Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Membrane transporters play a key role in determining exposure of a cell or organism to a variety of solutes, including nutrients, cellular by-products, environmental toxins, drugs, and other xenobiotics. A major challenge in transporter biology is to understand the structural basis for substrate recognition. Knowledge of the structural basis for substrate recognition is not only important in understanding the initial steps of transport but is also critical in designing and using drugs because membrane transporters are major determinants of drug response.

During evolution, as organisms have increased in complexity, transporters have increased in number and diversity of function. Although many studies have compared the sequences of transporters among organisms and highlighted sequence similarities and differences, this phylogenetic analysis has been used primarily to establish evolutionary relationships among transporters. Only a few studies on comparative functional analysis of transporter homologs have identified critical amino acids and structural domains involved in substrate recognition (Barker and Blakely, 1996; Barker et al., 1998). Functional comparisons of transporter homologs among organisms may also aid in understanding structural elements responsible for other processes such as coupling mechanisms involved in active transport.

Nucleoside transporters are integral membrane proteins that mediate both the uptake and release of nucleosides and many synthetic nucleoside analogs across plasma membranes. A major class of nucleoside transporters, concentrative nucleoside transporters (CNTs), mediate the active intracellular influx of nucleosides by coupling to the sodium or proton gradients across the plasma membrane (Baldwin et al., 1999).

Based on substrate selectivity, the concentrative nucleoside transporters can be divided into two major subtypes (Belt et al., 1993). One type (N1 or CNT2) is generally purine selective; the other type (N2 or CNT1) is generally pyrimidine selective. Genes for both CNT1 and CNT2 (also termed SPNT) have been cloned from rat and human; their protein products are predicted to be approximately 65% identical in amino acid sequence (Crawford et al., 1990; Dagnino et al., 1991). Both types accept uridine and adenosine. Mammalian CNT family members usually transport nucleosides by coupling to sodium (Che et al., 1995; Ritzel et al., 1997) whereas nucleoside transport systems in bacteria usually transport nucleosides by coupling to protons (Craig et al., 1994).

There is growing evidence to show that nucleoside transporters with broader substrate selectivities exist in nature. First, in isolated tissues from several mammalian species, a broadly selective nucleoside transporter, N3, which accepts both purine and pyrimidine nucleosides, and N4 and N5, transporters with distinct substrate selectivities, have been characterized (Wu et al., 1992, 1994; Huang et al., 1993). Second, by construction of chimeras, we engineered a broadly selective N3-type nucleoside transporter from rCNT1 and rCNT2 (Wang and Giacomini, 1997, 1999a). An N3-like transporter was also created by changing a single amino acid at position 318 of rCNT1 (Wang and Giacomini, 1999b). Recently, a nucleoside transporter with broad selectivity has been cloned from hagfish (GenBank accession number AF132298).

The free-living nematode Caenorhabditis elegans is a simple model system to study a variety of complex biological problems. Based on its genome sequence (Anonymous, 1998), many transporters in C. elegans have been cloned and characterized, including transporters for dopamine (Jayanthi et al., 1998), glutamate (Kawano et al., 1996), phosphate (Lee et al., 1999), organic anions and cations (George et al., 1999; Wu et al., 1999), monoamines (Duerr et al., 1999) and oligopeptides (Fei et al., 1998). Functional studies of these transporters show that, in C. elegans, some transporters are sodium-dependent (Jayanthi et al., 1998; Lee et al., 1999) whereas some others are proton-dependent (Fei et al., 1998; Duerr et al., 1999; Wu et al., 1999).

Based on the gene F27E11, two CNT transporters were predicted to exist in C. elegans. After multiple sequence alignments and examination of critical amino acids involved in substrate recognition, we hypothesized that one of CNT homologs encoded by gene F27E11 may function as a broadly selective nucleoside transporter. We cloned the cDNA for this transporter, termed CeCNT3, and determined its functional characteristics. CeCNT3 exhibited novel characteristics in terms of both its driving force and selectivity. Namely, unlike CNT transporters in mammals, CeCNT3 was sensitive to proton gradients but not to sodium gradients. The transporter was selective for both naturally occurring purine and pyrimidine nucleosides except for cytidine. These unique functional characteristics permit us to pursue a better understanding of the amino acid residues involved in both substrate recognition and driving forces of the transporters in the CNT family and shed light on the evolution of functional domains in this major family of transporters.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

RNA Isolation and cDNA Preparation. Total RNA was isolated from C. elegans (strain N2) with Trizol reagent (Life Technologies) according to the manufacturer's protocol, then was taken as the template to synthesize cDNA by reverse transcribed polymerase chain reaction (reverse transcription-PCR) using oligo(dT) primer (Life Technologies). CeCNT3 was then amplified by nested PCR technique using the cDNA as template. These primers were designed specially to amplify the potential nucleoside transporter coded by the genomic sequence in the F27E11.2 locus (GenBank accession number AF016413). The first pair of primers were 5'-ttagaggatccgcggtgccgaccaatca-3' (sense), 5'-aatttttatacaatagaataatatatgtac-3' (antisense). The second pair of primers were 5'-cgaccaatcagcgatgcgc aatgggcgtgg-3' (sense), 5'-aaaaactttgaatttatccgctaggaatccgg-3' (antisense). PCR was performed with the following conditions: 94°C for 5 min; 30 cycles of 94°C for 1 min, 62°C for 2 min, and 72°C for 2 min. The reaction was then extended by incubation at 72°C for an additional 15 min. A PCR product of 1.8 kilobases (encoding CeCNT3) was obtained, subcloned into the pGEM-T vector (Promega, Madison, WI) and sequenced at the Biomolecular Resource Center at the University of California, San Francisco.

Sequence Analysis. Multiple alignment and motif prediction were carried out using the Genetics Computer Group (GCG) software package (Wisconsin Package, Version 9, Madison, WI). Hydropathy analysis was performed using DNA Strider (Version 1.2). BLAST network at the National Center for Biotechnology Information was used in database searching.

Expression and Transport Assay in Xenopus Laevis Oocytes. To obtain higher expression, CeCNT3 was subcloned into vector pOX (a gift from Andrew T. Gray, University of California, San Francisco), which contains the 5' and 3' untranslated regions of the X. laevis -globin gene flanking the insert (Jegla and Salkoff, 1997; Chavez et al., 1999). CeCNT3 cRNA was synthesized using T3 polymerase (Stratagene, La Jolla, CA) following the manufacturer's protocol. Oocytes were harvested and treated as described previously (Wang et al., 1997). The cRNA (50 nl; ~0.4 ng/nl) or water was injected individually into defolliculated oocytes. Oocytes were incubated at 18°C for 30 to 40 h and then the uptake assays were performed in 100 µl of transport buffer [2 mM KCl, 1 mM CaCl₂, 10 mM HEPES, or 10 mM 2-(N-morpholino)ethanesulfonic acid (MES)] containing different concentrations of ³H-labeled nucleosides (Moravek Biochemicals, Brea, CA). To test the effect of sodium-gradient on the activity of CeCNT3, uptake assays were carried out in transport buffer containing 0 to 100 mM NaCl with choline chloride to maintain the iso-osmolality. In experiments testing the effect of proton-gradient on the activity of CeCNT3, oocytes were preincubated in choline buffer at pH 7.4 and then transferred to transport buffers with pH ranging from 5.0 to 8.0. To study the effects of FCCP, a proton ionophore, on the activity of CeCNT3, oocytes were preincubated in choline buffer (pH 8.0, no FCCP) for 3 h. Oocytes were then incubated in choline buffer (pH 5.5 or 8.0, with or without 5 µM FCCP) for 0, 30, and 60 min, and the uptake of thymidine or inosine was measured in sodium buffer at pH 5.5 and pH 8.0. In the pH range of 5.0 to 6.5, 10 mM MES was used in the uptake buffer; 10 mM HEPES was used in the pH range of 7.0 to 8.0. In inhibition studies, chemicals at tested concentrations were added to the uptake buffer individually. After 30 min of incubation at 25°C, oocytes were washed five times in 3 ml of ice-cold choline buffer and lysed individually in 10% SDS. The amount of radiolabeled nucleoside transported into each oocyte was determined by liquid scintillation counting. Most studies were carried out in transport buffer containing 100 mM NaCl, pH 6.5. A pH of 6.5 was selected because preliminary studies suggest that uptake of nucleosides was high at pH 6.5 and that this pH did not destroy the viability of the oocytes.

Site-Directed Mutagenesis. Single (tyrosine 332 to phenylalanine, Y332F) and double (threonine 327 to serine and valine 328 to leucine, T327S/V328L) mutations were constructed with QuickChange Site-directed Mutagenesis Kit (Stratagene), using wild-type CeCNT3 cDNA as the template. The sequences of Y332F and T327S/V328L mutants were confirmed by DNA sequencing at the Biomolecular Resource Center at the University of California, San Francisco. Expression and transport assays of Y332F and T327S/V328L mutants were the same as described above for wild-type CeCNT3.

Statistics and Data Analysis. Groups of 8 to 10 cRNA-injected or water-injected oocytes were used for each experiment. Uptake values are expressed as mean ± S.E. For kinetic studies of CeCNT3, uptake rates (V) determined at different substrate concentrations (S) were fit to the Michaelis-Menten equation: V = V_max × S/(K_m + S) + K_d × S, where V_max is the maximal uptake rate, K_m is the Michaelis-Menten constant (the substrate concentration at V_max/2), and K_d is the nonsaturable first-order rate constant. Fits were carried out using a nonlinear least-squares regression-fitting program (Kaleidagraph v. 3.0; Abelbeck/Synergy Software, Reading, PA). Kinetic experiments were repeated several times in different batches of oocytes; data for one representative experiment are presented in this study. For inhibition study, statistical analysis was carried out by comparing the uptakes from tested compounds with the uptake from controls in the same experiments using a two-tailed, two-sample equal variance t test. Results with the probability of p < 0.05 were considered significantly different.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Nucleotide and the Deduced Amino Acid Sequences of CeCNT3. A BLAST search of the C. elegans genome database with the cDNA sequence of rCNT1 revealed only two homologs of the CNT family. These two transporters are encoded by the genes, F27E11.1 and F27E11.2. We cloned the full-length cDNA of both homologs, using primers designed to anneal to the 5' and 3' ends of the predicted exons (GenBank accession number AF016413). Because the activity of the transporter encoded by F27E11.1 was low and we could not make any conclusions about its selectivity, we focused on the other transporter, CeCNT3 (encoded by gene F27E11.2) in this study. The nucleotide sequence and deduced amino acid sequences of CeCNT3 were deposited in GenBank with accession no. AF162674. As shown in Fig. 1A, the cDNA is 1826 bp long with an open reading frame of 1728 bp (including the termination codon). Because there is an in-frame stop codon upstream of the ATG and because the cloned transporter was functional, we concluded that the entire open reading frame was obtained. The sequence predicts a protein of 575 amino acid residues with a molecular mass of 63.4 kDa. Two fragments (158 bp and 48 bp) in the 5' region of the cDNA (GenBank accession number AF016413) predicted from the genomic sequence are absent from our cloned CeCNT3 cDNA, suggesting that either these two fragments were mistakenly predicted as exons or that there is alternative splicing. The initiation codon of CeCNT3 is preceded by a consensus Kozak sequence (A/GXXATG) (Kozak, 1986). Using Kyte-Doolittle (1982) hydropathy analysis, CeCNT3 is predicted to have 11 to 14 transmembrane domains (TMD) (Fig. 1B). The 14 putative TMDs are highlighted (Fig. 1A). Multiple alignments show that, at the amino acid sequence level, this transporter is related to mammalian CNTs (~30% identity) (Fig. 2A) (Huang et al., 1994; Che et al., 1995; Ritzel et al., 1997; Wang et al., 1997) and bacterial CNTs (~14-24% identity) (data not shown) (Craig et al., 1994; Fleischmann et al., 1995; Kunst et al., 1997; Tomb et al., 1997), but is not related to the equilibrative nucleoside transporter family (data not shown). We also compared the identity of the amino acid residues in the regions that have been found to be important in determining the purine and pyrimidine selectivity of hCNT and rCNT (Wang and Giacomini, 1997, 1999b; Loewen et al., 1999). In this particular region, CeCNT3 shows ~44% identity to hCNT2 and rCNT2 and ~33% identity to hCNT1 and rCNT1 (Fig. 2B). The most divergent regions between CeCNT3 and mammalian CNT are in the N-terminal and C-terminal domains. There are three putative N-linked glycosylation sites (Asn-171, Asn-506, and Asn-546) and five potential protein kinase C phosphorylation sites (Thr-25, Thr-98, Ser-123, Ser-523, and Ser-563). The presence of presumptive protein kinase C sites suggests that CeCNT3 may be regulated by phosphorylation.

View larger version (49K): [in this window] [in a new window]   Fig. 1.   Nucleotide and deduced amino acid sequences of CeCNT3. A, cDNA of CeCNT3 is 1826 bp. The open reading frame of CeCNT3 is flanked by 52 bp of 5'-untranslated region and 45 bp of 3'-untranslated region. The three potential N-linked glycosylation sites and five potential protein kinase C phosphorylation sites are labeled with * and #, respectively. The 14 putative TMDs are highlighted. B, Kyte-Doolittle hydropathy plot of CeCNT3 with a window of 11 amino acid residues.

View larger version (76K): [in this window] [in a new window]   Fig. 2.   Sequence analysis. A, multiple alignment of CeCNT3, hCNT1, rCNT1, hCNT2, and rCNT2. Amino acid residues with 100% identity among the transporters are highlighted. Periods indicate the gaps to create the alignment. B, putative transmembrane domain 8 and 9 of hCNT1, rCNT1, hCNT2, and rCNT2 and the corresponding region of CeCNT3 (amino acids 275-335) were compared. In this particular region, there are nine conserved amino acid differences between CNT1 and CNT2. Of these nine residues, CeCNT3 differs from both CNT1 and CNT2 at one position; CeCNT3 is identical to CNT2 at seven positions and to CNT1 at one position. Shaded positions indicate conserved residues that are identical to CeCNT3; bold positions indicate conserved residues that differ from CeCNT3.

Substrate Selectivity of CeCNT3. Functional studies of CeCNT3 were carried out using naturally occurring nucleosides. Oocytes injected with CeCNT3 cRNA were incubated in transport buffer containing 10 µM ³H-labeled nucleosides for 30 min. Initial studies established that the uptake of both thymidine and inosine was linear with time up to 3 h (Fig. 3). A 30-min time point was selected for further kinetic studies. Compared with control oocytes (water-injected), the uptake of inosine, thymidine, uridine, guanosine, and adenosine was significantly increased (20 to 30 fold) (Fig. 4). In contrast, the uptake of cytidine was not increased significantly. These data indicate that CeCNT3 exhibits a unique substrate selectivity, transporting inosine, thymidine, uridine, adenosine and guanosine, but not cytidine.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Time course of transport of inosine and thymidine by CeCNT3. A and B, uptake of [3H]inosine (10 µM) and [3H]thymidine (10 µM) into CeCNT3 cRNA-injected (solid line) or water-injected (dotted line) oocytes as the function of time was measured. Transport buffer containing 100 mM NaCl at pH 6.5 was used.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Nucleoside selectivity of CeCNT3. Uptake assays of [3H]inosine (10 µM), [3H]thymidine (10 µM), [3H]uridine (10 µM), [3H]cytidine (10 µM), [3H]adenosine (10 µM) and [3H]guanosine (10 µM) into CeCNT3 cRNA-injected (solid bars, left of pairs) or water-injected (gray bars, right of pairs) oocytes were carried out in transport buffer containing 100 mM NaCl, pH 6.5. Values are expressed as mean ± S.E.

Inhibition Profile of Nucleoside Transport. To further determine the characteristics of CeCNT3, we examined the inhibitory effects of naturally occurring nucleosides (inosine, thymidine, uridine, guanosine, adenosine, and cytidine), nucleobases (hypoxanthine, thymine), ribose, thymidine monophosphate (5'TMP) and nitrobenzylthioinosine (NBMPR) on CeCNT3-mediated transport of inosine and thymidine. We observed that all of the naturally occurring nucleosides, except cytidine, significantly inhibited the transport of both inosine and thymidine (Fig. 5A, 5B). Although cytidine significantly inhibited the transport of thymidine (p < 0.05), the inhibitory effect was much lower than that of other naturally occurring nucleosides.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Inhibition profile of CeCNT3. A and B, uptakes of [3H]inosine (10 µM) and [3H]thymidine (10 µM) into CeCNT3 cRNA-injected oocytes or water-injected (H2O) were measured in the absence of inhibitors (control) or in the presence of naturally occurring nucleosides (inosine, thymidine, uridine, cytidine, adenosine, and guanosine), 5'TMP, hypoxanthine, thymidine, ribose, and NBMPR. Final inhibitor concentrations were 1 mM, except NBMPR (100 µM). *, significantly inhibited the CeCNT3-mediated inosine or thymidine uptakes (p < 0.05).

Kinetic Studies of CeCNT3. CeCNT3-mediated transport of inosine and thymidine was saturable, with apparent K_m values of 15.2 ± 5.3 µM and 11.0 ± 2.4 µM, respectively. The apparent V_max values of inosine and thymidine were 3.20 ± 0.50 pmol/oocyte/30 min and 3.85 ± 2.88 pmol/oocyte/30 min, respectively (Table 1). Although differences in the kinetics of inosine and thymidine transport were observed between batches of oocytes, within the same batch of oocytes, the apparent K_m value of thymidine was always lower than that of inosine; at the same (nonsaturating) concentration, thymidine was always transported at greater rate than inosine.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Kinetic Study. A and B, in the absence of inhibitor (solid line) or in the presence of inhibitor (dotted line), CeCNT3-mediated uptakes of [3H]inosine (10 µM) and [3H]thymidine (10 µM) as a function of the substrate concentration were carried out in transport buffer containing 100 mM NaCl, pH 6.5. For inosine (A), 40 µM thymidine was used as inhibitor; for thymidine (B), 40 µM inosine was used as inhibitor. Under the same experimental conditions, uptake of [3H]inosine and [3H]thymidine into water-injected oocytes was less than one-tenth of the uptake into CeCNT3 cRNA-injected oocytes (data not shown).

Effects of Sodium-Gradient and Proton-Gradient on the Activity of CeCNT3. In initial studies, we observed that CeCNT3-mediated transport of inosine into oocytes functioned well in media at pH 7.4 containing either 100 mM NaCl or 100 mM choline chloride, with uptake rates of 6.74 ± 0.56 and 6.33 ± 0.69 pmol/oocyte/30 min, respectively, indicating that transport was not dependent on sodium. No stimulatory effect of sodium up to concentrations of 100 mM on the activity of CeCNT3 was observed (Fig. 7A).

View larger version (14K): [in this window] [in a new window]   Fig. 7.   Effects of sodium and pH on the activity of CeCNT3. A, sodium dependence of CeCNT3. Uptake of [3H]thymidine (10 µM) in CeCNT3 cRNA-injected (solid line) or water-injected (dotted line) oocytes were measured in transport buffer containing 0 to 100 mM NaCl with choline chloride to maintain the iso-osmolality. B, pH dependence of CeCNT3. Uptake of [3H]thymidine (10 µM) in CeCNT3 cRNA-injected (solid line) or water-injected (dotted line) oocytes were measured in transport buffer containing 100 mM NaCl with pH ranging from 5.0 to 8.0. In both A and B, values are expressed as mean ± S.E.

View larger version (13K): [in this window] [in a new window]   Fig. 8.   Effects of FCCP on the uptake of thymidine. Oocytes were incubated in choline buffer at pH 5.5 with 5 µM FCCP (), pH 5.5 without 5 µM FCCP (), pH 8.0 with 5 µM FCCP () or pH 8.0 without 5 µM FCCP () for 0, 30, and 60 min after having been preincubation in choline buffer at pH 8.0 for 3 h. Uptake of thymidine (30 min) was then performed in sodium buffer at pH 5.5 and 8.0. Under the same experimental conditions, uptake of [3H]thymidine into water-injected oocytes was less than one tenth of the uptake into cRNA-injected oocytes (data not shown).

Site-Directed Mutagenesis. Multiple alignments of the amino acid sequence of CeCNT3 with hCNT and rCNT in the region (between amino acids 275 and 335 of CeCNT3) shown previously to be critical in substrate recognition and binding (Wang and Giacomini, 1997, 1999b; Loewen et al., 1999) were performed. These alignments show that, of the nine conserved residues in this particular region, CeCNT3 has seven residues identical with those of hCNT2 and rCNT2 (except the tyrosine at 332, which is identical with hCNT1 and rCNT1) (Fig. 2B). Because CeCNT3 transports not only purine nucleosides but also thymidine, we hypothesized that the tyrosine at position 332 of CeCNT3 may facilitate the recognition and subsequent transport of thymidine as well as other nucleosides. We mutated this residue to the equivalent residue, phenylalanine, in CNT2. Compared with that of the wild-type transporter, the uptake of thymidine mediated by Y332F mutant was decreased (Fig. 9A). At equivalent concentrations, thymidine is the nucleoside transported most rapidly by the wild-type transporter (Fig. 4); however, it is the nucleoside transported most slowly by the Y332F mutant. Kinetic studies revealed that the K_m value of thymidine is increased from 11.4 ± 2.1 µM to 31.7 ± 5.5 µM (Fig. 9B). Further studies demonstrated that the K_m value of inosine was also increased (Table 2), suggesting that the tyrosine in this position is indeed involved in nucleoside recognition.

View larger version (17K): [in this window] [in a new window]   Fig. 9.   Nucleoside selectivity and kinetic study of Y332F mutant. A, uptake assays of [3H]inosine (10 µM), [3H]thymidine (10 µM), [3H]uridine (10 µM), [3H]cytidine (10 µM), [3H]adenosine (10 µM) and [3H]guanosine (10 µM) into Y332F mutant cRNA-injected (solid bars) or water-injected (gray bars) oocytes were carried out in transport buffer containing 100 mM NaCl, pH 6.5. Values are expressed as mean ± S.E. B, wild-type CeCNT3-mediated (solid circles) and Y332F mutant mediated () uptakes of [3H]thymidine as a function of substrate concentration were carried out in transport buffer containing 100 mM NaCl, pH 6.5. Under the same experimental conditions, uptake of [3H]thymidine into water-injected oocytes was less than one tenth of the uptake into cRNA-injected oocytes (data not shown).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We report here the cloning of a cDNA from C. elegans coding for a novel member, CeCNT3, of the concentrative nucleoside transporter family. We demonstrate that CeCNT3 has unique functional characteristics.

CeCNT3 has a unique substrate selectivity compared with other concentrative nucleoside transporters. In particular, in contrast to the cloned purine (CNT2 or SPNT) and pyrimidine (CNT1) selective transporters, CeCNT3 transports all naturally occurring nucleosides tested with the exception of cytidine (Fig. 4). Based upon the competitive interaction between thymidine and inosine (Fig. 6), our data suggest that there is a single recognition site in CeCNT3 for both purine and pyrimidine nucleosides. Thus, the molecular basis for broad selectivity resides in a single permeation pathway rather than multiple pathways.

Molecular Determinants of Substrate Discrimination. CeCNT3 is ~30% identical with mammalian CNTs (hCNT and rCNT) (Fig. 2A). Like hCNT and rCNT, it also has 11 to 14 putative transmembrane domains (Fig. 1). CeCNT3 thus has the overall structural features of a concentrative nucleoside transporter.

Proton Dependence. Another unique characteristic of CeCNT3 is that in contrast to mammalian transporters in the CNT family, nucleoside transport is dependent on the proton gradient, but not on the sodium gradient (Figs. 7 and 8). Concentrative transporters use ion-coupled electrochemical energy stored in a transmembrane gradient to actively transport substrates into or out of cells. In bacteria as well as in plants, a proton is the preferred coupling ion and the proton gradient is maintained by H⁺/K⁺-ATPase. In vertebrates, sodium is the preferred coupling ion of many coupled transporters because of the evolutionary substitution of H⁺/K⁺-ATPase by Na⁺/K⁺-ATPase in animal cells (Hediger, 1994; Nelson, 1994). To date, only a few mammalian plasma membrane transporters (e.g., the oligopeptide transporters, PepT1 and PepT2) have been described that use protons as the coupling ion (Fei et al., 1994; Leibach and Ganapathy, 1996). In the case of ion-coupled (also known as concentrative) nucleoside transporters, the NupC family, in bacteria such as Escherichia coli, Bacillus subtilis, and Helicobacter pylori, consists of several proton-coupled nucleoside transporters (Craig et al., 1994). In contrast, all documented concentrative nucleoside transporters in vertebrates are sodium-dependent (Crawford et al., 1990; Dagnino et al., 1991; Wu et al., 1992, 1994; Huang et al., 1993, 1994; Che et al., 1995; Wang et al., 1997).