VMA12 is essential for assembly of the vacuolar H(+)-ATPase subunits onto the vacuolar membrane in Saccharomyces cerevisiae.

vma12 mutants of the yeast Saccharomyces cerevisiae, which were originally identified as calcium-sensitive (cls) mutants that were also respiratory deficient (Pet-), have a defect in vacuolar membrane H(+)-ATPase activity (Ohya, Y., Umemoto, N., Tanida, I., Ohta, A., Iida, H., and Anraku, Y. (1991) J. Biol. Chem. 266, 13971-13977). The VMA12 gene was cloned by complementation of the growth defects of vma12 mutants. The nucleotide sequence of the gene predicts a polypeptide of 215 amino acids (25.2 kDa) with two putative membrane-spanning domains. A null vma12 mutant, constructed by chromosomal deletion of the gene, is viable but has completely lost the vacuolar membrane H(+)-ATPase activity and exhibits the same growth defects as observed for the original vma12 mutants. Synthesis and targeting of the subunits of the H(+)-ATPase in the delta vma12 mutant cells were examined by Western blotting analyses of whole cell and vacuolar membrane protein extracts. None of the peripheral membrane subunits that we analyzed (the 69-, 60-, 42-, and 27-kDa subunits) was detected in the vacuolar membrane fractions, although the cellular levels of these polypeptides appeared to be normal. The 100- and 17-kDa integral membrane subunits of the enzyme were absent or present at a substantially reduced level in mutant vacuolar membrane fractions. Anti-Vma12p antibodies recognized a vacuolar protein with the expected molecular mass of 25 kDa. However, the Vma12 protein was not detected in the vacuolar membrane ATPase complex that had been solubilized with a zwitterionic detergent, ZW3-14, and purified by glycerol gradient centrifugation (Kane, P. M., Yamashiro, C. T., and Stevens, T. H. (1989) J. Biol. Chem. 264, 19236-19244). These results indicate that the VMA12 gene product is not a component of the active vacuolar ATPase complex and instead suggest that this protein is required during the process of assembly and/or targeting of the enzyme complex to the vacuolar membrane.

The nucleotide sequence(s) reported in thispaper has been submitted to the GenBankTM/EMBL Data Bank with accession numberfs) Dl 1472.
$ Present address: The Institute of Physical and Chemical Research (RIKEN), Hirosawa, Wako-shi, Saitama 351-01, Japan.  264,[19236][19237][19238][19239][19240][19241][19242][19243][19244]. These results indicate that the VMA12 gene product is not a component of the active vacuolar ATPase complex and instead suggest that this protein is required during the process of assembly andlor targeting of the enzyme complex to the vacuolar membrane.
The yeast vacuole functions as a primary storage compartment for various ions and solutes and regulates cytosolic ion and pH homeostasis (1,2). In addition, this acidic compartment contains a number of hydrolases and serves as a digestive organelle similar to the animal cell lysosome (3). The physiological importance of the vacuole, particularly under stressed conditions, has become clear from studies of yeast mutants that are defective in vacuolar functions and biogenesis (4)(5)(6)(7)(8)(9).
The lumen of the vacuole is acidified by an electrogenic H+-translocating ATPase (H*-ATPase) residing on the membrane of the organelle (10). The activity of the H+-ATPase appears to be indispensable for many vacuolar functions. The enzyme generates a proton motive force across the membrane that drives various secondary transport systems on the membrane (11)(12)(13) and serves to acidify the vacuolar lumen, which may be important for vacuolar hydrolase function (2, 3).
The yeast vacuolar membrane H+-ATPase is a well characterized member of the "V-type" ATPases. V-type ATPases have also been identified in various endocytic and exocytic membrane compartments of eucaryotic cells (14). The V-type ATPases characterized thus far are multisubunit complexes with similar subunit compositions (14, 15). The yeast enzyme has a functional molecular mass of more than 500 kDa (16) and consists of a t least eight polypeptides with apparent molecular masses of 100, 69, 60, 42, 36, 32, 27, and 17 kDa' (17)(18)(19)(20). Genes for all these subunits, except for the 32-kDa subunit, have been isolated and sequenced (18,19,(21)(22)(23)(24)(25)(26)(27). The predicted primary structures of the subunits are very similar to those determined for the subunits of higher eucaryotic cells, illustrating that the yeast enzyme serves as a good model for studies on the structure and function of the V-type ATPase.
Studies on deletion mutants of the genes for the ATPase subunits ( VMA genes) have revealed that the mutants exhibit a common, characteristic set of growth phenotypes. The mutant cells did not grow in YPD medium (1% yeast extract ' The names of the subunits of the enzyme were according to Kane e f al. (17). The 69-, 60-, and 17-kDa subunits are referred to as subunits a, b, and c, respectively, in the papers by Anraku and his colleagues (20,23,26,32,53).
In this paper, we report the isolation and characterization of a gene, VMAl2, that was originally identified as CLSlO (30,34). Nucleotide sequence predicts, and biochemical studies confirm, that the VMAl2 gene product (Vma12p)' is a polypeptide of 25 kDa. Deletion of the VMAl2 gene resulted in the loss of vacuolar membrane H'-ATPase activity. Further analysis revealed that the ATPase subunits failed to assemble onto the vacuolar membrane in the umal2 mutant cells. Finally, although Vmal2p is detectable in the vacuolar membrane fraction in wild-type cells, this polypeptide did not cosediment with the ATPase activity when the enzyme was isolated by glycerol gradient fractionation. These results suggest that Vmal2p is essential for the assembly of the H+-ATPase, although it is not itself a component of the purified vacuolar enzyme.

EXPERIMENTAL PROCEDURES
Materials-Enzymes for recombinant DNA methods were purchased from Takara Sbuzo (Kyoto). Modified T7"polymerase was from U. S. Biochemical Corp. [a-32P]dCTP (-110 TBq/mmol) were from ICN K&K Laboratories Inc. Other chemicals were as described by Uchida et al. (20). Bafilomycin A, was a generous gift from Dr. Karlheinz Altendorf (University of Osnabruck).
Plasmids and Recombinant DNA Methods-pBluescript KS' (Stratagene) was used for DNA manipulation in Escherichia coli. Yeast single copy plasmid pRS315 (LEU2) (35) was used for subcloning and complementation analysis of the VMAl2 gene. pRS305 (LEU21 (35) and pJJ281 (TRPl) (37) were used as sources of yeast selectable marker genes. Plasmid isolation, gel electrophoresis, ligation, restriction enzyme analysis, and E. coli transformation were done as described by Ausubel et al. (38). Yeast genomic DNA was isolated as described by Holm et al. (39). Yeast transformation was carried out by the lithium acetate method of Ito et al. (40) with a modification by Rodriguez and Tait (41).
spheroplast method according to Burgers and Percival (43). Leu+ transformants were selected on synthetic complete medium lacking leucine. We obtained only one Leu+ transformant by 10 batches of transformation (1 X lo9 cells). This transformant could grow in both the Ca2+ and the glycerol medium. pNUVA401, the plasmid contained in the transformant, was further analyzed. The reason for the unusually low transformation efficiency is unknown, but we suspect that the mutant cell is so sensitive to the zymolyase treatment that only the Vma+ transformant survived under the transformation conditions. For the following subcloning and complementation analysis, we used the lithium acetate method (see previous section), which was found to give higher transformation efficiency.
Nucleotide Sequence Analysis-Nucleotide sequence of the VMAl2 gene was determined for both strands by the dideoxy chain termination method (44) with the modification of Tabor and Richardson (45). Proteins with sequence similarities to Vmal2p were searched in NBRF (release 31) and SWISS protein data base (release 20) using the FASTP algorithm (46).
Disruption of the VMAl2 Gene-The 1.2-kb SpeI-XhoI fragment containing the VMA12 gene was cloned into pBluescript KS+. Deletion of the coding region of the gene was done using a polymerase chain reaction (PCR) as described by Imai et al. (47). Two synthetic oligonucleotide probes, AAGTGAAAAGACAGACTTG (bases 619-601, see Fig. 2) and CCGGAATTCTAAGCAACATTACACTG (bases 897-916 with a linker sequence (underlined) to introduce an EcoRI site at the end of the PCR product), were used to amplify a gapped plasmid with a deletion in the VMAl2 gene. The primers are designed in inverted tail-to-tail directions. The PCR product was digested with EcoRI and was self-ligated to yield pNUVA437, which contains the VMAZ2 gene with a deletion in the coding sequence (bases 488-496). Then the Nsp(7524)V fragment of the VMAl2 gene (bases 277-483) was deleted from pNUVA437 to yield pNUVA435. A 1-kb EcoRI fragment from pJJ281 containing the TRPl gene was inserted into the EcoRI site in the insert of pNUVA435. The disrupted allele of the umal2 gene was cut out from the vector by SpeI-XhoI digestion and used to transform a wild-type strain, YPH500, to yield RH202.
Southern Hybridization-Chromosomal DNA was digested with XhoI and SpeI, separated on agarose gels, and blotted onto nylon membrane filters in 0.4 M NaOH. Blots were hybridized with the 1.2kb XhoI-SpeI VMAl2 fragment labeled with horseradish peroxidase, and the bound DNA fragment was detected using a chemiluminescent substrate (Amersham Corp.). Hybridization was done according to the supplier's recommendation.
SDS-PAGE and Western Blotting-Protein extracts of whole cells and vacuolar membrane vesicles were prepared as described by Kane et al. (48). SDS-polyacrylamide gel electrophoresis was performed by the method of Laemmli (49). Immunoblots were prepared and probed as described (17). Blots were probed with primary antibodies diluted 1:1,000. Alkaline phosphatase-conjugated goat anti-mouse or antirabbit antibodies (Promega Biotec) were used as second antibodies, and bound alkaline phosphatase was visualized by the addition of 5bromo-4-chloro-3-indolyl phosphate and p-nitro blue tetrazolium (Sigma). Apparent molecular masses of proteins were determined relative to prestained molecular weight standards (Bethesda Research Laboratories).
Antibodies-The 877-base pair BstUI-SpeI fragment in the VMAl2 gene was cloned into the E. coli expression vector pEXPl (pMH2O) (50). The Vmal2p antigen (amino acids 21-215) produced from this plasmid was purified from E. coli and injected into New Zealand White rabbits as previously described (51). Rabbit anti-Vmal2p antibodies were affinity purified against antigen expressed from plasmid pMH2O as described (52). Monoclonal antibodies that recognize the 100-kDa (7B1), 69-kDa (llE6), 60-kDa (13Dll1, and 42-kDa (7A2) vacuolar ATPase subunits were prepared as described (17, washed membranes were suspended in the same buffer (1 mg of protein/ml), and 5 volumes of chloroform/methanol(2:1) were added to the suspension. The suspension was incubated on ice for 1 h with occasional vortexing. The suspension was then centrifuged, and the upper aqueous phase and interphase were removed. Proteins in the organic phase were recovered by evaporating the solvent and were subjected to SDS-PAGE.
Other Methods-Preparation of vacuolar membrane vesicles and purification of the vacuolar membrane H+-ATPase were done as described previously (20,53). Vacuolar ATPase, dipeptidyl aminopeptidase B (54), and a-mannosidase (53) activities were determined as previously described.

RESULTS
Isolation of the VMA12 Gene-The urnal2-1 mutant was originally isolated as a Ca2+-sensitive mutant (clsl0-1) that also exhibited a respiratory deficient phenotype (Pet-) (30). The VMAIZ gene was isolated by complementation of the growth defect of the mutant cell. A vmal2 mutant strain, NUY33 (vrnal2-1, leuZ), was transformed with a yeast genomic DNA library carried on the yeast multicopy vector YEpl3 (LEUZ). Leu+ transformants were selected on synthetic complete plates lacking leucine and tested for growth on YPG and YPD supplemented with 100 mM CaC12. A single clone, pNUVA401, restored the growth of NUY33 on both the glycerol and the Ca2+ plates in a plasmid-dependent manner. Fig. 1 shows the restriction map of pNUVA401 and various subclones. Subcloning and complementation analysis using a yeast single copy vector (pRS315) indicated that the complementing activity resides within the 1.2-kb XhoI-SpeI region of the insert (Fig. 1). Southern blotting analysis of the chromosomal DNA from wild-type cells indicated that VMA12 is a single copy gene in yeast. To confirm that the isolated clone contained the VMAl2 gene, we analyzed linkage between the cloned DNA and urnal2-1 by integration mapping. The 2.6kb XhoI-Sac1 fragment from pNUVA412 ( Fig. 1) was cloned into the vector pRS305, which contains a LEU2 marker. The

T P C K T A T K X ? G Z A " T T~T A W L H Y S E D F K K K L E F L K Y Q E Q E C l l G U T A " 3 S N T A P T L E Y Q S M V K R S K S V F S L Q E D D
G4PCTVYIXITIWIUIPTWATAWXXXTT"  Leu-), indicating that the cloned gene is tightly linked to the VMAIZ locus. We thus concluded that we cloned the authentic VMAl2 gene.

~A T I A~A A~W A T . W T A~M~T Z G T A I C 4 X -T G A P A -T
Nucleotide Sequence of the VMAlZ Gene-The nucleotide sequence of the 1.2-kb XhoI-SpeI fragment in pNUVA420 was determined for both strands. A single open reading frame of 215 codons, capable of encoding a protein of 25.2 kDa, was found in the region (Fig. 2). Because no other open reading frame with significant length was found in this region, we concluded that the open reading frame specifies the VMAl2 gene product (Vmal2p). Fig. 3 shows a hydropathy plot of the predicted Vmal2p sequence drawn by the method of Kyte and Doolittle (55). At the C-terminal region of the sequence, two stretches of hydrophobic residues are present, suggesting that Vmal2p is an integral membrane protein.
No significant sequence similarities have been found between the Vmal2p  (N(Z)) and the EcoRI sites was replaced with the TRPZ gene. Detail of the construct is described under "Experimental Procedures." B, Southern blotting analysis of the VMAZZ locus in wild-type and the null vmaZ2 mutant cells. Chromosomal DNA from YPH500 (wild) and RH202 (AvmaZ2:TRPZ) cells was digested by XhoI and SpeI, resolved in an agarose gel, blotted onto a nylon membrane, and probed with the 1.2kb XhoI-SpeI fragment of the VMAZP gene. and any proteins in NBRF (release 31) or SWISS (release 20) protein data base. The VMAl2 gene has proven to be identical with the VPH2 gene: vph mutants were isolated by labeling mutagenized cells with a pH-sensitive fluorescent dye, 6carboxyfluorescein, and screening for mutant cells that are defective in vacuolar acidification (33). It seems likely that uph mutants may include other uma genes.
Disruption of the VMAl2 Gene-The chromosomal locus of the gene was disrupted by the method of Rothstein (56). The 1.2-kb XhoI-SpeI fragment containing the VMAl2 gene was cloned into the vector pBluescript KS+. The disrupted allele of the gene was constructed on the plasmid by replacing the coding region of the gene with the TRPl gene fragment (Fig.  4A). Almost the entire coding region was deleted in the construct. The disruption allele (Aumal2:TRPl) was then liberated from the vector by XhoI-SpeI digestion and introduced into a wild-type diploid, YPH501 (trplltrpl). A Trp+ transformant was picked, and substitution of one copy of the chromosomal VMAI2 gene was confirmed by Southern blotting analysis (data not shown). The Aumal2:TRPl/VMAl2 diploid cells were sporulated, and tetrads were dissected. All the tetrads analyzed (17 sets) yielded four viable spores, indicating that disruption of the gene was not lethal (data not shown). For the following analysis, we used a haploid Aumal2:TRPl mutant, RH202 (Fig. 4), in combination with the wild-type parental strain YPH500. The Aumal2 mutant cells exhibited growth phenotypes identical to those of uma mutants defective for H+-ATPase subunit genes (VMAI, VMA2, VMA3, and VMA4).5 They did not grow in YPG medium, which contains glycerol as a sole carbon source, YPD medium supplemented with 100 mM CaC12, or YPD buffered at neutral pH (Fig. 5).Vacuolar membranes isolated from the Aumal2 mutant cells lacked bafilomycin Al-sensitive ATPase activity (Table I), indicating that Vmal2p is essential for expression of this enzyme's activity. The morphology of the vacuoles in the mutant cells appeared normal. However, the mutant cells failed to accumulate the fluorescent dye quinacrine in their vacuoles (data not shown). Since quinacrine is known to concentrate within acidic membrane compartments (57), this result indicates that the Aumal2 mutant is defective in vacuolar acidification.
Assembly of the Vacuolar Membrane H+-ATPase in the Avmal2 Mutant Cells-To investigate the role of Vmal2p in expression of the ATPase activity, we examined the synthesis A. Bachhawat and E. W. Jones, personal communication.
' VMAI is the same gene as TFPZ, VMAP equals VAT2, and VMA 1 2 equals TFP3 (18,19).  and localization of the subunits of the enzyme in the Auma12 mutant cells. We analyzed four of the peripheral membrane subunits (the 69-, 60-, 42-, and 27-kDa subunits) and one of the integral membrane subunits (the 100-kDa subunit). Assembly of another integral membrane subunit, the 17-kDa proteolipid subunit, onto the vacuolar membrane was also assessed by examining the proteins in chloroform/methanol extracts of vacuolar membrane vesicles. Fig. 6 shows the results of the Western blotting analysis. The steady-state levels of these subunits were not affected by the Avmald mutation except that the level of the 100-kDa subunit was 5-10-fold reduced in the mutant cells (Fig. 6A). However, we could not detect any of these ATPase subunits in vacuolar membranes from Aumal2 cells (Fig. 6B). In addition, the level of the 17-kDa subunit was decreased in Aumal2 mutant vacuoles (Fig. 6C). These phenotypes are similar to those observed for mutant cells deleted for the VMA3 gene, which encodes the 17-kDa subunit, in that both the peripheral and integral membrane subunits were not detected in Auma3 vacuolar membranes (26,48).
Detection of the VMAl2 Gene Product-Rabbit antiserum against E. coli-expressed Vmal2p was generated and used to detect the polypeptide in yeast cells. Fig. 7 shows the result of Western blotting analysis to detect Vmal2p in whole cell extracts and vacuolar membrane fractions. The affinity-purified anti-Vmal2p antibodies recognized a polypeptide with an apparent molecular mass of 25 kDa in both fractions. The identity of the cross-reacting polypeptide was confirmed by the lack of the 25-kDa species in Aumal2 extracts. These results indicate that a t least a portion of Vmal2p associates with the vacuolar membranes. We next examined whether Vmal2p is a component of the vacuolar membrane H+-ATPase. The enzyme was isolated by the method of Uchida et al. (20). Proteins in vacuolar membrane vesicles were solubilized with a zwitterionic detergent, ZW3-14, and size-fractionated by centrifugation through a fractions. Vacuolar membrane vesicles from YPH500 and RH202 cells were extracted with chloroform/methanol (2:l) as described under "Experimental Procedures." Proteins in the organic phase were resolved by SDS-PAGE on 13.5% acrylamide gels and detected by staining with Coomassie Brilliant Blue. Vma3p is the major proteolipid in yeast vacuoles (26,32). Protein bands with higher apparent masses are of aggregates of the proteolipid. glycerol density gradient. Fig. 8A shows the distribution of the ATPase and dipeptidyl aminopeptidase B activities on a Coomassie Blue stained gel. Western blots of the glycerol gradient fractions were probed with affinity-purified anti-Vmal2p antibody and a monoclonal antibody specific for the 60-kDa subunit of the vacuolar membrane H'-ATPase (Fig.  8B). This blot revealed that Vmal2p did not cosediment with the peak fraction of the ATPase activity (or the 60-kDa subunit) but instead migrated to lower density fractions with dipeptidyl aminopeptidase B, a 120-kDa vacuolar membrane protein (17,50). It is also noteworthy that there is no abundant band of Vmal2p at 25 kDa visible by Coomassie staining through the glycerol gradient fractions. However, because of the variability of protein staining by Coomassie Blue, we cannot reliably quantitate the level of Vmal2p in yeast cells by this method. If the Coomassie Blue staining is an accurate indicator of the level of VmalPp, then the results of Fig. 8 would indicate that Vmal2p is less abundant in vacuolar membranes than bona fide vacuolar membrane H'-ATPase subunits. The results presented in Fig. 8 are also consistent with the possibility that Vmal2p was detached from the vacuolar H'-ATPase complex during the course of purification and that this polypeptide is required for regulation or coupling of the activity of ATP hydrolysis to proton pumping.

DISCUSSION
In this paper, we report the isolation of the yeast VMAl2 gene, which is essential for the expression of the vacuolar membrane H'-ATPase. The vmal2 mutant was originally isolated as a type IV cls (or Pet-ck, cls7-clsll) mutant, clsl0 (30,34). The VMAl2 gene was cloned by complementation of the umal2-1 mutation and sequenced. The nucleotide sequence of the gene predicts an integral membrane polypeptide of 25.2 kDa.
Disruption of the VMA12 gene was not lethal but conferred upon yeast cells a Ca2+ and pH sensitivity, as well as a respiratory deficient phenotype. Vacuolar membrane vesicles isolated from the Puma12 mutant cells were devoid of ATPase activity, and the vacuoles in these mutant cells did not accumulate the fluorescent dye quinacrine, indicating that Avma12 mutant vacuoles were not acidified. The phenotypes exhibited by the Avmal2 mutant were indistinguishable from those of other vma mutants, suggesting that screening new Pet-cls mutants may identify additional VMA gene products that are required for the expression and molecular organization of the vacuolar membrane H'-ATPase (30,32).
We were interested in determining what effect the loss of Vmal2p had on the synthesis and assembly of the subunits of the vacuolar membrane H'-ATPase. Our results show that the synthesis and/or stability of the 100-kDa integral membrane subunit was reduced 5-10-fold in Aumal2 cells, whereas the steady-state levels of the peripheral membrane subunits (69, 60, 42, and 27 kDa) were unaffected. We then analyzed vacuolar membrane vesicles prepared from Aumal2 cells and determined that none of these ATPase subunits was associated with the vacuolar membrane in this mutant. The level of the 17-kDa subunit in vacuolar membranes was also decreased in the mutant cells. We concluded, therefore, that Vmal2p is essential for the normal synthesis and/or stability of the integral membrane portion of the V-ATPase (100-and 17-kDa polypeptides) and for the targeting of both the peripheral and integral membrane ATPase subunits to the vacuolar membrane. These results are similar to the findings for Avma3 and Avmall mutant cells (33,48). Deletion of VMA genes encoding the integral membrane polypeptides, Vma3p or Vmallp, from yeast cells affects the targeting of the A FIG. 8. Detection of Vmal2p in glycerol gradient fractions. Solubilized vacuolar membrane vesicles were applied to a 2040% glycerol gradient and fractionated as previously described (20). Twenty-two fractions of -500 p1 each were collected from the bottom of the centrifuge tube. Each fraction was assayed for ATPase and dipeptidyl aminopeptidase B activities. The fractions exhibiting maximal enzyme activities are shown. The proteins present in each fraction were prepared for gel electrophoresis as described (20). A constant percentage of each fraction was separated by SDS-PAGE on 10% acrylamide gels. Total proteins present in each fraction were detected by staining with Coomassie Brilliant Blue (A ). Western blots of the same glycerol gradient fractions were probed with anti-Vma2p and anti-Vmal2p antibodies (R). The proteins above the 69-kDa marker are being recognized by the anti-Vmal2p antibody. However, they are not VMAl2-encoded protein, because they are still present in Aurnal2 cells.

6+
Qnremaining peripheral and integral membrane subunits of the ATPase onto the vacuolar membrane. In contrast, results obtained for Aumal or Auma2 mutant cells demonstrate that deletions of the genes encoding these peripheral membrane subunits affect only the assembly of the remaining peripheral subunits onto the vacuolar membrane (26,48). The 100-and 17-kDa integral membrane VMA gene products are present at wild-type levels in vacuoles from either Aumal or Auma2 cells. Therefore, the phenotypes associated with the Avmal2 mutant are similar to mutants lacking VMA integral membrane polypeptides in that the assembly of all of the H+-ATPase subunits onto the vacuolar membrane is disrupted.
Finally, we were interested in determining whether Vmal2p was a subunit of the vacuolar membrane H+-ATPase. Our results indicate that although this protein associated with vacuolar membranes, Vmal2p did not copurify with glycerol gradient-purified vacuolar membrane H+-ATPase. Instead, Vmal2p sediments in a manner similar to other small integral membrane proteins (<150 kDa). We therefore conclude that Vmal2p is not a subGnit required by the vacuolar membrane There are several possibilities as to the function of Vmal2p. one possibility is that Vmal2p is actually a component of the vacuolar membrane H+-ATPase that is required for either proton pumping or regulation of the enzyme activity but is detached from the enzyme complex in the course of purification. Another possibility is that Vmal2p facilitates the assembly and/or targeting of the H+-ATPase subunits onto the vacuolar membrane. This possibility necessitates that proteins other than the prominent subunits of the vacuolar membrane H+-ATPase are required for the expression of this enzyme's activity. Examples of proteins that affect the assembly of enzyme complexes but are not subunits of the final complex have been reported. Recently, Ackerman and coworkers (58-62) reported the isolation of four genes, ATPlO-ATP13, that are required for the expression of the mitochondrial ATPase activity but are not structural genes of the enzyme. Of these, ATP13 is required for transcription of at least one of the ATPase subunits, subunit 9 (58). The ATP10-ATP12 gene products are required for the assembly of the functional Fo (ATPIO) or F1 complex (ATP11 and ATP12) (59-62). In addition, COX10 and COXll, which are essential for the assembly of the yeast cytochrome oxidase complex, do not encode subunits of this multisubunit enzyme (63).
We have found that at least a portion of Vmal2p associates with the vacuolar membrane. However, Vmal2p may also reside in membrane compartments other than the vacuole. If Vmal2p functions in assembly and/or targeting of the H' -ATPase complex to the vacuolar membrane, then determining the subcellular distribution of Vmal2p will be of particular interest. For example, an intriguing possibility is that Vmal2p could function as an inhibitor/regulator of the vacuolar H' -ATPase, as well as an assembly/targeting factor. If, as we predict, the vacuolar membrane H+-ATPase is assembled onto membranes in an early secretory pathway organelle, one could envisage that Vmal2p could associate with the H'-ATPase t o control H' translocation and thus regulate organelle pH in the various secretory pathway compartments that the H+-ATPase traverses en route to the vacuole. In this scenario, Vmal2p could reside predominantly in membrane compartments outside the vacuole (e.g. endoplasmic reticulum, Golgi bodies, and/or endosomal membranes), where the H' translocation activity of the vacuolar H'-ATPase may need to be very tightly controlled. Therefore, further studies on the function and localization of Vmal2p might be important in establishing where and when the subunits of the vacuolar H+-ATPase are assembled onto the membrane to form the active enzyme complex.
The characterization of the VMA12 gene product and the subsequent effects of umal2 deletions in yeast have generated fundamental questions regarding the biogenesis, assembly, and molecular organization of the vacuolar membrane H' -ATPase complex. Future efforts will be directed toward elucidating the function of Vmal2p in the synthesis and assembly of this multisubunit complex onto the vacuolar membrane. These studies should further our general understanding of the coordinated assembly of multisubunit enzyme complexes in membranes.