Subunit D (Vma8p) of the Yeast Vacuolar H+-ATPase Plays a Role in Coupling of Proton Transport and ATP Hydrolysis*

To investigate the function of subunit D in the vacuolar H+-ATPase (V-ATPase) complex, random and site-directed mutagenesis was performed on the VMA8 gene encoding subunit D in yeast. Mutants were selected for the inability to grow at pH 7.5 but the ability to grow at pH 5.5. Mutations leading to reduced levels of subunit D in whole cell lysates were excluded from the analysis. Seven mutants were isolated that resulted in pH-dependent growth but that contained nearly wild-type levels of subunit D and nearly normal assembly of the V-ATPase as assayed by subunit A levels associated with isolated vacuoles. Each of these mutants contained 2–3 amino acid substitutions and resulted in loss of 60–100% of proton transport and 58–93% of concanamycin-sensitive ATPase activity. To identify the mutations responsible for the observed effects on activity, 14 single amino acid substitutions and 3 double amino acid substitutions were constructed by site-directed mutagenesis and analyzed as described above. Six of the single mutations and all three of the double mutations led to significant (>30%) loss of activity, with the mutations having the greatest effects on activity clustering in the regions Val71–Gly80 and Lys209–Met221. In addition, both M221V and the double mutant V71D/E220V led to significant uncoupling of proton transport and ATPase activity, whereas the double mutant G80D/K209E actually showed increased coupling efficiency. Both a mutant showing reduced coupling and a mutant with only 6% of wild-type proton transport activity showed normal dissociation of the V-ATPase complexin vivo in response to glucose deprivation. These results suggest that subunit D plays an important role in coupling of proton transport and ATP hydrolysis and that only low rates of turnover of the enzyme are required to support in vivo dissociation.

The V-ATPases are multisubunit complexes composed of two domains (1)(2)(3)(4)(5)(6)(7). The V 1 domain is a 570-kDa peripheral complex composed of eight different subunits (subunits A-H) of molecular mass 70 -14 kDa that is responsible for ATP hydrolysis. The V 0 domain is a 260-kDa integral complex composed of five different subunits (a, d, c, cЈ, and cЉ) of 100 to 17 kDa that is responsible for proton translocation. The V-ATPases are structurally and evolutionarily related to the F-ATPases of mitochondria, chloroplasts, and bacteria that normally function in ATP synthesis (12)(13)(14)(15)(16)(17). Thus sequence homology can be detected between the nucleotide-binding subunits of the V-ATPase (A and B) and the nucleotide-binding subunits of the F-ATPase (␤ and ␣) (18,19). Sequence homology also exists between the proteolipid c subunits of the two classes (13,20), although no other sequence similarities have been identified in the remaining subunits.
An important question is the mechanism of coupling of proton transport and ATP hydrolysis in the V-ATPases. For the F-ATPases, the availability of high resolution structural data (21)(22)(23)(24) together with a series of studies demonstrating rotation within the F-ATPase complex (25)(26)(27)(28) have led to the rotary model for coupling (14,29,30). In this model, ATP hydrolysis within the ␣ 3 ␤ 3 head of the F 1 domain drives rotation of a central ␥ subunit, which is tightly linked to a ring of c subunits in the F 0 domain. Rotation of the ring of c subunits relative to the a subunit of F 0 (which is held fixed relative to the ␣ 3 ␤ 3 head by a peripheral stator) in turn leads to unidirectional proton transport. Central to this coupling mechanism is the role of the ␥ subunit, which has been shown to rotate relative to ␣ 3 ␤ 3 hexamer (25)(26)(27). In addition, mutations have been identified in the ␥ subunit that lead to uncoupling of ATP hydrolysis and proton transport by the F-ATPases (31,32).
Whereas no subunits in the V-ATPase complex show homology to the ␥ subunit, two subunits (D and E) are predicted from sequence analysis to have a similarly high ␣-helical content (33,34). To investigate the possible role of subunit D in coupling within the V-ATPase, random mutagenesis of the gene encoding subunit D in yeast (VMA8 (35)) was performed. Mutations that lead to significant loss of V-ATPase function in yeast lead to a conditional growth phenotype in which cells are unable to grow at neutral pH (36 -38). Sequencing and analysis of mutants isolated by this procedure together with analysis of subsequent mutations constructed by site-directed mutagenesis have led us to suggest that subunit D plays a role in coupling of proton transport and ATP hydrolysis by the V-ATPases. In addition, we have further analyzed the activity and coupling requirements for in vivo dissociation of the V-ATPase, a process believed to play an important role in regulation of V-ATPase activity in the cell (39,40).

EXPERIMENTAL PROCEDURES
Materials-Leupeptin, aprotinin, and pepstatin were obtained from Roche Molecular Biochemicals. Tran 35 S-label was purchased from ICN Biochemicals. Zymolase 100T was from Sekagaku America Inc. Concanamycin A was purchased from Fluka Chemical Corp. Yeast extract, dextrose, peptone, and yeast nitrogen base were from Difco. Molecular biological reagents were from New England Biolabs, Promega, and Life Technologies, Inc. 9-Amino-6-chloro-2-methoxyacridine (ACMA) was from Molecular Probes, Inc. All other chemicals were of analytical grade and most were from Sigma.
Mutagenesis-Random mutagenesis of VMA8 gene was performed using a PCR-based method as described previously (41). Briefly, the entire encoding region of VMA8 flanked by additional 5Ј-and 3Ј-noncoding sequence (approximately 50 bases in length) in pKH3203 was amplified by PCR under mutagenic conditions. The first 10 cycles of PCR were done in the presence of 0.25 mM Mn 2ϩ and 4.25 mM Mg 2ϩ . The product was then diluted 1:100 into the same buffer but containing 3 mM Mg 2ϩ in the absence of Mn 2ϩ and amplified for another 30 cycles. The PCR products were purified using a Qiagen PCR purification kit.
Yeast Transformation and Selection-The PCR products of VMA8, which were produced under the mutagenic conditions described above, were used to co-transform yeast strain KHY105 using an in vivo recombination method as described previously (41). Unique BamHI and XbaI sites flanking the VMA8-coding sequence were introduced into pKH3203 by site-directed mutagenesis. pKH3203 was cleaved with BamHI and XbaI, and the large 6.6-kilobase pair fragment lacking the VMA8 gene was purified by agarose gel electrophoresis. The mutagenized PCR products (0.2 g) and the 6.6-kilobase pair vector fragment (1 g) were mixed and used to co-transform the KHY105 strain using the lithium acetate method (42). The transformants were then selected on Ura Ϫ plates. To screen for VMA8 mutants leading to defective V-ATPase function, colonies were replica-plated on YPD plates buffered by 50 mM KH 2 PO 4 and succinate to pH 5.5 and 7.5. Mutants unable to grow at pH 7.5 were selected, and the mutant plasmids were recovered and used to retransform KHY105 to verify that the observed phenotype was due to a mutation in VMA8. The mutant forms of VMA8 constructed by site-directed mutagenesis were subcloned into pRS316 and transformed into KHY105 followed by selection of transformants on Ura Ϫ plates.
Analysis of Subunit D Expression and V-ATPase Assembly-Whole cell lysates were prepared from overnight cultures in Ura Ϫ medium as described previously (43), and the proteins were separated by SDS-PAGE on 10% acrylamide gels. The expression of subunit D was detected by Western blotting using a polyclonal antibody raised against Vma8p (a generous gift of Dr. Tom Stevens). Assembly of the V-ATPase was assessed by measurement of the amount of subunit A present on isolated vacuolar membranes and by immunoprecipitation of the V-ATPase complex from whole cell lysates. Vacuolar membrane vesicles were isolated as described previously (44). Following separation of the proteins on 12% acrylamide gels, Western blotting was performed using the monoclonal antibody 8B1-F3 against the A subunit (obtained from Molecular Probes, Inc.). It has previously been shown that the V 1 domain (including the A subunit) only associates with the vacuolar membrane if V-ATPase assembly is normal. Western blots were developed using a horseradish peroxidase-conjugated secondary antibody and visualized by a chemiluminescent detection system from Kirkegaard & Perry Laboratories.
Metabolic labeling of the V-ATPase using Tran 35 S-label, solubilization, and immunoprecipitation of V-ATPase using 8B1-F3 were carried out as described previously (45). Cells were grown overnight in methionine-free media to an absorbance at 600 nm of 0.6 -0.8. The cells were converted to spheroplasts by treatment with Zymolyase and incubated with Trans 35 S-label (50 Ci/5 ϫ 10 6 spheroplasts) for 1 h at 30°C. Spheroplasts were pelleted, washed, and lysed in PBS (135 mM NaCl, 2 mM KCl, 10 mM sodium phosphate, 1.75 mM potassium phosphate (pH 7.4)) with 1% C 12 E 9 and 1 mM DSP. The V-ATPase complex was then immunoprecipitated using the antibody 8B1-F3 against subunit A and protein A-Sepharose followed by separation of proteins by SDS-PAGE on 12% acrylamide gels and autoradiography.
In Vivo Dissociation of the V-ATPase in Response to Glucose Deprivation-Dissociation of the V-ATPase in response to glucose depletion was detected as described previously (39) with some modifications. The ⌬vma8 yeast strain expressing the wild-type or mutant forms ofVMA8 was grown overnight to an absorbance at 600 nm of 0.6 -0.8. The cells were converted to spheroplasts and incubated in YEP media or YEP media containing 2% glucose for 40 min at 30°C. Spheroplasts were pelleted and lysed in PBS containing 1% C 12 E 9 and 1 mM DSP. The V-ATPase complex was then immunoprecipitated using 8B1-F3 and protein A-Sepharose followed by separation on 12% acrylamide gels and transfer to nitrocellulose. Western blotting was then performed separately using the antibodies 8B1-F3 against the A subunit and 10D7 against the 100-kDa a subunit of the V 0 domain. Dissociation results in the disappearance of the 100-kDa a subunit from the complex immunoprecipitated with the antibody against the V 1 A subunit. Western blots were developed using a horseradish peroxidase-conjugated secondary antibody and visualized by a chemiluminescent detection system from Kirkegaard & Perry Laboratories.
Other Assays-Protein concentrations were determined by the method of Lowry et al. (46). ATPase activity was measured using a coupled spectrophotometric assay as described previously (47). V-ATPase activity was defined as the ATPase activity of isolated vacuoles (10 g of protein) that was inhibitable by 0.2 M concanamycin (48). Vacuoles isolated from the ⌬vma8 strain expressing the wild-type VMA8 gene had a specific activity of 1.3-1.5 mol of ATP/min/mg protein. Typically, concanamycin inhibited approximately 90 -95% of the ATPase activity in isolated vacuoles. ATP-dependent proton transport was measured by quenching of ACMA fluorescence using a Perkin-Elmer LS50B spectrofluorimeter (47). Vacuoles isolated from the ⌬vma8 strain expressing the vector alone showed no ATP-dependent quenching of ACMA fluorescence.

RESULTS
To evaluate the functional role of subunit D in the V-ATPase complex, PCR-based random mutagenesis of the VMA8 gene encoding subunit D in yeast was performed. Disruption of V-ATPase function leads to a conditional growth phenotype in yeast in which cells are unable to grow at pH 7.5 but retain the ability to grow at pH 5.5 (36 -38). Following mutagenesis and transformation, transformants were selected for the inability to grow at pH 7.5. Screening of approximately 10,000 transformants resulted in the selection of 29 colonies unable to grow at neutral pH. These mutants were then further screened for stable expression of Vma8p by Western blotting of whole cell lysates using a polyclonal antibody specific for the yeast subunit D. Of the initial isolates, 15 colonies showed nearly normal levels of subunit D in whole cell lysates. Plasmids from these mutants were recovered, amplified in Escherichia coli, and reintroduced into the vma8-deficient strain. Fourteen of the mutant plasmids still led to the vma Ϫ phenotype and were sequenced. All of the mutant plasmids contained multiple point mutations. Seven plasmids, containing either 2 or 3 mutations, were subjected to further analysis. Previous attempts to isolate mutants in VMA8 causing the vma Ϫ phenotype had also resulted in only mutants containing multiple point mutations. 2 Fig. 1 shows a Western blot of whole cell lysates isolated from each of the mutant strains using a polyclonal antibody directed against Vma8p. As can be seen, all seven of the mutants showed nearly normal levels of subunit D relative to cells expressing the wild-type VMA8 gene. Several of the mutants showed slightly lower mobility (i.e. higher apparent molecular weight) relative to the wild-type protein, which might be due to altered SDS binding as a result of the mutations introduced. To evaluate the effects of the mutations on activity of the V-ATPase, vacuoles were isolated from cells expressing the wildtype and mutant forms of subunit D, and ATPase activity and proton transport were measured using a coupled spectrophotometric assay and uptake of the fluorescent dye ACMA, respectively. V-ATPase activity was defined as that fraction of the ATPase activity that was inhibited by 0.2 M concanamycin (48), typically approximately 90 -95% in isolated vacuoles. ATPase activities are expressed relative to that measured for vacuoles isolated from cells expressing the wild-type VMA8 gene (1.3-1.5 mol of ATP/min/mg of protein). As can be seen from Table I, all of the mutants showed loss of at least 58% of ATPase activity and 60% of ACMA quenching. Generally the loss of proton transport paralleled loss of ATPase, although for mutants 2 and 2Ј, loss of proton transport was slightly greater than loss of ATPase, whereas for mutant 3Ј the opposite was true. Several point mutations appeared in more than one mutant, including G80D, I188N, E220V, and M221V, and the mutations generally clustered in two regions, from Val 71 to Gly 80 and from Lys 209 to Met 221 .
To determine whether loss of V-ATPase activity was due to a disruption of assembly of the V-ATPase complex, Western blot analysis of purified vacuoles was performed using a monoclonal antibody directed against the A subunit of the V 1 domain. As has previously been shown (43), disruption of assembly of the V-ATPase complex leads to release of the entire V 1 domain from the vacuolar membrane, resulting in the loss of A subunit staining by Western blot. As can be seen in Fig. 2, all of the mutants showed nearly normal levels of A subunit on the vacuolar membrane. As a further test of assembly and as a measure of stability of the V-ATPase complex, the mutant strains were also metabolically labeled with Tran 35 S-label followed by cell lysis, detergent solubilization, and immunoprecipitation of the V-ATPase complex using the monoclonal antibody 8B1-F3 against the A subunit. As can be seen in Fig. 3, most of the mutants showed normal stability of the V-ATPase complex, displaying the full complement of V-ATPase subunits. Mutant 2Ј, however, showed a dramatic reduction in stability of the V-ATPase complex, as indicated by the near complete absence of V 0 subunits (i.e. subunits a and c) in the immunoprecipitate. Mutant 2 also showed somewhat reduced levels of subunit d in the immunoprecipitate. Because these mutants showed either slight or no reduction in the levels of A subunit present on vacuolar membranes (Fig. 2), these mutations in the D subunit most likely lead to reduced stability of the V-ATPase complex (i.e. inability to survive detergent solubilization and immunoprecipitation) rather than to a defect in assembly of the V-ATPase.
Because all of the mutants isolated from the initial screen contained multiple point mutations within the VMA8 gene, it was necessary to determine what effect each of the single mutations would have on activity of the V-ATPase. Fourteen single mutations were constructed in the VMA8 gene by sitedirected mutagenesis. In addition, three double mutations, containing two of the three point mutations in mutants 2, 4, and 2Ј were also constructed. The yeast strain disrupted in the VMA8 gene was then transformed with the wild-type gene and each of the mutant plasmids and the transformants analyzed as described above. All of the single mutants and the G80D/E220V and L149V/D249G double mutants showed a wild-type growth phenotype (i.e. normal growth at pH 7.5), whereas the V71D/ E220V double mutant showed a partial growth defect (that is it grew slower at pH 7.5 than the wild-type strain but not as slowly as the deletion strain). Western blot analysis of whole cell lysates (Fig. 4) indicated that all of the mutants showed nearly wild-type levels of Vma8p. Western blot analysis of  Table I). The proteins were separated by SDS-PAGE on 10% acrylamide gels and transferred to nitrocellulose. Western blotting was then performed using a polyclonal antibody against subunit D as described under "Experimental Procedures." a V-ATPase activity was defined as the ATPase activity of isolated vacuoles (10 g of protein) that was inhibitable by 0.2 M concanamycin. Values are expressed relative to that measured for vacuoles isolated from the ⌬vma8 strain expressing the wild type VMA8 gene (1.3-1.5 mol of ATP/min/mg protein). Typically, concanamycin inhibited approximately 90 -95% of the ATPase activity in isolated vacuoles. ATP hydrolysis was measured using a coupled spectrophotometric assay as described previously. Values represent the average of at least two independent determinations with the numbers in parentheses corresponding to the average deviation from the mean.
b Proton transport of isolated vacuoles (2.5 g of protein) was measured as ATP-dependent quenching of ACMA fluorescence as previously described (47). Values are expressed relative to that measured for vacuoles isolated from the ⌬vma8 strain expressing the wild type VMA8 gene (defined as 100%). Vacuoles isolated from the ⌬vma8 strain expressing the vector alone showed no ATP-dependent quenching of ACMA fluorescence. Values represent the average of at least two independent determinations with the numbers in parentheses corresponding to the average deviation from the mean.  Table I). The proteins were separated by SDS-PAGE on 12% acrylamide gels and transferred to nitrocellulose. Western blotting was then performed using the monoclonal antibody 8B1-F3 against subunit A as described under "Experimental Procedures." vacuoles isolated from each of the mutant strains also showed approximately wild-type levels of the A subunit, indicating normal assembly of the V-ATPase in each of the mutants (Fig.  5).
Measurement of ATPase activity and ACMA quenching in vacuoles isolated from each of the mutant strains (Table II) revealed that many of the mutations were without effect on either ATPase activity or proton transport. Six of the single mutations, however, reduced ATPase activity by greater than 35% and proton transport activity by greater than 30%. Of these, two mutations (V71D and M221V) had particularly large effects on activity, with V71D reducing ATPase activity and pumping by 70 and 58%, respectively, whereas M221V reduced ATPase activity and pumping by 50 and 70%, respectively. All of the double mutants constructed (including G80D/E220V, L149V/D249G, and V71D/E220V) also showed dramatically reduced activity, with the V71D/E220V mutant showing the lowest activity (20% of wild-type ATPase activity and no measurable proton transport). In two cases (M221V and the V71D/ E220V double mutant), the reduction in proton transport activity was significantly greater than the reduction in ATP hydrolysis, suggesting a partial uncoupling of these two activities. It is also interesting to note that for the G80D/K209E double mutant isolated in the first round, the greater reduction in ATPase activity than proton transport suggests an increase in coupling efficiency for this mutant. These results suggest that subunit D plays a role in coupling of proton transport and ATPase activity by the V-ATPase.
To determine the effect of these mutations on stability of the V-ATPase complex, cells were metabolically labeled, and the V-ATPase was solubilized and immunoprecipitated as described above. As can be seen in Fig. 6, all of the single and double mutants showed the normal complement of V-ATPase subunits except for two. Both V71D and the V71D/E220V double mutant showed greatly reduced stability relative to the wild-type enzyme. The L149V/D249G double mutant also showed slightly reduced levels of V 0 subunits in the immunoprecipitate. As noted above, however, these mutations are more likely to have reduced stability rather than assembly of the V-ATPase complexes. were prepared from the ⌬vma8 strain expressing the wild-type (WT) VMA8 gene in pRS316, the pRS316 vector alone (Vector), or the mutant forms of VMA8 in pRS316 as indicated. The proteins were separated by SDS-PAGE on 10% acrylamide gels and transferred to nitrocellulose. Western blotting was then performed using a polyclonal antibody against subunit D as described under "Experimental Procedures."

FIG. 5. Effect of site-directed mutations of VMA8 on the association of subunit A with the vacuolar membrane.
Vacuolar membranes (10 g of protein) were prepared from the ⌬vma8 strain expressing the wild-type (WT) VMA8 gene in pRS316, the pRS316 vector alone (Vector), or the mutant forms of VMA8 in pRS316 as indicated. The proteins were separated by SDS-PAGE on 12% acrylamide gels and transferred to nitrocellulose. Western blotting was then performed using the monoclonal antibody 8B1-F3 against subunit A as described under "Experimental Procedures." a V-ATPase activity was defined as the ATPase activity of isolated vacuoles (10 g of protein) that was inhibitable by 0.2 M concanamycin. Values are expressed relative to that measured for vacuoles isolated from the ⌬vma8 strain expressing the wild type VMA8 gene (1.3-1.5 mol of ATP/min/mg of protein). Values represent the average of at least two independent determinations with the numbers in parentheses corresponding to the average deviation from the mean.

TABLE II V-ATPase activity and proton transport of vacuoles isolated from cells expressing mutant forms of vma8 derived by site-directed mutagenesis
b Proton transport of isolated vacuoles (2.5 g of protein) was measured as ATP-dependent quenching of ACMA fluorescence as described previously (47). Values are expressed relative to that measured for vacuoles isolated from the ⌬vma8 strain expressing the wild type VMA8 gene (defined as 100%). Values represent the average of at least two independent determinations with the numbers in parentheses corresponding to the average deviation from the mean.

FIG. 3. Effect of random mutations of VMA8 on the assembly and stability of the V-ATPase complex.
The ⌬vma8 strain expressing the wild-type (WT) VMA8 gene in pRS316, the pRS316 vector alone (Vector), or the mutant forms of VMA8 in pRS316 were grown overnight in methionine-free media to an absorbance at 600 nm of 0.6 -0.8. The cells were converted to spheroplasts by treatment with Zymolyase and incubated with Trans 35 S-label (50 Ci/5 ϫ 10 6 spheroplasts) for 1 h at 30°C. Spheroplasts were pelleted, washed, and lysed in PBS with 1% C 12 E 9 and 1 mM DSP. The V-ATPase complex was then immunoprecipitated using the monoclonal antibody 8B1-F3 against subunit A and protein A-Sepharose followed by separation of proteins by SDS-PAGE on 12% acrylamide gels and autoradiography.
Finally, dissociation of the V-ATPase complex has been shown to occur in yeast in response to glucose deprivation (39) and has been suggested to play a role in regulation of V-ATPase activity both in yeast (39) and higher eukaryotes (40). Glucosedependent dissociation of the V-ATPase in yeast has been shown to require catalytic activity (49,50), but the level of activity required for dissociation has not been investigated. In addition, because no uncoupled mutants of the V-ATPase have previously been isolated, it has not been possible to assess the requirement for tight coupling of the in vivo dissociation process. To address these questions, a mutant possessing very low activity (mutant 2) and a partially coupled mutant (M221V) were incubated in the presence or absence of glucose followed by detergent solubilization and immunoprecipitation of the V-ATPase complex using the monoclonal antibody 8B1-F3 against the A subunit. Following SDS-PAGE, Western blotting was performed on the immunoprecipitates using antibodies against both the A subunit and the 100-kDa a subunit of the V 0 domain. As can be seen in Fig. 7, incubation of cells expressing the wild-type VMA8 gene in the absence of glucose resulted in dissociation of the V 1 and V 0 domains as reflected in the reduced amount of the 100-kDa a subunit immunoprecipitated using the antibody against subunit A. Interestingly, both mutant 2 and the M221V mutant showed comparable levels of dissociation, suggesting that neither high levels of activity nor tight coupling between proton transport and ATPase activity is required for glucose-dependent dissociation of the V-ATPase in vivo. Because mutant 2 showed altered levels of subunit d on immunoprecipitation of the complex using antibodies against subunit A (Fig. 3), it was possible that the observed dissociation might not reflect the normal reversible dissociation observed in vivo. We therefore tested the reversibility of the observed dissociation on readdition of glucose. We found normal reassembly of the V-ATPase (as reflected in coimmunoprecipitation of subunits a and A) when glucose was added back to mutant 2 cells that had been glucose-depleted (data not shown). This suggests that mutant 2 is competent to undergo glucose-dependent dissociation in vivo.

DISCUSSION
Electron microscopy of the V-ATPases (51-54) and F-AT-Pases (55-57) has revealed both an overall similarity in structure of the two complexes and a number of striking differences. Thus, both complexes are composed of a peripheral head piece (V 1 or F 1 ) attached to a membrane-bound domain (V 0 or F 0 ) by both a central and peripheral stalk. The existence of both central and peripheral stalks is an essential feature of the rotary model of coupling that has been proposed for the F-ATPases (14,29,30). For the F-ATPases, the central stalk is composed of the highly ␣-helical ␥ subunit together with the ⑀ subunit (23,58), whereas the peripheral stalk is composed of the ␦ subunit together with the soluble domain of subunit b (59,60). The central stalk has been shown to rotate during ATP hydrolysis relative to the ␣ 3 ␤ 3 hexameric head of F 1 (25)(26)(27). This in turn drives rotation of a ring of c subunits relative to the a subunit in F 0 (28). The peripheral stalk functions to hold the ␣ 3 ␤ 3 hexamer rigid relative to the a subunit during rotation of the central stalk. Electron microscopic images of the V-ATPases (51-54) suggest a more complex structure, possibly involving the existence of multiple peripheral stalks (54), although the location of particular subunits within these structures has not yet been identified. Although no V-ATPase subunits are homologous to the ␥ subunit of the F-ATPase, two subunits (D and E) show a similarly high content of predicted ␣-helix (33,34). Previous mutagenesis studies of the gene encoding subunit E in yeast (VMA4) have identified a temperature-sensitive mutation in VMA4 (D145G) affecting assembly of the V-ATPase complex (61) but no mutations directly affecting activity of an assembled enzyme. Cross-linking studies of the bovine coated vesicle enzyme have identified contacts between subunits D and F (62), consistent with the D/F subcomplex that can be identified in yeast strains lacking any of the other V 1 subunits (63). By contrast subunit E is able to crosslink to a variety of other subunits, including subunits C, G, H, and the 100-kDa subunit a of the V 0 domain (62). For this reason we have previously suggested that subunit D may be the V-ATPase counterpart to subunit ␥. Nevertheless, no functional information had been obtained suggesting that subunit D might play a similar role to the ␥ subunit in the V-ATPases.
To address the function of subunit D in the V-ATPase complex, random mutagenesis of the VMA8 gene was performed. Interestingly, all of the vma8 mutants isolated that displayed a vma Ϫ phenotype (i.e. inability to grow at neutral pH) contained multiple point mutations in the VMA8 gene. These mutants displayed 7-42% of wild-type ATPase activity and 0 -40% of FIG. 6. Effect of site-directed mutations of VMA8 on the assembly and stability of the V-ATPase complex. The ⌬vma8 strain expressing the wild-type (WT) VMA8 gene in pRS316, the pRS316 vector alone (Vector), or the mutant forms of VMA8 in pRS316 were grown overnight in methionine-free media to an absorbance at 600 nm of 0.6 -0.8. The cells were converted to spheroplasts, and the spheroplasts were incubated with Trans 35 S-label, pelleted, washed, and lysed as described in the legend to Fig. 3. The V-ATPase complex was then immunoprecipitated using the monoclonal antibody 8B1-F3 followed by SDS-PAGE and autoradiography as also described. FIG. 7. Effect of mutations of VMA8 on the in vivo dissociation of the V-ATPase in response to glucose depletion. The ⌬vma8 strain expressing the wild-type (WT) or the mutant forms of VMA8 indicated were grown overnight to an absorbance at 600 nm of 0.6 -0.8. The cells were converted to spheroplasts and incubated in YEP media or YEP media containing 2% glucose for 40 min at 30°C. Spheroplasts were pelleted and lysed in PBS containing 1% C 12 E 9 and 1 mM DSP. The V-ATPase complex was then immunoprecipitated using the monoclonal antibody 8B1-F3 and protein A-Sepharose followed by separation by SDS-PAGE on 12% acrylamide gels and transfer to nitrocellulose. Western blotting was then performed separately using the antibodies 8B1-F3 against the A subunit and 10D7 against the 100-kDa a subunit of the V 0 domain as described under "Experimental Procedures." wild-type proton transport activity. To identify the amino acid changes responsible for the observed effects on activity, a series of single point mutations were constructed by site-directed mutagenesis of the VMA8 gene. Of these, two had the most dramatic effects on activity. V71D near the N terminus reduced both pumping and ATPase activity by 60 -70%, whereas M221V near the C terminus showed only 50% of wild-type ATPase activity and 30% of wild-type proton pumping. This partial uncoupling of proton transport and ATP hydrolysis was also observed for the V71D/E220V double mutant, which had 20% residual ATPase activity but no proton transport. These mutants represent the first "uncoupled" mutants identified for the V-ATPases and suggest that subunit D plays a role in coupling of proton transport and ATP hydrolysis. That the double mutant G80D/K209E shows a greater reduction in ATPase activity than proton transport suggests that the wildtype enzyme may not be as tightly coupled as possible.
Interestingly, many of the mutations affecting activity are clustered in the regions Val 71 -Gly 80 near the N terminus and Lys 209 -Met 221 near the C terminus (the latter region being 35-45 residues from the C terminus). In addition, several mutants bearing mutations in both of these regions show more than an additive reduction in activity (for example, the G80D/ K209E mutant has a much lower activity than predicted from the effects of each of the single mutations). One possibility is that these regions interact in the folded structure of the protein. Such an interaction between the N-and C-terminal portions of the protein would be predicted if subunit D, like the ␥ subunit of the F-ATPase (21)(22)(23)(24), forms an extended helical rod that folds back on itself. Alternatively, each region may represent an important contact region between subunit D and other subunits in the V-ATPase complex. Subunit ␥ of the F-ATPases makes important contacts with the ␤ subunit in four regions of the protein. Residues in the region 80 -90 amino acids from the N terminus are in close contact with the "DELSEED" sequence of ␤ (21), and introduction of cysteine residues at these sites can be used to cross-link ␤ and ␥ (25,58). Mutations at both Met 23 as well as Asn 269 and Thr 273 cause uncoupling of proton transport and ATP hydrolysis (31,32), whereas mutations in the region 236 -246 suppress the uncoupling phenotype (64) (these latter being 40 -50 residues from the C terminus of ␥). The x-ray crystal structure of F 1 also places the 236 -246 region near the DELSEED sequence, whereas residues 269 and 273 appear to interact with Asp 316 of ␤ (21). Thus for both subunit D and subunit ␥, the regions 70 -80 residues from the N terminus and 40 -50 residues from the C terminus appear to be structurally or functionally important. Whether these regions of subunit D actually interact with subunit A in the V-ATPase complex will require further analysis using, for example, crosslinking or isolation of second-site suppressors. It should be noted, however, that mutations have been identified in other subunits of the F 1 complex (for example the ␦ subunit (65)) that also cause uncoupling of proton transport and ATP hydrolysis.
A central question concerns the mechanism of regulation of V-ATPase activity in vivo. Reversible dissociation of the V 1 and V 0 domains has been shown to occur in both yeast (39) and insects (40) and has been suggested to play a role in controlling V-ATPase activity in the cell. Mutations that cause inactivation of the V-ATPase have been shown to block dissociation of the V-ATPase in response to glucose deprivation in yeast (49,50). Moreover, a mutant capable of binding nucleotides but catalytically inactive showed no glucose-dependent dissociation, indicating that nucleotide binding was not sufficient to promote dissociation in vivo (50). In the current study we observed that a mutant possessing only 6% of wild-type proton transport and 10% of wild-type ATPase activity was still able to dissociate in response to glucose depletion, indicating that only a very low level of activity is necessary for dissociation to occur. It may be that the enzyme must pass through a particular conformational state in its catalytic cycle to be competent for dissociation. Moreover, tight coupling of ATP hydrolysis and proton transport is not necessary for dissociation. The V71D/ E220V double mutant possessing 20% of wild-type ATPase activity but no proton transport activity would be particularly interesting to investigate in this regard, but unfortunately this mutant was not stable to detergent solubilization and immunoprecipitation (Fig. 6), making it unsuitable for these studies.
In conclusion, we have provided evidence that subunit D plays a role in coupling of proton transport and ATPase activity by the V-ATPases. We have also further refined the activity requirements of in vivo dissociation of the V-ATPase. Further work will be required to identify the subunits in the V-ATPase complex with which subunit D functionally interacts.