INTRODUCTION

Yeast cells take up thiamin from the extracellular environment by an active transport system with a pH optimum of 4.5 and a K_m value of 0.18 µM, and thiamin in the cells is concentrated ~10,000-fold over extracellular levels (1). The thiamin transport system is more specific for the chemical structure of the pyrimidine moiety than the thiazole moiety of thiamin (2). Saccharomyces cerevisiae also secretes a thiamin-repressible acid phosphatase encoded by PHO3 gene (3, 4) in a periplasmic space. Since yeast cells cannot take up thiamin phosphate esters, the PHO3 protein with a high affinity for thiamin phosphate esters appears to hydrolyze them before the uptake (5, 6), and then thiamin is converted to thiamin pyrophosphate (TPP)¹ by thiamin pyrophosphokinase (EC 2.7.6.2) encoded by the THI80 gene (7). TPP is an important cofactor in the energy metabolism (8) and is a negative effector of the regulation mechanism of thiamin metabolism in yeast cells (9).

We report here the isolation and characterization of S. cerevisiae THI10 gene and provide evidence that the THI10 gene encodes for a thiamin transport carrier protein. The predicted THI10 protein is highly hydrophobic and shows significant sequence similarities to yeast uracil and allantoin transport proteins. The thiamin transport activity and the thiamin binding activity in the yeast plasma membrane fraction in thi10 null strain cells were restored when the THI10 open reading frame (ORF) was expressed by yeast GAL1 promoter. The regulation of the THI10 gene expression was also investigated.

EXPERIMENTAL PROCEDURES

Table I shows the S. cerevisiae strains used in this study. The thiamin transport mutant was isolated as a strain resistant to pyrithiamin after the chemical mutagenesis with ethyl methanesulfonate as described (10). The E. coli strain DH5 was used to amplify plasmids. The media and the growth conditions for the yeast and bacterial cells were as described previously (7). Glucose in the yeast minimal medium was replaced by 1% galactose for inducing the THI10, YOR071c, and YOR192c transcription from GAL1 promoter. Yeast strains auxotrophic for thiamin as thi2 and thi3 mutants were cultured in the minimal medium with thiamin at a concentration of 10⁸ M, which does not cause thiamin repression.

Table I. Yeast strains used Strain Genotype Source X2180-1A MATa SUC2 mal gal2 CUP1 YGSCa IFO10482 MATa his4-519 gal2 IFOb YPH499 MATa ura3-52 his3-200 leu2-1 trp1-63 ade2-101 lys2-801 Y. Ohyac YPH500 MAT ura3-52 his3-200 leu2-1 trp1-63 ade2-101 lys2-801 Y. Ohyac T50-1B MATa thi10-1 ura3-52 his3-200 leu2-1 trp1-63 lys2-801 This study NKC6 MAT thi10::HIS3 ura3-52 his3-200 leu2-1 trp1-63 ade2-101 lys2-801 This study O58-M5 MATa thi2(pho6) gal4 Y. Oshimad TRS3 MAT thi3 leu2-3,112 gal2 Our stock T36-2A MATa thi80-2 pho2-8 pho4-1 trp1 gal2 Our stock ND10 MATa/MAT thi10-1/thi10::HIS3 ura3-52/ura3-52 his3-200/his3-200 leu2-1/leu2-1 trp1-63/trp1-63 lys2-801/lys2-801 +/ade2-101 This study a Yeast Genetic Stock Center, University of California, Berkeley, CA. b Institute for Fermentation, Osaka, Yodogawa-ku, Osaka 532, Japan. c University of Tokyo. d University of Osaka.

Mating, sporulation, and dissection of yeast cells were carried out according to the standard procedures for yeast genetics (11). The thiamin transport activity on agar plates was detected using the staining method based on the reduction of triphenyltetrazolium chloride that is known to be taken up by yeast cells via the thiamin transport system (12).

A gene library of S. cerevisiae constructed by partial digestion of the genomic DNA with Sau3AI and its ligation with YEp13 (13) at the unique BamHI site was purchased from the American Type Culture Collection (Rockville, MD). Shuttle vector pRS316 (14) was used for expression in yeast as a single-copy vector. To construct the thi10 disrupted mutant, the 4.0-kilobase pair (kb) EcoRI-SphI fragment of pTTS1 (see Fig. 1) containing the THI10 gene was ligated into the 3.1-kb pUC118 vector (15) to yield pUC118-THI10, and a 1.4-kb XbaI-XhoI fragment containing yeast HIS3 gene of pJJ217 (16) was inserted into pUC118-THI10 at the corresponding sites within the THI10 coding sequence. The linear 3.0-kb SpeI-SpeI fragment of the resultant plasmid pUC-thi10::HIS3 was used to transform the haploid yeast YPH500. His⁺ colonies were selected, and gene disruption was confirmed by Southern blotting. To express THI10, YOR071c, and YOR192c genes from the GAL1 (17) promoter, a yeast expression vector pYES2 (Invitrogen) carrying the GAL1 portion of the divergent GAL1/GAL10 promoter region from S. cerevisiae and the yeast CYC1 (18) transcription termination signal sequence, a 2-µm origin of replication was used. For subcloning of THI10, the ORF was amplified by polymerase chain reaction (PCR, Perkin-Elmer Cetus Instruments) from a plasmid pTTS6 (see Fig. 1): the 5-end primer, CTCGAATTCATGAGTTTCGGTAGTAAAGT was immediately upstream of the THI10 ATG with the Ec0RI linker, and the 3-end primer, CTCGCATGCCTAAGCAGCTTTTTCACTGG, was downstream of the THI10 TAG with the SphI linker. For isolation of YOR071c and YOR192c genes by PCR, specific primers targeted to these untranslated regions were designed according to the sequences of the GenBank^TM data base. GTGTTGCTACCGCTATTTGTTCTTCGG (forward, 225 bp upstream from the presumed initiating codon) and TCTCTTGCTATATGGATGGCATAATTAGC (reverse, 259 bp downstream from the stop codon) were used for YOR071c, and ACCATTCACATAAGATAGACTGTTCCGC (forward, 561 bp upstream from the presumed initiating codon) and AATCAATGAAAATATCATCCTCAACAGAAATA (reverse, 252 bp downstream from the stop codon) were used for YOR192c gene. These amplified products from the genome DNA of wild-type yeast strain YPH500 were used as the templates for a second PCR to obtain the following ORFs. Forward primers were immediately upstream of the ATG with the HindIII linker, CTCAAGCTTATGAGCTTCAGTAGTATAGT (YOR071c) and CTCAAGCTTATGAGTTTCGGTACGAGAAT (YOR-192c), and the reverse primers were immediately downstream of the stop codons with the EcoRI linker, CTCGAATTCTCAGACAACTTGTTGAGTGG (YOR071c) and CTCGAATTCTTAGGCAATTTGTTTTTCAC (YOR192c). All PCR products were purified using Wizard^TM PCR Preps (Promega), and the obtained ORFs treated with the appropriate restriction enzymes were inserted at the corresponding sites of pYES2. General methods for DNA and RNA manipulation, blot analysis, and yeast or bacterial transformation were as described elsewhere (7, 15).

The DNA sequence was determined by primer walking using synthesized nucleotides on an automated laser fluorescent DNA sequencer (Pharmacia LKB Biotechnology Inc.) as described (7). The sequencing was performed twice for each primer on both orientations using the purified plasmid pTTS6 (see Fig. 1) as a template. Nucleotide and deduced amino acid sequences were analyzed with programs from a GENETYX software package (Software Development, Tokyo).

The uptake of the radioactive thiamin into the yeast cells was measured as described previously (1). The thiamin binding activity was determined by an equilibrium dialysis against 10⁷ M [¹⁴C]thiamin overnight in the cold as described (19) after subtracting nonspecific thiamin binding performed in the same manner except that 10⁴ M nonlabeled thiamin was added in the outer solution of dialysis containing the labeled thiamin. Yeast plasma membrane was prepared by a sucrose gradient centrifugation method from the crude membrane fraction as previously reported (20). Protein concentration in membrane preparations was measured by the procedure of Markwell et al. (21) using bovine serum albumin as standard.

DNA probes were labeled with [-³²P]dCTP (3,000 Ci/mmol) purchased from ICN Biomedicals, and thiazole-2-[¹⁴C]thiamin hydrochloride (24.3 Ci/mol) was from Amersham International. All other chemicals were purchased from commercial suppliers.

RESULTS

A yeast strain T50-1B resistant to pyrithiamin, a strong thiamin antagonist bearing high affinity for thiamin transport system in yeast, is a recessive mutant defective in thiamin transport activity. T50-1B was transformed with a yeast genomic library and plated on the agar medium without thiamin and leucine. The plates were incubated at 30 °C for 3 days, and a total of 15,000 Leu⁺ colonies were examined for the thiamin transport activity by staining with triphenyltetrazolium chloride. Only one colony was found to have the thiamin transport activity (by turning the colony red), and by the plasmid-curing experiment, the colony was confirmed to have a plasmid complementing the mutation of T50-1B. A plasmid prepared from the transformant was used to transform Escherichia coli DH5 to Amp^r. The plasmid obtained from the Amp^r, pTTS1 (Fig. 1), had an insertion of a 5.2-kb fragment in YEp13. The insert DNA contained an ORF (see below), and thus the gene obtained here was named THI10. Transformant carrying pTTS1 exhibited restoration of the thiamin transport activity (Table II).

Table II. Thiamin transport activity and thiamin binding activity in yeast plasma membrane fraction After the yeast cells were harvested at the early exponential phase in minimal medium without thiamin, the thiamin transport activity and the thiamin binding activity in the plasma membrane fraction were determined as described under "Experimental Procedures." Disrupted NKC6 transformed with pYES2 or pYES2 derivatives was cultured under the same conditions except that 2% glucose in the medium was replaced by 1% galactose and 106 M thiamin. Strain Plasmid Activity Thiamin transport Thiamin binding pmol/106 cells/min pmol/mg protein X2180-1A None 105 50.6 TM1-1 None ndc ntd T50-1B YEp13a nd nt pTTS1 123 nt NKC6 None nd nt pYES2b nd nd pYES2-THI10 82.5 32.6 pYES2-YOR071c 11.3 nt pYES2-YOR192c 33.9 nt a 2-µm-based multicopy vector used to construct the gene library. b Yeast expression vector bearing GAL1 promoter. c nd, not detected. d nt, not tested.

To define the limits of the THI10 gene, various subclones were constructed using the pRS316 vector from the cloned plasmid pTTS1. The yeast transformants carrying the subcloned plasmids were tested for their thiamin transport activities (Fig. 1). Plasmid pTTS6 containing the 2.6-kb SpeI-NdeI fragment complemented the mutation of T50-1B, whereas plasmid pTTS5, constructed by subcloning the 2.3-kb SpeI-XhoI fragment, did not significantly suppress the phenotype of the mutation. Therefore, we concluded that the 2.6-kb SpeI-NdeI region contains the THI10 gene and that the nucleotide sequence of this region was determined using pTTS6 as a DNA template. As shown in Fig. 2, an ORF of 1,794 bp starting at position 548 and ending at 2,341 was found. A putative TATA box was found at positions 404-408, and two CCAAT transcription elements (22) were located at positions 101-105 and 529-533 in the 5-upstream region. The polyadenylation signal AATAAA was found at positions 2,350-2,355, 2,375-2,380, 2,507-2,512, and 2,572-2,577 in the 3-noncoding region. The yeast intron splice signal, TACTAAC (23), was not found. GenBank^TM data base search revealed that the nucleotide sequence of THI10 is completely identical to that of a DNA coding an unknown protein (L8083.2 gene, accession number U19027), which has been mapped to chromosome XII. We also found YOR071c (1,794 bp of ORF, accession number Z74979) and YOR192c (1,797 bp, accession number Z75100) genes, which are both located at S. cerevisiae chromosome XV and code for proteins of unknown function. The nucleotide sequences of both genes match over the entire length of THI10 with no gaps and about 84% sequence identity (data not shown), suggesting that THI10, YOR071c, and YOR192c genes are from a common evolutionary origin.

The predicted protein encoded by the THI10 ORF contains 598 amino acid residues with a calculated molecular weight of 66,903. By comparison with sequences in the SWISS-PROT data base, the predicted THI10 protein was found to have significant sequence similarity to allantoin (DAL4) (24) and uracil transporter (FUR4) (25) of S. cerevisiae over almost the entire region with 30.2 and 27.9% identity, respectively (Fig. 3). However, neither allantoin nor uracil inhibited [¹⁴C]thiamin uptake into yeast cells when each base of 1,000 times the concentration of labeled thiamin was added to the assay mixture for thiamin transport activity (data not shown). Hydropathy analysis of the deduced THI10 protein indicated that it consists of many potential transmembrane stretches, with the NH₂- and COOH-terminal regions being mostly hydrophilic (Fig. 4). According to the algorithm developed by Kyte and Doolittle (26), 10 segments of 19 sequential residues with an average hydropathy above +1.2 would span the membrane. The average hydropathy values of the transmembrane candidates are as follows: II (amino acids 78-96), 1.61; III (amino acids 107-125), 1.37; IV (amino acids 127-145), 1.25; V (amino acids 175-193), 1.45; VI (amino acids 198-216), 2.02; VII (amino acids 278-296), 2.04; VIII (amino acids 330-348), 1.26; X (amino acids 402-420), 1.88; XI (amino acids 448-466), 1.21; XII (amino acids 488-506), 1.89. Although less extended, there are two hydrophobic segments: I (amino acids 53-71) with the value of 1.04 and IX (amino acids 361-379) with the value of 0.82. Like many other transporter proteins, the NH₂ terminus of the THI10 gene product apparently does not contain any structure that resembles a signal peptide. Potential N-linked glycosylation sites, Asn-X-(Ser/Thr) (27), were found at amino acid positions 39, 323, and 390. They were all located in hydrophilic regions at the NH₂ terminus or between two adjacent membrane-spanning regions.

To further confirm the identity of the cloned gene, we disrupted the corresponding gene locus. The HIS3 fragment was inserted within the THI10 coding region, and the disrupted thi10 fragment was introduced into the his3 strain (Fig. 5A). DNA was extracted from a His⁺ transformant strain (NKC6) and simultaneously digested with NdeI and SpeI. The digest was separated by agarose gel electrophoresis and Southern blotted. Fig. 5B shows that the THI10 locus of NKC6 was correctly disrupted (lane 2). NKC6 could not uptake [¹⁴C]thiamin in the transport assay (Table II), indicating that the THI10 protein is indispensable for the thiamin transport system in S. cerevisiae. Hence, we examined whether or not THI10 was the wild-type allele of the original mutation. NKC6 was crossed with T50-1B. The resultant diploid (ND10) was resistant to pyrithiamin, indicating that complementation did not occur. ND10 was then sporulated, and its tetrads were analyzed. The transport activity of all 10 tetrads was <1.0% that of the wild-type level (data not shown). Thus, the original isolate was concluded to be thi10 mutant.

We have previously reported the occurrence of a thiamin binding activity in the plasma membrane of S. cerevisiae (20, 28). The thiamin binding activity is repressed by exogenous thiamin, and the dissociation constant (K_d) value for thiaminis of 0.11 µM is close to the apparent K_m (0.18 µM) of thiamin transport in S. cerevisiae, but the entity of the protein in the cell membranes remains unclear. We investigated, therefore, the effect of the disruption of THI10 gene on the thiamin binding activity in the yeast plasma membrane fraction. As shown in Table II, thi10 null strain completely lost the thiamin binding activity, indicating that the THI10 protein is indispensable for the binding activity as well as the thiamin incorporation. Then we tried to produce the entire THI10 protein in thi10 null strain from the GAL1 promoter. A vector pYES2-THI10 was constructed as described under "Experimental Procedures" and transformed into NKC6, and the transformed cells were cultivated in galactose medium with 10⁶ M thiamin. We found that the plasma membrane fractions prepared from the transformed cells contained high thiamin binding activity, and the cells producing the THI10 protein restored the thiamin transport activity (Table II). These findings suggested that the THI10 protein is located in the plasma membrane of yeast cells and has the recognition site for thiamin. From these results and the similarity of THI10 protein with DAL4 and FUR4 proteins, we concluded that the THI10 gene of S. cerevisiae is a structural gene for a thiamin transport carrier protein.

Although YOR071c and YOR192c genes are extremely similar to the THI10 gene, it is not clear whether both genes are actually expressed and concerned in thiamin transport in yeast cells. Then, we isolated YOR071c and YOR192c genes by PCR from yeast genome DNA and expressed both ORFs from the GAL1 promoter in thi10 null strain in the same way as with TH110 expression. As shown in Table II, NKC6 cells producing YOR071c or YOR192c protein showed the thiamin transport activities of 13.7% and 41.0%, respectively, compared with the cells expressing THI10 protein. These results suggested that both gene products function together in a thiamin transporter complex with THI10 product or that these genes code a thiamin analog transporter bearing a low affinty for thiamin.

The thiamin transport activity of S. cerevisiae is coordinately repressed with the enzymes involved in thiamin metabolism by exogenous thiamin (29). We determined that the uptake and synthesis of thiamin are controlled positively by a regulatory gene, THI3 (10), and negatively by the intracellular TPP level (9). However, a mutant defective in another positive regulatory gene, THI2, associated with the expression of activities of the enzymes involved in thiamin biosynthesis and thiamin-repressible acid phosphatase retains sufficient thiamin transport activity (31). To determine how THI10 expression is regulated, we investigated the effect of thiamin in the medium on the abundance of THI10 mRNA using the wild-type strain as well as the thi2, thi3, and thi80 mutants. Cells were cultured under either low or high concentrations of thiamin, and total RNA was extracted and Northern-blotted using THI10 and URA3 DNAs as probes. As shown in Fig. 6, THI10 mRNA was detected as a single 1.9-kb band, which was almost consistent with the coding region. THI10 mRNA of the wild-type strain grown in low thiamin medium was expressed abundantly, but a high concentration of thiamin almost completely repressed THI10 mRNA (Fig. 6, lanes 1 and 2). The thi3 mutant did not exhibit the THI10 transcript (Fig. 6, lanes 7 and 8). On the other hand, the decrease in the THI10 mRNA level by thiamin was slight in the thi2 mutant defective in thiamin biosynthesis and the thi80 mutant carrying a level of thiamin pyrophosphokinase activity that was about one-tenth of that in the wild-type strain (Fig. 6, lanes 3-6). This effect was thought to be caused by an insufficient increase in the intracellular concentration of the co-repressor, TPP. These results indicated that THI10 expression is controlled at the mRNA level in S. cerevisiae by the TPP signal transduction pathway including the THI3 protein.

8

7

DISCUSSION

We cloned the THI10 gene by complementation of the phenotype of the thiamin transport defective mutant. The following findings strongly suggested that THI10 encodes a thiamin transport carrier protein. The amino acid sequence deduced from the nucleotide sequence of the gene possesses the structural characteristics of an integral membrane protein, and the thiamin transport activity as well as the thiamin binding activity in the plasma membrane fraction of thi10 null strain was restored when the THI10 ORF was expressed by the GAL1 promoter. This is the first report of the gene-encoding thiamin transport carrier protein for any species.

The predicted protein shows significant sequence similarity to allantoin (DAL4) and uracil transporter (FUR4) of S. cerevisiae. Andre (32) recently reported that the transporters of purines, pyrimidines, and derivatives can be classified into two superfamily (FUR and FCY) families. The FUR family contains DAL4, FUR4, YBL042c (uridine permease), and THI10, which is described as the L8083.2 gene in the review. The THI10 protein certainly seems to recognize the pyrimidine moiety of thiamin (2, 33). However, these four gene products appear not to share function. Cooper et al. (34) reported that allantoin accumulation was negligible in the dal4 mutant strain even though it possessed a wild-type FUR4 allele. We demonstrated here that both allantoin and uracil even at high concentrations did not inhibit thiamin uptake into yeast cells. On the other hand, adenosine and cordycepin, an analog of adenosine, share a common transport system with thiamin in growing S. cerevisiae cells (35). Among nucleosides other than adenosine, only cytidine was taken up by the thiamin transport system but to a level that was only 5.9% that of adenosine. It is probable that the THI10 product participates in the uptake of adenosine in S. cerevisiae.

Although the expression of YOR071c or YOR192c gene from GAL1 promoter partially compensated for the deficiency of THI10 protein in yeast cells, the functions of these genes remain unknown. For elucidation of the relationship between these genes and thiamin transport in yeast, it is essential to ascertain whether the disrupted strains of these genes abolish the thiamin transport activity and whether the expression of these genes is controlled by the thiamin regulatory pathway. Recently, Ruml and Silhankova (36) reported a recessive mutation, thp1, leading to complete loss of thiamin uptake in S. cerevisiae that was mapped to chromosome VII. It has not been clear whether the THP1 gene encodes a structural gene of the thiamin transporter or a positive regulatory factor specific for expression of THI10.

The results of Northern blot hybridization indicated that the THI10 gene expression of S. cerevisiae is regulated at the mRNA level by TPP, similar to other genes involved in thiamin metabolism (7, 37). The expression of the PHO3 gene is known to be regulated at the transcriptional level, and we analyzed the region responsible for the transcriptional activation of the PHO3 gene in response to thiamin in the medium (38). The presumed upstream activating sequence, AAA(A/C)TCAA, is found in the 5-upstream noncoding regions of PHO3 (3) and THI6 (37) genes by two copies for each gene. The THI6 gene product is a bifunctional enzyme with thiamin-phosphate pyrophosphorylase (EC 2.5.1.3) and 4-methyl-5--hydroxyethylthiazole kinase (EC 2.7.1.50) activities (30). On the other hand, there is a TTGATTT sequence at positions 238-244 of the THI10 5-upstream noncoding region (Fig. 2). This sequence is an inverse complement of the AAATCAA. Although the physiological significance of these sequences is not clear, a finding that the THI10 does not require the positive factor THI2 for its expression suggests that the regulatory mechanism of THI10 expression is different from those of PHO3, THI6, and THI80. The expressions of these structural genes, except for THI10, are considered to be controlled by both THI2 and THI3 factors. We expect that the detailed analysis of THI10 gene expression will facilitate elucidation of the TPP signal transduction pathway in yeast thiamin metabolism.