Identification of an Amino Acid Transporter Associated with the Cystinuria-related Type II Membrane Glycoprotein*

We identified an amino acid transporter that is associated with the cystinuria-related type II membrane glycoprotein, rBAT (related to b0,+ amino acid transporter). The transporter designated BAT1 (b0,+-type amino acid transporter 1) from rat kidney was found to be structurally related to recently identified amino acid transporters for system L, system y+L, and system x−C, which are linked, via a disulfide bond, to the other type II membrane glycoprotein, 4F2hc (4F2 heavy chain). In the nonreducing condition, a 125-kDa band, which seems to correspond to the heterodimeric complex of BAT1 and rBAT, was detected in rat kidney with anti-BAT1 antibody. The band was shifted to 41 kDa in the reducing condition, confirming that BAT1 and rBAT are linked via a disulfide bond. The BAT1 and rBAT proteins were shown to be colocalized in the apical membrane of the renal proximal tubules where massive cystine transport had been proposed. When expressed in COS-7 cells with rBAT, but not with 4F2hc, BAT1 exhibited a Na+-independent transport of cystine as well as basic and neutral amino acids with the properties of system b0,+. The results from the present investigation were used to establish a family of amino acid transporters associated with type II membrane glycoproteins.

Two type II membrane glycoproteins with single transmembrane domains have been implicated in amino acid transport via the plasma membrane (1). The first is rBAT 1 (related to b 0,ϩ amino acid transporter), which is also known as D2 or NBAT (2)(3)(4). It induces the Na ϩ -independent transport of cystine as well as basic and neutral amino acids with the properties of system b 0,ϩ when expressed in Xenopus oocytes. rBAT was found to be related to the genetic disease cystinuria, in which defects in amino acid reabsorption in the renal proximal tubules lead to urinary loss of cystine and basic amino acids (5). Because rBAT is not typical of transporters with multiple transmembrane structures, it was proposed that rBAT itself is not a transporter but is associated with unidentified amino acid transporters to activate them (1). The second one is 4F2hc (the heavy chain of the 4F2 antigen), which was originally identified as a cell surface antigen that is up-regulated upon lymphocyte activation (6,7). Because of the amino acid sequence similarity between rBAT and 4F2hc, experiments were performed in which 4F2hc is expressed in Xenopus oocytes to show that 4F2hc also induces amino acid transport (8,9).
The 4F2 antigen is a heterodimeric complex composed of two subunits, a heavy chain and a light chain that are linked via a disulfide bond (6,7). Recently, by performing expression cloning, we identified a 4F2 light chain to be an amino acid transporter (10). Transporters for amino acid transport systems L, y ϩ L and x Ϫ C were shown to be 4F2 light chains which require 4F2hc for their functional expression (10 -18).
We report here the identification of the first rBAT-associated transporter. It is structurally related to 4F2hc-associated transporters, thereby establishing a family of amino acid transporters associated with type II membrane glycoproteins.

EXPERIMENTAL PROCEDURES
cDNA Cloning-The cDNA for a mouse expressed sequence tag (Gen-Bank TM /EBI/DDBJ accession number AA162896) showing nucleotide sequence similarity to rat LAT1 was obtained from the integrated and molecular analysis of genomes and their expression. The ϳ0.8-kb EcoRI fragment was excised from the cDNA (IMAGE cDNA clone number 578502) and labeled with [ 32 P]dCTP ( T7 Quick Prime, Amersham Pharmacia Biotech) for use as a probe for screening a cDNA library prepared from rat kidney poly(A) ϩ RNA using Superscript Choice System (Life Technologies, Inc.) (19,20). Screening of the cDNA library and isolation of positive plaques were performed as described (19,20). The cDNA was sequenced in both directions by the dye terminator cycle sequencing method (Perkin-Elmer and Applied Biosystems). Transmembrane regions of proteins were predicted based on SOSUI algorithm (15,21).
Northern Analysis-High stringency Northern blot analysis of rat poly(A) ϩ RNA (3 g/lane) was performed as described previously (10,22). The EcoRI fragment of BAT1 cDNA corresponding to 1-793 base pairs was labeled with 32 P using T7 Quick Prime kit (Amersham Pharmacia Biotech).
Anti-peptide Antibody-Oligopeptides (QMLMEVVPPEKDPEC) corresponding to amino acid residues 474 -487 of BAT1 and (SD-VDTHAVSLEKGEC) to amino acid residues 629 -642 of rat rBAT (2) were synthesized. The C-terminal cysteine residues were introduced for conjugation with keyhole limpet hemocyanine. Anti-peptide antibodies were produced as described elsewhere (23). For immunohistochemical * This work was supported in part by grants from the Ministry of Education, Science, Sports and Culture of Japan (grants-in-aid for Scientific Research and High-Tech Research Center), the Scientific Research Promotion Fund of the Japan Private School Promotion Foundation, the Japan Science and Technology Corporation, the Kato Memorial Bioscience Foundation, the Research Fund of Mitsukoshi Health and Welfare Foundation (1998), the Uehara Memorial Foundation, and the Toyota Physical & Chemical Research Institute. 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB029559.
Western Blotting-Kidney membranes were prepared as described elsewhere (25). Protein samples were heated at 100°C for 5 min in sample buffer in either the presence or absence of 5% 2-mercaptoethanol and subjected to sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (26). The separated proteins were transferred electrophoretically to a Hybond-P polyvinylidene difluoride transfer membrane (Amersham Pharmacia Biotech). The membrane was treated with diluted anti-rBAT antiserum (1:40,000) or anti-BAT1 antiserum (1:4,000), and then with horseradish peroxidase-conjugated anti-rabbit IgG as the secondary antibody (Jackson ImmunoResearch Laboratories, Inc.) (26). The signal was detected by an ECL plus system (Amersham Pharmacia Biotech). To verify the specificity of immunoreactions by absorption experiments, the membranes were treated with primary antibodies in the presence of antigen peptides (50 g/ml).
For amino acid uptake measurements, cells were used three days after plating (90ϳ100% confluence) on 24-well plates. Dulbecco's modified phosphate-buffered saline (Dulbecco's PBS) containing 14 C-amino acids was added after preincubation of the cells in Dulbecco's PBS for 10 min at 37°C. The reaction was terminated by removing the uptake medium followed by washing three times with ice-cold Dulbecco's PBS. Then, the cells were solubilized with 0.1 N NaOH, and radioactivity was counted. Because [ 14 C]cystine (100 M) uptake was linearly dependent on the incubation time up to 5 min, for all the experiments, uptakes were measured for 2 min, and the values were expressed as picomoles/ milligram protein/minute. K m and V max of amino acid substrates were determined using the Eadie-Hofstee equation based on the BAT1-mediated amino acid uptakes measured at 3, 10, 30, 100, 300, and 1,000 M. The BAT1-mediated amino acid uptakes were calculated as differences between the means of uptakes of the COS-7 cells transfected with cDNAs and those of the mock-transfected controls.
For the measurements of the uptake of radiolabeled amino acids in the present study, four wells were used for each data point. Each data point in the figures represents the mean Ϯ S.E. of uptake values (n ϭ 4). To confirm the reproducibility of the results, three separate experiments were performed for each measurement using different batches of COS-7 cell transfectants except for K m and V max determination. The results from representative experiments are shown in the figures.

RESULTS
Structural Features of BAT1-A cDNA clone with a 1,746base pair insert was isolated from rat kidney cDNA library. It contained an open reading frame from nucleotides 170 to 1,630 encoding a 487-amino acid protein designated BAT1 (b 0,ϩ -type amino acid transporter 1). The amino acid sequence of BAT1 exhibited remarkable homology to those of rat system L transporters LAT1 (43% identity) (10) and LAT2 (43%) (15), human y ϩ L transporters y ϩ LAT1 (42%) (16) and KIAA0245/y ϩ LAT2 (44%) (16), and a mouse system x Ϫ C transporter xCT (43%) (18). The amino acid sequence of BAT1 also exhibited significant homology to those of system y ϩ transporters CAT1ϳ4 (ϳ30%) from mice and humans (5). As shown in Fig. 1, 12 transmembrane regions were predicted from the amino acid sequence. There is a conserved cysteine residue (BAT1 amino acid residue 144) in the putative extracellular loop between predicted transmembrane domains 3 and 4 (28).
Tissue Distribution and Localization of Expression-In Northern blot analysis, a 1.9-kb message was expressed in kidney, jejunum, and ileum for BAT1 ( Fig. 2A). In the immu-nohistochemical analysis of rat kidney tissue, BAT1 immunoreactivity was detected in the apical membrane of the proximal tubules (Fig. 2B, left panel). It was stronger in the proximal convoluted tubules than in the proximal straight tubules, with   FIG. 1. Deduced amino acid sequence of BAT1. The predicted transmembrane regions of BAT1 numbered 1ϳ12 are shown by lines above the sequence. A potential tyrosine kinase-dependent phosphorylation site is located at residue 99 (indicated by #). Protein kinase C-dependent phosphorylation sites are predicted at residues 5, 51, 169, 345, and 399, among which those at residues 5 and 345 are predicted to be located intracellularly (indicated by *). A potential cAMP-dependent phosphorylation site is located at residue 350. Arrows indicate proximal straight tubules with strong rBAT immunoreactivity and with faint, but significant, BAT1 immunoreactivity. faint but significant immunostaining detected in the proximal straight tubules (Fig. 2B, left panel: arrow). rBAT immunoreactivity was also detected in the apical membrane of the proximal tubules (Fig. 2B, right panel). The expression of rBAT was strongest in the proximal straight tubules. The expression of rBAT was still detected in the convoluted portion of the proximal tubules (Fig. 2B, right panel: arrowhead), although it was weak compared with that in the straight portion (Fig. 2B, right  panel: arrow). Colocalization of BAT1 and rBAT proteins was demonstrated in the apical membrane, in particular, in the convoluted portion of the proximal tubules (Fig. 2B, right and left  panels: arrowhead).
Protein Characterization under Nonreducing and Reducing Conditions-Western blot analyses were performed on the membrane fraction prepared from rat kidney in the presence or the absence of 2-mercaptoethanol (Fig. 3A). For BAT1, 251-, 125-, and 38-kDa bands were detected in the absence of 2-mercaptoethanol (nonreducing condition). In the presence of 2-mercaptoethanol (reducing condition), the 251-and 125-kDa bands disappeared, and a 41-kDa band was detected instead. The 38-kDa band seemed to be nonspecific, because it was not affected by the presence of antigen peptides in the absorption experiments, whereas the 251-, 125-, and 41-kDa bands disappeared (Fig. 3A). For rBAT, 231-, 125-, 86-, and 62-kDa bands were detected under the nonreducing condition. Under the reducing condition, the 231-and 125-kDa bands disappeared, and a 90-kDa band was detected instead. The 86-and 62-kDa bands seemed to be nonspecific because they were not affected by the presence of antigen peptides in the absorption experiments (Fig. 3A).
Functional Expression in COS-7 Cells-Because Xenopus oocytes are not suitable for the functional expression of BAT1 due to the abundant expression of endogenous transporters associated with rBAT (2-4), the functional properties of BAT1 were determined by transient expression in COS-7 cells. When BAT1, rBAT, or 4F2hc was solely expressed in COS-7 cells, significant [ 14 C]L-cystine uptake was not detected compared with mock-transfected controls. Large [ 14 C]L-cystine uptake was, however, detected when BAT1 was coexpressed with rBAT but not with 4F2hc, indicating that BAT1 requires rBAT, not 4F2hc, for its functional expression (Fig. 3B). Thus, in subsequent experiments to determine the functional characteristics of BAT1, BAT1 was coexpressed with rBAT in COS-7 cells.
BAT1-mediated [ 14 C]L-cystine uptake was dependent on neither Na ϩ nor Cl Ϫ (data not shown). The uptake was saturable and followed the Michaelis-Menten kinetics (see below for K m values). Results from uptake measurements of 14 C-labeled amino acids indicated that BAT1 transported L-cystine, L-lysine, L-arginine, L-ornithine, L-leucine, L-phenylalanine, and L-tyrosine at high levels and other neutral amino acids at low levels (Fig. 4). The uptake values for glycine and proline were particularly low. Acidic amino acids were not transported by BAT1. The K m /V max values of BAT1 for L-cystine, L-lysine, L-arginine, L-ornithine, L-leucine, and L-phenylalanine were 305 Ϯ 43.4 M/1,970 Ϯ 380 pmol/mg of protein/min (mean Ϯ S.E. of three separate experiments), 463 M/2,990 pmol/mg of protein/min, 203 M/2,570 pmol/mg of protein/min, 387 M/ 2,530 pmol/mg of protein/min, 269 M/1,330 pmol/mg of protein/min, and 190 M/1,730 pmol/mg of protein/min, respectively. Consistent with the radiolabeled amino acid uptakes, [ 14 C]L-cystine uptake was inhibited by L-isomers of cystine as well as basic and neutral amino acids in the experiments in which [ 14 C]L-cystine uptake (50 M) was measured in the presence of nonlabeled amino acids (5 mM) (data not shown). System A-selective inhibitor ␣-(aminomethyl)isobutyric acid and system L-selective inhibitor 2-aminobicyclo-(2, 2, 1)-heptane-2carboxylic acid had no or weak inhibitory effect on [ 14 C]Lcystine uptake, consistent with the properties of system b 0,ϩ .

DISCUSSION
In the present study, we have identified a novel protein (BAT1) that is structurally related to amino acid transporters for system L, system y ϩ L and system x Ϫ C (10 -18). Although these transporters require 4F2hc for their functional expression, we found that BAT1 is associated with rBAT and not with 4F2hc to express its function.
In the Western blot analyses for BAT1 and rBAT, we showed that two corresponding bands (125 and 251 kDa/231 kDa) were detected under the nonreducing condition, whereas under the reducing condition, the bands shifted to smaller sizes, 41 kDa for BAT1 and 90 kDa for rBAT, which corresponded to the sizes of BAT1 and rBAT proteins (2,3,29), respectively (Fig. 3A). This suggests that BAT1 and rBAT are linked via disulfide bonds. The 125-kDa band seems to correspond to the heterodimeric complex, whereas the larger bands probably represent multimerized proteins. For transporters that are associated with 4F2hc, it is proposed that a conserved cysteine residue in the extracellular loop between putative transmembrane domains 3 and 4 is responsible for the disulfide bond formation between 4F2hc and transporters (28). The corresponding cysteine residue is also conserved for BAT1 (Cys-144). Because rBAT supports the functional expression of BAT1, whereas 4F2hc does not (Fig. 3B), it is interesting to know what structural feature is responsible for the recognition of partner molecules.
BAT1 is expressed predominantly in kidney and small intestine ( Fig. 2A). The pattern of expression of BAT1 correlates well with that of rBAT (2,3,22). We were able to demonstrate the colocalization of two proteins in the renal proximal tubules (Fig. 2B). BAT1, as well as rBAT, exists in the apical membrane of the proximal tubules, not in the basolateral membrane where 4F2hc localizes (13,30,31), which is consistent with the observation that BAT1 is functionally associated with rBAT and not with 4F2hc (Fig. 3B). The expression of rBAT is the highest in straight convoluted tubules where the expression of BAT1 is fairly low (Fig. 2B) (22,30,31). It is, therefore, speculated that still unidentified transporters exist in the proximal tubules, which are associated with rBAT.
When coexpressed with rBAT in COS-7 cells, BAT1 mediates the Na ϩ -independent transport of cystine as well as basic and neutral amino acids with the properties of system b 0,ϩ (32). The substrate selectivity of BAT1 expressed with rBAT in COS-7 cells is basically consistent with that observed when rBAT is solely expressed in Xenopus oocytes where rBAT is assumed to associate with unidentified oocyte endogenous transporters to activate them (2, 3). The K m values for cystine, lysine, arginine, ornithine, leucine, and phenylalanine lie between ϳ200 and ϳ500 M. Although these values are higher than those measured in Xenopus oocytes expressed solely with rBAT, they are still regarded as high affinity transports (2,3,(33)(34)(35), indicating that the BAT1-rBAT heterodimeric complex is responsible for at least one of the high affinity cystine transport systems in the renal proximal tubules. It has been proposed that the genetic defects of cystine transporters in the renal proximal tubules are responsible for cystinuria (5). The mutation of the rBAT gene was found to be related to autosomal recessive type I cystinuria, whereas type II and type III cystinuria were proposed to be due to mutations of the genes for rBAT-associated transporters (5). It is, therefore, necessary to determine the implications of BAT1 in cystinuria.
In conclusion, we have identified the first rBAT-associated transporter BAT1, which is, surprisingly, structurally related to 4F2hc-associated transporters, thereby establishing a family of amino acid transporters associated with type II membrane glycoproteins. This finding will facilitate the understanding of the mechanisms of molecular association between transporters and their partner type II membrane glycoproteins. The results from the present investigation also suggest the presence of still unidentified transporters that are associated with rBAT.