A Proton Pump ATPase with Testis-specific E 1-Subunit Isoform Required for Acrosome Acidification*

The vacuolar-type H (cid:1) -ATPases (V-ATPases) are a fam-ily of multimeric proton pumps involved in a wide variety of physiological processes. We have identified two novel mouse genes, Atp6e1 and Atp6e2 , encoding testis-specific ( E 1) and ubiquitous ( E 2) V-ATPase subunit E isoforms, respectively. The E 1 transcript appears about 3 weeks after birth, corresponding to the start of meio-sis, and is expressed specifically in round spermatids in seminiferous tubules. Immunohistochemistry with iso-form-specific antibodies revealed that the V-ATPase with E 1 and a 2 isoforms is located specifically in developing acrosomes of spermatids and acrosomes in mature sperm. In contrast, the E 2 isoform was expressed in all tissues examined and present in the perinuclear com-partments of spermatocytes. The E 1 isoform exhibits 70% identity with the E 2, and both isoforms functionally complemented a null mutation of the yeast counterpart VMA4 , indicating that they are bona fide V-ATPase subunits. The chimeric enzymes showed slightly lower K m ATP than yeast V-ATPase. Consistent with the temper- ature-sensitive growth of (cid:2)

Vacuolar H ϩ -ATPase (V-ATPase), 1 an ATP-dependent proton pump, is one of the ubiquitous eukaryotic enzymes. It is present in endomembrane organelles such as vacuoles, lysosomes, endosomes, Golgi apparatus, chromaffin granules, and coated vesicles (1)(2)(3) and is also found in the plasma membranes of specialized cells including osteoclasts and renal epithelial cells (4,5). V-ATPase is required for diverse cellular processes including receptor-mediated endocytosis, renal acid-ification, bone resorption, neurotransmitter accumulation, and activation of acid hydrolases (2).
V-ATPase (V 1 V 0 ) has a similar structure and mechanism to F-type ATPase (ATP synthase, F 1 F 0 ), and their ATP-dependent conformational changes are transmitted between the peripheral complex (V 1 or F 1 ) and the proton pore (V 0 or F 0 ) through a number of subunits forming a stalk (1)(2)(3). Deletion of mammalian V o subunit c, encoded by a single gene (6,7), has been shown to cause an embryonic lethal phenotype (8,9). One of the most important questions is how a ubiquitous V-ATPase can function in a wide variety of physiological processes.
The diverse functions of V-ATPase may be established by utilizing a specific subunit isoform(s), its basic functional structure being maintained. Multiple subunit isoforms have been found for the largest subunit, a, of the V 0 sector in nematode (10), chicken (11), mouse (12)(13)(14)(15), and man (16). a4 is specifically expressed in renal intercalated cells (14,17), and its mutations cause renal acidosis (17), whereas a defect of a3, a component of the osteoclast plasma membrane enzyme (12), results in osteopetrosis (18). The only reported isoforms of the mammalian V 1 sector is the B subunit. B1 is specifically expressed in kidney and cochlea (19), and B2 is expressed ubiquitously.
In this study, we have isolated a mouse gene, Atp6e1, coding for a novel isoform (named E1) of the V-ATPase E subunit. E1 exhibits high similarity to ubiquitously expressed E2 (Atp6e2) and is expressed exclusively in the germ cells of testis. We found that a V-ATPase with E1 and a2 isoforms is located specifically in the acrosomal membrane of sperm. This unique V-ATPase is required for an acidic intra-acrosomal pH of 5.3 ϩ 0.1 (20), which is necessary for processing protease zymogen essential for fertilization.

EXPERIMENTAL PROCEDURES
Isolation of E1 and E2 cDNAs-poly (A) ϩ RNAs from mouse testis were prepared using TRIZOL ® (Invitrogen) and Oligotex-dT30 (Takara) by standard methods. The cDNA encoding for E1 was amplified by reverse transcription-PCR (High Fidelity RNA PCR kit, Takara) with primers 5Ј-cctgcctctgcccccgtttctgagagcagccatg-3Ј (GH93, forward) and 5Ј-gaggaagatgacaccccacccccaccccag-3Ј (GH94, reverse), designed from available mouse expressed sequence tags (EST). The resulting PCR product was cloned into pBluescript, and then the sequences of two independent clones were verified by ABI Prism® BigDye Terminator cycle sequencing (Applied Biosystems). The full-length cDNA was isolated by means of rapid amplification of cDNA ends. The cDNA encoding the E2 isoform was obtained from mouse EST clone accession number BF385623 and then sequenced.
Northern Blot Analysis, Southern Analysis, and in Situ Hybridization-Aliquots of poly(A) ϩ RNA (2 g) from testis, ovary, or uterus were subjected to gel electrophoresis and then transferred to a nylon membrane (Hybond N; Amersham Biosciences) (21). The mouse multiple tissue blots were purchased from CLONTECH. The probes corresponding to the ϩ491 to ϩ957-bp region of E1 and the ϩ716 to ϩ1148-bp region of E2 were used for hybridization. Southern analysis was performed as previously described (14), with probes corresponding to the ϩ491 to ϩ813-bp region of E1 and the ϩ716 to ϩ1148 bp region of E2 cDNAs.
Digoxigenin (DIG)-11-UTP-labeled single-stranded RNAs were prepared with the DIG RNA-labeling mixture and the corresponding T3 or T7 RNA polymerase (Roche Molecular Biochemicals). The 466-and 432-bp fragments from the 3Ј-untranslated regions of E1 and E2, respectively, were cloned into pBluescript (Stratagene) and used for preparing probes. In situ hybridization was performed on 8-m cryo-sections from the testes of ICR male mice (8ϳ10 weeks old) (22). Alkaline phosphatase was detected after incubation for 2-24 h at 37°C, which was terminated with Tris-EDTA. The slides were rinsed in water, counterstained with methyl green and then mounted with Crystal Mount (Biomeda).
Measurement of ATPase Activity-ATPase activity was measured using a coupled spectrophotometric assay as described previously (27) with several modification. To determine the K m ATP and V max for V-ATPase complexes containing the yeast Vma4p or mouse E1 and E2, ATPase activity was measured over a range of ATP concentrations from 0.05-1.5 mM. Vacuolar membrane vesicles isolated from the vma4⌬ strain expressing Vma4p, E1, or E2 (10 g protein) were incubated in ATPase assay buffer (50 mM NaCl, 30 mM KCl, 20 mM HEPES-NaOH, pH 7.0, 0.2 mM EGTA, 10% glycerol, 1 mM MgCl 2 , 4 mM MgSO 4 , 1.5 mM phosphoenolpyruvate, 0.35 mM NADH, 20 units/ml pyruvate kinase, and 10 units/ml lactate dehydrogenase) with 0.1% dimethyl sulfoxide or 1 M concanamycin A at 25°C for 10 min. The assay was started by addition of ATP, and the absorbance at 340 nm was measured continuously using a Hitachi UV-visible recording spectrophotometer (UV-FIG. 1. Alignment of the amino acid sequences of subunit E isoforms and Southern analysis of Atp6e1 and Atp6e2 genes. a, the amino acid sequences of the mouse E1 and E2 isoforms and yeast Vma4p are aligned. The numbers represent the positions of amino acid residues. Identical residues are indicated by shadowing, and the regions used for antibody preparation are underlined. Asp-145 in Vma4p is indicated (asterisk). b, genomic Southern blot analysis of the Atp6e1 and Atp6e2 genes. Mouse genomic DNA (10 g) was digested with various endonucleases. After electrophoresis and blotting, the filter was hybridized with radioactive probe of E1 or E2. kb, kilobases. 2500 PC). V-ATPase was defined as the fraction of the ATPase activity inhibited by 1 M concanamycin A. ATP-dependent proton transport activity was measured by quenching of acridine orange fluorescence using a Hitachi F-3000 fluorescence spectrophotometer (28).
Preparation of Membrane Fractions and Western Blotting-Tissues were dissected out from two ICR male mice (10 weeks old) at 4°C, suspended in 5 mM MOPS buffer, pH 7.0, containing 0.25 M sucrose, 0.1 mM MgSO 4 , 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and complete protease inhibitor mixture (Roche Molecular Biochemicals), and then homogenized with a Wheaton homogenizer. The lysates, obtained on centrifugation at 1,000 ϫ g for 5 min, were centrifuged at 8,000 ϫ g for 10 min. The supernatants were centrifuged at 100,000 ϫ g for 60 min, and the pellets suspended in the above buffer were used as membrane fractions. Yeast cells (YPH499) transformed with either pKT-N-myc-E1 or pKT-N-myc-E2 were lysed as previously described (30), the total lysates being used as positive controls.
Solubilization and Immunoprecipitation of Testis V-ATPase-The immunoprecipitation was performed as previously described (14). Briefly, the membrane fraction obtained from testes was incubated at 4°C for 2 h in solubilization buffer (50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 2% n-octyl-␤-glucopyranoside and 200 mM NaCl) and then centrifuged at 100,000 ϫ g for 30 min. The supernatant (solubilized V-ATPase) was incubated with 1 g of anti-a2 or preimmune IgG at 4°C overnight and then with protein A-Sepharose beads for 2 h at room temperature. After washing the beads six times with the above buffer, the immunoprecipitates were boiled in the sample buffer (31) and subjected to Western blotting.
Immunohistochemistry, Fluorescence Microscopy, and LysoTracker Staining-Adult male ICR mice were used as the source for the isolation of testes, spermatogenic cells, and sperm (32,33). Immunohistochemistry of testis cryo-sections (6 m) was performed as previously described (14). Sperms or spermatogenic cells were attached to a poly-L-lysine-coated slide glass in 1 ml of Krebs-Ringer bicarbonate buffer (32) and then fixed for 1 h in 4% paraformaldehyde in 0.1 M phosphatebuffered saline, pH 7.2 (PBS). The cells were rinsed three times in PBS and then permeabilized for 30 min at room temperature in 0.2% Triton X-100 in PBS. After blocking for 1 h with 10% goat serum and 0.2% gelatin in PBS, the cells were incubated with primary antibodies at 4°C overnight. After three rinses with PBS containing 0.02% Tween 20 (PBST), the samples were incubated with fluorescein isothiocyanate (FITC)-conjugated secondary antibodies for 40 -60 min at room temperature. The samples were washed three times with PBST and then stained with 10 ng/ml 4,6-diamidino-2-phenylindole (Molecular Probes) for 1 min. The slides were washed and mounted in a drop of Vectashield mounting medium (Vector Labs) and then examined under a Zeiss Axioplan2 epifluorescence microscope.
For double immunostaining, the fixed sperm were first labeled with anti-a2 antibodies as described above. After three rinses with PBST, the samples were incubated with Cy3-conjugated secondary antibodies FIG. 2. Complementation of the yeast ⌬vma4 mutation by mouse E1 and E2. a, a yeast mutant (⌬vma4) harboring the vector, E1 or E2 expression plasmid, or the VMA4 gene on a single copy plasmid (VMA4) was streaked onto a yeast extract/peptone/dextrose plate buffered to pH 5.0 or 7.5 and then incubated for 4 days at 30°C before photographing. b, the same amount of ⌬vma4 cells (5 ϫ 10 4 ) harboring pKT10-E1 or pKT10-E2 was removed from the culture and serially diluted 5-fold 7 times (right to left for each dilution), and then 4 l of the diluted mixture was spotted on a yeast extract/peptone/dextrose plate, pH 7.5, and incubated at 30°C or 37°C for 3 days. c, electrochemical H ϩ gradient (⌬pH) formation in yeast vacuolar vesicles was assayed with fluorescence dye acridine orange (28). Vesicles (10 g of protein) were incubated in 1 ml of 50 mM NaCl, 30 mM KCl, 20 mM HEPES-NaOH, pH 7.0, 0.2 mM EGTA, 10% glycerol, 1 mM MgCl 2 , 4 mM MgSO 4 , and 1 M acridine orange with 0.1% dimethyl sulfoxide or 1 M concanamycin A at 25 or 37°C for 20 min and assayed for ⌬pH formation at the corresponding temperature. ATP (0.5 mM sodium-salt, open arrowheads) and 5 mM NH 4 Cl (closed arrowheads) were added at the times indicated. a ATPase activities were measured on isolated vacuolar membranes containing Vma4p, E1, or E2 (10 g of membrane protein) over a range of ATP concentrations from 50 M to 1.5 mM as described under "Experimental Procedures." Activities were measured in the absence and presence of 1 M concanamycin A, and the results shown represent the concanamycin A-sensitive fraction of the activity. K m and V max were calculated from Lineweaver-Burk plots of reciprocal of initial ATP concentration versus reciprocal of initial rate of ATPase activity (mol/min/mg). b ATP-dependent proton transport activities were estimated from the initial rate of ATP-dependent fluorescence (⌬F) quenching in the presence of 0.5 mM ATP using the fluorescence dye acridine orange in the absence or presence of 1 M concanamycin A. The transport assay buffer was the same as that used for measurement of ATPase.
c The coupling ratio was calculated from the initial rate of fluorescence quenching (⌬F) and the ATPase activity in the presence of 0.5 mM ATP.
(Jackson Immuno Research) for 40 -60 min at room temperature, extensively washed with PBST, and blocked again with 10% goat serum and 0.2% gelatin in PBS for 1 h. The samples were then incubated for 2 h with FITC-labeled anti-E1 antibodies (FluoReporter® FITC protein labeling kit, Molecular Probes). The double-labeled sperm were then washed, stained with 4,6-diamidino-2-phenylindole, and mounted for microscopy as described above.
For LysoTracker staining, sperm were isolated as above and then incubated with 75 nM LysoTracker Red DND-99 (Molecular Probes) in buffer for 30 min at 33°C under 5% CO 2 . After washing three times with Krebs-Ringer bicarbonate buffer, the samples were observed by fluorescence microscopy.
Electron Microscopy-The pre-embedded silver enhancement immunogold method was used as described previously (14). The seminiferous tubules were fixed in 4% paraformaldehyde in PBS overnight at 4°C. Cryo-sections (6 m) were reacted with primary antibodies overnight followed by incubation with colloid gold (1.4-nm diameter)-conjugated secondary antibodies. The gold labeling was intensified using a silver enhancement kit (Nano Probes). Mouse multiple tissue and embryo stage Northern blots (CLONTECH) (2 g of poly (A) ϩ RNA/lane) were hybridized with the 32 P-labeled E1 or E2 probes (a). Two g of poly(A) ϩ RNA/lane from ovary, uterus, and testis (b) or testes of mice of different ages (weeks (w); c) was subjected to electrophoresis, blotted onto a membrane, and then hybridized with a radioactive probe. The sizes of the E1 (1.2 kilobases (kb)) and E2 transcripts (1.4 kilobases) correspond to the sizes of the isolated cDNAs (arrows). a ATPase assays were carried out on isolated vacuolar membranes at the indicated temperature after incubation at the same temperature for 20 min. The V-ATPase activities represent the concanamycin A-sensitive portions at the 0.5 mM ATP.

Identification and Isolation of a Novel
(E subunit) amino acid sequence revealed the presence of related mouse EST clones in GenBank TM data bases and the RIKEN full-length cDNA project data base (34). In particular, the mouse EST clones fall into two categories that are homologous to Vma4p, indicating the presence of two genes. Because one class of the EST was detected solely in testis, we have isolated the full-length cDNA from the adult mouse testis and named the encoded protein the E1 isoform. E2 cDNA was obtained by sequencing EST clone accession number BF385623. We designated the genes coding for E1 and E2 as Atp6e1 and Atp6e2, respectively. According to the FANTOM data base of RIKEN, we found that the Atp6e1 cDNA has been mapped to chromosome 17 (ϳ47.15 centimorgans). This region is part of a segment syntenic with the location of human Atp6e1 on 2p16-p21. 2 The E1 and E2 isoforms comprise 226 amino acid residues, exhibit more than 70% identity to each other, and show ϳ33% identity to Vma4p (Fig. 1a). Conserved residue Asp-139 corresponds to Asp-145 in VMA4, in which mutations affect the stability of Vma4p (35) (Fig. 1a). Southern analysis confirmed that mouse E1 and E2 are encoded by two distinct genes (Fig. 1b).
Mouse E1 and E2 Are Able to Complement the Yeast ⌬vma4 Mutant-The open reading frame of E1 or E2 was placed under the control of the yeast TDH3 constitutive promoter, and the resulting plasmids were introduced into cells lacking endogenous Vma4p. Expression of mouse E1 or E2 complemented the ⌬vma4 mutation (Fig. 2a). This result indicates that mouse E1 and E2 are functional V-ATPase subunits. Interestingly, we found that yeast cells expressing E1 showed a temperaturesensitive growth phenotype, whereas E2 functioned at a high temperature similar to authentic yeast VMA4 (Fig. 2b).
To compare the kinetic properties of V-ATPase complexes containing Vma4p, E1, or E2, vacuolar membranes were isolated from ⌬vma4 cells expressing the E1, E2, or Vma4p. The K m ATP was ϳ310 M for the complex containing Vma4p and ϳ150 M for the complex containing E1 or E2 (Table I). No significant kinetic difference between the E1-and E2-containing enzymes was observed. Because the yeast cells expressing E1 or E2 showed different temperature sensitivity (Fig. 2b), we examined the ATPase activities of the complex containing E1 or E2 at 25°C and 37°C. The ATPase activities of both complexes increased approximately 2-fold when the assay temperature was shifted from 25 to 37°C (Table II), indicating that their catalytic activities are not temperature-sensitive.
We then measured the ATP-dependent proton transport activities using quenching assay of a fluorescence dye acridine orange. The ATP-dependent proton transport activity at 25°C in vacuolar membranes isolated from E1-or E2-expressing cells was almost the same as that observed for Vma4p (Fig. 2c). The ratio of proton transport to ATP hydrolysis for either E1-or E2-containing complex was 65% that of Vma4p-containing enzyme at (Table I). However, we found that the E1-containing complex was completely inactive in proton transport after incubation at 37°C for 20 min, whereas the E2-or Vma4pcontaining complex still retained significant proton transport activity (Fig. 2c). These results suggest that the V-ATPasecontaining E1 isoform exhibited temperature-sensitive uncoupling of proton transport and ATP hydrolysis.
Expression of E1 mRNA Is Developmentally Regulated during Spermatogenesis-As shown in Fig. 3, the 1.4-kilobase E2 transcript was detected from all tissues examined, whereas on the contrary, the E1 transcript (1.2 kilobases) was only ex- Membrane fractions from murine brain, kidney, spleen, liver, and testis together with ones from yeast expressing E1myc and E2-myc fusion proteins were subjected to gel electrophoresis in the presence of sodium dodecyl sulfate. Immunoblotting using the purified polyclonal rabbit IgG against each isoform is shown. The arrows indicate the positions of E1 and E2 estimated from their cDNA sequences. pressed in testis. E2 was expressed from early embryonic stages, whereas E1 expression was only detectable in adult testis (Fig. 3a). In contrast to the expression in testis, E1 was not expressed in female reproductive organs (Fig. 3b). Because the testis undergoes postnatal development during which progressively more mature germ cells appear, we examined the E1 expression during development. E1 mRNA was not observed in the testis up to 2 weeks after birth but became detectable at 3 weeks, gradually increased, and reached a plateau level at 8 weeks (Fig. 3c).
We then investigated whether or not E1 expression is specific to restricted germ cells and regulated during spermatogenesis. Frozen sections of 8 -10-week-old mouse testes were hybridized with E1 antisense probes (Fig. 4). E1 mRNA was expressed in the middle zone between the center and the basal lamina of seminiferous tubules (Fig. 4, c and e), where late spermatocytes and early round spermatids are abundant. E2 mRNA was detected at a high level in spermatogonia, the mitotic dividing cells, and a low level in spermatocytes (Fig. 4f). E1 mRNA was not detectable in Sertoli cells, spermatogonia, or spermatozoa. These results clearly indicate that the expression of E1 and E2 is regulated differently in testis.
E1 Is Specifically Expressed in the Acrosomal Membrane-The expression pattern of E1 suggests its involvement in germ cell-specific function(s). We then examined the localization of the isoforms using antibodies raised against specific peptides of E1 and E2 (see Fig. 1 for sequences). Immunoblotting with the purified antibody against E1 revealed a 33,000-M r protein specifically in testis, whereas the E2 antibody recognized a band (32,000 M r ) for all tissues examined (Fig. 5). The molecular weights of both bands correspond to those predicted from the respective cDNAs.
For testicular spermatogenic cells, three typical staining patterns were observed for E1 (Fig. 6, a-c). The E1 isoform was located specifically in the developing acrosomes of elongated spermatids (Fig. 6a, arrows) and proacrosomal granules of round spermatids (Fig. 6b, arrows). Weak staining of E1 was observed in ϳ 20% of the seminiferous tubules (Fig. 6c). This difference is possibly due to the different maturation stages of the seminiferous tubules. The E2 isoform was located in the perinuclear region of all cells except spermatozoa (Fig. 6d).
We observed the localization of three a subunit isoforms (a1, a2, and a3) that are expressed in testis, as we have shown previously (12). Immunostaining with the specific antibodies revealed that these isoforms exhibited different distributions in seminiferous tubules. The a1 isoform was expressed in almost all cells except the spermatozoa (Fig. 6e). Interestingly, the isoform localized in the acrosomes of elongated spermatids was a2 (Fig. 6f, arrows), and a3 was expressed specifically in Sertoli cells (Fig. 6g, arrows), which produce the nutrients required for the completion of spermatogenesis (36).
In epididymal sperm, the E1 isoform was located specifically in the acrosomes, which can be visualized by LysoTracker staining because of their luminal acidic pH (Fig. 7). Lyso-Tracker staining was completely abolished when sperm were treated with 10 nM bafilomycin A 1 , a specific inhibitor of V-ATPase (Fig. 7c), indicating that acrosome acidification is carried out solely by V-ATPase. The V-ATPase A subunit and the a2 isoform were observed in acrosome (Fig. 7, e and f). Other a subunit isoforms (a1 and a3) were not detectable in mature acrosomes (Fig. 6, e and g).
The localization of the E1 in acrosomal membrane was confirmed by immunoelectron microscopy. The testes were fixed and processed for immunogold labeling with the anti-E1 antibodies. The gold particles (arrows) were observed on the outer and inner acrosomal membranes (Fig. 7n).
The co-localization of E1 and a2 in acrosomal membrane was further examined by double-staining experiments using anti-a2 and FITC-labeled anti-E1 antibodies. The staining signals for E1 and a2 in the acrosome were completely merged (Fig. 7,  j-m). The result indicated that the V-ATPase with isoforms E1 and a2 is required for acrosome acidification.
In addition, immunoprecipitation experiment was carried out to confirm the presence of both E1 and a2 isoforms in the same V-ATPase complex. The octylglucoside-solubilized V-ATPase fraction from testis membranes was incubated with anti-a2 IgG, and the immunoprecipitate was subjected to polyacrylamide gel electrophoresis. The precipitate contained the E1 isoform and subunit c of the V 0 sectors, indicating that both E1 and a2 are present in the same enzyme (Fig. 8). 6. Localization of E and a isoforms in testis seminiferous tubules. Six-m cryo-sections were immunostained with antibodies against E1 (a-c), E2 (d), a1 (e), a2 (f), and a3 (g). Three typical staining patterns were observed for E1 (a-c). The E1 isoform localized in the acrosome regions of spermatocytes, spermatids, and sperm is indicated (arrows). The acrosomal localization of a2 and the Sertoli cell-specific localization (arrows) of a3 are also shown (f and g, respectively). Bar, 25 m.

DISCUSSION
We have identified a novel isoforms (E1and E2) of the mouse V-ATPase E subunit, and the corresponding genes, Atp6e1 and Atp6e2, respectively. In contrast to the ubiquitously expressed E2, E1 was expressed specifically in testis and localized in proacrosomal granules in spermatocytes and acrosomes in de-veloping spermatids and mature sperm. We also found that the A subunit of V 1 and a2 of V 0 were localized in the same membranes (Figs. 6 and 7), indicating that a unique V-ATPase is present in acrosomes. The E1 as well as the E2 isoform functionally complemented the yeast counterpart vma4 null mutation, indicating that mouse isoforms can assemble with other V-ATPase subunits to form a functional enzyme complex.
The acrosome is known to be an acidic secretory vesicle containing hydrolytic enzymes that are involved in the passage of the sperm across the zona pellucida. Thus, the assembly of a proton pump is an essential step for the biogenesis of this unique organelle. The presence of V-ATPase in acrosomes was predicted in a recent review (1) and suggested by a preliminary experiment (37). The present results clearly indicate that a V-ATPase with specific isoforms E1 and a2 is solely responsible for acrosome acidification. Our results show that E1 is induced concomitantly with the onset of acrosome biogenesis, whereas the ubiquitous E2 isoform is present in spermatocytes but completely absent in late spermatids and mature sperm.
Interestingly, the a subunit that is localized on the acrosomal membrane is a2, an isoform localized specifically to Golgi apparatus, not other a isoforms. 3 This is consistent with the previous morphological observation that the acrosome is formed through the fusion of vesicles derived from the Golgi apparatus (38,39). Thus, the assembly of V-ATPase with E1 3 T. Toyomura and M. Futai, manuscript in preparation.

FIG. 7. Localization of E1 isoform in sperm.
A phase-contrast image (a) is shown together with a living epididymal sperm stained with LysoTracker Red (b). A sperm was stained with LysoTracker after incubation for 30 min in Krebs-Ringer bicarbonate buffer containing 10 nM bafilomycin A 1 (c). For immunochemical identification of the E1, A, and a2 subunits, sperm were fixed and labeled with anti-E1 (d), anti-A (e), and anti-a2 (f) antibodies. Merged images with DNA visualized using the fluorescent dye 4,6-diamidino-2-phenylindole are also shown (gi). The slight differences observed in the acrosome staining were due to the orientation of individual sperm attached to the slide glasses. The double immunolabeling is also shown with FITC-conjugated anti-E1 (j), anti-a2 (k), merged image of j and k with 4,6-diamidino-2-phenylindole staining (l), and the corresponding phasecontrast image (m). Electron microscopic localization of E1 isoform in acrosome membranes of testis sperm (n) is shown by electron dense silver-enhanced immunogold particles (see arrows, for examples). N, nucleus; IM, inner membrane; OM, outer membrane of acrosome. Bar, 500 nm.
FIG. 8. Co-immunoprecipitation of E1 and a2 from solubilized testis membrane. The octylglucoside-solubilized membrane fraction was incubated with anti-a2 (␣-a2) or control IgG (IgG) in the buffer used for the solubilization of V-ATPase (IP). The immunoprecipitates were subjected to 12 and 15% polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate, blotted onto nitrocellulose sheets, and incubated with anti-E1 and anti-c antibodies, respectively. The positions of the E1 and c subunits are indicated by arrows. As a control, the brain membrane fraction was applied to the electrophoresis (membrane). and a2 isoforms is acrosome-specific and essential for its physiological function.
Previous studies indicated that the E subunit is essential for the assembly and activity of the yeast V-ATPase (40).
Tomashek et al. (41) demonstrated that the stability of the yeast E subunit protein depended on the presence of the G subunit (VMA10) and proposed that the two V 1 subunits may interact as a part of a "stalk" in the V-ATPase. It was reported recently that the E subunit interacts with aldolase and may be involved directly in the coupling of V 1 V 0 assembly and glucose metabolism (42). Therefore, E isoforms may have different roles in V-ATPase activity and assembly. The yeast ⌬vma4 cells expressing the E1 isoform were temperature-sensitive, whereas those with the E2 could grow at 37°C. Unexpectedly, the yeast V-ATPase containing the E1 isoform exhibited a temperature-sensitive defect in coupling between proton transport and ATP hydrolysis. The temperature sensitivity may be consistent with the lower optimal temperature (ϳ33°C) for the culture of spermatid cells (32,33). The present results also indicate that E subunit is essential for energy coupling, and V-ATPase became temperature-sensitive because the yeast Vma4p was replaced by mouse E1. This notion is also supported by the structure model of V-ATPase which located E subunit at the second stalk connecting catalytic V 1 and proton pathway V 0 sectors (43).
The presence of E isoforms (named E1 and E2) was observed recently in plants including pea, mung bean, and Arabidopsis thaliana (44). The two isoforms of had different K m and V max values. The plant E1 was hardly detectable in leaves and cotyledons, whereas E2 was expressed ubiquitously. Their expression is different, but neither of them shows highly restricted expression similar to that of mouse E1.
We have shown that a unique V-ATPase with E1 and a2 isoforms is involved in the acidification of the acrosomal lumen. Our results also indicate that apparently ubiquitous enzymes become acrosome-specific upon assembly with subunit isoforms E1 and a2. a3 and a4 may play similar roles in osteoclasts and renal intercalated cells (12,14,15). Studies on subunit isoforms are important for understanding the physiological functions of V-ATPase in various acidic compartments.