The structure and function of the fungal V-ATPase

The fungal vacuole is a prominent organelle that functions as a storage site for amino acids, Ca²⁺, storage carbohydrates, inorganic phosphate and numerous hydrolases (Fig. 1) (Klionsky et al. 1990). A hallmark of fungal vacuoles is that they are acidic compartments involved in the turnover of cellular macromolecules. In the yeast Saccharomyces cerevisiae, the hydrolase activities in the vacuole are required for diploid cells to progress normally through sporulation and meiosis. The goal of this chapter is to present an overview of what is known about the enzyme complex responsible for acidifying the fungal vacuole and to summarize the current view of the function of acidification of the vacuolar network.

The vacuolar H⁺-ATPase (V-ATPase) is responsible for acidification of the fungal vacuolar network (Klionsky et al. 1990; Raymond et al. 1992). This enzyme functions by coupling the hydrolysis of cytoplasmic ATP to the translocation of protons across the vacuolar membrane. The V-ATPase has been isolated from several fungi, and the enzymes isolated from Neurospora crassa and the yeast Saccharomyces cerevisiae have been extensively characterized. In all cases, the enzyme is a multisubunit complex of at least 500×10³ relative molecular mass (M_r) (Hirata et al. 1989) composed of both membrane-bound polypeptides and peripherally associated subunits. Assigning functions to each of the polypeptides associated with the purified, enzymatically active V-ATPase complex remains a challenge for the future.

The molecular cloning of the genes encoding the V-ATPase subunits has provided an avenue for the detailed characterization of the structure and function of the various polypeptides. Subunit-encoding genes were first isolated and characterized for the Neurospora crassa V-ATPase 69 and 59×10³M_r polypeptides (Bowman et al. 1988a,b). These investigations were followed by the isolation of the genes for the corresponding yeast subunits (Shih et al. 1988; Nelson et al. 1989; Hirata et al. 1990; Yamashiro et al. 1990). The availability of the yeast genes provided an opportunity for the complete genetic dissection of the fungal V-ATPase subunit composition and function through the use of yeast mutants lacking individual subunits as a result of gene disruptions. These genetic analyses have allowed each biochemically defined subunit to be individually scrutinized for its in vivo requirement for function and assembly of the V-ATPase complex. These topics are considered in detail in this paper and the other papers of this fungal V-ATPase chapter.

The fungal V-ATPase is typically purified beginning with vacuolar membranes that have been separated from cytoplasmic proteins and other organelle membranes, and washed extensively to remove lumenal vacuolar hydrolases (Uchida et al. 1985). Following detergent solubilization of vacuolar membranes, most procedures require sedimentation of the membrane proteins through either a sucrose or a glycerol density gradient. These procedures take advantage of the large size of the V-ATPase complex relative to other vacuolar membrane proteins and the fact that this protein is rather abundant in vacuolar membranes. These biochemical separation procedures have been sufficient to define a number of polypeptides that co-purify with V-ATPase activity (Kane, et al. 1989), and the V-ATPase complexes from N. crassa and yeast have been found to contain similar-sized putative subunits (Kane and Stevens, 1992; Bowman et al. 1992). The V-ATPase subunits from either fungal source behave as if they are associated with either an integral membrane sector (or V₀) or a peripheral membrane sector (or V₁).

The N. crassa V₀ membrane sector is composed of 100, 40 and 16 × 10³M_r polypeptides (Bowman et al. 1992), whereas the V₀ portion of the yeast V-ATPase contains polypeptides of 100, 36, 17 and 16×10³M_r (C. M. Bauerle, M. N. Ho, M. A. Lindorfer and T. H. Stevens, in preparation). There are also similarities in the polypeptides that form the V i peripheral membrane sector of these two enzymes. When the V i sector from N. crassa was dissociated from the vacuolar membrane, polypeptides of 69, 59, 48, 30 and 17 × 10³M_r were identified (Bowman et al. 1989). By comparison, the peripheral membrane portion of the yeast V-ATPase includes polypeptides of 69, 60, 54,42,32 and 27×10³M_r (Kane and Stevens, 1992; M. N. Ho, K. Hill, M. A. Lindorfer, and T. H. Stevens, in preparation; M. N. Ho, R. Hirata, N. Umemoto, Y. Ohya, T. H. Stevens and Y. Anraku, in preparation). With the exception of the biochemically well characterized 69 and 60× 10³M_r peripheral membrane subunits and the l6×10³M_r V₀ proteolipid, it is has been difficult to assess the requirement and/or function of the remaining polypeptides without the aid of either genetics or biochemical reconstitution. Genetic approaches have been pursued in both yeast and Neurospora crassa (see Bowman et al. 1992).

Genetic analysis of the V-ATPase has proceeded rapidly in yeast. The first efforts focused on the cloning and disruption of the 60× 10³M_r subunit (Nelson et al. 1989; Nelson and Nelson, 1990; Yamashiro et al. 1990). These studies indicated that the 60×10³M_r subunit was required for V-ATPase function and that the V-ATPase was required for normal growth rates and for growth over a wide pH range. In particular, yeast cells lacking the 60×10³M_r subunit grew about 50% slower than the wild-type cells under optimal conditions, failed to grow at neutral pH, and the vacuoles of the mutants had a neutral pH (Yamashiro et al. 1990). Subsequent genetic analysis indicated that V-ATPase mutants failed to grow on non-fermentable carbon sources (Foury, 1990), were sensitive to elevated Ca²⁺levels in the growth medium (Ohya et al. 1991), and contained elevated cytoplasmic Ca²⁺ levels (Ohya et al. 1991). Despite all of the physiological effects resulting from loss of the V-ATPase, it has become clear that yeast cells can grow without the V-ATPase and very likely without acidification of the vacuolar network.

Two approaches in yeast have yielded a large number of genes that effect the expression of the V-ATPase. The first was a reverse genetic approach, which involved the biochemical isolation of V-ATPase subunits, determination of peptide sequence by micro-sequencing, and cloning of the gene using degenerate oligonucleotides. This approach has yielded the VMA2, VMA3 and VMA6 genes, which encode the 60, 16 and 36×10³M_r V-ATPase subunits (see Table 1), respectively. The second approach used genetic screens to isolate yeast mutants that exhibited phenotypes predicted for cells lacking a functional V-ATPase, such as a failure to acidify the vacuole (Preston et al. 1989; Manolson et al. 1992), sensitivity to Ca²⁺ (Ohya et al. 1991), failure to grow at neutral pH (M. N. Ho, K. Hill, M. A. Lindorfer and T. H. Stevens, in preparation), as well as sensitivity to the drug trifluoperazine (Shih et al. 1988, 1990). These approaches yielded the TFP1 (also isolated as VMA1), VPH1, VMA13, VMA5 and VMA11 (also isolated as TFP3) genes, which encode the 69, 100, 54, 42 and 17×10³M_r V-ATPase subunits (Table 1), respectively. These approaches also yielded a number of genes that encode polypeptides required for V-ATPase assembly or function; however, these polypeptides may not be actual subunits of the V-ATPase enzyme complex (Ohya et al. 1991; R. Hirata, N. Umemoto, M. N. Ho, Y. Ohya. T. H. Stevens and Y. Anraku, in preparation; M. N. Ho, K. Hill, M. A. Lindorfer and T. H. Stevens, in preparation). Table 1 summarizes the genes and polypeptides required for yeast V-ATPase function, which have been identified either genetically or biochemically.

Evidence that the fungal V-ATPase contains integral membrane (V₀) and peripheral membrane (V₁) sectors comes from structural, biochemical and genetic investigations. The V₁ polypeptides were shown to be stripped from vacuolar vesicles by procedures involving either washing the membranes with 50–100 mmol l⁻¹ KNO₃ in the presence of MgATP (Bowman et al. 1989; Kane et al. 1989) or cold-inactivation in the presence of MgATP (Beltrán et al. 1992). Whether the vacuolar membranes were from N. crassa or yeast, a similar set of polypeptides was removed from the membranes by these treatments. Interestingly, the nitrate treatment of N. crassa vacuolar membranes not only removed the peripherally associated V-ATPase subunits, but the ball and stalk structures (Vi complex) that could be visualized by electron microscopy were correspondingly stripped off the membrane by this treatment (Bowman et al. 1989, 1992). In an elegant combination of biochemical and structural approaches, Bowman et al. (1992) found that the V₁ complex stripped from vacuolar membranes retained most of the structural features seen when the V₁ sector was attached to the vacuolar membrane. These studies lend solid support to the model that the V₁ and V₀ sectors form distinct subcomplexes within the V-ATPase.

Additional support for the two-sector model comes from studies of the state of assembly of the yeast V-ATPase complex in mutants lacking individual subunits of the enzyme. In vma mutants lacking an individual subunit of the V₁ sector (69, 60, 42 or 27 ×10³M_r), subcellular fractionation and immunolocalization studies indicate that there is a complete failure to assemble any of the V₁ subunits onto the vacuolar membrane (M. N. Ho, K. Hill, M. A. Lindorfer and T. H. Stevens, in preparation; Kane et al. 1992; reviewed in Kane and Stevens, 1992). Interestingly, these polypeptides are stable in the cytoplasm of these vma mutant cells, but they do not associate with the vacuole. In contrast, vma mutants lacking a V₁ subunit localize the 100, 36 and 16×10³M_r polypeptides to the vacuolar membrane. These V₀ subunits are stable in cells lacking a V₁ subunit, presumably because they are assembled into a V₀ complex independently of the V₁ complex.

The state of V-ATPase assembly has also been investigated in mutants lacking components of the V₀ sector. Yeast vma mutants lacking either the 16 or 36×10³M_r subunits (yma3 or vma6 mutants, respectively) contained wild-type levels of the V₁ polypeptides, but these V₁ components did not associate with the vacuolar membrane (Noumi et al. 1991; Umemoto et al. 1991; C. M. Bauerle, M. N. Ho, M. A. Lindorfer and T. H. Stevens, in preparation). These studies also found that the absence of a V₀ subunit destabilized the remainder of the V₀ polypeptides, suggesting that all of the membrane components are required for the stability of the V₀ sector. Interestingly, in the absence of a Vo subunit (vma3 mutant) the V₁ polypeptides appear to be assembled into a subcomplex of approximately 400×10³M_r (Kane, 1992).

An exception to the behavior of cells lacking V i sector polypeptides is the yeast mutant lacking the 54×10³M_r subunit encoded by the VMA13 gene (M. N. Ho, R. Hirata, N. Umemoto, Y. Ohya, T. H. Stevens and Y. Anraku, in preparation). vmal3 mutants appear to contain normal levels of all V-ATPase subunits that can be monitored (100, 69, 60,42, 36, 27 and 16× 10³M_r subunits) and assemble all of the remaining Vi polypeptides onto the vacuolar membrane. Whereas the V-ATPase complex appears to assemble in the absence of the 54×10³M_r subunit, it is much less stably associated with the vacuolar membrane than wild-type V-ATPase. The role of this VMA13 protein in the function of the yeast V-ATPase has yet to be determined.

Other VMA genes have been identified that are required for the function of the yeast V-ATPase (Ohya et al. 1991; M. N. Ho, K. Hill, M. A. Lindorfer and T. H. Stevens, in preparation). Whereas little is known about the VMA22 or VMA23 genes (Table 1), the VMA21 gene is predicted to encode a hydrophobic 77 amino acid protein (K. Hill and T. H. Stevens, unpublished data). More is known about the VMA12 gene and its encoded product (R. Hirata, N. Umemoto, M. N. Ho, Y. Ohya, T. H. Stevens and Y. Anraku, in preparation). The VMA12 gene encodes a 25×10³M_r polypeptide, which is predicted to have two membrane-spanning domains. Yeast cells lacking the VMA 12 protein behave like V₀ sector mutants, in that the V₀ subunits are destabilized and the V₁ polypeptides fail to associate with the vacuolar membrane. Interestingly, whereas the VMA12 protein is associated with the vacuolar membrane, it does not co-purify with the detergent-solubilized V-ATPase complex, indicating that this polypeptide is not required for ATP hydrolysis and probably is not a subunit of the active ATPase. In summary, these genetic approaches have successfully identified a large number of yeast V-ATPase subunits as well as at least one factor required for V-ATPase function that is not a subunit of the complex.

The yeast S. cerevisiae is the only organism for which mutants completely lacking the V-ATPase have been reported. Thus, our knowledge of the in vivo function of the fungal V-ATPase is limited to yeast, an organism that survives without a V-ATPase. With this limitation in mind, this section will provide a brief summary of the function of the yeast V-ATPase.

Yeast vma mutants lacking the V-ATPase have been shown to have no measurable ATPase activity associated with their vacuoles and seem to lack any obvious alternative means to pump protons across the vacuolar membrane. In vivo measurements of the vacuolar lumenal pH yield a value of 7.1 (as opposed to a pH of 6.1 for wild-type vacuoles) (Preston et al. 1989; Yamashiro et al. 1990), indicative of a complete collapse of the pH gradient across the vacuolar membrane. These studies have been interpreted to indicate that the V-ATPase is solely responsible for pumping protons into the vacuole, and that yeast cells survive without an acidified vacuole (Yamashiro et al. 1990). An alternative model has been proposed by Nelson and Nelson (1990), in which acidification of the yeast vacuole is required for yeast cell viability, and that vma mutant cells acidify their vacuoles by delivering protons to the vacuole by fluid-phase endocytosis. This model was proposed to explain the pH-sensitive growth phenotype of yeast vma mutants. Thus far, there has been no experimental test of this alternative model.

Anraku and colleagues (Anraku et al. 1992a,b) have shown that vacuoles isolated from vma mutants failed to accumulate Ca²⁺ and basic amino acids in an ATP-dependent manner (Ohya et al. 1986, 1991). Because wild-type vacuoles are very efficient at this ATP-dependent uptake (Ohsumi and Anraku, 1981, 1983; Sato et al. 1984a,b; Kitamoto et al. 1988), these results have led to the suggestion that vma mutant vacuoles are incapable of Ca²⁺ and amino acid storage in vivo (Anraku et al. 1992a). This suggestion has gained experimental support from the observation that cytosolic Ca²⁺ concentration is elevated approximately sixfold in vma mutant cells (Ohya et al. 1991).

The role of acidification by the V-ATPase in membrane traffic and protein sorting has also been investigated. Studies on the uptake of the yeast peptide pheromone a-factor by receptor-mediated endocytosis indicate that delivery of a-factor to the vacuole, and its proteolytic degradation in this compartment are normal in yeast cells defective for acidification (Singer and Riezman, 1990). These experiments suggest that acidification of the yeast vacuolar network is not required for endocytic traffic from the plasma membrane to the vacuole.

The delivery of newly synthesized proteases to the vacuole is effected in yeast mutants lacking the V-ATPase. Early studies demonstrated that two soluble vacuolar hydrolases were sorted to the vacuole with 75–80% efficiency (Yamashiro et al. 1990), indicating that vma cells sort vacuolar proteins much more efficiently than mutants isolated as deficient in vacuolar protein sorting (vps mutants; Raymond et al. 1992). In addition, the vacuolar membrane protein alkaline phosphatase was found to be present only in the vacuolar membrane in vma mutants, when investigated by indirect immunofluorescence (Yamashiro et al. 1990). Kinetic and steady-state experiments indicated that the vacuolar forms of the vacuolar proteases were proteolytically activated in vma vacuoles (Yamashiro et al. 1990).

A more recent investigation of vacuolar protein sorting in vma mutants suggested that the defects in vacuolar protein sorting may be more severe than originally reported (Klionsky et al. 1992a, b). These investigators observed a similar degree of missorting of soluble vacuolar hydrolases into the surrounding medium as had been reported previously, but detected a higher level of the intracellular precursor forms of both soluble and integral membrane vacuolar hydrolases. However, the wild-type strain used in the recent investigations exhibited a three-to fourfold slower processing rate for vacuolar hydrolase precursors (Klionsky et al. 1992a). In addition, subcellular fractionation studies were not conducted to determine whether the intracellular precursors of endogenous vacuolar hydrolases were in fact accumulating in a non-vacuolar organelle in vma cells. Therefore, it remains possible that other than secretion of 20–25% of newly synthesized soluble vacuolar hydrolases, vma mutant cells exhibit only a kinetic delay in the processing of hydrolase precursors (for further discussion see Klionsky et al. 1992).

There are a number of important conclusions about the fungal V-ATPase that can be reached with our current level of knowledge. The V-ATPase is structurally very similar to the FiFo-ATPase. The V-ATPase is composed of a V_o and a V₁ portion, and each of these sectors appears to be capable of assembling into a sub-complex when the other sector is absent. The yeast V-ATPase is composed of at least ten subunits, and genetic approaches have identified at least four additional factors that constitute either novel subunits or factors required for assembly, targeting and/or regulation of the V-ATPase. Finally, the yeast V-ATPase is absolutely required for acidification of the vacuole, and acidification of the vacuolar network is necessary for efficient vacuolar protein sorting.

Many questions remain to be answered regarding the fungal V-ATPase. The mechanism of assembly and targeting of the V-ATPase to the vacuolar membrane is unknown, as is the subcellular distribution of the V-ATPase through the endomembrane system (endoplasmic reticulum, Golgi complex, endosomes, etc.). Much has yet to be learned about the roles of the various V-ATPase subunits in ATP hydrolysis and proton pumping. There are also a large number of structural issues relating to this multisubunit enzyme that have yet to be addressed. The complex physiological defects associated with loss of the V-ATPase are almost a complete mystery; in particular, why are yeast vma mutants incapable of growth on non-fermentable carbon sources? The proteins that function together with the fungal V-ATPase to maintain a vacuolar membrane pH gradient (such as a Cl⁻ channel; Wada et al. 1992) have yet to be well characterized. A number of the issues raised here are addressed in the five papers that follow in this fungal V-ATPase chapter.

This work was supported by a grant from the National Institutes of Health (GM 38006) and an American Cancer Society Faculty Research Award.