Defective H(+)-ATPase of hygromycin B-resistant pma1 mutants fromSaccharomyces cerevisiae.

Mutations in the plasma membrane H(+)-ATPase gene (PMA1) of Saccharomyces cerevisiae that confer growth resistance to hygromycin B have been shown recently to cause a marked depolarization of whole cell membrane potential (Perlin, D. S., Brown, C. L., and Haber, J. E. (1988) J. Biol. Chem. 263, 18118-18122). In this report, the biochemical and genetic properties of H+-ATPases from four prominent hygromycin B-resistant pma1 mutants, pma1-105, pma1-114, pma1-147, and pma1-155, are described. Single base pair changes were identified in pma1-105, pma1-114, and pma1-147 that resulted in amino acid substitutions of Ser-368----Phe, Gly-158----Asp, Pro-640----Leu, respectively. An A----G transition mutation at -39 in the 5'-untranslated region of the mRNA of pma1-155 was also found. This mutation creates an out-of-Frame upstream AUG initiation codon that apparently reduces normal translation of PMA1. DNA sequence analysis of PMA1 from strain Y55 identified 9 base pair substitutions that resulted in 6 amino acid changes in nonconserved regions when compared to the published sequence for strain S288C. Plasma membranes of three of the four pma1 mutants contained normal amounts of H(+)-ATPase; membranes from pma1-155 contained enzyme at 62% of the wild-type level. The kinetics of ATP hydrolysis were most strongly altered for enzymes from pma1-105 and pma1-147 which showed changes in both Km and Vmax. A striking pH dependence for these parameters was found for enzyme from pma1-105 which resulted in a precipitous decline in Km and Vmax below pH 6.5. ATP hydrolysis by enzymes from pma1-105 and pma1-147 was insensitive to inhibition by vanadate. These enzymes, in contrast to wild-type and vanadate-sensitive mutant enzymes, were poorly protected from trypsin-induced inactivation by MgATP and vanadate or Pi alone. These results are pertinent to the mechanism of vanadate-induced enzyme inhibition and suggest that Ser-368 and Pro-640 influence the affinity of the phosphate-binding site for Pi. All mutant enzymes catalyzed ATP-induced pH gradient formation following purification and reconstitution into liposomes. Finally, these results further demonstrate the usefulness of hygromycin B as a generalized screening tool for isolating diverse plasma membrane ATPase mutants.

publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. I To whom correspondence should be addressed.
The fungal plasma membrane contains an H'-ATPase that maintains the electrochemical proton gradient necessary for ion and nutrient transport and also plays an important role in intracellular pH regulation (1,2). The H+-ATPase has been extensively characterized from the yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe, as well as from the ascomycete fungus Neurospora crassa (1,2). It consists of a single subunit of about M , = 100,000 (3-5), forms a phosphorylated intermediate (6)(7)(8), is sensitive to inhibition by vanadate (9), and cycles between at least two distinct conformational states during its reaction cycle (10,11). The primary amino acid sequence of the H+-ATPase has been deduced from the DNA sequence of its structural gene ( P M A l ) and shown to share regions of direct homology with the Ca2+-ATPase, Na+,K'-ATPase and H'-K'-ATPase of animal cells Stains of yeast carrying mutations in PMAl are of considerable importance because they may provide valuable information about specific amino acid residues and/or protein structure domains participating in ATP hydrolysis, proton transport, and energy coupling. The cloning of P M A l (12), along with the development of a suitable gene expression system (15), has now made it possible to apply site-directed mutagenesis techniques to produce targeted amino acid substitutions. With this approach, Serrano and colleagues (16) have started to map different functional domains of the H+-ATPase.
In contrast to a targeted mutagenesis approach, our approach has been to isolate random mutants with the prospect that mutations will arise in different regions of PMAl and alter separate functional properties of the enzyme. Random mutagenesis, unlike site-directed mutagenesis, requires no assumptions to be made about essential residues. Recently, we described the isolation of a large number of UV-induced pmal mutations from S. cereuisiae by selecting for resistance to the aminoglycoside antibiotic hygromycin B (17). These mutants exhibited a variety of phenotypes including growth sensitivity to weak acids, low pH medium (pH < 3.5) and NH:. However, one of the most interesting properties of these mutants was a generalized depolarization of cellular membrane potential which was postulated to account for the cellular resistance of pmal mutants to hygromycin B (18).
Since the H'-ATPase is primarily responsible for the highly hyperpolarized membrane potential state in fungi (19,20), it is of considerable interest to understand how specific genetic alterations influence catalysis and hence, enzyme function. Preliminary evidence indicated that the kinetic properties of the H+-ATPase were altered (17).

EXPERIMENTAL PROCEDURES
Materials-All culture media supplies were from Difco. Trypsin, trypsin inhibitor, acrylamide (99.9%), bisacrylamide, MEGA-8,' and octyl glucoside detergents were from Boehringer Mannheim. lZ5I-Protein A (30 pCi/mg) was from Amersham Corp. Crude phosphatidylserine was obtained as a Folch Fraction I11 extract from Sigma. Acetone/ether-washed asolectin was from Avanti Polar Lipids. Deoxycholate was purchased from Kodak. All restriction endonucleases were purchased from New England Biolabs. Oligonucleotides used for sequencing were made on a Cyclone DNA Synthesizer (Milligen/ Biosearch).
Yeast Strains and Cell Culture-All pmal mutant strains of yeast used in this study were derived from parental wild-type strain Y55 (HO gal3 MALI SUCI) as described by McCusker et al. (17). Cells were grown in 10-liter batches of yeast extract peptone dextrose at 30 "C until mid-log phase and harvested by centrifugation, as described previously (11).
Cloning Mutant Alleles of PMAI-Mutant alleles of pmal-105, pmal-141,pmal-ll4,pmaI-147,pmal-155, and the wild-type PMAI from S. cereuisiae strain Y55 were cloned by gap repair (21). Plasmid YCp50-PMA1 containing a 5-kilobase pair HindIII fragment including the PMAl gene (12) was obtained from G. Fink (Whitehead Institute for Biomedical Research, Cambridge, MA). The coding region of PMAl was removed from this plasmid by partial digestion with KpnI and complete digestion with XbaI. The ends of the 8.4kilobase pair fragment were made blunt with Klenow fragment and deoxynucleoside triphosphates, and XbaI linkers were added before ligation to create plasmid pSH10. When pSHlO was linearized by restriction digestion with XbaI and transformed into u r d -I , pmal strains, the free ends of PMAl flanking region DNA invaded homologous chromosomal DNA, served as primers for repair synthesis to repair the gap and produced an autonomously replicating repaired plasmid containing a copy of the mutant allele.
Mapping the Site of Mutation within PMAl-Restriction enzyme digests of clones pmal-I05 and pmal-141 indicated that the clones contained a new EcoRI restriction site at the same position. The novel EcoRI site thus mapped the mutations in these genes very precisely. The region surrounding the new site was sequenced by the dideoxy method (22) with modifications for sequencing doublestranded template with Sequenase according to the suggestions of the manufacturer (United States Biochemical Corp.). Mutations in the remaining alleles were mapped using a modification of the single step gene replacement technique (23). Mutant strains were transformed with DNA fragments containing portions of the wild-type PMAI gene, and the ability of these fragments to rescue the mutant phenotypes was measured. URA3 present in the 3"nontranslated portion of PMAI enabled chromosomal integration of these fragments to be selected ( Fig. 1). Plasmid pSHll was digested with the restriction endonuclease HindIII (which yields a 6-kilobase pair URA3-containing fragment that includes the entire PMAl gene) or with HindIII and a second restriction endonuclease with a recognition site within PMAI coding sequence. This produced a set of DNA fragments containing progressively smaller portions of PMAI. By assessing which fragments rescued the mutants, the search for the position of the mutations was narrowed sufficiently so that only a small portion of each gene had to be sequenced.
Plasma Membrane Isolation-Plasma membranes were isolated by passing 75 g of cells resuspended in 300 ml of homogenization buffer (50 mM Tris-HC1, pH 7.5, 0.3 M sucrose, 5 mM Na2EDTA, 1 mM EGTA, 5 mg/ml bovine serum albumin, 2 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 2.5 pg/ml chymostatin) through a French pressure cell at 20,000 p.s.i. The lysate was adjusted to pH 7.25 with 1 M Tris and then centrifuged at 3500 X g for 5 min. The supernatant was removed and centrifuged at 14,000 X g for 20 min. The supernatant was again removed and then centrifuged at 300,000 X g for 1 h. The pellet from this high speed spin was resuspended in 100 ml of membrane wash buffer (10 mM Tris-HC1, pH 7.0, 1 mM EGTA, and 10% glycerol (w/v)) plus 0. fluoride, homogenized vigorously with a glass homogenizer and tightfitting teflon plunger and centrifuged at 300,000 X g for 1 h. The pellet was resuspended in membrane wash buffer at 10 mg/ml and frozen at -80 "C. When being prepared for enzyme purification, plasma membranes were washed, as above, in membrane wash buffer containing 1 mM phenylmethylsulfonyl fluoride and 2 mM DTT. These membranes were stored in buffer containing 10 mM Tris-HC1, pH 7.0, 1 mM EGTA, 25% (v/v) glycerol, and 2 mM DTT. All preparative operations were performed at 4 "C. The final pellet was highly enriched with plasma membranes and contained less than 1% mitochondrial (azide-sensitive) ATPase contamination.
Purification and Reconstitution of H+-ATPase-A modification of the enzyme purification procedure of Koland and Hammes (24) was used to isolate normal and mutant forms of the H+-ATPase. This procedure was developed to address problems of enzyme instability and excessively proton leaky reconstituted vesicles. Yeast plasma membranes were resuspended at 2 mg/ml in a solubilization buffer consisting of 10 mM HEPES-KOH, pH 7.2, 0.1 M KC1, 45% glycerol, 0.2 mM EDTA, 1 mM DTT. Crude phosphatidylserine and acetone/ ether washed asolectin were added to this suspension at a final concentration of 1 mg/ml. The suspension was stirred slowly at 4 'C and deoxycholate was added slowly to a final concentration of 0.5% (w/v). The clarified suspension was centrifuged at 303,000 X g for 75 min. The upper translucent layer of the pellet was removed, resuspended at 2 mg/ml in solubilization buffer containing 0.3 M KC1, and then re-extracted with 0.4% (w/v) deoxycholate, as before. The suspension was centrifuged at 303,000 X g for 75 min. The upper translucent portion of the pellet was again removed and resuspended in solubilization buffer containing 1.0 M KC1. The suspension was centrifuged as above for 90 min. The KC1-washed pellet was resuspended at 1 mg/ml in solubilization buffer containing 2 mg/ml phosphatidylserine and asolectin and then extracted with 1.5% MEGA-8 detergent. The suspension was centrifuged at 198,000 X g for 90 min. The pellet was resuspended at 1 mg/ml in solubilization buffer containing 3 mg/ml phosphatidylserine and 1.5 mg/ml asolectin and used immediately for reconstitution.
Reconstitution was performed by a modification of the method described by Newman and Wilson (25). Octyl glucoside was added to the resuspended MEGA-8 washed pellet, above, at a final concentration of 0.8% (w/v), and the suspension was mixed by gentle vortexing. The suspension was centrifuged at 198,000 X g for 60 min, and the supernatant was removed. 1 ml of this fraction was added to 1 ml of solubilization buffer, 0.57 ml of 50 mg/ml sonicated asolectin, and 0.06 ml of 50 mg/ml phosphatidylserine. (For mutant enzymes, 1.5 ml of octyl glucoside-extracted supernatant and 0.5 ml of solubilization buffer were used.) The suspension was adjusted to 0.8% (w/v) octyl glucoside, placed on ice for 5 min, and then diluted to 25 ml with buffer containing 10 mM HEPES-KOH, pH 6.8, 300 mM KCl, and 1 mM DTT. The proteoliposomes formed spontaneously and were collected by centrifugation at 303,000 X g for 60 min. The reconstituted liposomes were washed by resuspension with dilution buffer and then centrifuged at 303,000 X g for 60 min. The final pellet was resuspended in 150 pl of the dilution buffer. SDS gel electrophoresis was used to verify that the reconstituted enzymes represented 80% or more of the total reconstituted protein and remained intact following incorporation into liposomes.
Abundance ofH+-ATPase in Plasma Membranes-The appearance of intact H+-ATPase in plasma membrane fractions was evaluated by SDS gel electrophoresis and Western blot techniques as described previously (11). The relative amount of intact enzyme was determined by scanning Coomassie Blue-stained SDS gels with a video densitometer (Bio-Rad 620). The digitized data was analyzed with an IBM PC-AT using a one-dimensional analysis program (Bio-Rad Corp.). The relative abundance of H'-ATPase was calculated by comparing total area for the intact, M, = 100,000, protein with total area represented by all membrane proteins.
Protection from Trypsin-induced Inactiuation-Protection studies were performed essentially as described by Perlin and Brown (11). A 7.0,5 mM MgSO,, 5 mM ATP, and Na3V04 or 20 mM HEPES-KOH, 1-ml reaction medium consisted of either 20 mM HEPES-KOH, pH pH 7.0, 5 mM MgSO,, and 150 mM KH2P04. Both reactions were initiated by the addition of 250 pg of trypsin and were allowed to proceed for 1-20 min (VO, protection) or 10 min (Pi protection) at 30 "C. The reactions were terminated by the addition of 2.5 mg/ml trypsin inhibitor and 5 mM EDTA. The P,-protected samples were washed as described previously (11) to remove excess P,. All treated membranes were then assayed for ATP hydrolysis.
ATP Hydrolysis Measurements-Routine ATP hydrolysis assays were performed at 30 "C in a 1-ml reaction medium containing 10 mM MES-Tris, pH 6.5,5 mM MgSO,, 5 mM ATP, and 25 mM NH4C1. Purified enzyme was assayed in the same medium but at pH 6.8 and in the presence of 0.5 mg/ml asolectin. Inorganic phosphate liberation from ATP was determined by the method of Ames (27). Kinetic determinations of K, and V, , , were performed as described by Bowman and Slayman (9) except that 50 mM MES-Tris buffer was used to maintain pH between pH 5.5 and 8.0. Assays were arranged to limit ATP hydrolysis to 5% of total ATP. by a modified Lowry assay (28) and by the Amido Black method (29) Protein Determinations-Protein determinations were performed which overcomes problems due to phospholipid and detergent interference.

RESULTS
Sequence of Wild-type PMAl Allele from Strain Y55"The wild-type allele of PMAl from strain Y55 was cloned by gap repair and sequenced in its entirety. We expected that sequence differences would exist between the published sequence from strain S288C (12) and the wild-type allele from Y55, the strain background in which all our hygromycin Bresistant mutants were isolated (17). Nine single nucleotide substitutions were found which resulted in 6 amino acid changes: Pro-74 + Leu, Val-209 + Ile, Lys-444 * Met, Ser-479 + Phe, Ala-480 + Val, and Ala-836 + Ser. A number of these amino acid changes are apparently nonconservative, but generally occur in positions not strongly conserved within the family of ATPases having phosphorylated intermediates (12).
Appearance of H+-ATPase in pmal Mutant Membranes-Plasma membranes derived from log-phasepmal mutant cells were evaluated by SDS gel electrophoresis (Fig. 2 A ) . Analysis of the protein profiles by video densitometry indicated that the H'-ATPase, as indicated by the appearance of the M, = 100,000 band, represented 9.1% of the total plasma membrane protein in wild-type, pmal-105 and pmal-147 membranes, 8.3% in pmal-114 membranes, and 5.6% in pmal-155 membranes. These relative proportions of enzyme have now been observed in more than nine different membrane preparations of each pmal mutant. A Western blot analysis (Fig. 2B) was performed with a specific anti-H+-ATPase antibody capable of detecting 20 or more proteolytic cleavage products (11). T o enhance the detection of proteolysis products, the portion of the blot containing mutant membranes was exposed nearly twice as long as the wild-type portion. The results confirm that the mutant enzymes are intact and, despite the presence of a single prominent breakdown product, show no signs of generalized proteolysis. The breakdown peptide apparent in pmal-105, pmal-114, and pmal-147 mutant membranes has also been observed in different preparations of wild-type membranes and is believed to be a product of the normal turnover pathway.* These results suggest that proteolysis is not likely to account for the reduced amount of H'-ATPase in pmal-155 membranes.
Kinetic Properties of pmal Mutant Enzymes-The kinetic properties of the pmal mutant enzymes are summarized in Fig. 3, A and B. Of the four mutants, only pmal-105 and pmal-147 mutant enzymes showed prominent decreases in K , and V,,,,,. The kinetic behavior of enzymes from pmal-114 andpmal-155 appeared most defective in Vmax, but these differences were diminished following normalization of these values for the amount of enzyme in the membrane relative to wild type. The same generalized kinetic properties found for these mutants were also observed in kinetic characterizations of nine other hygromycin B-resistant pmal mutant enzymes (data not shown).
The effect of pH on kinetic parameters K, and Vb,, was evaluated for enzymes from pmal-105, which is unable to grow at low pH, and pmal-155 which shows no growth inhi-* D. S. Perlin, unpublished data. The portion of the blot containing mutant membranes was exposed 1.8 X as long as wild type to allow detection of proteolytic breakdown products.
The arrows indicate the position of intact H+-ATPase ( M , = 100,000). . Over the pH range 5.5-8.0, there was little difference in kinetic behavior between enzyme from wild type and pmal-155. In contrast, enzyme from the low pH-sensitive mutant, pmal-105, showed a precipitous decline in both K, and Vmax at pH values below pH 6.5 (Fig.  4, A and B ) . These experiments were limited to a 2.5-unit pH range, since pH values above pH 8.0 and below pH 5.5 were strongly destabilizing in both wild-type and mutant enzymes (data not shown). Vanadate Sensitiuity of Mutant Enzymes-The sensitivity of wild-type and pmal mutant enzymes to inhibition of ATP hydrolysis by vanadate is shown in Fig. 5 . Enzymes from pmal-114 and pmal-155 were strongly inhibited by vanadate and showed inhibition profiles identical to that of wild type. In contrast, enzyme from pmal-105 was insensitive to vanadate, whereas enzyme from pmal-147 was only weakly sensitive to vanadate over a concentration range that produced 95% inhibition of wild-type enzyme. This result suggests that for pma-105 and pmal-147 enzymes, vanadate is unable to form a transition state complex which locks the enzyme in an E2 (E2-P or E2. P) conformational state. To examine this possibility, the ability of vanadate to induce a stable E2 conformational state that would confer protection from trypsin-induced inactivation was examined for the mutant enzymes. It was observed previously that trypsin treatment of wild-type H'-ATPase caused a rapid loss of ATP hydrolysis and this effect was strongly protected by the formation of a vanadate-induced E2 conformational state (11). Fig. 6 indicates that wild-type enzyme and vanadate-sensitive mutant enzyme pmal-114 were strongly protected from trypsin-induced inactivation, whereas vanadate-insensitive enzymes, pmal-105 andpmal-147, were poorly protected from trypsin treatment. The level of protection paralleled the degree of vanadate-insensitive ATP hydrolysis (Fig. 5 ) and supports the suggestion that vanadate insensitivity correlates with an inability to attain a stabilized E 2 (E2-P or E2. P) intermediate species.
Effects on Phosphate Binding-If vanadate acts via an interaction at a phosphate binding site, then it is possible that vanadate insensitivity results from a change in the affinity of this site for phosphate. A direct assessment of phosphate binding is not possible because the loose affinity of this site for Pi (K, = 177 mM (10)) precludes such a measurement.
indirect assay utilizing phosphate-induced protection from trypsin proteolysis (11). The results in Fig. 7 show that 150 mM Pi was able to induce protection from trypsin-induced inactivation more strongly in vandate-sensitive enzymes, wild-type and pmal-155, than in vanadate-insensitive enzymes, pmal-105 and pmal-147. This result suggests that a change in phosphate binding has occurred in the vanadateinsensitive enzymes.
Proton Transport by Mutant Enzymes-It was reported previously from measurements of whole cells that hygromycin B-resistant pmal mutants are depolarized in cellular membrane potential but maintain near normal levels of proton transport (18). Since the evaluation of proton transport in whole cells relied on measurements of glucose-induced medium acidification, which reflects the combined action of both the H+-ATPase and organic acid transport, it was important to independently verify that pmal mutant enzymes actively catalyze proton transport. To accomplish this objective, wildtype and mutant enzymes were purified and reconstituted in liposomes. The specific ATPase activities for reconstituted mutant enzymes pmal-105, pmal-147, and pmal-114 were 8.8, 11.7, and 10.0 pmol Pi mg" min"; wild-type specific activities were in the range 27.7-32.8 Fmol Pi mg" min". The relative purity of reconstituted enzymes, as determined from SDS gel electrophoresis, was greater than 80% for wildtype and mutant enzymes. Each mutant enzyme was reconstituted a t least two times from separate membrane preparations and the results, as illustrated below, were always consistent.
Mutant enzymespmaI-105,pmal-I47, andpmal-114 readily formed MgATP-dependent pH gradients in reconstituted vesicles, as measured by fluorescence quenching of the pH gradient probe acridine orange (Fig. 8, A-C). In this assay, reconstituted vesicles were suspended in buffer containing ATP and the fluorescent probe acridine orange. Enzymemediated proton transport was initiated by the addition of Mg2+ and assessed by a continuous measurement of relative fluorescence. The formation of interior acid pH gradients, as indicated by the quenching of acridine orange fluorescence, reached a steady state after several minutes. The rate and extent of pH gradient formation was always lower with the mutant enzymes relative to wild type and this was expected from their decreased rates of ATP hydrolysis (see above). However, it was observed in three separate reconstitution experiments that pmal-114 enzyme always formed smaller pH gradients (Fig. 8C) relative to the other mutant enzymes, despite having a similar rate of ATP hydrolysis (10 pmol Pi mg" min").
The possibility that this enzyme is partially uncoupled from ATP hydrolysis is currently being investigated by a detailed analysis of H+/ATP ~toichiometry.~ The addition of 10 p M vanadate during steady-state pH gradient formation had no effect on vanadate-insensitive enzymes, pmal-105 andpmal-147 (Fig. 8, A and B ) , but caused a rapid reversal of the pH gradient in vanadate-sensitive enzymes pmal-114 and wild type (Fig. 8C). In each case, the pH gradient was completely collapsed by NH,Cl. The H'-ATPase catalyzes electrogenic proton transport and it was found that pH gradient formation was optimal in the presence of valinomycin which eliminated interior positive membrane potential formation by allowing compensating charge movement from the K'-loaded vesicles. When the mutant enzymes were allowed to form transient membrane potentials by initiating proton transport in the absence of valinomycin, as illustrated for pmal-105 (Fig. 8A), there was a marked decline in their apparent rate of proton transport D. Seto-Young and D. S. Perlin, unpublished data. relative to wild type. A subsequent addition of valinomycin restored pH gradient formation to its optimal level. This behavior raises the possibility that the pmal mutant enzymes may have an altered sensitivity to membrane voltage.

DISCUSSION
Membrane Potential-altering Mutations-The results in this study demonstrate that mutations in PMAl which cause depolarization of cellular membrane potential (18) are widely distributed throughout the gene (Fig. 1). The mutant enzymes were present in near wild-type quantities in the membrane except for pmal-155 which showed 62% of the wild-type level. An A + G mutation at -39 base pairs within the promoter region of pmal-155 creates an AUG sequence at position -41 in the PMAl mRNA that should be the site of translational initiation. However, only a tripeptide should be produced since an UAA codon is found in frame at position -32. Previous experiments in yeast (30) have shown that only the first AUG is used to initiate translation and that there is little or no reinitiation at downstream AUG sites, even when the first AUG is followed by a termination signal. Our results suggest either that the first AUG is poorly recognized or that reinitiation does occur, as the level of PMAl protein is only reduced by about 40%. The failure to detect significant pro-teolytic breakdown products in the membrane (Fig. 2B) suggests that decreased translational efficiency of PMAl is likely to account for the reduced amount of enzyme detected.
The kinetic properties of pmal-105, pmal-114, pmal-147, and pmal-155 mutant enzymes are diverse (Figs. 3 and 4). However, a reduction in membrane-associated H+-ATPase activity (except for pmal-114 enzyme) appears to be a common feature and has been observed in kinetic analyses of nine other enzymes from hygromycin B-resistant pmal mutants.' A decrease in total membrane-associated H+-ATPase activity, resulting from either altered translational efficiency (pmal-155) or from direct protein structure modifications (pmal-105 and pmal-147), appears sufficient to alter the cellular membrane potential and hence confer growth resistance to hygromycin B. Capieaux et al. (31) reached a similar conclusion from a study where mutations within upstream UASR~C sequences, which control transcription of PMAl, decreased the level of H'-ATPase in the membrane and conferred hygromycin B resistance to cells.
A close correlation between membrane potential and membrane-associated H'-ATPase activity assumes that ATP hydrolysis and proton transport are strictly coupled and that decreased proton transport leads to depolarization of the cellular membrane potential. This assumption is supported by studies of Dio-9-resistantpmal mutants (32,33). However, we recently reported for pmal-105 andpmal-147 that proton transport, as deduced from whole cell medium acidification, was not significantly different from wild type (18) despite the fact that these mutants show reduced levels of membraneassociated ATPase activity (Fig. 3). Thus, predictions of net proton transport and steady-state membrane potential from measurements of membrane-associated H'-ATPase activities can be misleading. It would be best to measure the coupling efficiency between ATP hydrolysis and proton transport for each mutant enzyme and recognize that membrane potential defects may reflect inherent changes in the enzyme-mediated charge-transfer mechanism (18). In this respect, recent patchclamp studies of isolated membrane patches from pmal-105 have demonstrated an ATP-dependent activation of voltagegated K' conductance that may reflect a direct participation of the H'-ATPase in K' transport (34). A detailed analysis of coupling (H'/ATP) stoichiometry, membrane voltage effects, and ion specificity by the mutant enzymes is currently in progress.
Vanadate-insensitive Enzymes-The inhibition of ATP hydrolysis by vanadate in P-type ATPases is generally considered a major characteristic feature of this class of enzyme. Vanadate is believed to bind at the site from which phosphate is released and enzyme inhibition results from the formation of a pentacoordinate complex at the site of phosphorylation (35)(36)(37). The importance of vanadate as a mechanistic probe suggests that mutants with altered vanadate sensitivity should provide important information about enzyme mechanism. In this study, two mutant H'-ATPases with amino acid substitutions of Ser-368 + Phe (pmal-105) and Pro-640 +-Leu (pmal-147) were found to be strongly resistant to inhibition by vanadate (Fig. 5 ) . Both amino acids are contained within conserved regions of the large hydrophilic domain which is believed to form all or part of the catalytic site (38). Vanadateinsensitive H'-ATPases have also been identified in Dio-9resistant pmal mutants from S. cerevisiae and S. pombe (32). The S. pombe mutant has been identified as a Gly-268 + Asp substitution (33) within a highly conserved region near the end of the second putative hydrophilic domain (Fig. 1). This domain is believed to be separated from the large hydrophilic domain by two membrane-spanning helices (38, 39).
On the basis of site-directed mutations, Glu-233 + Gln and Asp-200 + Asn, which influence the level of phosphorylated intermediate at Asp-378, Serrano (16,38) speculated that the two hydrophilic regions encompassing vanadate-insensitive loci Gly-268 and Ser-368 interact to form a "phosphatase" domain that would function in the hydrolysis of the aspartyl phosphate intermediate. However, the conferral of vanadate insensitivity is not localized to this domain since Pro-640, which also confers vanadate-insensitive enzyme behavior, lies within the putative ATP binding site (16). In addition, the relative position of an amino acid within a specific protein structure domain cannot account for vanadate insensitivity, since Asp-638 + Asn and Asp-378 + Asn, Glu, and Thr mutations which reduce ATP hydrolysis and lie close to vanadate-insensitive loci have no effect on vanadate sensitivity (16). Although the importance of Gly-268, Ser-368, and Pro-640 in vanadate-insensitive enzyme behavior cannot yet be defined precisely, it appears likely that they may play an essential role, either directly or indirectly, in phosphate binding.
In these pmal mutants, vanadate interactions at the phosphate-binding site may have been altered by a structural defect within the site or from a decrease in the steady-state level of EZ conformational intermediate necessary for Pi binding (33,36,39). Such behavior was supported by a decrease in phosphate-induced protection from trypsin-induced inactivation for vanadate-insensitive enzymes, pmal-105 and pmal-147 (Fig. 7 ) .
pH Dependence of Enzyme Activity-As shown in Fig. 4, the kinetic properties of wild-type enzyme and pmal-105 mutant enzyme vary greatly with pH. Wild-type enzyme shows a relatively constant K, and changing V, , , over the pH range 5.5-8.0. In contrast, pmal-105, which is unable to grow at low pH (17), shows a significant decline in K, and V, , , over this same range (Fig. 4, A and B). This kinetic behavior may be explained by proposing that a Ser-368 +-Phe mutation in pmal-105 results in exposure of a normally buried residue with an acidic pK,, such as glutamate or aspartate. Ionization of this residue at acidic pH values would inhibit the enzyme by altering the steady-state distribution of El and Ez catalytic intermediates. Acidic pH (below pH 6.0) has been proposed to stabilize a low affinity ATP conformation of the Ca'+-ATPase (40) and Na+,K'-ATPase (41) and such an effect may be mimicked in thepmal-105 enzyme.
Conclusion-The detailed genetic and biochemical characterizations of pmal mutant enzymes, pmal-105, pmal-114, pmal-147, and pmal-155, in this study confirm that hygromycin B is a valuable selection tool for isolating random and diverse pmal mutants. Unlike more popular site-directed mutagenesis approaches, random mutagenesis requires no assumptions to be made about essential residues. The major potential advantage of this approach is that numerous and diverse pmal mutants can be isolated which affect different partial catalytic reactions. Finally, through a detailed examination of proton transport by pmal mutant enzymes and a comprehensive revertant analysis of vanadate-insensitive pmal mutants, we hope to better define electrogenic proton transport and ATP hydrolysis by the H'-ATPase.