Mutations in the CYS4 Gene Provide Evidence for Regulation of the Yeast Vacuolar H+-ATPase by Oxidation and Reduction in Vivo *

The vma41-1 mutant was identified in a genetic screen designed to identify novel genes required for vacuolar H+-ATPase activity in Saccharomyces cerevisiae. The VMA41 gene was cloned and shown to be allelic to theCYS4 gene. The CYS4 gene encodes the first enzyme in cysteine biosynthesis, and in addition to cysteine auxotrophy, cys4 mutants have much lower levels of intracellular glutathione than wild-type cells. cys4mutants display the pH-dependent growth phenotypes characteristic of vma mutants and are unable to accumulate quinacrine in the vacuole, indicating loss of vacuolar acidificationin vivo. The vacuolar proton-translocating ATPases (V-ATPase) is synthesized at normal levels and assembled at the vacuolar membrane in cys4 mutants, but its specific activity is reduced (47% of wild type) and the activity is unstable. Addition of reduced glutathione to the growth medium complements the pH-dependent growth phenotype, partially restores vacuolar acidification, and restores wild type levels of ATPase activity. TheCYS4 gene was deleted in a strain in which the catalytic site cysteine residue implicated in oxidative inhibition of the yeast V-ATPase has been mutagenized (Liu, Q., Leng, X.-H., Newman, P., Vasilyeva, E., Kane, P. M., and Forgac, M. (1997) J. Biol. Chem. 272, 11750–11756). This catalytic site point mutation suppresses the effects of the cys4 mutation. The data indicate that the acidification defect of cys4 mutants arises from inactivation of the vacuolar ATPase in the less reducing cytosol resulting from loss of Cys4p activity and provide the first evidence for the modulation of V-ATPase activity by the redox state of the environment in vivo.

Many organelles of the vacuolar network of eukaryotic cells, including the vacuoles/lysosomes, Golgi apparatus, endosomes, clathrin-coated vesicles, synaptic membrane vesicles, chromaffin granules, and other secretory vesicles, are acidified by a single class of proton pumps, the vacuolar proton-translocating ATPases (V-ATPases) 1 (1). Vacuolar H ϩ -ATPases are multi-subunit complexes with an overall structure and subunit composition very similar to the F 1 F 0 -ATPases of bacteria, chloroplasts, and the inner mitochondrial membrane (2). The V 1 sector of the V-ATPase, which contains the ATP-binding sites, is a cytoplasmically-oriented complex of peripheral subunits, while the V 0 sector consists of integral membrane subunits and contains the proton pore (3). The electrochemical gradient generated by the V-ATPases is crucial for processes such as protein sorting, zymogen activation, receptor-mediated endocytosis, and the transport of ions, amino acids, and other metabolites (4,5).
V-ATPases are highly conserved between fungi, plants, and animals. Thirteen different polypeptides, ranging in molecular mass from 10 to 100 kDa have been identified as subunits of the V-ATPase of the yeast Saccharomyces cerevisiae. The genes encoding all of these subunits have been cloned (Ref. 6 and references therein, see also, Refs. 7 and 8). The products of four other genes (VMA12, VMA21-23) are also required for assembly of the yeast V-ATPase even though they are not part of the final complex (9 -11).
V-ATPases are present in several distinct locations within a single cell. Very little is known about how enzyme activity is regulated in vivo to maintain different organelles within a single cell at different specific pH values or to adjust organelle acidification in response to changing extracellular conditions. Several mechanisms have been proposed for the regulation of vacuolar acidification by V-ATPases (see Ref. 12 for a recent review). Reversible dissociation of the peripheral V 1 and integral V 0 domains in response to changes in growth conditions (13,14) has indicated that disassembly and reassembly may be a means of regulating V-ATPase activity in vivo. Regulation of acid secretion by changes in the density of V-ATPase in the apical membrane has been demonstrated in intercalated cells in the kidney (15). Reversible disulfide bond interchange (16 -18), changes in the degree of coupling between ATP hydrolysis and proton pumping (19,20), and changes in membrane potential (21,22) have also been suggested as possible means of regulating V-ATPase activity.
Biochemical studies on the enzyme isolated from bovine clathrin-coated vesicles have indicated that reversible sulfhydryl-disulfide bond interconversion within the catalytic subunit may play a role in controlling V-ATPase activity in vivo (16 -18). Specifically, these studies show that disulfide bond formation between conserved cysteine residues near the nucleotide-binding site of the catalytic subunit results in inactivation of the V-ATPase and that this inactivation can be reversed by a disulfide interchange within the catalytic subunit. Furthermore, Dschida and Bowman (23) showed that reducing agents have a stabilizing effect on the V-ATPase from Neuro-spora crassa and oxidizing agents are potent inhibitors of the V-ATPase in vitro. These results suggest that the redox state of the immediate environment may be an important regulator of vacuolar ATPase activity, but this has not been demonstrated in vivo.
In a genetic screen designed to identify novel genes affecting V-ATPase activity, we isolated a mutation in the CYS4 gene. We report here that the product of the CYS4 gene is required for the in vivo activity, but not the biosynthesis and assembly of the yeast vacuolar H ϩ -ATPase, when cells are grown in rich medium. Mutations in CYS4 lead to a decrease in the cellular concentration of reduced glutathione as a result of impaired cysteine biosynthesis. Our results provide the first in vivo evidence in support of previous results (17,23) which suggest that the V-ATPase may be inactivated in a less reducing environment and that this inactivation involves a highly conserved catalytic site cysteine.

EXPERIMENTAL PROCEDURES
Materials-Restriction endonucleases were purchased from New England Biolabs and Boehringer Mannheim. Taq DNA polymerase was purchased from Boehringer Mannheim. Zymolyase 100T was purchased from ICN. 35 S-dATP was purchased from NEN Life Science Products. 1-kb DNA ladder and prestained protein molecular mass standards were obtained from Life Technologies, Inc. Synthetic oligonucleotide primers for polymerase chain reaction and sequencing were obtained from Genosys. Zwittergent 3-14 (ZW3-14) was obtained from Calbiochem. All other reagents were purchased from Sigma.
Strains, Media, and Microbiological Techniques-Yeast strains used in this study and their genotypes are listed in Table I. Yeast cells were grown aerobically in media prepared as described by Sherman et al. (24) or Yamashiro et al. (25), except that 50 mM MES and 50 mM MOPS were used to buffer YEPD, pH 7.5, containing 50 mM calcium chloride. Sporulation medium (SPIII-22) was prepared as described (26) except that p-aminobenzoic acid was omitted from the supplement mixture. cys4 mutant cells were grown on supplemented minimal medium (SD) or YEPD media to which 30 g/ml reduced glutathione (GSH) was added to give 0.1 mM GSH final. For glutathione depletion, cells were grown in YEPD, pH 7.5, containing 50 M 1-chloro-2,4-dinitrobenzene.
Cloning of the VMA41 Gene-Yeast strain YOY14-4Ba was transformed with a yeast genomic library as described (27). The yeast genomic library, constructed by cloning a yeast partial Sau3A genomic DNA into the BamHI site of the CEN plasmid, YCp50, was a kind gift from Dr. Saul Honigberg at Syracuse University. Transformants carrying a URA3-containing plasmid capable of complementing the Vma Ϫ growth phenotype of YOY14 were selected directly on SD-ura (supplemented minimal medium lacking uracil) plates buffered to pH 7.5. Plasmids were isolated from transformants as described (28) and retransformed into YOY14-4Ba to confirm the phenotype. Various fragments ( Fig. 1) of the complementing DNA were subcloned into pRS316 (29) and tested for complementation. All DNA manipulations were done as described by Sambrook et al. (30).
Disruption of CYS4 Gene-A null cys4 strain was constructed by the one-step allele replacement method (31). The 870-base pair AgeI fragment within the CYS4 open reading frame in plasmid pYO49 was replaced by a 2.1-kb HpaI fragment containing the LEU2 gene. The resulting plasmid (pYO50) was digested with ApaI and SacII to release the LEU2-disrupted allele from the vector and the linear DNA fragment generated was used to transform yeast strain SF838-5Aa. Leu ϩ transformants were selected and disruption of the CYS4 locus was confirmed by polymerase chain reaction from chromosomal DNA using synthetic oligonucleotides 5Ј-GGTAGAATTCATCCTTCCAG-3Ј and 5Ј-GATAA-CATCAGTGACCTTAGC-3Ј. Isolation of yeast genomic DNA for polymerase chain reaction analysis was carried out as described by Nasmyth and Reed (32) except that DNA was treated with RNase A for 25 min at 37°C and 5 min at 65°C before the final precipitation.
Construction of the YOY13-2Ca Strain-The CYS4 locus was disrupted in a vma1⌬::URA3 strain (SF838-5Aa vma1⌬::URA3) as described above. The resulting vma1⌬ cys4⌬ strain was crossed to yeast strain PNY1 (33) to obtain a heterozygous diploid yeast strain (YOY13) with only one functional copy of the CYS4 gene and no wild-type allele of the VMA1 gene (the only functional copy is the C261V mutant allele, designated here as vma1-51, integrated at the URA3 locus). YOY13 was sporulated and tetrads were dissected on YEPD, pH 5.0, plates. cys4 spores, identified by their cysteine auxotrophy, were selected. Whole cell lysates were prepared from selected (cys4) spores and analyzed by Western blotting using monoclonal antibody 7D5 directed against the 69-kDa VMA1 gene product. This antibody is able to detect the product of the vma1-51 mutant allele of VMA1 (33). Spores in which the vma1-51 mutation had co-segregated with the cys4-⌬1 mutation were selected. One such spore, YOY13-2Ca, was used in the experiments described here.
Quinacrine Vital Staining-Vacuolar accumulation of quinacrine was assessed as described by Roberts et al. (34). Once stained, cells were visualized within 10 min using a Zeiss Axioskop Routine immunofluorescence microscope. Cells were viewed under Nomarski optics to observe normal cell morphology and under a fluorescein isothiocyanate filter with a 100 ϫ objective to observe vacuolar staining.
Purification of the Yeast V-ATPase-Solubilization of vesicles and purification of the vacuolar H ϩ -ATPase were performed basically as described (3,51) with the following modifications. 0.5-1 mg of solubilized vesicles were layered on a 12 ml of 20 -50% (w/v) glycerol gradient and centrifuged at 200,000 ϫ g for 8 h in a Beckman Ti-75 rotor. Sixteen 700-l fractions were collected and analyzed for ATPase activity to identify fractions containing peak ATPase activity. Fractions were diluted 1:1 with water and protein precipitated by addition of an equal volume of 20% trichloroacetic acid. Precipitated proteins were solubilized in 50 l of cracking buffer (50 mM Tris-HCl, pH 6.8, 1 mM EDTA, 8 M urea, 5% SDS, 5% ␤-mercaptoethanol), separated on a 10% SDSpolyacrylamide gel, and detected by silver staining or Western blotting. To determine the specific activity of the purified V-ATPase, protein was precipitated as described above and resuspended in cracking buffer lacking ␤-mercaptoethanol. Protein concentration was then determined using Bio-Rad DC Protein Assay kit.
DNA Sequencing-Plasmid DNA for sequencing was purified using the QIAprep-spin Plasmid Kit from QIAGEN. Sequencing was done by the dideoxy chain termination method using Sequenase R sequencing kit with Sequenase version 2.0 (U. S. Biochemical Corp.) and 35 S-dATP.
Other Methods-Indirect immunofluorescence microscopy and preparation of vacuolar membrane vesicles were performed as described by Roberts et al. (34). Vacuolar ATPase activity was measured in a coupled enzyme assay as described previously (37). The effects of various oxidizing and reducing agents on V-ATPase activity were tested by preincubating 5-g vesicles with the indicated agent in 100 l of buffer (15 mM MES-Tris, pH 7.0, 4.8% glycerol). At the end of the incubation period, the whole reaction mixture was assayed for ATPase activity as described (37). Equivalent concentrations of the oxidizing and reducing agents were shown not to inhibit the enzymes of the coupled ATPase assay system. Purification of the V-ATPase was carried out as described. 2 Protein was determined by the method of Lowry (38) and silver staining was performed as described (39).

RESULTS
Identification of the VMA41 Gene as CYS4 -The MEY14 strain, carrying the vma41-1 mutation was obtained in a genetic screen designed to identify genes required for vacuolar membrane ATPase activity in yeast. 2 The screen was based on the set of growth phenotypes characteristic of mutants with loss of vacuolar membrane V-ATPase activity (vma mutants). The Vma Ϫ growth phenotypes include inability to grow in medium buffered to pH 7 or above, medium containing 100 mM CaCl 2 , or medium containing a non-fermentable carbon source. After backcrossing to the original MEY14 strain to remove background mutations, one Vma Ϫ spore, YOY14-4Ba, was selected for further analysis. In addition to Vma Ϫ phenotypes, both the MEY14 mutant and the YOY14-4Ba strain are unable to grow on minimal medium, suggesting the cosegregation of the Vma Ϫ phenotype and an undetermined nutritional auxotrophy. The VMA41 gene was cloned by complementation of the pH-dependent growth phenotype of the YOY14-4Ba strain. The YOY14-4Ba mutant strain was transformed with a yeast DNA library on a single copy (CEN) plasmid (YCp50). Transformants were selected on SD-ura, pH 7.5, medium and 15 independent Ura ϩ Vma ϩ transformants were obtained. Plasmids were recovered from transformants and re-checked for complementation. All 15 plasmids were able to restore growth at pH 7.5 to the YOY14-4Ba strain. Restriction endonuclease analyses revealed that all 15 contain the same 11-kb yeast DNA insert ( Fig. 1). Various subclones of plasmid pMEY14-1 were generated in the yeast shuttle vector pRS316. Analyses of these sublones indicated that a 2-kb XbaI-SphI fragment is sufficient for complementation ( Fig. 1). A 350-base pair region internal to this fragment was sequenced and used to search for homology to any sequences in the GenEMBL data base. This analysis revealed that the sequenced region (indicated by the arrows in Fig. 1) lies within the reported nucleotide sequence of the yeast NHS5 gene for ␤-thionase, also known as the STR4 or CYS4 gene for cystathionine-␤-synthase (40,41). For clarity, the CYS4 nomenclature is used throughout this report. The CYS4 gene product is essential for cysteine metabolism; as a result, cys4 mutants exhibit a cysteine-dependent growth phenotype (40). A LEU2-disrupted copy (plasmid pYO50) of the CYS4 open reading frame was constructed as shown in Fig. 1. Haploid yeast strain SF838-5A␣ was transformed with ApaI/SacIIdigested pYO50 and stable Leu ϩ transformants selected. DNA was extracted from three independent transformants and the parental wild type strain and analyzed by polymerase chain reaction to confirm the disruption of the CYS4 locus in the transformants. The results (not shown) indicate that the 968base pair fragment, expected from the wild type CYS4 locus, was replaced by a larger, 2.3-kb fragment in the transformants, indicating that the CYS4 gene has been disrupted in these cells. The growth phenotypes of cys4⌬ cells are indistinguishable from those of the original YOY14-4Ba mutant. In addition to the characteristic Vma Ϫ growth phenotypes, cys4⌬ cells are unable to grow on minimal medium without externally sup-plied cysteine. These results suggest that a single mutation is responsible for both the Vma Ϫ phenotypes and the cysteine auxotrophy of the mutant cells.
To determine if the Vma Ϫ growth phenotype of the vma41 strain is directly related to its cysteine auxotrophy, YOY14-4Ba and cys4⌬ cells were supplied with cysteine or reduced glutathione. Under these conditions, the mutant cells were able to grow on minimal medium. Interestingly, externally supplied cysteine was able to partially complement the growth phenotype of YOY14 strain on YEPD medium buffered to pH 7.5 ( Fig.  2), with glutathione giving better complementation than cysteine. Conversely, depletion of intracellular glutathione in wild type cells by addition of 50 M 1-chloro-2,4-dinitrobenzene (57) prevented growth of the cells on YEPD medium buffered to pH 7.5 (data not shown) under conditions where growth of cells on YEPD buffered to pH 5.0 continued. These results indicate that glutathione deficiency results in a pH-dependent (Vma Ϫ ) growth phenotype in yeast.
To confirm that vma41-1 is indeed an allele of CYS4, diploid strain YOY12/pYO38, obtained from a cross between cys4-⌬1 and YOY14-4Ba/pYO38 was sporulated and the resulting tetrads dissected. We found that YOY12 cured of the plasmid pYO38 was unable to sporulate, even after 2 weeks in sporulation medium. This is consistent with previous results indicating that glutathione auxotrophic mutants of Schizosaccharomyces pombe are defective in sporulation (42). Tetrad analysis indicated a 4:0 segregation of the Vma Ϫ phenotype in YOY12 spores since all Ura Ϫ spores were Vma Ϫ and all Ura ϩ spores were also Vma ϩ . Moreover, when Ura ϩ spores lost the plasmid pYO38, they became Vma Ϫ . Also, YOY12 diploids lacking plasmid pYO38 exhibit a pH-dependent growth phenotype. These results confirm that vma41-1 is allelic to CYS4.
Characterization of the V-ATPase from cys4 Mutants-All known vma mutants are unable to accumulate the fluorescent weak base, quinacrine, in their vacuoles as a result of loss of vacuolar acidification (25). Quinacrine vital staining was used to assess vacuolar acidification in YOY14-4Ba (vma41-1) and cys4⌬ mutant cells. Our results (Fig. 3) show that mutants lacking a functional CYS4 gene are unable to accumulate quinacrine in their vacuoles, indicating loss of vacuolar acidification in these mutants. Since GSH was able to complement the pH-dependent growth phenotype of cys4 mutants, it may also be able to restore quinacrine accumulation in the vacuole. Fig.  3 shows partial restoration of vacuolar acidification by GSH as indicated by partial vacuolar staining with quinacrine in the presence of GSH.
To begin to understand the basis of the Vma Ϫ growth and vacuolar acidification defects of cys4 mutants, we examined the steady-state levels of several V-ATPase subunits and one specific assembly factor in these cells. Whole cell protein extracts prepared from YOY14-4Ba and cys4⌬ cells were analyzed by Western blotting. The results show that the 69-, 60-, 54-, 42-, and 27-kDa peripheral subunits of the V-ATPase and the 25-kDa (Vma12p) V-ATPase assembly factor are present in cys4 mutant cells at normal levels compared with wild type cells (Fig. 4A). Addition of GSH to the growth medium does not increase the steady state levels of the V-ATPase subunits in cys4-⌬1 cells (Fig. 4A). These results indicate that, even in the absence of additional extracellular glutathione, there is sufficient cysteine in the mutants during growth in rich medium to support normal levels of subunit biosynthesis. Previous results (43) have shown that the presence of V-ATPase subunits at normal levels does not always imply assembly of the enzyme. Therefore, we examined the assembly status of the V-ATPase in cys4 mutants. Indirect immunofluorescence microscopy, using antibody against the 60-kDa peripheral subunit of the V-ATPase (13D11) was performed as described (34). The 13D11 antibody recognizes the 60-kDa peripheral subunit by itself, as part of a V 1 subcomplex, or in the fully assembled V 1 V 0 -ATPase. A ring-like staining corresponding to the vacuolar membrane was detected in the cys4 mutants (data not shown). These results indicate that the V-ATPase is assembled in cys4 mutants.
To address whether the Vma Ϫ mutant phenotypes of cys4 mutants are a direct result of the loss (or impairment) of vacuolar ATPase activity, vacuolar membrane vesicles were isolated from wild type and cys4-⌬1 cells and assayed for concanamycin A-sensitive ATPase activity. Concanamycin A is a specific inhibitor of V-ATPases (44). Table II shows that the specific V-ATPase activity of vacuolar membrane vesicles isolated from cys4-⌬1 cells is less than 50% of that of vesicles isolated from the isogenic wild type strain. However, if cys4-⌬1 cells were supplemented with 30 g/ml GSH during growth, the V-ATPase activity on isolated vesicles is restored to nearwild type levels (Table II). Glutathione supplementation does not have any effect on the V-ATPase activity of vesicles from wild type cells (results not shown). Western analysis of proteins in isolated vacuolar membranes confirm that the reduced level of V-ATPase activity in cys4-⌬1 cells is not a result of reduced amounts of V-ATPase subunits on these membranes. As shown in Fig. 4B, all the subunits monitored are present at wild type levels in the vacuolar membranes of cys4⌬ cells. Addition of GSH to the growth medium does not lead to increased levels of subunits on the membrane. These observations strongly suggest that the cys4 mutation does not affect the assembly or targeting of the vacuolar H ϩ -ATPase, but has a direct effect on the catalytic activity of the enzyme.
Mutation of Cysteine 261 in the Catalytic Subunit Suppresses Effects of the cys4-⌬1 Mutation-In vitro studies have shown that the clathrin-coated vesicle V-ATPase can be regulated by oxidation and reduction (17). In particular, it was demonstrated that reducing agents activate the enzyme in vitro by promoting disulfide bond interchange between cysteine residues located near the catalytic site (17). Glutathione is the most abundant low molecular weight thiol in the cell, constituting Ͼ95% of cellular thiols (45). Most of the glutathione in yeast cells is in the reduced (GSH) form; the thiol:disulfide (i.e. GSH:GSSG) balance is between 50 and 60, similar to other eukaryotic cells (46,47). Since cys4 mutants have been shown to have reduced cellular glutathione concentration (46), and our results (above) indicate that the V-ATPase is expressed at levels similar to wild type in cys4-⌬1 cells, we reasoned that the vma phenotypes of cys4 mutants may result from accumulation of the V-ATPase in an oxidized, inactive form in the less reducing environment resulting from low cellular GSH levels. To test this hypothesis, we used a yeast mutant in which the sulfhydryl-reactive cysteine residue has been mutated. There are three perfectly conserved cysteine residues in the catalytic (A) subunit of all V-ATPases (48). In yeast, these residues are Cys-261, Cys-284, and Cys-538 (numbers without intein); and in bovine they are at positions 254, 277, and 532, respectively. Cys-254 in the bovine enzyme has been identified as the sulfhydryl-reactive residue responsible for inhibition by NEM, and, together with Cys-532 has been implicated in the formation of the inactivating disulfide bond within the A subunit (18). Following this model, mutations that prevent the formation of the proposed disulfide bond are expected to abolish the sensitivity of the V-ATPase to the redox state of the environment. Taiz et al. (48) and Liu et al. (33) have demonstrated that mutation of the cysteine residue at position 261 in yeast to valine abolished the inactivation of the V-ATPase by sulfhydryl reagents and the reversible inactivation in response to oxidizing agents. We constructed a haploid yeast strain (YOY13-2Ca) in which the CYS4 gene has been disrupted and the sole copy of the VMA1 gene (vma1-51) has the critical Cys-261 residue changed to valine by site-directed mutagenesis (Ref. 33, see "Experimental Procedures"). The VMA1 gene product is stably expressed in both the PNY1 strains carrying the C261V mutation in the VMA1 gene and the YOY13-2Ca strain carrying both the C261V mutation and a deletion in the CYS4 gene and the two strains exhibit similar growth on YEPD medium buffered to pH 7.5 (33). 3 Isolated vacuolar membrane vesicles from YOY13-2Ca and the isogenic PNY1 strains were assayed for V-ATPase activity. Both PNY1 and YOY13-2Ca strains have the same level of V-ATPase activity (Table III) as cells carrying a wild type VMA1 allele. Moreover, in contrast to the cys4-⌬1 mutant, the V-ATPase activity of YOY13-2Ca is not increased by the addition of reduced glutathione to the growth medium, indicating that the C261V mutant enzyme, which is no longer sensitive to inhibition by oxidizing agents (33,48), is no longer influenced by the cys4-⌬1 mutation. Western analysis of vacuolar membrane vesicle proteins shows that a number of other V-ATPase subunits are present at normal levels in the PNY1 and YOY13-2Ca strains (results not shown). Our results suggest that the cys4 mutation impairs activity of the yeast vacuolar ATPase by altering the thiol/redox balance in the cell, and that the altered redox balance is sensed (at least in part) by the catalytic subunit of the V-ATPase in vivo.
Biochemical Properties of the V-ATPase in Vacuoles from cys4⌬ Mutants-Analysis of point mutations causing partial defects in V-ATPase activity (49) have indicated that the full pH-dependent growth phenotype requires the loss of Ͼ75% of V-ATPase activity on the vacuole. However, cys4 mutants exhibit full Vma Ϫ phenotypes but have approximately 47% wild 3 Y. E. Oluwatosin and P. M. Kane, unpublished results.

FIG. 4. Immunoblots of V-ATPase subunits in whole cell protein extracts and vacuolar membrane vesicles.
Whole cell extracts and solubilized vacuolar vesicles were prepared as described (35).

TABLE II Vacuolar ATPase activities of vacuolar membrane vesicles isolated
from wild type and cys4-⌬1 cells Vacuolar membrane vesicles were prepared as described under "Experimental Procedures" and the ATPase activity that is sensitive to 100 nM concanamycin A determined at 37°C. Activity is represented as mean Ϯ S.D. (n samples). All three strains were grown in YEPD buffered to pH 5.0 prior to vacuole isolation; ϩGSH samples had an additional 30 g/ml reduced glutathione added to the growth medium.  III The V-ATPase of YOY13-2Ca cells is insensitive to modulation by GSH Vacuolar membrane vesicles were prepared as described under "Experimental Procedures" and the ATPase activity that is sensitive to 100 nM concanamycin A determined at 37°C. Activity is represented as mean Ϯ S.D. (n samples). The isogenic wild-type strain used in these experiments (NO11-2) is one in which the normal chromosomal copy of the VMA1 gene has been deleted and a wild type copy integrated at the URA3 locus (33,49). This strain gives somewhat lower ATPase activities than the strain shown in Table II (33,49). All strains were grown in YEPD buffered to pH 5.0 prior to vacuole isolation; the ϩGSH sample had an additional 30 g/ml reduced glutathione added to the growth medium. type V-ATPase activity on isolated vacuolar membranes. It is possible that the in vivo activity of the V-ATPase is much lower than that observed in vitro (on isolated vacuoles). One explanation for the difference in activity is that the enzyme could be activated by disulfide interchange stimulated by atmospheric oxygen during the process of vacuolar isolation. If this is true, preserving the redox state of the enzyme, for example, by excluding oxygen from the buffers used for vacuole preparation, might yield an enzyme with specific V-ATPase activity that is less than 47% of wild type. Feng and Forgac (17) have reported that a significant fraction of the vacuolar proton pumps exist in an inactive, disulfide-bonded form in native bovine clathrin-coated vesicles. To test this possibility, we prepared vacuolar vesicles from wild type and cys4-⌬1 cells using degassed buffers and carried out all homogenization steps under nitrogen. In addition, isolated vesicles were stored under nitrogen. This treatment did not have significant effect on the V-ATPase from S. cerevisiae wild type or cys4-⌬1 mutant cells (data not shown). This result does not prove that the yeast V-ATPase is not subject to activation during purification; further experiments and/or optimization of the conditions may be needed to demonstrate this. To further investigate the enzymatic properties of the V-ATPase in vacuolar membranes from cys4-⌬1 mutants grown in the absence of added extracellular glutathione, we examined the effects of several chemical agents that have been linked to oxidative inhibition of V-ATPase activity in vitro in other systems (16,23) on the yeast enzyme. Our results (Table IV) show that the enzymatic behavior of the residual V-ATPase activity from cys4-⌬1 vacuolar vesicles is similar to that of the wild type enzyme. (The "residual" ATPase activity represents the activity in vacuoles isolated from cys4-⌬1 mutants grown in the absence of additional glutathione.) V-ATPase activity is stable at 0°C or room temperature (23°C) for up to 9 h with no activation or inhibition observed (Table IV). The sensitivity of V-ATPase activity from cys4-⌬1 membranes to NEM and concanamycin A, two specific inhibitors of V-ATPases, is identical to that of the wild type enzyme (Table IV and data not shown). Previous evidence indicates that NEM inhibits V-ATPase activity by modification of Cys-261 of the catalytic subunit (16, 33, 48). Therefore, the inhibition of the V-ATPase in cys4-⌬1 mem-branes suggests Cys-261 is available for NEM binding in the population of enzyme responsible for the residual ATPase activity. Oxidizing agents (iodate and hydrogen peroxide) inhibit the ATPase activity of both enzymes at similar concentrations. Enzyme activity is partially restored by treatment with DTT after iodate or peroxide treatment (Table IV). DTT alone activates ATPase activity up to 30%. In contrast to previous results which show that GSH activates the N. crassa V-ATPase (23), addition of GSH at concentrations ranging from 500 M to 30 mM did not activate the yeast V-ATPase in vitro (Table III and data not shown). In fact, our in vitro experiments show that GSH reproducibly inactivates the yeast V-ATPase in a concentration-dependent manner. Notably, treatment of isolated vesicles with GSH did not restore wild type V-ATPase activity to the enzyme isolated from cys4-⌬1 mutant (Table IV), indicating that intracellular GSH (manipulated by extracellular GSH concentrations) may not act directly on the V-ATPase to restore activity in the cys4 mutants.
To determine if the cys4-⌬1 mutation causes any structural defects in the V-ATPase, we purified the V-ATPase from vacuolar membranes isolated from wild type and cys4-⌬1 cells by glycerol gradient centrifugation as described (50,51). Our results suggest that the V-ATPase from cys4 mutant cells may be structurally different from the V-ATPase from wild type cells. First, in cys4-⌬1 vacuolar membranes, 39% of the specific V-ATPase activity is lost upon detergent solubilization, a treatment that slightly increases the specific activity of the enzyme from wild type membranes. Second, the residual V-ATPase activity of solubilized cys4-⌬1 vacuolar membranes is completely lost after glycerol gradient centrifugation. No activity is detected in any of the fractions collected from cys4-⌬1 membranes (data not shown) but solubilized vacuolar membranes from wild type cells give a sharp peak of V-ATPase activity centered around fraction 9. These data indicate that even though vacuolar vesicles from cys4-⌬1 cells contain a considerable V-ATPase activity, this activity is significantly less stable than that of wild type vesicles. To check whether the observed loss of V-ATPase activity upon glycerol gradient centrifugation is a result of dissociation of the V-ATPase into partial complexes or individual subunits, proteins in fractions obtained from glycerol gradient were precipitated, separated by SDS-PAGE, and analyzed by silver staining. Fig. 5 shows that the bulk of the assembled V-ATPase in solubilized vacuolar membranes from wild type (Fig. 5A) and cys4-⌬1 (Fig. 5B) cells fractionates to similar portion of the gradient. Previous results (58) have shown that in yeast mutants lacking a functional copy of the VMA13 gene product, the V-ATPase is assembled but inactive. Fraction 9, which contains the peak of V-ATPase activity in wild type membranes was analyzed by Western blotting, using polyclonal antisera against the 54-kDa Vma13p and monoclonal antibody against the 69-kDa Vma1p. Fig. 5C shows that Vma13p and Vma1p are present at wild type levels in the V-ATPase purified from cys4-⌬1 cells. These results indicate that the loss of activity in the V-ATPase purified from cys4-⌬1 cells is not due to loss of the 54-kDa Vma13p. Western blots of the 69-, 60-, 42-, and 27-kDa V 1 subunits and the 100-kDa V 0 subunit (data not shown) confirmed that fraction 9, which contains the peak of V-ATPase activity in wild type membranes, also represents a single peak of V 1 subunits. V 0 subunits peak at both fractions 9 and 12-13, consistent with previous results (35). The solubilized cys4-⌬1 membranes show a very similar pattern, indicating that the V-ATPase subunits are present at comparable levels in the purified V-ATPase from cys4-⌬1 and wild type cells, and there do not appear to be multiple populations of partially assembled enzyme in the cys4-⌬1 vacuolar membranes.

TABLE IV
Properties of the V-ATPase of vacuolar membrane vesicles isolated from wild type and cys4-⌬1 cells Preparation of vacuolar membrane vesicles and determination of concanamycin A-sensitive ATPase activity are as described in the legend to Table II. Initial incubations were performed with 5 g of vesicle protein in a total of 100 l of incubation buffer (15 mM MES-Tris, pH 7.0, 4.8% glycerol) and incubation mixtures were added to a 1 ml of coupled enzyme assay mixture. The specific activities of untreated vesicles were 3.4 and 1.6 mol of ATP/min/mg of protein for wild type and cys4-⌬1, respectively. For NEM, then DTT samples, vacuoles were treated with 100 M NEM for 10 min followed by 100 mM DTT for 30 min. For H 2 O 2 , then DTT samples, vacuoles were treated with 10 mM H 2 O 2 for 10 min followed by 100 mM DTT for 30 min. For NaIO 3 , then DTT samples, vacuoles were treated with 10 mM NaIO 3 for 5 min followed by 100 mM DTT for 30 min. Finally, the V-ATPase was purified from the YOY13-2Ca strain. Western analysis and silver staining of SDS-PAGE gels indicate that the polypeptide composition of the V-ATPase purified from YOY13-2Ca vacuolar membranes is indistinguishable from that of the enzyme purified from an isogenic wild type strain (data not shown). However, the specific ATPase activity of the V-ATPase purified from YOY13-2Ca cells was 5.2 mol/min/mg compared with 8.0 mol/min/mg for the enzyme from the isogenic wild type strain. DISCUSSION In a genetic screen designed to identify novel genes affecting the function of the yeast vacuolar H ϩ -ATPase, we obtained a yeast strain YOY14-4Ba containing the vma41-1 mutation. Complementation analysis revealed that VMA41 is allelic to CYS4. Our results indicate that although the CYS4 gene product is not essential for assembly of the V-ATPase, it is required for full enzyme activity. These data provide the first in vivo evidence for the regulation of vacuolar H ϩ -ATPase activity by the redox state of its environment.
Effect of Loss of Cys4p Activity on Cellular Thiol Balance-The yeast CYS4 gene was identified as the gene responsible for cystathionine-␤-synthase activity (40,41,52). The human homolog of CYS4 (CBS) is able to restore cysteine-independent growth to a yeast strain in which the CYS4 gene has been disrupted (39), indicating that Cys4p activity is conserved in evolution. Cys4p is involved in cysteine biosynthesis through the transulfuration pathway in S. cerevisiae where it catalyzes the first of two reactions in the biosynthesis of cysteine from homocysteine (40,53). Loss of Cys4p activity confers a cysteine dependence which cannot be relieved by supplementation with other organic sulfur compounds, such as methionine (40), but is effectively relieved by glutathione (41). The observation that reduced glutathione as an exogenous source of sulfur is better than cysteine has been attributed to the poor uptake (40) and toxicity (45) of cysteine. In addition to protein synthesis, cysteine is required for glutathione synthesis; therefore, because cys4 mutants are unable to synthesize cysteine, they will be deficient in glutathione synthesis. In agreement with this, cys4 mutants have been shown to have greatly reduced cellular GSH (13.6% of wild type; Ref. 46). Since GSH represents more than 95% of the non-protein thiol pool in wild type yeast cells (47), GSH-deficient strains should have a much less reducing cytosol than wild type cells.
Glutathione is the most abundant low molecular weight peptide and the most prevalent cellular thiol (45). GSH functions in metabolism, transport, cellular protection, and catalysis, particularly in the reduction of the disulfide linkages of proteins and other molecules, and plays an essential role in cell growth (42,45). The ability of yeast cys4 mutants to grow in YEPD (which contains 1% yeast extract) buffered at pH 5.0 is consistent with the fact that the yeast extract in the medium contains glutathione (42). Our results showing that cys4-⌬1 cells grown in YEPD, pH 5.0, have reduced levels of V-ATPase activity on isolated vacuolar membranes despite the presence of some glutathione in the medium can be explained by reports on S. pombe, indicating that glutathione synthesis is more efficient than its uptake (42). In these studies, despite the presence of ϳ0.25 mM glutathione in the medium, the GSH level in gcs1⌬ cells (equivalent of S. cerevisiae gsh1⌬ mutants, which lack one enzyme involved in glutathione biosynthesis; Ref. 42) was only 20% that of wild type cells. It is possible that uptake from the medium provides enough cysteine to allow synthesis of normal levels of V-ATPase subunits, but does not provide enough GSH to fully overcome the inhibition of the V-ATPase in cys4-⌬1 cells. Further addition of GSH to the medium may stimulate uptake by a concentration effect to the point where the intracellular GSH level is enough for partial activation of the V-ATPase. Chaudhuri et al. (42) have reported that supplementation of Gsh Ϫ mutants of S. pombe with reduced glutathione (up to 0.25 mM) restores a Gsh ϩ growth phenotype but not pigmentation in adenine-limiting medium. The emerging theme from studies of GSH metabolism is that there is a minimum critical level of GSH required for growth and protein synthesis which is lower than that required for the other functions of GSH in metabolism, detoxification, and transport (47).
Effects of Loss of Cys4p Activity on V-ATPase Activity-We show here that cys4 mutants exhibit growth phenotypes characteristic of the loss of V-ATPase activity in vivo. In addition to the cysteine-dependent growth phenotype characteristic of cys4 mutants (40,41), the vma41-1 and cys4-⌬1 mutants display Vma Ϫ growth phenotypes ( Fig. 2 and data not shown) and are unable to accumulate quinacrine in their vacuoles (Fig. 3), indicating loss of vacuolar acidification. Isolated vacuolar ves- FIG. 5. Glycerol gradient purification of the V-ATPase. The ATPase was solubilized and purified from (A) wild type and (B) cys4-⌬1 cells. The V-ATPase was purified from vacuolar membrane vesicles as described under "Experimental Procedures." Proteins in fractions obtained from the glycerol gradient were precipitated by adding an equal volume of 20% of trichloroacetic acid and incubating on ice for 30 min. Precipitated proteins were resuspended in cracking buffer (50 mM Tris-HCl, pH 6.8, 8 M urea, 1 mM EDTA, 5% SDS, and 5% ␤-mercaptoethanol) and separated on a 10% SDS-polyacrylamide gel. Protein bands were visualized by silver staining. Positions of known V-ATPase subunits are indicated on the left. Asterisk indicates the position of an unidentified protein that consistently copurifies with V-ATPase activity. Bands from wild type vacuoles at ϳ28, 30, and 45 kDa did not consistently cofractionate with ATPase activity. The positions of protein molecular weight standards are indicated on the right. Vacuolar ATPase activity peaked in gradient fraction 9 of the gradient from wild type vacuoles (A). No ATPase activity was detected in any fraction of the gradient from cys4-⌬1 vacuoles (B). C, detection of the 69-kDa Vma1p and 54-kDa Vma13p in gradient fraction 9 by Western blot analysis. icles appear to contain fully assembled V-ATPase complexes, but have reduced V-ATPase activity. Significantly, the vma growth and acidification defects of the cys4 mutants can be suppressed by addition of GSH to the growth medium, and vacuoles isolated from cys4-⌬1 strain after growth in the presence of extracellular GSH have wild type levels of V-ATPase activity.
Vacuolar ATPases have been shown to be sensitive to the redox state of their environment in vitro (16 -18, 23, 33). The effects of the cys4-⌬1 mutant on the yeast V-ATPase activity can be suppressed by the previously characterized vma1-51 mutation (33,48), in which the highly conserved cysteine 261, which is predicted to be at the catalytic site of the yeast 69-kDa subunit (Vma1p), has been replaced by valine (Table III). The vma1-51 mutation has previously been demonstrated to render the yeast V-ATPase insensitive to oxidative inactivation in vitro, presumably by preventing formation of an inactivating disulfide bond involving cysteine 261 of the catalytic subunit (33). This result indicates that the functional and structural defects of the V-ATPase in cys4 mutants may be derived from oxidative inactivation involving disulfide bond formation in the catalytic subunit of the enzyme in vivo as a result of reduced levels of cytosolic GSH. Although enzymatic characterization of the V-ATPase from cys4-⌬1 vacuolar membranes indicated that the enzyme could not be readily reactivated by reducing agents in vitro (Table IV), this result may not be surprising, because even after a brief oxidative inactivation, dependent on cysteine 261, of the yeast V-ATPase in vitro, ATPase activity could only be partially restored by reducing agents (33). If, as proposed, reactivation of the enzyme after oxidative inactivation requires disulfide interchange in the enzyme (18), this interchange may well be more difficult to induce than a simple sulfhydryl reduction.
There is evidence, however, that the effects of the cys4 mutations on the V-ATPases may not be mediated solely through disulfide bond formation by the catalytic site cysteine 261 in vivo. Vacuolar membranes isolated from cys4-⌬1 mutants grown in the absence of added extracellular glutathione contain a residual V-ATPase activity (47% of wild-type) that is sensitive to NEM, suggesting that cysteine 261 is not disulfidebonded in the population of enzyme responsible for this activity. This residual activity could be derived from spontaneous activation of a population of the V-ATPase during vacuolar isolation, as reported by Feng and Forgac (17) for the bovine clathrin-coated vesicle enzyme. The residual activity found in the yeast cys4-⌬1 membranes proved to be unstable to detergent solubilization and glycerol gradient purification, however, indicating that the enzyme complexes responsible for the activity are functionally defective, even though we detected no obvious structural defect in the isolated enzyme. Furthermore, previous results have shown that the onset of the full pH-dependent growth phenotype requires the loss of more than 75% of V-ATPase activity at the vacuole (49). The fact that cys4 mutants exhibit the full range of Vma Ϫ phenotypes, including pH-dependent growth and total loss of quinacrine accumulation, strongly suggests that the residual activity in cys4-⌬1 vacuoles is not functional in vivo. The fact that the specific activity of the V-ATPase purified from the YOY13-2Ca strain is less than that of the comparable wild type strain is additional evidence for an effect of the cys4 mutation not mediated through disulfide bond formation involving cysteine 261 at the catalytic site.
Taken together, these results indicate that the loss of V-ATPase activity in cys4 mutants is the result of two separate effects of the mutation on the V-ATPase. The ability of the vma1-51 mutation to suppress the V-ATPase activity defect indicates that one of these effects is oxidation of cysteine 261 in the catalytic subunit. This oxidative inactivation appears to be reversed or suppressed by addition of excess extracellular GSH to yield an enzyme capable of function in vitro, based on restoration of wild type levels of ATPase activity at the vacuole, and in vivo, based on the partial restoration of quinacrine staining shown in Fig. 3. A second population of enzyme is responsible for the 47% residual activity of isolated cys4-⌬1 membranes. This enzyme population exhibits a structural defect that has not been fully defined but does not appear to involve cyteine 261. It is unclear whether this enzyme population is ever functional in vivo. We detected no obvious structural defect in the V-ATPase purified from cys4 strains, but due to the size of the V-ATPase (at least 750-kDa), we might not detect small differences in subunit composition between the V-ATPase from wild type and cys4 membranes. Therefore, it is possible that the loss of V-ATPase activity may be due to the absence of some unidentified subunit(s) of the V-ATPase as a result of oxidative inhibition in vivo.
Oxidation and Reduction as a Means of Enzyme Regulation-Several mechanisms have been proposed for the regulation of vacuolar H ϩ -ATPases (reviewed in Ref. 12). These include disassembly and reassembly of peripheral V 1 and integral V 0 domains, recruitment of enzyme in vesicles to the site of action, regulation of parallel ion channels that compensate for the electrogenicity of the V-ATPase, changes in the degree of coupling between ATP hydrolysis and proton pumping, and reversible disulfide-bond interchange between conserved cysteine residues near the catalytic site.
The occurrence of reactive sulfhydryl groups is a common feature of enzymes that are inactivated by oxidizing agents such as iodoacetamide or N-ethylmaleimide and are stabilized or activated by reducing agents such as dithiothreitol. Oxidative inhibition in vitro has been demonstrated for several enzymes including the V-ATPase of bovine clathrin-coated vesicles (16), the N. crassa V-ATPase (23), and the rabbit muscle phosphofructokinase (54). In a recent study, it was shown that the loss of intracellular GSH causes a reversible impairment of the Na ϩ /H ϩ antiporter of Madin-Darby canine kidney cells (55). This type of result has led to the suggestion that modulation of the thiol/disulfide ratio in vivo may serve as a "third messenger" in response to signals such as cAMP levels (54).
Our findings provide the first in vivo evidence for the regulation of a V-ATPase by thiol/disulfide exchange, but do not necessarily address how important this regulation mechanism may be for the activity of the yeast V-ATPase in vivo in wild type cells. The oxidative stress imposed by the cys4 mutations, which revealed the evidence of oxidative inactivation of the yeast V-ATPase in vivo, probably exceeds the stresses encountered by a wild type cell under normal growth conditions. Nevertheless, a variety of biochemical studies, combined with the conservation of the catalytic site cysteine corresponding to Cys-261 of the yeast Vma1p, indicate that oxidative inactivation is a general characteristic of V-ATPases. A number of other eukaryotic cells possessing high levels of V-ATPases may encounter conditions of severe oxidative stress regularly (for example, macrophages undergoing an oxidative burst), and our results suggest that oxidative inactivation of the V-ATPase could be a very important mechanism for regulation of V-ATPase activity in these cells.