a4, a unique kidney-specific isoform of mouse vacuolar H+-ATPase subunit a.

The vacuolar-type H+ -ATPase (V-ATPase) translocates protons across membranes. Here, we have identified a mouse cDNA coding for a fourth isoform (a4) of the membrane sector subunit a of V-ATPase. This isoform was specifically expressed in kidney, but not in the heart, brain, spleen, lung, liver, muscle, or testis. Immunoprecipitation experiments, together with sequence similarities for other isoforms (a1, a2, and a3), indicate that the a4 isoform is a component of V-ATPase. Moreover, histochemical studies show that a4 is localized in the apical and basolateral plasma membranes of cortical alpha- and beta-intercalated cells, respectively. These results suggest that the V-ATPase, with the a4 isoform, is important for renal acid/base homeostasis.

In this study, we have isolated a mouse cDNA coding for a fourth isoform (a4) of V-ATPase subunit a. Isoform a4 exhibits high similarities with the a1, a2, and a3 isoforms and is expressed exclusively in kidney. Furthermore, a4 was localized immunochemically in the apical and basolateral plasma membranes of cortical ␣and ␤-intercalated cells, respectively. These results suggest that the V-ATPase with the a4 isoform is a kidney-specific proton pump important for acid/base homeostasis.

EXPERIMENTAL PROCEDURES
cDNA Cloning and Nucleic Acid Blotting-A mouse EST (expressed sequence tag) clone, 2099716 (19), coding for a part of the a4 isoform, was sequenced. To obtain the 5Ј-region of the a4 cDNA, total RNA was prepared from C57BL/6J male mouse kidneys (age, 8 weeks; Japan SLC). Reverse transcriptase-polymerase chain reaction was carried out using gene-specific primers, and the product was ligated with the EST clone to create a full-length cDNA for the a4 isoform. The nucleotide sequence reported in this study will appear in the DDBJ, EMBL, and GenBank TM data bases with the accession number AB050903.
Northern blot analysis was carried out with multiple tissue blots (CLONTECH) as described by Toyomura et al. (7). A probe was prepared from the cDNA clone (between ϩ1660 and ϩ2167 bp, numbering from the first letter of the initiation codon), and labeled with [␣-32 P]dCTP using rediprime TM II random prime labeling system (Amersham Pharmacia Biotech). Each filter was hybridized with the probe using ExpressHyb Hybridization Solution (CLONTECH) at 68°C for 60 min.
Genomic DNA (10 g) from C57BL/6J mice was digested with restriction enzymes, subjected to agarose gel electrophoresis, and then blotted onto a filter. A DNA fragment (between ϩ792 and ϩ1073 bp) of the a4-coding region was used to prepare a labeled probe. Hybridization was performed as described above, and radioactivity was analyzed with a BAS-1000 (Fuji Film).
Preparation of Kidney Cortex Membrane Fraction-All operations were carried out at 4°C. Kidney cortex (about 2.7 g) was obtained from ten ICR mice (age, 8 weeks), and suspended in 14 ml of 10 mM HEPES-KOH pH 7.4 containing 0.25 M sucrose, 10 mM KCl, 5 mM MgCl 2 , 1 mM EDTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and complete protease inhibitor mixture (amount recommended by the Company) (Roche Diagnostics). The suspension was homogenized in a Wheaton homogenizer. The supernatant fraction, obtained by centrifugation at 500 ϫ g for 5 min, was centrifuged at 10,000 ϫ g for 15 min. The supernatant was centrifuged at 100,000 ϫ g for 30 min. The precipitate was suspended in 2.4 ml of PBS containing 1 mM EDTA and 10% glycerol and stored at Ϫ80°C until use.
Immunochemistry-The synthetic peptide KHQKSQLQSFTIHE-DAVEGDH (positions 665-685 of the a4 isoform a region, not homologous to a1, a2, or a3) was used to immunize albino rabbits for antibodies against the a4 isoform. The resulting serum was applied to a peptideconjugated column, and antibodies were further purified using a recombinant protein-conjugated column. Antibodies against human Cl Ϫ / HCO 3 Ϫ exchanger (AE1) were obtained by injecting a human polypeptide (His 834 -Val 911 ) into a rabbit (20). Immunoblot was carried * 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) AB050903.
out as described previously (7). For immunoprecipitation experiments, kidney cortex membranes (2 mg/ml) were incubated at 4°C for 2 h in 50 mM Tris-HCl (pH 7.5), 1 mM EDTA, 2% octylglucoside, and 200 mM NaCl. The supernatant fraction, obtained by centrifugation at 100,000 ϫ g for 30 min, contained solubilized V-ATPase. Histology-ICR mice (age, 8 -10 weeks; Japan SLC) were anesthetized and perfused briefly with PBS, pH 7.4 and then fixed with 4% paraformaldehyde in PBS for 15 min. The kidneys were removed, incubated in the same solution overnight at 4°C, and cut transversely into 5-mm thick blocks. They were successively infiltrated with 30% sucrose in PBS, embedded in OCT compound (Miles), and stored frozen. Frozen blocks were sectioned at 4-m thickness and mounted on MAScoated slides (Matsunami Glass). The sections were rinsed with PBS containing 0.05% Tween 20 (PBST), and incubated for 30 min with 0.3% H 2 O 2 in methanol. They were blocked with PBS containing 1.5% normal goat serum and incubated at 4°C with anti-a4 antibodies (1 g/ml) in PBST containing 0.1% bovine serum albumin. After washing with PBST, the sections were further incubated with a biotinylated goat anti-rabbit IgG (VECTOR Laboratories) in PBS containing 1.5% normal goat serum, and then developed using VectaStain Elite ABC reagent (VECTOR Laboratories). After washing with PBST, the sections were stained with 0.02% 3,3Ј-diaminobenzidine and 0.005% H 2 O 2 in 0.05 M Tris-HCl (pH 7.6) for 1 to 2 min. The immunostained sections were counterstained with hematoxylin and mounted with Eukitt (Kindler). Renal sections were also stained with fluorescein isothiocyanate-labeled peanut agglutinin (10 g/ml) (VECTOR Laboratories) for 30 min at room temperature, and mounted by PermaFluor (Shandon) after washing with PBST.
Electron Microscopy-The pre-embedding silver enhancement immunogold method was used, as described previously (21). Mice were anesthetized with ether and perfusion-fixed for 10 min with 4% paraformaldehyde and 0.2% picric acid in 0.1 M sodium phosphate (pH 7.4). The fixed kidneys were removed and incubated for another 50 min in the same solution. Cryo sections (6 m in thickness) were reacted with 5 g/ml anti-a4 antibodies overnight, followed by incubation with colloidal gold (1.4-nm diameter)-conjugated secondary antibodies. The gold labeling was intensified using a silver enhancement kit (Nano Probes).

RESULTS
Identification of the a4 Isoform of V-ATPase Subunit a-Three subunit a isoforms have been identified in mouse and chicken (7,13,15). This finding prompted us to search for the fourth mammalian isoform. Analysis of the mouse kidney EST library (19) led us to identify a clone 2099716, coding for a protein highly homologous to mouse a1, a2, or a3 isoforms. The clone encoded an open reading frame of 280 amino acid residues, but apparently lacked a 5Ј-terminal region. The 5Ј-region was obtained by reverse transcriptase-polymerase chain reaction from mouse kidney RNA.
The entire cDNA was 3014 bp (not including polyadenylation) and contained a coding region for 833 amino acid residues with two potential N-linked glycosylation sites (Asn 367 and Asn 489 ). The predicted protein exhibited 63, 54, and 48% identity with the mouse a1, a2, and a3 isoforms, respectively. A hydropathy plot of the protein suggested a structure with nine transmembrane regions (Fig. 1, I-IX), i.e. similar to other isoforms. The amino-terminal domain (ϳ60 residues) and all transmembrane regions (except VI) of the protein were highly similar to other isoforms (a1, 74.2%; a2, 65.6%; a3, 64.6%). Mutational analysis of the yeast VPH1 gene showed 11 amino acid residues that are essential for activity, assembly, or intracellular sorting of subunit a (22). All of them except His 729 (yeast numbering) were completely conserved in all mouse isoforms (Fig. 1). Replacement of His 729 affects V-ATPase activity (22), but this residue was replaced by Asn in all mouse isoforms (Fig. 1, asterisk). Based on the structure and sequence similarities, the protein encoded by the cDNA was named the a4 isoform. Southern blot analysis of mouse genomic DNA gave  isoforms (a1, a2, a3, and a4) and human ATP6N1B were aligned to obtain maximal homology. Boxed residues are identical in the four isoforms. Putative transmembrane domains (IϳIX) were defined from hydropathy analysis. The shaded residues correspond to those essential for yeast subunit a (Vph1p) (23). Yeast His 729 corresponds to the asparagine residue (asterisk) in mouse isoforms. a single band ( Fig. 2A), indicating that only one gene for a4 is present in the mouse genome.
Kidney-specific Expression of the a4 Gene-It was of interest to determine which tissue(s) expresses the a4 gene. A 3.3-kb single transcript was detectable exclusively in kidney (Fig. 2B), i.e. no signals were observed in heart, brain, spleen, lung, liver, skeletal muscle, or testis even after longer exposure of the filter (data not shown). These results were consistent with the EST clone 2099716 (19) as isolated from the kidney library. These findings suggest that the a4 isoform is a kidney-specific subunit a. During embryonic development, the a4 gene was transcribed from 15-dpc embryos, whereas no significant signal was detectable in 7-and 11-dpc (Fig. 2B).
Detection of the a4 Isoform in Kidney-Antibodies were generated using a synthetic peptide for immunochemical studies on the localization of the a4 isoform. The affinity-purified antibodies specifically recognized a single 94-kDa protein in a lysate of yeast cells carrying an a4 expression plasmid (Fig.  2C); i.e. they did not react with other isoforms (data not shown). The position of the band matched the molecular weight (95,603) calculated from the deduced amino acid sequence. Bands corresponding to 95ϳ106 kDa were observed for kidney membranes (Fig. 2C), whereas no significant bands were detectable for other tissues. The multiple bands may correspond to glycosylated forms because bovine subunit a is known to acquire N-linked oligosaccharides (23). The a4 isoform was detected mainly in the cortical membrane fraction of kidney (Fig. 2D). Other isoforms and subunits including subunit A that have a catalytic site were distributed equally in cortical and medullar fractions ( Fig. 2D and data not shown).
Immunoprecipitation of V-ATPase with the a4 Isoform-Membranes were obtained from kidney cortex and treated with octylglucoside. The soluble fraction was incubated with anti-bodies against the a4 isoform, and the immunoprecipitate was subjected to polyacrylamide gel electrophoresis in the presence of SDS. As shown by immunoblotting (Fig. 3), the precipitate contained subunit A of membrane extrinsic V 1 sector and the c subunit of V o , suggesting that the a4 isoform is a component of kidney cortical V-ATPase.
Localization of the a4 Isoform in Intercalated Cells of the Cortical Collecting Ducts-Immunohistochemical analysis was carried out to identify the renal cells expressing the a4 isoform. No significant signal was observed in glomeruli or proximal and distal convoluted tubules (Fig. 4A). The a4 isoform was strongly expressed in the cortical collecting ducts (Fig. 4A) and found specifically on the apical and basolateral surfaces of certain epithelial cells (Fig. 4B, arrowheads). The signal diffused through the cytoplasm was often observed in the cortical collecting ducts (Fig. 4B, arrow). Two types of epithelial cells (principal and intercalated cells) are present in cortical collecting ducts (24). Immunoelectron microscopic analysis revealed that the a4 isoform was concentrated on the apical (Fig. 5A) and basolateral (Fig. 5B) surfaces of intercalated cells detected as mitochondria-rich epithelial cells (24), whereas no signal was found in principal cells (Fig. 5C).
In the cortical collecting ducts, two kinds (␣ and ␤) of intercalated cells are responsible for proton and bicarbonate secretion, respectively (8). The ␣ and ␤ cells have V-ATPase localized on their apical and basolateral plasma membranes, respectively (25). Peanut lectin agglutinin (PLA) is associated with the apical membranes of rabbit ␤ cells but not with those of ␣ cells (26). Although it is not a specific marker in rodent, double immunostaining indicated that all intercalated cells expressing the a4 isoform at the basolateral surface were apical PLApositive (Fig. 4C). Electron micrographs (Fig. 5) showed that the cells expressing the a4 isoforms in the basolateral surface exhibit extensive invagination in the basolateral membranes and a few microvilli in the apical surface (Fig. 5B). This morphology is reported to be characteristic of rat ␤-intercalated cells (24,27). Considering both the immunofluorescence and immunoelectron microscopy evidence, we suggest that the mouse ␤ cells have basolateral a4 isoforms.
The ␣ cells have Cl Ϫ /HCO 3 Ϫ exchanger (AE1) on their basolateral membranes (28 -30). All intercalated cells having the FIG. 2. Kidney-specific expression of the mouse a4 isoform. A, genomic Southern blot analysis of the a4 isoform. Mouse genomic DNA (10 g) was digested with various endonucleases and electrophoresed on an agarose gel. After blotting, the filter was hybridized with radioactive probe. B, Northern blot analysis of the a4 isoform. Poly(A) ϩ RNAs (2 g) of various adult tissues and whole embryos were hybridized with radioactive probe. The blot was also hybridized with a control probe of ␤-actin. Arrowhead indicates the position of the transcripts. C, kidney-specific presence of the a4 isoform. Total proteins (20 g) from various mouse tissues and yeast cells expressing a4 (a4 (in yeast)) were separated by the gel electrophoresis in the presence of SDS and then incubated with antibodies against a4. The arrowhead indicates the position of the a4 subunit that was not glycosylated. D, presence of the a4 isoform in the renal cortex and medulla. Total proteins (30 g) of whole kidney, medulla, or cortex were separated by gel electrophoresis, and then incubated with antibodies against the a3, a4, or A subunit. FIG. 3. Association of the a4 isoform with other subunits of V-ATPase. Octylglucoside-solubilized fraction (100 g of protein) was incubated for 1 h with purified antibodies (1 g) against the a4 isoform (␣-a4) or control IgG (IgG) in the buffer used for solubilization of V-ATPase (IP). Immunoprecipitates were subjected to 8% (for detecting A subunit) or 12% (for c subunit) polyacrylamide gel electrophoresis and blotted onto nitrocellulose. They were incubated with antibodies against A (A) or c (c) subunit for immunodetection. Positions for isoforms are indicated by arrowheads. As a control, the kidney membrane fraction was applied to the gel (membranes). Other procedures were described previously (11).
apical a4 isoform were basolateral AE1-positive, indicating that ␣ cells have apical a4 (Fig. 4D). These results indicate that the V-ATPase with the a4 isoform is localized specifically on the apical and basolateral surfaces of ␣and ␤-intercalated cells, respectively.

DISCUSSION
The subunit a is forming the intrinsic membrane V o sector of V-ATPase. Three a isoforms (a1, a2, and a3) were found previously in chicken, mouse, cow, and human (7,(13)(14)(15)(16)(17)(18). We have identified a fourth subunit a isoform (a4) of mouse V-ATPase in this study. Transcripts of the a4 isoform were found in 15-and 17-dpc embryos, but not in 7-and 11-dpc embryos. The differentiation of mouse kidney proceeds after 14-dpc when primitive glomeruli are observed. At the same time, the number of collecting tubules also increases in association with the development of glomeruli (31). Considered with the finding that the a4 isoform was expressed in collecting ducts, these observations suggest that a4 gene expression is closely related to the differentiation of renal collecting tubules.
Renal intercalated cells in cortical collecting ducts have been classified into at least two cell types (␣ and ␤) (8,24,32). For acid/base homeostasis, ␣and ␤-intercalated cells are thought to be involved in proton and bicarbonate secretion, respectively (8,32). The ␣ cells express V-ATPase on their apical membranes and AE1 on their basolateral membranes (8, 28 -30), suggesting that the V-ATPase with the a4 isoform is required for apical proton secretion in ␣ cells.
On the other hand, ␤ cells are thought to be mirror images of ␣-intercalated cells, with apical Cl Ϫ /HCO 3 Ϫ exchanger and basolateral V-ATPase (8, 32). As described above, a4 was clearly detectable on the ␤ cell basolateral membrane. A new Cl Ϫ / HCO 3 Ϫ exchanger (AE4) was recently been found in the apical surface of ␤ cells (33), suggesting that a combination of AE4 and V-ATPase with a4 is important for bicarbonate secretion from apical membranes of ␤ cells.
Smith et al. (34) have reported recently that the mutations in the human ATP6N1B gene cause recessive distal renal tubular acidosis. The ATP6N1B gene product is highly expressed in kidney and exhibits 85.8% identity with the mouse a4 isoform, suggesting that the ATP6N1B gene codes for a human counterpart of the mouse a4 isoform (Fig. 1). The ATP6N1B product has been suggested to be present on the apical surface of FIG. 4. Presence of the a4 isoform in intercalated cells of cortical collecting ducts. A, presence of the a4 isoform in collecting ducts. A section of kidney cortex was stained with antibodies against the a4 isoform: CCD, cortical collecting duct; GL, glomeruli; PCT, proximal convoluted tubules; or DCT, distalconvoluted tubules. Scale, 50 m. B, presence of the a4 isoform in the apical and basolateral surface of epithelial cells of collecting ducts. The boxed area of A was enlarged. The a4 isoform was detected on apical (black arrowhead) and basolateral (red arrowhead) surface or diffusely in cytoplasm (arrow). Scale, 20 m. C, presence of the a4 isoform in the basolateral surface of ␤ cells. The ␤ cell apical membrane stained with fluorescent peanut lectin had basolateral a4 isoform (␤, arrows). Scale, 10 m. D, presence of the a4 isoform in the apical surface of the ␣ cell. Two serial sections were stained with antibodies, the apical a4 isoform (a4) and basolateral AE1 being detected. Scale, 20 m. intercalated cells in cortical collecting ducts. In contrast, the mouse a4 isoform was clearly detectable not only on the apical surface of ␣-intercalated cells but also on the basolateral surface of ␤ cells, implying that human a4 may be also present on the basolateral surface of ␤ cells.
Two isoforms (B1 and B2) of subunit B have been identified in the mammalian V-ATPases V 1 sector (35,36). The B1 isoform is expressed specifically in kidney (35)(36)(37), whereas B2 is ubiquitously observed. Mutations in the human B1 gene also cause distal renal tubular acidosis (37), similar to that in the a4 gene (34). These findings suggest that the B1 and a4 form of V-ATPase are essential for the regulation of renal acid/base homeostasis.