Transmembrane Topography of the 100-kDa a Subunit (Vph1p) of the Yeast Vacuolar Proton-translocating ATPase*

The membrane topography of the yeast vacuolar proton-translocating ATPase a subunit (Vph1p) has been investigated using cysteine-scanning mutagenesis. A Cys-less form of Vph1p lacking the seven endogenous cysteines was constructed and shown to have 80% of wild type activity. Single cysteine residues were introduced at 13 sites within the Cys-less mutant, with 12 mutants showing greater than 70% of wild type activity. To evaluate their disposition with respect to the membrane, vacuoles were treated in the presence or absence of the impermeant sulfhydryl reagent 4-acetamido-4′-maleimidylstilbene-2,2′-disulfonic acid (AMS) followed by the membrane permeable sulfhydryl reagent 3-(N-maleimidylpropionyl) biocytin (MPB). Three of the 12 active cysteine mutants were not labeled by MPB. The mutants E3C, D89C, T161C, S266C, N447C, K450C, and S703C were labeled by MPB in an AMS-protectable manner, suggesting a cytoplasmic orientation, whereas G602C and S840C showed minimal protection by AMS, suggesting a lumenal orientation. Factor Xa cleavage sites were introduced at His-499, Leu-560, and Pro-606. Cleavage at 560 was observed in the absence of detergent, suggesting a cytoplasmic orientation for this site. Based on these results, we propose a model of the a subunit containing nine transmembrane segments, with the amino terminus facing the cytoplasm and the carboxyl terminus facing the lumen.

The V-ATPases from fungi, plants, and animals are structurally very similar and are composed of two functional domains (1). The V 1 domain is a peripheral complex of molecular mass of 570 kDa composed of eight different subunits of molecular mass 70 -14 kDa (subunits A-H) that is responsible for ATP hydrolysis. The V 0 domain is a 260-kDa integral complex composed of five subunits of molecular mass 100 -17 kDa (subunits a, d, c, cЈ, and cЉ) that is responsible for proton translocation. This structure is similar to that of the ATP synthases (or F-ATPases) that function in ATP synthesis in mitochondria, chloroplasts and bacteria (12)(13)(14)(15)(16)(17), and sequence homology between these classes of ATPase has been demonstrated for both the nucleotide binding subunits (A, B, ␣, and ␤) (18,19) and the proteolipid subunits (subunits c, cЈ, and cЉ) (20,21). No sequence homology has been identified for any of the remaining subunits.
The 100-kDa subunit of the V-ATPase is an integral membrane protein possessing an amino-terminal hydrophilic domain and a carboxyl-terminal hydrophobic domain containing multiple putative membrane spanning segments (22)(23)(24). In yeast, the 100-kDa subunit is encoded by two genes, VPH1 and STV1 (23,24). Vph1p is targeted to the central vacuole, whereas Stv1p is normally targeted to some other intracellular membrane, possibly Golgi or endosomes (24).
We have previously demonstrated using a combination of site-directed and random mutagenesis that Vph1p contains several buried charged residues, mutation of which significantly alters proton transport activity of the V-ATPase complex (25,26). These results have led us to suggest that the 100-kDa subunit is the V-ATPase homolog to the a subunit of the F-ATPases. Mutational studies have identified several critical buried charged residues in the last two transmembrane segments of the F-ATPase a subunit that are important for proton translocation (27)(28)(29)(30), although more recent studies have suggested that some of these residues can be replaced without complete loss of function (31).
Although considerable information has been obtained concerning the membrane topography of the F-ATPase a subunit (32,33), essentially no information has yet been reported concerning the folding of the V-ATPase a subunit. In the present study, we have employed a combination of cysteine-scanning mutagenesis and chemical modification together with introduction of factor Xa cleavage sites in putative loops to investigate the topographical arrangement of the V-ATPase a subunit.

EXPERIMENTAL PROCEDURES
Materials and Strains-Zymolyase 100T was obtained from Seikagaku America, Inc. Bafilomycin A1 was a kind gift from Dr. Karlheinz Altendorf (University of Osnabruck). Factor Xa protease and protease inhibitors were from Roche Molecular Biochemicals. 9-Amino-6-chloro-2-methoxyacridine (ACMA), 3-(N-maleimidylpropionyl) biocytin (MPB), and 4-acetamido-4Ј-maleimidylstilbene-2,2Ј-disulfonic acid (AMS) were purchased from Molecular Probes. The monoclonal antibody 10D7 * This work was supported by National Institutes of Health Grant GM34478. Fluorescence facilities were provided by National Institutes of Health Grant DK 34928. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
against the yeast V-ATPase a subunit (34) was also from Molecular Probes. NeutrAvidin and immunoblotting reagents were obtained from Pierce. Escherichia coli and yeast culture media were purchased from Difco Laboratories. Restriction endonucleases, T4 DNA ligase and other molecular biology reagents were from Life Technologies, Inc., Promega, and New England Biolabs. ATP, phenylmethylsulfonyl fluoride, and most other chemicals were purchased from Sigma.
Mutagenesis-The Cys-less Vph1p was generated by substitution of all seven endogenous cysteine residues with serine. Single cysteine mutants were then constructed on the Cys-less Vph1p background. Mutations were introduced into either the wild type VPH1 or the Cys-less VPH1 using the Altered Sites II in vitro mutagenesis system (Promega) following the manufacturer's protocol. The Cys-less VPH1 and the single cysteine mutants were then confirmed by DNA sequencing.
The factor Xa cleavage site corresponding to the sequence IEGR was introduced immediately amino-terminal to His-499 and Pro-606, and a tandem pair of sites was introduced immediately amino-terminal to Leu-560. Mutations were introduced into the wild type VPH1 using the Altered Sites II system and confirmed by DNA sequencing.
Transformation and Selection-Yeast cells (MM112) lacking functional endogenous Vph1p and Stv1p were transformed using the lithium acetate method (36) with the wild type VPH1 in pRS316 as a positive control, pRS316 vector alone as a negative control, the Cys-less form of VPH1, each of the single cysteine mutants of VPH1, and the mutants bearing the factor Xa cleavage sites also in pRS316. The transformants were selected on uracil minus (Ura Ϫ ) plates as described previously (37). Growth phenotypes of the mutants were assessed on yeast extract-peptone-dextrose plates buffered with 50 mM KH 2 PO 4 or 50 mM succinic acid to either pH 7.5 or pH 5.5. For testing calcium sensitivity, yeast extract-peptone-dextrose plates buffered with 50 mM MES and 50 mM MOPS, pH 7.5, were supplemented with 50 mM calcium chloride.
Isolation of Vacuolar Membrane Vesicles-Vacuolar membrane vesicles were isolated as described previously (25). For the labeling experiments, vacuolar membranes enriched in V 0 domains were prepared by inclusion of a glucose starvation step after spheroplasting using the protocol of Kane (38). This allowed immunoprecipitation of the V 0 domain directly using the monoclonal antibody 10D7 (which recognizes the a subunit in the V 0 domain but not in the intact V 1 V 0 complex (34)) and reduced variability due to dissociation of the V-ATPase complex. Briefly, after the spheroplasting step, the spheroplasts were incubated in the same medium but without glucose for 20 min at 30°C. The subsequent steps were the same as those employed in the normal vacuolar membrane vesicle preparation, as described previously (25).
Chemical Labeling and Blocking-The vacuolar membrane vesicles were collected by centrifugation at 100,000 ϫ g for 10 min at 4°C and washed using labeling buffer (10 mM Tris-Mes (pH 7.0), 0.25 mM MgCl 2 , and 1.1 M glycerol). Membrane vesicles were then resuspended in 200 l of labeling buffer and divided into two tubes. Where indicated, 0.25% Zwittergent 3-14 was added to permeabilize the vesicles. AMS (100 M) was added to one tube, and both samples were incubated for 5 min at 10°C. Samples were then transferred to ice and diluted 5-fold with labeling buffer followed immediately by addition of 250 M MPB and incubation for 15 min at 25°C. Dilution was employed rather than washing to stop reaction with AMS to avoid variability due to losses associated with an additional pelleting step. The labeling reaction was then stopped by addition of 15 mM 2-mercaptoethanol.
Detergent Solubilization, Immunoprecipitation, and Detection of MPB Labeling of the a Subunit-After MPB labeling, vesicles were pelleted by centrifugation at 100,000 ϫ g for 15 min and solubilized in ice-cold phosphate-buffered saline containing 1% C 12 E 9 , and the V 0 domain was immunoprecipitated using the mouse monoclonal antibody 10D7 specific for Vph1p plus protein A-Sepharose. In cases where glucose starvation was not employed after spheroplasting, immunoprecipitation was carried out using the monoclonal antibody 8B1-F3 (Molecular Probes) specific for the 70-kDa A subunit, which precipitates both V 1 and the intact V-ATPase complex. Samples were then subjected to SDS-PAGE on 10% acrylamide gels and transferred to nitrocellulose membranes for MPB detection as described below.
For MPB detection, the blots were probed with horseradish peroxidase-conjugated NeutrAvidin and developed using the Supersignal UL-TRA chemiluminescent system (Pierce). After MPB detection, blots were stripped and reprobed with the mouse monoclonal antibody 10D7 against the 100-kDa subunit to ensure the presence of equal amounts of the a subunit. For immunoblot analysis, original blots or stripped blots were probed with 10D7, followed by horseradish peroxidase-conjugated secondary antibody (Bio-Rad) as described previously (25). Blots were developed using a chemiluminescent detection method obtained from Kirkegaard & Perry Laboratories.
Factor Xa Cleavage of Wild Type and Vph1p Mutants Bearing Factor Xa Cleavage Sites-Vacuolar membrane vesicles (50 g protein) prepared from cells transformed with wild type VPH1 or VPH1 mutants bearing the factor Xa sites as described above were pelleted and resuspended in buffer containing 400 mM potassium iodide, 1.6 mM MgATP, and 20 mM Tris-Mes (pH 7.0) and incubated on ice for 60 min. The purpose of the potassium iodide treatment was to remove the V 1 domain, thus exposing any factor Xa sites that may be shielded by the presence of V 1 . Membrane vesicles were then sedimented at 100,000 ϫ g for 45 min and resuspended in factor Xa buffer (100 mM NaCl, 1 mM CaCl 2 , 1 M glycerol, 50 mM Tris-HCl (pH 7.4), 2 mM dithiothreitol). The resuspended vesicles were divided into three aliquots. One received no factor Xa, whereas two received 2 g of factor Xa protease. One of the samples receiving the factor Xa also received 1% C 12 E 9 . All samples were incubated overnight at 4°C followed by addition of Laemmli sample buffer and separation of half of each sample on a 9% polyacrylamide gel. The proteins were then blotted to nitrocellulose and probed with the monoclonal antibody 10D7 against the 100-kDa subunit followed by horseradish peroxidase-conjugated secondary antibody (Bio-Rad). We have observed that the antibody 10D7 recognizes a site in the amino-terminal half of the protein, 2 so that the size of the fragments generated by factor Xa cleavage that would be recognized by 10D7 could be predicted. Blots were developed using a chemiluminescent detection method obtained from Kirkegaard and Perry Laboratories.
Other Procedures-Protein concentrations were determined by the method of Lowry et al. (39). ATP-dependent proton transport was measured in transport buffer (25 mM MES/Tris, pH 7.2, 5 mM MgCl 2 ) using the fluorescence probe ACMA in the presence or absence of 10 nM bafilomycin A 1 as described previously (40). SDS-polyacrylamide gel electrophoresis was carried out as described by Laemmli (41).

Construction of the Cys-less Form of Vph1p
-To probe the membrane topology of the 100-kDa a subunit of the V-ATPase using a cysteine modification approach, we first needed to construct a Cys-less form of Vph1p in which all seven of the endogenous cysteine residues were replaced. Site-directed mutagenesis of VPH1 was performed to replace each of the endogenous cysteine residues with serine. The final construct was confirmed by DNA sequencing. Transformation of a yeast strain in which both VPH1 and STV1 are disrupted with the Cys-less form of Vph1p gave rise to a strain showing a wild type growth phenotype. That is, cells were able to grow at both pH 5.5 and 7.5, even in the presence of elevated Ca 2ϩ . This is in contrast to the deletion strain that displays a vma Ϫ phenotype (i.e. unable to grow at pH 7.5 or in the presence of elevated Ca 2ϩ (5,42,43)). Isolation of vacuoles from the wild type and Cys-less strains and measurement of bafilomycin-sensitive ATPdependent proton transport revealed that the Cys-less form of Vph1p gave rise to a V-ATPase complex that possessed approximately 78 Ϯ 16% of wild type levels of transport activity (Table  I). Thus, the Cys-less Vph1p represents an appropriate genetic background in which to construct mutants bearing unique cysteine residues.
Vph1p Mutants Bearing Single, Unique Cysteine Residues-Unique cysteine residues were introduced into the Cys-less form of Vph1p at positions 3, 89, 161, 266, 447, 450, 561, 564, 602, 703, 761, 763, and 840 by site-directed mutagenesis, and the mutations were confirmed by DNA sequencing. The mutant forms of Vph1p were then expressed in the deletion strain, and the growth phenotype was analyzed. All mutants displayed a wild type growth phenotype. Isolation of vacuoles and measurement of bafilomycin-sensitive ATP-dependent proton transport (Table I) revealed that all of the single cysteine mutants except G763C displayed greater than 70% of wild type levels of activity, with G763C having approximately 50% of wild type activity. Western blot analysis of isolated vacuoles using a monoclonal antibody specific for Vph1p (10D7) showed that all the mutants expressed nearly wild type levels of Vph1p (data not shown). These results suggest that, except for G763C, none of the introduced cysteine residues significantly perturb the structure or function of Vph1p.
Labeling and Protection of Cysteine Mutants of Vph1p-To assess the orientation of the cysteine residues introduced into Vph1p with respect to the membrane, modification by two sulfhydryl reagents was employed. The first reagent, MPB, is a membrane permeable reagent that can react with sulfhydryl groups exposed on both sides of the vacuolar membrane (33,44). The second reagent, AMS, is a membrane impermeant reagent that should react only with sulfhydryl groups on the exposed cytoplasmic surface of the vacuolar membrane (32,33,44).
Previous studies indicate that intact vacuoles isolated from yeast are both uniformly sided with the cytoplasmic side exposed and sealed with respect to small molecules (45). The specific activity of the V-ATPase in the vacuolar membrane vesicles employed in the current study (3-4 mol of ATP/ min/mg of protein) is very similar to the values we have measured in intact vacuoles (46), suggesting that very little of the V-ATPase is oriented with the cytoplasmic face sequestered inside the vesicles. Moreover, the vacuolar membrane vesicles used here display a ratio of proton transport (as measured by ACMA uptake) to bafilomycin-sensitive ATPase activity that is nearly the same as that we measure for either intact vacuoles or purified, bovine brain clathrin-coated vesicles. We have previously shown that coated vesicles are both tightly sealed and uniformly oriented with the cytoplasmic surface exposed on the outside (47). These results suggest that the vacuolar membrane vesicles employed here have the same properties.
Labeling of cysteine residues exposed on the cytoplasmic surface of Vph1p by MPB should be blocked by prior treatment with the impermeant reagent AMS. By contrast, cysteine residues exposed on the lumenal surface of the protein should show labeling by MPB that is unchanged by pretreatment with AMS. This strategy has been successfully employed to determine the orientation of cysteine residues in several membrane proteins, including MDR (44) and subunit a of the F-ATPase (32,33).
Labeling of Vph1p by MPB is measured by detergent solubilization, immunoprecipitation, SDS-PAGE, and Western blotting using horseradish peroxidase-conjugated avidin. To avoid possible protection of cysteine residues by the V 1 domain, immunoprecipitation of the V 0 domain was carried out using the monoclonal antibody 10D7 specific for Vph1p. Because 10D7 recognizes the a subunit only in the free V 0 domain (not the intact V 1 V 0 complex (34)), it is necessary to optimize the fraction of V 0 in the free state. This was accomplished by including a brief glucose starvation step after spheroplasting of the yeast, a procedure that Kane (38) has shown causes an increase in the fraction of free V 0 domain in the vacuolar membrane from approximately 25% to 75%.
Isolated vacuoles were first incubated in the presence or absence of 100 M AMS for 5 min at 10°C. These conditions were selected to allow reaction of AMS with cytoplasmically oriented cysteine residues but to minimize its permeation across the membrane and are similar to the conditions previously employed in related studies (32,33). Vacuoles were then diluted 5-fold and reacted with 250 M MPB for 15 min at 25°C. Dilution rather than washing was employed to stop reaction with AMS to eliminate additional variability in yield associated with pelleting and resuspending the membranes. The reaction was then stopped by addition of 15 mM 2-mercaptoethanol and the V-ATPase solubilized and immunoprecipitated as described above.
Preliminary results indicated that four of the introduced cysteine residues at positions 561, 564, 761, and 763 were not labeled by MPB, and these mutants were therefore not further pursued. Fig. 1 shows the labeling of the a subunit by MPB with and without prior treatment with AMS for each of the remaining nine cysteine mutants, as well as the Cys-less control. Fig. 1A shows the results obtained for the cysteine residues introduced into the amino and carboxyl-terminal ends of the a subunit, whereas Fig. 1B shows the results for cysteine residues introduced into the loop regions. Because labeling of G602C by MPB in the absence of AMS was reduced approximately 3-fold (by densitometry) relative to the other loop cysteine residues, a longer exposure was employed for this mutant so that the intensity of the band in the absence of AMS would more closely approximate the other cysteine mutants. This facilitates a more accurate comparison of their relative labeling in the presence and absence of AMS. Western blotting using the antibody 10D7 indicated that the reduced labeling of G602C by MPB is not due to a reduced amount of the a subunit (data not shown).
As can be seen in Fig. 1, seven of the nine cysteine residues (E3C, D89C, T161C, S266C, N447C, K450C, and S703C) showed very effective blocking of MPB labeling by pretreatment with AMS, suggesting that these residues are exposed on the cytoplasmic surface of the protein. By contrast, two of the nine cysteine residues (G602C and S840C) showed very little protection of MPB labeling by AMS, suggesting that these residues are exposed on the lumenal side of the membrane. These results were reproducible even when the temperature used in blocking by AMS was raised to 20°C. a ATP-dependent proton transport (as assessed by ACMA quenching) was measured on aliquots of purified vacuolar membrane vesicles containing 2 g of protein as described under "Experimental Procedures." Activities are expressed relative to the ⌬vph1 ⌬stv1 strain expressing the pRS316 plasmid containing the wild-type VPH1 gene (defined as 100%), which was completely inhibited by 10 nM bafilomycin. No ATPdependent ACMA quenching was observed in vacuolar membrane vesicles isolated from cells transformed with the vector alone. Representative values for each mutant are shown.
To test whether G602C and S840C are not protected by AMS due to the inability of the reagent to cross the membrane, reaction with AMS and labeling by MPB was repeated for these two mutants and one of the other cysteine mutants (S703C) in the presence of 0.25% Zwittergent 3-14 in order to allow access of the AMS to the lumenal side of the membrane. As can be seen in Fig. 2, the presence of the detergent allowed AMS to effectively block G602C and S840C from reaction with MPB, consistent with a lumenal orientation of these residues.
Analysis of Mutants Bearing Factor Xa Cleavage Sites-To further probe the topography of the 100-kDa subunit, factor Xa protease cleavage sites were introduced into three putative loops at positions His-499, Leu-560, and Pro-606. Table I shows the effect of introduction of the four amino acid site (IEGR) preceding His-499 and Pro-606 and the tandem eight amino acid sequence (IEGRIEGR) preceding Leu-560. As can be seen, only the single factor Xa site at position 606 significantly reduced proton transport activity (by approximately 55% relative to wild type), whereas the other two mutants showed near wild type levels of activity. When tested for cleavage by factor Xa in the absence and presence of detergent (1% C 12 E 9 ), only one of the sites (at position 560) showed significant cleavage. As can be seen in Fig. 3, treatment of the mutant bearing the tandem sites at position 560 with factor Xa resulted in the generation of a 65-kDa fragment in either the absence or presence of detergent, whereas no such fragment was observed for the wild type protein. These results suggest that the tandem sites at position 560 are exposed on the cytoplasmic side of the membrane. DISCUSSION We have used cysteine-scanning mutagenesis together with chemical modification by membrane permeable and impermeable reagents, as well as introduction of protease cleavage sites, to analyze the membrane topography of the 100-kDa a subunit of the yeast V-ATPase. The cysteine-scanning method depends upon the successful construction of a functionally active Cysless form of the protein together with mutants containing unique cysteine residues that do not significantly perturb ac-tivity. In the case of the a subunit, this required the simultaneous replacement of seven endogenous cysteine residues. In addition, because the a subunit is part of a multisubunit complex, the mutations introduced must not alter the interactions between the a subunit and the remaining V-ATPase subunits. We observed that both the Cys-less form of the a subunit and all but one of the single cysteine-containing mutants give rise to complexes showing near wild type levels of activity, suggesting that these changes do not significantly alter the structure or function of the a subunit. In the case of G763C, a loss of approximately 50% of proton transport activity was observed.
The labeling strategy employed depends upon the membrane permeability of MPB and the membrane impermeability of AMS. This allows MPB to react with cysteine residues on both sides of the membrane, whereas AMS reacts only with cysteine residues exposed on the surface of the membrane vesicles. In the case of yeast vacuolar membrane vesicles used in the present study, the exposed surface is the cytoplasmic side of the membrane. The orientation and sealed state of the vesicles employed is supported by the similar ratio of proton transport to ATP hydrolysis that we have observed when compared with other sealed, well oriented vesicles, including intact yeast vacuoles (46) and purified clathrin-coated vesicles (47). This strategy has been successfully used to study the membrane topography of several other integral membrane proteins, including in particular subunit a of the E. coli F-ATPase (32,33). These latter studies have led to the current five membrane spanning model for the F-ATPase a subunit.
In addition to the cysteine modification approach, we have also introduced factor Xa cleavage sites into three putative loop regions of the protein. Although the mutant bearing a factor Xa site at position 606 showed only approximately 50% of wild type levels of activity, the other two mutants showed nearly normal proton transport, suggesting minimal perturbation of structure.
The results obtained in the current study have led to the FIG. 1. Labeling of a subunit cysteine mutants by MPB and protection by AMS in the absence of detergent. A, vacuolar membrane vesicles (35 g of protein) isolated from the Cys-less mutant and the single cysteine mutants at the amino and carboxyl termini (E3C, D89C, T161C, S266C, and S840C) were incubated in the presence or absence of 100 M AMS for 5 min at 10°C, followed by 5-fold dilution and labeling with 250 M MPB for 15 min at 25°C. The membranes were then solubilized with C 12 E 9 , and the a subunit immunoprecipitated using the monoclonal antibody 10D7 and protein A-Sepharose. Samples were separated by SDS-PAGE, transferred to nitrocellulose, and probed with horseradish peroxidase-NeutrAvidin. Blots were developed using the Supersignal Ultra system from Pierce. B, vacuolar membrane vesicles (45 g of protein) isolated from the Cys-less mutant, and each of the a subunit mutants containing single cysteine residues in the loop regions (N447C, K450C, G602C, and S703C) were treated as described in A. Because G602C showed approximately 3-fold lower labeling by MPB (by densitometry) in the absence of AMS relative to the other loop cysteine residues, a longer exposure is shown for this mutant such that the intensity of the band in the absence of AMS is comparable to that observed for the other loop cysteine mutants. This allows a more direct comparison of the effect of AMS pretreatment on MPB labeling. FIG. 2. Labeling of G602C, S703C, and S840C mutants by MPB and protection by AMS in the presence of detergent. Vacuolar membrane vesicles (45 g protein) isolated from the G602C, S703C, and S840C mutants were incubated in the presence or absence of 100 M AMS followed by dilution and labeling with 250 M MPB as described in the legend to Fig. 1 except that 0.25% Zwittergent 3-14 (ZW) was added to the samples prior to addition of AMS where indicated. Solubilization with C 12 E 9 , immunoprecipitation of Vph1p, SDS-PAGE, and blotting were then performed as described in Fig. 1.   FIG. 3. Cleavage of a Vph1p mutant bearing tandem factor Xa sites at position 560 by factor Xa. Vacuolar membrane vesicles (50 g of protein) isolated from wild type (WT) cells or cells bearing a mutant form of Vph1p containing tandem factor Xa sites at position 560 (2fXa-560) were treated with potassium iodide and MgATP to remove the V 1 domain and then incubated in the absence or presence of C 12 E 9 (1.0%) and factor Xa (2 g of protein) as described under "Experimental Procedures." The samples were then separated by SDS-PAGE, transferred to nitrocellulose, and blotted with the monoclonal antibody 10D7 specific for Vph1p as described. The fragment observed for the Vph1p mutant bearing the tandem factor Xa sites at position 560 has a molecular mass of approximately 65 kDa, corresponding to the predicted molecular mass of the amino-terminal fragment generated by factor Xa cleavage at position 560. model for the folding of the V-ATPase a subunit shown in Fig.  4. In addition, a hydropathy plot for Vph1p together with the location of each of the putative membrane spanning segments is shown in Fig. 5. The protein is proposed to span the bilayer nine times, with the amino terminus on the cytoplasmic side of the membrane and the carboxyl terminus on the lumenal side. All four of the cysteine residues introduced into the aminoterminal soluble domain showed a labeling pattern consistent with a cytoplasmic orientation whereas S840C at the carboxyl terminus showed labeling characteristic of a lumenal orientation. This suggests that the amino and carboxyl termini are on opposite sides of the membrane, requiring that the protein span the bilayer an odd number of times.
In our original folding model for the a subunit (25), we had postulated that the amino-terminal soluble domain was exposed on the lumenal side of the membrane based upon the increased labeling observed for the bovine a subunit using membrane impermeant reagents when the membrane was disrupted with detergents (47). This result may be explained by a change in conformation of the protein in the presence of detergent such that previously shielded sites on the cytoplasmic side of the membrane become exposed. In addition, because the a subunit is not synthesized with an amino-terminal leader sequence, a cytoplasmic orientation for the amino terminus is more consistent with what is known concerning the requirements for translocation of hydrophilic protein segments across the endoplasmic reticulum membrane during biosynthesis.
A lumenal orientation of the carboxyl terminus is also at odds with our previously proposed model, in which we suggested a cytoplasmic orientation for the carboxyl terminus (26). This was based on the identification by random mutagenesis of a cluster of five mutations between residues Leu-800 and Gly-814 that disrupted attachment of the V 1 and V 0 domains, suggesting that this region may be important in assembly of the V-ATPase complex (26). It is nevertheless possible that disruption of assembly in these mutants results from conformational changes in the structure of the 100-kDa subunit or the V 0 domain that prevent attachment of V 1 to V 0 . Interestingly, our current model places nearly all of the residues that have been observed to disrupt assembly, including Asp-425, Lys-538, and Arg-735, as well as the five mutations between residues 800 and 814 on the lumenal side of the membrane. It is possible that these residues may form a lumenal domain that senses the intravesicular pH and conveys this information through conformational changes in the 100-kDa subunit to the cytoplasmic domain of the complex.
With respect to the loop regions, four of the introduced cysteine residues (at positions 561, 564, 761, and 763) are not labeled by MPB, suggesting that they are shielded from reaction by interaction with other V 0 subunits or other regions of the 100-kDa subunit itself. It is also possible that one or more of these cysteine residues may be shielded from reaction with MPB by interaction with the lipid bilayer. Of the remaining introduced cysteine residues, three appear to have a cytoplasmic orientation (N447C, N450C, and S703C) whereas G602C appears to be lumenal. The cytoplasmic orientation of residues 447 and 450 requires that the hydrophobic region between residues 405 and 443 span the bilayer twice rather than once (as originally postulated (25)), placing Asp-425 on the lumenal side of the membrane. A lumenal orientation for Gly-602 and a cytoplasmic orientation for Ser-703 is consistent with a single membrane span between these sites. That Ser-703 is cytoplas- We have also constructed several mutations at Arg-735, including R735K, R735L, and R735A, because an arginine near the carboxyl terminus of the F-ATPase a subunit appears to play a role in proton transport (29,31). All of the Arg-735 mutants show a defect in either stability of the a subunit or assembly of the V-ATPase complex (data not shown), so that we are not able to assess the role of this residue in proton transport by the V-ATPase. The site of trypsin cleavage of the bovine a subunit from the cytoplasmic side of the membrane (48) is also shown by an arrow. The borders of the putative transmembrane segments are indicated by the residue numbers. mic is also consistent with the observation that the bovine V-ATPase a subunit can be cleaved by low concentrations of trypsin at a site immediately upstream of this residue in intact coated vesicles (48). Thus, the large soluble loop between residues 653 and 727 is likely to be oriented toward the cytoplasm.
Of the three factor Xa sites introduced into the 100-kDa subunit, protease cleavage was observed only for the tandem sites introduced at position 560. The fact that this cleavage happened in intact vacuolar membrane vesicles as well as in the presence of detergent indicates that this site is exposed on the cytoplasmic side of the membrane. For the sites introduced at positions 499 and 606, no cleavage by factor Xa was observed in either the presence or absence of detergent, again suggesting that these sites may be shielded through interaction with other regions of the protein. Nevertheless, the observation that all three of these mutants show significant proton transport activity suggests that all three sites are located in loop regions between transmembrane segments in the a subunit. This conclusion comes from results obtained with lac permease, where it has been observed that factor Xa sites introduced into loops between transmembrane segments generally do not affect activity, whereas sites introduced within transmembrane segments invariably lead to loss of both correct folding and transport activity. 3 It is important to emphasize that the model shown in Fig. 4 is only a working model for the topography of the V-ATPase a subunit based upon the first available data presented here concerning the disposition of this polypeptide with respect to the membrane. The labeling and proteolysis approaches taken in the current study, like all other methods of analyzing the topography of membrane proteins, are subject to several pos-sible sources of error. For instance, it is possible that local differences in the environment of individual cysteine residues, in addition to their disposition with respect to the membrane, may influence their relative reactivity toward the two sulfhydryl reagents employed. We have shown that permeabilization of the membranes with detergent allows the lumenally oriented cysteine residues to be effectively blocked by AMS. Nevertheless, it will be important that the proposed model be further tested, both by introduction of additional labeling and cleavage sites and by employing other approaches to determine topography, such as epitope tagging and preparation of site-specific antibodies that recognize the native protein. The current study, however, represents a first step in determining the folding pattern of this important V-ATPase subunit.