Plasma Urate Level Is Directly Regulated by a Voltage-driven Urate Efflux Transporter URATv1 (SLC2A9) in Humans*

Hyperuricemia is a significant factor in a variety of diseases, including gout and cardiovascular diseases. Although renal excretion largely determines plasma urate concentration, the molecular mechanism of renal urate handling remains elusive. Previously, we identified a major urate reabsorptive transporter, URAT1 (SLC22A12), on the apical side of the renal proximal tubular cells. However, it is not known how urate taken up by URAT1 exits from the tubular cell to the systemic circulation. Here, we report that a sugar transport facilitator family member protein GLUT9 (SLC2A9) functions as an efflux transporter of urate from the tubular cell. GLUT9-expressed Xenopus oocytes mediated saturable urate transport (Km: 365 ± 42 μm). The transport was Na+-independent and enhanced at high concentrations of extracellular potassium favoring negative to positive potential direction. Substrate specificity and pyrazinoate sensitivity of GLUT9 was distinct from those of URAT1. The in vivo role of GLUT9 is supported by the fact that a renal hypouricemia patient without any mutations in SLC22A12 was found to have a missense mutation in SLC2A9, which reduced urate transport activity in vitro. Based on these data, we propose a novel model of transcellular urate transport in the kidney; Remunurate is taken up via apically located URAT1 and exits the cell via basolaterally located GLUT9, which we suggest be renamed URATv1 (voltage-driven urate transporter 1).

Hyperuricemia is a significant factor in a variety of diseases, including gout and cardiovascular diseases. Although renal excretion largely determines plasma urate concentration, the molecular mechanism of renal urate handling remains elusive. Previously, we identified a major urate reabsorptive transporter, URAT1 (SLC22A12), on the apical side of the renal proximal tubular cells. However, it is not known how urate taken up by URAT1 exits from the tubular cell to the systemic circulation. Here, we report that a sugar transport facilitator family member protein GLUT9 (SLC2A9) functions as an efflux transporter of urate from the tubular cell. GLUT9-expressed Xenopus oocytes mediated saturable urate transport (K m : 365 ؎ 42 M). The transport was Na ؉ -independent and enhanced at high concentrations of extracellular potassium favoring negative to positive potential direction. Substrate specificity and pyrazinoate sensitivity of GLUT9 was distinct from those of URAT1. The in vivo role of GLUT9 is supported by the fact that a renal hypouricemia patient without any mutations in SLC22A12 was found to have a missense mutation in SLC2A9, which reduced urate transport activity in vitro. Based on these data, we propose a novel model of transcellular urate transport in the kidney; Remunurate is taken up via apically located URAT1 and exits the cell via basolaterally located GLUT9, which we suggest be renamed URATv1 (voltage-driven urate transporter 1).
Urate (uric acid), an end product of purine metabolism in humans because of the genetic silencing of hepatic uricase, is now recognized as a natural antioxidant that has neuroprotective properties (1). Despite its beneficial role, elevation of the serum urate level is correlated with gout, hypertension, and cardiovascular and renal diseases (1,2). The kidney plays a dominant role in maintaining plasma urate levels through the excretion process; it eliminates ϳ70% of the daily urate production (3). Therefore, it is important to understand the mechanism of renal urate handling because underexcretion of urate has been demonstrated in the majority of hyperuricemia patients (4).
Since urate is a weak acid at physiological pH (pK a , 5.75), it hardly permeates the plasma membrane of cells in the absence of transport proteins (3). In 2002, we identified a long hypothesized urate-anion exchanger, URAT1, 2 encoded by SLC22A12, that localized on the apical side of the renal proximal tubule (5). Despite several potential candidate proteins for urate transport such as UAT (uric acid transporter), OAT1 (organic anionic transporter 1), OAT3, OAT4, OATv1/NPT1 (sodium phosphate transporter 1), MRP4 (multidrug resistance-associated protein), and OAT10 (6 -10), URAT1 is the sole transporter whose physiological role in renal urate reabsorption is established, based on the fact that lossof-function mutations in URAT1 cause renal hypouricemia (5). However, it is not known how urate taken up via URAT1 exits from the tubular cell (11). Moreover, there are patients with renal hypouricemia who had no mutation in SLC22A12, suggesting the existence of a non-URAT1-mediated urate reabsorption system (12,13). Here we report a previously unknown urate transporter on the basolateral side of the renal proximal tubule, which is likely to act in tandem with URAT1 for urate reabsorption in its physiological role in vivo in humans.
Functional Characterization of GLUT9-GLUT9 isoform 1 and 2 cDNAs were purchased from OriGene Technologies (isoform 2) and Open Biosystems (isoform 1). In vitro transcription and injection of capped cRNA into oocytes were performed as described previously (5,16). Oocytes were maintained in Barth's buffer for 2-3 d at 18°C before use. The ND96 buffer contained (in mM) 96 NaCl, 2 KCl, 1 MgCl 2 , 1.8 CaCl 2 , and 5 HEPES buffer (pH 7.4). In the 0 Cl Ϫ bath, Cl Ϫ was replaced with an equimolar amount of gluconate. Kinetic parameter for the uptake of urate was estimated from the following equation: v ϭ * This work was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan, the Japan Society for the Promotion of Science, the Takeda Science Foundation, the Salt Science Research Foundation (Grant 0721), the Gout Research Foundation of Japan, and the Shimabara Foundation and Kyorin University School of Medicine (Collaborative Project 2008). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
where v is the rate of substrate uptake (pmol/h/oocyte), S is the substrate concentration in the medium (M), K m is the Michaelis-Menten constant (M), and V max is the maximum uptake rate (pmol/h/oocyte). These kinetic parameters were determined by the Eadie-Hofstee plot.
The trans-stimulation experiments were done as described previously (17). The experiments were performed using three batches of oocytes, and results from the representative experiments are expressed as mean Ϯ S.E. Statistical significance was determined by Student's t test.
Mutation Analysis and Construction of Mutant cDNA-For the GLUT9 sequence determination in renal hypouricemia patients and normal control subjects, we used the primers described by S. Li with slight modification (18). Institutional approval was obtained at each participating site. High molecular weight genomic DNA was extracted from peripheral whole blood cells and was amplified by PCR. The PCR prod-ucts were sequenced in both directions using a 3130xl genetic analyzer (Applied Biosystems). To generate a GLUT9 mutant (P412R), we performed site-directed mutagenesis using the QuikChange site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions (19). The mutagenic oligonucleotide primers for generation of P412R mutant were 5Ј-CACGCC-CCCTGGGTCCGCTACCTGAGT-ATCGTG-3Ј (forward) and 5Ј-CAC-GATACTCAGGTAGCGGACCC-AGGGGGCGTG-3Ј (reverse). Proper construction of the mutated cDNA was confirmed by complete sequencing.

RESULTS
To identify novel renal urate transporters, we performed a homology search (Blastp) against the Swiss-Prot protein data base in the National Center for Biotechnology Information (NCBI) using the human and mouse sequences for URAT1/Urat1 (SLC22A12/ Slc22a12) (5,20) and human OAT4 (SLC22A11) (21). Surprisingly, we found that several members of the facilitated glucose transporter (SLC2) family (GLUT6, -9, -10, -12, and -14) have remote similarities to SLC22A11. Among these molecules, we decided to characterize one of the extended (class II) SLC2 family members named GLUT9 (which is encoded by SLC2A9) because it localizes mainly in the kidney and the liver (22). Human GLUT9 was originally identified as a gene of unknown function (23). Although glucose transporter activity of GLUT9 was demonstrated, it does not seem as efficient as the classical (class I) glucose transporter GLUT4 (24,25). Human GLUT9 has two splice variants: isoform 1 (NM_020041) and isoform 2 (NM_001001290) (supplemental Fig. S1A). The difference between these two isoforms lies in the presence of the first exon only in isoform 1, which results in differential targeting in polarized Madin-Darby canine kidney cells (25).
First, we examined the membrane expression and urate transport activities of both isoforms of GLUT9 using the Xenopus oocyte expression system. Both isoforms of GLUT9 were expressed on the plasma membrane when GLUT9 cRNAs were injected into oocytes (supplemental Fig. S2A), and both isoforms had equivalent urate transport activity (supplemental Fig. S2B). We used isoform 2 for further characterization. The uptake rate of [ 14 C]urate in oocytes expressing GLUT9 was ]urate in water-injected (control) oocytes and GLUT9-expressing oocytes was measured during a 120-min incubation. B, transport kinetics of GLUT9. Urate transport was measured in control oocytes and GLUT9-expressing oocytes over a urate concentration range of 10 -1,500 M. The GLUT9-specific transport activity in GLUT9-expressing oocytes was calculated by subtracting the transport activity in control oocytes. These GLUT9-specific transport activities were used in kinetic analysis. Inset, Eadie-Hofstee plot. C, [ 14 C]urate uptake by GLUT9 with K ϩ replacement of external Na ϩ . The transport rate of [ 14 C]urate (20 M) in control oocytes and GLUT9-expressing oocytes was measured for 1 h in the presence or absence of extracellular Na ϩ , K ϩ , and Cl Ϫ . ***, p Ͻ 0.001 when compared with ND96. D, the pH dependence of GLUT9-mediated urate transport. The [ 14 C]urate (20 M) transport rate was measured in control oocytes (open bars) and GLUT9-expressing oocytes (filled bars) in the ND96 solution. To prepare uptake solutions with different pH levels, MES-NaOH (5.5), HEPES-NaOH (6.5 and 7.4), and Tris-HCl (8.5) were used as the buffer systems. At acidic extracellular pH values in the ND96 bath, the net urate uptake rate was increased. ***, p Ͻ 0.001 when compared with pH 7.4. All data are mean Ϯ S.E. (error bars) with n ϭ 8 -10.
9-fold higher than that in control oocytes, whereas much lower uptake rates of [ 14 C]glucose and [ 14 C]fructose were observed (supplemental Fig. S2C). GLUT9 did not show any significant uptake of representative organic anionic substrates such as para-aminohipurate, estrone sulfate, or salicylate, nor of substrates known as interactors of classical renal urate transport systems (including URAT1) such as lactate, nicotinate, ␤-hydroxybutyrate, or salicylate (supplemental Fig. S2C). Thus, GLUT9 had a narrower substrate specificity than URAT1.
Next, we examined the urate transport properties of GLUT9. Xenopus oocytes expressing GLUT9 exhibited time-dependent transport of [ 14 C]urate (Fig. 1A). The GLUT9-mediated uptake of urate manifested saturable kinetics and followed the Michaelis-Menten equation. The Eadie-Hofstee plot yielded a K m of 365 Ϯ 42 M and a V max of 5,521 Ϯ 291 pmol/h/oocyte (mean Ϯ S.E. of five separate experiments) for urate, indicating that GLUT9 has a high affinity for urate similar to URAT1 (Fig. 1B). Elimination of extracellular Na ϩ did not affect urate transport by GLUT9, indicating that it did not involve a direct Na ϩ urate cotransport. GLUT9 activity was sensitive to membrane potential because the elevation of external K ϩ (which depolarizes the plasma membrane of a Xenopus oocyte) facilitated urate uptake (Fig. 1C). This voltage sensitivity should favor the efflux of urate from the tubular cells because of a net negative charge within the cell. We then measured the effect of an outward Cl Ϫ gradient (intracellular to extracellular) on urate uptake. Imposing a Cl Ϫ gradient by complete removal of external Cl Ϫ did not accelerate the urate uptake via GLUT9, indicating that GLUT9 does not have the exchange mechanism for inorganic Cl Ϫ observed in URAT1 (Fig. 1C). In addition, unlike URAT1, a dependence of urate transport activity on extracellular pH was observed for GLUT9 (Fig. 1D).
To further investigate the substrate selectivity of GLUT9, an inhibition study was performed. The cis-inhibitory effect of various compounds at 1 mM (except for benzbromarone, at 50 M) on GLUT9-mediated [ 14 C]urate (10 M) uptake was examined (Table 1). GLUT9 again exhibited pharmacological properties different from those of URAT1 (5). Although the urate transport via GLUT9 was inhibited strongly by urate, hexoses such as glucose and fructose, monocarboxylates such as lactate, nicotinate, and orotate, or ketone bodies such as acetoacetate and ␤-hydroxybutyrate did not inhibit urate uptake via GLUT9. These results are consistent with the initial observation that   7.4). The values were expressed as percentages of uptake under control conditions (no inhibitors). Data are mean Ϯ S.E. with n ϭ 8 -10. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001, compared with the uptake in control.

Substrates (1 mM) Uptake (% of control)
Control ( urate appeared to be the only substrate for GLUT9 (supplemental Fig. S2C). We also tested the influence of uricosuric substances (5,26). Probenecid and benzbromarone (50 M) cis-inhibited urate uptake moderately. Interestingly, pyrazinoate (PZA), the active metabolite of pyrazinamide widely used as an antituberculosis agent and known as an interactor with URAT1 (5), did not exert any inhibitory effect on urate uptake via GLUT9. Losartan, an angiotensin II receptor blocker, moderately cis-inhibited urate uptake. These results demonstrated that it exhibited responsiveness to various drugs potentially affecting renal urate handling different from that of URAT1, indicating that GLUT9 can be a potential target for novel uricosuric agents.
Consistent with our aforementioned idea that GLUT9 may act as an efflux transporter under a physiological condition, GLUT9-expressing oocytes preloaded with [ 14 C]urate showed a time-dependent efflux of radioactivity when incubated in the standard uptake solution ( Fig. 2A). Given its predominantly basolateral membrane localization in proximal tubular cells (25), we propose that GLUT9 is responsible for transferring urate from the cell to the peritubular interstitium as a second step in reabsorption following the initial step mediated by URAT1 (Fig. 2B). When the efflux of radioactivity from the oocytes preloaded with [ 14 C]urate was compared in the absence and presence of extracellular urate, glucose, and fructose, the efflux was stimulated by the extracellular urate but not by glucose or fructose (supplemental Fig. S3B). Therefore, urate transport via GLUT9 would not be altered by the normal serum levels of these sugars.
Idiopathic renal hypouricemia (or, simply, renal hypouricemia: Mendelian Inheritance in Man (MIM) 220150) is a hereditary disease characterized by abnormally increased renal urate clearance and caused by an isolated inborn error of membrane transport for urate in the renal proximal tubule (27). As mentioned previously, there are renal hypouricemia patients who have no mutation in SLC22A12 (12,13). If GLUT9 acts in the urate reabsorption pathway in tandem to URAT1, it is possible that such patients carry mutation(s) in the SLC2A9 gene.
One of these patients (a 36-year-old female) had a plasma urate level of 2.4 mg/dl (the normal value is ϳ5 mg/dl), similar to the values found in renal hypouricemia patients with a URAT1 heterozygous mutation such as W258X. Analysis of the GLUT9-coding region from the genomic DNA of the patient revealed the presence of a heterozygous C to G alteration at nucleotide 1296 within exon 11 of SLC2A9 (supplemental Figs. S1 and 3A), changing proline 412 (CCC) to arginine (CGC). This residue (Pro-412) was located in loop 9 facing the extracellular side (supplemental Fig. S1B) and was conserved in both GLUT5 and GLUT9 in SLC2 family members (28). This P412R mutation was not identified in 50 randomly chosen control Japanese individuals.
To confirm the function of mutant GLUT9, we analyzed urate transport activity after injection of wild-type or mutated GLUT9 cRNA into oocytes and found that this substitution (P412R) significantly reduced urate transport activity (Fig. 3B). Expression level of both wild-type and mutant GLUT9 was comparable in the plasma membrane by immunostaining (Fig. 3C). FIGURE 3. Mutation within SLC2A9 is associated with idiopathic renal hypouricemia. A, electropherograms of partial sequences of exon 11 of SLC2A9 showing the heterozygous point mutation P412R found in a 36-year-old female patient exhibiting clinical features compatible with idiopathic renal hypouricemia. B, a disease-associated GLUT9 mutant from the patient significantly decreased the urate transport activity. The uptake of urate (20 M) was measured in oocytes injected with wild-type or mutated cRNA. Data are mean Ϯ S.E. with 9 or 10 oocytes per group. ***, p Ͻ 0.001 when compared with wild type. C, subcellular localization of the wild-type and P412R mutant in oocytes. Immunodetection with an anti-GLUT9 antibody showed that the wild-type and P412R proteins are expressed at the plasma membrane, whereas fluorescence levels were undetectable in oocytes injected with water (control).

DISCUSSION
In this study, a sugar transport facilitator family protein GLUT9 was found to act as a voltage-driven urate transporter. This characteristic is well suited for urate efflux transporter from the cell. This finding will complete a model of urate reabsorption in the renal tubular cell, where urate in the urinary lumen is taken up via URAT1 and intracellular urate exits from the cell to the interstitium/blood space via GLUT9 (Fig. 2B). Based on our model (Fig. 2B), we predict that intact function of both URAT1 and GLUT9 is necessary for normal urate reabsorption in the renal proximal tubule since both transporters act on the same pathway in tandem. This prediction was supported in vivo by the existence of mutations with reduced function in SLC2A9 in a patient with idiopathic renal hypouricemia who had no mutations in SLC22A12. Because this mutation (P412R) replaces a non-charged residue proline with a positively charged residue arginine, we speculate that it affected the interaction of GLUT9 with its substrate urate.
While we were in the process of functional characterization of GLUT9, a few genetic studies have been published that relate variations in/near SLC2A9 to the plasma urate level. For example, Li et al. (18) reported the association of serum urate levels with common genetic variants in the SLC2A9 gene. In addition, our results should give strong (patho)physiological support to two recent whole genome association studies that demonstrated that the serum urate levels were correlated with single nucleotide polymorphisms in or very close to SLC2A9 (29,30). Taken together, proteins encoded by SLC2A9 functioned as urate transporters in the kidney, mainly acting to excrete reabsorbed urate from the tubular cell, and their urate transport activity affected the plasma urate level. Therefore, we propose that the proteins encoded by SLC2A9 be called URATv1 (voltage-driven urate transporter 1), instead of GLUT9.