VMA11 and VMA16 encode second and third proteolipid subunits of the Saccharomyces cerevisiae vacuolar membrane H+-ATPase.

The vacuolar membrane H+-ATPase (V-ATPase) of the yeast Saccharomyces cerevisiae is composed of peripheral catalytic (V1) and integral membrane (V0) domains. The 17-kDa proteolipid subunit (VMA3 gene product; Vma3p) is predicted to constitute at least part of the proton translocating pore of V0. Recently, two VMA3 homologues, VMA11 and VMA16 (PPA1), have been identified in yeast, and VMA11 has been shown to be required for the V-ATPase activity. Cells disrupted for the VMA16 gene displayed the same phenotypes as those lacking either Vma3p or Vma11p; the mutant cells lost V-ATPase activity and failed to assemble V-ATPase subunits onto the vacuolar membrane. Epitope-tagged Vma11p and Vma16p were detected on the vacuolar membrane by immunofluorescence microscopy. Density gradient fractionation of the solubilized vacuolar proteins demonstrated that the tagged proteins copurified with the V-ATPase complex. We conclude that Vma11p and Vma16p are essential subunits of the V-ATPase. Vma3p contains a conserved glutamic acid residue (Glu137) whose carboxyl side chain is predicted to be important for proton transport activity. Mutational analysis of Vma11p and Vma16p revealed that both proteins contain a glutamic acid residue (Vma11p Glu145 and Vma16p Glu108) functionally similar to Vma3p Glu137. These residues could only be functionally substituted by an aspartic acid residue, because other mutations we examined inactivated the enzyme activity. Assembly and vacuolar targeting of the enzyme complex was not inhibited by these mutations. These results suggest that the three proteolipid subunits have similar but not redundant functions, each of which is most likely involved in proton transport activity of the enzyme complex. Yeast cells contain V0 and V1 subcomplexes in the vacuolar membrane and in the cytosol, respectively, that can be assembled into the active V0V1 complex in vivo. Surprisingly, loss-of-function mutations of either Vma11p Glu145 or Vma16p Glu108 resulted in a higher degree of assembly of the V1 subunits onto the V0 subcomplex in the vacuolar membrane.

The vacuolar membrane H ؉ -ATPase (V-ATPase) of the yeast Saccharomyces cerevisiae is composed of peripheral catalytic (V 1 ) and integral membrane (V 0 ) domains. The 17-kDa proteolipid subunit (VMA3 gene product; Vma3p) is predicted to constitute at least part of the proton translocating pore of V 0 . Recently, two VMA3 homologues, VMA11 and VMA16 (PPA1), have been identified in yeast, and VMA11 has been shown to be required for the V-ATPase activity. Cells disrupted for the VMA16 gene displayed the same phenotypes as those lacking either Vma3p or Vma11p; the mutant cells lost V-ATPase activity and failed to assemble V-ATPase subunits onto the vacuolar membrane. Epitope-tagged Vma11p and Vma16p were detected on the vacuolar membrane by immunofluorescence microscopy. Density gradient fractionation of the solubilized vacuolar proteins demonstrated that the tagged proteins copurified with the V-ATPase complex. We conclude that Vma11p and Vma16p are essential subunits of the V-ATPase. Vma3p contains a conserved glutamic acid residue (Glu 137 ) whose carboxyl side chain is predicted to be important for proton transport activity. Mutational analysis of Vma11p and Vma16p revealed that both proteins contain a glutamic acid residue (Vma11p Glu 145 and Vma16p Glu 108 ) functionally similar to Vma3p Glu 137 . These residues could only be functionally substituted by an aspartic acid residue, because other mutations we examined inactivated the enzyme activity. Assembly and vacuolar targeting of the enzyme complex was not inhibited by these mutations. These results suggest that the three proteolipid subunits have similar but not redundant functions, each of which is most likely involved in proton transport activity of the enzyme complex. Yeast cells contain V 0 and V 1 subcomplexes in the vacuolar membrane and in the cytosol, respectively, that can be assembled into the active V 0 V 1 complex in vivo. Surprisingly, loss-of-function mutations of either Vma11p Glu 145 or Vma16p Glu 108 resulted in a higher degree of assembly of the V 1 subunits onto the V 0 subcomplex in the vacuolar membrane.
The V-type ATPase consists of two structural domains, V 1 and V 0 , both of which are composed of several different subunits (4,7,12). The peripheral V 1 domain possesses the nucleotide binding site(s) required for ATP hydrolysis (5,7,12,13). The integral V 0 domain translocates protons across the membrane and anchors the peripheral V 1 domain to the membrane (4,7,12). The yeast V-ATPase is composed of at least seven V 1 subunits (69, 60, 54,42,32,27, and 14 kDa) and three V 0 subunits (100, 36, and 17 kDa) (14 -26). The 36-kDa subunit is not an integral membrane protein but stably associates with the V 0 subcomplex and thus is considered to be a nonintegral component of the V 0 domain (20,27). Recently, another essential subunit of 13 kDa was identified (28). The membrane disposition of this subunit has not yet been unambiguously determined.
Yeast cells also express genes that encode proteins with sequence similarity to the 100-kDa (Vph1p) and the 17-kDa (Vma3p) 2 V-ATPase subunits (29 -31). STV1 encodes a polypeptide of 102 kDa with 54% amino acid identity to Vph1p (14,29). Stv1p is localized to nonvacuolar membranes, possibly endosomal and/or Golgi membrane, and is not required for vacuolar membrane V-ATPase activity. Overexpression of Stv1p induces mislocalization of this protein to the vacuolar membrane and can partially complement the phenotypes associated with the loss of V-ATPase activity in ⌬vph1 mutant cells (29). From these observations, it has been proposed that yeast cells possess another V-type ATPase acidifying organelles (containing Stv1p) other than vacuoles and that the 100-kDa subunit isoforms are responsible for targeting of the V-type ATPases to different organelles (29).
VMA11 and VMA16 genes can encode polypeptides of 17 and 23 kDa with 56 and 35% amino acid identity to Vma3p, respectively (24,25,30,31). Vma3p is a hydrophobic polypeptide chemically characterized as a proteolipid (soluble in chloroform/methanol) and is thought to constitute all or part of the proton translocating pore in V 0 (5,24,25). Homologous subunits are found universally in V-type ATPase complexes characterized from various membrane sources, and all amino acid sequences determined for the subunits contain a conserved glutamic acid residue predicted to be critical for proton transport activity (32). VMA11 was cloned by complementation of the growth defect of a calcium sensitive mutant, cls9 (30, 33). Unlike STV1, VMA11 is required for the activity and assembly of the vacuolar membrane V-ATPase complex (30, 33). VMA16 was originally identified as PPA1, an open reading frame adjacent to the MAS2 gene (31). Initial gene disruption analysis demonstrated that Vma16p is important for cell growth (31), but the physiological function of this protein remained unclear.
In this work, the function of Vma11p and Vma16p was studied by examining the cellular localization of these proteins and by characterizing phenotypes of the cells carrying mutant forms of each protein. Our results indicate that Vma11p and Vma16p are novel V 0 subunits of the V-ATPase complex essential for enzyme activity. We also report here that both proteins contain a glutamic acid residue important for their function as found for Vma3p (34), and inactivating mutations of these residues influence the assembly status of the enzyme complex.
Disruption of the VMA11 Gene-Null vma11 mutants were constructed as follows. A 1.8-kb EcoRV-SpeI fragment containing the VMA11 gene was cloned into pBluescript KS ϩ (Stratagene), creating pRHA150. pRHA150 was digested with XhoI and HindIII, blunted with Klenow, and religated to remove the ClaI and HincII sites originating from the multicloning site of pBluescript KS ϩ . The 0.6-kb HincII-ClaI (blunt) fragment of pRHA151 was replaced with a 0.85-kb EcoRI-BglII (blunt) fragment of pJJ280 (TRP1) (37) to create ⌬vma11::TRP1 (pRHA163). The ⌬vma11 fragment in pRHA163 was isolated from the plasmid by digestion with EcoRV and SpeI and used to construct the disruption strain YRH11a by the method as described previously (38).
Isolation and Disruption of the VMA16 (PPA1) Gene-Two oligonucleotide primers, atcgaagttgttttcagtctc and acatgtatctcagatatctca were synthesized based on the reported sequence of the VMA16 (PPA1) gene (31) and used to amplify the gene fragment by polymerase chain reaction from yeast chromosomal DNA in the presence of digoxigenin-11-dUTP. A yeast genomic DNA library (18) was screened by hybridization with the digoxigenin-labeled VMA16 fragment to isolate the full-length VMA16 gene. A 2.0-kb NcoI (blunt ended)-EcoRI fragment containing the VMA16 gene was cloned into the SmaI-EcoRI site of pUC119 (39) to yield pRHP152. A 0.6-kb BamHI-EcoRV fragment within the VMA16 gene was replaced with a 0.94-kb BamHI-HincII TRP1 fragment from pJJ281 (37) to create pRHP165. The ⌬vma16::TRP1 fragment was liberated from the plasmid by HpaI-BglI digestion and used to disrupt the chromosomal VMA16 gene by the method as described in Ref. 38.
Epitope Tagging of Vma3p, Vma11p, and Vma16p-VMA3::HA gene fragment was constructed as follows. A 1.6-kb EcoRI fragment containing the VMA3 open reading frame was cloned into pBluescript KS ϩ , and the resultant plasmid was mutagenized to introduce AatII and NheI restriction sites near the 3Ј end of the VMA3 open reading frame with a mutagenic primer, ctagctagcagacgacgtcttgagtagccctggagttca (40). Then the AatII-NheI fragment, encoding the last four amino acids of Vma3p, was replaced with a synthetic oligonucleotide duplex coding for the last four amino acids followed by a nine amino acid HA epitope (YPYDVPDYA) (41). A 1-kb AgeI-EcoRI fragment containing a part of the VMA3::HA gene fragment was cloned into a yeast integrating vector pRS306 (36), digested with HindIII, and used to substitute for the chromosomal VMA3 gene by two step gene replacement (38). Site-directed mutagenesis following the method of Kunkel (42) was performed to epitope tag VMA11 at the extreme C terminus. The primer ttttaaacttttgactcacgcatagtcaggaacatcatatgggtattcagagcctc was used to introduce the sequence encoding the HA epitope just prior to the stop codon of VMA11. The generation of pLG36 (pRS316-VMA11::HA) was determined by the introduction of a unique NdeI site within the oligonucleotide sequence (underlined). A 2-kb KpnI-SacI fragment from pLG36 containing VMA11::HA was subcloned into pRS306 creating pLG40. The chromosomal copy of VMA11 was replaced by VMA11::HA to generate the yeast strain LGY11 by transformation of SF838 -1D with linearized pLG40 (BglII) and two-step gene replacement.
The sequence encoding the HA epitope was introduced into VMA16 before the stop codon by site-directed mutagenesis using the primer tggtttgagcgcttacgcatagtcaggaacatcatatgggtactgaaattcagaagc. The creation of pLG32 (pRS316-VMA16::HA) was confirmed by the presence of a unique NdeI restriction enzyme site (underlined). pLG34 was generated by subcloning a 1.9-kb SacI-KpnI fragment from pLG32 containing VMA16::HA into pRS306. The chromosomal copy of VMA16 was replaced by VMA16::HA to generate the yeast strain LGY10 by transformation with linearized pLG34 (BstUI) as described above.
Site-directed Mutagenesis of the VMA11 gene-vma11 genes with a mutation at the glutamic acid at position 145 (Glu 145 ) were constructed by site-directed mutagenesis of pRHA150. Mutagenic primers used were, accatataaccctaacacgtcagagaaaattaga (E145D), tataaccctaaaactagtgagaaaattagaat (E145L), and taccatataaccctagtacttgagagaaaattagaat (E145Q). Each primer introduces a unique restriction site (underlined) to screen for the introduced mutation. A 0.85-kb EcoRI-BglII (blunt) fragment from pJJ280 (TRP1) (37) was inserted at the SnaBI site of the resultant mutant plasmids as a marker for mutant selection. The SnaBI site is located ϳ450 base pairs 5Ј of the initiating ATG, and insertion of the TRP1 gene into the wild type VMA11 gene at this position did not affect the cell growth or the V-ATPase activity (data not shown). The mutant vma11 fragments were used to substitute for the chromosomal VMA11 gene in YPH499 to yield RHA116 -118. vma11 mutants were selected by Trp ϩ phenotype, and introduction of the mutations was confirmed by Southern blot analysis of chromosomal DNA.
Protein Preparation, SDS-PAGE, and Western Blot Analysis-Preparation of vacuolar membrane fractions and purification of the V-ATPase were previously described (3,5). When vacuolar membranes were prepared from strains (RHP1100-RHP1108) expressing Vma16p from pRS316 (URA3 ϩ )-based plasmids, cells were grown in YNBD (Ϫuracil) supplemented with 0.5% casamino acids (Difco). Each vacuolar membrane fraction was assayed for the vacuolar membrane marker, dipeptidyl aminopeptidase B (DPAP-B) (43), and a purification index, which was expressed as the ratio of the specific DPAP-B activity in the vacuolar membrane to the activity in spheroplast lysate, was determined. Vacuolar membrane fractions with a purification index of Ͼ30 (5) were used throughout this study. Protein extracts of whole cell and vacuolar membrane fractions were prepared as described by Hill and Stevens (44). Proteolipid subunits were purified by chloroform/methanol extraction of whole cell extracts or vacuolar membranes (45). Crude membrane and cytosolic fractions were prepared as follows. Yeast cells were spheroplasted and lysed with a Dounce homogenizer in a buffer containing 50 mM Tris-HCl, pH 7.5, 0.2 M Sorbitol, 1 mM EDTA, 2 g/ml each of antipain, aprotinin, leupeptin, chymostatin, pepstatin, and 0.5 mM phenylmethanesulfonyl fluoride. The homogenate was centrifuged at 500 ϫ g for 5 min to remove unbroken cells and then fractionated into membrane and cytosolic fractions by centrifugation at 100,000 ϫ g for 1 h. Proteins were solubilized and subjected to SDS-PAGE analysis as described in Ref. 20. Immunoblots were prepared and probed as described (46).
Other Methods-Recombinant DNA manipulations were performed as described (38,47). Fluorescence microscopy was done essentially as described (44). Affinity purified anti-HA polyclonal serum was used at a 1:40 dilution. Vacuolar acidification in vivo was examined by quinacrine staining of the cells (25). V-ATPase activity (ATPase activity sensitive to 0.1 M bafilomycin A 1 ) of wild type and site-directed mutants was assayed at 30°C as described (5). The reaction mixture (100 l) contained 25 mM Tris-Mes (pH 6.9), 5 mM MgCl 2 , 25 mM KCl, 5 mM ATP-2Na ϩ , and vacuolar membrane vesicles (10 -20 g of protein). Prior to the assay of ATPase activity, the reaction mixture (without ATP) was incubated with or without 0.1 M bafilomycin A 1 for 20 min on ice, and then the reaction was started by adding ATP. Bafilomycin A 1 was added as an ethanolic solution, and control activities were determined in the presence of an equivalent amount of ethanol. All samples contained 1% ethanol, and this concentration of ethanol did not inhibit the enzyme activity. V-ATPase activity of the membrane fractions from LGY11 (VMA11::HA), LGY10 (VMA16::HA), and SF838 -1D was assayed as described (3).
Materials-Enzymes for recombinant DNA methods were purchased from Takara Shuzo Co., Ltd. (Kyoto, Japan) or New England Biolabs, Inc. (Beverly, MA). Bafilomycin A 1 was from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Monoclonal antibodies were obtained from Molecular Probes, Inc. (Eugene, OR). Other chemicals were as described by Uchida et al. (5).

Construction of a ppa1 (vma16) Null
Mutant-Vma3p and its homologue, Vma11p, have been shown to be required for activity and assembly of the V-ATPase complex (30). To determine whether another Vma3p homologue, Ppa1p (Vma16p) (31), is also required for expression of the V-ATPase activity, we constructed and characterized a ppa1 null mutant. One of the two copies of the PPA1 gene in a wild type diploid strain, YPH501 (36), was disrupted by a DNA fragment replacing ϳ80% of the PPA1 reading frame with the TRP1 gene (Fig. 1A). The resulting PPA1/⌬ppa1::TRP1 diploid cells were sporulated, tetrads were dissected, and spores were checked for viability and segregation of the TRP1 marker. Spores were grown on YPD medium buffered at pH 5.0. Spore viability for the heterozygous diploid was almost identical to that of the wild type parent cells, and no anomalous segregation pattern was observed (data not shown). These results indicate that the ⌬ppa1 mutant is viable. PPA1 (VMA16) Is Required for the V-ATPase Activity-⌬ppa1 cells displayed growth phenotypes characteristic of mutants disrupted for VMA3, VMA11, or other genes required for the V-ATPase activity (VMA genes) (4, 23) (Table II). These include inability to grow in YPD medium buffered at pH 7.5 (17,48), YPD medium containing 100 mM CaCl 2 (33,48), or medium containing glycerol as a sole carbon source (YPG) (33). The ⌬ppa1 cells also failed to accumulate quinacrine in their vacuoles, which indicates that the mutant vacuole is not acidified (Table II). In addition, vacuolar membrane fractions isolated from the ⌬ppa1 cells lacked bafilomycin A 1 -sensitive ATPase activity (Table II). These results suggest that the PPA1 gene is required for V-ATPase activity. Thus, the PPA1 gene is an additional member of the VMA genes. The name PPA (for proteolipid of proton ATPase) seems to be confusing because the name is used by another gene with unrelated function (PPA2; inorganic pyrophosphatase). We propose here to rename the gene VMA16. VMA16 (PPA1) Is Required for the Assembly of the V-ATPase Complex-⌬vma3 and ⌬vma11 mutant cells fail to assemble V-ATPase subunits onto the vacuolar membrane (25,30,49). In addition, steady state levels of two integral subunits (100-and 17-kDa V 0 subunits) are substantially lowered in these mutants as compared with wild type cells (30, 49). Fig. 2 shows the steady state levels and distribution of the V-ATPase subunits in wild type and ⌬vma16 mutant cells. As observed for ⌬vma3 and ⌬vma11 mutant cells, none of the V-ATPase subunits were detected in the ⌬vma16 mutant vacuolar membrane fraction.
In addition, the levels of the integral subunits (100-and 17-kDa V 0 subunits) were decreased significantly in the ⌬vma16 cells. Loss of Vma16p did not affect the cellular levels of the V-ATPase subunits that are peripherally associated with the membrane (all the V 1 and the 36-kDa V 0 subunit). 3 Pulse-chase experiments of the 100-kDa subunit in ⌬vma16 cells showed that the subunit was synthesized normally compared with wild type but degraded more quickly by nonvacuolar protease(s) in the mutant cells (data not shown). An increased rate of degradation of the 100-kDa subunit was also observed in ⌬vma3 and ⌬vma11 cells (data not shown).
Construction of Strains Expressing Functional Epitopetagged Vma11p and Vma16p-The observed phenotypes of ⌬vma3, ⌬vma11, and ⌬vma16 mutant cells were indistinguishable from one another. The simplest interpretation of these results would be that Vma11p and Vma16p are second and third proteolipid subunits of the V-ATPase. However, a possible function for these proteins in the assembly and/or vacuolar targeting of the enzyme complex could also account for the phenotypes of the mutant cells (44 -46). To further investigate the function of Vma11p and Vma16p, we examined the cellular localization of these proteins. Vma11p and Vma16p were tagged at the C terminus with a nine-amino acid epitope of the influenza virus hemagglutinin (HA) (41) (Fig. 1A) to allow detection of the proteins by anti-HA antibodies. Recombinant DNA fragments coding for the tagged proteins (Vma11p-HA and Vma16p-HA) were constructed, and the chromosomal copies of the respective genes were replaced with the tagged alleles (VMA11::HA and VMA16::HA). Resultant strains expressing Vma11p-HA (LGY11) or Vma16p-HA (LGY10) grew as well as wild type cells (SF838 -1D) on YPD medium buffered to pH 7.5, YPD supplemented with 100 mM CaCl 2 , and YPG. Vacuolar acidification was normal in these strains as indicated by quinacrine uptake into the organelles (data not shown), and the V-ATPase specific activity in isolated vacuolar membrane fractions was similar to that in the wild type membranes (1.83, 1.14, and 1.23 mol ATP hydrolyzed/min/mg protein for SF838 -1D, LGY11, and LGY10, respectively). We conclude that the tagged Vma11p and Vma16p were functional.
b Vacuolar acidification was monitored by quinacrine staining in vivo. c Bafilomycin A 1 -sensitive ATPase activity in vacuolar membrane fractions (nmol ATP hydrolyzed/min/mg protein). The ATPase activity was assayed at 30°C as described under "Experimental Procedures." The vacuolar membrane fractions were isolated from cells grown in YNBD medium to minimize plasmid loss during cell growth. This presumably is the reason for the low V-ATPase activity measured in this series of experiments as compared with that in other experiments (cultured in rich medium (YPD), Table III). The purity of the fractions was not affected by growing the cells in YNBD medium. d ND, not determined.

HA-tagged Vma11p and Vma16p
Are Localized on the Vacuolar Membrane-Localization of Vma11p-HA and Vma16p-HA was analyzed by indirect immunofluorescence microscopy. Anti-HA antibodies bound to the outline of vacuoles in LGY11 (VMA11::HA) and LGY10 (VMA16::HA) cells (Fig. 3). The staining pattern is similar to that of RHA374 expressing a functional HA-tagged Vma3p (Vma3p-HA), and no nonspecific staining was observed for the wild type cells. The discrete punctate pattern around the vacuole observed for Vma11p and Vma16p is often seen when probing for other vacuolar membrane proteins, including the 100-and 69-kDa V-ATPase subunits (19,25). These results indicate that Vma11p-HA and Vma16p-HA are localized in the vacuolar membrane.  Fig. 4). We conclude that these polypeptides represent the tagged proteins because they were uniquely found in each membrane fraction. Protein bands with higher molecular masses are presumably aggregates of the tagged proteins, because no cross-reactive polypeptide was detected in the blot of the wild type proteins (Fig. 4, lane 1). Vma3p-HA migrated slower than Vma11p-HA although the calculated mass of Vma3p (16.4 kDa) (24,25) is smaller than that of Vma11p (17.0 kDa) (30). The reason for the unexpected behavior of these proteins in the gel is currently unknown.
The signal intensity of Vma16p-HA (Fig. 4, lane 4) was similar to that of Vma11p-HA (lane 3) and was 5-10 times lower than that of Vma3p-HA (lane 2; eight times less amount of proteins was loaded in lane 2). Although we cannot estimate the exact molar ratio of the three proteins by this method, this result suggests that Vma16p and Vma11p exist at lower levels than Vma3p in the vacuolar membrane. The tagged proteins gave only weak signals in both immunofluorescence microscopy and Western blot analysis when expressed in cells containing normal vacuolar protease activities, though the cells still display Vma ϩ phenotypes (data not shown). The C-terminal ends of these proteins may be facing inside the lumen of the vacuole, and the tag might be removed by the vacuolar proteases. However, other interpretations of the data cannot be ruled out at this time.
HA-tagged Vma11p and Vma16p Are Components of the V-ATPase Complex-We next examined whether Vma11p and Vma16p are included in the V-ATPase complex. Vacuolar membrane proteins from LGY11 and LGY10 cells were solubilized with ZW3-14 and fractionated through a 20 -50% glycerol density gradient. Fractions from the gradients were assayed for V-ATPase activity and subjected to Western blot analysis with anti-V-ATPase subunit antibodies. Fig. 5 shows the distribution of two V-ATPase subunits (100-kDa V 0 and 69-kDa V 1 subunits) and the tagged proteins in the gradient. As reported previously (18,20), V-ATPase subunits fractionate into two peaks, one that cofractionates with the ATPase activity and contains both the V 0 and V 1 subunits (V 0 V 1 ) (LGY11, fractions 8 -10; LGY10, fractions 7-9) and the other that migrates to lower density near the dipeptidyl aminopeptidase B protein (DPAP-B; ϳ120-kDa vacuolar membrane protein (43)) and contains only the V 0 subunits (LGY11, fraction 12; LGY10, fraction 11). The second peak in the lower density fractions is predicted to represent the V 0 subcomplex not assembled with V 1 subunits (20,27,50). The distribution patterns of Vma11p-HA and Vma16p-HA matched exactly with those of the 100- (Fig. 5) and 36-kDa (Ref. 20 and data not shown) V 0 subunits, indicating that the two proteins are physically associated with both the active V 0 V 1 complex and the V 0 subcomplex in the vacuolar membrane.
Vma11p Glu 145 and Vma16p Glu 108 Are Important for the V-ATPase Activity-The glutamic acid residue at position 137 (Glu 137 ) in Vma3p lies in the center of the predicted fourth transmembrane domain of the subunit (Fig. 1, A and B) (24,25,34). Mutational analysis of this residue has suggested that the carboxyl side chain at this position is critical for proton translocating activity (34). We examined whether Vma11p and Vma16p also contain an acidic residue functionally similar to Vma3p Glu 137 . Vma11p Glu 145 , Vma16p Glu 108 , and Vma16p Glu 188 (Fig. 1, A and B) were modified by site-directed mutagenesis. No other acidic residues are found in the predicted transmembrane domains of the two proteins. vma11 mutant strains (vma11 E145D,E145L,E145Q) were constructed by replacing the chromosomal copy of the gene with the mutant alleles. vma16 mutants were constructed by introducing mutant genes (vma16 E108D,E108L,E108Q,E108V and vma16 E188D,E188Q,E188V) into RHP110 (⌬vma16) on a low copy, centromere-based plasmid, pRS316 (36). Tables II and III summarize the growth phenotypes and V-ATPase activities of the mutant cells compared with wild type cells. Of the three residues examined, mutants of the Vma11p Glu 145 and Vma16p Glu 108 displayed phenotypes similar to cells expressing Vma3p Glu 137 . These residues could be functionally substituted by an aspartic acid residue, but other mutations altering these residues (vma3 E137Q,E137V,E137K (see Ref. 34), vma11 E145L,E145Q, and vma16 E108L,E108Q,E108V) (Tables II  and III) completely eliminated the V-ATPase activity. On the other hand, all of the Vma16p Glu 188 mutants produced a functional Vma16p, which indicates that this residue is not essential for the activity of this protein. These results suggest that Vma11p Glu 145 and Vma16p Glu 108 may be functionally equivalent to Vma3p Glu 137 .
Inactivating Mutations of Vma11p Glu 145 and Vma16p Glu 108 Promote Assembly of V 1 Subunits onto the V 0 Subcomplex in the Membrane-The phenotypes of the mutants of Vma11p Glu 145 and Vma16p Glu 108 could be interpreted as indicating that the carboxyl side chains of these residues are important for their activity. However, the inactivating mutations of the residues might alter the secondary or tertiary structure of these polypeptides, thus inhibiting the assembly and/or targeting of the V-ATPase complex. We therefore examined whether the V-ATPase complex was assembled onto the vacuolar membrane in the vma11 Glu 145 and vma16 Glu 108 mutant cells. Neither of the mutations inhibited the assembly of the V-ATPase subunits onto the vacuolar membrane (Fig. 6,  A and B), although the amounts of membrane-bound V 1 subunits appear to be increased in the inactive vma11 and vma16 mutant cells (Fig. 6, A, lanes 3 and 4, and B, lanes 3-5) and will be discussed in a later section. Solubilization and density gra-dient fractionation of the vma11 E145L mutant vacuolar membrane proteins showed that the V-ATPase subunits in the mutant membrane were assembled almost exclusively into a high molecular weight complex (Fig. 7, RHA117, fractions 4 -6) containing all the known V-ATPase subunits. We found that vma16 E108L mutant vacuolar membrane also contained a fully assembled V 0 V 1 complex (data not shown). These results a Mutant cells were grown on the medium as in Table II. RHA116 grew as well as RHA115 on the selective media, whereas RHA117 and RHA118 displayed growth phenotypes similar to YRH11a.
b Vacuolar acidification was monitored by quinacrine staining in vivo. c Bafilomycin A 1 -sensitive ATPase activity in vacuolar membrane fractions (nmol ATP hydrolyzed/min/mg protein). The ATPase activity was assayed at 30°C as described under "Experimental Procedures." indicate that the vma11 Glu 145 and vma16 Glu 108 mutations do not inhibit the assembly and vacuolar targeting of the V-ATPase subunits. The mutant complex appeared to migrate to denser fractions than the wild type complex (Fig. 7, A and B). The reason for this difference in the behavior of these complexes is currently unknown. Although sometimes one or two additional polypeptides were observed by Coomassie staining of the peak fraction of the mutant V-ATPase complex, these proteins were not consistently present in the peak fractions. It is most likely that these polypeptides are the breakdown products of the V-ATPase subunits. However, we cannot exclude the possibility that these proteins are specifically associated with the mutant complex and thus changing its mobility in the density gradient.
Surprisingly, the inactivating Vma11p Glu 145 and Vma16p Glu 108 mutations resulted in a higher degree of assembly of the V 1 subunits to the V 0 subcomplex in the vacuolar membrane (Fig. 6, A, lanes 3 and 4, and B, lanes 3-5). The levels of V 0 subunits in vacuolar membranes were kept at relatively unchanged levels (Fig. 6, A and B). These mutations did not affect the steady state levels of the V-ATPase subunits (data not shown) but appeared to change the distribution of the 60-kDa V 1 subunit (Fig. 6C) from cytosol to the membrane fractions. Fig. 7 shows that the second peak of the V 0 subunits (RHA115, fractions 8 -10) was not prominent in the mutant membrane (RHA117, fraction 8). Similar changes in the ratio of V 0 V 1 and V 0 complexes resolved in glycerol gradient were observed for the vma16 E108L mutant vacuolar membranes (data not shown). The lighter gradient fractions of the vma11 E145L mutant in Fig. 7B appeared to contain fewer proteins than the wild type fractions. Some of the proteins in the mutant membranes might not have been solubilized efficiently under the conditions that we used for solubilization of the V-ATPase complexes. However, densitometric analyses of the Coomassie Blue-stained gels indicated that the ratio of the total amounts of 69-kDa V 1 subunit to the 36-kDa V 0 subunit in the gradient fractions remained unchanged from that in the membrane (data not shown), indicating that there was no specific loss of V 0 subunits during solubilization and fractionation on the gradient. These results suggest that the assembly and/or retention of V 1 subunits to the V 0 subcomplex on the vacuolar membrane was promoted in the vma11 E145L and vma16 E108L mutant cells.
To further investigate the assembly status we also compared the effects of nitrate on the wild type and mutant V-ATPase complexes. Nitrate (50 -100 mM) causes the stripping of V 1 from V 0 only in the presence of Mg 2ϩ and ATP (3). The dissociation of V 1 by these reagents is thought to be triggered by a conformational change induced by the binding of ATP to a nucleotide binding site in V 1 (20). V 1 subunits in the vma11 E145L mutant complex were released from the membrane by the treatment with nitrate and Mg-ATP (Fig. 8). Incubation in a buffer without Mg-ATP or nitrate was not effective at stripping V 1 from the membranes (data not shown). Therefore, the mutant enzyme complex is likely to retain a structure that can FIG. 7. Detection of V 0 V 1 and V 0 complexes in glycerol gradient fractions containing solubilized vacuolar membrane proteins. 850 g of vacuolar membranes prepared from either RHA115 (A, wild type) or RHA117 (B, vma11 E145L) were solubilized in 1% ZW3-14 (w/v) and fractionated by centrifugation for 14 h at 177,000 ϫ g through a 10-ml 20 -50% (2.5% steps) glycerol gradient. 750-l fractions were collected, and aliquots (100 l each) were assayed for activity of the V-ATPase and DPAP-B. Proteins were precipitated from each fraction by the addition of 5% trichloroacetic acid (final concentration), separated by SDS-PAGE and detected by staining with Coomassie Blue (Coom.). The numbers in italics under the lanes represent V-ATPase and DPAP-B activities for the corresponding fraction. Lanes without numbers on the bottom contained activity less than 1 nmol/min/ml fraction. Distribution of the V-ATPase subunits through the fractions was examined by Western blot analysis with antibodies specific for the 100-, 69-, and 36-kDa V-ATPase subunits. Gradient fractions are identified with even numbers starting at the bottom of the gradient. bind ATP and execute the conformational change triggered by the nucleotide binding. DISCUSSION In this paper, we showed that Vma11p and Vma16p are novel proteolipid subunits of the yeast V-ATPase. The major evidence that supports our conclusion is as follows: 1) Cells disrupted for either VMA11 or VMA16 are defective in the activity and assembly of the V-ATPase. 2) The functional HAtagged Vma11p and Vma16p reside on the vacuolar membrane and cofractionate with the V-ATPase complex when the enzyme is solubilized and purified by glycerol density gradient centrifugation. 3) Loss-of-function mutants (e.g. vma11 E145L) of the two proteolipid subunits contain an inactive but fully assembled V-ATPase complex in the vacuolar membrane, suggesting that the functions of the two proteins are required during the catalytic reaction. Therefore, the yeast V-ATPase contains three different proteolipid subunits in its complex. It is unlikely that these subunits are isoforms of different V-ATPases, because mutations affecting any of these proteins completely inhibit the V-ATPase activity in the vacuolar membrane (25,30,34,48).
No polypeptide corresponding to Vma11p (17 kDa) or Vma16p (23 kDa) has yet been identified by previous purification studies (3,5). This might be due to poor staining of these subunits by Coomassie Blue as has been observed for the proteolipid subunits of the V-and F 0 F 1 -type ATPases (25,51). The levels of the HA-tagged subunits suggest that Vma11p and Vma16p are present at lower levels than Vma3p in the vacuolar membrane, although the exact molar ratio of the three proteolipid subunits in the enzyme complex remains to be determined. This might also be the reason that only Vma3p has been identified by previous biochemical studies of the V-ATPase.
In this work, we demonstrated that the three yeast V-ATPase proteolipid subunits exhibit many common properties. The primary structures of these proteins are similar to one another, and all three subunits are essential for activity and assembly of the enzyme complex (24,25,30,48). Furthermore, each contains a glutamic acid residue that lies in a predicted transmembrane domain and is required for V-ATPase activity (Glu 137 in Vma3p, Glu 145 in Vma11p, and Glu 108 in Vma16p) ( Fig. 1) (34). These results indicate that the three subunits share a common function, which is most likely to be involved in the proton transport activity of the enzyme complex.
The phenotypes of the mutants altering the conserved glutamic acid residues in Vma3p, Vma11p, and Vma16p are similar but not identical. Generally, mutations at Vma3p Glu 137 appeared to be more severe than those of the other two. For example, Vma3p E137D appears to be only partially active (34), whereas Vma11p E145D and Vma16p E108D are fully functional. Only vma3 E137Q, but not vma11 E145Q and vma16 E108Q, inhibits cell growth in YNBD medium buffered to pH7.5 when each of the mutant genes is introduced into wild type haploid cells on a low copy, centromere-based plasmid (pRS316). 4 This dominant negative phenotype produced by the vma3 E137Q mutation, together with the relative abundance of Vma3p-HA in the vacuolar membrane, may suggest that Vma3p is the major proteolipid subunit in the enzyme complex. The coated vesicle proton pump is reported to contain six copies of the 17-kDa proteolipid subunits (52), and binding of N,NЈdicyclohexylcarbodiimide to a single copy of this subunit is sufficient to abolish the enzyme activity (53). Providing that both wild type and mutant subunits are expressed in the same cells and incorporation of a single mutant 17-kDa subunit inhibits the enzyme activity, the fraction of the active ATPase complex is expected to increase as the copy number of the subunit decreases. In this scenario, the defects resulting from mutations in Vma3p would be more detrimental. Of course, the mutant phenotypes might reflect the difference in the functions of the conserved glutamic acid residues in the three proteins. It will be important in the future to establish the function, subunit interaction, and stoichiometry of these proteins individually.
Our present work raises many questions about the proteolipid subunit content of V-type ATPases in other organelles and organisms. The ϳ20-kDa V 0 subunits present in the V-type ATPases of bovine chromaffin granules and coated vesicles have been reported to exhibit similar biochemical properties to those of the 17-kDa proteolipid subunits (52,54) and thus are candidates for a "second" proteolipid subunit in their enzyme complexes. It should be noted that cDNA fragments capable of encoding all or part of Vma16p homologues have been isolated from plant, mouse, human, and nematode. 5 Although yeast cells appear to have two V-ATPase complexes, one containing Vph1p and the other containing Stv1p (14,29), it is highly likely that all three proteolipid subunits function in both V-ATPase complexes, because ⌬vma3, ⌬vma11, and ⌬vma16 mutants each display phenotypes identical to ⌬stv1⌬vph1 double mutants (25,29,30,48).
The inactivating mutations of Vma11p Glu 145 and Vma16p Glu 108 do not inhibit the assembly and targeting of the V-ATPase subunits but instead promote the binding of the V 1 subunits to the V 0 subcomplex on the vacuolar membrane, resulting in an increase in the population of inactive but fully assembled V 0 V 1 complex. These results suggest that the equilibrium of assembly/disassembly of the V 1 subunits to the V 0 subcomplex is altered in the mutant cells. Reversible assembly of the V 1 subunits to the V 0 domain in vivo has been found to be responsive for physiological changes. Kane (27) has shown that dissociation and reassembly of V 1 and V 0 can occur rapidly in vivo in response to changes in nutrient conditions. Sumner et al. (55) reported that the inactivation of the Munduca sexta plasma membrane V-ATPase during molting parallels the dissociation of V 1 subunits from V 0 in the membrane. In this context, it will be interesting to investigate whether the V-ATPase complex senses changes in a proton motive force loaded across the vacuolar membrane or an ATP potential in the cytosol under different nutritional conditions.
In summary, Vma11p and Vma16p are novel proteolipid subunits of the yeast V-ATPase, which show properties similar to Vma3p. Despite the structural and functional similarities of these polypeptides, our results reveal that a single V-ATPase complex must contain at least one copy of each of the three proteolipids. The three proteolipid subunits must therefore participate cooperatively in proton transport through the V 0 sector and thus regulate vacuolar acidification. Our future studies will address these issues.