) and in T. pyriformis (47). C. acidi-urici was shown to contain nucleoside phosphorylase, guanase, and xanthine dehydrogenase (xanthine to uric acid) as well as enzymes degrading xanthine to ureidoimidazolecarboxylic acid, aminoimidazolecarboxylic acid, aminoimidazole, and finally to glycine, formate, and ammonia ( ). The induced synthesis of uricase has been demonstrated in yeast (174). Adenosine kinase has been isolated from yeast (32, 100, 175): Adenosine + ATP = AMP + ADP Kornberg et al. (99) have also found yeast capable of converting purine bases to nucleotides via pyrophosphorylase: Purine + PRPP = Purine nucleotide + PP This reaction appears to be a rather general one in that most purine and pyrimidine derivatives are utilized. It has been found in E. coli as well as in avian and mammalian tissues (99, 216). trans - N - Deoxyribosylase (trans - N -glycosidase) activity has been observed in Lactobacillus helveticus and E. coli (93, 103, 125). Roush and Betz (173) investigated the details of this reaction in L. helveticus. The enzyme was found to catalyze an exchange reaction with both purine and pyrimidine deoxyribosides and either purine or pyrimidine bases: Purine, or pyrimidine, deoxyriboside + purine2 or pyrimidine2 = Purine2 or pyrimidine2 deoxyriboside + purine1 or pyrimidine, The indications are, however, that the distribution of this enzyme may be rather limited as it could not be demonstrated in Lactobacillus delbrueckii, ten genera of yeast, two molds, six bovine organs, rat liver, or the hepato-pancreas of crayfish. 5'-AMP ribosidase, an enzyme which splits 5'-AMP to adenine and R-5-P, has been demonstrated in A. vinelandii (89). The reaction is irreversible and the enzyme appears to differ from the pyrophosphorylase of yeast (99), and the adenosine kinase of yeast (100, 175). 2. Other evidencefor interconversion mechanisms. Brown and Roll (26) have summarized the ability of several microbial species to convert adenine and guanine to polynucleotide purines. Other examples have been published more recently (40, 90, 126, 158, 219). Four groups are represented as summarized below: a. Obtain both nucleic acid adenine and guanine from either adenine or guanine: Escherichia coli, Plasmodium aurelia, Salmonella typhimurium, Aerobacter aerogenes, Ochromonas malhamensis, Lactobacillus casei, and Streptococcus faecalis. b. Utilize adenine for the synthesis of both nucleic acid adenine and guanine. Cannot utilize guanine for the synthesis of nucleic acid adenine:

18 326 MOAT AND FRIEDMAN [VOL. 24 Lactobacillus leichmannii, Staphylococcus aureus, and Tetrahymena geleii. c. Utilize guanine for the synthesis of both nucleic acid adenine and guanine. Cannot utilize adenine for the synthesis of nucleic acid guanine: Neurospora crassa and Torulopsis utilis. d. Neither adenine nor guanine can be utilized for the synthesis of both nucleic acid adenine and guanine: Corynebacterium diphtheriae. Information concerning the utilization of other purine bases or nucleosides and nucleotides is often revealing with regard to interconversion mechanisms. For example, Dalby and Holdsworth (40) found that hypoxanthine can serve as the sole source of purines for C. diphtheriae indicating that this was the only purine which could be converted to both nucleic acid adenine and guanine. S. faecalis can apparently utilize any purine for growth, indicating the ability to convert all purines to nucleic acid adenine and guanine. In the presence of folic acid, purines are synthesized and the organism does not require exogenous purines for growth (90). E. coli, which does not require purines for growth, utilized any purine except uric acid when synthesis was blocked by mutation (80). Incorporation of C'4-labeled adenine and guanine into nucleic acid purines (97) also indicated that E. coli interconverted adenine and guanine readily in either direction and apparently utilized preformed purines in preference to synthesizing them. Mutagenic methylpurines, on the other hand, were not incorporated into nucleic acids to any significant extent (96). A. aerogenes appears to possess comparable flexibility in the utilization of purines for the synthesis of polynucleotide adenine and guanine (8, 128). Unfortunately, comparable information is not available for all of the species listed. The equal utilization of purine ribotides and free bases for the growth of L. casei as compared with the poor utilization of ribosides suggested that nucleotides were not synthesized from purine bases via the intermediate formation of ribosides but were converted directly to ribotides (26). Although the possibility of a permeability barrier to the ribosides was not completely eliminated, the ready utilization of nucleotides bears out such an interpretation. E. coli, by comparison, degraded nucleotides to the purine stage prior to utilization of the components (10). 2,6-DAP, an unnatural purine, is in the somewhat anomalous position of being utilized by L. casei for the synthesis of nucleic acid adenine and guanine regardless of whether the conditions are such that DAP is stimulatory or inhibitory to growth. Under inhibitory conditions a significant decrease in the extent of interconversion of adenine and guanine was demonstrated (52). At high concentrations of DAP, only adenine reversed the inhibition. Utilization of ribotides was even more strongly inhibited in that adenine nucleotides could not relieve the inhibition. A DAP-resistant strain exhibited a decreased ability to utilize adenine and DAP as sources of PNA purines, as indicated by the inability of adenine to reverse the action of DAP under these circumstances. These findings suggested that adenine and DAP are metabolized via the same route in the wild type. The mutant has apparently lost the ability to convert adenine and DAP to the nucleotide level (52). In yeast it was shown (94) that DAP was used for the synthesis of nucleic acid guanine, but not of adenine. In the presence of guanine and DAP, the incorporation of formate into guanylic acid was repressed, and its incorporation into adenylic acid was increased. Studies on a purine-requiring mutant of A. aerogenes (P-14) which exhibited a specific requirement for guanine or DAP (8), revealed that the synthesis of guanine, but not of adenine, was lost through mutation. Later work (148) proved that nucleotide guanine was synthesized from IMP via XMP in A. aerogenes, implying that mutant P-14 cannot convert XMP to GMP. This conclusion was substantiated by the accumulation of XMP by the mutant. GMP can presumably be synthesized from guanine as indicated by the stimulation of growth in the presence of guanine. Utilization of DAP for growth in place of guanine further suggests that DAP is converted to the nucleotide stage via the same pathway as guanine. These findings are in agreement with the observation that DAP is converted almost exclusively to polynucleotide guanine in rats and other mammals (26). It would appear that DAP stimulates growth if it is deaminated and converted to nucleic acid guanine and inhibits growth if it is converted directly to a nucleotide of DAP. The nucleotide may then inhibit the function of nucleotide coenzymes. Conversion of small amounts of DAP to guanine may have little bearing on the inhibi-

19 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES 327 tory action when the inhibitor is present in sufficiently high concentrations. 6-Mercaptopurine (6-MP), closely analogous in structure to hypoxanthine and adenine, also has been shown to inhibit the growth of L. casei. Growth could be restored by any of the physiological purine bases or by the nucleotides of adenine and guanine. A 6-MP-resistant strain of L. casei, which could not use hypoxanthine for growth, used adenine or adenylic acid poorly, but grew well on xanthine, guanine, or guanylic acid. In the wild strain, 6-MP markedly inhibited the conversion of adenylic acid b to RNA guanine but the conversion of free adenine to RNA guanine was almost completely unaffected. The resistant mutant accumulated hypoxanthine riboside (inosine) and free hypoxanthine, indicating its inability to convert hypoxanthine to the ribonucleotide (48-51). The primary metabolic block which confers 6-MP resistance appears to be the loss of the ability to convert hypoxanthine to IMP. Further evidence was presented by Brockman et al. (23), who showed that 6-MP was incorporated into nucleosides and nucleotides only by sensitive strains. This reaction may, therefore, be considered to be a major site of action of 6-MP, although inhibition of reactions involving the six position (42) may also be of equal importance (see section below for further discussion of the mechanism of action of 6-MP in mammalian and cancer tissues). Such studies serve not only to determine the site of action of carcinostatic agents but to elucidate the metabolic interactions which occur in the organism (133). Obviously, inhibition of one interconversion reaction may not prevent nucleic acid synthesis completely because of the multiplicity of interconversion mechanisms which are operative in most systems. B. Interconversions in Mammalian, Avian, and Cancer Tissues Most of the recent information concerning the interconversion of purines by mammalian, avian, and cancer tissues has been derived from studies designed to determine the chemotherapeutic value or the mechanism of action of purine derivatives against experimental tumors as compared with their action on normal tissues. It will, therefore, be convenient to consider the evidence for interconversion mechanisms of all of these various types together. 1. Specific enzymatic mechanisms. Nucleases and other enzymes attacking nucleic acid components have been reviewed (180). The synthesis of nucleic acids has also been reviewed recently (179). Reactions which provide mechanisms for the interconversion of purines are included in the composite chart shown in figure 4. It is important to emphasize that many of the enzymatic actions on purines and their derivatives are essentially irreversible, limiting the number of pathways by which purines may be interconverted. For example, nucleotidases, nucleosidases, adenylic and guanylic acid deaminases, adenosine deaminase, adenase, guanase, and xanthine oxidase have been shown to act primarily in the direction of degradation. However, insofar as they may provide for the formation of purine derivatives which can be incorporated into nucleic acid adenine and guanine, they are of importance in relation to nucleic acid synthesis. The enzymes concerned with the direct conversion of IMP to AMP and GMP have been discussed as a part of the biosynthetic pathway. The formation of GMP from IMP via XMP has been demonstrated in bone marrow and calf thymus (4-6) and in pigeon liver (101, 102). Pyrophosphorylases, which convert purines directly to nucleotides in the presence of PRPP, were found in avian and mammalian tissues by Kornberg et al. (99) and Korn et al. (98, 216). Lukens and Herrington (121) showed that a pyrophosphorylase from beef liver converted hypoxanthine, guanine, and 6-MP to the nucleotides, but was inactive with AICA, xanthine, uric acid, 8-azaguanine, 2, 6-DAP, or orotic acid. Kornberg et al. (99) and Remy et al. (168) have also demonstrated the mechanism of synthesis of PRPP: ATP + R-5-P = 5'-AMP + PRPP The enzyme, purified from pigeon liver, was also demonstrated in mammalian liver and microorganisms. Adenosine kinase, which catalyzed the formation of adenylic acid from adenosine in the presence of ATP, has also been found in mammalian tissues (32). Lee (107) has obtained highly purified preparations of 5'-adenylic acid deaminase from rabbit skeletal muscle. 8- Hydroxyadenine has been shown to be an intermediate in the conversion of adenine to 2,8- dihydroxyadenine by xanthine oxidase (222). Quantitative measurement of the levels of purine metabolizing enzymes has revealed im-

20 328 MOAT AND FRIEDMAN [VOL. 24 portant differences in interconversion mechanisms in tumors as compared with host tissues. Bergel et al. (22) demonstrated that xanthine oxidase levels were lower in mouse mammary carcinoma than in normal host tissues. The arrest of these tumors by administration of xanthine oxidase suggested that the control of purine metabolism by the concentration of enzyme present may have far-reaching effects. De Lamirande et al. (43) investigated the levels of activity and intracellular distribution of 5'- nueleotidase, nueleoside phosphorylase, adenosine deaminase, guanase, adenase, xanthine oxidase, and uricase in normal liver and Novikoff hepatoma. The action of 5'-nucleotidase, nueleoside phosphorylase, guanase, and adenase were greatly diminished whereas that of adenosine deaminase was increased in transplanted hepatoma as compared with normal liver. Xanthine oxidase and uricase activity could not be demonstrated in tumor transplants. Their findings suggest that the catabolism of purines and purine derivatives in Novikoff hepatoma is generally impaired and is completely blocked at the stage of xanthine and hypoxanthine. The findings of these two groups are in good agreement regarding the inability of the tumors studied to degrade purines via xanthine oxidase. Conservation of purine metabolites by tumor cells via the deletion of catabolic enzymes represents an important concept as to the mechanism of tumor development which is currently being investigated with regard to other neoplasms (19, 185). 2. Metabolism of unnatural purine derivatives. The metabolic utilization of purine derivatives may have bearing on their effectiveness as antipurines as shown by the findings with regard to 8-azaguanine. This compound and other 8-azapurines have been investigated extensively for their effect upon a number of biological systems. Its incorporation into nucleic acid appears to be related to its inhibitory action (132, 133, 135). AICA inhibits the deamination of 8-azaguanine by guanase, affording an explanation for the increased toxicity of 8-azaguanine toward mice in the presence of AICA. Prevention of the degradation of the inhibitor also explains the increased effectiveness of the combination of AICA and 8-azaguanine in inhibiting a sarcoma (31), the increased incorporation of guanine into the liver nucleic acids of CAF, mice (31) and into the polynucleotides of Sarcoma 180, Adenocarcinoma 755, and human sarcoma (18). AICA was also found to inhibit the conversion of 2,6-diaminopurine to nucleic acid guanine in both tumor and normal tissues. It had previously been observed by Bennett et al. (20, 21) that, relative to control tissues, all tumors incorporated guanine poorly, whereas 2,6-diaminopurine was utilized for the synthesis of nucleic acid guanine by tumors at least as well as by intestine or liver. They concluded that tumors contained high levels of guanase and degraded it rapidly. Presumably guanase acts less readily upon 2, 6-diaminopurine, allowing its conversion to the nucleotide stage where it may be converted to nucleic acid guanine by deamination at the 6 position. 3. Other evidence for interconversion mechanisms. Mammalian species have been shown to vary considerably in their ability to utilize preformed purines for the synthesis of nucleic acid adenine and guanine. By labeling experiments it has been established that guanine is utilized only sparingly by Sherman rats. C57 black mice, on the other hand, utilized guanine quite readily. Both species used dietary adenine for the synthesis of both nucleic acid adenine and guanine (26). In hamsters, a comparison of utilization by host tissue as compared with heterologous tumor implants revealed that each tissue and tumor exhibited a characteristic pattern of incorporation of PNA precursors (12). In general, tumors utilized preformed purines to a much smaller extent than the intestine, liver, kidney, or spleen. Glycine was incorporated into PNA purines equally by tumor and host tissues. In the presence of tumors, host tissues exhibited an increased incorporation of hypoxanthine into PNA adenine. In addition to suggesting potential pathways available for purine utilization, these findings also indicated that an inhibitor of purine biosynthesis, in the presence of hypoxanthine, might be expected to reduce PNA synthesis in the tumors whereas intestinal synthesis would be maintained at normal levels by an exogenous supply of hypoxanthine. Investigations on the mechanism of action of 6-mercaptopurine (6-MP) and 6-thioguanine (6-TG), which exhibits cross-resistance with 6-MP (145), have provided evidence for the metabolic interconversion of purines by a variety of mammalian tissues and tumors. Most of the evidence is in agreement with that obtained from microbial inhibition studies, i.e. 6-MP and 6-TG must be converted to the nucleotide stage in order

21 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES to be inhibitory (108, 155, 189) and in this form prevent the utilization of hypoxanthine for adenine synthesis. Simultaneous administration of azaserine and 6-MP or 6-TG is necessary for maximal tumor inhibition (177), suggesting that bypass mechanisms are operative if the 6-substituted purines are utilized alone. 6-TG is incorporated into the nucleotides of Ehrlich ascites tumors and into the intestine, liver, and bone marrow of Swiss mice. In mouse tissues, approximately equal amounts were found in nucleotides and thiouric acid (145). LePage (108) has demonstrated that both derived resistant cells and naturally occurring thioguanine-resistant lines of Ehrlich ascites carcinoma metabolize 6-TG to the nucleotide level. The presence of more active degradative enzymes for 6-TG appeared to be the biochemical basis for the resistance of the sublines. In L-1210 leukemia, 6-PM prevents the incorporation of hypoxanthine into adenine but not into guanine (42), suggesting that the agent acts on the conversion of hypoxanthine to the nucleotide (IMP). Glycine incorporation into nucleotide adenine is also inhibited without action on its incorporation into nucleotide guanine, indicating that metabolic blockage also occurs between IMP and AMP. Resistance to 6-MP or 6-TG may arise via several possible mechanisms: (a) Inability to form the nucleotides, (b) inability to convert nucleotides to the nucleic acid level, or (c) increased degradation by guanase and xanthine oxidase (108). 4. Reutilization of host purines by tumors. In the case of tumor-host protein relationships, the concept of the tumor as a nitrogen trap has been developed. The tumor is in positive nitrogen balance even when the host is starving (203). It is felt that this same situation exists with respect to nucleic acid precursors, but only a few detailed studies have been made in attempting to study this possibility. Dancis and Balis (41) concluded from several experimental situations with Sarcoma 180 in Swiss albino male mice that there was relatively little reutilization of labeled host purines for the resynthesis of tumor nucleic acids. Henderson and LePage (85) implanted Sarcoma 180 into mice 48 hr after injecting labeled adenine. After several days the tumor acquired significant radioactivity in the acid-soluble and nucleic acid purines. TA3 ascites tumor cells in CAF1 mice also acquired significant amounts of radioactivity from labeled host purines. Since both normal and malignant cells are known to 329 synthesize purines de novo ( ), it is difficult to assess the importance of reutilization of host purines by tumors and of species differences among hosts. Variations in this process may be a partial explanation for the differences in effectiveness of antimetabolites as chemotherapeutic agents. V. GENERAL DISCUSSION AND SUMMARY The specific reactions leading to the biosynthesis of purines from elementary precursors have been well documented for only a few biological systems. Avian liver and microorganisms have been studied most extensively and it is now possible to chart a reasonably complete pathway of reactions from elementary precursors to the completed purine molecule. Several steps have never been studied in detail in microbial systems and represent important areas for further investigation. Recent work indicates, however, that these steps are identical in microbial systems and presumably will be described in detail in the near future (83). Relatively few of the steps shown in table 1 have been delineated in mammalian and cancer cells. Nevertheless, the similarity of precursors (glycine, formate, aspartate, and carbon dioxide) and the demonstration of formylglycinamide ribonucleotide, 5-amino-4-imidazolecarboxamide ribonucleotide, and adenylosuccinate as intermediates is taken by most as sufficient evidence to indicate that an identical (or very nearly identical) pathway exists in mammalian and tumor systems. In fact, most of the data presented to date suggest that the pathway is, for all practical purposes, identical in all biological forms which are capable of synthesizing purines. The remarkable similarity in the de novo pathway in all biological forms makes it seem that a search for agents effective against this pathway which would selectively inhibit the development of one form while having limited action on another would stand little chance for success. This conclusion is belied by the demonstrated effectiveness of agents such as aminopterin, amethopterin, azaserine, and 6-diazo-5- oxo-l-norleucine against both experimental and human neoplasms. Thus, the remarkable similarity in the de novo pathway may be offset by differences in the degree to which metabolic recycling and feedback mechanisms may be operative in one organism as contrasted to another. Contribution of carbon from histidine

22 330 MOAT AND FRIEDMAN [VOL. 24 to purine synthesis and the utilization of purine nitrogen and carbon for histidine synthesis indicate that the precursor pattern may be markedly altered by the presence of competing substrates. Feedback inhibition of the synthetic pathway by preformed purines and the alterations in purine synthesis resulting from inhibition of the succinate-glycine cycle represent examples of the manner in which metabolic interactions may affect more than one pathway. The metabolic interconversion of purines also appears to be basically similar in all of the biological systems which have been investigated to date. However, as compared with the synthetic pathway, many individual variations have been demonstrated. Certain reactions which potentially could occur (e.g. inosine to xanthosine, guanosine to xanthosine) do not appear to exist in any system. At least they have not been reported. Others, while present in microorganisms as well as higher forms, represent "one-way streets" in that they may be considered irreversible (e.g. reactions mediated by xanthine oxidase, inosine 5'-phosphate dehydrogenase, adenylic deaminase). Of the reactions shown in figure 4, not all may be present in a given organism. Thus, the available pathways from the purines to nucleic acid adenine and guanine are limited by the irreversibility of certain reactions and the inherent complement of interconversion mechanisms. Organisms which cannot synthesize purines from elementary precursors may vary widely in their ability to utilize adenine, guanine, or other purines (xanthine, hypoxanthine, and even unnatural purines such as 2, 6-diaminopurine and 6-mercaptopurine) as sources of polynucleotide purines. It is also a relatively universal finding that systems capable of the synthesis of purines preferentially utilize preformed purines whenever they are available. In so doing, the synthetic pathway is usually curtailed via feedback mechanisms. Certain systems present a remarkable flexibility in the utilization of purines for nucleic acid synthesis whereas others are quite limited in their capacity. Distinct differences between various biological systems with regard to interconversion mechanisms offer a fruitful rationale for chemotherapy which is verified by the inhibitory action of several purine antagonists, particularly against cancer systems. It is obvious, however, that all of the detailed differences in purine metabolism between various species have yet to be elucidated. Differences in susceptibility of various leukemic cell types to antipurines and de novo inhibitors (217) serve as examples. Marked differences in susceptibility of tumors to anticancer agents may reflect metabolic variations which have yet to be elucidated. For example, spontaneous tumors (mouse mammary carcinoma) were found to be virtually unaffected by agents which exhibited significant effects on transplanted tumors, suggesting that cancers which arise under more or less natural environmental conditions may bear little metabolic resemblance to transplantable tumors ( ). Further investigation is necessary before it can be concluded that all living forms synthesize purines or interconvert preformed purines via identical metabolic pathways. The particular metabolic maze by which a given organism may form nucleic acid purines is governed not only by the direct pathways involved, but by the presence or absence of other metabolic cycles which may either compete with the purine pathway or direct alternate sources of precursors into it. The quest for chemotherapeutic agents against cancer has stimulated much of the recent interest in elucidating the physiological interactions which affect nucleic acid synthesis. This impetus has resulted in the rapid development of our knowledge of this subject to a point where metabolic pathways may be charted which were completely unknown a short time ago. Microorganisms have played an important part in the demonstration of these pathways and we now have a much clearer picture of how purines are synthesized and how preformed purines are utilized for nucleic acid synthesis. Microorganisms also represent excellent tools for the investigation of many of the metabolic subtleties which still remain to be elucidated. VI. ACKNOWLEDGMENTS The authors wish to express their gratitude to Dr. Joseph S. Gots, Department of Microbiology, University of Pennsylvania School of Medicine, for his constructive comments and suggestions regarding the subject of purine biosynthesis and interconversion, and to Miss Nancy Lloyd and Mrs. Barbara Keller for secretarial and clerical services. VII. REFERENCES 1. AARONSON, S The purine requirement of Staphylococcus flavocyaneus. J. Gen. Microbiol., 12,

23 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES AARONSON S. AND NATHAN, H. A Utilization of imidazole counterparts of purines in microbial systems. Biochim. et Biophys. Acta, 15, AARONSON, S. AND RODRIGUEZ, E Utilization of purines and their derivatives by Gaffkya homari. J. Bacteriol., 74, ABRAMS, R. AND BENTLEY, M Biosynthesis of nucleic acid purines. II. Role of hypoxanthine and xanthine compounds. Arch. Biochem. Biophys., 58, ABRAMS, R. AND BENTLEY, M Transformation of inosinic acid to adenylic acid and guanylic acid in a soluble enzyme system. J. Am. Chem. Soc., 77, ABRAMS, R. AND BENTLEY, M Biosynthesis of nucleic acid purines. III. Guanosine-5'-phosphate formation form xanthosine-5'-phosphate and L-glutamine. Arch. Biochem. Biophys., 79, ABRAMS, R., HAMMARSTEN, E., AND SHEMIN, D Glycine as a precursor of purines in yeast. J. Biol. Chem., 173, BALIS, M. E., BROOKE, M. S., BROWN, G. B., AND MAGASANIK, B The utilization of purines by purineless mutants of Aerobacter aerogenes. J. Biol. Chem., 219, BALIS, M. E., HYLIN, V., COULTAS, M. K., AND HUTCHISON, D. J Metabolism of resistant mutants of Streptococcus faecalis. III. The action of 6-mercaptopurine. Cancer Research, 18, BALIS, M. E., LARK, C. T., AND LUZzATI, D Nucleotide utilization by Escherichia coli. J. Biol. Chem., 212, BALIS, M. E., LEVIN, D. H., AND LUZZATI, D A purine-histidine relationship in E. coli. J. Biol. Chem., 216, BALIS, M. E., VAN PRAAG, D., AND BROWN, G. B Studies on the metabolism of human tumors. I. Pentosenucleic acid synthesis in tumor-bearing hamsters. Cancer Research, 15, BALLIO, A. AND SERLUPI-CRESCENZI, G Isolation of adenylsuccinic acid from Penicillium chrysogenum. Nature, 179, BEN-ISHAI, R., BERGMAN, E. D., AND VOL- CANI, B. E Ribosidation of 4- amino-imidazole-5-carboxamide by Escherichia coli. Nature, 168, BEN-ISHAI, R., VOLCANI, B. E., AND BERG- MAN, E. D The synthesis of the purine nucleus by Escherichia coli, a study of the mode of action of sulfa drugs. Experientia, 7, BEN-ISHAI, R., VOLCANI, B., AND BERG- MAN, E. D Amino-imidazole-5- carboxamide, precursor of the purines in Escherichia coli. Arch. Biochem. Biophys., 32, BENNETT, L. L., JR., SCHABEL, F. M., JR., AND SKIPPER, H. E Studies on the mode of action of azaserine. Arch. Biochem. Biophys., 64, BENNETT, L. L., JR., AND SKIPPER, H. E Searches for exploitable biochemical differences between normal and cancer cells. III. Effects of aminoimidazolecarboxamide on purine metabolism. Cancer Research? 17, BENNETT, L. L., JR., AND SKIPPER, H. E Conservation of purines in neoplastic tissue. Proc. Am. Assoc. Cancer Research, 3, BENNETT, L. L., JR., SKIPPER, H. E., STOCK, C. C., AND RHOADS, C. P Searches for exploitable biochemical differences between normal and cancer cells. I. Nucleic acid purine metabolism in animal neoplasms. Cancer Research, 15, BENNETT, L. L., JR., SKIPPER, H. E., TOOLAN, H. W., AND RHOADS, C. P Searches for exploitable biochemical differences between normal and cancer cells. II. Nucleic acid purine metabolism in human tumors. Cancer Research, 16, BERGEL, F., BRAY, R. C., HADDOW, A., AND LEWIN, I Enzymatic control of purines by xanthine oxidase. In Ciba Foundation symposium on the chemistry and biology of purines, pp Edited by G. E. W. Wolstenholme and C. M. O'Connor. Little, Brown & Co., Boston. 23. BROCKMAN, R. W., SPARKS, M. C., AND SIMPSON, M. S A comparison of the metabolism of purines and purine analogs by susceptible and drug-resistant bacterial and neoplastic cells. Biochim. et Biophys. Acta, 26, BRoQUIST, H. P. AND SNELL, E. E Studies on the mechanism of histidine synthesis in lactic acid bacteria. J. Biol. Chem., 180, BROWN, G. B Chemical pathways of nucleic acid biosynthesis. Federation Proc., 15, BROWN, G. B. AND ROLL, P. M Biosynthesis of nucleic acids. In The nucleic acids, vol. 2, pp Edited by E. Chargaff and J. M. Davidson. Academic Press, Inc., New York.

24 332 MOAT AND FRIEDMAN [VOL BUCHANAN, J. M The effect of azaserine and 6-diazo-5-oxo-L-norleucine on the biosynthesis of inosinic acid de novo. Texas Repts. Biol. and Med., 15, BUCHANAN, J. M., FLAKS, J. G., HARTMAN, S. C., LEVENBERG, B., LUKENS, L. N., AND WARREN, L The enzymatic synthesis of inosinic acid de novo. In Ciba Foundation symposium on the chemistry and biology of purines, pp Edited by G. E. W. Wolstenholme, and C. M. O'Connor. Little, Brown & Co., Boston. 29. BUCHANAN, J. M., LEVENBERG, B., FLAKS, J. G., AND GLADNER, J. A Interrelationships of amino acid metabolism with purine biosynthesis. In A symposium on amino acid metabolism, pp Edited by W. D. McElroy and B. Glass. Johns Hopkins Press, Baltimore. 30. BUCHANAN, J. M., LEVENBERG, B., MELNICK, D., AND HARTMAN, S. C The specific action of azaserine on enzymes concerned with purine biosynthesis In The leukemias: etiology, pathophysiology, and treatment. Henry Ford Hospital international symposium, pp Academic Press, Inc., New York. 31. CARLO, P. AND MANDEL, H. G The effect of 4-amino-5-imidazole-carboxamide on the toxicity of 8-azaguanine. Cancer Research, 14, CAPUTTO, R The enzymatic synthesis of adenylic acid: adenosinekinase. J. Biol. Chem., 189, CARTER, C. E Metabolism of purines and pyrimidines. Ann. Rev. Biochem., 25, CARTER, C. E. AND COHEN, L. H Enzymatic synthesis of adenylosuccinic acid. J. Am. Chem. Soc., 77, CARTER, C. E. AND COHEN, L. H The preparation and properties of adenylosuccinase and adenylosuccinic acid. J. Biol. Chem., 222, CHAMBERLAIN, N., CUTTS, N. S., AND RAIN- BOW, C The formation of pigment and arylamine by yeasts. J. Gen. Microbiol., 7, CHAMBERLAIN, N. AND RAINBOW, C The formation of diazotizable amine and hypoxanthine by a yeast: possible implications in the biosynthesis of purines. J. Gen. Microbiol., 11, CHRISTMAN, A. A Purine and pyrimidine metabolism. Physiol. Revs., 32, COLLIER, H. 0. J. AND HUSKINSON, P. L The effects of potential antipurines on a purine-requiring strain of Escherichia coli. In Ciba Foundation symposium on the chemistry and biology of purines, pp Edited by G. E. W. Wolstenholme and C. M. O'Connor. Little, Brown & Co., Boston. 40. DALBY, A. AND HOLDSWORTH, E Growth factors for Corynebacterium diphtheriae strain Dundee. J. Gen. Microbiol., 15, DANCIS, J. AND BALIS, M. E The reutilization of nucleic acid catabolites. J. Biol. Chem., 207, DAVIDSON, J. K Mechanism of action of 6-mercaptopurine in L-1210 leukemia. Proc. Am. Assoc. Cancer Research, 3, DE LAMIRANDE, G., ALLARD, C., AND CAN- TARO, A Purine-metabolizing enzymes in normal rat liver and Novikoff hepatoma. Cancer Research, 18, Di CARLO, F. J., SCHULTZ, A. S., AND FISHER, R. A Yeast nucleic acid. III. The effect of glycine on yeast proliferation and nucleic acid biosynthesis. Arch. Biochem. Biophys., 27, DRYSDALE, G. R., PLAUT, G. W. E., AND LARDY, H. A The relation of folic acid to formate metabolism in the rat: formate incorporation into purines. J. Biol. Chem., 193, EDMONDS, M., DELLUVA, A. M., AND WILSON, D. W The metabolism of purines and pyrimidines by growing yeast. J. Biol. Chem., 187, EICHEL, H. J Purine-metabolizing enzymes of Tetrahymena pyriformis. J. Biol. Chem., 220, ELION, G. B. AND HITCHINGS, G. H Biochemical effects of 6-mercaptopurine. In Ciba Foundation Symposium on the chemistry and biology of purines, pp Edited by G. E. W. Wolstenholme and C. M. O'Connor. Little, Brown & Co., Boston. 49. ELION, G. B., SINGER, S., AND HITCHINGS, G. H The purine-metabolism of a 6-mercaptopurine-resistant Lactobacillus casei. J. Biol. Chem., 204, ELION, G. B., SINGER, S., AND HITCHINGS, G. H Antagonists of nucleic acid derivatives; synergism in combinations of biochemically related antimetabolites. J. Biol. Chem., 208, ELION, G. B., SINGER, S., AND HITCHINGS, G. II Microbial effects of 6-mercaptopurine. Ann. New York Acad. Sci., 60, ELION, G. B., SINGER, S., HITCHINGS, G. H.,

25 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES 333 BALIS, M. E., AND BROWN, G. B Effects of purine antagonists on a diaminopurine-resistant strain of Lactobacillus casei. J. Biol. Chem., 202, FLAKS, J. G., ERWIN, M. J., AND BUCHANAN, J. M Biosynthesis of the purines. XVI. The synthesis of adenosine 5'-phosphate and 5-amino-4-imidazolecarboxamide ribotide by a nucleotide pyrophosphorylase. J. Biol. Chem., 228, FLAKS, J. G., ERWIN, M. J., AND BUCHANAN, J. M Biosynthesis of the purines. XVIII. 5-Amino-l-ribosyl-4-imidazolecarboxamide 5'-phosphate transformylase and inosuccinase. J. Biol. Chem., 229, FLAKS, J. G., WARREN, L., AND BUCHANAN, J. M Biosynthesis of the purines. XVII. Further studies of the inosinic acid transformylase system. J. Biol. Chem., 228, Fox, C. L., JR Production of a diazotizable substance by Escherichia coli during sulfonamide bacteriostasis. Proc. Soc. Exptl. Biol. Med., 51, FRIEDMAN, H. AND MOAT, A. G A comparison of nutritional and genetic blocks in the synthesis of purines by yeasts, molds, and bacteria. Arch. Biochem. Biophys., 78, FRIEDMAN, S. AND GOTS, J. S Deamination of isoguanine by Escherichia coli. Arch. Biochem. Biophys., 32, FRIEDMAN, S. AND GOTS, J. S The purine and pyrimidine metabolism of normal and phage-infected Escherichia coli. J. Biol. Chem., 201, FRIES, N Further studies on mutant strains of Ophiostoma which require guanine. J. Biol. Chem., 200, GILES, N. H., PARTRIDGE, C. W. H., AND NELSON, N. J The genetic control of adenylsuccinase in Neurospora crassa. Proc. Natl. Acad. Sci. U. S., 43, GOLDSTEIN, J. AND GOTS, J. S Histidine as a requirement for feed back inhibition of purine synthesis in Escherichia coli. Bacteriol. Proc., 1959, GOLDTHWAIT, D. A Phosphoribosylamine, a precursor of glycinamide ribotide. J. Biol. Chem., 222, GOLDTHWAIT, D. A Purine nucleotide biosynthesis and neoplasia. In The leukemias: etiology, pathophysiology, and treatment. Henry Ford International Symposium, pp Academic Press, Inc., New York. 65. GOLDTHWAIT, D. A. AND BENDICH, A Effects of a folic acid antagonist on nucleic acid metabolism. J. Biol. Chem., 196, GOLDTHWAIT, D. A., PEABODY, R. A., AND GREENBERG, G. R The biosynthesis of the purine ring. In Amino acid metabolism, pp Edited by W. D. McElroy and B. Glass. Johns Hopkins Press, Baltimore. 67. GOLDTHWAIT, D. A., PEABODY, R. A., AND GREENBERG, G. R On the occurrence of glycinamide ribotide and its formyl derivative. J. Biol. Chem., 221, GOLLUB, E. G. AND GOTS, J. S Purine metabolism in bacteria. VI. Accumulations by mutants lacking adenylosuccinase. J. Bacteriol., 78, GOTS, J. S The accumulation of 4- amino-5-imidazolecarboxamide by a purine-requiring mutant of Escherichia coli. Arch. Biochem. Biophys., 29, GOTS, J. S Occurrence of 4-amino-5- imidazolecarboxamide as a pentose derivative. Nature, 172, GOTS, J. S Purine metabolism in bacteria. V. Feed-back inhibition. J. Biol. Chem., 228, GOTS, J. S. AND CHU, E. C Studies on purine metabolism in bacteria. I. The role of p-aminobenzoic acid. J. Bacteriol., 64, GOTS, J. S. AND GOLLUB, E. G Purine metabolism in bacteria. IV. L-Azaserine as an inhibitor. J. Bacteriol., 72, GOTS, J. S. AND GOLLUB, E. G Sequential blockade in adenine biosynthesis by genetic loss of an apparent difunctional deacylase. Proc. Natl. Acad. Sci. U. S., 43, GOTS, J. S. AND LOVE, S. H Purine metabolism in bacteria. II. Factors influencing biosynthesis of 4-amino-5-imidazolecarboxamide by Escherichia coli. J. Biol. Chem., 210, GREENBERG, G. R Role of folic acid derivatives in purine biosynthesis. Federation Proc., 13, GREENBERG, G. R. AND JAENICKE, L On the activation of the one-carbon unit for the biosynthesis of purine nucleotides. In Ciba Foundation symposium on the chemistry and biology of purines, pp Edited by G. E. W. Wolstenholme and C. M. O'Connor. Little. Brown & Co., Boston. 78. GREENBERG, G. R. AND SPILMAN, E Isolation of 5-amino-4-imidazolecarbox-

26 334 MOAT AND) FRIEDMAN [VOL. 24 amide riboside. J. Biol. Chem., 219, GREENLEES, J. A. AND LEPAGE, G. A Purine biosynthesis and inhibitors in ascites cell tumors. Cancer Research, 16, GUTHRIE, R Studies of a purinerequiring mutant strain of Escherichia coli. J. Bacteriol., 57, HARTMAN, S. C. AND BUCHANAN, J. M Biosynthesis of the purines. XXI. 5- Phosphoribosylpyrophosphate amidotransferase. J. Biol. Chem., 233, HARTMAN, S. C. AND BUCHANAN, J. M Biosynthesis of the purines. XXII. 2- Amino - N - ribosylacetamide - 5' - phosphate kinosynthase. J. Biol. Chem., 233, HARTMAN, S. C. AND BUCHANAN, J. M Nucleic acids, purines, pyrimidines (nucleotide synthesis). Ann. Rev. Biochem., 28, HARTMAN, S. C., LEVENBERG, G., AND Bu- CHANAN, J. M Biosynthesis of the purines. XI. Structure, enzymatic synthesis, and metabolism of glycinamide ribotide and (a-n-formyl) -glycinamide ribotide. J. Biol. Chem., 221, HENDERSON, J. F. AND LEPAGE, G. A Utilization of host purines by transplanted tumors. Cancer Research, 19, HEPPEL, L. A Nucleotides and nucleosides. In Chemical pathways of metabolism, vol. II, pp Edited by D. M. Greenberg. Academic Press, Inc., New York. 87. HEPPEL, L. A., HURWITZ, J., AND HORECKER, B. L Adenine deaminase of Azotobacter vinelandii. J. Am. Chem. Soc., 79, HEPPEL, L. A. AND RABINOWITZ, J. C Enzymology of nucleic acids, purines, and pyrimidines. Ann. Rev. Biochem., 27, HURWITZ, J., HEPPEi,, L. A., AND HORECKER, B. L Enzymic cleavage of adenylic acid to adenine and ribose-5-phosphate. J. Biol. Chem., 226, HUTCHISON, D. J Metabolism of resistant mutants of Streptococcus faecalis. I. Isolation and characterization of the mutants. Cancer Research, 18, JAENICKE, L Occurrence of N'0- formyltetrahydrofolic acid and its general involvement in transformylation. Biochim. et Biophys. Acta, 17, JOKLIK, W. K The occurrence of adenine- and adenvl-suceinic acid in mammalian liver. Biochim. et Biophys. Acta, 22, KALCKAR, H. M The enzymes of nucleoside metabolism. Fortschr. Chem. org. Naturstoffe, 9, KERR, S. F. AND CHERNIGOY, F On the biosynthesis of ribonucleic acid purines and their interconversion in yeast. J. Biol. Chem., 200, KOCH, A. L The kinetics of glycine incorporation by Escherichia coli. J. Biol. Chem., 217, KOCH, A. L The metabolism of methyl purines by Escherichia coli. I. Tracer studies. J. Biol. Chem., 219, KOCH, A. L., PUTNAM, F. W., AND EvANS, E. A., JR The purine metabolism of Escherichia coli. J. Biol. Chem., 197, KORN, E. D., REMY, C. N., WASILEJKo, H. C., AND BUCHANAN, J. M Biosynthiesis of the purines. VII. Synthesis of nucleotides from bases by partially purified enzymes. J. Biol. Chem., 217, KORNBERG, A., LIEBERMAN, I., AND SIMMS. E. S Enzymatic synthesis and properties of 5-phosphoribosylpyrophosphate. J. Biol. Chem., 215, KORNBERG, A. AND PRICER, W. E., JR Enzymatic phosphorylation of adenosine and 2,6-diaminopurine riboside. J. Biol. Chem., 193, LAGERKVIST, U Biosynthesis of guanosine 5'-phosphate. I. Xanthosine 5'- phosphate as an intermediate. J. Biol. Chem., 233, LAGERKVIST, U Biosynthesis of guanosine 5'-phosphate. II. Amination of xanthosine 5'-phosphate by purified enzyme from pigeon liver. J. Biol. Chem., 233, LAMPEN, J. O. AND JONES, M. J The growth promoting and anti-suilfonamide activity of p-aminobenzoic acid and related compounds for Lactobacillus arabinosus and Streptobacterium plantarwui. J. Biol. Chem., 170, LAMPEN, J Metabolism of nucleic acid components in bacteria. Bacteriol. Revs., 16, LAMPEN, J. 0. AND WANG, T. P The mechanism of action of Lactobacillus pentosus nucleosidase. J. Biol. Chem., 198, LARDY, H. A. AND PEANASKY, R. 1953

27 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES 335 Metabolic functions of biotin. Physiol. Revs., 33, LEE, Y. P '-Adenylic acid deaminase. I. Isolation of the crystalline enzyme from rabbit skeletal muscle. J. Biol. Chem., 227, LEPAGE, G. A Incorporation of thioguanine into tumor nucleic acids. Proc. Am. Assoc. Cancer Research, 3, LEPAGE, G. A., GREENLEES, J., AND FER- NANDES, J. F Studies of nucleic acid synthesis in ascites tumor cells. Ann. N. Y. Acad. Sci., 63, LEPAGE, G. A. AND SARTORELLI, A. C Purine synthesis in ascites tumor cells. Texas Repts. Biol. and Med., 15, LEVENBERG, B. AND BUCHANAN, J. M Biosynthesis of the purines. XII. Structure, enzymatic synthesis and metabolism of 5- aminoimidazole ribotide. J. Biol. Chem., 224, LEVENBERG, B. AND BUCHANAN, J. M Biosynthesis of the purines. XIII. Structure, enzymatic synthesis, and metabolism of (a-n-formyl)-glycinamidine ribotide. J. Biol. Chem., 224, LEVENBERG, B., MELNICK, I., AND BUCHANAN, J. M Biosynthesis of the purines. XV. The effect of aza-l-serine and 6-diazo- 5-oxo-L-norleucine on inosinic acid biosynthesis de novo. J. Biol. Chem., 225, LICHSTEIN, H. C Functions of biotin in enzyme systems. In Vitamins and hormones, vol. IX, pp Academic Press, Inc., New York LIEBERMAN, I Enzymatic synthesis of adenosine-5'-phosphate from inosine-5'- phosphate. J. Biol. Chem., 223, LITTLEFIELD, J. W. AND DUNN, D. B The occurrence and distribution of thymine and three methylated-adenine bases in ribonucleic acids from several sources. Biochem. J., 70, LONES. D. P., RAINBOW, C., AND WOODWARD, J. D A diazotizable amine produced by yeast: its chemical nature and factors affecting its accumulation. J. Gen. Microbiol., 19, LOVE, S. H. AND GOTS, J. S Purine metabolism in bacteria. III. Accumulation of a new pentose-containing arylamine by a purine-requiring mutant of Escherichia coli. J. Biol. Chem., 212, LOVE, S. H Synthesis of purine intermediates by a cell-free extract of Escherichia coli. J. Bacteriol., 72, LUKENS, L. N. AND BUCHANAN, J. M Biosynthesis of the purines. XXIII. The enzymatic synthesis of N-(5-amino-1- ribosyl imidazolylcarbonyl) - L - aspartic acid 5'-phosphate. J. Biol. Chem., 234, (cf., ibid., pp ) LUKENS, L. N. AND HERRINGTON, K. A Enzymatic formation of 6-mercaptopurine ribotide. Biochim. et Biophys. Acta, 24, LUZZATI, D. AND GUTHRIE, R Studies of a purine- or histidine-requiring mutant of Escherichia coli. J. Biol. Chem., 216, LYNEN, F., KNAPPE, V., LORCH, E., JUTTING, G., AND RINGELMAN, E Die biochemische Funktion des Biotins. Angew. Chem., 71, MACLEOD, P. R. AND LARDY, H. A Metabolic functions of biotin. II. The fixation of carbon dioxide by normal and biotin-deficient rats. J. Biol. Chem., 179, McNUTT, W. S., JR The enzymatically catalyzed transfer of the deoxyribosyl group from one purine to another. Biochem. J., 50, McNUTT, W. S., JR Incorporation of the carbon skeleton of adenosine into the purine nucleosides of ribonucleic acid and deoxyribonucleic acid by Neurospora. J. Biol. Chem., 233, MAGASANIK, B Guanine as a source of the nitrogen 1-carbon 2 portion of the imidazole ring of histidine. J. Am. Chem. Soc., 78, MAGASANIK, B Nutrition of bacteria and fungi. Ann. Rev. Microbiol., 11, MAGASANIK, B. AND BOWSER, H. R Pathways of histidine degradation in microorganisms. In Amino acid metabolism, pp Edited by W. D. McElroy, and B. Glass. Johns Hopkins Press, Baltimore MAGASANIK, B., MOYED, H. S., AND GEHR- ING, L. B Enzymes essential for the biosynthesis of nucleic acid guanine. Inosine 5'-phosphate dehydrogenase of Aerobacter aerogenes. J. Biol. Chem., 226, MAGASANIK, B., MOYED, H. S., AND KA- RIBIAN, D Interconversion of purines in the biosynthesis of nucleic acid adenine and guanine and of histidine. J. Am. Chem. Soc., 78, MANDEL, H. G Incorporation of 8-

28 9336 MOAT AND FRIEDMAN[ [VOL. 24 azaguanine and growth inhibition in Bacillus cereus. J. Biol. Chem., 225, MANDEL, H. G., SUGARMAN, G. I., AND APTER, R. A Fractionation studies of Bacillus cereus containing 8-azaguanine. J. Biol. Chem., 255, MANSON, L. A. AND LAMPEN, J The metabolism of hypoxanthine desoxvriboside in animal tissues. J. Biol. Chem., 191, MATTHEWS, R. E. F The biological effects of 8-azapurines. In Ciba Foundation Symposium on the chemistry and biology of purines, pp Edited by G. E. W. Wolstenholme and C. M. O'Connor. Little, Brown & Co. Boston MELNICK, I. AND BUCHANAN, J. M Biosynthesis of the purines. XIV. Conversion of (a-n-formyl)-glycinamide ribotide to (a-n-formyl)-glycinamidine ribotide; purification and requirements of the enzyme system. J. Biol. Chem., 225, MILLER, R. W., LUKENS, L. N., AND BUCHA- NAN, J. M The enzymatic cleavage of 5-amino-4-imidazole-N-succinocarboxamide ribotide. J. Am. Chem. Soc., 79, MILLER, Z., MCINTOSH, N., HANNUM, M Metabolism of 4-amino-5-imidazole carboxamide in vitro. II. Utilization by actively growing tissues. J. Biol. Chem., 225, MILLER, Z. AND WARREN, L Studies on the metabolism of 4-amino-5-imidazolecarboxamide in vitro. I. Utilization by normal tissue preparations. J. Biol. Chem., 205, MITCHELL, H. K. AND MCELROY, W. D Adenosine deaminase from Aspergillus oryzae. Arch. Biochem., 10, MITOMA, C. AND SNELL, E. E The role of purine bases as histidine precursors in Lactobacillus casei. Pro. Natl. Acad. Sci. U. S., 41, MOAT, A. G. AND NASUTI, F The role of biotin in the utilization of aminoimidazole for purine biosynthesis. Federation Proc. 19, MOAT, A. G., WILKINS, C. N., JR., AND FRIED- MAN, H A role for biotin in purine biosynthesis. J. Biol. Chem., 223, MOORE, E. C. AND LEPAGE, G. A In vivo sensitivity of normal and neoplastic mouse tissues to azaserine. Cancer Research, 17, MOORE, E. C. AND LEPAGE, G. A The metabolism of 6-thioguanine in normal and neoplastic tissues. Cancer Research, 18, MOYED, H. S. AND FRIEDMAN, M Altered feedback control; a basis for resistance to an antimetabolite. Federation Proc., 18, M1OYED, H. S. AND MAGASANIK, B The role of purines in histidine biosynthesis. J. Am. Chem. Soc., 79, MOYED, H. S. AND MAGASANIK, B Enzymes essential for the biosynthesis of nucleic acid guanine: xanthosine-5'-phosphate aminase of Aerobacter aerogenes. J. Biol. Chem., 226, NEIDLE, A. AND WAELSCH, H Participation of glutamine in the biosynthesis of histidine. J. Am. Chem. Soc., 78, NIMMO-SMITH, R. H The function of folic acid in the biosynthesis of purine and pyrimidine derivatives. In Ciba Foundation Symposium on the Chemistry and Biology of Pteridines, p J. & A. Churchill, Ltd., London NOGUCHI, T Biosynthesis of nucleic acids. I. The role of 4-amino-5-imidazolecarboxamide as a precursor of polynucleotide purines. Pharmacol. Bull. Japan, 4, NOGUCHI, T Biosynthesis of nucleic acids. II. Role of antimetabolites as inhibitors of polynucleotide biosynthesis. Pharmacol. Bull. Japan, 4, NOGUCHI, T. AND MIURA, Y The incorporation in vitro of 4-amino-5-imidazole carboxamide into the polynucleotides of pigeon liver cells. J. Biol. Chem., 223, PARTRIDGE, C. W. H. AND GILES, N. H Identification of major accumulation products of adenine-specific mutants of Neurospora. Arch. Biochem. Biophys., 67, PATTERSON, A. R. P The formation of 6-mercaptopurine nucleotides in mouse ascites tumor cells. Proc. Am. Assoc. Cancer Research, 3, PATERSON, A. R. P. AND ZBARSKY, S. H In vitro synthesis of purines by rat intestinal mucosa. Biochim. et Biophys. Acta, 18, PEABODY, R. A., GOLDTHWAIT, D. A., AND GREENBERG, G. R The structure of glycinamide ribotide. J. Biol. Chem., 221, PIERCE, J. G. AND LORING, H. S Growth requirements of a purine-deficient

29 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES 337 strain of Neurospora. J. Biol. Chem., 160, POMPER, S Purine-requiring and pyrimidine-requiring mutants of Saccharomyces cerevisiae. J. Bacteriol., 63, RABINOWITZ, J. C Purine fermentation by Clostridium cylindrosporum. III. 4-Amino-5-imidazolecarboxylic acid and 4- aminoimidazole. J. Biol. Chem., 218, RABINOWITZ, J. C. AND BARKER, H. A Purine fermentation by Clostridium cylindrosporum. I. Tracer experiments on the fermentation of guanine. J. Biol. Chem., 218, RABINOWITZ, J. C. AND BARKER, H. A Purine fermentation by Clostridium cylindrosporum. II. Purine transformations. J. Biol. Chem., 218, RABINOWITZ, J. C. AND PRICER, W. E., JR Purine fermentation by Clostridium cylindrosporum. IV. 4-Ureido-5-imidazolecarboxylic acid. J. Biol. Chem., 218, RATNER, S Arginine metabolism and interrelationships between the citric acid and urea cycles. In Amino acid metabolism, pp Edited by W. D. McElroy and B. Glass. Johns Hopkins Press, Baltimore REICHARD, P Biosynthesis of purines and pyrimidines. In The nucleic acids, pp Edited by E. Chargaff and J. N. Davidson. Academic Press, Inc., New York REID, J. C. AND LANDEFELD, M. A Synthesis of purines and methyl groups from carbon atoms of histidine. Arch. Biochem. Biophys., 34, REID, J. C., LANDEFELD, M. A., AND SIMP- SON, J. L Metabolism of L-histidine-2-C14 in the normal and hepatomabearing rat. J. Natl. Cancer Inst., 12, REMY, C. N., REMY, W. T., AND BUCHANAN, J. M Biosynthesis of the purines. VIII. Enzymatic synthesis and utilization of a-5-phosphoribosylpyrophosphate. J. Biol. Chem., 217, REVEL, H. R. B. AND MAGASANIK, B Utilization of the imidazole carbon 2 of histidine for the biosynthesis of purines in bacteria. J. Biol. Chem., 233, RICHERT, D. A., EDWARDS, S., AND WESTER- FELD, W. W On the determination of liver xanthine oxidase and the respiration of the rat. J. Biol. Chem., 181, ROBERTS, R. B., ADELSON, P. H., COWIE, D. B., BOLTON, E. T., AND BRITTEN, R. J Studies of biosynthesis in Escherichia coli. Carnegie Inst. Wash. Publ. No. 607, ROSE, W. C Purine metabolism. Physiol. Revs., 3, ROUSH, A. H. AND BETZ, R. F Purification and properties of trans-n-deoxyribosylase. J. Biol. Chem., 233, ROUSH, A. II. AND DOMNAS, A. J Induced biosynthesis of uricase in yeast. Science, 124, SABLE, H. Z Phosphorylation of ribose and adenosine in yeast extracts. Proc. Soc. Exptl. Biol. Med., 75, SARTORELLI, A. C. AND LEPAGE, G. A The development and biochemical characterization of resistance to azaserine in a TA3 ascites carcinoma. Cancer Research, 18, SARTORELLI, A. C. AND LEPAGE, G. A Inhibition of ascites cell growth by combinations of 6-thioguanine and azaserine. Cancer Research, 18, SARTORELLI, A. C., LEPAGE, G. A., AND MOORE, E. C Metabolic effects of 6-thioguanine. I. Studies on thioguanineresistant and -sensitive Ehrlich ascites cells. Cancer Research, 18, SCHLENK, F Biosynthesis of nucleosides and nucleotides. In The nucleic acids: chemistry and biology. vol. 2, pp Edited by E. Chargaff and J. N. Davidson. Academic Press, Inc., New York SCHMIDT, G Nucleases and enzymes attacking nucleic acid components. In The nucleic acids: chemistry and biology, vol. I, pp Edited by E. Chargaff and J. N. Davidson. Academic Press, Inc., New York SCHOLLER, J. AND BITTNER, J. J Further studies on chemotherapeutic agents in spontaneous mammary adenocarcinomas of mice and in transplants of recent origin. Cancer Research, 18, SCHOLLER, J., BITTNER, J. J., AND PHILLIPS, F. S Chemotherapeutic studies with transplants of spontaneous mammary tumors of mice growing in various hosts. Cancer Research, 17, SCHOLLER, J., PHILLIPS, F. S., STERNBERG, S. S., AND BITTNER, J. J A comparative study of chemotherapeutic agents

30 338 MOAT AND FRIEDMAN [VOL. 24 in spontaneous mammary adenocarcinomas of mice and in transplants of recent origin. Cancer, 9, SCHULMAN, M. P Purines and pyrimidines. In Chemical pathways of metabolism, vol. II, ch. 14. Edited by D. M. Greenberg. Academic Press, Inc., New York SCOTT, J. F., TAFT, E. B., AND LETOURNEAU, N Conservation of nucleic acid purines by Ehrlich ascites tumor cells. Proc. Am. Assoc. Cancer Research, 3, SHEMIN, D Metabolism of glycine. The succinate-glycine cycle. Federation Proc., 15, SHIVE, W B-Vitamins involved in single carbon unit metabolism. Federation Proc., 12, SKEGGS, H. R., DRISCOLL, C. A., TAYLOR, H. N., AND WRIGHT, L. D The nutritional requirements of Lactobacillus bifidus and Lactobacillus leichmannii. J. Bacteriol., 65, SKIPPER, H. E On the mechanism of action of 6-mercaptopurine. Ann. N. Y. Acad. Sci., 60, SKIPPER, H. E., BELL, M. J., AND CHAPMAN, J. B Partial reversal of the antileukemic action of folic acid antagonists by nucleic acids. Cancer, 4, SLOTNICK, I. J. AND SEVAG, M. G An investigation of the natural occurrence of 4-amino-5-imidazolecarboxamide in several strains of Escherichia coli. Arch. Biochem. Biophys., 57, STEPHENSON, M. AND TRIM, A. R The metabolism of adenine compounds by Bacterium coli with a micro-method for the estimation of ribose. Biochem. J., 32, STETTEN, M. R. AND Fox, C. L., JR An amine formed by bacteria during sulfonamide bacteriostasis. J. Biol. Chem., 161, SUTTON, W. B., SCHLENK, F., AND WERKMAN, C. H Glycine as a precursor of bacterial purines. Arch. Biochem. Biophys., 32, TALMAN, E. L., LABBE, R. F., ALDRICH, R. A., AND SEARS, D Porphyrin metabolism. V. The metabolism of purines in experimental porphyria. Arch. Biochem. Biophys., 80, TOMISEK, A. J., KELLY, H. J., REID, M. R., AND SKIPPER, H. E Chromatographic studies of purine metabolism. III. Effects of A-methopterin on formate-c'4 utilization in mice bearing susceptible and dependent L1210 leukemia. Arch. Biochem. Biophys., 78, TOMISEK, A. J., KELLY, H. J., REID, M. R., AND SKIPPER, H. E Chromatographic studies of purine metabolism. IV. Reversal of azaserine inhibition by phenylalanine. Quarterly Progress Report, Southern Research Inst., Birmingham, Ala TOMISEK, A. J., KELLY, H. J., AND SKIPPER, H. E Chromatographic studies of purine metabolism. 1. The effect of azaserine on purine biosynthesis in E. coli using various C'4-labeled precursors. Arch. Biochem. Biophys., 64, TOTTER, J. R. AND BEST, A. N The metabolism of formate-c14 by rabbit bone marrow in vitro. Arch. Biochem. Biophys., 54, TOTTER, J. R., KELLY, B., DAY, P. L., AND EDWARDS, R. R The metabolism of glycine by folic acid-deficient chick liver homogenates. J. Biol. Chem., 186, TOTTER, J. R., VOLKIN, E., AND CARTER, C. E Incorporation of isotopic formate into the nucleotides of ribo- and desoxyribonucleic acids. J. Am. Chem Soc., 73, TSCHUDY, D. P., MARSHALL, M., GRAFF, A., AND GRAFF, S Studies of nitrogen metabolism in human cancer using isotopically labeled L-aspartic acid. Cancer, 11, TYNER, E. P., HEIDELBERGER, C., AND LE- PAGE, G. A In vivo studies on the incorporation of glycine-2-c'4 into proteins and nucleic acid purines. Cancer Research, 12, WAELSCH, H. AND MILLER, H Certain aspects of the enzymatic breakdown of histidine. In Amino acid metabolism, pp Edited by W. D. McElroy and B. Glass. Johns Hopkins Press, Baltimore WAHBA, A. J., RAVEL, J. M., AND SHIVE, W Involvement of aspartic acid in purine biosynthesis. Biochim. et Biophys. Acta, 14, WAHBA, A. J. AND SHIVE, W A role of aspartic acid in purine biosynthesis. J. Biol. Chem., 211, WANG, T. P. AND LAMPEN, J The cleavage of adenosine, cytidine, and xanthosine by Lactobacillus pentosus. J. Biol. Chem., 192, WARREN, L. AND BUCHANAN, J. M Biosynthesis of the purines. XIX. 2-Amino-

31 1960] BIOSYNTHESIS AND INTERCONVERSION OF PURINES 339 N-ribosylacetamide 5'-phosphate (glycinamide ribotide) transformylase. J. Biol. Chem., 229, WARREN, L., FLAKS, J. G., AND BUCHANAN, J. M Biosynthesis of the purines. XX. Integration of enzymatic transformylation reactions. J. Biol. Chem., 229, WEAVER, J. M. AND SHIVE, W Involvement of thymidine in the utilization of 5-amino-4-imidazole-carboxamide. J. Am. Chem. Soc., 75, WEBB, M Aminopterin inhibition in Aerobacter aerogenes: alanine and valine accumulation during the inhibition and their utilization on recovery. Biochem. J., 70, WHITFIELD, P. R Accumulation of adenine-succinic acid by an adenine-requiring mutant of Neurospora crassa. Arch. Biochem. Biophys., 65, WILLIAMS, A. M The in vitro incorporation of nucleic acid precursors into leukemic tumor leukocytes. Proc. Am. Assoc. Cancer Research, 3, WILLIAMS, A. M. AND LEPAGE, G. A Purine metabolism in mouse ascites tumor cells. I. Effect of preformed purines on in vitro incorporation of glycine-2-c14. Cancer Research, 18, WILLIAMS, A. M. AND LEPAGE, G. A Purine metabolism in mouse ascites tumor cells. II. In vitro incorporation of preformed purines into nucleotides and polynucleotides. Cancer Research, 18, WILLIAMS, W. J. AND BUCHANAN, J. M Biosynthesis of the purines. V. Conversion of hypoxanthine to inosinic acid by liver enzymes. J. Biol. Chem., 203, WINZLER, R. J Differences in nucleic acid metabolism between normal and leukemic human leukocytes. Ann. N. Y. Acad. Sci., 75, WOESSNER, J. F., JR., BACHHAWAT, B. K., AND COON, M. J Enzymic activation of carbon dioxide. II. Role of biotin in the carboxylation of,3-hydroxyisovaleryl coenzyme A. J. Biol. Chem., 233, WOOD, R. C. AND STEERS, E Study of the purine metabolism of Staphylococcus aureus. J. Bacteriol., 77, WOODS, D. D Metabolic relations between p-aminobenzoic acid and folic acid in microorganisms. In Ciba Foundation symposium on the chemistry and biology of pteridine, pp Edited by G. E. W. Wolstenholme and M. P. Cameron. Little, Brown & Co., Boston WOOLEY, D. W. AND PRINGLE, R. B Formation of 4-amino-5-carboxamideimidazole during growth of Escherichia coli in the presence of 4-aminopteroyl glutamic acid. J. Am. Chem. Soc., 72, WYNGAARDEN, J. B. AND DUNN, J. T Hydroxyadenine as the intermediate in the oxidation of adenine to 2,8-dihydroxyadenine by xanthine oxidase. Arch. Biochem. Biophys., 70, WYNGAARDEN, J. B. AND ASHTON, D. M The regulation of activity of phosphoribosylpyrophosphate amidotransferase by purine ribonucleotides: a potential feedback control of purine biosynthesis. J. Biol. Chem., 234, YURA, T Genetic studies with bacteria. Evidence of nonidentical alleles in purine-requiring mutants of Salmonella typhimurium. Carnegie Inst. Wash. Publ. No. 612,