Site-directed mutagenesis of the yeast V-ATPase A subunit.

To investigate the function of residues at the catalytic nucleotide binding site of the V-ATPase, we have carried out site-directed mutagenesis of the VMA1 gene encoding the A subunit of the V-ATPase in yeast. Of the three cysteine residues that are conserved in all A subunits sequenced thus far, two (Cys284 and Cys539) appear essential for correct folding or stability of the A subunit. Mutation of the third cysteine (Cys261), located in the glycine-rich loop, to valine, generated an enzyme that was fully active but resistant to inhibition by N-ethylmalemide, 7-chloro-4-nitrobenz-2-oxa-1,3-diazole, and oxidation. To test the role of disulfide bond formation in regulation of vacuolar acidification in vivo, we have also determined the effect of the C261V mutant on targeting and processing of the soluble vacuolar protein carboxypeptidase Y. No difference in carboxypeptidase Y targeting or processing is observed between the wild type and C261V mutant, suggesting that disulfide bond formation in the V-ATPase A subunit is not essential for controlling vacuolar acidification in the Golgi. In addition, fluid phase endocytosis of Lucifer Yellow, quinacrine staining of acidic intracellular compartments and cell growth are indistinguishable in the C261V and wild type cells. Mutation of G250D in the glycine-rich loop also resulted in destabilization of the A subunit, whereas mutation of the lysine residue in this region (K263Q) gave a V-ATPase complex which showed normal levels of A subunit on the vacuolar membrane but was unstable to detergent solubilization and isolation and was totally lacking in V-ATPase activity. By contrast, mutation of the acidic residue, which has been postulated to play a direct catalytic role in the homologous F-ATPases (E286Q), had no effect on stability or assembly of the V-ATPase complex, but also led to complete loss of V-ATPase activity. The E286Q mutant showed labeling by 2-azido-[32P]ATP that was approximately 60% of that observed for wild type, suggesting that mutation of this glutamic acid residue affected primarily ATP hydrolysis rather than nucleotide binding.

The vacuolar (H ϩ )-ATPases (or V-ATPases) are a class of ATP-dependent proton pumps that play an important role in a variety of cellular processes, including receptor-mediated endocytosis, intracellular targeting, macromolecular processing and degradation, and coupled transport (1)(2)(3)(4)(5)(6)(7)(8)(9). The V-ATPases are located in both intracellular compartments and in the plasma membrane of certain specialized cells (7, 10 -12). In yeast, acidification of the central vacuole by the V-ATPase serves to activate degradative enzymes and to drive uptake of solutes such as Ca 2ϩ and amino acids (5).
Both the A and B subunits participate in nucleotide binding by the V-ATPases (36 -40) and show approximately 25% sequence identity with the ␤ and ␣ subunits of the F-ATPases (20,(41)(42)(43)(44)(45)(46). The F-ATPases normally function in ATP synthesis in mitochondria, chloroplasts, and bacteria (47)(48)(49)(50)(51), and a recent x-ray crystal structure of the F 1 domain has revealed a hexameric arrangement of ␣ and ␤ subunits, with the catalytic sites located primarily on ␤ and the noncatalytic sites located primarily on ␣ (52). Several lines of evidence suggest that the catalytic nucleotide binding sites of the V-ATPase are located on the A subunit. First, ATP-protectable labeling of the A subunit by agents such as NEM 1 and NBD-Cl correlates with inhibition of activity (for references, see (1)). Second, modification of a single A subunit cysteine residue (Cys 254 of the bovine A subunit) by NEM, cystine, or by disulfide bond formation with Cys 532 leads to inactivation of the V-ATPase (38,53,54). Third, labeling of the A subunit by the photoaffinity analog 2-azido-[ 32 P]ATP correlates well with inhibition by this reagent (39). Finally, the A subunit, unlike the B subunit, possesses several consensus sequences, including the glycine-rich loop, which appear to be critical for ATP hydrolysis by the homologous F-ATPase ␤ subunit (55)(56)(57).
To further probe the structure of the catalytic nucleotide binding sites on the V-ATPase, we have carried out site-directed mutagenesis of the VMA1 gene in yeast. We have focused first on the three highly conserved cysteine residues (Cys 261 , Cys 284 , and Cys 539 ), two of which have been postulated to be involved in regulation of V-ATPase activity through disulfide bond formation (53,54). We have also tested the effects of mutation of residues in the glycine rich loop region and of an acidic residue which has been postulated to play a direct role in ATP hydrolysis.

EXPERIMENTAL PROCEDURES
Materials and Strains-Zymolyase 100T was obtained from Sekagaku America, Inc. 2-Azido-[ 32 P]ATP labeled in the ␤ and ␥ positions (58) was a kind gift from Dr. Richard Cross, Department of Biochemistry and Molecular Biology, SUNY, Syracuse. Concanamycin A was obtained from Fluka, Inc. ATP and most other chemicals were purchased from Sigma.
All mutants were confirmed by DNA sequencing using the dideoxy method (59) and subcloned into the YIp5 vector using BamHI and SalI sites along with wild type VMA1. No other mutations were detected in the final product.
Transformation-The wild type plasmid (YIp5-VMA1), vma1 mutant plasmids, or YIp5 vector alone were linearized with ApaI to target integration of the constructs to the URA3 locus. Yeast cells (SF838-5A␣ vma1⌬-8) were transformed with the linearized plasmids using the lithium acetate method (60). The transformants were selected on Ura Ϫ plates as described previously (61). Chromosomal DNA was isolated from the transformed yeast cells and the VMA1 gene was amplified by a polymerase chain reaction. The presence of the expected mutations in VMA1 was confirmed by sequencing the polymerase chain reaction products. Then the mutants were tested for growth on pH 7.5 or 5.5 YPD plates buffered with 50 mM KH 2 PO 4 and 50 mM succinic acid (62).
Purification of the V-ATPase Complex-The yeast integrated with wild type plasmid (YIp5-VMA1), mutants, or the vector YIp5 alone (as a negative control) were cultured overnight in 1-liter of YPD (pH 5.5) to log phase. Vacuoles were isolated from the different transformants as described previously (63). For purification of the V-ATPase, vacuolar membranes were washed three times with EDTA and solubilized with Zwittergent 3-14, and the V-ATPase was isolated by density gradient sedimentation on 20 -50% glycerol gradients as described previously (64).
Targeting and Processing of Carboxypeptidase Y (CpY)-The procedures used for growth and metabolic labeling of yeast cells were essentially as described previously (65,66). Cells were grown in yeast nitrogen base medium buffered at pH 5.5 to mid-log phase. Approximately 7 ϫ 10 7 cells were collected and converted to spheroplasts as described previously (66). The spheroplasts were labeled with Tran 35 S-label (250 Ci) for 20 min at 30°C and chased for various times in medium containing 8 mM methionine and 4 mM cysteine, and aliquots containing 2 ϫ 10 7 cells were removed. Samples were separated into intracellular (spheroplast) and extracellular (periplasm plus medium) fractions by centrifugation and then precipitated with trichloroacetic acid. Samples were solubilized in buffer containing 1% SDS and 3 M urea, and immunoprecipitations were carried out using an antibody against CpY (the kind gift of Dr. Daniel Klionsky) and protein A-Sepharose as described previously (65). The immunoprecipitated proteins were eluted by heating in Laemmli sample buffer at 75°C for 8 min followed by SDS-PAGE on an 8% acrylamide gel and autoradiography.
Quinacrine Staining of Acidic Compartments-ϳ5 ϫ 10 6 log-phase yeast cells were harvested, resuspended in 500 l of YPD buffered with 50 mM Na 2 HPO 4 (pH 7.6) containing 200 M quinacrine and incubated at room temperature for 5 min as described previously (63). Cells were sedimented at 10,000 ϫ g for 5 s, washed once with 500 l of 2% glucose buffered with 50 mM Na 2 HPO 4 (pH 7.6), and resuspended in 100 l of the same solution. 10-l samples were applied to a concanavalin Acoated microscope slide and viewed immediately in the fluorescence microscope using a fluorescein filter.
Labeling of the V-ATPase by 2-Azido-[ 32 P]ATP-About 5 g of purified V-ATPase were incubated with 0.3 mM 2-azido-[ 32 P]ATP (100 cpm/ pmol) and 5 mM MgSO 4 in the absence or presence of 3 mM ATP at 4°C in the dark for 20 min, followed by irradiation of the sample with a short wavelength mineral light UV lamp at a distance of 2 cm at 4°C for 20 min. To each sample was added 5 ϫ concentrated Laemmli sample buffer followed by SDS-PAGE on a 12% acrylamide gel and autoradiography. Quantitation was done using a PhosphorImager system (Molecular Dynamics).
Other Procedures-ATPase activity was measured using a coupled spectrophotometric assay in the absence or presence of 0.5 M concanamycin A (63) with the modification of using 0.35 mM NADH instead of 0.5 mM NADH. SDS-PAGE was carried out as described by Laemmli (67), and silver staining was performed by the method of Oakley et al. (68). Protein concentrations were determined by Lowry et al. (69) assay.
For the measurement of yeast growth curves, overnight yeast cell cultures were diluted into pH 7.5 or 5.5 YPD media buffered with 50 mM KH 2 PO 4 and 50 mM succinic acid, and the absorbance at 600 nm was measured starting at a value of 0.5.
For immunoblot analysis, whole cell lysates and solubilized vacuoles were prepared as described elsewhere (70), subjected to SDS-PAGE, and probed with the mouse monoclonal antibody 8B1-F3 against the 69-kDa subunit (Molecular Probes, Inc.), followed by horseradish peroxidase-conjugated secondary antibody (Bio-Rad). Blots were developed using a chemiluminescent detection method obtained from Kirkegaard & Perry Laboratories. Quantitation was carried out using an IS-1000 digital imaging system (Alpha Innotech Corp.).

RESULTS
Nine separate mutations have been constructed in the VMA1 gene encoding the yeast A subunit and these mutant genes have been expressed in a vma1⌬ strain in which the VMA1 gene has been disrupted. Mutations were constructed in three classes of A subunit residues. First, mutations were made in three highly conserved cysteine residues (Cys 261 , Cys 284 , and Cys 539 ), which are located, respectively, in the glycine-rich loop, the GXGER consensus sequence, and the C-terminal domain. Second, mutations were constructed in two other residues (Gly 250 and Lys 263 ) postulated to be important on the basis of their location in the glycine-rich loop region. Finally, Glu 286 was mutated to test its possible role in catalysis of ATP hydrolysis. Of the mutations tested, only one (C261V) gave a wild type growth phenotype, whereas the remaining mutations (C284S, C284V, C539S, C539T, C539D, G250D, K263Q, and E286Q) all showed the growth phenotype of the deletion strain; that is, they were unable to grow at pH 7.5 but were able to grow at pH 5.5. Comparison of the bafilomycin-sensitive ATPase activity in vacuoles isolated from the wild-type, mutant, and deletion strains revealed that all of the mutations except the C261V mutation resulted in complete loss of functional V-ATPase activity. The specific activity of the C261V mutant (1.06 Ϯ 0.10 mol of ATP/min/mg of protein) was indistinguishable from the wild type control (0.98 Ϯ 0.08 mol of ATP/min/mg of protein).
To determine the basis for the observed loss in V-ATPase activity for the remaining eight mutations, Western blot analysis using a monoclonal antibody specific for the A subunit was performed on both whole cell homogenates and isolated vacuoles. The presence of the A subunit on the vacuolar membrane provides an initial measure of the competence of the mutant protein to support assembly of the V-ATPase complex. As can be seen in Fig. 1, mutations in the remaining two conserved cysteine residues (C284S, C284V, C539S, C539T, and C539D) led to the absence of a detectable signal in whole cell homoge-nates, suggesting that these mutations resulted in a loss of stability of the A subunit. Alternatively, one of these mutations may have disrupted the epitope recognized by the monoclonal antibody (8B1-F3). As a further test, Western blot analysis on whole cell homogenates was also carried out using a second monoclonal antibody (7D5) directed against the A subunit. As with the antibody 8B1-F3, no immunoreactive band was observed for any of the C284 or C539 mutants (data not shown). While these results are not conclusive because the epitopes recognized by these antibodies have not been mapped, they suggest that mutations at these two cysteine residues lead to destabilization of the A subunit. By contrast, C261V, E286Q, and K263Q gave nearly normal levels of A subunit in the cell homogenate, and the former two mutations showed normal association of the A subunit with the vacuolar membrane. Quantitation of the Western blot in Fig. 1B by densitometry indicates that the K263Q mutation resulted in a 10% decrease in the amount of A subunit associated with the vacuole. The G250D mutation resulted in a 60% decrease in the amount of A subunit in whole cells and the complete loss of V-ATPase assembly on the vacuolar membrane.
As a further test for assembly and stability of the V-ATPase complex, vacuolar membranes were solubilized with Zwittergent 3-14, and the V-ATPase was isolated by density gradient sedimentation as described previously (64). As can be seen in Fig. 2, normal assembly of the V-ATPase as assessed by the presence of the complete complement of V-ATPase subunits in the purified enzyme was observed for both the C261V and E286Q mutants. The K263Q mutant, despite the presence of 90% of the wild type level of A subunit on the vacuolar membrane (Fig. 1B), showed greatly reduced levels of assembly, suggesting that the K263Q mutation leads to reduced stability of the V-ATPase. None of the remaining mutants showed any detectable levels of stable V-ATPase complex.
Because the C261V mutant showed wild type levels of V-ATPase activity, we proceeded to further characterize this mutant. We have previously suggested that Cys 254 of the bovine V-ATPase A subunit (corresponding to Cys 261 in yeast) is the residue responsible for sensitivity of the V-ATPases to sulfhydryl reagents, such as NEM (38). Moreover, we have suggested that disulfide bond formation between Cys 254 and Cys 532 (corresponding to Cys 539 in yeast) leads to reversible inhibition of V-ATPase activity (53,54). To test these hypotheses, we have tested the sensitivity of the C261V mutant to both NEM and oxidation. As can be seen in Fig. 3, the C261V mutant is both more resistant to inhibition by NEM and more resistant to oxidation by hydrogen peroxide. In addition, this mutant is also more resistant to inhi-bition by NBD-Cl. NBD-Cl is an ATP protectable inhibitor of the V-ATPases (71) that is capable of reacting reversibly with tyrosine and cysteine residues or irreversibly with lysine residues (72). The observed resistance of the C261V mutant to inhibition by NBD-Cl suggests that Cys 261 may be the site of inhibition of the V-ATPases by this reagent.
We have also suggested that disulfide bond formation between Cys 254 and Cys 532 may play a role in regulation of V-ATPase activity in vivo (53,54). Because correct targeting and processing of the soluble vacuolar protein CpY is partially disrupted in vma mutants (65,73), a role of vacuolar acidification in intracellular targeting in yeast has been suggested. We therefore tested the effect of the C261V mutation, which should block formation of the inhibitory disulfide bond and therefore lead to a constitutively active V-ATPase, on targeting and processing of CpY. As can be seen in Fig. 4, the time course for processing of the precursor of CpY to the mature form and the fraction of immature CpY secreted into the media is indistinguishable in the wild type and C261V mutant, whereas a marked reduction in processing is observed in the deletion strain. The fact that some CpY is secreted even in wild type cells is most likely due to the lower V-ATPase activity observed in the vacuoles isolated from vma1⌬ cells containing an integrated copy of the wild type A subunit gene at the URA3 locus relative to the parental VMA1 strain. Thus disulfide bond formation in the A subunit does not appear to be required for correct targeting and processing of CpY.
We have compared three additional parameters in the wild type and the C261V mutant cells staining with the fluorescent weak base quinacrine, fluid phase endocytosis of Lucifer Yellow, and growth at neutral and acidic pH. As can be seen in Fig. 5, the intensity and pattern of staining observed with quinacrine, which labels acidic intracellular compartments, was markedly reduced in the deletion strain but indistinguishable in the wild type and C261V mutants. Similarly, although the deletion strain showed a marked reduction in fluid phase uptake of Lucifer Yellow, the C261V mutant was identical in uptake to the wild type (data not shown). Finally, the growth characteristics of the wild type and mutant cells were identical at both pH 7.5 and 5.5, in contrast to the deletion strain which showed near normal growth at pH 5.5 but greatly reduced growth at pH 7.5 (Fig. 6). Thus, the inability of the V-ATPase A subunit to form an inhibitory, intramolecular disulfide bond between Cys 261 and Cys 539 FIG. 1. Immunoblot analysis yeast A subunit levels in cell lysates (A) and isolated vacuoles (B). Yeast cell lysates (40 g of protein) or isolated vaculoes (5 g of protein) were prepared as described under "Experimental Procedures" and then subjected to SDS-PAGE on a 12% acrylamide gel followed by transfer to nitrocellulose and Western blotting using the monoclonal antibody 8B1-F3 directed against the yeast A subunit and horseradish peroxidase-conjugated secondary antibody.

FIG. 2. SDS-PAGE profile of purified V-ATPase from wild type and mutant A subunits.
Yeast vacuoles were prepared and washed and the V-ATPase isolated by detergent solubilization and density gradient sedimentation as described under "Experimental Procedures." 50-l aliquots of the fractions that typically contain the peak of V-ATPase activity (fraction 8 -10) were then applied to a 12% acrylamide gel followed by SDS-PAGE and silver staining as described in the text.
has no obvious consequences on vacuolar acidification, membrane traffic, or cell physiology in yeast.
As noted above, the E286Q mutant shows normal levels of A subunit and normal assembly, but leads to a complete loss of V-ATPase activity. To test whether the observed loss of activity in this mutation resulted from an inhibition of nucleotide binding, labeling of purified V-ATPase complexes containing the wild type and E286Q mutant A subunits with 2-azido-[ 32 P]ATP was carried out in the presence and absence of ATP. As can be seen in Fig. 7, the E286Q mutant showed reduced but still detectable labeling by 2-azido-[ 32 P]ATP, which remained ATP protectable. Quantitation revealed a 40% reduction in labeling of the E286Q mutant relative to the wild type. This result suggests that the primary defect in this mutant is reduced ATP hydrolysis rather than reduced nucleotide binding. DISCUSSION Information concerning the identity and function of residues present at the catalytic nucleotide binding site of the V-ATPase has thus far been very limited. This information is particularly important since the F-ATPase nucleotide binding subunits, for which high resolution structural data is available (52), are only 20 -25% identical to the corresponding subunits of the V-ATPase. We have presented data that Cys 254 in the glycinerich loop region of the bovine A subunit is the residue responsible for the sensitivity of the V-ATPases to sulfhydryl reagents (38) and have demonstrated that Cys 254 is able to disulfide bond with Cys 532 in the C-terminal domain, leading to reversible inactivation of the V-ATPase (54). This latter finding indi-cates that these two residues, despite their large separation in the primary sequence, are within 5-6 Å of each other in the tertiary structure of the protein. The data presented in the present study support the formation of an inhibitory disulfide bond between these two cysteine residues in yeast as well as mammals since mutation of Cys 261 in the yeast A subunit (corresponding to Cys 254 in the bovine protein) to valine generates a V-ATPase complex, which is fully active but resistant to inhibition by NEM and oxidation. Interestingly, the C261V mutant is also resistant to inhibition by NBD-Cl, a reagent which inhibits the F-ATPases by reaction at an active site tyrosine residue (72). Our data suggests that NBD-Cl inhibits the V-ATPases primarily through reaction with the cysteine residue in the glycine-rich loop.
It is clear that V-ATPases are present in multiple intracel- FIG. 4. Effect of the C261V mutation on targeting and processing of CpY. Yeast cells were grown, converted to spheroplasts, and labeled for 20 min at 30°C with Tran 35 S-label as described under "Experimental Procedures." The spheroplasts were then incubated for the indicated chase times at 30°C followed by separation into intracellular (I) and extracellular (E) fractions by sedimentation. The fractions were then precipitated with trichloroacetic acid, solubilized with SDS, and urea, and the CpY was immunoprecipitated using an antibody against CpY and protein A-Sepharose. The immunoprecipitated proteins were then eluted with Laemmli sample buffer and separated by SDS-PAGE on a 8% acrylamide gel. Two forms of the precursor of CpY (p2 and p1) as well as the mature form (m) are observed, as previously reported (65).  (1), suggesting that there must be mechanisms for regulating V-ATPase activity in vivo. We have proposed a model in which disulfide bond formation between Cys 254 and Cys 532 controls activity of the vacuolar proton pump within the cell (53,54). In this model, the V-ATPase is maintained in an inactive state in less acidic compartments by disulfide bond formation, and reversibly activated by thio-disulfide exchange upon reaching an organelle requiring more acidification. This model is supported by the observation that at least 50% of the V-ATPase in native clathrin-coated vesicles exists in the disulfide bonded state (53). Moreover, this estimate is a lower limit because following isolation of coated vesicles, continued incubation in the absence of oxygen or reducing agents results in conversion of the V-ATPase to the fully reduced form, presumably as a result of an internal thio-disulfide exchange within the A subunit (54). Other organelles, such as the Golgi apparatus, may also contain an inactive, oxidized population of V-ATPases that help ensure the organelle pH remains above that of later compartments in the biosynthetic pathway such as the lysosome/vacuole. To test this model further, we have investigated targeting and processing of the soluble vacuolar protein CpY in C261V mutant cells, which lack the capacity for oxidative inactivation of the V-ATPase. Vacuolar targeting of CpY is receptor-mediated, and appears to involve sequential steps of binding of CpY in the Golgi and release in a prevacuolar compartment followed by receptor recycling (74), much like the route followed by mannose 6-phosphate receptors in mammalian cells (75). Targeting and processing of CpY is partially disrupted in vma mutants which lack a functional V-ATPase (65,73), suggesting that organelle acidification plays some role in these processes. If oxidative inactivation of V-ATPases in the Golgi apparatus is important for regulating the pH of the Golgi apparatus, we would predict that the constitutively active C261V mutant would have a lowered Golgi apparatus pH, which might result in mistargeting and/or incomplete processing of CpY. Our results, however, indicate that targeting and processing of CpY appear to be normal in the C261V mutant, suggesting either that disulfide bond formation is not being employed to suppress V-ATPase activity in the Golgi apparatus (at least in yeast), or that the CpY receptor binding and release steps do not depend directly on pH differences between the Golgi apparatus and later compartments.
As further tests of the role of disulfide bond formation in regulation of vacuolar acidification in yeast, we have compared the rate of fluid phase endocytosis, the staining pattern observed with the fluorescent dye quinacrine, and the rate of cell growth in the wild type and C261V mutant cells. Anraku and co-workers (33) have shown that endocytic uptake of Lucifer Yellow is impaired in vma mutants, suggesting that vacuolar acidification is linked to normal fluid phase endocytosis. We observe no difference between the wild type and C261V mutant in the rate of Lucifer Yellow uptake. We also find no difference in quinacrine staining, suggesting that introduction of this mutation does not cause a large proliferation of acidic compartments in the mutant cells. This might be predicted if disulfide bond formation were important in maintaining the V-ATPase in an inactive state in other intracellular compartments, such as the endoplasmic reticulum, Golgi apparatus, or endosomes. It should be noted, however, that because of the intense quinacrine staining of the central vacuole, increased staining of other, less prominent organelles within the cell might be obscured. Finally, comparison of the growth rates of the wild type and C261V mutant at pH 7.5 and 5.5 reveals no difference, indicating that the inability to form inactivating disulfide bond in the V-ATPase does not impair cell growth or division. These results suggest that either disulfide bond formation does not play a role in regulating vacuolar acidification in yeast or that the full phenotypic consequences of constitutively activating the V-ATPase in yeast have not yet been identified.
The V-ATPase A subunit contains two additional cysteine residues that are conserved in all A subunit sequences, Cys 284 and Cys 539 . Cys 284 is located in the GXGER consensus sequence that mutagenesis studies have suggested is important in the F-ATPase ␤ subunit (55,76), while Cys 539 corresponds to Cys 532 of the bovine A subunit that participates in disulfide bond formation. Previous mutagenesis studies have suggested that mutation of these cysteine residues to serine leads to loss of ATPase activity in the intact vacuole (77), but provided no information concerning the molecular basis for the observed inactivation. The current studies suggest that mutation of Cys 284 and Cys 539 to any of a variety of residues results in destabilization of the A subunit such that no detectable protein is present in whole cell homogenate, although the possibility that these mutations lead to a loss of antibody binding to the A subunit cannot be eliminated.
In addition to the conserved cysteine residues, three additional sites in the A subunit have been investigated. Mutation of the lysine in the glycine-rich loop sequence did not block assembly of the V-ATPase complex as assessed by the presence of the A subunit on the vacuolar membrane, but did destabilize the complex since very little intact V-ATPase was observed following detergent solubilization and isolation by density gradient sedimentation. Mutation of Gly 250 at the beginning of the glycine-rich loop led to both significantly lower levels of A subunit and to the complete absence of V-ATPase assembly. These results suggest that an intact P-loop region may be important for assembly and/or stability of the V-ATPase complex. These results are in contrast to those obtained upon mutation of the corresponding region of the F-ATPase ␤ subunit, where a direct effect on catalysis with little if any effect on assembly has most frequently been observed (55)(56)(57).
The final residue targeted for mutation (Glu 286 ) corresponds to an acidic residue in the F-ATPase ␤ subunit that has been implicated to play an important role in catalysis from both chemical modification (49) and site-directed mutagenesis (55,76). This residue has been proposed to participate in ATP hydrolysis by abstracting a proton from the water molecule involved in nucleophilic attack on the terminal phosphate (52). Mutation of this residue to glutamine in the V-ATPase A subunit leads to complete loss of activity with no discernable effect on stability of the A subunit or assembly of the V-ATPase complex. Moreover, only a 40% decrease in labeling by 2-azido-[ 32 P]ATP was observed in the E286Q mutant, suggesting that the effect of this mutation on nucleotide binding was minimal. These results suggest that, as with the F-ATPases, this residue plays an important role in ATP hydrolysis by the V-ATPases. Random mutagenesis studies of the A subunit have revealed that even substitution of an aspartic acid at position 286 led to a complete loss of activity despite proper complex assembly (78), supporting the critical nature of this residue.
It is interesting to note that unlike the bovine V-ATPase, where both the A and B subunits are labeled by 2-azido-[ 32 P]ATP, only the A subunit is labeled in yeast (Fig. 7). Because of the preference of 2-azido-[ 32 P]ATP for modification of tyrosine residues (79), this result suggests that the yeast B subunit lacks a tyrosine residue at the nucleotide binding site which is present in the bovine protein. Comparison of the bovine and yeast B subunit sequences over the region which is labeled by 2-azido-[ 32 P]ATP (starting at Gly 375 ) (39) reveals three tyrosine residues that are present in the bovine protein and absent in the yeast protein. These are (using the bovine numbers): Tyr 378 , Tyr 451 , and Tyr 474 . The residues corresponding to Tyr 451 and Tyr 474 in the F-ATPase ␣ subunit are far from the nucleotide binding sites on either ␣ or ␤ (52), whereas the residue corresponding to Tyr 378 is in close proximity to ATP bound at the catalytic site of F 1 . These results suggest that Tyr 378 may participate in nucleotide binding by the V-ATPase. Additional studies will be required to further elucidate the identity and role of residues at the nucleotide binding sites of the V-ATPase family of proton pumps.