Characterization of a Temperature-sensitive Yeast Vacuolar ATPase Mutant with Defects in Actin Distribution and Bud Morphology*

The 27-kDa E subunit, encoded by theVMA4 gene, is a peripheral membrane subunit of the yeast vacuolar H+-ATPase. We have randomly mutagenized theVMA4 gene in order to examine the structure and function of the 27-kDa subunit. Cells lacking a functional VMA4 gene are unable to grow at pH > 7 or in elevated concentrations of CaCl2. Plasmid-borne, mutagenized vma4 genes were screened for failure to complement these phenotypes. Mutants producing Vma4 proteins detectable by immunoblot were selected; one (vma4–1 ts ) is temperature conditional, exhibiting the Vma− phenotype only at elevated temperature (37 °C). Sequencing revealed that a single point mutation, D145G, was responsible for the phenotypes of thevma4-1 ts allele. The unassembled 27-kDa subunit made in the vma4-1 ts cells is rapidly degraded, particularly at 37 °C, but can be protected from degradation by prior assembly into the V-ATPase complex. In purified vacuolar vesicles from the mutant cells, the peripheral subunits are localized to the vacuolar membrane at decreased levels and a comparably decreased level of ATPase activity (14% of the activity in wild-type vesicles) is observed. When vma4-1 ts mutant cells are shifted to pH 7.5 medium at 37 °C, the cells become enlarged and exhibit multiple large buds, elongated buds, and other abnormal morphologies, together with delocalization of actin and chitin, within 4 h. These phenotypes suggest connections between the vacuolar ATPase, bud morphology, and cytokinesis that had not been recognized previously.

translocating ATPase (H ϩ -ATPase) is a multisubunit complex that acidifies the yeast vacuole (4,5), and biochemical and genetic characterization indicates that it consists of at least 13 polypeptides (3). The enzyme complex is divided into a peripheral cytoplasmic domain, V 1 , which contains the sites for ATP hydrolysis, and an integral membrane domain, V o , which forms the proton channel. With the exception of the vph1⌬ and stv1⌬ mutants, which lack two different isoforms of the 100-kDa subunit (6,7), disruption of any of the ATPase subunit genes in yeast results in an identical set of Vma Ϫ (vacuolar membrane H ϩ -ATPase) phenotypes, including a complete loss of vacuolar acidification and bafilomycin A 1 -sensitive ATPase activity in isolated vacuoles, an inability to grow in medium buffered higher than pH 7 (with optimal growth at pH 5), sensitivity to high extracellular Ca 2ϩ concentration, and a number of other physiological alterations (8 -10).
The overall physiological consequences of loss of vacuolar acidification have recently been examined in a number of organisms other than yeast, either by construction of a chromosomal disruption of a V-ATPase subunit gene (11,12), or by growth of cells in the presence of concanamycin A, a specific inhibitor of V-ATPases (13,14). Disruption of the gene encoding the B subunit of the Drosophila V-ATPase resulted in a larval-lethal phenotype, with evidence that many or most different cell types required the gene product for growth (12), and disruption of the catalytic A subunit in Neurospora crassa also proved to be lethal (11). Growth of N. crassa in the presence of inhibiting concentrations of concanamycin A resulted in a conditional phenotype; growth of wild-type cells was inhibited at pH 7, but cells were able to grow in medium buffered to pH 5.8. This growth phenotype could be suppressed by mutations in the plasma membrane proton pump, but under all conditions, the cells showed gross morphological alterations in the presence of concanamycin A (14). Dictyostelium also exhibited gross morphological alterations in its endomembrane system when cells were grown in the presence of concanamycin A (13). These experiments suggest that V-ATPases play a critical but poorly defined role in overall cell physiology and morphology, in addition to their specific roles in endocytosis and protein sorting.
Even though the yeast Vma Ϫ growth phenotype has been used in screening for mutants showing defects in vacuolar acidification (8,15,16), the underlying mechanism for the growth arrest of yeast cells at elevated pH or in high concentrations of CaCl 2 is still not known. The cell division cycle (CDC) of the yeast Saccharomyces cerevisiae involves several sequential morphogenetic events before cytokinesis, including formation of a ring of chitin (the "bud scar") at the site of the previous budding events, and concentration of patches of actin filaments at the bud site (17). During bud formation, the location of the pre-bud site is specified by the mating type of the cell and the site of the previous cytokinesis, and the cell polarity is maintained during bud growth. A very large number of genes have been identified that affect progression of these morphological changes with the cell cycle. Cells with defects in either CDC42 or CDC24 genes, which are essential for the establishment for cell polarity, lose the asymmetric location of actin filaments and the cell is arrested as large, round, unbudded cell (18,19). During the pre-bud site assembly, the actin rearrangement is triggered by the activations of CDC28 and CLNs (20). Mutant cells defective in CDC3, CDC10, CDC11, or CDC12 are unable to complete cytokinesis but undergo multiple cycles of budding, DNA synthesis, and nuclear division. The buds formed in these mutants are abnormally elongated (21). There are also indications that alterations in cytoplasmic pH may be linked to the cell cycle. Some plasma membrane H ϩ -ATPase (pma1) mutants fail to complete bud enlargement and cytokinesis, and the cells accumulate nucleated small buds (22). A significant change in cytoplasmic pH has been demonstrated during the early part of the cell cycle (23), so the ATPasemediated regulation of internal pH may play a central role in regulating cell division.
The VMA4 gene encodes the 27-kDa E subunit of the peripheral (V 1 ) sector of the yeast vacuolar ATPase (24). Null mutants (vma4⌬) revealed that the 27-kDa subunit is essential for the assembly and activity of the vacuolar ATPase (15). The null vma4 mutants are viable, but exhibit a typical Vma Ϫ phenotype (24). Vacuolar membranes prepared from the vma4⌬ mutants lacked the 69-, 60-, 42-kDa and other peripheral subunits, whereas a significant portion of the 100-kDa subunit (and/or the 75-kDa breakdown product) was still present on the vacuolar membrane (15), suggesting that the 27-kDa subunit is required for stable assembly of the V 1 sector onto the vacuolar membrane. Tomashek et al. (25) demonstrated that stability of the Vma4 protein depended on the presence of the VMA10 gene product and that the VMA4 and VMA10 gene products could be cross-linked, suggesting that these two V 1 subunits may interact as part of a "stalk" in the V-ATPase.
In this study, we have randomly mutagenized the VMA4 gene in order to study the structure and function of the 27-kDa subunit. One of the mutants (vma4-1 ts ) is temperature sensitive, and only shows the Vma Ϫ phenotype at 37°C. The 27-kDa subunit was made in this mutant, but rapidly degraded at the nonpermissive temperature. Assembly of the mutant 27-kDa subunit into V-ATPase complexes protected it from degradation and the assembled complexes showed ATPase activity. When the vma4-1 ts cells were shifted to medium buffered to pH Ͼ 7 at 37°C, the cells show abnormal bud morphologies and the delocalization of actin and chitin, suggesting the growth arrest of vma4-1 ts at elevated pH may result from the defects in secretion and polarized growth. These experimental results reveal a possible connection between the vacuolar ATPase, bud formation, and polarized growth during the cell division cycle of budding yeast that could be mediated by intracellular pH or Ca 2ϩ changes.  (29), to give deletion strains SF838-5A␣ vma4⌬ and SF838-1D␣ vma4⌬. Deletion of VMA4 was confirmed by polymerase chain reaction of chromosomal DNA prepared from yeast cells (30), using primers: 5Ј-TAGGTATA-CAAGCTGCTG and 5Ј-TTTCGGCCGACTTGTCCCTCGTTGCT. DNA sequencing was performed using a Sequenase kit (U. S. Biochemical Corp.) and the dideoxy chain termination method using the same primers (31). The plasmid pvma4-1 (consisting of the pVMA4 plasmid with the D145G mutation in the VMA4 gene) was isolated from a yeast strain expressing the temperature conditional Vma Ϫ phenotype after screening as described below. A strain containing the vma4-1 ts allele integrated at the VMA4 locus (JWY1) was generated by a two-step gene replacement (32). The vma4-1 ts allele was cut from the pRS315 vector with XhoI and SacII and then ligated to integrating vector pRS305 (28) digested with SalI and SacII to form plasmid pJW6. pJW6 was digested with SphI to linearize the plasmid. SF838-5A␣ vma4⌬ cells were transformed with the digested plasmid, and integrants were selected by growth on supplemented minimal medium lacking leucine. The integrants were then grown non-selectively (on YEPD, pH 5.0 at 25°C), and cells that became auxotrophic for uracil were identified by growth on plates containing 5-fluoroorotic acid (27). Integration of the vma4-1 ts allele at the VMA4 locus to generate strain JWY1 (with loss of the vma4⌬::URA3 allele) were confirmed by the presence of a fragment of the appropriate size by polymerase chain reaction from genomic DNA, followed by sequencing of this polymerase chain reaction fragment. The strain overproducing the vma4-1 ts allele (JWY1/2-vma4-1 ts ) was generated by first cloning the BamHI fragment containing the vma4-1 ts allele from pRS315 into YEp24 which had been digested with BamHI to form plasmid 2-vma4-1 ts , and then transforming the JWY1 strain with this plasmid and selecting uracil prototrophs. Mutagenesis of VMA4 Gene and Screening for Vma Ϫ Cells-The VMA4 gene on the pRS315 plasmid was propagated in XL 1-Red cells, an Escherichia coli strain which is deficient in three of the primary DNA repair pathways (mutS, mutD, mutT) (Stratagene). The mutagenized plasmids were transformed to the vma4⌬ yeast strain by an overnight lithium acetate procedure (33). Transformants were selected on supplemented minimal medium lacking leucine. Vma Ϫ cells were screened by checking the growth on YEPD, pH 7.5, 60 mM CaCl 2 and YEPD, pH 5.0, plates at 25, 30, and 37°C. Transformants that grew on pH 5.0 but failed to grow on pH 7.5 Ca 2ϩ plates were selected for further characterization as described (34).
Immunoprecipitation-To study the assembly of vacuolar ATPase, immunoprecipitations were carried out under nondenaturing conditions in the presence of 0.6 mM dithiobis(succinimidylpropionate) added at the time of solubilization as described (35), except that cells were converted to spheroplasts at room temperature and shaken in supplemented minimal medium lacking methionine containing 1.2 M sorbitol for 20 min at 25 or 37°C before labeling. ATPase complexes were immunoprecipitated with the 8B1 monoclonal antibody (36). To study the turnover of Vma4 protein, cells were converted to spheroplasts and labeled as for the nondenaturing immunoprecipitations and the chase was initiated by addition of unlabeled methionine and cysteine to 0.33 mg/ml each. At each time point, spheroplasts were pelleted by centrifugation and lysed in 100 l of pre-warmed immunoprecipitation buffer (10 mM Tris-HCl, pH 8, 1% Triton X-100, 1% SDS, 20 mM EDTA) by incubation at 75°C for 20 min, the lysis mixture was diluted to 0.5 ml with water, then the mixture was placed on ice and pretreated with Protein A-Sepharose Cl-4B. After centrifugation to pellet the Protein A-Sepharose, the supernatant was incubated overnight with 3 l of rabbit anti-27-kDa polyclonal antiserum (a generous gift from Tom Stevens and Margaret Ho). Protein A-Sepharose was then added to precipitate the immune complexes, and immunoprecipitated proteins were washed, solubilized, and analyzed by SDS-PAGE as described (35). Duplicate samples for each time point were quantitated on a Molecular Dynamics PhosphorImager (Model 425E).
Isolation of Vacuolar Vesicles and ATPase Activity Assay-Vacuolar vesicles were prepared as described (37), except that the temperaturesensitive mutant cells were grown overnight at 25°C and all of samples were converted to spheroplasts at room temperature. JWY1/2-vma4-1 ts cells and SF838 -5A␣ cells transformed with YEp24 were grown in supplemented minimal medium lacking uracil to an optical density at 600 nm (OD 600 ) of 0.8 -1.0 before harvesting. The measurement of vacuolar ATPase activity was performed on a Beckman DU640 spectrophotometer using a coupled enzyme assay at 25°C (38). To examine the thermal stability of the V-ATPase, both types of vesicles were incubated at 25 and 37°C for 30 and 60 min, then the activity was assayed at 25°C and samples were taken for Western blotting. In addition, the stability of the activity in the mutant vesicles was monitored by performing the coupled enzyme assay at 37°C. No decrease in activity was observed over 30 min. Protein concentrations were determined by the Lowry assay (39). Concanamycin A-sensitive ATPase activity was determined by comparing the ATPase activity with and without a 20-min preincubation with 100 nM concanamycin A (40).
Western Blotting-To determine the levels of the 69-, 60-, 42-and 27-kDa subunits in whole cells, the cells were grown overnight at 25°C, then diluted to equivalent density in YEPD, pH 5.0, medium and grown at both 25 and 37°C. Equal numbers of cells (ϳ10 7 ) were removed at various times, and whole cell lysates were prepared and analyzed by SDS-PAGE and immunoblotting as described previously (41). Vacuolar proteins were also detected as described previously (41). The 10D7, 7D5, 13D11, and 7A2 monoclonal antibodies were used to detect the 100-, 69-, 60-, and 42-kDa subunits (41), and polyclonal antisera raised against the 27-kDa subunit were used to detect Vma4p (15).
Fluorescence Microscopy-For staining with quinacrine, cells were grown under various conditions at 25 or 37°C to logarithmic phase, then 0.5 ϫ 10 7 cells were harvested, washed with phosphate-buffered saline, pH 7.5, containing 2% glucose, incubated with 200 M quinacrine, pH 7.5, for 5 min, and visualized immediately as described (37). To visualize actin in whole cells (42), cells were fixed in 3.7% formaldehyde for 1 h after a defined period of growth at 25 or 37°C. Cells were sonicated, pelleted, washed, and resuspended in phosphate-buffered saline. Then fixed cells were incubated with 0.6 M rhodamine-phalloidin for 2 h in the dark. Finally, cells were washed extensively with phosphate-buffered saline and resuspended in mounting medium (90% glycerol, 0.1 mg/ml -phenylenediamine). Immunofluorescence micrographs were obtained by using an Axioskop (Zeiss) microscope under fluorescein isothiocyanate optics for observation of quinacrine staining and rhodamine optics for observation of actin with a ϫ 100 objective. To detect chitin localization, cells were grown overnight at 25°C to the mid-logarithmic phase, then diluted to the same density in YEPD, pH 5.0, or YEPD, pH 7.3, medium and grown at 25 or 37°C. Constant numbers of cells (0.3 ϫ 10 7 ) were taken at various times and briefly sonicated in microcentrifuge tubes to disperse clumps, then pelleted and washed with water. Chitin labeling of bud scars was observed after incubation of cells in a 300 g/ml solution of calcofluor for 5 min in the dark (42). For visualization of nuclei, 1 g/ml DAPI was included in the mounting medium (42). Chitin and DAPI staining were observed using a UV filter set. Photographs were taken using Tmax 400 film (Kodak). Images were printed on Rapitone paper and arranged using Adobe Photoshop 3.0. Morphology of the cells was observed under Nomarski optics. In scoring populations of wild-type and mutant cells under various conditions, at least 150 cells for each condition were analyzed and large budded cells were defined as cells where the diameter of the smallest spheroid was greater than half of the diameter of the largest one.

Identification of a Temperature Conditional Mutation in
VMA4 -In order to better characterize the structural and functional roles of the VMA4 gene product, we randomly mutagenized the gene in vitro and then screened for mutations that affected vacuolar H ϩ -ATPase function in vivo. Deletion of the VMA4 gene leads to the loss of growth on medium buffered to pH greater than 7 or medium containing high concentrations of CaCl 2 , with the most severe growth defects occurring in medium that has both elevated calcium concentrations and elevated pH (8,9). The Vma Ϫ phenotype can be complemented by transforming the vma4⌬ strain with the plasmid carrying VMA4 gene, as shown in Fig. 1A. We randomly mutagenized the VMA4 gene by transforming the mutator E. coli strain XL 1-Red with a plasmid-borne copy of the gene. The random mutation rate in the mutator strain was reported to be about one base change per 2000 nucleotides (43). The mutagenized vma4 plasmids were used to transform a vma4⌬ yeast strain. Transformants were selected by leucine prototrophy and mu-tant plasmids failing to complement Vma Ϫ phenotypes were identified. About 2,300 transformants were screened and 6 colonies exhibiting a Vma Ϫ phenotype were selected. From this collection, only 5 mutants producing stable Vma4 protein based on Western blot analysis were identified. When the mutants were screened further for temperature dependence of the pHsensitive growth phenotype, one of these mutants (vma4-1 ts ) proved to be temperature conditional, exhibiting the Vma Ϫ phenotype only at elevated temperature (37°C). As shown in Fig. 1A, vma4-1 ts cells grew normally at the permissive temperature (25°C or 30°C). When cells were grown at 37°C, however, the vma4-1 ts cells failed to grow on YEPD, pH 7.5, ϩ 60 mM CaCl 2 medium, but remained capable of growth in pH 5.0 medium. The vma4-1 ts mutation was then integrated into the genome by pop-in, pop-out gene replacement (32). The resulting mutant (JWY1) showed the same temperature and pH conditional growth phenotype as the vma4-1 ts mutation borne on a CEN-plasmid. Wild-type cells transformed with the vma4-1 ts mutant on a CEN-plasmid showed normal growth at pH 7.5 at 37°C, suggesting that the mutation is recessive.
Sequencing of the vma4-1 ts mutation revealed one nucleotide change in the open reading frame of the VMA4 gene, resulting in change at Asp 145 3 Gly. This aspartate is perfectly conserved in eight genes for the 27-kDa subunit that have been cloned from widely divergent organisms ( Fig. 1B) (44,45), even though it lies in a generally poorly conserved region of the VMA4 gene.
Biochemical Analysis of the vma4-1 ts Mutant-Production of 27-kDa subunit protein in cells grown at 30°C was one of the criteria in our mutant screen, but we investigated whether the temperature dependence of the vma4-1 ts mutant might be related to protein stability. Western blot analysis of whole cell lysates revealed that vma4-1 ts cells growing at the permissive temperature (25°C) have a somewhat lower steady-state level of the Vma4 protein (27-kDa) than wild-type cells, but mutant cells grown overnight at 37°C have no detectable 27-kDa subunit ( Fig. 2A). Both cell lines have comparable levels of the 69-kDa V 1 subunit at both temperatures ( Fig. 2A). The time course for the loss of the 27-kDa subunit with a temperature shift is shown in Fig. 2B. After a shift to 37°C, there was a dramatic loss of the 27-kDa subunit in vma4-1 ts cells over several hours, with about 50% of the protein seen at 25°C remaining after 8 h at 37°C, and none detectable after 24 h. Other peripheral subunits of the V-ATPase, the Vma1 (69-kDa), Vma2 (60-kDa), and Vma5 (42-kDa) proteins remained at the same levels in whole cell lysates from vma4-1 ts cells at both permissive and nonpermissive temperatures (Fig. 2B). Wild-type cells show no decrease in the level of the 27-kDa subunit during incubation at 37°C (Fig. 2C).
The steady state level of any protein is determined by a combination of the rate of synthesis and the rate of decay. We explored the instability of the Vma4 protein further by biosynthetically labeling yeast cells and determining the rate of decay of the labeled protein. Short pulse and chase experiments, followed by immunoprecipitation of the denatured subunit with a polyclonal antibody, showed the labeled 27-kDa subunit in wild-type cells was quite stable over a subsequent 60-min chase at either 25 or 37°C. Wild-type cells labeled at 37°C initially incorporated approximately 40% more 35 S into the 27-kDa subunit than cells labeled at 25°C, based on PhosphorImager quantitation of the immunoprecipitated 27-kDa protein, but cells had comparable levels of labeled 27-kDa subunit at both temperatures after a 60-min chase (Fig. 3A). In contrast, the newly synthesized 27-kDa subunit made in the vma4-1 ts mutant was rapidly degraded at both 25 and 37°C (Fig. 3B). In addition, there was a substantially lower amount of 27-kDa subunit after the initial 5-min pulse at 37°C than there was after the 25°C pulse, suggesting that a substantial part of the 27-kDa subunit was being degraded in less than 5 min at 37°C. At both 25 and 37°C, there appeared to be a fraction of the 27-kDa subunit that was stable for 60 min. If the mutant 27-kDa subunit is protected from degradation by assembly into the V-ATPase complex, then we might expect that a stable fraction could be "rescued" from degradation by the competing process of assembly at early times (see below). These results suggest that the difference in the rate of decay at 25 and 37°C of the 27-kDa subunit immediately after synthesis, combined with the small difference in the stable fractions at longer times, must be sufficient to account for the substantial difference in the steady state level of the protein at the two different temperatures seen in Fig. 2. If the 27-kDa subunit is assembled into the V-ATPase complex very rapidly (within 5 min after synthesis) as preliminary results suggest, 2 then the rate of decay immediately after synthesis would be particularly important, and at 25°C there might be much more assembly of V-ATPase complexes than at 37°C.
We wished to examine the properties of the V-ATPase containing the mutant 27-kDa protein further, but vacuoles from yeast strains carrying the mutant allele on a low copy plasmid contained very low levels of the 27-kDa subunit, even when the vacuoles were isolated at 25°C. We therefore constructed a strain that would overproduce the vma4-1 ts allele by transforming JWY1, which has the vma4-1 ts allele integrated at the VMA4 locus, with a 2-plasmid carrying the vma4-1 ts allele. This strain contained a higher steady-state level of 27-kDa 2 M. Tarsio and P. Kane, unpublished data.

FIG. 2. The 27-kDa subunit is destabilized in the vma4-1 ts mutant.
A, steady-state levels of the 27-and 69-kDa subunits in wild-type (1) and vma4-1 ts mutant (2) cells grown overnight in YEPD, pH 5, medium at the indicated temperature. B and C, time course of decay of the mutant 27-kDa subunit at 37°C. The vma4-1 ts mutant (SF838-1D␣vma4⌬/pvma4-1 ts ; B, and wild-type cells (SF838-1D␣; C, were grown overnight at 25°C, then diluted to the same density in YEPD, pH 5.0, medium, and growth continued at 25 or 37°C for the indicated times. For A-C, samples corresponding to a constant number of cells were removed from each culture and whole cell lysates were prepared after the indicated period of growth. The lysates were separated by SDS-polyacrylamide gel electrophoresis and V-ATPase subunits were identified by Western blotting. The 69-, 60-, and 42-kDa subunits were detected with monoclonal antibodies 8B1, 13D11, and 7A2, respectively, and the 27-kDa subunit was detected with subunit-specific polyclonal antisera as described under "Experimental Procedures." FIG. 1. Genetic characteristics of the vma4-1 ts mutant. A, the vma4-1 ts mutant exhibits a Vma Ϫ phenotype at elevated temperature. Cells were streaked as indicated on YEPD, pH 7.5, ϩ 60 mM CaCl 2 or YEPD, pH 5.0, plates and incubated at 25°C or 37°C for 3 days. Strains shown are wild-type (SF838-1D␣); vma4⌬ (SF838-1D␣vma4⌬); vma4⌬//pVMA4 (SF838-1D␣vma4⌬ carrying the wild-type VMA4 gene on the pRS315 (CEN) plasmid); vma4⌬/pvma4 ts (SF838-1D␣vma4⌬ carrying the vma4-1 ts allele on the pRS315 plasmid). B, multiple sequence alignment of the region surrounding the mutation in the yeast vma4-1 ts mutant. The vma4-1 ts has a single amino acid change at Asp 145 3 Gly; the aspartate mutated is conserved in eight genes for the 27-kDa subunit that have been cloned from different species.
protein at 25°C because of increased levels of synthesis, even though the rate of decay of the protein in pulse-chase studies (Fig. 3C) was fairly similar to the strains containing the vma4-1 ts at low copy (Fig. 3B). Significantly, the strain overproducing the vma4-1 ts allele showed a temperature conditional Vma Ϫ phenotype as severe as the strain carrying the integrated vma4-1 ts allele or the vma4⌬ strain carrying this allele on a low copy plasmid (data not shown). Using the overproducing strain, we confirmed that assembly of the 27-kDa subunit into V-ATPase complexes protected the subunit from degradation at 37°C by immunoprecipitation of the V-ATPase complex under nondenaturing conditions (Fig. 3D). The mutant strain labeled for 60 min at 25°C contains assembled V-ATPase complexes that are stable through a subsequent 60-min chase at either 25 or 37°C. In contrast, the mutant strain labeled for 60 min at 37°C contains very little 27-kDa subunit, although, surprisingly, the other V-ATPase subunits can be coprecipitated with the 69-kDa subunit. These results indicate that assembly of the 27-kDa subunit synthesized at 25°C is much more efficient than assembly of the subunit synthesized at 37°C.
In order to examine further the stability and activity of ATPase complexes assembled in the vma4-1 ts mutant, vacuolar membranes were prepared from wild-type cells and mutant cells overproducing the vma4-1 ts allele grown at 25°C. The concanamycin A-sensitive ATPase specific activity of the mutant vacuolar vesicles was 0.27 mol of P i /min/mg, when assayed at 25°C, and vesicles from the wild-type strain had a specific activity of 2.0 mol of P i /min/mg when assayed under similar conditions. The thermal stability of the ATPase activity in both membranes was assessed by incubating the membranes at 25 and 37°C for 30 and 60 min. The ATPase activity was stable in both types of membranes at both temperatures for up to 60 min. Western blot analysis of vacuolar membranes prepared from the vma4-1 ts cells revealed that the amount of 69-, 60-, 42-, and 27-kDa subunits present on the membrane is significantly decreased relative to wild-type, whereas the 100-kDa subunit and its 75-kDa breakdown product are present at the same level (Fig. 4). Quantitative comparison of the levels of 27-and 69-kDa subunit in the wild-type and vma4-1 ts mem-branes (Fig. 4A) indicated the mutant membranes contained approximately 13% the level of these subunits present the wild-type membranes. This result suggests that the reduction in ATPase activity in the mutant vacuolar membranes can be fully accounted for by the reduction in the levels of V 1 subunits. The decrease in the levels of all of the peripheral subunits suggests that although these subunits appear to be able to associate with the V o subunits on a 1-2 h time scale, based in the immunoprecipitations in Fig. 3D, these associations are not stable enough to survive through the several hours required for vacuole isolation. Incubation of the vacuolar membranes at 37°C did not significantly affect the levels of any of the subunits in the wild-type or mutant cells (Fig. 4B).
Loss of Acidification at Elevated Temperature in vma4-1 ts Mutants-Uptake of the fluorescent lysosomotropic amine quinacrine into the yeast vacuole provides a qualitative measure of vacuolar acidification in vivo (37). In order to determine the approximate rate at which vacuolar acidification is lost in the vma4-1 ts mutant after a shift to 37°C, we stained cells grown under various conditions with quinacrine and visualized the staining by fluorescence microscopy (Fig. 5). As shown in Fig. 5, G and I, wild-type cells show bright staining with quinacrine (left) that colocalizes with vacuoles visualized under Nomarski optics (right) whether they are incubated at 37°C in pH 7.3 medium or at 25°C in pH 5 medium. vma4⌬ mutants show no quinacrine uptake under either condition (Fig. 5, H  and J, left), despite the presence of vacuoles of normal mor-phology (right). vma4-1 ts mutants accumulate quinacrine into the vacuole and show staining comparable to that of wild-type cells when grown at 25°C (Fig. 5, A and C). After 1 h of growth at 37°C, however, the vma4-1 ts cells showed slightly diminished quinacrine staining (Fig. 5B), and after 3 h at 37°C, quinacrine staining of the vacuole was completely gone (Fig. 5, D-F). After 3 h at 37°C, some of the vma4-1 ts cells appeared to show some cytoplasmic staining (Fig. 5, D and F), but this staining was excluded from the vacuole and was also seen in a small population of vma4⌬ cells. These results suggest that the V-ATPase had lost its ability to maintain a proton flux across the vacuolar membrane in the vma4-1 ts mutant after 3 h at 37°C, regardless of whether the mutant was grown at pH 7.3 (Fig. 5D), pH 5.0 (Fig. 5E), or in 100 mM CaCl 2 (Fig. 5F). We cannot eliminate the possibility that mutant cells growing at pH 5.0 maintain some acidification of the vacuole by a means other than proton pumping by the V-ATPase (for example, endocytosis (46)) because the actual quinacrine staining must be carried out at elevated pH to allow deprotonation of the amine.
The vma4-1 ts Cells Also Show Defects in Bud Morphology and Cytokinesis-Although the Vma Ϫ phenotype has been used frequently for the identification of mutants lacking vacuolar H ϩ -ATPase activity, the basis of the pH and Ca 2ϩ -sensitive growth accompanying loss of V-ATPase activity is not well understood. The vma4-1 ts mutant is the first temperature conditional vma mutant to be characterized, and it could be an ideal tool for examining the onset of Vma Ϫ phenotypes and determining the biological basis of these phenotypes. To examine the onset of the Vma Ϫ phenotypes, haploid vma4-1 ts cells were first grown overnight to logarithmic phase at 25°C in medium buffered to pH 5, then cells were transferred to pH 7.3 medium and shifted to 37°C. Control experiments were done by growing the cells in pH 5 medium at 25 and 37°C and in pH 7.3 medium at 25°C. Microscopic analysis of cells shifted to pH 7.3 medium and incubated at 37°C revealed that up to 25% of the vma4-1 ts cells showed abnormally elongated or multiple buds after 4 h (Fig. 6). Under similar conditions, only 4% of vma4⌬ cells had abnormal buds, and almost none of the wildtype cells showed the aberrant morphologies. There was a pronounced increase in the number of unbudded cells after vma4⌬ cells were shifted to pH 7.3 medium for 24 h (Fig. 6), although even at pH 5.0, the vma4⌬ mutant appears to contain a higher proportion of unbudded cells than the other strains.
In order to determine whether the aberrant morphologies quantitated in Fig. 6 are a result of cell death rather than a more direct cellular consequence of the shift to elevated pH, we examined whether the mutant strains incubated at elevated temperature at pH 7.5 could return to growth at pH 5.0. The results of this experiment are shown in Fig. 7A. Wild-type, vma4⌬ mutant, and vma4-1 ts mutant cells were grown in liquid culture (YEPD buffered to pH 5.0, 25°C or YEPD buffered to pH 7.5, 37°C) for the times indicated and a constant number of cells were removed from the culture, serially diluted, and plated on YEPD, pH 5.0, plates. As demonstrated in Fig. 7A, the vma4-1 ts mutant cells are able to return to growth in numbers comparable to wild-type even after 24 h incubation under the nonpermissive conditions. Growth curves from this mutant strain (Fig. 7B) indicate that the growth rate of the mutant significantly declines relative to wild-type by 4 h of incubation at 37°C in pH 7.5 medium, and there is almost no growth after 8 h incubation. The vma4⌬ mutant exhibits an almost immediate loss of growth upon a shift to pH 7.5 at 37°C (Fig. 7B), but in contrast to the vma4-1 ts mutant, a large proportion of the arrested cells are unable to return to growth at pH 5, even after 6 h incubation. These results indicate that the morphological phenotypes observed in the vma4-1 ts mutant can be attributed directly to changes in response to loss of V-ATPase activity at elevated pH and not to secondary effects of cell death.
In budding yeast, bud sites are selected in an axial pattern in haploid cells, so cells bud near the site of the previous budding. Wild-type yeast cells exhibit a ring of chitin at the neck of the emerging bud that remains on the mother cell after each cell division as a bud scar, as well as faint chitin staining over the body of the mother cell. Deposited chitin can be stained with the fluorescent dye calcofluor (Fig. 8). The distribution of chitin in wild-type cells was not altered by changes in temperature or pH (Fig. 8, A and B). In pH 5.0 medium at 25°C, both vma4⌬ and vma4-1 ts cells appear to show normal patterns of budding and chitin deposition, with perhaps slightly more staining of the cell body than in wild-type cells (Fig. 8, C and E). In pH 7.3 medium, after a shift to the restrictive temperature for 16 h, the vma4-1 ts cells showed a bright diffuse chitin distribution over the entire cell surface, including the buds (Fig. 8F). In cells containing multiple buds (Fig. 8F), the buds appeared to emerge adjacent to each other, suggesting that an axial budding pattern was largely maintained. After as little as 2 h at 37°C in pH 7.3 medium, the vma4-1 ts cells showed chitin delocalization (data not shown). vma4⌬ cells show a similar but less pronounced chitin delocalization, and tend to have a random budding pattern even though most of them arrested as unbudded cells (Fig. 8D). Both the vma4-1 ts and the vma4⌬ mutants became significantly larger after growth at elevated pH and temperature. These results suggest that at elevated pH, cells lacking vacuolar H ϩ -ATPase activity lose the ability to properly localize chitin, perhaps resulting from a loss of directed secretion.
The presence of multiple, malformed buds under the nonpermissive conditions suggest that the vma4-1 ts mutants exhibit a cell cycle defect under the nonpermissive conditions for growth. In order to characterize this defect more fully, we visualized the cellular DNA in the vma4-1 ts mutant with DAPI (Fig. 8, G and  H). Under the permissive conditions (pH 5.0, 25°C, Fig. 8G), vma4-1 ts mutants exhibit a fairly typical correlation of bud size and nuclear migration: small budded cells exhibit a single nucleus in the mother, large-budded cells exhibit nuclei in both the bud and the mother, and a number of cells show intermediate stages of nuclear migration. Under the nonpermissive conditions (Fig. 8H), most of the multiply budded cells appeared to contain nuclei in the buds, indicating that nuclear division and nuclear migration had occurred without completion of cytokinesis. A smaller proportion of the mutant cells appear binuclear, suggesting nuclear division occurred followed by a defect in nuclear migration in some cases.
Defects in chitin localization and bud morphology are often seen in cells with defects in actin localization. In wild-type yeast cells, actin localization is tightly coupled to the cell cycle (20). Filamentous actin is primarily localized to the bud site and the growing buds during the early stages of the cell cycle, and at the bud neck immediately before cytokinesis. Wild-type cells stained with rhodamine/phalloidin, which detects filamentous actin, showed a fairly typical distribution of actin at either 25 or 37°C and at either pH 5 (Fig. 9G) or pH 7.3 (Fig.  9H). The vma4-1 ts mutants also showed a normal actin distribution in pH 5.0 medium at 25°C (Fig. 9E) and normal morphology when viewed under Nomarski optics (Fig. 9F). After only 2 h under the nonpermissive conditions (pH 7.3, 37°C), however, the vma4-1 ts mutant strains had begun to show elongated buds when visualized under Nomarski optics (Fig. 9B), and cortical actin patches covering the entire cell body (Fig.  9A). After 16 h under the nonpermissive conditions, these phenotypes were more pronounced in the vma4-1 ts cells (Fig. 9, C  and D), and in addition, the actin staining was markedly brighter, with actin filaments extending into the malformed or elongated buds, and the cells were significantly enlarged. The vma4⌬ cells (Fig. 9J) were also enlarged and showed bright, delocalized cortical actin patches after 16 h at pH 7.3, 37°C.
The morphological defects of the vma4-1 ts strain were both pH-and temperature-sensitive. The most prominent morphological defects, including the presence of elongated and multiple buds and delocalization of actin and chitin were considerably diminished even at pH 6.7-6.8. Some of the phenotypes (abnormal bud morphology and chitin delocalization) were also  (SF838-1D␣vma4⌬, black bar), and vma4-1 ts cells (SF838-1D␣vma4⌬/pvma4-1 ts , gray bar) were grown to log phase at 25°C and then shifted to YEPD, pH 5, at 25°C or YEPD, pH 7.3, at 37°C for the indicated time before examining their morphology under Nomarski optics. Percentages of cells with the following morphologies are shown: UB, unbudded; SB, small budded (daughter cell less than 50% of the diameter of the mother cell); LB, large budded (daughter cell greater than 50% of the diameter of the mother cell); and AM, showing abnormal morphology, specifically multiple buds and/or elongated buds. At least 150 cells from each strain under each condition were scored to obtain the indicated percentages. exhibited by vma4-1 ts cells incubated at 37°C in unbuffered medium containing elevated (100 mM) concentrations of calcium (data not shown). In general, both the growth defects of the mutant and the morphological defects were less severe in calcium-containing medium than they were at elevated pH.

DISCUSSION
The yeast vacuolar H ϩ -ATPase closely resembles the vacuolar-type ATPases from other fungi, plant, and animal cells (3,4). Although all of the vacuolar H ϩ -ATPases that have been characterized appear to have at least eight subunits, the functions of only three of these subunits, the catalytic ATP-hydrolyzing subunit of approximately 70-kDa, the regulatory ATPbinding subunit of approximately 60-kDa, and the 17-kDa proteolipid involved in formation of the proton pore, have been clearly defined (3,5). These three subunits exhibit homology to subunits of the F 1 F 0 -ATPase that reflects an evolutionary relationship between the two enzymes (47).
The 26 -30-kDa E subunit of the vacuolar ATPase has been cloned from a number of species (Ref. 45, and references therein). The secondary structure of all of the sequenced E subunits is predicted to be predominantly ␣-helical, with conserved NH 2 and COOH termini, even though the overall sequence of Vma4 protein is not highly conserved (showing roughly 25% overall identity) (44,48). Although the primary sequence of the Vma4 subunit does not show significant similarity to any subunit of the F-type ATPase, it has been suggested that this subunit may have an analogous role in the V-type ATPase to that of the ␥-subunit in the F-type ATPase, which also has highly conserved ␣-helical structure near the COOH and NH 2 termini (44). The ␥-subunit is part of the stalk region of F 1 -ATPases and interacts directly with ␣and ␤-subunits through the ␣-helical regions (49), presumably to help communicate conformational changes in these subunits to the proton pore. Tomashek et al. (25) have provided recent structural evidence that the yeast Vma4p may also behave as a stalk subunit in combination with the Vma10p, Vma7p, and Vma8p subunits of the V-ATPase.
Asp 145 of S. cerevisiae VMA4 is highly conserved among sequences cloned from other species, and is near the beginning of the COOH-terminal ␣-helices predicted by computer analysis (44). The D145G mutation responsible for the phenotypes of the vma4-1 ts mutant appears to act primarily by destabilizing the unassembled protein. Although the mutant 27-kDa subunit is less stable than wild-type at both the permissive and nonpermissive temperatures, a larger proportion of the protein appears to be immediately degraded at 37°C and the fraction of the protein that is not immediately degraded is not competent for assembly. At 25°C, the stable fraction of the protein appears to be fully competent for assembly and, in fact, to be protected by assembly from degradation during a subsequent FIG. 7. A, vma4-1 ts mutant cells retain viability during incubation in YEPD, pH 7.5, at 37°C. Wild-type (SF838-1D␣), vma4⌬ (SF838-1D␣vma4⌬), and vma4-1 ts (SF838-1D␣vma4⌬/pvma4-1 ts ) cells were grown in liquid cultures under permissive conditions, YEPD, pH 5.0, at 25°C, then diluted into either YEPD, pH 5.0, and grown at 25°C or YEPD, pH 7.5, and grown at 37°C. At the indicated times, 10 6 cells were removed from each culture, the cells were 5-fold serially diluted 8 times (top to bottom for each condition) and then a constant fraction of the diluted mixture was spotted on YEPD, pH 5.0, plates and grown at 30°C. B, growth curves of wild-type (ࡗ, छ), vma4⌬ (f, Ⅺ), and vma4-1 ts (q, E) cells in YEPD medium buffered to pH 5.0 at 25°C (filled symbols) or YEPD medium buffered to pH 7.5 at 37°C (open symbols). All three cell types were grown at pH 5.0 under permissive conditions, then diluted into the indicated medium at time 0. Samples were removed at the indicated times and cell density quantitated from the absorbance at 600 nm (OD 600 nm ). incubation at 37°C (Fig. 3). Complementation of the Vma Ϫ phenotype implies that V-ATPases assembled in the vma4-1 ts mutant at 25°C must exhibit some V-ATPase activity, and mutant vacuoles isolated from cells grown at 25°C contain 14% of the concanamycin A-sensitive activity of wild-type vacuoles and a comparably reduced level of the peripheral subunits (Fig.  4). Previous results have indicated that complementation of Vma Ϫ phenotype requires approximately 25% of the wild-type V-ATPase activity (34), but the vacuole isolation takes several hours and it is possible that there is more loss of activity from the mutant enzyme during this time than there is from the wild-type enzyme. Incubation of the mutant vacuolar membranes at 37°C indicated that both the ATPase activity and the V 1 subunits present in these membranes were stable for at least 1 h, but we cannot eliminate the possibility that there is some loss of ATP-driven proton pumping during incubation at elevated temperature. Consistent with this possibility, in Fig. 5 there seems to be some loss of quinacrine staining in the mutant strain after 1 h at 37°C. When the mutant V-ATPase is synthesized at 37°C, assembly of newly synthesized 27-kDa subunit is almost completely eliminated and V-ATPase complexes lacking the 27-kDa subunit but containing most or all of the other subunits appear, at least in relatively short-term experiments (Fig. 3D). In contrast, immunoprecipitations from the vma4⌬ strain with antibodies against the 60-and 69-kDa subunits detected only a complex containing these two subunits (50) after 60 min labeling, although we have more recently detected an interaction with the 100-kDa subunit as well. 3 We have not fully explored the possibility that the unstable 27-kDa subunit synthesized at 37°C in vma4-1 ts cells supports assembly of different partial complexes than those present in the vma4⌬ strain; however, the fact that the mutation is recessive suggests that any partial complexes formed are not functional.
The vma4-1 ts strain exhibits distinct phenotypic differences from the vma4⌬ strain. vma4⌬ cells lose viability much more rapidly during an incubation at pH 7.5 than the vma4-1 ts cells. Whole cells were labeled with rhodamine/phalloidin in order to stain F-actin, as described under "Experimental Procedures" in frames A, C, E and G-J, and H. The morphology of the same vma4-1 ts mutant cells shown in A, C, and E, are shown under Nomarski optics in B, D, and F. All pictures were photographed by using identical magnifications, and exposure times for the immunofluorescence micrographs were also the same. Images were arranged using Adobe Photoshop. Genotypes of strains are the same as in Fig. 7. matic morphological phenotypes, including elongated and multiple buds, while vma4⌬ cells accumulated primarily unbudded cells (Figs. 6, 8, and 9). There are several possible explanations for these differences. First, there may be a physiological difference between "long-term" loss of vacuolar H ϩ -ATPase function in the vma4⌬ mutant and the more sudden loss of ATPase function caused by shifting the vma4-1 ts mutant to elevated temperature. As described below, the effects of the vma mutants on morphology probably result from changes in cellular ion balances. It may be that other ion channels or pumps are recruited to partially compensate for loss of the V-ATPase during long-term growth of vma4⌬ mutant, but there is no time for such an adaptation when the vma4-1 ts mutant is shifted to elevated temperature. Consistent with this explanation, Bowman et al. (14) have reported that mutations in the plasma membrane proton pump of Neurospora can partially compensate for the defects of a strain grown in concanamycin A, and Temesvari et al. (13) have also observed adaptation to concanamycin A in Dictyostelium. It could be that some of these adaptations, which help sustain the vma4⌬ mutant yeast strain at pH 5.5, may accelerate its demise at pH 7.3. Alternatively, the long-term growth defects of the vma4⌬ strain, which cause retarded growth even under optimal conditions, pH 5.5, may weaken the cells to such an extent that they rapidly decline at pH 7.3, while the vma4-1 ts mutants are able to display more dramatic morphological aberrations because they maintain viability. A third possibility is that the vma4-1 ts mutant is not a true "loss of function" mutation, but this possibility seems inconsistent with the observation that the mutation is recessive.
The pH-dependent growth of vma mutants has been extensively exploited in screening for vma mutants, but the physiological reason that cells lacking vacuolar H ϩ -ATPase activity lose viability at elevated extracellular pH is not understood. Characterization of the vma mutants has focused primarily on changes inside the vacuolar lumen or in the later stages of the vacuolar network, including loss of organelle acidification, defects in protein sorting or processing, and depletion of Ca 2ϩ and metabolite stores (8,10,51,52), but none of these changes has been directly connected to a pH-dependent growth defect. Based on the phenotypes of the vma4-1 ts mutant at 37°C, it appears that the vacuolar H ϩ -ATPase is also essential for several functions occurring outside the vacuolar network, including maintenance of a polarized actin distribution, normal bud morphology, and cytokinesis when cells are at elevated pH (pH Ͼ 7).
The observation that the vma4-1 ts mutant affected these processes under the nonpermissive conditions was somewhat surprising but not unprecedented. Previous studies have suggested a role for the vacuolar H ϩ -ATPase in actin distribution and cell morphology. vma mutants exhibit synthetic lethality with several end mutants, including the end3⌬ and end4⌬ mutants, and this has been attributed to a requirement for endocytosis in mutants lacking a functional V-ATPase (46). An additional explanation, however, is that both the end mutants and the vma mutants impair certain functions of the actin cytoskeleton, affecting its roles in cell growth as well as in endocytosis, such that the combination of mutants is lethal. Specifically, the END4 gene is allelic to SLA2, a gene identified from synthetic lethal interactions with actin mutants. sla2⌬ mutants resemble the vma4-1 ts mutants in that cells become enlarged and show actin delocalization at 37°C (53). end3⌬ mutants also show delocalization of actin and chitin and budding defects indicative of disrupted actin function (54). Growth of Neurospora in the presence of concanamycin A results in a number of dramatic morphological phenotypes that could be associated with cytoskeletal disruption (14). In addition, certain mutations in the PMA1 gene, which encodes the yeast plasma membrane H ϩ -ATPase, cause defects in cellular morphology and cytokinesis, suggesting these processes might be generally connected to pH homeostasis. In the pma1-105 mutant, cells arrest with multiple, small nucleated buds (22), and two mutants containing PMA1 promoter mutations that reduce activity of the enzyme appear to cause an elongated bud morphology (55).
It is unlikely that a V-ATPase subunit plays a direct role in morphological or cell cycle processes, but loss of V-ATPase function has the potential to change a number of other cellular parameters that have been linked directly to the cell cycle. The yeast vacuolar H ϩ -ATPase is involved in Ca 2ϩ and pH homeostasis and in sequestration of a number of other metal ions. The vacuole is the major Ca 2ϩ store in yeast cells, and uptake of Ca 2ϩ into the vacuole is mediated by a Ca 2ϩ /H ϩ antiporter driven by the proton gradient that is established by the vacuolar H ϩ -ATPase (56). vma mutants have been shown to have elevated levels of cytoplasmic Ca 2ϩ , apparently as a result of inability to take up Ca 2ϩ into the vacuole. A vma4⌬ mutant was shown to have a cytosolic free ionized calcium level ([Ca 2ϩ ] c ) of 1.8 M when cells were preincubated with glucose and exposed to 10 mM extracellular CaCl 2 (57); in the absence of added extracellular CaCl 2 , other vma mutants exhibit a [Ca 2ϩ ] c of 900 nM under conditions when the wild-type [Ca 2ϩ ] c was 150 nM (8). The role played by the vacuolar H ϩ -ATPase in regulating cytosolic pH in yeast has not been clearly defined, but vacuolar H ϩ -ATPases are involved in pH homeostasis in other systems (58). Wild-type yeast cells maintain a relatively constant cytosolic pH with extracellular pH values varying from 3.0 to 7.5 (59), and are also able to grow over a wide range of pH values. It has not been established whether vma mutants lose control of cytosolic pH at elevated pH (Ͼ7) where growth of the mutants is affected. The vacuole is also a store for a number of other divalent cations and other metabolites, most of which require the vacuolar pH gradient for uptake (60), so loss of vacuolar H ϩ -ATPase activity would be predicted to elevate cytosolic levels of these ions and metabolites as well. Because the vacuolar pH gradient is a common currency for movement of a number of ions from cytosol to vacuole, it will be difficult to separate the roles of the vacuolar H ϩ -ATPase in pH regulation and regulation of other ion movements.
Cytosolic pH and Ca 2ϩ concentrations can potentially affect yeast cell morphology and cell cycle progression in a number of ways (61,62), many of which could account for the phenotypes of the vma4-1 ts mutant. Depletion of intracellular and extracellular Ca 2ϩ has been shown to generate first a pause in the cell cycle at G 1 , and subsequently a full arrest at G 2 /M, resulting in an accumulation of small-budded 2N cells (61). This result suggests that Ca 2ϩ regulation may be important at multiple places in the cell cycle, and therefore, the influence of various regulators of cytosolic Ca 2ϩ , including the vacuolar H ϩ -ATPase, may vary at different points in the cell cycle. The morphological defects of the vma4-1 ts mutant, including multiple and elongated buds that contain nuclei, suggest a defect in G 2 /M phase. The essential targets for Ca 2ϩ are not clear, but a number of potential targets have been identified. CDC24, which is essential for establishment of cell polarity and organization of actin toward the bud site (19), is predicted to be a Ca 2ϩ -binding protein based on its sequence (63). Interestingly, a vma5 mutation was recently demonstrated to show synthetic lethality with the cdc24-4 allele (64). Actin itself exhibits Ca 2ϩ binding properties that may influence its function, and a number of actin-binding proteins also appear to bind Ca 2ϩ (65). The link between intracellular pH in regulation of cellular differentiation and proliferation is also well established in a number of cell types (66,67). Although intracellular pH changes are difficult to measure in yeast cells and changes during the cell cycle have not been studied extensively, it has been demonstrated that starved yeast cells re-entering the cell cycle experience a rise in cytosolic pH of as much as 0.55 pH units (cytosolic pH 6.75 rising to pH 7.3) at approximately the time of DNA synthesis (23). One particularly intriguing target of pH regulation, given the phenotypes of the vma4⌬ and vma4-1 ts mutants, is actin itself. In Dictyostelium, intracellular pH may affect the dynamics of the cytoskeleton by modulating the interaction of EF-1␣, which acts as an F-actin-binding protein, with F-actin (68,69). In hamster lung fibroblast cells, the activation of the Na ϩ -H ϩ exchanger, NHE1, which plays a central role in intracellular pH homeostasis, is necessary for RhoA-mediated assembly of stress fibers (70).
Many of the defects seen in the vma4-1 ts mutant cells, including actin and chitin delocalization, enlargement of cells, and abnormal bud morphology are very similar to those seen in yeast actin and actin-binding protein mutants (Ref. 71, and references therein). One relatively simple hypothesis, consistent with the phenotypes of the vma4-1 ts mutants, is that when extracellular pH values rise above 7, the vacuolar H ϩ -ATPase plays an essential role in maintaining proper actin localization and that loss of organization in the actin cytoskeleton results in the other phenotypes of the vma4-1 ts mutant and other vma mutants. In future experiments we will attempt to address the specific connections between the vacuolar H ϩ -ATPase and actin functions.