Metabolism of Pyrimidines and Pyrimidine Nucleosides by Salmonella typhimurium

Transcription

1 JOURNAL OF BAcTERIOLOGY, Apr. 1972, p Copyright American Society for Microbiology Vol. 110, No. 1 Printed in U.S.A. Metabolism of Pyrimidines and Pyrimidine Nucleosides by Salmonella typhimurium CHRISTOPH F. BECK, JOHN L. INGRAHAM, JAN NEUHARD,1 AND ELISABETH THOMASSEN Department of Bacteriology, University of California, Davis, California Received for publication 24 November 1971 The pathways by which uracil, cytosine, uridine, cytidine, deoxyuridine, and deoxycytidine are metabolized by Salmonella typhimurium are established. The various 5-fluoropyrimidine analogues are shown to exert their toxic effects only after having been converted to the nucleotide level, and these conversions are shown to be catalyzed by the same enzymes which similarly convert the natural substrates. Methods for isolating mutant strains blocked in various steps of metabolism of pyrimidine bases and nucleosides are described. The pathway of de novo biosynthesis of the four essential pyrimidine nucleoside triphosphates by enteric bacteria is now well established (Fig. 1), but the routes by which they are able to utilize exogenously supplied pyrimidine bases and nucleosides are only partially understood (19). These auxiliary or salvage pathways are quite extensive. Some years ago it was established that each of the six pyrimidine bases and nucleosides were equally effective as precursors of the pyrimidine moieties of ribonucleic acid (RNA) and deoxyribonucleic acid (DNA) (6, 14), and it is further known that each of the six bases and nucleosides is able to meet the total pyrimidine demands of a pyrimidine auxotroph. Elucidation of the pyrimidine salvage pathways, in addition to adding to our general knowledge of the metabolism of enteric bacteria, is important from several points of view. Such knowledge is essential to the intelligent selection and use of radioactive compounds for the labeling of nucleic acids and their precursors; it provides a basis for understanding the mechanism of action of pyrimidine analogues, and for isolating mutants (often by positive selection techniques) blocked in the various steps of the salvage pathways. Previously we have reported the chromosomal location of a number of genes encoding enzymes of the salvage pathway (5). In this paper, by using mutant strains, we establish the various pathways by which the six pyrimidine bases and nucleosides are metabolized by Salmonella typhimurium. I Permanent address: Enzyme Division, University Institute of Biological Chemistry B, Copenhagen K, Denmark. MATERIALS AND METHODS Media. The 007 medium (7) was used as a basal salts medium. Glucose was added to 0.1%. Amino acid supplements were added at a final concentration of 50 Ag/ml; and pyrimidines at the following concentrations: uracil and cytosine, 20 ;g/ml; uridine and cytidine, 40 ;g/ml; thymine and thymidine, 20,ug/ml. Plates in which cytidine or cytosine was the total source of nitrogen contained the normal minimal salts medium without inorganic nitrogen but with cytidine or cytosine added at 200,ug/ml. In media in which nucleosides were the total carbon source, glucose was replaced with the nucleoside at 500 jg/ml. Nutrient broth (Difco) was used as an enriched medium. When nucleoside-requiring strains were grown in enriched medium, the nucleoside was added at 40 jig/ml. Mutagenesis and genetics. Mutagenesis employed N-methyl-N-nitroso-N'-nitroguanidine (1). The procedures for penicillin counterselection (25) and for genetic manipulation (5) were as previously described. Bacterial strains. S. typhimurium, strain LT2, was the original parent of all strains used in these experiments. Table 1 lists the strains used and their genotypes. Table 2 lists the selections used for individual mutations. Table 3 verifies the genotypes. Preparation of enzyme extracts. Cells were grown ovemight in 250-ml batches of glucose minimal media containing 0.15% norite-treated, vitamin-free casein hydrolysate (Nutritional Biochemical Corp.) and 50 ug of uridine and cytidine/ml. Uridine and cytidine were not included in cultures grown for assays of cytosine deaminase (cod) and thymidine kinase (tdk). Cells were harvested by centrifugation, washed once in 0.9% NaCl, resuspended in 5 ml of buffer, and sonically treated three times for 20 sec with a Biosonic oscillator. Extracts were centrifuged for 15 min at 17,000 x g at 3 C, and the supernatant liquid was treated with one-fifth volume of 10% streptomycin sulfate. After centrifugation (15 219

2 220 BECK ET AL. J. BACTERIOL. min at 17,000 x g) the extract was dialyzed for 2 hr For each enzyme the same buffer was used during in the cases of cytosine deaminase, thymidine ki- sonic treatment, dialysis, and assay. The following nase, and uridine kinase, and overnight in the cases buffers were used: (i) for cytidine deaminase, uridine of cytidine deaminase, uridine phosphorylase, and phosphorylase, and UMP pyrophosphorylase: 0.05 M uridine monophosphate (UMP) pyrophosphorylase. tris(hydroxymethyl)aminomethane (Tris)-hydrochlodCTP dctp ~ pymrg ' dutp dttp ' pytj dcdp (! -CDP UD d DP dtdp DE NOVO I CMP ( UMP dumpp--dtmp CdR CR cdd UR' UdR TdR C'U 1-I- idr+lp T FIG. 1. Pathways of biosynthesis of pyrimidine nucleotides and of utilization of exogenous pyrimidine bases and nucleosides. Genetic symbols indicate: cdd, cytidine deaminase; cod, cytosine deaminase; pyrg, cytidine triphosphate (CTP) synthetase; pyrh, uridine monophosphate (UMP) kinase (unpublished data); tdk, thymidine kinase; thy, thymidylate synthetase; tpp, thymidine phosphorylase; udk, uridine kinase; udp, uridine phosphorylase; upp, UMP pyrophosphorylase. Numbered reactions refer to functions for which mutations are unknown in Salmonella typhimurium; they are: (1) nucleoside diphosphate reductase; (2) deoxycytidine triphosphate (dctp) deaminase; (3) deoxyuridine triphosphate (dutp) pyrophosphatase. TABLE 1. Strains used Strain Genotype Parent Derivationa JL1019 pyrc1502 LT2 NG KP1074 pyrci502,cdd-9 JL1019 NG, FCdR KP1077 pyrc1502,cod-5 JL1019 Sp, FC KP1091 pyrc1502,cdd-9,cod-8 KP1074 Sp, FC KP1119 pyrc1502,cdd-9,cod-8,udp-8 KP1091 NG KP1124 pyrc1502,cdd-9,cod-8,tpp-1 KP1091 NG KP1125 pyrcl5o2,cdd-9,cod-8,udp-8,upp-17 KP1119 Sp, FU KP1130 pyrcl5o2,cdd-9,cod-8,udp-8,upp-17,udk-6 KP1125 Sp, FUR KP1146 pyrci502,cdd-9,cod-8,udp-8,udk-6 KP1130 Trans JL1193 pyrci502,cdd-9,cod-8,tpp-1,udp-11 KP1124 NG JL1208 pyrcl5o2,cdd-9,cod-8,udk-6,mete338 KP1146 Trans JL1210 pyrc15o2,cdd-9,cod-8,udp-8,udk-6,udc3b KP1 146 NG, FUR JL1215 pyrcl5o2,cdd-9,cod-8,upp-17,mete338 KP1125 Trans JL1211 pyrcl5o2,cdd-9,cod-8,tpp-1,udp-11,argb69 JL1193 Sp, transc JL1221 pyrcl5o2,cdd-9,cod-8,tpp-1,udp-11,argb69,tdk-1 JL1211 Sp, FUdR a Abbreviations used: NG, mutagenesis by N-methyl-N-nitroso-N'-nitroguanidine; FCdR, selected for being resistant to 5-fluorodeoxycytidine; Sp, spontaneous mutant (no mutagenesis); FC, selected for being resistant to 5-fluorocytosine; trans, constructed by transduction; FUR, selected for being resistant to 5-fluorouridine; FUdR, selected for being resistant to 5-fluorodeoxyuridine. For gene designation see legend of Fig. 1. b Genetic designation udc indicates the phenotype of complete inability to utilize uridine as pyrimidine source in a udp- background. c Strain selected in two steps: a thy mutation was introduced by resistance to trimethoprim (24) and then replaced with argb69 by cotransduction.

3 VOL. 110, 1972 PYRIMIDINE AND PYRIMIDINE NUCLEOSIDE METABOLISM TABLE 2. Selection of mutants Counterselection in Gene Required background Resistance to:a presence of:a, b Location on Salmonella Pyrimidine Carbon map (min) source source cdd None FCdRC 68d pyr- CdR cod pyr+ FC 106e pyr- C tdk tpp - FUdRc 53e tpp None TdR 0g dra-' TdR upp- FU + UdRg udc pyr+,udk-,udp- FUR pyr-,udk-,udp- UR udk cdd- FCRc 66'e pyr+,upp- FUR udp None UR 122d pyr-,tpp- UdR upp pyr+ FU 77d pyr- U a 5-Fluoro analogues were added to selection plates at 10 ug/ml; deoxythymidine (TdR) was added at 200,gg/ml. Abbreviations: U, uracil; C, cytosine; UR, uridine; CR, cytosine, UdR, deoxyuridine; CdR, deoxycytidine; the prefix F denotes the 5-fluoro analogue. b Penicillin counterselection as described in text. c Selection done in presence of 20 gg of uracil/ml. d Reference 5. e Unpublished data. ' dra is the gene encoding deoxyribose-5-phosphate aldolase. g Reference 2. ride at ph 7.3, 10 mm MgCl2, and 2 mm 2-mercaptoethanol; (ii) for cytosine deaminase: 0.05 M Trishydrochloride at ph 7.3; (iii) for uridine kinase and thymidine kinase: 0.1 M Tris-hydrochloride at ph 7.8, 10 mm MgCl2, and 2 mm 2-mercaptoethanol. Cytidine deaminase (cdd). Assay was made in a 0.5-ml reaction mixture containing 0.05 M Tris-hydrochloride at ph 7.3, 10 mm MgCl2, 2 mm 2-mercaptoethanol, 1 mm cytidine (or 2'-deoxycytidine), and enzyme. The reaction mixture was incubated at 37 C, and at appropriate times 0.1-ml samples were taken into 0.9 ml of ice-cold 0.5 M perchloric acid. The precipitated protein was removed by centrifugation, and absorbancy of the supematant fluid was determined at 290 nm. A difference of absorbancy of 1.01 (1-cm light path) is equivalent to a change in substrate concentration of 0.1 mm. Cytosine deaminase (cod). Assays were made in 0.5 ml of reaction mixture containing: 0.05 M Trishydrochloride at ph 7.3, 1.0 mm cytosine, and enzyme. The reaction mixture was incubated at 37 C, and at appropriate times 0.1-ml samples were taken into 0.9 ml of ice-cold 0.5 M perchloric acid. The precipitated protein was removed by centrifugation, and absorbancy of the supernatant fluid was read at 282 nm. A difference of absorbancy of 0.73 (1-cm light path) is equivalent to a change in substrate concentration of 0.1 mm. Thymidine kinase (tdk). Assays were made in 50 Aliters of reaction mixture containing: 0.1 M Tris- 221 hydrochloride at ph 7.8, 10 mm MgCl2, 2 mm 2- mercaptoethanol, 3 mm adenosine triphosphate (ATP), 1 mm thymidine-2-14c (1 mci/mmole), 1 mm KCl, and enzyme. The reaction mixture was incubated at 37 C and at appropriate times, 10 laiters of reaction mixture was spotted on thin-layer plates coated with poly(ethylene imine)-cellulose (PEI) (21), and dried with hot air to stop the reaction. Plates were washed with methanol, dried, and developed stepwise in methanol (3 cm) and water (10 cm). The areas where the samples were applied to the plates (containing the phosphorylated products) were cut out and counted in a Packard Tricarb liquid scintillation spectrometer in vials containing 7 ml of toluene base scintillation mixture. Uridine kinase (udk). Assay was based on observations of L. Finch (personal communication) who suggested the use of guanosine triphosphate (GTP). Assays were made in 50 gliters of reaction mixture containing: 0.1 M Tris-hydrochloride at ph 7.8, 10 mm MgCl2, 2 mm 2-mercaptoethanol, 3 mm GTP, 1 mm [2-4C] nucleoside (uridine or cytidine at 1 mci/mmole), 1 mm KCl. Reaction mixture was incubated at 37 C, and at 20 min a 25-gliter sample was taken and spotted to PEI-cellulose thin-layer plates and processed as described for thymidine kinase. Uridine phosphorylase (udp). Assays were made in 1.0 ml of reaction mixture containing: 0.05 M Trishydrochloride at ph 7.3, 10 mm MgCl2, 2 mm 2- mercaptoethanol, 10 mm sodium arsenate, 5 mm uri-

4 222 BECK ET AL. J. BACTERIOL. dine, and enzyme. The reaction mixture was incubated at 37 C, and at appropriate times 0.2-ml samples were taken into 0.8 ml of ice-cold 0.5 M perchloric acid. The precipitated protein was removed by centrifugation; 0.5 ml of the supernatant fluid was mixed with 0.5 ml of the 1.0 M NaOH, and absorbancy was measured at 290 nm. A difference in absorbancy of (1-cm light path) is equivalent to a change in substrate concentration of 0.1 mm. For each assay a parallel reaction was run in the absence of arsenate, and the amount of conversion under these conditions was subtracted to obtain the reported values of uridine phosphorylase activity. UMP pyrophosphorylase (upp). The procedure employed is a modification of that of Molloy and Finch (16). Assays were made in 50 Mliters of reaction mixture containing: 0.05 M Tris-hydrochloride at ph 7.3, 10 mm MgCl2, 2 mm 2-mercaptoethanol, 0.1 mm uracil-2-14c (10 mci/mmole), 0.2 mm Mg phosphoribosyl pyrophosphate, 0.25 M sodium ethylenediaminetetraacetic acid, and enzyme (preincubated for 5 min in 3 mm GTP at 37 C). Reaction mixture was incubated at 37 C, and at appropriate times 10-uliter samples were taken, spotted to PEIcellulose thin-layer plates, and processed as described under thymidine kinase. Calculation of enzyme activities. For uridine kinase, several enzyme concentrations were routinely assayed, and specific activity was computed from the linear portion of a plot relating activity and enzyme concentration. In all other assays, activities were computed from the linear part of plots relating extent of conversion and time. Protein was determined by the method of Lowry et al. (15). TABLE 3. RESULTS AND DISCUSSION Heat sensitivity of de novo biosynthesis. All of the mutant strains used in this study carry a mutation pyrc1502 which confers heat lability to the encoded enzyme, dihydro-orotase (EC ; L-4,5-dihydro-orotate amidohydrolase). Thus, at 42 C, the enzyme is nonfunctional, allowing one to test which metabolic steps are required for the utilization of specific exogenous pyrimidine bases and nucleosides; at 30 C the strains are prototrophic, allowing one to test the toxicity of analogues in the absence of added natural pyrimidines which might otherwise antagonize the action of the analogues. It also allows one to maintain strains in which all routes for the utilization of exogenous pyrimidines have been blocked by mutation. Uracil mboin. Uracil is a reactant for two reactions (Fig. 1) catalyzed by enzymes contained in wild-type S. typhimurium: (i) uridine phosphorylase (EC ; uridine: orthophosphate ribosyltransferase) and (ii) UMP pyrophosphorylase (EC ; UMP: pyrophosphate phosphoribosyl transferase). But only the latter enzyme plays a significant role in synthesizing uracil ribonucleotides from exogenous uracil; the former enzyme (uridine phosphorylase) functions only catabolically. The essentiality of UMP pyrophosphorylase for the synthesis of UMP from exogenous ur- Certification of mutations Enzymatic evidence Wild type Mutant Gene Encoded enzyme Genetic evidence Stan astrain Specific activ- Allele Sain StanSpecific Activ- Allele itya cdd Cytidine deaminase JL cdd-9 JL (CdR) cod Cytosine deaminase JL cod-5 KP cod-8 KP pyrc Dihydro-orotase pyrc1502 Cotransduces with pyrc-7" tdk Thymidine kinase JL tdk-l JL1221 <0.05 tpp Thymidine tpp-l Cotransduces with phosphorylase upp-201 e udk Cytidine kinase KP udk-6 KP Uridine kinase KP udk-6 KP1130 <0.1 udp Uridine JL udp-8 KP1125 < 1 phosphorylase udp-i1 JL1193 < 1 upp UMP: pyrophos- JL upp-17 KP1125 < 0.04 phorylase a Specific activity expressed as nanomoles of substrate utilized per minute per milligram of protein. "This transduction is feasible owing to the heat sensitivity of pyrc Reference 23. itya

5 VOL. 110, 1972 PYRIMIDINE AND PYRIMIDINE NUCLEOSIDE METABOLISM acil can be seen by comparing the growth responses of strains JL1215 and KP1091 (Table 4). The strains differ with respect to uracil metabolism; JL1215 carries a mutation, upp- 17, which renders UMP pyrophosphorylase nonfunctional, whereas KP1091 carries a wildtype allele of the gene. At 42 C, uracil satisfies the pyrimidine requirement of KP1091 but not of JL1215. Thus UMP pyrophosphorylase is essential for the utilization of exogenous uracil; the route through uridine phosphorylase (udp) and uridine kinase (udk) does not function biosynthetically. By comparing the effect of analogues on the growth of these two strains at 30 C (Table 4), we can further conclude that the pyrimidine analogue 5-fluorouracil (FU) is incorporated into the nucleotide pool by the same route as uracil. FU exerts its toxicity by being converted into 5-fluorouracil triphosphate (FUTP) (12), which, in turn, is incorporated into RNA rendering it nonfunctional. Strain JL1215, which lacks UMP pyrophosphorylase, is completely resistant to FU; strain KP1091 which contains it is sensitive. Cytosine metabolism. Previously we have shown (18) that mutant strains which are unable to synthesize cytidine triphosphate (CTP) owing to a genetic block in the conversion of UTP to CTP (pyrg) require exogenous cytidine for growth; the requirement, however, cannot be met by exogenous cytosine, i.e. S. typhimurium is incapable of converting cytosine to either cytidine or cytidine monophosphate (CMP). We concluded from these results that the only route by which cytosine is utilized is through deamination by cytosine deaminase (EC ; cytosine aminohydrolase) to yield uracil. Hence, one would predict that a pyrimidine auxotroph lacking cytosine deaminase could not have its pyrimidine requirement met by cytosine. A comparison of strains JL1019 and KP1077 (Table 4) establishes the validity of this prediction. The mutation cod-5 also confers resistance to the analogue 5-fluorocytosine (FC) but not to FU at 30 C, so we can conclude that FC is metabolized by the same route as cytosine, namely, it is only converted to FU. The enzyme is capable of deaminating cytosine at a sufficiently rapid rate to allow maximal growth rates when cytosine supplies the total pyrimidine demands of the cell, but unlike Escherichia coli (Jan Neuhard, unpublished data) our S. typhimurium LT-2 strains cannot use cytosine as a total nitrogen source. Deoxycytidine metabolism. Enteric bacteria have been shown to lack a deoxycytidine 223 kinase (13, 17), and hence this deoxynucleoside is unable to serve as a direct source of cytosine deoxyribonucleotides. Both E. coli and S. typhimurium contain an inducible cytidine deaminase (EC ; cytidine aminohydrolase) for which deoxycytidine (CdR) appears to be the preferred substrate (26) and will use deoxycytidine as a total source of pyrimidine. Our data establish that the deamination of CdR to deoxyuridine (UdR) is the essential first step in the utilization of CdR as a pyrimidine source. Strain KP1074 which carries a mutation (cdd-9) encoding an inactive cytidine deaminase is unable to use CdR but can use UdR as a total pyrimidine source; JL1019 which produces an active cytidine deaminase can utilize both CdR and UdR as a total source of pyrimidine (Table 4). Strain KP1074, but not JL1019, is resistant to the analogue 5- fluoro-2'-deoxycytidine (FCdR) but is sensitive to 5-fluoro-2'-deoxyuridine (FUdR). Hence we can conclude that for FCdR to be toxic it must first be converted to FUdR and, further, that this conversion is catalyzed by cytidine deaminase. The deamination of CdR by cytidine deaminase is sufficiently rapid to meet the total nitrogen requirements of the cell from the ammonia produced by the reaction. Deoxyuridine metabolism. UdR is known to undergo two reactions in enteric bacteria; the enzyme, thymidine kinase (tdk) (EC ; ATP: thymidine 5'-phosphotransferase), is capable of phosphorylating UdR [as well as deoxythymidine (TdR)] to form dump (20), and the inducible enzyme, thymidine phosphorylase (tpp) (EC ; thymidine: orthophosphate deoxyribosyltransferase), splits UdR (as well as TdR) to form uracil and deoxyribose-1-phosphate. Via the latter reaction, UdR may serve as a total source of pyrimidine (through formation of uracil) and as a total source of carbon and energy (through formation of deoxyribose-1-phosphate). UdR serves as a total source of carbon and energy only if the cell contains an active thymidine phosphorylase (9). We note from Table 4 that strain KP1091 can grow on UdR as a sole source of carbon and energy, whereas strain KP1124, which lacks thymidine phosphorylase (tpp-1), is unable to grow on UdR as a carbon source. We further notice that KP1124 retains its ability to utilize UdR as a pyrimidine source. In confirmation of the recent results of Beacham and Pritchard (4) with E. coli, we found that the residual ability of tpp- strains to utilize UdR as a pyrimidine source is dependent on the activity of the en-

6 224 BECK ET AL. J. BACTERIOL. L.r 09 S < X u I+++I++I+I U~~~~~~~~ >gce Q~~~~~~~~~+ + + e 3, e ~~~~~ >0- Z cz E _ co co,~ U I+I+I i++ E i + + +t +^+ + _ 0.~+ 0~~~~~~~~~~~~1 -AXtt tr t I sr Q c) c; cz~q Q~~~~~~~~~~~~b. o O 8 4,rV- C14 c ts t3 cici ri,0413 'V Q Q cj 10. '- Q ci.~~~ ~ ~ ~ 3-6C-4C-1C4 C -I ~ Ur,s n8hohu O) 0 D" "- w- X ;; P- v- ; r--

7 VOL. 110, 1972 PYRIMIDINE AND PYRIMIDINE NUCLEOSIDE METABOLISM zyme uridine phosphorylase (udp) which apparently has activity for the alternate substrate, UdR. JL1193 which lacks both thymidine and uridine phosphorylases cannot utilize UdR either as a source of carbon or as a source of pyrimidine (Table 4). By comparing the responses of strains blocked in the reactions catalyzed by uridine and thymidine phosphorylases we can further conclude that the UdR analogue, FUdR, exerts its toxicity by two different biochemical mechanisms. (i) It is split to yield FU; (ii) it is phosphorylated to FdUMP which is a competitive inhibitor of thymidylate synthetase (8) and thus causes thymine starvation. Resistance to the former route of toxicity can be obtained by blocking the breakdown of FUdR to FU (introducing mutations in tpp and udp) or alternatively preventing the reutilization of FU (introducing a mutation in upp or adding uracil to the medium which effectively competes with FU and prevents its conversion to FUMP). Starting with a strain (JL1193; tpp-1, udp-11) which lacks the first route leading to FUdR toxicity, we have been able to isolate a strain (JL1221) which is completely resistant to FUdR owing to an additional mutation in the gene encoding thymidine kinase (tdk-1). Subsequently we have found that it is not necessary to block both uridine and thymidine phosphorylases in order to isolate mutants in tdk by resistance to FUdR. Such mutants can be isolated directly in strains producing normal uridine phosphorylase simply by adding exogenous uracil which antagonizes the FU generated by phosphorolytic cleavage of FUdR. However, a block in thymidine phosphorylase is required for this selection even in the presence of exogenous uracil, probably because tpp+ strains phosphorolyze FUdR too rapidly to maintain toxic intracellular concentrations of the analogue. Uridine metabolism. Uridine (UR) undergoes two known reactions in enteric bacteria. It is converted to UMP by uridine kinase (3) (EC ; ATP: uridine 5'-phosphotransferase) and to uracil (U) (22) by the inducible uridine phosphorylase (11). Thus there are two routes by which uridine can give rise to uridine triphosphate (UTP) through UMP: (i) by direct conversion to UMP through the action of uridine kinase; or (ii) by phosphorolysis to yield uracil and subsequent conversion to UMP (UMP pyrophosphorylase). Both of these routes function in vivo. Pyrimidine auxotrophs which lack the first route owing to genetic blocks in udk, the gene encoding uridine kinase, can utilize uridine by the second route to satisfy their requirement for pyrimidine (see strain JL1208, Table 4). Alternatively, strains which are blocked in the second route owing to mutations in genes encoding uridine phosphorylase (udp), UMP pyrophosphorylase (upp), or both, are able to use the first route to supply their pyrimidine requirements (Table 4, strains KP1119, JL1215, and KP1146, respectively). If both routes are blocked by introducing negative mutations in udk, udp, and upp, the strain is unable to use uridine as a pyrimidine source, as is strain KP1130 (Table 4). Having established that two routes for UTP biosynthesis from UR are operative in vivo in S. typhimurium, we are in a position to determine the distribution of flow between these two possible routes. We have done this by studying the growth yield from uridine of various mutants (Fig. 2). Strains which lack UMP pyrophosphorylase (upp) are unable to reutilize uracil which is derived from UR by uridine phosphorylase, and hence, in these strains the UR-to-U-to-UMP route does not contribute to the growth yield. The growth yield, therefore, is a direct measurement of the flow through the UR to UMP route. The growth yield of S [ g / I,_mmm--~ KP-KID KP // _ - ~~~(M-.t1ff) $ - > JL Z0 n5 TIME (MN.) FIG. 2. Growth yield of various mutants growing on uridine as a total source of pyrimidine at 42 C, at which temperature all are pyrimidine auxotrophs. Strain JL1215 lacks uridine monophosphate (UMP) pyrophosphorylase (upp-17). Strain KP1125 lacks both UMP pyrophosphorylase (upp-17) and uridine phosphorylase (udp-8). Strain KP1091 contains both UMP pyrophosphorylase and uridine phosphorylase. All cultures contained 6 jg of uridine/mi as the sole source of pyrimidine.

8 226 BECK ET AL. J. BACTERIOL. strain KP1091 (udk+, udp+, upp+) is a measure of the total utilization of UR (100%). Strain JL1215 carries a mutation in UMP pyrophosphorylase (upp-17) thus precluding the reutilization of uracil. We notice (Fig. 2) that the growth yield of JL1215 from UR is only about 25% of that of KP1091, and we conclude that under these conditions 75% of uridine added is broken down to uracil; i.e., the major route of utilization of UR is through uracil. Blocking the breakdown of UR by introducing a mutation in udp (strain KP1125) might be expected to return the growth yield on UR to 100%. But such is not the case (Fig. 2); rather the growth yield of such a doubly blocked strain is only 80% of wild type, suggesting that, even in a strain lacking uridine phosphorylase, 20% of the uridine is still broken down to uracil. This conclusion is also shown by the fact that KP1146, which lacks both uridine kinase and uridine phosphorylase, is still able to utilize uridine as a pyrimidine source. We have some evidence that there is a second function, in addition to uridine phosphorylase, which is capable of breaking down uridine, and we have isolated mutants (JL1210) starting from a udp- strain (KP1146) which are unable to utilize UR as a pyrimidine source. We do not know what function has been lost by the second mutation (udc); however, we have established by transduction that udc is a different locus from udp. The UR analogue, 5-flurouridine (FUR), is metabolized by the same routes as exist for uridine. Strains lacking either uridine kinase (Table 4, strain JL1208) or UMP pyrophosphorylase (KP1215) remain sensitive to FUR, but strains lacking both of these enzymes (KP1130) and hence both routes of metabolism of UR are resistant to FUR at 30 C. JL1210, which is unable to utilize UR owing to mutation in udp, udk, and udc, is also resistant to FUR. Cytidine metabolism. Cytidine (CR) enters into two enzyme reactions. (i) It is phosphorylated by cytidine kinase to form (CMP) (18); and (ii) it is deaminated by cytidine deaminase to form UR (26). The latter enzyme is quite active in Salmonella as shown by the fact that CR is capable of serving as a total source of nitrogen. We have found that cytidine and uridine kinases are a single protein encoded by a single gene (udk). Two observations lead to this conclusion. (i) Strains carrying mutations in udk are resistant not only to FUR but also to the cytidine analogue 5-fluorocytidine (FCR). (ii) Strains carrying mutations in udk lack both uridine and cytidine kinase activities (Table 3). One might predict that a pyrimidine auxotroph which lacks cytidine deaminase (cdd) would have its pyrimidine requirement only partially satisfied by cytidine; i.e., that cytidine could satisfy the requirement for cytosine nucleotides but not for uracil nucleotides. But such is not the case. KP1074 (pyrc1502, cdd-9) can utilize cytidine as a total source of pyrimidine (Table 4). Clearly, therefore, there is a route from CR to uracil nucleotides which does not go through cytidine deaminase; i.e., there is a second CR-to-UMP route. We can ask if cytosine is an intermediate of this route and the answer is no, because we know (see Cytosine metabolism section) that the only route of metabolism of cytosine is by way of cytosine deaminase (cod), and strains lacking both cytidine and cytosine deaminases are still capable of using cytidine as a total source of pyrimidine (compare strains KP1091 and KP1074, Table 4). The possibility of deamination at the nucleotide level is eliminated by the finding (Table 4) that strains carrying mutations in cdd, cod, and udk (JL1208) are able to utilize cytidine as a total source of pyrimidine. Thus, direct deamination of CR to UR appears to be the only remaining possibility, and we have established that the route, in fact, flows through UR because strain JL1210 (cdd, udp, udc, udk), which is unable to grow on UR, is also unable to grow on CR as a total pyrimidine source (Table 4). Surprisingly, strains lacking both cytidine deaminase and UMP pyrophosphorylase cannot have their entire pyrimidine requirement satisfied by CR (strain JL1215, Table 4). Thus even in a strain lacking cytidine deaminase the route must be: CR -- UR d U - UMP. It is reasonable to ask why uridine kinase is not capable of directly converting UR to UMP under these conditions. An explanation might be that CR, when present in excess, competes effectively with UR for uridine kinase and thereby prevents the direct conversion of UR to UMP. In fact we have shown that the growth of the JL1125 (upp, cdd) pyrimidine auxotroph on UR is inhibited by CR. The enzyme which deaminates CR in strains lacking cytidine deaminase remains unknown. Conclusions. (i) The routes by which exogenous pyrimidine bases and nucleosides are metabolized are summarized in Fig. 1. (ii) Certain enzymes catalyze two reactions as shown in Fig. 1. Uridine kinase phosphorylates both uridine and cytidine. Thymidine kinase phosphorylates both thymidine and

9 VOL. 110, 1972 PYRIMIDINE AND PYRIMIDINE NUCLEOSIDE METABOLISM deoxyuridine. Thymidine phosphorylase splits both thymidine and deoxyuridine. Cytidine deaminase deaminates both cytidine and deoxyctidine. (iii) Certain steps shown in Fig. 1 can be catalyzed by two enzymes. Deoxyuridine is split by thymidine phosphorylase and to a lesser extent by uridine phosphorylase. Uridine can be phosphorolytically cleaved by uridine phosphorylase and can be split by some other unknown activity. Cytidine is converted to uridine by cytidine deaminase and by another unknown activity. (iv) Exogenously added uridine is mainly broken down to uracil before being incorporated into UTP, and hence radioactive uridine is an unnecessarily complicated label and should be avoided in critical experiments. (v) The 5-fluoro analogues FU, FC, FUR, FCR, FUdR, and FCdR are metabolized by the same enzymes which metabolize U, C, UR, CR, UdR, and CdR, respectively. The 5-fluoro analogues are toxic only after having been converted to nucleotides. (vi) FUdR exerts its toxic effect as a consequence of two biochemical routes: by being converted to FdLJMP and by being converted to FU. (vii) The 5-fluoro analogues can be used to select positively for a number of mutations in genes governing pyrimidine metabolism (see Table 2). (viii) Although the conclusions presented here concern S. typhimurium, cumulative evidence indicates they also apply to E. coli with only slight quantitative differences. Similar studies on utilization of exogenous pyrimidine by Saccharomyces cerevisiae have been made by Grenson (10). ACKNOWLEDGMENTS We thank John Roth and Kenneth Sanderson for supplying some of the strains used in this study and Gerard O'Donovan for helpful discussions and for supplying certain chemicals. Most of the 5-fluoropyrimidines were supplied by William Scott of Hoffmann-La Roche, Inc. We thank Marjorie Ingraham for preparing the figures. One of us (J.N.) was a recipient of a NATO travel grant. The investigation was supported by Public Health Service Grant Al from the National Institute of Allergy and Infectious Diseases. LITERATURE CITED 1. Adelberg, E. A., M. Mandel, and G. C. C. Chen Optimal conditions for mutagenesis by N-methyl-N'- nitro-n-nitrosoguanidine in Escherichia coli K-12. Biochem. Biophys. Res. Commun. 18: Ahmad, S. I., and R. H. Pritchard A map.of four genes specifying enzymes involved in catabolism of nucleosides and deoxynucleosides in Escherichia coli. Mol. Gen. Genet. 104: Anderson, E. P., and R. W. Brockman Feedback 227 inhibition of uridine kinase by cytidine triphosphate and uridine triphosphate. Biochim. Biophys. Acta 91: Beacham, I. R., and R. H. Pritchard The role of nucleoside phosphorylase in the degradation of deoxyribonucleosides by thymine-requiring mutants of E. coli. Mol. Gen. Genet. 110: Beck, C. F., and J. L. Ingraham Location on the chromosome of Salmonella typhimurium of genes governing pyrimidine metabolism. Mol. Gen. Genet. 111: Bolton, E. T., and A. M. Reynard Utilization of purine and pyrimidine compounds in nucleic acid synthesis by Escherichia coli. Biochim. Biophys. Acta 13: Clark, D. J., and 0. Maaloe DNA replication and the division cycle in Escherichia coli. J. Mol. Biol. 23: Cohen, S. S., J. G. Flaks, H. D. Barner, M. R. Loeb, and J. Lichtenstein The mode of action of 5-fluorouracil and its derivatives. Proc. Nat. Acad. Sci. U.S.A. 44: Fangman, W. L Specificity and efficiency of thymidine incorporation in Escherichia coli lacking thymidine phosphorylase. J. Bacteriol. 99: Grenson, M The utilization of exogenous pyrimidines and the recycling of uridine-5'-phosphate derivatives in Saccharomyces cerevisiae, as studied by means of mutants affected in pyrimidine uptake and metabolism. Eur. J. Biochem. 11: Hammer-Jespersen, K., A. Munch-Peterson, P. Nygaard, and M. Schwartz Induction of enzymes involved in the catabolism of deoxyribonucleosides and ribonucleosides in Escherichia coli K12. Eur. J. Biochem. 19: Horowitz, J., and E. Chargaff Massive incorporation of 5-fluoro-uracil into a bacterial ribonucleic acid. Nature (London) 184: Karlstrom, H Inability of Escherichia coli B to incorporate added deoxycytidine, deoxyadenosine, and deoxyguanosine into DNA. Eur. J. Biochem. 17: Lichtenstein, J., H. D. Barner, and S. S. Cohen The metabolism of exogenously supplied nucleotides by Escherichia coli. J. Biol. Chem. 235: Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193: Molloy, A., and L. R. Finch Uridine-5'-monophosphate pyrophosphorylase activity from Escherichia coli. Fed. Eur. Biochem. Soc. Lett. 5: Neuhard, J Pyrimidine nucleotide metabolism and pathways of thymidine triphosphate biosynthesis in Salmonella typhimurium. J. Bacteriol. 96: Neuhard, J., and J. Ingraham Mutants of Salmonella typhimurium requiring cytidine for growth. J. Bacteriol. 95: O'Donovan, G. A., and J. Neuhard Pyrimidine metabolism in microorganisms. Bacteriol. Rev. 34: Okazaki, R., and A. Komberg Deoxythymidine kinase of Escherichia coli. I. Purification and some properties of the enzyme. J. Biol. Chem. 239: Randerath, K., and E. Randerath Ion-exchange chromatography of nucleotides on poly(ethylene imine)-cellulose thin layers. J. Chromatogr. 16: Razzell, W. E., and H. G. Khorana Purification and properties of a pyrimidine deoxyriboside phosphorylase from Escherichia coli. Biochim. Biophys. Acta 28: Sanderson, K. E Current linkage map of Salmonella typhimurium. Bacteriol. Rev. 34:

10 228 BECK ET AL. J. BACTERIOL. 24. Stacey, K. A., and E. Simson Improved method for the isolation of thymine-requiring mutants of Escherichia coli. J. Bacteriol. 90: Tai, P-C., D. P. Kessler, and J. L. Ingraham Cold-sensitive mutations in Salmonella typhimurium which affect ribosome synthesis. J. Bacteriol. 97: Wang, T. P Cytosine nucleoside deaminase from Escherichia coli, p In S. P. Colowick and N. 0. Kaplan (ed.), Methods in enzymology, vol. 2. Academic Press Inc., New York.