J. Cell Sci. Suppl. 11, 161-178 (1989)Printed in Great Britain @ The Company of Biologists Limited 1989

161

Protein sorting in yeast: the role of the vacuolar proton- translocating ATPase

P A T R IC IA M . K A N E , C A R L T . Y A M A S H IR O , JO E L H . R O T H M A N a n d T O M H . S T E V E N S

Institute o f Molecular Biology, University o f Oregon, Eugene, Oregon 97403, USA

SummaryWe are investigating the physiological roles of organelle acidification in yeast by two different approaches. First, we have identified two mutants which are defective in acidification of the yeast lysosome-like vacuole from among a collection of mutants which mis-sort soluble vacuolar proteins to the cell surface. These mutants have been helpful in identifying other cellular functions linked to acidification, such as the activation of vacuolar zymogens. We have complemented this classical genetic approach to acidification with direct biochemical and reverse genetic studies on the yeast vacuolar proton-translocating ATPase (H +-ATPase), the enzyme responsible for vacuolar network acidification. Our biochemical characterization of this enzyme indicates that it is a multisubunit complex with many structural similarities to other vacuolar H +-ATPases. Like the other vacuolar H +-ATPases characterized, it also shares some structural features with the F iF 0-type ATPases of mitochondria, chloroplasts, and Escherichia coli. We are currently cloning the genes for the subunits of the yeast vacuolar H+-ATPase. Mutagenesis of the cloned genes will allow us to determine the phenotype of yeast cells expressing a vacuolar H +-ATPase altered in well controlled ways. We are also beginning to investigate how the subunits of the vacuolar H+-ATPase are assembled into the enzyme complex and targeted to their proper cellular location.

IntroductionThe acidification of the ‘vacuolar network’ of eukaryotic cells has been shown to play a critical role in a wide variety of cellular processes, including receptor-mediated endocytosis, protein sorting and targeting, activation of zymogens, and the assembly of secretory granules (Mellman et al. 1986). Many membrane-bound organelles, including endosomes, lysosomes, the Golgi apparatus, secretory vesicles, and clathrin-coated vesicles, appear to maintain an acidic internal pH, and the acidifi­cation of all of these compartments is mediated by a single type of proton- translocating ATPase, the vacuolar H+-ATPase (Mellman et al. 1986).

The vacuolar H+-ATPases are distinct from the FiFo-type ATPases of Escherichia coli, mitochondria, and chloroplasts and the EiE2-type ATPases of the plasma membrane based on their lack of immunological crossreactivity and comparison of the available amino acid sequence data (Bowman & Bowman, 1986; Bowman et al. 1988a,c; Manolson et al. 1986). The vacuolar H +-ATPases are distinguished by their sensitivity to bafilomycin Ai (Bowman et al. 19886), iV-ethyl maleimide, and relatively low concentrations of KNO3 and KSCN, and by their insensitivity to azide and vanadate (Bowman & Bowman, 1986). The vacuolar H+-ATPases described so

Key words: acidification, H+-ATPase, vacuolar network, endocytosis.

162 P. M. Kane and othersfar are large multisubunit complexes of molecular weight 400-750K (K = 103M T) (Mandala & Taiz, 1985; Bowman et al. 1986; Arai et al. 1988). Although the reported numbers and stoichiometries of subunits vary, the characterized complexes generally have in common three subunits: one each of molecular weights approxi­mately 70 and 60K, which may contain catalytic and regulatory ATP binding domains, and a 15-20K -dicyclohexyl-carbodiimide (DCCD)-binding proteo- lipid believed to be involved in proton transport (Bowman et al. 1986; Manolson et al. 1986). The 70 and 60K subunits appear to be highly conserved across a wide range of species (Manolson et al. 1986). It is not yet clear if the reported discrepancies in subunit composition for different vacuolar H+-ATPases reflect genuine species- or organelle-specific variations or are simply the result of different purification methods. The mechanisms by which the vacuolar H+-ATPases are targeted specifically to multiple cellular organelles and maintained in these organelles is completely unknown.

It has been well established that the yeast Saccharomyces cerevisiae uses many of the same mechanisms as other eukaryotic cells for organelle assembly and protein traffic, making it a good model system for the study of these processes (Schekman, 1985; Rothman & Stevens, 1988). We are investigating the role of vacuolar acidification in protein targeting, interorganelle transport, and organelle biogenesis using the wide range of experimental approaches available in yeast. As shown in other eukaryotic cells, vacuolar acidification appears to be involved in a diverse collection of cellular functions in yeast. We have identified two mutants which are defective in acidification of the yeast lysosome-like vacuole. Such mutants not only provide insight into the range of processes linked to vacuolar acidification, but also provide one means of identifying the cellular components required for acidification of the vacuolar network. Yeast also possesses a vacuolar H +-ATPase which is very similar to the vacuolar H +-ATPases of other eukaryotic cells (Uchida et al. 1985). We are using biochemistry and reverse genetics to approach the problem of vacuolar acidification by purifying and characterizing the yeast vacuolar H+-ATPase and identifying the genes which encode this complex. With this information, we hope to assess directly the consequences of completely disrupting the ATPase complex or subtly altering its structure and to address the problem of how the ATPase is assembled and transported to its proper cellular locations.

Cellular functions of organelle acidification in yeast: identification of two acidification-defective mutantsThe pH gradient across the membranes of acidic intracellular organelles can be collapsed by the addition of lysosomotropic amines (Mellman et al. 1986). When animal cells are treated with concentrations of lysosomotropic agents which disrupt acidification of the vacuolar network, receptor-mediated endocytosis, regulated exocytosis, and the sorting and targeting of newly synthesized proteins are also disrupted (Mellman et al. 1986; Moore et al. 1983; von Figura & Hasilik, 1986; Caplan et al. 1987). When wild-type yeast cells were treated with 400 mM-ammonium

Table 1. Cellular consequences of acidification defectsRole of the yeast vacuolar H+-ATPase 163

CellsSecretion of proteinase A

Quinacrinestaining

Normalized vacuolar ATPase activity

Intracellular PrA in precursor form

Wild-type Wild-type +

400 mM-Nl vpl3 vpl6 vp!8

+

+++

+ + +

+ +

1-0n.d.

0-060-060-72

n.d.

++

Secretion of PrA and intracellular accumulation of proPrA were assessed by immunoprécipitation of intracellular and extracellular fractions from cells pulse-labeled with 35SO+2- as described (Rothman et al. 1988). The level of vacuolar quinacrine accumulation was determined by incubating cells in the presence of 200 jUM-quinacrine in YEPD buffered to pH 7-7 and then observing the cells by fluorescence microscopy (Weisman et al. 1987 ; Rothman et al. 1988). For measurement of vacuolar ATPase activity, yeast vacuolar membranes were purified as described (Uchida et al. 1985), and ATPase activity was assayed using a coupled enzyme assay with an ATP regeneration system (Lotscher et al. 1984). ATPase activities were determined in the presence and absence of inhibitors of the plasma membrane and mitochondrial ATPases (100 |UM- NaVC>3 and 2mM-NaN3), and in all cases < 5 % of the ATPase activity was affected by these inhibitors. ATPase activities were normalized to DPAP-B activities to adjust for varying vacuolar yields from different strains; the normalized activity of wild-type cells was then set equal to 1-0.

Adapted from Rothman et al. 1989b. n.d., not determined.

chloride, they mislocalized a substantial proportion of the newly synthesized vacuolar protein, proteinase A (PrA), to the cell surface rather than delivering it to the vacuole (Table 1), suggesting that acidification of the vacuolar network is necessary for proper sorting of yeast vacuolar proteins.

This laboratory has previously reported the isolation of a large collection of yeast mutants that secrete multiple soluble vacuolar proteins including carboxypeptidase Y (CPY) and PrA (Rothman & Stevens, 1986; Rothman et al. 1989a). Like I-cell fibroblasts, these vacuolar protein localization (vpl) mutants fail to target these enzymes to the vacuole, causing them to travel into the default pathway leading to secretion. The collection of VPL gene products could include many different components of the sorting machinery; among the possible candidates are a receptor which would be responsible for binding to vacuolar enzymes and preventing them from entering the late secretory pathway, and proteins responsible for transporting this receptor and its bound ligand to the vacuole. The observation that collapsing the pH gradient across intracellular membranes also caused secretion of vacuolar proteins suggested that some of the vpl mutants might encode proteins involved in vacuolar acidification.

Representative mutants from each of the 19 vpl complementation groups were screened by fluorescence microscopy for accumulation of the fluorescent lysosomo­tropic dye quinacrine, as a measure of vacuolar acidification (Weisman et al. 1987). Yeast cells treated with ammonium ion failed to accumulate quinacrine in the vacuole under conditions where untreated cells exhibited brightly labeled vacuoles, indi­cating that the accumulation of the dye within the vacuole is dependent on vacuolar

164 P. M. Kane and othersacidification (Weisman et al. 1987; Rothman, 19896). The results for several of the mutants are summarized in Table 1. Most of the mutants (as represented by vpl8 in Table 1) accumulated only slightly lower levels of quinacrine within their vacuoles than wild-type cells. However, two of the mutants, vpl3 and vpl6, exhibited greatly reduced quinacrine labeling even though they appeared to contain normal vacuoles when visualized under Nomarski optics. These results suggest that vpl3 and vpl6 mutants are deficient in a cellular component necessary for establishing or maintain­ing an acidic pH in the vacuolar lumen.

The zymogen maturation of vacuolar hydrolases in the vpl mutants was also examined, and the results obtained are consistent with the conclusions from quinacrine staining. All the available data support the hypothesis that the proteolytic conversion of proPrA to PrA occurs by an auto-catalytic mechanism which is initiated by the acidic pH of the vacuole (Ammerer et al. 1986; Woolford et al. 1986; Rothman & Stevens, 1988). Mature PrA then catalyzes the activation of a large group of vacuolar zymogens (Rothman & Stevens, 1988). When we tested the ability of the various vpl mutants to convert the intracellular portion of proPrA to PrA in vivo, several of the strains were found to accumulate proPrA. As shown in Table 1, failure to proteolytically process proPrA correlated well with low levels of quinacrine labeling; the intracellular PrA was predominantly in the precursor form in vpl3 and vpl6 cells and in the mature form in the other vpl mutants.

The cause of the acidification defect in these cells was investigated further by isolating vacuoles from various vpl mutants and measuring ATP hydrolysis in the presence of sodium azide and sodium vanadate, specific inhibitors of the mitochon­drial and plasma membrane ATPases, respectively. The results are included in Table 1 and indicate that the ATPase activity of the vacuolar membranes is reduced to a degree comparable to the deficiency in acidification in vpl3 and vpl6, while another representative vpl mutant (vpl8) showed no reduction in ATPase activity. The levels of the 69 and 60K subunits of the vacuolar H +-ATPase present in the isolated vacuoles were measured by Western analysis using antisera raised against the beet tonoplast H +-ATPase subunits (see below). Both the 69 and 60K subunits were present at diminished levels in vpl3 and vpl6 relative to wild-type. These results suggest that the acidification defects in vpl3 and 6 are directly related to a deficiency in vacuolar H+-ATPase activity.

The exact function of the VPL3 and VPL6 gene products is not clear. One obvious possibility is that these genes encode subunits of the vacuolar H+-ATPase and that the absence of one subunit can hinder assembly of the ATPase complex. We have determined that the VPL3 gene does not encode the 69 or 60K subunit of the yeast vacuolar H+-ATPase by deleting the gene and showing that both subunits are still present in the vacuole, although at reduced levels (C. K. Raymond & C. T. Yamashiro, unpublished data). The VPL6 gene has now been cloned and similar experiments are in progress. It will be possible to address this issue more conclusively as the entire subunit composition of the yeast vacuolar H +-ATPase becomes better defined.

It is also possible that the VPL3 and VPL6 genes encode cellular components

Role of the yeast vacuolar H +-ATPase 165necessary for the proper assembly and/or targeting of the ATPase subunits. Technical difficulties have precluded measurement of vacuolar H+-ATPase protein levels or activity in whole cell lysates, so it is not yet clear if the enzyme is mislocalized in these mutants. Another vacuolar membrane protein, dipeptidyl aminopeptidase B (DPAP-B) (Suarez Rendueles et al. 1981; Bordallo et al. 1984), appears to be correctly localized in vpl3 and vpl6 cells, suggesting that these mutants do not have a general defect in the sorting of vacuolar membrane proteins. However, the structure of the H +-ATPase may present some unique problems in assembly and targeting (discussed below). While it appears that the acidification-defective vpl mutants described here have identified proteins that are fairly intimately connected with the vacuolar H+-ATPase, it might also be possible to identify additional mutants with defects in accessory molecules, such as other ion transporters which regulate the pH of acidic compartments more indirectly, by screening for acidifi­cation defects. The power of this classical genetic approach is that it can both extend our understanding of the cellular functions dependent on acidification and permit the identification of the full range of components necessary for establishing and maintaining an acidic pH.

Biochemical characterization of the yeast vacuolar H+-ATPaseWe are complementing the genetic approach to vacuolar acidification described above with a direct biochemical characterization of the yeast vacuolar H+-ATPase. A partial purification of the enzyme has been described previously (Uchida et al. 1985). The H+-ATPase complex was reported to contain only three subunits, of molecular weights 89, 64, and 19-5K, and the smallest subunit was shown to bind DCCD. We have repeated the described purification procedure, which consists of isolation of yeast vacuoles, removal of loosely associated protein by low salt washes, solubiliz­ation of the washed vacuolar vesicles with the zwitterionic detergent ZW3-14, and fractionation of the solubilized proteins by glycerol gradient centrifugation. After the gradient centrifugation step, ATPase activity was well separated from activity of DPAP-B. A Coomassie-stained gel of the glycerol gradient fraction exhibiting the highest ATPase activity is shown in Fig. 1A. This fraction contains two major polypeptides of molecular weights 69 and 60K, as well as several minor bands of molecular weights 100, 43 , 36, 32, and 27K. Figs IB and 1C show immunoblots of the same glycerol gradient fraction, probed with polyclonal antibodies raised against the 67 and 57K subunits of the beet tonoplast (vacuolar) H+-ATPase (Manolson et al. 1985). (Antibodies were a gift from M. F. Manolson and R. J. Poole.) The antibodies crossreact specifically with the 69 and 60K polypeptides of the partially purified yeast vacuolar H+-ATPase. Antibodies against the Neurospora crassa 70K subunit (a gift from B. J. and E. J. Bowman; Bowman et al. 1988c) also crossreacted with the yeast 69K polypeptide. Extensive immunological crossreactivity between vacuolar H +-ATPases from animal, plant, and fungal sources has been reported previously (Manolson et al. 1986; Bowman et al. 1986), and these results suggest that

166 P. M. Kane and othersA B C

Mr x 1CT3116 —

9 7 -

66 -

4 5 -

Mr x 10-3

- 6 9

- 60

2 9 -

Fig. 1. The yeast vacuolar H +-ATPase was partially purified by glycerol gradient centrifugation as described (Uchida et al. 1985). The ATPase activity in each gradient fraction was measured (Lotscher et al. 1984), and the protein in each fraction was precipitated with trichloroacetic acid and analyzed by SD S-P A G E . A. Coomassje- stained gel of the gradient fraction exhibiting maximum ATPase activity. Molecular weight standards are shown at left. B ,C. Western blots of the gradient fraction shown inA. The protein was transferred to nitrocellulose and probed with antibody against the beet tonoplast 67K subunit (B) and 57K subunit (C).

the structure of the yeast vacuolar H +-ATPase is fundamentally similar to that of other vacuolar H +-ATPases.

Further support for this conclusion is provided by inhibitor labeling studies. The 70K class of subunits from several vacuolar H + -ATPases has been reported to be labeled with [ I4 C]NBD-C1 (7-chloro-4-nitrobenzo-2-oxa-l,3-diazole), an ATP ana­log which binds covalently to ATP-binding sites and inhibits ATPase activity (Bowman et al. 1986; Randall & Sze, 1987; Arai et al. 1987a; Uchida et al. 1988). The 69K subunit of the yeast vacuolar H +-ATPase can also be labeled with [1 4 C]NBD-C1 in an ATP-protectable manner, suggesting that it is the catalytic subunit of the enzyme (Mandala & Taiz, 1986; Uchida et al. 1988). D CCD is another inhibitor of ATPase activity, which is believed to bind to the proton ‘pore’ (Bowman & Bowman, 1986) and to inhibit ATPase activity because of the tight coupling of A TP hydrolysis and proton translocation. When unsolubilized vacuolar vesicles or the glycerol-gradient-purified yeast vacuolar H +-ATPase are reacted with 20 ,um-[1 4 C ]D C C D , a single polypeptide of molecular weight 17K is labeled. Like the

Role of the yeast vacuolar H +-ATPase 16715-20K subunits of other vacuolar H +-ATPases (Sebald etal. 1979; Rea etal. 1987; Kaestner et al. 1988), this 17K peptide behaves chemically as a proteolipid and can be extracted very specifically from the vacuolar membrane by treatment with 2 : 1

(v/v) CHCI3: MeOH (our unpublished results).The yeast vacuolar H+-ATPase is also sensitive to relatively low concentrations of

chaotropic anions such as K N 0 3, with an ID50 for KNO3 of approximately 75 mM (our unpublished results). Although the sensitivity of vacuolar-type H+-ATPases to chaotropic anions such as nitrate and isothiocyanate has been well-established (Bowman & Bowman, 1986), it has only recently become apparent that these agents may act by removing some of the H+-ATPase subunits from the membrane (Rea et al. 1987; E. J. and B. J. Bowman, personal communication). Fig. 2A shows the kinetics of inhibition of the yeast vacuolar H+-ATPase by 1 0 0 mM-KN0 3 in the presence of 5 mM-MgATP, and Fig. 2B demonstrates that the kinetics and extent of release of the 69K subunit from the vacuolar membrane correlate closely with the inhibition of ATPase activity. Similar results were obtained for the 60K subunit. The 60 and 69K subunits of the H+-ATPase are the predominant proteins which are stripped from the vacuolar membrane under these conditions, but smaller amounts of proteins of molecular weights 27, 33, 36, and 43K also appear in the supernatant after KNO3 treatment. The molecular weights of these proteins match well with those of the minor components in the glycerol gradient-purified H+-ATPase, and the implications of this result are discussed below. Interestingly, the 69 and 60K subunits are not removed from the membrane by K N 0 3 in the absence of MgATP or in the presence of the non-hydrolyzable ATP analog, AMP-PNP, suggesting that the binding, and possibly the hydrolysis, of MgATP induces a conformation of the enzyme which is competent for stripping.

The results described above are summarized in Table 2. These results indicate that the yeast vacuolar H+-ATPase is a ‘typical’ vacuolar H+-ATPase and help to establish yeast as an appropriate model system for our studies of vacuolar acidification. They also extend the previous structural characterization of the yeast enzyme, and a proposed structural model which incorporates these data is shown in Fig. 3. The main features of the model are now discussed. (1) We believe that all of the peptides which appear in the glycerol gradient fraction exhibiting the maximum ATPase activity (Fig. 1) are strong candidates for genuine subunits of the enzyme. We have consistently seen this pattern and stoichiometry of bands in different preparations of the enzyme, including preparations which had very high specific activity. This proposed subunit composition closely resembles the proposed compo­sitions of at least two other vacuolar H+-ATPases, the clathrin-coated vesicle H+- ATPase (Arai et al. 1987a) and the bovine kidney H+-ATPase (Gluck & Caldwell, 1987). It is possible that in the earlier studies on the yeast vacuolar H+-ATPase too little material was fractionated by SDS-PAGE to permit visualization of the minor bands. (2) The orientation of the H+-ATPase in the vacuolar membrane is postulated from its function, which is to utilize ATP (presumably cytoplasmic since there is no evidence of an ATP/ADP translocase in the vacuolar membrane) to acidify the interior of membrane-bound organelles. This conclusion is supported by

A168

B

Time

Cu

0 30T im e (m in)

60

-N itra te + Nitrate

Pellet Sup. Pellet Sup.

0 30 60 0 30 60 0 30 60 0 30 60 /Wr x 10 3

200

97-4

68-0

Fig. 2. Washed vacuolar membranes were incubated with 5 mM-MgATP at 37 °C in the presence or absence of 100mM -KN03. A. Aliquots of the incubation mixture were removed at the indicated times and assayed for ATPase activity as described in Fig. 1.B. Aliquots of the mixture were removed and centrifuged at 4°C to pellet the membranes. The protein released into the supernatant was precipitated with TCA. Both the pellet and supernatant fractions were solubilized, separated by SD S-P A G E , and blotted onto nitrocellulose. The blot shown was probed with antibody against the beet tonoplast 67K subunit. Sup., supernatant.

the ability to strip the putative catalytic (69K) subunit from vacuolar vesicles, since these vesicles were previously shown to be predominantly right side out (Ohsumi & Anraku, 1981). (3) The coordinate removal of the 27, 33, 36 and 43K peptides with the 69 and 60K subunits suggests that these peptides may be associated to form an Fi-like complex which is stripped from the vacuolar membrane by treatment with 100 mM-KNO3 . An additional implication of the K N O 3 stripping experiments is that the stripped subunits must be peripheral proteins. We have confirmed this for the 69

+ 100mM-

nitra te

P. M. Kane and others

— lOOmM-

nitra te

Table 2. Properties of yeast vacuolar H +-ATPase subunitsRole of the yeast vacuolar H +-ATPase 169

Mr ( X 10 3) Inhibitor Stripped by Proposedof subunit labeling 100m M K N 03? function

100 - - ?69 [14C]NBD-C1 + Catalytic ATP-binding subunit60 — + Regulatory ATP-binding subunit*43 - + ?36 - + ?32 - + ?27 - + ?17 [14C]DCCD - Proton‘pore’

Vacuolar membranes were isolated and washed in low salt as described (Uchida et al. 1985). The washed membranes were labeled with 50 jUM-[14C]NBD-C1 for 60min on ice in the presence and absence of 5 mM-MgATP. The vesicles were then pelleted, solubilized, and separated by SD S-P A G E under non-reducing conditions. [14C]NBD-C1 labeled proteins were identified by autoradiography of the dried gel. Although several proteins appear to be labeled at this concentration of [14C]NBD-C1, the most prominent band that labeled in an ATP-protectable manner was at 69K. In the DCCD labeling experiments, washed vacuolar membranes were incubated with 100/iM [14C]DCCD for 60-90 min at approximately 25 °C and labeling was assessed as described above. The major labeled protein had a molecular weight of 17K and was extracted quantitatively into 2 :1 (v/v) C H C ^M eO H . A protein of similar molecular weight was labeled when the glycerol gradient-purified enzyme was treated with 2 0 /xm [14C]DCCD. For the KN O 3- stripping experiments, washed vacuolar vesicles were incubated with 100mM-KN03 in the presence of 5 mM-MgATP for 75 min at 30 °C. The pellet and supernatant fractions were analyzed as described in Fig. 2.

* Proposed based on similarity to the beet tonoplast H+-ATPase 57K subunit (Manolson et al. 1985).

H+

Fig. 3. Proposed subunit structure of the yeast vacuolar H+-ATPase.

170 P. M. Kane and othersand 60K subunits by treating vacuolar membranes with alkaline sodium carbonate (conditions which appear to dissociate all but integral membrane proteins) and showing that the immunoreactive 69 and 60K subunits appear quantitatively in the supernatant. (4) Several lines of evidence suggest that the 17K subunit is an integral membrane protein involved in proton translocation. The 17K subunit remains with the membrane pellet following alkaline sodium carbonate treatment, consistent with its chemical behavior as a proteolipid. Data obtained for other vacuolar H+-ATPases indicate that the DCCD-labeling proteolipid is present in multiple (approximately 6 ) copies (Arai et al. 19876) and that the isolated subunit can be reconstituted to form a proton pore (Sun et al. 1987). (5) The structure shown in Fig. 3 is strikingly similar to the structure of the F¡Fq class of ATPases (with the exception of the 100K subunit which has no counterpart in FiFo). This overall structural similarity is strongly supported by similarities in the amino acid sequence of the a and /? subunits of F i and the 70 and 60K classes of subunits of several vacuolar H+-ATPases, suggesting that all of these subunits may have a common ancestor (Bowman et al. 1988a,c; Zimniak et al. 1988; Manolson et al. 1988; Denda et al. 1988). Further support for such a model is provided by recent studies on the subunit stoichiometry of the clathrin-coated vesicle H+-ATPase, which shows strong similarities with the F iF 0

subunit stoichiometries (Arai et al. 1988), and by electron micrographs of the bovine kidney (Brown et al. 1987) and yeast (E. Gogol, C. T . Yamashiro, and P. M. Kane, unpublished data) vacuolar H+-ATPases which suggest an overall structural resemblance to FjFo-ATPases.

Relating structure and function of the vacuolar H+-ATPasePhysiological roles of vacuolar acidificationUsing the molecular cloning and reverse genetic techniques available in yeast, in combination with the biochemical information described in the last section, we can begin to look at the cellular role of acidification directly by genetically manipulating the vacuolar H+-ATPase. We are currently cloning the genes for the 69, 60, and 17K subunits. N-terminal amino acid sequences for the 69 and 60K subunits were obtained by stripping the subunits from the vacuolar membrane with KNO3 as described above, separating by SDS-PAGE, and blotting the separated proteins onto polyvinyldiene difluoride membrane (Matsudaira, 1987) for sequencing. The 17K proteolipid was sequenced after extraction in 2 :1 (v/v) CHCI3: MeOH with no further purification. Synthetic oligonucleotides were designed based on the amino acid sequences and used to probe a yeast genomic DNA library. The gene for the 60K subunit has been cloned by this method, and we are characterizing several positive clones identified by the oligonucleotides against the 17 and 69K sequences.

After the H+-ATPase subunit genes have been cloned, we can mutagenize them in vitro, replace the wild-type alleles with the mutagenized copies, and determine the phenotype of yeast cells expressing a vacuolar H+-ATPase altered in well-controlled ways. Initial studies will focus on deleting one of the subunit genes (initially the 60K subunit gene). We predict that deletion of the 69, 60, or 17K subunit gene will

Role of the yeast vacuolar H +-ATPase 171dramatically disrupt the structure of the H+-ATPase (discussed below), resulting in a mutant with no functional vacuolar H+-ATPase. Characterization of such ATPase- deficient mutants should yield insight into the multiple roles of vacuolar acidification and help clarify the results obtained with the acidification-defective vpl mutants described above.

Based on results with the vpl mutants (described above and in Rothman et al. 19896), we predict that ATPase-deficient mutants will be defective in the sorting of soluble vacuolar proteins. Observation of a vpl phenotype in ATPase-deficient mutants would confirm a direct link between vacuolar network acidification and vacuolar protein sorting. It will also be interesting to observe whether the sorting of vacuolar membrane proteins, such as DPAP-B, is perturbed in an ATPase-deficient mutant since there is evidence that soluble and membrane proteins use distinct mechanisms for vacuolar sorting (Rothman & Stevens, 1986, 1988). It is not clear at present how acidification of the vacuolar network could exert an effect on vacuolar protein sorting in yeast. In mammalian cells, where the sorting of soluble lysosomal proteins appears to be mediated by the mannose 6-phosphate receptor, the lysosomal enzyme sorting process is thought to mirror the endocytic pathway: the mannose 6- phosphate receptor binds its ligand at neutral pH and carries it until it reaches an acidic compartment destined for fusion with the lysosome (Kornfeld, 1986; Brown et al. 1986). In an acidic environment, the ligand-receptor interaction is weakened, the enzyme bound for the lysosome is released, and the receptor recycles back to a pre­sorting compartment. Although yeast cells do not appear to utilize the mannose 6- phosphate pathway in vacuolar enzyme delivery, a similar receptor-mediated pathway based on a non-carbohydrate determinant has been proposed (Yalls et al. 1987; Rothman & Stevens, 1986). While the precise mechanistic role for acidification may be difficult to define, an ATPase-deficient mutant could be helpful in dissecting the pathway of vacuolar protein sorting if, for example, the failure to acidify the vacuolar network trapped a sorting receptor in a single cellular compartment (Brown .et al. 1986).

The results with the acidification-defective vpl mutants also suggest that many functions of the yeast vacuole may be affected by a failure in acidification. As described above, the activation of proPrA, and therefore all the subsequent steps in the zymogen activation cascade, appear to be dependent on the acidic pH of the vacuole (Woolford et al. 1986; Ammerer et al. 1986; Rothman & Stevens, 1988). A variety of antiporters utilize the H+-gradient to pump basic amino acids and Ca2+ into the vacuole (Ohsumi & Anraku, 1981; Anraku, 1987). The physiological importance of these transporters has not been determined conclusively, but it has been suggested that vacuolar amino acid stores may be important during nitrogen starvation (Anraku, 1987) and that the yeast vacuole may play a role in osmoregula­tion in the cell, similar to the role played by plant tonoplasts (Banta et al. 1988). Banta et al. (1988) have also observed that a . vpt (vacuolar protein targeting; Bankaitis et al. 1986) mutant which is deficient in vacuolar acidification by quinacrine staining is also highly sensitive to low pH media, suggesting that vacuolar acidification is important in the regulation of intracellular pH. Since the acidic pH of

172 P. M. Kane and othersthe vacuole seems to be a central feature in so many of its functions, an ATPase- deficient mutant may in fact help define the physiological roles of the vacuole in the yeast cell.

Functions dependent on acidic compartments other than the vacuole might also be affected in ATPase-deficient mutants. Vacuolar protein sorting may actually fall into this category, since sorting appears to be performed in a pre-vacuolar compartment, presumably an equivalent of the late Golgi or trans-Golgi network which has been defined for mammalian cells. In mammalian cells, the acidic pH of the endosome appears to be critical for endocytosis (Mellman et al. 1986), so it is possible that an ATPase-deficient yeast mutant might also be defective in endocytosis. Fluid-phase endocytosis has been demonstrated in yeast using soluble fluorescent dyes as endocytic markers (Makarow, 1985; Riezman, 1985). Measurements of FITC -dex- tran uptake suggested that the marker may encounter a low pH pre-vacuolar compartment (Makarow & Nevalainen, 1987), but subsequent studies have indicated that this marker is somewhat unreliable for measurement of endocytic uptake (Preston et al. 1987). Receptor-mediated endocytosis has been implicated in the uptake of the mating pheromone a-factor (Chvatchko et al. 1986), and it has been suggested that «-factor endocytosis is important at some stages in the signal transduction pathway of the mating response (Reizman et al. 1986). In fact, the broad range of functions which may be linked to vacuolar acidification suggests that a complete deficiency in vacuolar H+-ATPase function might be lethal to the cell. If this proves to be the case, many of the same issues can be addressed by isolating temperature-sensitive mutations in the vacuolar H+-ATPase subunit genes and observing the onset of the various phenotypes upon a shift to the non-permissive temperature.

Assembly and targeting of the vacuolar H + -ATPaseExamination of the yeast vacuolar H+-ATPase mutants could also provide infor­mation on how the enzyme is assembled and targeted. Based on studies on FxFq- ATPases and other multisubunit complexes, it seems highly unlikely that the cell will be able to assemble a fully functional vacuolar H+-ATPase without one of the subunits (Schneider & Altendorf, 1986; Futai & Kanazawa, 1986). This may be especially true for the 69, 60, and 17K subunits, which appear to perform well- defined functions in the enzyme (Table 2) and may be present in multiple copies in the assembled enzyme (Arai et al. 1988). However, it is possible that deletion of the various subunits could result in different phenotypes, since these deletions might give rise to distinct partially assembled complexes. For example, the cytoplasmic orientation of the 69 and 60K subunits and the ability of these subunits to be stripped from the membrane suggests that these subunits might be pre-assembled into a cytoplasmic complex which is then attached to the proteolipid portion of the enzyme. If this model is correct, then deletion of the 17K proteolipid would result in an accumulation of Fj-like complexes in the cytoplasm. In contrast, deletion of the 60 or 69K subunit gene could give rise to both partially assembled or unassembled

Role of the yeast vacuolar H +-ATPase 173cytoplasmic subunits and, possibly, an ‘unplugged’ proteolipid proton pore in the vacuolar membrane.

As our biochemical understanding of the vacuolar H+-ATPase increases, some of the unique problems associated with its assembly are becoming clear. All of the vacuolar proteins investigated so far, including both soluble and membrane proteins, utilize the initial stages of the secretory pathway in their transport to the vacuole (Stevens et al. 1982; Roberts et al. 1989). However, at present there is no evidence that the 69 and 60K subunits of the yeast vacuolar H+-ATPase are transported through the secretory pathway. These two subunits are cytoplasmically-oriented, peripheral membrane proteins of the vacuole. Therefore, if portions of the protein are ever translocated into the endoplasmic reticulum lumen, perhaps forming an ‘anchor’ during transport through the initial stages of the secretory pathway, these portions of the protein must be cleaved off in the vacuole. There is little evidence of this type of processing in the published sequences of the 69 and 60K subunits from other systems (Bowman et al. 1988a,c\ Zimniak et al. 1988; Manolson et al. 1988). The molecular weights of the purified proteins are very similar to the molecular weights predicted from the sequence of the cloned genes. A few amino acids appear to be cleaved from the N terminus of these proteins, but there is no evidence of a signal sequence at the N terminus and a small amount of N-terminal proteolytic processing of cytoplasmic proteins is commonly seen. The available sequence for the 60K subunit of the yeast vacuolar H+-ATPase is consistent with these observations from other systems. None of the subunits of the gradient-purified yeast vacuolar H+- ATPase receives N-linked glycosylation, based on the lack of sensitivity to EndoF treatment (our unpublished results). While this result is inconclusive in the absence of amino acid sequence information showing unmodified glycosylation sequences, it provides additional support for the contention that much of the H+-ATPase never encounters the enzymes of the secretory pathway. The 17K proteolipid is currently the best candidate among the H+-ATPase subunits for transit through the early secretory pathway and sorting to the vacuole. Although it does not appear to be glycosylated, it is an extremely hydrophobic protein, thus it would not be surprising if it lacked glycosylation sites. In the absence of other evidence, the fact that it is an integral membrane protein argues that it might transit the early secretory pathway along with other vacuolar membrane proteins such as DPAP-B (Roberts et al. 1989). It is possible then that the complete vacuolar H+-ATPase is assembled from a combination of components that are translated on cytoplasmic ribosomes and components which enter the early secretory pathway and are sorted to the vacuole.

The mechanisms by which these components might assemble are completely unknown. The closest analogy in the cell is the assembly of multisubunit mitochon­drial enzymes such as F iF 0-ATPases from a combination of nuclear and mitochon­drial gene products (Douglas et al. 1986). However, the analogy is still somewhat weak because nuclear-encoded FjFo-ATPase subunits must cross a double mem­brane before assembling in the mitochondrial matrix (Douglas et al. 1986), while the 69 and 60K subunits of the vacuolar H +-ATPase could be competent for assembly immediately after translation. The hypothesis that the cytoplasmically synthesized

174 P. M. Kane and otherssubunits could assemble into a partial complex before becoming membrane-bound is an attractive one. We have developed a collection of monoclonal antibodies which recognize both the native vacuolar H+-ATPase and the material stripped from the vacuolar membranes by KNO3. These antibodies could be very useful in studying assembly if they prove to be capable of recognizing partial ATPase complexes. If the regulation of subunit stoichiometries and vacuolar H+-ATPase assembly is post- transcriptional, it may also be possible to accumulate partially assembled complexes by altering the gene dosage of the various subunit genes.

The question of how the vacuolar H+-ATPase is assembled is closely associated with the question of how it is targeted to its proper locations in the vacuolar network. Proper targeting of the enzyme would appear to be of fundamental importance given the multiple roles of organelle acidification described above. The mechanisms for properly targeting this enzyme may be complex because it has been shown, most clearly in animal cells, that all acidic compartments are not alike. Instead, a protein appears to encounter progressively decreasing pH as it travels from the plasma membrane to the lysosome in the endocytic pathway and through the Golgi apparatus to the lysosome in the biosynthetic pathway (Mellman et al. 1986). This combination of targeting and pH regulation could be achieved by the presence of different isoforms of the vacuolar H+-ATPase in different organelles. These organelle-specific forms of the enzyme could have varying forms of a single subunit or have different numbers of subunits, but in either case the variable subunits would both carry targeting information and modulate the activity of the enzyme. There is no evidence yet to prove or disprove this model. A second model for targeting and regulation proposes a single form of the vacuolar H+-ATPase, which is present throughout the vacuolar network, but requires that the activity of the enzyme be regulated by other factors which are present in the various organelles. There is some evidence that this second type of mechanism may play a part in regulating the pH of the endocytic pathway in mammalian cells. It has been suggested that the plasma membrane Na+K +-ATPase can modulate the pH of early endosomes by creating an electrochemical gradient which inhibits the activity of the vacuolar network H+- ATPase in these cells (Cain & Murphy, 1988). As the Na+K +-ATPase is sorted away from the H+-ATPase and recycled back to the cell surface, the inhibition of the H+- ATPase is relieved and the pH of the late endosome and lysosome is lowered. A similar mechanism could be envisioned for the biosynthetic pathway, in which the vacuolar H+-ATPase is assembled early in the secretory pathway, but other newly synthesized transporters modulate its activity until they are sorted away from it into the late secretory pathway.

ConclusionsOur current model for the assembly and cellular roles of the yeast vacuolar H+- ATPase is shown in Fig. 4. For simplicity, the H+-ATPase is depicted as assembling from membrane and cytoplasmic components in the late Golgi, where sorting of soluble vacuolar (V) proteins from secretory (S) proteins takes place. The H+-

Role of the yeast vacuolar H +-ATPase 175

Fig. 4. Model for the assembly and cellular roles of the yeast vacuolar H+-ATPase. Symbols: t, vacuolar H+-ATPase; ■ , ‘Fo-like’ (proton channel) portion of the H + ATPase; # , ‘Fi-like’ (ATPase) portion of the H+-ATPase; V, soluble vacuolar protein;S, secretory protein.

ATPase is then transported with other proteins destined for the vacuole to a pre- vacuolar compartment and from this compartment to the vacuole. At the vacuole, the acidic environment created by the H+-ATPase promotes zymogen activation and drives a series of antiporters for Ca2+ and basic amino acid uptake.

Many unanswered questions still surround the processes of vacuolar acidification and protein sorting. We have described a combination of genetic and biochemical approaches directed at both characterizing the yeast vacuolar H+-ATPase as the central player in vacuolar acidification and elucidating the physiological roles of acidification. Using the broad range of approaches available in yeast, we will continue to dissect the components responsible for establishing and maintaining acidic compartments and to investigate the many cellular processes linked to acidification.

This work was supported by an American Heart Association, Oregon Affiliate, postdoctoral fellowship to P .M .K ., National Institutes of Health predoctoral traineeships to J.H .R . and

176 P. M. Kane and othersC .T .Y ., and by grants from the National Institutes of Health and the Searle Scholars Foundation to T .H .S .

ReferencesAmm erer, G ., H un ter , C . P ., R othman, J . H ., Va l l s , L. A. & S teven s, T . H. (1986). PEP4

gene of Saccharomyces cerevisiae encodes proteinase A, a vacuolar enzyme required for processing of vacuolar precursors. Molec. cell. Biol. 6, 2490-2499.

An raku , Y. (1987). Active transport of amino acids and calcium ions in fungal vacuoles. In Plant Vacuoles (ed. B. Marin), pp. 255-265. New York, London: Plenum Press.

Ara i, H ., B ern e , M. & F orgac, M. (19876). iV,Ar'-dicyclohexylcarbodiimide-binding proteolipid of the vacuolar H+-ATPase. J . biol. Chem. 262, 11006-11011.

A ra i, H., B ern e , M ., T e r r es , G ., T er r es , H., P uopolo , K . & F orgac, M. (1987a). Subunit composition and A TP site labeling of the coated vesicle proton-translocating adenotriphospha- tase. Biochemistry 26, 6632-6638.

A rai, H ., T e r r e s , G ., P ink, S. & F o rg a c , M. (1988). Topography and subunit stoichiometry of the coated vesicle proton pump. jf. biol. Chem. 263 , 8796-8802.

B an kaitis, V. A., J ohnson , L . M . & E m r , S. D . (1986). Isolation of yeast mutants defective in protein targeting to the vacuole. Proc. natn. Acad. Sci. U.S.A. 83, 9075-9079.

B a n ta , L . M ., R obinson , J . S ., K lio n sk y , D . J . & Em r, S. D . (1988). Organelle assembly in yeast: Characterization of yeast mutants defective in vacuolar biogenesis and protein sorting. J . cell Biol. 107, 1369-1383.

B o r d a l lo , C., S ch w en ck e, J . & S u a re z R e n d u e le s , M . S . (1984). Localization of the thermosensitive X-prolyl dipeptidyl aminopeptidase in the vacuolar membrane of Saccharo­myces cerevisiae. FEBS Lett. 173 , 199-203.

Bowman, B . J . , A lle n , R ., W esch ser, M. & Bow m an, E. J . (1988a). Isolation of genes encoding the Neurospora vacuolar ATPase. Analysis of vma-2 encoding the 57kDa polypeptide and comparison to v m a -l .J . biol. Chem. 263 , 14002-14007.

Bowman, B . J . & Bow m an, E . J . (1986). H+-ATPases from mitochondria, plasma membranes, and vacuoles of fungal cells. J . Membr. Biol. 94 , 8 3 -9 7 .

B owman, E. J . , M andala, S ., T a iz , L . & B owman , B . J . (1986). Structural studies of the vacuolar membrane ATPase from Neurospora crassa and comparison with the tonoplast membrane ATPase from Zea mays. Proc. natn. Acad. Sci. U.S.A. 83, 48-52.

Bowman, E. J . , S ieb ers , A. & A lte n d o r f , K. (19886). Bafilomycins: a new class of inhibitors of membrane ATPases from microorganisms, animal cells, and plant cells. Proc. natn. Acad. Sci. U.S.A. 85 , 7972-7976.

Bowman, E. J . , T en n ey , K . & Bowman, B . J . (1988c). Isolation of genes encoding the Neurospora vacuolar ATPase. Analysis of vma-1 encoding the 66kDa subunit reveals homology to other ATPases. .7. biol. Chem. 263 , 13 994-14001.

B ro w n , D ., G lu c k , S. & H artw ig , J . (1987). Structure of the novel membrane-coating material in proton-secreting epithelial cells and identification as an H+-ATPase. J . Cell Biol. 105 ,

1637-1648.B row n, W. J . , G ood hou se, J . & F a rq u h a r , M. G . (1986). Mannose-6-phosphate receptors for

lysosomal enzymes cycle between the Golgi complex and endosomes. J . Cell Biol. 103 ,

1235-1247.Cain , C . C . & M urphy, R. F . (1988). A chloroquine-resistant Swiss 3T3 cell line with a defect in

late endocytic acidification. J . Cell Biol. 106 , 269-277.C a p la n , M . J . , S to w , J . L ., Newman, A. P ., M ad ri, J . , A n d erson , H. C ., F a r q u h a r , M . G .,

P a la d e , G . E . & Jam ieson, J . D . (1987). Dependence on pH of polarized sorting of secreted proteins. Nature, Lond. 329 , 632-635.

C h vatch k o , Y ., H o w a ld , I. & Riezm an, H. (1986). Two yeast mutants defective in endocytosis are defective in pheromone response. Cell 46 , 355-364.

D en d a , K ., K o n ish i, J . , Oshima, T ., D a te , T . & Y o sh id a , M. (1988). The membrane- associated ATPase from Sulfolobus acidocaldarius is distantly related to F t -ATPase as assessed from the primary structure of its a-subunit. J . biol. Chem. 263 , 6012-6015.

D o u g la s , M . G ., McCammon, M . T . & V a s sa ro tti , A. (1986). Targeting proteins into mitochondria. Microbiol. Rev. 50 , 166-178.

Role of the yeast vacuolar H +-ATPase 177Fux Ai, M. & K a n a z a w a , H . (1986). Use of isolated subunits of F j from Escherichia coli for

genetic and biochemical studies. Meth. Enzym. 126, 588-594.G lu c k , S. & C a ld w e ll , J . (1987). Immunoaffinity purification and characterization of vacuolar

H +-ATPase from bovine kidney. J . biol. Chem. 262 , 15780-15789.K a e s tn e r , K . H ., R a n d a l l , S . K . & S z e , H. (1988). N, N'-dicyclohexylcarbodiimide-binding

proteolipid of the vacuolar H+-ATPase from oat roots. J . biol. Chem. 263 , 1282-1287.K o r n fe ld , S. (1986). Trafficking of lysosomal enzymes in normal and disease states. J . clin.

Invest. 77 , 1 -6 .L o ts c h e r , H .-R ., d eJo n g , C. & C ap ald i, R. A. (1984). Modification of the F 0 portion of the H +-

translocating adenosinetriphosphatase of Escherichia coli by the water-soluble carbodiimide 1- ethyl-3-[3-(dimethylamino)propyl]carbodiimide and effect on the proton channeling function. Biochemistry 23 , 4128-4134.

M akarow , M . (1985). Endocytosis in Saccharomyces cerevisiae'. Internalization of O'-amylase and fluorescent dextran into cells. EMBO jf. 4 , 1861-1866.

M akarow , M . & N e v a la in en , L . T . (1987). Transport of a fluorescent macromolecule via endosomes to the vacuole in Saccharomyces cerevisiae. J . Cell Biol. 104 , 67-75.

M a n d a la , S . & T a iz , L. (1985). Partial purification of a tonoplast ATPase from corn coleoptiles. PI. Physiol. 78 , 327-333.

M a n d a la , S. & T a iz , L . (1986). Characterization of the subunit structure of the maize tonoplast ATPase. J . biol. Chem. 261 , 12850-12856.

M a n o lso n , M . F . , O u e l le t te , B. F . F . , F il io n , M. & P o o le , R. J . (1988). cDNA sequence and homologies of the 57 kDa nucleotide binding subunit of the vacuolar ATPase from Arabidopsis. J . biol. Chem. 263 , 17987-17994.

M anolson , M . F ., Percy, J . M ., Ap p s , D. K ., X ie , X -S ., S ton e, D . K . & Poole, R . J . (1986). Endomembrane H +-ATPases from plants and animals show immunological cross-reactivity. In Membrane Proteins: Proceedings o f the Membrane Protein Symposium, (ed. S . C. Goheen, L . Hjelmeland, M . M cNamee & R. Gennis). Bio-Rad. Richmond, CA.

M a n o lso n , M . F . , R ea , P. A. & P o o le , R . J . (1985). Identification of 3-0-(4-benzoyl)benzoyl- adenosine 5 '-triphosphate- and ¿V,A''-dicyclohexylcarbodnmide-binding subunits of a higher plant H +-translocating ATPase. Jf. biol. Chem. 260 , 12273-12279.

M a tsu d a ira , P. (1987). Sequence from picomole quantities of proteins electroblotted onto polyvinyldiene difluoride membranes. J . biol. Chem. 262 , 10035-10038.

M ellm a n , I ., F u ch s, R. & H e l e n iu s , A. (1986). Acidification of the endocytic and exocytic pathways. A. Rev. Biochem. 55, 663-700.

M o o re , H-P. H., Gum biner, B. & K e l ly , R . B. (1983). Chloroquine diverts ACTH from a regulated to a constitutive secretory pathway in AtT-20 cells. Nature, Lond. 302 , 434-436.

Ohsumi, Y. & A n rak u , Y . (1981). Active transport of basic amino acids driven by a protonmotive force in vacuolar membrane vesicles of Saccharomyces cerevisiae. J . biol. Chem. 256 ,

2079-2082.P re sto n , R. A., M urphy, R. F . & Jo n e s , E. W. (1987). Apparent endocytosis of fluorescein

isothiocyanate-conjugated dextran by Saccharomyces cerevisiae reflects uptake of low molecular weight impurities, not dextran. J . Cell Biol. 105 , 1981-1987.

R an d a ll , S. K . & Sze, H. (1987). Probing the catalytic subunit of the tonoplast H +-ATPase. J . biol. Chem. 262, 7135-7141.

R e a , P. A., G riffith , C. J . & S a n d ers, D . (1987). Purification of the jV,A?'-dicyclohexylcarbodii- mide-binding proteolipid of a higher plant tonoplast H+-ATPase. J . biol. Chem. 262, 14745-14752.

Riezm an, H. (1985). Endocytosis in yeast: several of the yeast secretory mutants are defective in endocytosis. Cell 40 , 1001—1009.

Riezm an, H ., C h vatch k o , Y . & D u lic , V. (1986). Endocytosis in yeast. Trends Biochem. Sei. 11,

325-328.R o berts, C. J . , Pohlig , G ., R othman, J . H. & S teven s, T . H. (1988). Dipeptidyl aminopepti-

dase B , an integral membrane protein of the yeast vacuole, utilizes early stages of the secretory pathway for transport. J . Cell Biol, (in press).

R othm an, J . H. & S tev en s , T . H. (1986). Protein sorting in yeast: mutants defective in vacuole biogenesis mislocalize vacuolar proteins into the late secretory pathway. Cell 47 , 1041-1051.

R othman, J . H. & S teven s, T . H. (1988). Protein sorting and biogenesis of the lysosome-like

178 P. M. Kane and othersvacuole in yeast. In Protein Transfer and Organelle Biogenesis, (ed. R. C. Das & P. W. Robbins), pp. 317-362. New York: Academic Press.

R othm an, J . H ., H ow ard , I. & S tev en s, T . H. (1989a). Characterization of genes required for protein sorting and vacuolar function in the yeast Saccharomyces cerevisiae. EMBO J . 8 (in press).

R othman, J . H., Y amashiro , C. T . , R aymond, C. K ., K ane P. M. & Steven s , T . H. (19896). Vacuolar acidification and H+-translocating ATPase activity are deficient in two yeast mutants that fail to sort vacuolar proteins. J . Cell Biol, (submitted for publication).

S chekman , R. (1985). Protein localization and membrane traffic in yeast. A. Rev. Cell Biol. 1, 115-143.

S ch n e id e r , E. & A lte n d o r f , K . (1986). Proton-conducting portion (Fo) from Escherichia coli ATP synthase: Preparation, dissociation into subunits, and reconstitution of an active complex. Meth. Enzym. 126 , 569-578.

S e b a ld , W ., G r a f , T . & L u k in s , H. B. (1979). The dicyclohexylcarbodiimide-binding protein of the mitochondrial ATPase complex from Neurospora crassa and Saccharomyces cerevisiae. Eur. J . Biochem. 93 , 587-599.

S te v e n s , T . H., Esm on, B. & Schekm an, R. (1982). Early stages in the yeast secretory pathway are required for transport of carboxypeptidase Y to the vacuole. Cell 30 , 439-448 .

S u a re z R e n d u e le s , M. P ., Sch w en ck e , J . , G a rc ia A lv a re z , N. & G a sco n , S . (1981). A new X-prolyl-dipeptidyl aminopeptidase from yeast associated with a particulate fraction. FEBS Lett. 131 , 296-300.

Su n , S -Z ., X ie , X -S . & S to n e , D . K . (1987). Isolation and reconstitution of the dicyclohexylcar- bodiimide-sensitive proton pore of the clathrin-coated vesicle proton translocating complex. J . biol. Chem. 262 , 14790-14794.

U chid a , E ., Ohsumi, Y . & A n rak u , Y . (1985). Purification and properties of a H+-translocating, Mg2+-dependent-adenosine triphosphatase from vacuolar membranes of Saccharomyces cerevi-. s ia e .J . biol. Chem. 260 , 1090-1095.

U ch id a , E ., Ohsumi, Y. & A n rak u , Y. (1988). Characterization and function of catalytic subunit a of H + -translocating adenosine triphosphatase from vacuolar membranes of Saccharomyces cerevisiae. J . biol. Chem. 263 , 45-51.

V a l l s , L . A., H u n te r , C. P ., R othm an, J . H. & S tev en s, T . H . (1987). Protein sorting in yeast: the localization determinant of yeast vacuolar carboxypeptidase Y resides in the propeptide. Cell 48 , 887-897.

von F igura , K . & Ha silik , A. (1986). Lysosomal enzymes and their receptors. A. Rev. Biochem. 55, 167-193:

W eism an, L . S ., B a c a l la o , R. & W ick n er, W . (1987). Multiple methods of visualizing the yeast vacuole permit evaluation of its morphology and inheritance during the cell cycle. J . Cell Biol. 105 , 1539-1547.

W oolford , C. A ., D a n ie ls , L . B ., Park , F . J . , J o n es , E . W ., van Ar sd e l l , J . N. & I n n is , M. A. (1986). The PEP4 gene encodes an aspartyl protease implicated in the posttranslational regulation of Saccharomyces cerevisiae vacuolar hydrolases. Molec. cell. Biol. 6, 2500-2510.

Zimniak, L -, D ittr ic h , P., G o g a rte n , J . P., K ibak, H. & T a iz , L . (1988). The cDNA sequence of the 69kDa subunit of the carrot vacuolar H +-ATPase. J . biol. Chem. 263 , 9102-9112.