Abstract
The human organic anion transporter hOAT1 (SLC22A6) contributes to the uptake of a range of small organic anions across the basolateral membrane of the renal proximal tubule and drives their urinary elimination. The aim of this study was to identify genetic variants of hOAT1 and to investigate potential effects on the functional properties of this transporter. Twenty single nucleotide polymorphisms (SNPs) in hOAT1 were identified in genomic DNA from 92 individuals of African, Asian, and Caucasian origin. Two SNPs encoded changes in amino acid sequence; arginine to histidine (residue 50) and lysine to isoleucine (residue 525). Significantly, these SNPs were only present in the samples of African origin. When expressed in Xenopus oocytes, wild-type R50-hOAT1 and the variants R50H-hOAT1 and K525I-hOAT1 all mediated the probenecid-sensitive uptake of the classic organic anion para-aminohippurate (PAH). Kinetic analysis indicated that the transport affinity (Km) for PAH was unchanged in the variants, compared with wild type. Interestingly, the Km for the nucleoside phosphonate analogs adefovir, cidofovir, and tenofovir seemed to be decreased in the R50H-hOAT1 variant compared with the wild type, whereas the kinetics of K525I-hOAT1 remained unchanged. In conclusion, this is the first study to identify variation of hOAT1 in a racially diverse sample and to investigate the functional properties of the resulting variants. Since hOAT1 has been suggested as the basis of nephrotoxicity induced by nucleoside phosphonate analogs, this study raises the intriguing possibility that individuals with genetic variation in hOAT1, such as R50H, may display different handling of these drugs.
The central role of the kidney in the elimination of potentially toxic xenobiotics from the blood into the urine is well documented. Early studies began with the whole animal observations (Rennick et al., 1977; Besseghir et al., 1981; Besseghir and Rennick, 1981), followed by many years of intact tissue (Groves et al., 1995; Chatsudthipong and Jutabha, 2001) and cellular studies (Saito et al., 1992; Chan et al., 1997). It is only recently that the molecular identity of the contributing transport proteins has been elucidated (Burckhardt and Wolff, 2000; Wright and Dantzler, 2004).
During secretion of organic anions by the renal proximal tubule, the anionic drug or xenobiotic must cross two plasma membrane barriers. It is first removed from the circulation across the basolateral membrane into the cell and then secreted across the apical membrane into the urine for elimination. Specific transport proteins are thought to be responsible for each of these steps. At the basolateral membrane, the main mechanism for organic anion uptake was shown to be organic anion/α-ketoglutarate exchange (Pritchard and Miller, 1996). The molecular counterpart of this uptake process was, until recently, considered to be organic anion transporter (OAT) 1, first isolated from rat kidney by expression cloning (Sekine et al., 1997; Sweet et al., 1997). Recent studies, however, indicate that a second transporter, OAT3, also operates by this mechanism and thus may contribute toward this step (Bakhiya et al., 2003; Sweet et al., 2003).
The first human organic anion transporter hOAT1 (SLC22A6), was identified by homology cloning in 1998 and identified as a 60.3-kDa protein (Reid et al., 1998; Cihlar et al., 1999; Hosoyamada et al., 1999). At the genomic level, hOAT1 is 8.2 kilobases, located on chromosome 11q13-1 and is composed of 10 exons, separated by nine introns (Bahn et al., 2000). Since its isolation, OAT1 transport has been characterized in great detail (Burckhardt and Wolff, 2000; Wright and Dantzler, 2004). Substrates for this transporter are wide ranging; from the classic small organic anion para-aminohippurate (PAH), to several clinically important drugs, herbicides, and endogenous substances. Indeed, OAT1's uptake of the nucleoside phosphonate analogs cidofovir and adefovir has been suggested to be the basis for the nephrotoxicity induced by clinical doses of these drugs (Ho et al., 2000).
Analysis of the human genome indicates that several types of genetic variation, including deletions, insertions, and single nucleotide polymorphisms (SNPs) exist. SNPs are sequence variations that occur when a single nucleotide in the genomic sequence is altered and are the most common form of variation, with approximately one SNP occurring per 1.9 kilobases of the genome (Chakravarti, 2001). Recently, several groups have turned their efforts to the study of variation in membrane transporters (Iida et al., 2001; Ito et al., 2001; Leabman et al., 2002; Schinkel and Jonker, 2002). This study, however, is the first to identify variation in a member of the OAT transporter family in a set of geographically diverse samples and to investigate the impact of this variation on transporter function.
Materials and Methods
Variant Identification. Seventy-two genomic DNA samples screened for variation in hOAT1 were obtained from the Human Genetic Cell repository, sponsored by the National Institutes of Health, housed at the Coriell Institute (Camden, NJ). These samples from individuals of African (15 African-American and nine African-Pygmy), Asian (four Indo-Pakistani, five native Taiwanese, five mainland China, three Cambodian, three Japanese, and four Melanesian), or Caucasian (seven Utah, five Druze, seven Eastern European, and five Russian) origin were selected for geographical diversity (Fritsche et al., 2000). An additional 20 samples were from the “DNA Polymorphism Discovery Resource” (National Institutes of Health; Collins et al., 1998). These samples are from anonymous U.S. residents selected to represent major racial groupings of the population: European-American, African-American, Mexican-American, Native-American, and Asian-American. In total, 92 individuals/184 chromosomes were screened. Because all samples came from commercially available cell lines from healthy anonymous donors, the protocol was considered exempt by the Lawrence Livermore National Laboratory Review Board under 10CRF745.101(b) 4.
The resequencing strategy was used to identify variants and is described elsewhere in detail (Mohrenweiser et al., 2002). Briefly, this involved the sequencing of the same genomic region in the 92 individuals to identify DNA sequence variation. PCR products containing the individual exons of hOAT1, plus adjacent intronic and noncoding regions, were generated using oligonucleotide primers specific to hOAT1 (Table 1) and Pfu DNA polymerase (Stratagene, La Jolla, CA). The PCR primers were located in the introns approximately 75 to 100 nucleotides 5′ and 3′ of the intron/exon boundary so that at least 50 nucleotides of high-quality intronic sequence could be obtained adjacent to each exon, with the exceptions noted in Table 1. Exon 1 was amplified as two overlapping products to accommodate the large size of this exon. Several introns were sufficiently small so that pairs of exons and the entire intron could be amplified as a single product. The ability to obtain high-quality sequence from both DNA strands was the limitation imposed on the size of the amplified product. 5′ binding sites for energy transfer (ET) DNA sequencing primers (forward, GTTTTCCCAGTCACGACG; reverse, AGGAAACAGCTATGACCAT) were added to each primer, and thus PCR products could be directly sequenced in both directions using the DYEnamic ET primers and sequencing system (Amersham Biosciences, Inc., Cleveland, OH). It should be noted that complete intronic and untranslated regions were not sequenced in this study. The locations of the PCR primers and SNPs in the current genomic sequence can be obtained by searching the database with a sequence homology algorithm (e.g., BLAST) using first the reference cDNA sequence to obtain the genomic sequence for SLC22A6 and then repeating the search using the primer sequence or the sequence surrounding the SNP.
Variant Reconstruction. The variants R50H-hOAT1 and K525I-hOAT1 were reconstructed by site-directed mutagenesis of the wild-type hOAT1 (pcDNA3.1-hOAT1) (Cihlar et al., 1999) using the QuikChange mutagenesis system (Stratagene). The mutagenic primers (QIAGEN, Valencia, CA) were as follows: R50H-hOAT1 (sense) 5′-CCCACCACTGCCACCCGCCTGCCG-3′, R50H-hOAT1 (antisense) 5′-CGGCAGGCGGGTGGCAGTGGTGGG-3′, K525I-hOAT1 (sense) 5′-GAGCAGGAAAGGGATACAGACGCGACAGC-3′, and K525I-hOAT1 (antisense) 5′-GCTGTCGCGTCTGTATCCCTTTCCTGCTC-3′. The wild-type and variant plasmids were purified (high-speed plasmid midi; QIAGEN), and the sequence was confirmed by automated sequencing (DYEnamic ET; Amersham Biosciences, Inc.) at the National Institute of Environmental Health Sciences core facility. The plasmids were linearized by XbaI digest (New England Biolabs, Beverly, MA), and capped cRNA was produced by in vitro transcription (T7-message machine; Ambion, Austin, TX).
Expression of Variants. Wild-type and variant hOAT1 cRNAs were expressed in Xenopus laevis oocytes as described previously (Cihlar et al., 1999). Briefly, stage V and VI oocytes were harvested from X. laevis (Xenopus One, Ann Arbor, MI) and isolated by collagenase digestion in calcium-free oocyte Ringer buffer (82.5 mM NaCl, 2.5 mM KCl, 1 mM MgCl2, and 5 mM HEPES, pH 7.6). Twenty-four hours postisolation, oocytes were injected with either 30 ng of cRNA or water as the control. Medium was changed every 24 h.
Transport Measurements. Seventy-two hours postinjection, groups of six to eight oocytes were incubated for 20 min at room temperature in oocyte-Ringer-2 buffer (82.5 mM NaCl, 2.5 mM KCl, 1 mM Na2HPO4, 3 mM NaOH, 1 mM CaCl2, 1 mM MgCl2, 1 mM pyruvic acid, and 5 mM HEPES, pH 7.6) containing various concentrations of either [3H]PAH (1 μCi/ml), [14C]adefovir [9-(2-phosphonylmethoxyethyl)-adenine] (1 μCi/ml), [3H]cidofovir [(S)-1-[3-hydroxy-2(phosphonylmethoxy)propyl]-cytosine] (1 μCi/ml), or [3H]tenofovir [(1R)-9-(2-phosphonylmethoxypropyl)-adenine] (1 μCi/ml). To calculate the diffusional component of uptake, substrate uptake at 10 and 200 μM was also measured in the presence of 1 mM probenecid. At 200 μM, the diffusional component was approximately 10 and 20% of total flux for PAH and the nucleoside phosphonates, respectively. These figures were subtracted from total uptake to determine the transporter-mediated component. Previous experiments determined that 20-min substrate uptake was within the initial linear range of hOAT1 transport. Individual oocyte radioactivity was measured by liquid scintillation spectroscopy with external quench correction. To determine transport Km, kinetic data were corrected for diffusion and subjected to nonlinear and linear (Lineweaver-Burk transformation) regression analysis. All animal experiments were conducted under protocols approved by the National Institute of Environmental Health Sciences Animal Care and Use Committee.
Statistics. Means were compared using one way analysis of variance with Dunnett's post test. Differences in mean values between control and test groups were considered significant when P < 0.05. Kinetic data are presented graphically as single representative experiments using six to eight oocytes per point and in tabular format as the mean of two to four independent experiments.
Chemicals. [3H]PAH (4 Ci/mmol) was purchased from Perkin-Elmer Life and Analytical Sciences (Boston, MA). [14C]Adefovir, [3H]cidofovir, and [3H]tenofovir and their unlabeled equivalents were purchased from Moravek Biochemicals (Brea, CA) or were kind gifts from Gilead Sciences (Foster City, CA). Unlabeled PAH and probenecid were obtained from Sigma-Aldrich (St. Louis, MO). All other chemicals were obtained from commercial sources and were of reagent grade.
Results
Genetic Variation in SLC22A6. The 10 exons of the hOAT1 gene, with their flanking intronic and untranslated regions, were directly sequenced from 184 chromosomes. In total, 20 SNPs were identified; five in untranslated regions, six in introns, and nine in exons. Of the nine sequence variants found in exons, two coded for amino acid changes (nonsynonymous), whereas seven did not (synonymous). At amino acid position 50, arginine (codon CGC) was changed to histidine (codon CAC), and at amino acid 525, lysine (codon AAA) was changed to isoleucine (codon ATA). The locations of the amino acid substitutions were assigned using GenBank sequence AF124373 as the reference. Table 2 shows the positions of the SNPs, the changes detected and the allele frequency. The highest frequency SNPs (0.27 and 0.30) were in the 5′-untranslated region of hOAT1, and seven of the 20 SNPs were >0.01 in the total chromosomes screened. The frequency of the two nonsynonymous SNPs, R50H and K525I, were 0.04 and 0.005, respectively, in the total sample and were heterozygous in all cases. However, if the data are divided into the subsamples of Caucasian, African, and Asian origin (Table 3), it can be seen that in the samples studied, these two SNPs were only detected in the chromosomes of African origin. Within this subsample of 48 chromosomes, the allele frequency is increased to 0.17 and 0.02 for R50H and K525I-hOAT1 genotypes, respectively. Interestingly, within the samples of African origin, the R50H variant was represented in both the African American and Pygmy groups (allele frequency of 0.28 and 0.1 within each group, respectively). The K525I variant was an individual of African American descent. Previous studies have identified genetic variation in hOAT1; Iida et al. (2001) identified eight variants in the hOAT1 gene in a Japanese sample of 96 chromosomes, and three of these SNPs were identified in the Asian samples of the present study (Table 2). The most comprehensive study to date of genetic variation in hOAT1 was carried out by the Pharmacogenetics and Pharmacogenomics Knowledge Base (Klein et al., 2001; www.pharmgkb.org); 31 variants were identified in an ethnically diverse population of 528 to 552 chromosomes. Of these variants, nine were also found in the present study, as indicated in Table 2.
Examination of the two nonsynonymous SNPs observed in the present study indicates that in contrast to lysine 525, arginine 50 is an evolutionary conserved residue located within a motif of six conserved residues (Fig. 1A). Within the predicted secondary structure of hOAT1, R50 seems to be located in the large extracellular loop between transmembrane domains 1 and 2, whereas K525 is located in the C terminus (Fig. 1B). The potential impact of these substitutions on protein structure and activity was explored with the “Sorting Intolerant from Tolerant” (SIFT) (Ng and Henikoff, 2002) and the “Polymorphism Phenotyping” (PolyPhen) (Sunyaev et al., 2001) algorithms. These algorithms predict the potential impact of amino acid substitutions based upon evolutionary conservation of sequence, localization within functional domains and the chemical and physical properties of the exchanged amino acid residues. Both algorithms predicted that the Ile substitution at reside 525 had a high probability of impacting function (PSIC score of 2.31 and SIFT score of 0.0), whereas the His substitution at residue 50 was predicted to have only a moderate negative impact (PSIC score of 1.82 and SIFT score of 0.17).
Functional Analysis of hOAT1 Variants. Initial experiments were conducted to determine whether the reconstructed variants R50H-hOAT1 and K525I-hOAT1 were functional. Uptake of 10 μM [3H]PAH into oocytes expressing either the wild-type hOAT1 or individual variants was measured in the presence and absence of the hOAT1 transport inhibitor probenecid (1 mM). Compared with water-injected control, oocytes expressing wild type, R50H, or K525I-hOAT1 mediated significant uptake of the substrate, and this uptake was almost abolished in the presence of probenecid (data not shown). Based on this, the transport properties of the two variants were then characterized in more detail by measuring transport kinetics for four hOAT1 substrates: PAH, adefovir, cidofovir, and tenofovir.
Figure 2 shows one representative experiment of the uptake of the classic organic anion PAH mediated by wild-type hOAT1, R50H, and K525I-hOAT1 over a range of concentrations (2–200 μM). Uptake is corrected for diffusion (measured in presence of 1 mM probenecid). The calculated Km for wild-type hOAT1 in this experiment was 5.1 ± 0.2 μM, which is consistent with the previously published Km for this transporter (Cihlar et al., 1999). Table 4 shows the mean Km values for the R50H and K525I-hOAT1 variants were 4.9 ± 0.6 and 4.7 ± 0.1 μM, respectively. The lack of statistical significant difference (P > 0.05) between these kinetic values indicates no change in the transport affinity for PAH in the R50H and K525I-hOAT1 variants, compared with wild-type hOAT1.
Kinetic studies were also conducted for three nucleoside phosphonate analogs. Figures 3, 4, and 5 show representative experiments of transporter-mediated adefovir, cidofovir, and tenofovir uptake for wild-type-hOAT1 and R50H and K525I-hOAT1 mutants over a range of concentrations (5–200 μM). Furthermore, Table 4 shows mean Km values for several independent experiments. Figures 3 and 4 show that again, hOAT1 had calculated Km values for adefovir and cidofovir consistent with the literature (Cihlar et al., 1999) of 20 ± 1.9 and 55 ± 9.1 μM, respectively; however, the R50H-hOAT1 construct displayed significantly reduced (P < 0.01) Km values of 11.4 ± 0.6 and 16.1 ± 2.7 μM, a decrease of approximately 40 and 70%, respectively. The Km for K525I-hOAT1 remained almost identical in both cases to the wild-type transporter. Figure 5 shows a representative experiment where kinetic constants were determined for a third nucleoside, phosphonate, tenofovir, which showed a similar trend. Mean Km values from multiple experiments are again shown in Table 4, and it can be seen that the R50H-hOAT1 mutant Km (14.6 ± 1.8 μM) was significantly decreased (P < 0.05) compared with wild-type hOAT1 (22.3 ± 0.7 μM), whereas the K525I variant was similar at 21.9 ± 2.5 μM. Vmax values for the transport of the three nucleoside phosphonate analogs were also decreased ∼50 to 60% for the R50H variant, compared with either the wild-type hOAT1 or the K525I variant.
Discussion
The last decade has seen an explosion in the interest and understanding of epithelial transport mechanisms. Dedicated research efforts have identified many transport proteins at the molecular level and have gone on to ascertain their tissue and cellular distributions, driving forces, and substrate specificities (van Montfoort et al., 2003). Now, following in the footsteps of the highly characterized cytochrome P450 enzymes, transporter research is moving in the direction of pharmacogenetics. The identification of genetic variation in transporters and the subsequent characterization of phenotype will add to the growing knowledge base in the understanding of interindividual response to xenobiotics.
The aim of the present study was to identify SNPs in the hOAT1 gene and to investigate the functional properties of any nonsynonymous SNPs detected. Among the 184 chromosomes screened, 20 single nucleotide polymorphisms were detected. Ten of these SNPs are new, with three being identified previously in Japanese samples (Iida et al., 2001) and nine identified in an ethnically diverse population (Klein et al., 2001). Two of the SNPs identified in our study resulted in an amino acid change: arginine to histidine at residue 50 and lysine to isoleucine at residue 525. As has been described for polymorphisms in other transporters and enzymes (Schaeffeler et al., 2001; Leabman et al., 2002), there was an uneven distribution of SNPs across racial groups; indeed, of the 184 samples studied, the two nonsynonymous SNPs detected were only found in the chromosomes of African descent.
In an effort to reveal potentially subtle differences in transporter function, analysis of transport kinetics of a range of substrates was studied. R50H-hOAT1 seemed to have an increased affinity for the nucleoside phosphonate analogs cidofovir, adefovir, and tenofovir compared with wild-type hOAT1, whereas its transport of PAH was unaffected. In contrast K525I-hOAT1 seemed to have similar transport characteristics to the wild-type-hOAT1, suggesting that this residue is not important in the determining the transport affinity of the substrates tested.
Both Vmax and Km changes were seen in the kinetics of the R50H variant relative to wild-type hOAT1 (both ∼50–60% lower) or the K525I variant, which was identical to wild type. However, although differences in Vmax were easily seen, their significance remains less certain, since they may reflect either changes in intrinsic transporter function or simply a different level of expression in the oocyte system. As shown in Figs. 2, 3, 4, 5, all three variants were expressed in the plasma membrane and mediated transport. Unfortunately, attempts to detect differences in variant expression using antibodies against hOAT1 were not successful, and both possibilities remain. However, this uncertainty serves to highlight the critical need to determine hOAT1 expression levels in people carrying the R50H SNP, before the full impact of this mutation on anionic drug transport can be determined.
There is no such complication in the interpretation of the changes in Km. These changes do demonstrate a substantial increase in substrate affinity for the R50H variant. This increase seemed to follow the rank order of substrate affinity for the wild-type hOAT1, i.e., the greatest increase in the affinity of the R50H variant was seen for those substrates demonstrating the weakest affinity for the wild-type transporter. In agreement with these data, several reports have shown that the functional properties of variant transporters vary with the substrate(s) tested (Kerb et al., 2002; Leabman et al., 2002; Ohashi et al., 2002). This study supports the clear need to study a range of substrates when investigating the phenotype of polymorphic transporters.
A recent study by Shu et al. (2003) suggested that the degree of chemical change and the evolutionary conservation of the amino acid changed in a polymorphism could be predictive of the functional consequences. For the nonsynonymous SNPs identified in this study, lysine to isoleucine is considered a radical chemical change [Grantham value of 102 (Grantham, 1974), PSIC score of 2.31 (Ng and Henikoff, 2002), and a SIFT score of 0.0 (Sunyaev et al., 2001)]; however, residue 525 does not seem to be conserved in OAT1 and OAT3 (Fig. 1A). It is interesting that this substitution was predicted to disrupt function, but functional differences were not detected. In contrast, however, arginine to histidine is not considered to be a substantial chemical change (Grantham value of 29, PSIC score of 1.82, and SIFT score of 0.17), but this residue does lie within a highly conserved region of hOAT1 (Fig. 1A). Indeed, R50 resides within the first extra cellular loop of hOAT1—a structure that is conserved throughout the families of organic anion and cation transporters (Burckhardt and Wolff, 2000; Wright and Dantzler, 2004). Furthermore, several residues within this loop have previously been shown to be important for transporter function (Bleasby et al., 2002). Thus in this study, evolutionary conservation of amino acid residues seems to be more predictive of function than chemical changes.
This study has focused primarily on the two nonsynonymous SNPs detected in hOAT1; however, seven synonymous SNPs were also detected. A high ratio of synonymous to nonsynonymous SNPs was also observed by Leabman et al. (2002) in the organic cation transporter hOCT2, and it was suggested that this could be indicative of a lack of tolerance to changes in transporters involved in the critical role of detoxification. In addition, several SNPs were also found in the 5′- and 3′-untranslated and intronic regions of hOAT1, despite only a small fraction of these regions being sequenced. It is possible that SNPs in these regions of hOAT1 could have important effects on protein expression and modulation; however, until more is known about the regulation of wild-type hOAT1, it is not possible to investigate the potential importance of SNPs in these regions.
As membrane transporter research moves toward the investigation of genetic variation, caution must be observed when interpreting the data, and the wider picture must be taken into account. Transporter expression and function can be modulated by several mechanisms, with genetic variation being just one of them. It must also be considered that many of these transporters are multispecific and thus a particular substrate could use more than one transporter to traverse a membrane (e.g., OAT1 and OAT3 at the basolateral membrane). Thus, under conditions where transporters are altered by polymorphisms, the pathway used for transport may change, but as is often observed with knockout animals (Jonker et al., 2003), a compensatory mechanism or alternative pathway may mean that no overall change would be detected in vivo. However, for an OAT1 substrate, if the rate-determining step for renal elimination is uptake across the basolateral membrane, then even moderate changes in the function of OAT1 could cause significant effects. Since hOAT1 has been suggested as the basis of nephrotoxicity of the nucleoside phosphonate analogs (Ho et al., 2000) and cidofovir does not seem to be a substrate for hOAT3 (A. Roy and T. Cihlar, personal communication), these studies raise the interesting possibility that individuals with genetic variation in hOAT1 such as R50H may handle these drugs differently.
Footnotes
-
This work was supported in part by The Wellcome Trust UK Grant GR061519MF. Work at Lawrence Livermore National Laboratory was performed under auspices of the U.S. Department of Energy by the University of California, contract no. W-7405-ENG-48, and supported in part by Interagency Agreement Y1-ES-8054-05 from National Institute of Environmental Health Sciences, National Institutes of Health. This work was presented previously, in part, in Drug Metabolism Rev (2002) 34 (Suppl 1):83.
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
doi:10.1124/jpet.105.084301.
-
ABBREVIATIONS: OAT, organic anion transporter; hOAT, human organic anion transporter, SNP, single nucleotide polymorphism; PAH, para-aminohippurate; PCR, polymerase chain reaction; ET, energy transfer; SIFT, sorting intolerant from tolerant; PSIC, position-specific independent counts.
-
↵1 Current address: Department of Drug Metabolism, Merck Research Laboratories, Rahway, NJ 07065.
-
↵2 Current address: Epidemiology Division, University of California, Irvine, Irvine, CA.
- Received January 31, 2005.
- Accepted May 19, 2005.
- The American Society for Pharmacology and Experimental Therapeutics