Introduction

Vacuolar-type proton-translocating ATPases (V-ATPases) are found in all eukaryotic cells, and acidify a number of different organelles, including lysosomes, endosomes, the Golgi apparatus, secretory vesicles and clathrin-coated vesicles (Forgac 1989). V-ATPases have been shown to play key roles in diverse physiological processes such as receptor-mediated endocytosis, protein degradation and processing, protein sorting and targeting, coupled transport of small molecules and ions across the vacuolar membrane (Anraku et al. 1992; Stevens and Forgac 1997), and membrane fusion of vesicles (Peters et al. 2001).

V-ATPases from fungi, plants and animals share a high degree of structural similarity. Like the F1F0-ATPases in mitochondria, V-ATPases are large multimeric enzymes consisting of more than 10 different subunits, and can be functionally divided into two structurally distinct domains. The catalytic V1 domain consists of eight subunits (A–H), whereas the V0 domain contains five subunits (a, d, c, c’, c”) (Stevens and Forgac 1997). The V1 subcomplex is assembled onto the V0 subcomplex, which is composed of both peripheral and integral membrane proteins, resulting in the formation of a functional enzyme complex (Kane et al. 1999; Kane and Parra 2000).

In the budding yeast Saccharomyces cerevisiae, at least 13 subunits are required for V-ATPase function (Graham et al. 2000; Arata et al. 2002). All of the V-ATPase subunit genes have been identified in S. cerevisiae. With the exception of VPH1 and STV1, which encode homologous subunit a proteins of the V0 sector (Manolson et al. 1992, 1994), all are present as single-copy genes in the yeast genome. The nucleotide-binding sites are located in the A and B subunits of the V1 sector, and the catalytic sites are found primarily in the A subunits encoded by VMA1/TFP1 (Hirata et al. 1990). Proton translocation is postulated to occur at the interface between the A subunit and the ring of proteolipid subunits (c, c’, c”) encoded by VMA3, VMA11 and VMA16 (Kawasaki-Nishi et al. 2003). Disruption of each of the genes encoding essential V-ATPase subunits (except for VPH1 and STV1) results in the same phenotype: the mutant yeast cells cannot grow at a pH higher than 7.0 (Nelson and Nelson 1990; Yamashiro et al. 1990) and are sensitive to low and high concentrations of calcium (Ohya et al. 1986, 1991) or heavy metals such as Zn2+, Mn2+ and Cu2+ (Bachhawat et al. 1993; Eide et al. 1993; Ramsey and Gadd 1997; Szczypka et al. 1997) in the culture medium. Vacuoles in such mutants are not acidified (Nelson and Nelson 1990; Yamashiro et al. 1990), and cytosolic Ca2+ levels are elevated (Ohya et al. 1991).

The fission yeast Schizosaccharomyces pombe, which is taxonomically and evolutionarily distant from the budding yeast (Russell and Nurse 1986), is genetically and physiologically well characterized. We have recently isolated and characterized an S. pombe mutant which is deficient for vacuolar function, and has a deletion in the vps33 + gene (Iwaki et al. 2003). Although vps33Δ cells are viable at low temperatures, the severity of the vps33 phenotype demonstrates that the presence of a normal vacuole, and correct targeting of cellular material to it, are essential for fission yeast cells (Iwaki et al. 2003). Several genes encoding subunits of V-ATPase ( vma1 + for subunit A; vma2 + for subunit B; vma3 + for the c subunit) have been cloned and sequenced (Toyama et al. 1991; Ghislain and Bowman 1992). Comparison of the S. cerevisiae genes for V-ATPase subunits with the genome sequence of fission yeast enabled us to identify the other V-ATPase subunit genes in S. pombe, and revealed that all of these are present in single copies (our unpublished results). However, little is known about the function of V-ATPases in S. pombe.

To elucidate the functional role of the V-ATPase in S. pombe, we have constructed disruptants for the vma1 + and vma3 + genes. The phenotypes of these vma deletion mutants demonstrate that vacuole acidification by V-ATPase is essential for intracellular ion homeostasis, mating and sporulation, and for endocytosis, vacuolar biogenesis and protein transport to the vacuole.

Materials and methods

Strains, media, and materials

The E. coli strain XL1-blue (Stratagene) was used for all cloning procedures. The wild-type S. pombe strain ARC039 ( h - leu1-32 ura4-C190T) was obtained from Dr. Y. Giga-Hama (Asahi Glass Co., Yokohama, Japan), while KJ100-7B ( h 90 leu1-32 ura4) was obtained from Dr. K. Tanaka (Tokyo University). The mutants cpy1Δ ( h + leu1-32 his2 ura4-D18 ade6-M216 cpy1::ura4 +) and vps34Δ ( h + leu1-32 ura4-D18 ade6-M216 vps34::ura4 +) were constructed as described previously (Takegawa et al. 1995; Tabuchi et al. 1997). Standard rich medium (YES or YPD), synthetic minimal medium (MM), and sporulation medium (MEA) for S. pombe cells were used (Moreno et al. 1991). Restriction enzymes and other DNA-modifying enzymes were purchased from either Takara Shuzo (Kyoto, Japan) or New England BioLabs (Beverly, Mass., USA). The Expres35S Protein Labeling Mix (NEG-072) for protein labeling was from NEN Life Science Products, and FM4-64 was obtained from Molecular Probes (Eugene, Ore., USA). All other chemicals were from Sigma Chemical or Wako Pure Chemicals Co. (Osaka, Japan).

Genetic methods and plasmid construction

The general genetic methods used have been described previously (Alfa et al. 1993). S. pombe cells were transformed by electroporation as detailed elsewhere (Suga et al. 2000; Suga and Hatakeyama 2001). The construction of pREP41-Hmt1-GFP was described in a previous report (Iwaki et al. 2003). The hmt1 cDNA was amplified and cloned into pTN197, which was derived from pREP41 (Nakamura et al. 2001). The plasmid pTN381 (GFP-psy1) was obtained from Dr. T. Nakamura (Osaka City University).

Gene disruptions

The vma1 + locus was disrupted in the wild-type strain by replacing an internal vma1 gene fragment with the S. pombe ura4 gene. In order to clone the vma1 gene from chromosomal DNA of S. pombe, two oligonucleotides, sense (5′-GGTTTGACACAATTCTGCATCAAACATGC-3′) and antisense (5′-CGCGATTTAATGTGCAAAATTACTTAGGAC-3′), were designed based on the genomic vma1 sequence (SPAC343.05) in GeneDB (Sanger Centre, Hinxton, UK; available at http://www.genedb.org/genedb/pombe/index.jsp). A fragment of 2.6 kb was recovered, and ligated into the vector pGEM-T EASY (Promega). An internal Hin dIII fragment in the cloned vma1 gene was then replaced by the 1.6-kb ura4 + cassette (Grimm et al. 1988). A linearized DNA fragment harboring this disrupted vma1 gene was used for transformation of wild-type haploid ARC039 cells. Transformants were plated on selective MM-Ura medium, and colonies in which the resident vma1 gene had been replaced were identified by PCR.

The S. pombe vma3 gene (SPAC1B3.14) was amplified from chromosomal DNA by PCR using appropriate primers (5′-GGAAAAACCCGATATCCGGTAGTCAAGG-3′ and 5′-TCAATACCAATGTGCCGATGCCTTTAGACG-3′); the Kpn I-BglII fragment within the vma3 ORF was removed, and replaced with the ura4 + gene cassette. A linearized DNA fragment carrying this disrupted vma3 gene was used for transformation of wild-type strains, and the disruption was verified by PCR.

Staining of vacuoles and fluorescence microscopy

Vacuoles were labeled with FM4-64 as described previously (Iwaki et al. 2003). Acidic compartments were stained with quinacrine as reported by Roberts et al. (1991). Log-phase yeast cells were harvested, resuspended in 0.5 ml of YPD buffered with 50 mM Na2HPO4 (pH 7.6) containing 200 μM quinacrine, and incubated at room temperature for 5 min. Cells were sedimented at 10,000× g for 5 s, washed once with 0.5 ml of 2% glucose buffered with 50 mM Na2HPO4 (pH 7.6), and resuspended in the same solution. Samples were viewed immediately with a fluorescence microscope. Fluid-phase endocytosis was by observed fluorescence microscopy after cells had been treated with Lucifer Yellow CH (Sigma). The procedure for staining with Lucifer Yellow has been described by Murray and Johnson (2001). Briefly, cells were grown to exponential phase in YES medium, and collected by centrifugation. Cells were incubated at 27°C for 60 min in 0.5 ml of YES medium containing 5 mg/ml of Lucifer Yellow. After washing of the cells, dye that had accumulated in the vacuole was visualized by fluorescent microscopy. Stained cells were observed with a fluorescence microscope (Model BX-60; Olympus, Tokyo, Japan) equipped with a U-MGFPHQ filter set (Olympus) for quinacrine, Lucifer Yellow CH and GFP, and a U-MWIG filter set (Olympus) for FM4-64. Images were captured with a Sensys Cooled CCD camera using MetaMorph (Roper Scientific, San Diego, Calif., USA) and were saved as Adobe Photoshop files on a Macintosh G4 computer.

Pulse-chase analysis and immunoblot analysis of the S. pombe CPY

For analyses of CPY processing, cells were pulse-labeled with Expres35S Protein Labeling Mix (NEN) for 15 min at 30°C, and chased at the same temperature for given periods. Immunoprecipitation of CPY was performed using a rabbit polyclonal antibody raised against S. pombe Cpy1p as described previously (Tabuchi et al. 1997). For the CPY colony-blot assay, freshly grown spots were replicated onto a nitrocellulose filter and incubated for 2 days; the assay was then performed as reported elsewhere (Cheng et al. 2002).

Assay for mating efficiency

For assay of mating efficiency, S. pombe cells were grown on YES medium, and then transferred to MEA plates. These were incubated for 3 days, and the cultures were examined under a microscope, in order to count the numbers of zygotes and spores. The mating efficiency (ME) was calculated with the formula ME=(2Z+2A+0.5S)÷(H+2Z+2A+0.5S), where H is the number of haploid cells, Z the number of zygotes; A the number of asci, and S the number of free spores.

Results

Vacuole acidification does not occur in fission yeast mutants deleted for V-ATPase genes

Comparison of the S. cerevisiae subunit A (Vma1p) protein sequence with subunit A from other organisms indicates overall conservation except within the VDE (Vma1-derived element) (Table 1). VDE is a 50-kDa segment that self-splices from the middle of the 119-kDa precursor protein (Hirata et al. 1990; Kane et al. 1990). The VDE is a specific endonuclease that can cleave this gene during meiosis (Gimble and Thorner 1992). To our knowledge, this self-splicing segment has been described in only two organisms, S. cerevisiae and Candida tropicalis (Kane et al. 1990; Gu et al. 1993). The sequence of subunit c is also conserved (Table 1). The amino acid sequence of the predicted product of S. pombe vma3 + shares 57% identity with the S. cerevisiae Vma3p sequence.

Table 1 Sequence comparison of V-ATPase subunits A and subunits c from various organisms
Full size table

To investigate the functional role of V-ATPase in fission yeast, we constructed deletion-mutants for vma1 + (Ghislain and Bowman 1992) and vma3 + (Toyama et al. 1992). We sought to determine the physiological consequences of loss of V-ATPase function within defined time parameters by studying vacuolar acidification using quinacrine fluorescence. Quinacrine is a weakly basic dye that accumulates in acidic compartments in response to proton gradients, and is frequently used to assess the state of vacuolar acidification (Umemoto et al. 1990; Roberts et al. 1991; Morano and Klionsky 1994). The wild-type strain as a control, vma1 -disrupted ( vma1Δ) cells and vma3 -disrupted cells ( vma3Δ) were treated with quinacrine. Wild-type cells showed strong fluorescence localized to the vacuoles (Fig. 1A). The vma1Δ and vma3Δ mutants did not show vacuolar fluorescence, indicating that the vacuole could not be acidified in these disruptants (Fig. 1A).

Fig. 1A, B
figure 1

General effects of disruption of V-ATPase genes. A Quinacrine staining of acidic compartments. Wild-type (WT), vma1Δ and vma3Δ cells were incubated with 200 μM quinacrine for 5 min at room temperature and washed as described in Materials and methods. Cells were then viewed immediately using either Nomarski optics or a fluorescence microscope. B Gene disruptants are sensitive to elevated pH and CaCl2. Wild-type (WT), vma1Δ and vma3Δ cells were grown at 30°C for 3 days on YES plates containing 100 mM CaCl2 or supplemented with 50 mM MOPS (pH 7.0) and adjusted to pH 7.0 with NaOH

Full size image

Disruption of S. cerevisiae VMA1/TFP1 or VMA3 affects cell growth at neutral pH (Nelson and Nelson 1990; Yamashiro et al. 1990) and confers sensitivity to divalent cations, such as Ca2+, Cu2+ and Zn2+ (Ohya et al. 1986, 1991; Bachhawat et al. 1993; Eide et al. 1993; Ramsey and Gadd 1997; Szczypka et al. 1997). A vmaA disruptant of Aspergillus oryzae, which lacks a catalytic subunit of the vacuolar H+-ATPase, also displays a pH-dependent growth defect (Kuroki et al. 2002). Similarly, a vma-1 mutant of Neurospora crassa, which is deficient for another V-ATPase catalytic subunit, cannot grow in medium buffered to pH 7.0 or above, or in medium supplemented with Zn2+ (Bowman et al. 2000). We therefore examined the ability of our mutants to grow at neutral pH. We found that S. pombe vma1Δ and vma3Δ cells grew poorly at pH 7.0 (Fig. 1B) and could not grow at pH 7.5 (data not shown). Compared to the wild-type strain, vma1Δ and vma3Δ also showed strong sensitivity to 100 mM CaCl2 (Fig. 1B), 1 mM CuCl2 (or CuSO4) and 5 mM ZnCl2 (data not shown).

The morphology of vacuoles in vma mutants of S. cerevisiae and Ashbya gossyppii has been reported to be normal (Yamashiro et al. 1990; Förster et al. 1999). Smaller or fragmented vacuoles were observed in the vmaA -disrupted strain of A. oryzae grown at pH 8.5 (Kuroki et al. 2002). Vacuoles in vma-1 mutant strains of N. crassa were irregular, often misshapen, and frequently multilamellar (Bowman et al. 2000). As determined by Nomarski optics (see Figs. 1 and 2), vma1Δ and vma3Δ cells showed abnormally large vacuoles. This analysis strongly suggested that vacuolar morphology was significantly altered by inactivation of V-ATPase in S. pombe.

Fig. 2A, B
figure 2

Effects of disruption of V-ATPase genes disruption on endocytosis. A Endocytosis of the fluorescent endocytic marker FM4-64 into the vacuole is inhibited in V-ATPase disruption mutants. Wild-type (WT), vma1Δ, vma3Δ, vps34Δ, and ste12Δ cells were stained with FM4-64 as described in Materials and methods, and photographed after a 90-min chase in YES medium. The panels on the left show views in Nomarski optics; the right panels show FM4-64 fluorescence in the same cells. B Accumulation of Lucifer Yellow CH is reduced in V-ATPase disruption mutants. Wild-type (WT), vma1Δ and vma3Δ cells were incubated for 60 min at 28°C in YES medium containing Lucifer Yellow (5 mg/ml). The washed cells were viewed with a fluorescence microscope ( right) and with Nomarski optics ( left)

Full size image

Endocytosis is affected in V-ATPase mutants

FM4-64 is a lipophilic styryl dye, and has been used as a marker for the endocytic pathway and for vacuoles in budding yeast (Vida and Emr 1995). We used this vital dye to study the vacuolar morphology in V-ATPase mutants. Wild-type, vma1Δ and vma3Δ cells were exposed to FM4-64 for 30 min, and then subjected to a 90-min chase in YES medium. Numerous vacuoles were visible in the wild-type strain. In vma1Δ and vma3Δ cells, intermediate compartments were stained (Fig. 2A), and no vacuole staining was observed after an additional 30- to 60-min chase (data not shown). FM4-64 has been shown to be endocytosed via a route involving the prevacuolar compartment (Vida and Emr 1995), and therefore the stained structures in vma1Δ and vma3Δ cells may be prevacuolar compartments. In contrast, vacuoles were stained under the same conditions in the mutants vps34Δ (Takegawa et al. 1995) and ste12Δ (Morishita and Shimoda 2000), which have large vacuoles. These observations demonstrate that the vacuolar membrane staining with FM4-64 does not correlate with an aberrant vacuolar morphology and that V-ATPase is involved in endocytosis.

Nonspecific fluid-phase endocytosis in yeast cells can be easily monitored microscopically using the membrane-impermeable fluorescent dye Lucifer Yellow CH. In S. cerevisiae, the vma1 deletion strain shows a marked reduction in the fluid-phase uptake of Lucifer Yellow (Liu et al. 1997), and the vma3 disruption strain is completely unable to take up the dye (Umemoto et al. 1990). To determine whether the V-ATPase mutants of S. pombe exhibit blocks in fluid-phase endocytosis, the accumulation of Lucifer Yellow was examined. Vacuoles of wild-type cells were seen as fluorescence-positive compartments as a consequence of endocytosis (Fig. 2B). There were no such brightly fluorescent compartments in either the vma1Δ cells or the vma3Δ cells. This observation suggests that the acidified vacuolar compartment is required for endocytosis in S. pombe cells.

Effects of disruption of genes for V-ATPase subunits on the delivery of vacuolar proteins

As shown above, the disruption of a V-ATPase subunit gene, either vma1 or vma3, causes loss of vacuolar acidification, and affects endocytosis and vacuolar morphology. We therefore examined whether this type of acidification-defective vacuole can function serve as a target for vacuolar proteases or vacuolar membrane proteins. In S. cerevisiae, a kinetic delay in the processing of carboxypeptidase Y (CPY) occurs (Umemoto et al. 1990; Klionsky et al. 1992; Morano and Klionsky 1994), and missorting of CPY into the periplasmic space and/or medium takes place, in V-ATPase mutants (Umemoto et al. 1990; Klionsky et al. 1992; Bonangelino et al. 2002). We examined the sorting of S. pombe carboxypeptidase Y (SpCpy1p) in the vma1Δ and vma3Δ strains. During the initial 15 min of labeling, the ER- and Golgi-specific precursor form (proCPY) and a small amount of the vacuole-specific mature form (mCPY) were produced in the wild-type cells. After a 30-min chase, proCPY is almost completely converted to the mature form (Fig. 3A; Tabuchi et al. 1997). The vma1Δ and vma3Δ mutants, in contrast, were defective in processing SpCpy1p. After a 30-min chase, much less proCPY had been converted into mCPY in the mutant than in the wild-type cells. To confirm the missorting of SpCPY to the cell surface in these mutants, we employed a CPY colony blot assay that directly tests cells for secretion of SpCpy1p. In wild-type cells, SpCpy1p is efficiently sorted to the vacuoles, and is therefore not detected by the assay. In contrast, the vma1Δ and vma3Δ cells showed secretion of SpCPY (Fig. 3B). These results indicate that V-ATPase activity is required for delivery of SpCpy1p to the vacuole in S. pombe.

Fig. 3A, B
figure 3

Disruption of genes for V-ATPase subunits results in mislocalization of the S. pombe CPY protein. A Processing of SpCPY in vivo. Wild-type (WT), vma1Δ and vma3Δ cells were pulse-labeled with Express-35S label for 15 min at 30°C and chased for 30 min. The immunoprecipitates were fractionated on an SDS-10% polyacrylamide gel. The autoradiograms of the fixed, dried gels are shown. The positions of proCPY (110 kDa) and mature CPY (mCPY: 32 kDa) are indicated. B Filter immunoblot for the detection of secreted SpCPY. The indicated strains were grown on MM plates in contact with the nitrocellulose filter at 28°C for 3 days. The filter was processed for immunoblotting using a rabbit polyclonal antibody against Cpy1p. cpy1Δ was used as a negative control and vps34Δ was used as a positive control for mis-sorting of Cpy1p

Full size image

A marker protein, Hmt1-GFP, was used as a probe to monitor the delivery of vacuolar membrane proteins (Iwaki et al. 2003). The hmt1 + gene encodes an ABC-type transporter protein required for cadmium tolerance (Ortiz et al. 1992). Fission yeast responds to cadmium stress by inducing the synthesis of phytochelatins, a family of glutathione-related peptides having the structure (γ-Glu-Cys) n -Gly (where n=2–11) (Cobbett 2000). The Hmt1 protein is located in the vacuolar membrane, and is responsible for the transport of phytochelatins and phytochelatin-Cd complexes into the vacuole (Ortiz et al. 1992, 1995). In wild-type cells, Hmt1-GFP fluorescence was found in the vacuolar membrane as described previously (Fig. 4, Iwaki et al. 2003). Comparison of fluorescence patterns of Hmt1-GFP with the corresponding Nomarski images showed that Hmt1-GFP resides predominantly on the vacuolar membrane in vma1Δ and vma3Δ cells (Fig. 4). In agreement with these observations, vma1Δ and vma3Δ cells were tolerant to 100 μM CdCl2 (data not shown). These results suggest that the vacuolar membrane protein Hmt1p is correctly transported to the vacuoles; however, the sorting of the soluble vacuolar protein CPY is significantly inhibited because the intracellular compartment is not adequately acidified in V-ATPase mutants of S. pombe.

Fig. 4
figure 4

Intracellular localization of Hmt1-GFP. Wild-type (WT), vma1Δ and vma3Δ cells were grown in MM-leu medium supplemented with thiamine (5µg/ml) at 28°C. The cells were transferred to MM-leu without thiamine medium for 16 h, and then incubated with water for 2 h to induce vacuolar fusion. The Hmt1-GFP fusion protein was visualized using Nomarski optics and fluorescence microscopy

Full size image

Mating efficiency is reduced in the vma3Δ mutant

In A. gossypii, the vma1 gene disruption mutant fails to form generative spores (Förster et al. 1999). The vma-1 mutant strains of N. crassa can not produce conidia, asexual spores, or perithecia (Bowman et al. 2000). The fission yeast exists as two mating types, h + and h -. When haploid cells of opposite types are shifted to a nitrogen-free medium they conjugate to form a diploid zygote, which then undergoes meiosis and sporulation (Egel 1989). To assess the influence of the defect in vacuolar acidification on these processes in S. pombe, a vma3 disruption mutant was derived from a homothallic strain, which was then streaked onto MEA medium and observed by microscopy. After incubation for 3 days, the mating efficiency (ME) of the wild-type h 90 cells was 71.1%, and that of the vma3Δ was 53.7% (Fig. 5A). Hence the vma3Δ cells were not sterile under these conditions. When homothallic strains were incubated in the MEL medium for 24 h under aerobic conditions, vma3Δ cells did not aggregate, and only a few mated diploids were observed (ME=9.3%). Under the same conditions, wild-type cells formed large aggregates including diploids, spore-containing zygotes and haploid cells. In this case, the ME was 30.5%.

Fig. 5A, B
figure 5

Mating and sporulation behavior of the homothallic strain with a disrupted vma3 gene ( A) and intracellular localization of Psy1-GFP ( B). A Wild-type KJ100-7B ( h 90 leu1-32 ura4) and vma3Δ ( h 90 leu1-32 ura4 vma3::ura4 +) cells were cultured on ME plates at 28°C for 3 days, and then observed using Nomarski optics. B Intracellular localization of Psy1-GFP in homothallic strains. Wild-type (WT) KJ100-7B and vma3Δ cells were grown in MM-leu medium at 28°C overnight, then observed using Nomarski optics and fluorescence microscopy

Full size image

Furthermore, the distribution of GFP-tagged syntaxin-like protein Psy1p was examined. In the wild-type strain, the fluorescence was found to be distributed in the plasma membrane in vegetative cells, and in the forespore membrane in mature spores (Fig. 5B, Nakamura et al. 2001). The same pattern was observed in the vma3Δ cells (Fig. 5B). These results suggest that the V-ATPase takes part in the mating of fission yeast cells, but the delivery of Psy1p to the forespore membrane is not inhibited in vma3Δ cells.

Discussion

Genes encoding V-ATPase subunits have been identified in a variety of organisms, including humans, mice, Drosophila melanogaster , Caenorhabditis elegans , higher plants and fungi. Deletion, or interference with the expression, of some of these genes resulted in a lethal phenotype in animals and Dictyostelium (Davies et al. 1996; Xie et al. 1996; Dow et al. 1997; Oka and Futai 2000; Pujol et al. 2001; Choi et al. 2003). Inactivation of V-ATPase subunits may be responsible for the lack of this enzyme activity in specific cells, such as osteoclasts in mammal and H-shaped excretory cells in C. elegans, thereby affecting their function and leading to embryonic or larval lethal phenotypes (Oka and Futai 2000; Scimeca et al. 2000; Pujol et al. 2001; Choi et al. 2003). In contrast, gene disruptions in fungi were not lethal but did cause a number of growth defects (Nelson and Nelson 1990; Yamashiro et al. 1990; Ohya et al. 1986, 1991; Bachhawat et al. 1993; Eide et al. 1993; Ramsey and Gadd 1997; Szczypka et al. 1997; Förster et al. 1999; Bowman et al. 2000; Kuroki et al. 2002).

In this study, we have investigated the functional and structural roles of the V-ATPase in vacuolar acidification and protein sorting to the vacuole in S. pombe, based on the functional analysis of two disruption mutants, vma1Δ and vma3Δ. We found that neither vma1 + nor vma3 + is essential for growth, but both are indispensable for vacuolar acidification in vivo.

Phenotypes of vmaΔ mutants in S. pombe

Deletion of either of these two V-ATPase subunits had pleiotropic effects on the growth of S. pombe cells. We found that vma1Δ cells and vma3Δ cells of S. pombe showed nearly the same phenotypes. In S. cerevisiae, cells lacking a V0 subunit fail to incorporate any V-ATPase subunits into the vacuolar membrane; however, the V1 domain is stable in the cytosol (Kane et al. 1992, 1999). Loss of one V0 subunit destabilizes the other V0 subunits (Kane et al. 1992) and an assembled V0 domain is required to recruit the V1 domain to the membrane. S. pombe vma3Δ cells might not express the V1 subunits in the vacuole, resulting in loss of function of the V-ATPase. We are currently examining the localization of the V1 domain in vma3Δ cells to clarify the assembly of V-ATPase in this mutant.

The vma1Δ and vma3Δ cells showed severe pH sensitivity and did not grow at pH 7.0 or more. The basis for pH sensitivity has been investigated in S. cerevisiae vma mutants. A role for endocytosis in acidification is supported by the observation that blocking of endocytosis in a V-ATPase-deficient strain results in cell death (Munn and Riezman 1994). When the cells are grown at basic pH, fluid-phase endocytosis may be insufficient to acidify endocytic compartments in the absence of V-ATPase activity. However, an NH4 +-dependent system has been reported to be responsible for vacuolar acidification in V-ATPase-deficient yeast (Plant et al. 1999). Rich growth media contain high concentrations of ammonium in equilibrium with the unprotonated form (ammonia) at basic pH. After reaching the vacuole, ammonia is protonated, thereby elevating the vacuolar pH.

Growth of the V-ATPase mutants of S. pombe was inhibited by the divalent cations Zn2+ and Cu2+. It has been reported that the endoplasmic reticulum transporter Zhf1p and the zinc metallothionein Zym1p contribute to zinc homeostasis in fission yeast by sequestering surplus zinc in vivo (Borrelly et al. 2002; Clemens et al. 2002). These two proteins are major determinants of zinc tolerance in fission yeast, but the sensitivity to Zn2+ of the vacuole-less mutant vps33Δ (Iwaki et al. 2003) and V-ATPase mutants indicates that vacuoles are also involved in zinc homeostasis. Recently, a vacuolar copper transporter, Ctr6p, was identified in fission yeast, although it was suggested that Ctr6p serves to mobilize stored copper from the vacuole to the cytosol under conditions in which copper is scarce (Bellemare et al. 2002). Which of these divalent cations are stored in the intracellular compartments (including vacuoles) of S. pombe, and how they are sequestered, has not been elucidated, but these results indicate that the V-ATPase is critical for storage and/or detoxification of Zn2+ and Cu2+.

In addition, high concentrations of Ca2+ in medium inhibited the growth of the vma1Δ and vma3Δ strains of S. pombe, as in the case of the corresponding S. cerevisiae mutants (Ohya et al. 1986, 1991). Tight control of cytosolic Ca2+ concentrations in budding yeast is achieved primarily via a vacuolar Ca2+/H+ exchanger (Vcx1p) driven by V-ATPase, and a vacuolar Ca2+ pump (Pmc1p; Förster and Kane 2000), and V-ATPase cooperates with calcineurin to regulate the concentration of free Ca2+ in the cytosol (Tanida et al. 1995). The sequestering of Ca2+ in fission yeast is carried out by the Ca2+/H+ antiporter(s) of intracellular membranes, and no Ca2+-ATPase activity has been detected; however, genes for putative Ca2+-ATPases have been found in the genome sequence of fission yeast (Okorokov et al. 2001). Our results suggest that acidification of intracellular compartments by V-ATPase has a significant effect on Ca2+ homeostasis mediated by Ca2+/H+ antiporters in S. pombe.

Vacuolar morphology is abnormal in vma mutants

The most obvious and pronounced effects of disruption of vma1 or vma3 in S. pombe are on vacuolar morphology. The enlarged vacuoles seen in V-ATPase mutant strains of S. pombe are similar to those observed in vps34Δ and ste12 cells (Takegawa et al. 1995; Morishita and Shimoda 2000) (Fig. 2). The S. pombe Vps34p encodes a phosphatidylinositol (PtdIns) 3-kinase, and the vps34Δ strain lacks PtdIns 3-kinase activity and shows a defect in vacuolar protein transport in S. pombe (Takegawa et al. 1995; Tabuchi et al. 1997). Therefore, the Vps34p/PtdIns 3-kinase facilitates anterograde protein transport from the Golgi to the vacuole through the regulated synthesis of PtdIns(3)P. The ste12 + gene has been cloned and was found to be homologous to the S. cerevisiae FAB1 gene (Yamamoto et al. 1995), which was originally identified in studies on mutants that displayed aberrant chromosomal segregation (Morishita et al. 2002). In cells lacking Ste12p, PtdIns(3,5)P2 is undetectable, and this result implies that the ste12 + gene encodes a PtdIns(3)P 5-kinase, which synthesizes PtdIns(3,5)P2 from PtdIns(3)P (Morishita et al. 2002). Increases in vacuole surface area in S. pombe vps34 and ste12 mutants may be a consequence of defects in the turnover or efflux of the vacuolar membrane. The precise role of vacuolar acidification in vacuolar membrane turnover has yet to be determined, and we are examining the PtdIns 3-kinase activity and the level of PtdIns(3)P and PtdIns(3,5)P2 in vma1 and vma3 mutant cells.

During the fusion of vesicles, V0 sectors from opposing membranes form complexes by docking and bilayer fusion (Peters et al. 2001; Nishi and Forgac 2002). The vma3Δ mutant of S. pombe was expected to be an ideal model for studying the role of the V0 sector in vacuolar fusion in vivo, since hypotonic stress caused a transitory fusion of vacuoles in S. pombe, affecting the volume and number of vacuoles (Bone et al. 1998). However, when the vma3Δ mutant was exposed to hypotonic stress, the volume and numbers of vacuoles were indistinguishable from those seen by Nomarski optics under normal conditions (data not shown). The large vacuoles found in vma3Δ mutants suggest that V0 sectors are not involved in vacuolar fusion in S. pombe cells.

Effects of the inactivation of V-ATPase on membrane trafficking in S. pombe

The S. cerevisiae V-ATPase plays a role in membrane trafficking from endosomes to vacuoles (Stevens and Forgac 1997). Targeting of newly synthesized vacuolar enzymes from the Golgi to the vacuole is dependent on vacuolar acidification. S. cerevisiae vma strains were found to secrete significant amounts of newly synthesized vacuolar proteases, including CPY (Umemoto et al. 1990; Klionsky et al. 1992, Bonangelino et al. 2002). In S. cerevisiae, the targeting of CPY requires its binding to the CPY receptor Vps10p in the Golgi, followed by delivery via vesicles to an endosomal compartment (Marcusson et al. 1994; Cooper and Stevens 1996). Within this compartment, the low pH is hypothesized to activate release of CPY into the lumen, and recycling of Vps10p back to the trans -Golgi (Marcusson et al. 1994; Cooper and Stevens 1996). Neutralization of this compartment results in the saturation of receptors with ligand, and the secretion of vacuolar enzymes via a secretory pathway. In this study, we have shown that S. pombe V-ATPase mutants showed a delay in the transport of SpCPY and also secreted SpCPY. The mechanism required for moving SpCPY to the vacuole is similar to that noted for CPY in S. cerevisiae (Takegawa et al. 2003). Comparison of the S. cerevisiae VPS10 gene with the fission yeast genome sequence has identified a homologous gene (SPBC16C6.06) (Takegawa et al. 2003). Characterization of this vps10 homologue and the distribution of Vps10p in vma1 and vma3 mutants should reveal the mechanism used for the sorting of vacuolar proteins in S. pombe.

In S. cerevisiae, vma mutants are defective in endocytosis (Umemoto et al. 1990; Liu et al. 1997; Perzov et al. 2002). Although the vacuolar delivery of internalized FM4-64 was strongly inhibited in V-ATPase mutants of S. pombe, Hmt1-GFP was transported normally to the vacuolar membrane in these cells. The delivery of Hmt1p to the vacuoles is not well understood, but this observation may indicate that the sorting pathway for Hmt1p is distinct from the sorting pathway for SpCPY and the endocytic process. In addition, the distribution of Psy1-GFP in the vma3Δ mutant strain was normal (Fig. 5). In the wild-type strain, Psy1-GFP fluorescence was found to be distributed in the plasma membrane in vegetative cells, and in the forespore membrane in mature spores (Nakamura et al. 2001). Two possible explanations for the localization of Psy1p were proposed. One is that Psy1p on the plasma membrane is degraded at metaphase II, and the Psy1p synthesized de novo is exclusively transported to forespore membranes via the conventional ER/Golgi pathway. The other explanation is that the plasma membrane Psy1p is internalized by endocytosis and transported to the forespore membrane. Our results showed that vma3Δ cells are defective in endocytosis, but transport of Psy1p to the forespore membrane is normal. This finding suggests that Psy1-GFP is incorporated into the forespore membrane not by endocytosis from the plasma membrane, but by transport of newly synthesized protein via the normal ER/Golgi pathway.

Mating efficiency is affected in the vma3Δ strain

Mating efficiency is reduced in the S. pombe vma3Δ mutant, suggesting that V-ATPase is vital for mating in fission yeast. We have three possible explanations for this mild defect in mating. First, missorting of vacuolar proteases may diminish the activity of vacuolar proteases which are required for the nitrogen starvation-induced protein degradation that accompanies sporulation. In S. cerevisiae, vma mutants secrete precursor forms of CPY and proteinase A (Klionsky et al. 1992). It may be possible for fission yeast vma mutants to secrete vacuolar proteases—Isp6p and other unidentified proteases. A change in vacuolar pH may be partly responsible for the reduced activity in vacuoles. Second, the defect may be due to reduced endocytosis. The role of endocytosis in sporulation remains completely unknown; however, some endocytosis mutants show a defect in sporulation in budding yeast (Whitacre et al. 2001). These mutants also have a defect in the actin cytoskeleton. Therefore, the third possible explanation is that the large vacuoles in V-ATPase mutants disturb the polarization of the cytoskeleton. The cytoskeletal changes that accompany conjugation in fission yeast have been described, with cortical actin patches moving to the point of cell extension (Petersen et al. 1998). However, when vma mutants of fission yeast were grown in rich medium, the actin distribution seemed to be almost the same as in wild-type cells (data not shown). Thus, this third explanation seems less plausible than the first two.