ABSTRACT
Nucleoside hydrolases catalyze the cleavage of N-glycosidic bonds in nucleosides, yielding ribose and the respective bases. While nucleoside hydrolase activity has not been detected in mammalian cells, many protozoan parasites rely on nucleoside hydrolase activity for salvage of purines and/or pyrimidines from their hosts. In contrast, uridine phosphorylase is the key enzyme of pyrimidine salvage in mammalian hosts and many other organisms. We show here that the open reading frame (ORF) YDR400w of Saccharomyces cerevisiae carries the gene encoding uridine hydrolase (URH1). Disruption of this gene in a conditionally pyrimidine-auxotrophic S. cerevisiae strain, which is also deficient in uridine kinase (urk1), leads to the inability of the mutant to utilize uridine as the sole source of pyrimidines. Protein extracts of strains overexpressing YDR400w show increased hydrolase activity only with uridine and cytidine, but no activity with inosine, adenosine, guanosine, and thymidine as substrates, demonstrating that ORF YDR400w encodes a uridine-cytidine N-ribohydrolase. Expression of a homologous cDNA from a protozoan parasite (Crithidia fasciculata) in a ura3 urk1 urh1 mutant is sufficient to restore growth on uridine. Growth can also be restored by expression of a human uridine phosphorylase cDNA. Yeast strains expressing protozoan N-ribohydrolases or host phosphorylases could therefore become useful tools in drug screens for specific inhibitors.
Nucleoside hydrolases cleave the N-glycosidic bond in nucleosides, converting them into ribose and the respective bases, which are available as precursors for auxotrophs. Most protozoan parasites are incapable of synthesizing purines de novo and are dependent on purine salvage from the blood of their hosts (20). In contrast, pyrimidine metabolism seems to be less conserved. Some parasitic protozoa (e.g., Toxoplasma gondii) appear to utilize both salvage and de novo synthesis as a source of pyrimidines, and plasmodia seem to rely on de novo synthesis; others, for instance, Trichomonas vaginalis, depend entirely upon salvage (19). Nucleoside hydrolase activity in mammalian cells, which possess an alternative enzymatic activity for the generation of uracil from uridine, has not been detected. The reaction (uridine + Pi ⇔ uracil + ribose-1-phosphate) is catalyzed by uridine phosphorylase (EC 2.4.2.3). This enzyme is frequently elevated in tumors, and inhibitors of uridine phosphorylase are being evaluated in combination with 5-fluorouracil in chemotherapy (26, 34). Differences in the ability to synthesize pyrimidines de novo and differences in salvage pathway enzymes are therefore potential drug targets for treatment of diseases caused by certain parasites and also for cancer therapy.
Three nucleoside hydrolases of protozoan origin have been cloned: an inosine-uridine nucleoside hydrolase from Crithidia fasciculata (17), a purine-specific inosine-adenosine-guanosine nucleoside hydrolase from Trypanosoma brucei subsp. brucei (31), and a nonspecific nucleoside hydrolase from Leishmania major (38).
Uridine nucleosidase (EC 3.2.2.3) activity in Saccharomyces cerevisiae has been identified (6), and the enzyme has been purified to homogeneity (28). A URH1 gene, encoding uridine hydrolase in S. cerevisiae, was proposed by genetic and biochemical approaches (18, 22, 24). Using 5-fluorocytidine as a selective agent, it was shown (24) that cytidine could also be a substrate for Urh1p, contradicting results obtained with the purified enzyme (28).
YDR400w is a predicted open reading frame (ORF) in the S. cerevisiae genome, which shows similarity to the inosine-uridine nucleoside hydrolase of C. fasciculata. To test whether the ORF YDR400w is indeed URH1, we disrupted this gene in a conditionally pyrimidine-requiring (ura3) S. cerevisiae strain which is additionally deficient in urk1 (encoding uridine kinase). The resulting strain was unable to utilize uridine as the sole source of pyrimidines. Using this Uri− phenotype, we tested whether heterologous genes can complement the urh1 defect. Expression of the inosine-uridine nucleoside hydrolase from C. fasciculata (17) or the human uridine phosphorylase (26) restored the Uri+ phenotype. We therefore conclude that ORF YDR400w encodes uridine-cytidine N-ribohydrolase (the URH1 gene) and propose that deletion strains of yeast expressing heterologous uridine ribohydrolases or phosporylases could be used for in vivo screening of drugs which are specific inhibitors of the respective heterologous enzymes.
MATERIALS AND METHODS
Plasmid constructions.Genomic DNA of S. cerevisiae strain AB1380 (4) was employed as a template for all plasmid construction steps using PCR. For construction of the Δurk1::TRP1 deletion plasmid pRM753, the synthetic oligonucleotides URK-UP (5′-TCAGCACGTTCTCGTCATC) and URK-DW (5′-TCTTCGGTCTAGTGATTCTTG) were used to amplify a 2,287-bp DNA fragment containing the URK1 gene. After EcoRI/XbaI digestion of the PCR product, a 2,096-bp fragment was inserted into pBluescript KS(+) cut with EcoRI/XbaI. The resulting plasmid, pRM528, was cut with NcoI and religated to delete an 883-bp fragment internal to the URK1 ORF (pRM554). Finally, a 932-bp EcoRI/SalI fragment of pJJ248 (21) containing the TRP1 gene was treated with Klenow enzyme (fill-in) and inserted into NcoI-linearized and Klenow enzyme-treated pRM554.
For construction of the Δurh1::LEU2 deletion plasmid pRM1165, the two oligonucleotides YDR400-FW (5′-TGAGCTCTGTTCACCACCACGTAA) and YDR400-RV (5′-ACTCGAGCAGAACCTGACCAAAG) were used to amplify a 1,255-bp fragment containing the URH1 gene. After SacI/XhoI digestion of the PCR product, a 1,243-bp fragment was inserted into pBluescript KS(+) cut with SacI/XhoI to give pRM1129. A 430-bp EcoRV/BamHI fragment inside the URH1 ORF was replaced by a 1,990-bp XbaI (Klenow fill-in)/BamHI fragment of pJJ252 (21) containing the LEU2 gene.
For construction of the Δurh1::kanMX plasmid pRM1381, a 1,487-bp EcoRI/BamHI fragment from pFA6a-kanMX6 (41) was inserted into pRM1129 cut with EcoRV/BamHI, resulting in the replacement of a 430-bp fragment inside the URH1 ORF by the kanMX module.
The expression vector pADH-FW (PADH1-BamHI-EcoRI-NdeI-XhoI-TADH1 LEU2 2μm) was constructed starting from the two-hybrid vector pGAD424 (GenBank accession no. U07647 ): the 695-bp HindIII fragment, comprising the GAL4 activation domain and the multiple cloning site, was deleted and replaced by a new multiple cloning site generated by the complementary oligonucleotides 5′-AGCTCGGATCCATCGAATTCCATATGCTCGAGC and 5′-AGCTGCTCGAGCATATGGAATTCGATGGATCCG.
For use as a positive control and to facilitate cloning, plasmid pTK2 was constructed. First, the EcoRI/SmaI fragment (2,541 bp), comprising the URA3 gene and the tetR gene of YIp5 (GenBank accession no. L09157 ), was cloned into pBluescript KS(+) cut with EcoRI/EcoRV to give pTK1. The URA3/tetR fragment of pTK1 was recovered by EcoRI/XhoI digestion, and the 2,566-bp fragment was cloned into pADH-FW cut with EcoRI/XhoI to give pTK2. The following expression vectors were constructed using pTK2: pRM1503, pRM1436, pRM1583, pRM2013, and pTK246.
(i) pRM1503 (PADH1-URH1-ATG1).The oligonucleotides ATG1-URH (AGAGCTCAACAATGGAATCTGCAGATTT) and YDR400-RV (ACTCGAGCAGAACCTGACCAAAG) were used to PCR amplify a 1,207-bp fragment containing the entire YDR400w ORF (378 amino acids), as proposed by the Saccharomyces Genome Database. After XhoI digestion of the PCR product, a 1,201-bp fragment was inserted into the yeast expression vector pTK2 cut with BamHI (blunted by Klenow fill-in) and XhoI.
(ii) pRM1436 (PADH1-URH1-ATG2).Oligonucleotides ATG2-URH (TGA GCTCAACAATGACTGTTAGTAA) and YDR400-RV were used to PCR amplify a 1,093-bp fragment containing the YDR400w ORF starting at an ATG codon, 38 codons downstream of the postulated initiation ATG of YDR400w. After XhoI digestion of the PCR product, a 1,087-bp fragment was inserted into pTK2 cut with BamHI (blunted by Klenow fill-in) and XhoI.
(iii) pRM1583 (PADH1-URH1-ATG1-FS).Using oligonucleotide ATG1-URH, a unique PstI site was introduced downstream of the first ATG, without altering the amino acid sequence. Plasmid pRM1503 was linearized with PstI, protruding 3′ termini were removed by bacteriophage T4 DNA polymerase, and blunt ends were religated. The predicted frameshift mutation between the first and second ATG was confirmed by sequencing.
(iv) pRM2013 (PADH1-CfIUNH).An NcoI (blunted by Klenow fill-in)/BamHI fragment of pET3d-IUNH (kindly provided by V. Schramm) comprising the complete ORF and additional 3′-untranslated region sequences of inosine-uridine nucleoside hydrolase from C. fasciculata (17) was cloned into pBluescript KS(+) cut with SalI (blunted by Klenow fill-in)/BamHI, generating pRM1679. A SalI (blunted by Klenow fill-in)/BamHI fragment of pRM1679 was cloned into pRM1436 cut with SacI (blunted by T4 polymerase)/BamHI to replace the N-terminal part of YDR400w with the IUNH cDNA (pRM1984). Subsequently, the remaining C-terminal part of YDR400w in pRM1984 was deleted by BamHI/XhoI digestion, Klenow fill-in, and religation, generating pRM2013.
(v) pTK246 (PADH1-hUP).A 1.2-kb fragment containing full-length human uridine phosphorylase cDNA (kindly provided by G. Pizzorno) was released from pBS-hUP (26) by an EcoRI/XhoI digestion and cloned into the respective sites of pTK2.
Multiple sequence alignments were performed using CLUSTAL_X (PAM250 residue weight table) (40) and edited with the GenDoc multiple sequence editor and shading utility. The program TargetP (14) was used for the prediction of subcellular location (http://www.cbs.dtu.dk/services/TargetP/ ).
Yeast strains.All the S. cerevisiae strains used in this study were derived from YYM8 (29). The relevant genotypes are shown in Table 1. YYM8 is derived from strain PL3 (33), in which transcription of the URA3 gene can be induced by substances with estrogenic activity via three estrogen-responsive elements in the promoter (3xERE-URA3). Strains harboring Δurk1::TRP1 and Δurh1::LEU2 deletions were constructed by a one-step gene replacement procedure (35) by transforming the corresponding parental strains with the ClaI/NotI fragment of pRM753 and the SacI/DraI fragment of pRM1165, respectively. The Δurh1::kanMX deletion was obtained by transformation with the PCR product with oligonucleotides YDR400-FW and YDR400-RV and pRM1381 as the template. Correct integration of the deletion constructs was confirmed by PCR analysis of genomic DNA isolated from several transformants.
Relevant genotypes of S. cerevisiae strains used in this study
Yeast media.Yeast strains were grown in yeast-peptone-dextrose or in synthetic complete (SC) medium (37), which was simplified by omitting methionine, tyrosine, phenylalanine, glutamic acid, aspartic acid, valine, and serine. Uracil (Fluka 35 94220) (URA), uridine (Sigma U-6381), cytosine (Sigma C-3506), and cytidine (Sigma C-9505) were prepared as 100× aqueous stock solutions (2,000 mg/liter; sterile filtered). Diethylstilbestrol (Sigma D-4628) was dissolved in ethanol. Geneticin-resistant transformants were selected on yeast-peptone-dextrose plates containing 200-mg/liter G418. Yeast strains to be tested for growth on agar plates were grown to exponential growth phase. Identical volumes (6 μl) of diluted cell suspension (A600 = 0.03) were spotted onto the plates.
Colorimetric assay for uridine hydrolase activity.The hydrolysis of uridine, cytidine, thymidine (Sigma T-9250), inosine (Sigma I-4125), guanosine (Sigma G-6752), and adenosine (Sigma O-1890) was monitored at 30°C by the release of reducing sugar from 5 mM nucleoside solutions as described (30). The amount of protein was determined by the dye-binding method of Bradford (Bio-Rad 35 500-0006) with bovine gamma globulin as a standard. Enzymatic activity data are expressed in units (1 U = 1 μmol of nucleoside per min), which are the means of three replicates.
RESULTS
Growth of urk1 mutants on uridine is YDR400w dependent.According to the model of pyrimidine metabolism in yeast (Fig. 1), the hypothetical uridine hydrolase Urh1p should bypass the uridine kinase (23) step and allow growth of urk1-deficient mutants on SC medium containing either uridine or cytosine as the sole source for pyrimidines. To test this model, we deleted the URK1 gene in strain YYM8 (29). This strain displays pyrimidine auxotrophy (ura3-Δ1), unless the second engineered copy of URA3 containing three estrogen-responsive elements in the promoter (trp1::3xERE-URA3) is activated by the human estrogen receptor in the presence of estrogenic substances (33). In strain YYM8 the trans-acting human estrogen receptor is driven by the yeast PGK promoter and provided on a 2μm HIS3 plasmid, YEp90-HEG0. Consequently, diethylstilbestrol (and other estrogenic substances) can restore de novo pyrimidine biosynthesis in all YYM8-derived strains.
Pyrimidine salvage pathway in S. cerevisiae. In the strains used, de novo biosynthesis of pyrimidines is conditionally blocked by a ura3 mutation before UMP. Reactions leading to the conversion of UTP into CTP and further to DNA and RNA synthesis are indicated merely by dashed lines. External uridine is taken up by uridine permease (encoded by FUI1). Uridine can be either directly converted into UMP by uridine kinase (URK1) or first converted into uracil by uridine ribohydrolase (encoded by URH1 = YDR400w) (this study) and then into UMP by uracil phosphoribosyltransferase (encoded by FUR1). Cytidine can be converted into uridine by cytidine deaminase (encoded by CDD1), and cytosine can be converted into uracil by cytosine deaminase (encoded by FCY1), respectively. Uptake of cytosine (and cytidine) is mediated by a cytosine-purine permease (encoded by FCY2).
Strain YZGA439 (Δurk1::TRP1) (Table 1) grows well on uridine and cytosine but is unable to utilize cytidine (Fig. 2). The growth of the urk1 strain on uridine (which is FUR1 dependent [data not shown]) indicates that the S. cerevisiae genome should encode an enzyme (Urh1p) with uridine N-ribohydrolase activity. In view of its sequence similarity (48% similarity and 23% identity [Fig. 3]) to the inosine-uridine nucleoside hydrolase from C. fasciculata (17), we suspected that the ORF YDR400w could be URH1.
Utilization of exogenous pyrimidines by salvage pathway mutants. Diluted suspensions of the following yeast strains were spotted onto plates with SC-HIS-URA supplemented with the indicated pyrimidines (final concentration, 20 mg/liter). Diethylstilbestrol (DES, 1 μg/liter) restores URA3 expression and de novo pyrimidine synthesis in the ura3 strain used (see text for details and Table 1). The strains and relevant genotypes are as follows: YYM8, YZGA439 (Δurk1::TRP1), YZRM12 (Δydr400w::LEU2), YZRM15 (Δurk1::TRP1 Δydr400w:: LEU2), YZRM18 (Δurk1::TRP1 Δydr400w::kanMX), and YZRM22 (p URA3). Note that urk1 ydr400 double mutants are unable to utilize uridine as the sole source of pyrimidines.
Sequence alignment of confirmed nucleoside N-ribohydrolases and homologous predicted proteins. Residues with a proposed role in catalysis are shown as white letters on a black background. An asterisk above such positions indicates involvement in chelating Ca2+ at the active site of Cf-IUNH; + below the sequence indicates involvement in hydrogen bonding to the 2′-, 3′-, and 5′-hydroxyl groups of ribose in Cf-IUNH (see Discussion). Amino acids which are highly conserved (greater than 90%; PAM250 residue weight table) are shaded in dark gray; conservation of greater than 75% is indicated by lighter gray shading. The predicted protein sequences show the following sequence identities when compared with Urh1p: C. fasciculata inosine-uridine nucleoside hydrolase (accession number U43371 ), 23%; T. brucei subsp. brucei purine-specific nucleoside hydrolase (AF017231), 13%; L. major nonspecific nucleoside hydrolase (39), 26%; E. coli Yeik (U000007), Ybek (AE000169), and Yaaf (D10483), 24, 25, and 24%, respectively. Two hypothetical genes from S. pombe (S. pombe 1, Z69795; S. pombe 2, U33010) show 23 and 21% identity, respectively.
To test this hypothesis, we additionally disrupted YDR400w in strain YZGA439. The resulting strains YZRM15 (Δurk1::TRP1 Δydr400w::LEU2) and YZRM18 (isogenic to YZRM15 except Δydr400w::kanMX, to allow the use of the LEU2 marker for complementation experiments) as well as the corresponding single knockout mutants YZGA439 (Δurk1::TRP1) and YZRM12 (Δydr400w::LEU2) were tested for uridine utilization (Fig. 2). In contrast to the single mutants YZGA439 (Δurk1::TRP1) and YZRM12 (Δydr400w::LEU2), in the ura3 background, the urk1 urh1 double mutants YZRM15 and YZRM18 are unable to grow on minimal medium supplemented with 20-mg/liter uridine. The transformants with a disrupted YDR400w ORF show the phenotype predicted for a urh1 mutant with respect to uridine utilization, indicating that YDR400w is indeed URH1. Lack of growth on cytidine is already embodied in the YYM8 strain and does not allow any conclusions on the substrate specificity of Urh1p (see Discussion).
N terminus of Urh1p and substrate specificity.A TBLASTN search was performed to reveal ORFs homologous to URH1 in other organisms. Besides the published sequences of the inosine-uridine nucleoside hydrolase (Cf-IUNH) from C. fasciculata (17), the purine-specific inosine-adenosine-guanosine nucleoside hydrolase from T. brucei subsp. brucei (31), and a nonspecific nucleoside hydrolase from L. major (38), significant similarities are also found in the ORFs of Escherichia coli and Schizosaccharomyces pombe (Fig. 3). The predicted protein sequences show the following similarities when compared with Urh1p: C. fasciculata (accession number U43371 ), 48%; T. brucei subsp. brucei (AF017231), 34%; L. major (38), 47%. The predicted E. coli proteins Yeik (U00007), Ybek (AE000169), and Yaaf (D10483) show 47, 46, and 43% similarity, respectively. These E. coli genes have recently been renamed, and the ribonucleoside hydrolase activity of the encoded proteins has been demonstrated (32). Two hypothetical genes from S. pombe (designated S. pombe 1 [Z69795] and S. pombe 2 [U33010]) show 49 and 44% similarity.
The predicted amino acid sequence of the ydr400w gene product (Urh1p) has an N-terminal extension of about 40 amino acids when it is aligned with other sequences of nucleoside hydrolases (Fig. 3). Yet since a second methionine is present at position 39, we hypothesize that this downstream ATG is the actual translation start. The ORFs, starting at either the first or the second methionine, were cloned into the 2μm-based yeast expression plasmid pADH-FW (PADH1-MCS-TADH1 LEU2). pRM1503 contains the 378-amino-acid ORF as proposed for Urh1p by database analyses; the ORF in pRM1436 lacks the N-terminal extension and starts at the second ATG codon. Both constructs were able to restore the Uri+ phenotype when transformed into YZRM18 (Δurk1::TRP1 Δydr400::kanMX), indicating that the ORF starting at the second ATG is sufficient for activity. In addition, we introduced a frameshift mutation into pRM1503 between the first and second ATG (pRM1583) (see Materials and Methods). The strain YZRM67 containing this plasmid was still capable of growth on uridine and had a uridine hydrolase activity which was 41% of that measured in the control strain (Fig. 4).
Growth of strains with different translation initiation sites in URH1. Growth was on SC-HIS-LEU supplemented with uracil or uridine, respectively. The positive control strain YZRM22 is prototrophic (URA3 gene on plasmid), and the negative control YZRM64 contains the empty vector (ADH1 promoter only). The constructs complementing the urh1 disruption in strains YZRM28 and YZRM24 start at the first or second ATG; in YZRM67 a frameshift mutation (FS) was introduced between ATG1 and ATG2. The relative uridine hydrolase activity of the strains is indicated (100% = 34 mU/mg of protein).
To test the substrate specificity of Urh1p, crude protein extracts of the yeast strains were prepared by glass bead lysis, and the activity was determined by following the formation of reducing sugar from 5 mM nucleoside solutions (30). YZRM24, a strain overexpressing Urh1p in a urh1 urk1 genetic background, and wild-type strain YZRM70 show activity towards both uridine and cytidine, with a preference for uridine. No significantly altered nucleoside hydrolase activity was detected with thymidine, inosine, guanosine, and adenosine. The knockout strain YZRM12 (Δurh1::LEU2) has only very low background activity with all the nucleosides tested (Fig. 5).
Nucleoside hydrolase activity towards different substrates. The activity (in milliunits per milligram of protein) present in strains with a different URH1 copy number was determined (log phase; SC-LEU). The strains used were YZRM12 (Δurh1), YZRM70 (URH1), and YZRM24 (2μm PADH1-URH1).
A nucleoside hydrolase from a protozoan parasite and human uridine phosphorylase can restore the Uri+ phenotype.The nucleoside hydrolase (expressed by Cf-IUNH) from the protozoan parasite C. fasciculata (17) was chosen to test whether expression of a heterologous cDNA can complement the growth defect on uridine caused by a urh1 mutation in a urk1 ura3 background. The insert of plasmid pET3d-IUNH was cloned behind the ADH1 promoter (pRM2013) (see Materials and Methods for details). Plasmid pRM2013 was introduced into strain YZRM18 (urk1 urh1), and several transformants were compared to strains containing the empty vector or a complementing yeast URH1 gene.
Figure 6 shows that the expressed Cf-IUNH is functional in yeast. The ability to utilize uridine as the sole source of pyrimidines is clearly restored, although growth of the strain is significantly slower.
Restoration of growth on uridine by expression of heterologous uridine ribohydrolase and phosphorylase. Starting with yeast cultures with an optical density at 600 nm of 0.1, 4 serial 1:6 dilutions were spotted on SC-URA-HIS-LEU plates supplemented with either uracil or uridine, where only two dilutions are shown. Dilutions 3 and 4 on the uracil-containing plate indicate equal loading of cells. Dilution steps 1 and 2 are shown on the uridine-supplemented plate. See Table 1 and the text for details.
Since uridine can also be converted into uracil by phosphorolytic cleavage, we wanted to test whether the expression of a host uridine phosphorylase cDNA could restore growth on uridine. Therefore, a full-length human cDNA encoding hUP (26) was cloned into the yeast expression vector (pTKD248); this plasmid was again used to transform YZRM18. Figure 6 shows that both the parasite uridine ribohydrolase and the human uridine phosphorylase can functionally replace URH1.
DISCUSSION
According to the pyrimidine salvage pathway model (Fig. 1), the functional URH1 gene in strain YZGA439 (ura3 Δurk1::TRP1) should bypass the urk1 step and allow growth on minimal medium containing either uridine or cytidine as the sole source of pyrimidines. Consequently, the double mutants YZRM15 (Δurk1::TRP1 Δydr400::LEU2) and YZRM18 (Δurk1::TRP1 Δydr400w::kanMX) should be unable to utilize either uridine or cytidine. While the first prediction was fulfilled, we observed that the YDR400w wild-type strain YZGA439 and all other strains derived from YYM8 were unable to grow on cytidine. The results of activity measurements of strains lacking or overexpressing Urh1p clearly demonstrate that the enzyme also has cytidine hydrolase activity. A likely explanation for the inability of our strains to grow on cytidine is its limited uptake. Transport of cytidine is mediated by a purine-cytosine permease, Fcy2p (7), which has a very low affinity (23) towards cytidine (Km = 1.1 × 10−3 M) compared to cytosine (Km = 1.67 × 10−6 M).
A significant amount of structure information is already available about the protozoan enzymes. Cf-IUNH requires the presence of three hydroxyl groups (2′-, 3′-, and 5′-hydroxyls) in the nucleoside substrates for efficient binding and catalysis (30). Residues involved in hydrogen bonding of the three hydroxyls in Cf-IUNH include Asp-10, Asp-14, Asp-15, Asn-39, and Asp242 for the 2′- and 3′-hydroxyls and Asn-160, Glu-166, and Asn-168 for the 5′-hydroxyl (9). All of these residues involved in substrate recognition are conserved in S. cerevisiae ORF YDR400w (Fig. 3). His241 in Cf-IUNH has been proposed to be involved in leaving group activation, which is conserved in YDR400w; the amino acid residues chelating Ca2+ at the active site are also conserved.
A surprising feature of the protein predicted from ORF YDR400w is the unique N-terminal extension. Although the shorter gene product starting at ATG2 is sufficient for enzyme activity, we cannot exclude that the longer form exists and has a biological role. It is not uncommon in baker's yeast that a single gene encodes isoforms of a protein by using different translation start sites. Prominent examples are, for instance, SUC2 (5), CAT2 (11), and LEU4 (2). Most frequently, the alternative ATGs are used for expression of gene products with a dual role in nuclear and mitochondrial DNA metabolism and tRNA synthesis. Examples are CDC9 (42), PIF1 (36) and (acting on tRNAs, for instance) HTS1, MOD5, and TRM1 (8, 13, 16).
In the case of MOD5, transcripts containing both ATGs are detectable, which are used alternatively. In our ADH1 promoter URH1 fusion constructs, the transcription initiation sites are determined by the ADH1 promoter (1). In a similar ADH1-MOD5 promoter fusion, the ATG preference was changed so that the first ATG was used almost exclusively. Introduction of a frameshift mutation between the ATGs resulted in an about 12-fold-smaller amount of Mod5p, starting at the second ATG (39). With a similar construct, we observed only a 2.4-fold reduction of Urh1p activity, indicating that the second ATG of URH1 is used to an unusually high extent.
The unaltered ratio of uridine:cytidine activity does not support the hypothesis that the isoforms have different substrate preferences (data not shown). According to results obtained with the program TargetP (14), it seems also very unlikely that the N-terminal extension functions as a mitochondrial or extracellular targeting signal. We have compared the growth of strains on plates with SC medium lacking histidine and uracil (SC-HIS-URA) containing uridine as the sole source of pyrimidines and glycerol as the sole nonfermentable carbon source. Strains containing only the short version of URH1 (starting at the second ATG) grew as well as other strains (data not shown), evidence against a mitochondrial role for the hypothetical long form. Yet, the modest overexpression (Fig. 5) of the leaderless protein from the ADH1 promoter could suppress a respiration-deficient phenotype, as was observed in other cases (8, 10). The observed growth phenotype also indicates that Urh1p is not secreted (to aid in the salvage of extracellular ribonucleotides): otherwise, cytidine should have been converted to cytosine, which is taken up very well and obviously allows growth (Fig. 2).
Nucleoside hydrolases are key enzymes in purine-pyrimidine salvage in protozoan parasites and are therefore potential targets for antiprotozoan drugs (20). The described urh1 urk1 strains with a conditional ura3 gene may provide a valuable tool for drug screening. The relevance of uridine ribohydrolase as a drug target depends largely on the genetic makeup of the pathogen (20). For instance, trypanosomatids, like their mammalian hosts, are able to synthesize pyrimidine nucleotides de novo and to use the salvage pathway. The differences in the salvage pathway could be important drug targets. Whereas uridine kinase activity plays a prominent role in mammals, this activity was not detected in extracts of trypanosomatids (19). While uridine phosphorylase is present in mammalian cells, nucleosidase activity is the predominant activity in protozoan parasites (19). In the model organism C. fasciculata, over 90% of nucleoside salvage occurs through the inosine uridine nucleoside hydrolase. We could show that the Crithidia enzyme can functionally replace the yeast enzyme, albeit not fully, as the growth of the strain is slower.
Growth on uridine medium that is dependent on the heterologous enzyme could potentially be a valuable tool to screen for specific inhibitors. A similar approach has been successfully employed in the search for inhibitors of dihydrofolate reductase of Cryptosporidium parvum, by making use of a strain with a deletion of the yeast gene that is complemented by the DHFR gene of the protozoan parasite (3).
A strain deficient in urk1 urh1 expressing a ribohydrolase gene of a protozoan parasite could be useful as a primary screening tool to test newly synthesized pyrimidine derivatives in yeast, instead of the use of elaborate and costly in vivo tests with the parasites (27). Also, the finding that expression of a uridine phosphorylase leads to a phenotypic change in yeast could be of interest for parasitology. The anaerobic flagellate Giardia lamblia, for instance, is totally dependent on pyrimidine salvage and utilizes a uridine-thymidine phosphorylase with properties significantly different from those of the host enzyme (25). It has also been proposed that inhibitors of uridine phosphorylase could be used as antischistosomal chemotherapeutic agents, since some inhibitors revealed clear differences between the enzyme from Schistosoma mansoni and from mice (12). Uridine phosphorylase inhibitors are also interesting compounds with respect to cancer therapy strategies employing 5-fluorouracil. The inhibitor benzyloxyacyclouridine, for instance, has been shown to increase the antitumor activity and the cytotoxic activity of 5-fluorouracil against cancer cells (34, 43).
In summary, yeast strains could be employed in the search for new inhibitors of either uridine ribohydrolases or phosphorylases. The specificity of candidate drugs could be tested by activating URA3 transcription (bypassing the need for pyrimidine salvage) with an estrogenic substance. Knockout of the ABC transporter-encoding genes PDR5 and SNQ2 could facilitate uptake of drugs. Furthermore, the strains could also become a useful tool for cloning genes of interest by transformation with a cDNA expression library of the respective protozoan parasite and selection of Uri+ transformants.
ACKNOWLEDGMENTS
This work was supported by a grant from the Austrian Ministry of Agriculture and Forestry (BMLF Zl.24.002/51-IIA/97).
We thank K. Kuchler (Department of Molecular Genetics, Vienna Biocenter, Vienna, Austria) for providing yeast strain YYM8, P. Chambon (Institut de Génétique et de Biologie Moléculaire et Cellulaire, Illkirch, France) for the permission to use YEp90-HEG0, and V. Schramm (Department of Biochemistry, Albert Einstein College of Medicine of Yeshiva University, New York, N.Y.) and G. Pizzorno (Department of Internal Medicine, Yale University School of Medicine, New Haven, Conn.) for providing plasmids pET3d-IUNH and pBS-hUP, respectively. We thank V. Cameron-Mills and G. Wiesenberger for critically reading the manuscript.
FOOTNOTES
- Received 14 September 2001.
- Accepted 2 January 2002.
- Copyright © 2002 American Society for Microbiology
