Evidence for a conserved 95–120 KDA subunit associated with and essential for activity of V-ATPases

The first accounts of V-ATPase structure did not report any associated subunits larger than the 70 kDa nucleotide binding subunit. Susceptibility of the 95–120 kDa subunit to proteolysis and to aggregation may be in part attributed to its absence in these early reports. More recently, the 95–120kDa subunit has been shown to co-purify with V-ATPases from several sources and to be essential to V-ATPase activity in both biochemical and genetic experiments. Here we review the recent evidence that a conserved 95–120 kDa integral membrane polypeptide is associated with V-ATPases from various organelles and is essential to the assembly and activity of the enzyme.

The first reports of V-ATPase structure (Uchida et al. 1985; Marin et al. 1985; Manolson et al. 1985) were generally one-step purifications on glycerol gradients or Sepharose columns resulting in low purifications and complex polypeptide mixtures. The subunits reported to be associated with these V-ATPases were conservatively limited to the polypeptides that could be specifically labeled with radioactive probes such as the 70 and 60kDa nucleotide binding subunits and the 16 kDa N,N′-dicyclohexylcarbodiirnide binding subunit. If one re-examines some of the protein gels from these first reports, a subunit of approximately 100×10³M_r can indeed be seen in the lanes containing V-ATPase activity (Manolson et al. 1985, Fig. 8). Nevertheless, there are more recent reports on bovine renal tubules (Gluck and Caldwell, 1987), rat Golgi apparatus (Moriyama and Nelson, 1989a) and rat liver lysosomes (Moriyama and Nelson, 1989b) in which purified V-ATPases from these sources did not contain a subunit migrating between 95 and 120× 10³M_r. For rat liver lysosome and Golgi apparatus V-ATPases, the absence of a 95–120×10³M_r polypeptide may reflect a genuine difference in the composition of V-ATPases from various sources. Gillespie et al. (1991) have since presented immunological evidence of a 115× 10³ M_r subunit associated with bovine renal tubule V-ATPase, suggesting that the absence of this subunit, as reported by Gluck and Caldwell (1987), may have resulted from the difficulty of visualizing the 95–120× 10³M_r subunit by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).

Susceptibility of the 95–120 kDa subunit to proteolysis during V-ATPase purification has been reported by Gillespie et al. (1991) and Kane et al. (1992). Adachi et al. (1990a) have shown that the 100×10³M_r subunit from purified bovine coated vesicle V-ATPase is quickly degraded to an 80× 10³M_r and a 21×10³M_r species by low concentrations of trypsin. To avoid the proteolysis problem when purifying the Saccharomyces cerevisiae vacuolar H⁺-ATPase, we have used a yeast strain (BJ926, MATαlMATa, trp1-1/+ +/his1, prcl-126/prcl-126pep4-3/pep4-3, prbl-1122/prbl-l 122 canl/canl gal2/gal2) in which the structural genes encoding the predominant yeast vacuolar hydrolases (protease A, protease B and carboxypeptidase Y) have been mutated. Furthermore, to reduce other cytosolic protease activities, we included 1 mol1⁻¹ EDTA, 2 mmol 1⁻¹ phenylmethylsulfonyl fluoride and 1μgml⁻¹ pepstatin A during the V-ATPase purification procedure (Manolson et al. 1992).

A second potential problem for visualizing the 95–120× 10³M_r V-ATPase subunit on SDS-PAGE is its tendency to aggregate when heated in protein-denaturing sample buffer (Gillespie et al. 1991; Manolson et al. 1992). Denaturing whole-cell extracts in the presence of 1% SDS and 0.6 mmol 1⁻¹ β-mercaptoethanol at temperatures over 62°C, or at 62°C for longer than 15 min, resulted in the failure to detect any signal on an immunoblot from the yeast 95 kDa V-ATPase subunit, Vphlp (Fig. 1). Although 45°C for 15 min and 62°C for 2.5 min appear to result in equal signals (Fig. 1), the latter treatment was found to be empirically more reproducible.

Purifications of V-ATPases that have resulted in an associated 95–120×10³M_r subunit are listed in Table 1. Monoclonal (Kane et al. 1992) and polyclonal (Manolson et al. 1992) antibodies to a yeast 95–100× 10³M_r polypeptide show co-purification of the antigen with V-ATPase activity whereas immunofluorescence microscopy (Kane et al. 1992) and immunogold electron microscopy (Manolson et al. 1992) have demonstrated the specificity of the antigen for the yeast vacuolar membrane. The V-ATPase nucleotide binding subunits appear to block epitope sites in vivo on the yeast 95–100×10³M_r polypeptide (Kane et al. 1992), further supporting the association of this polypeptide with the yeast vacuolar H⁺-ATPase.

Specific removal of the 116 and 38×10³M_r subunits from the purified clathrin-coated V-ATPase did not affect Ca²⁺-activated ATP hydrolysis yet eliminated Mg²⁺-supported ATP hydrolysis and proton pumping activity (Xie and Stone, 1988). Mutations in the structural gene encoding the yeast 95 kDa subunit, VPH1, result in a pH of 6.9 in the vacuolar lumen instead of the wild-type pH of 6.2 (Preston et al. 1989). Vacuoles purified from strains bearing the vphl-1 mutation had no measurable bafilomycin-sensitive ATP hydrolysis or proton-pumping activity (A. K. Bachhawat, M. F. Manolson, D. G. Murdock, D. Garmen and E. W. Jones, in preparation) and had barely detectable levels of the 60 and 69 kDa peripherally bound nucleotide-binding V-ATPase subunits (Manolson et al. 1992). In contrast, when either of the structural genes encoding the 60 and 69 kDa V-ATPase (VAT2, TFP1) is disrupted, the 95 kDa subunit is correctly targeted and stably inserted into the vacuolar membrane as long as the 16 kDa subunit is present (Kane et al. 1992). Taken together, these results suggest that the 95 and 16 kDa subunits are epistatic to the 60 and 69 kDa subunits for V-ATPase assembly.

The 100×10³M_r subunit of the coated vesicle proton pump was labeled by the hydrophobic reagent 3-(trifluoro-methyl)-3-(m-iodophenyl)diazirine, implying that a portion of the polypeptide is buried within the membrane bilayer (Arai et al. 1988). Failure of chaotropic agents such as alkaline Na₂CO₃ (Kane et al. 1992; Manolson et al. 1992) or 0.3 mol 1⁻¹ KI (Adachi et al. 1990b) to strip the 95–100×10³M_r subunit from membranes further suggests that this subunit is an integral membrane protein. This biochemical evidence is supported by sequence analysis of the 116 kDa subunit of the clathrin-coated vesicle/synaptic vesicle H⁺-ATPase (Perin et al. 1991) and the 95 kDa subunit of the yeast vacuolar H⁺-ATPase (Manolson et al. 1992). In both cases, the carboxyl terminal half of the protein contains at least six regions of sufficient length and hydrophobicity to be considered as membrane-spanning regions.

As shown is Table 1, a 95–120 kDa subunit has been found in V-ATPases from a variety of sources. Antibodies directed against the 120× 10³M_r subunit of the chromaffin-granule V-ATPase recognized a 115× 10³M_r polypeptide in the renal tubule V-ATPase, implying shared epitopes between the two proteins (Gillespie et al. 1991). There are now five sequences that appear to encode this V-ATPase subunit; the 116 kDa polypeptide of the rat clathrin-coated vesicle/synaptic vesicle proton pump (Perin et al. 1991, accession number M58758), the 95 kDa subunit (Vphlp) of the yeast vacuolar H⁺-ATPase (Manolson et al. 1992, accession number M89778), the predicted polypeptide encoded by the Similar Io VPHI yeast gene STVJ (M. F. Manolson, D. Proteau, R. A. Preston, B. T. Roberts, M. A. Hoyt and E. W. Jones, in preparation), the TJ6 mouse immune suppressor factor (Lee et al. 1990, accession number M31226) and the Caenorhabditis elegans proton pump homologue (Sulston et al. 1992, accession number Z11115). Multiple sequence alignments demonstrate the extensive homology over the entire open reading frames of these five sequences (Fig. 2). Of the five sequences, only the rat 116 kDa polypeptide and the yeast 95kDa subunit (Vphlp) have been directly linked to their respective V-ATPase enzyme. The TJ6 mouse immune suppressor factor and the Caenorhabditis elegans proton pump homologue are thought to be V-ATPase subunits solely on the basis of sequence homology. Although the Similar To VPHI yeast gene product, Stvlp, was initially identified merely on the basis of sequence homology, there is now genetic evidence that it is a functional homolog of Vphlp (M. F. Manolson, D. Proteau, R. A. Preston, B. T. Roberts, M. A. Hoyt and E. W. Jones, in preparation).

Although there is considerable evidence that a 95–120 kDa subunit is associated with a variety of V-ATPase, its specific function is unknown. Data showing that this one subunit contains domains exposed to the cytoplasm (Adachi et al. 1990a) and to the membrane bilayer (Adachi et al. 1990b; Kane et al. 1992; Manolson et al. 1992) have led Perin et al. (1991) to speculate that its function is in coupling ATP hydrolysis (carried out by the cytoplasmic subunits) to proton translocation (carried out by the intramembranous subunits). Yeast contain several membrane-bound organelles that require electrochemical gradients across their membranes (vacuole, clathrin-coated vesicles, peroxisomes and Golgi complex). The presence of at least two functional homologues of the 95-100 kDa V-ATPase subunit in yeast (Vphlp and Stvlp) has led Manolson et al. (1992) to hypothesize that this subunit may be involved in targeting or regulating V-APases for specific endomembrane organelles. In Saccharomyces cerevisiae, the 95 kDa subunit, Vphlp, is essential for the in vivo assembly and activity of the vacuolar H⁺-ATPase.

This work was supported by a grant from the Medical Research Council of Canada to M.F.M., grants from the National Institutes of Health, GM29713 and DK18090 (E.W.J.), and in part by a grant from the Pittsburgh Supercomputing Center through the NTH Division of Research Resources cooperative agreement U41 RR04154.