INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein [SNAP] receptor) proteins comprise distinct families of membrane-associated receptors that are components of the vesicle docking and fusion machinery in eukaryotes (62; reviewed in references 35 and 54). Evidence from evolutionarily divergent systems, e.g., from yeast to mammals, suggests that the role of SNAREs in protein transport is highly conserved (reviewed in references 7 and 21). SNAREs act throughout the secretory pathway to confer the trafficking of cargo-containing carrier vesicles and may also participate in compartmental organization. SNAREs found on vesicular compartments (v-SNAREs) are proposed to interact with specific cognate receptors (t-SNAREs) present on target compartments to form a prefusion SNARE complex. Thus, v-SNARE-t-SNARE (v-t SNARE) assembly confers vesicle docking; one of the first events leading to membrane fusion. Moreover, recent work has suggested that together v- and t-SNAREs may provide the minimal requirements necessary for conferring both membrane association and bilayer fusion, using a liposome-based in vitro assay (70). However, there is still some controversy surrounding the role of SNAREs in the fusion event itself (13, 65).

To confer specificity to protein transport, SNARE assembly leading to membrane fusion is expected to occur only on appropriate target membranes. Yet, the v- and t-SNAREs involved in Golgi and post-Golgi trafficking events passage through the early secretory pathway and are likely to reside together on identical endoplasmic reticulum (ER)- and Golgi-derived transport vesicles. Therefore, some mechanism(s) should prevent nonproductive SNARE partnering from occurring between opposing membranes early in the pathway. Likewise, it is reasonable to assume that other cellular mechanisms regulate specific v-t SNARE interactions in order to confer both temporal and spatial regulation of membrane docking and fusion, perhaps at the level of the target compartments themselves. Thus, certain constraints may prevent nonproductive SNARE partnering early in the pathway and, perhaps, ready SNAREs for assembly upon reaching their appropriate target membranes. Moreover, additional steps to remove these constraints may be prerequisites for docking and fusion to proceed in vivo. These steps should include dissociation of preformed SNARE complexes, as proposed earlier by Ungermann et al. (64), as well as the removal of other inhibitory constraints placed on the individual SNARE elements. Such constraints may include negative-acting SNARE regulatory proteins, which we have designated SNARE-masters (reviewed in reference 30). By definition, these regulators are expected to associate directly with specific v- or t-SNAREs and act to downregulate trafficking functions by modulating their entry into SNARE complexes. Thus, dissociation of SNARE regulators from SNAREs, or their inactivation, is expected to precede complex assembly and membrane fusion.

A conserved family of SNARE regulators which function in constitutive and regulated secretory systems is that of yeast Sec1 and its homologs found in higher organisms (1, 23, 26, 34, 47, 55). SEC1 was identified in the original sec mutant screen (45) and encodes a soluble protein that interacts with members of the Sso/syntaxin family of t-SNAREs (2, 23, 47, 48). Studies on mammalian Sec1 proteins have shown that they are not components of the SNARE complex and act to restrict SNARE partnering, probably by preventing t-SNAREs from assembling into binary complexes (24, 34, 48). In addition, studies in vivo have shown that the overexpression of rop1, a Drosophila homolog of Sec1, inhibits neurotransmitter release at the neuromuscular junction in larvae (60). Thus, Sec1-like molecules may act as potential SNARE-masters for the Sso/syntaxin family members by virtue of their ability to inhibit t-SNARE functioning.

A possible regulator of v-SNAREs in regulated exocytosis is synaptophysin, a membrane protein from synaptic vesicles (37, 71) which forms complexes with members of the synaptobrevin/VAMP (vesicle-associated membrane protein) family (10, 19, 69). Interestingly, these complexes were shown not to contain any of the known t-SNARE partners (19, 22, 69), suggesting that the synaptophysin-synaptobrevin interaction may prevent synaptobrevin from undergoing assembly into the ternary complex. The mechanism by which these proteins disassociate to allow for formation of the fusion complex remains unknown.

A large class of potential SNARE regulators consists of the Rab GTPases, which appear to confer SNARE complex assembly. Rabs have been suggested to regulate the specificity in membrane trafficking steps, due to their distinct subcellular localizations and interactions with SNARE components (reviewed in reference 46). Rabs have been implicated at the level of SNARE activation, and in the case of ER-Golgi transport in yeast, the Ypt1 GTPase has been proposed to bind directly to the Sed5 t-SNARE and to displace the Sec1 homolog, Sly1 (42). Genetic studies performed with yeast have demonstrated that overexpression of v- and t-SNAREs can suppress the temperature sensitivity of mutant Rab proteins that are specific to the transport step on which the SNAREs function (9, 16, 40). Thus, Rab proteins not only may regulate the fidelity of docking and fusion but could serve to dissociate proteins that act as negative regulators for SNARE complex formation (30). More recently, Rabs have been suggested to act in the tethering of vesicles to their acceptor compartments, prior to SNARE assembly (12, 65).

Another class of SNARE regulators includes the NSF/Sec18 and SNAP/Sec17 protein families which function upon membrane transport at various stages of the secretory pathway. These proteins were originally proposed to mediate both the disassembly of ternary SNARE complexes and ATP-dependent membrane fusion but more recently have been proposed to prime the docking step in SNARE assembly (64, 74). Such proteins appear to act as general SNARE regulators and, thus, disassemble preformed v-t SNARE complexes present in the same membrane (64) and, potentially, regulate either the association or dissociation of other, more specific, SNARE regulators. In the latter example, a novel SNARE regulator that participates in homotypic vacuolar fusion was identified. LMA1 (72, 73) is a heterodimer consisting of thioredoxin and the I^B₂ protease B inhibitor, which cooperates with (and may be initially bound to) Sec18 to release Sec17 and to stabilize the activated t-SNAREs (64, 74). Moreover, LMA1, like Sec17/Sec18, may act on different levels of the secretory pathway in yeast (5).

As no structural homologs of the synaptophysin family are encoded by the yeast genome, it is unclear whether negative regulators (aside, perhaps, from Sec1) act upon constitutive exocytosis in lower eukaryotes. To identify potential SNARE-masters for SNAREs which function in exocytosis, we used the two-hybrid system to screen for proteins that interact with the Snc v-SNAREs. The yeast Snc1 and Snc2 proteins (28, 49) are archetypal exocytic v-SNAREs which bear high structural homology to members of the synaptobrevin/VAMP/cellubrevin family of vesicle-associated membrane proteins (6, 44, 63) and are engaged in similar functions (14, 29, 49; reviewed in references 7 and 21). Like their brethren, Snc proteins localize to secretory vesicles (49) and interact physically with t-SNAREs from the plasma membrane (e.g., Sec9, Sso1, and Sso2) to form a ternary SNARE complex in vitro (9, 53). This was confirmed in vivo by genetic studies which showed that cells bearing disruptions in both SNC genes (a conditional lethal effect) and possessing a temperature-sensitive (ts) allele of those late-acting SEC genes which are involved in membrane fusion (e.g., sec17-1, sec9-4, and sso2-1) are inviable (14, 17, 29). Structure-function analyses have shown that the region of these SNAREs required to mediate exocytosis and cell viability localizes to the -helical portion of the cytoplasmic domain (29). This region is conserved evolutionarily and is required for VAMP to form SNARE complexes in vitro (11, 36) and in permeabilized cells (50).

Here, we describe the identification of a protein that acts as a potential SNARE-master for the Snc v-SNAREs. The Vsm1 (v-SNARE-master 1) protein was identified by using the yeast two-hybrid system and was found to coimmunoprecipitate preferentially with the Snc2 v-SNARE. Deletion of VSM1 in yeast results in the enhanced secretion of proteins into the medium, and overexpression of the VSM1 gene was found to (i) inhibit the secretion and growth of cells bearing a mutant Sec9 t-SNARE, (ii) block rescue of the sec9-4 mutant allele by overexpression of the Snc1 v-SNARE, and (iii) result in the accumulation of low-density secretory vesicles (LDSVs) at the bud tips. On the basis of these results, we propose that Vsm1 may function as a negative regulator of exocytosis in yeast.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Media, DNA, and genetic manipulations. DNA endonucleases and modification enzymes were used as recommended by the suppliers. Molecular cloning techniques were performed as described by Sambrook et al. (56). DNA sequencing was performed by the dideoxynucleotide chain termination method (57). PCRs and subcloning of PCR products were carried out as described previously (27).

Yeast strains. Yeast strains used in this study are listed in Table 1. A VSM1 disruption strain in a wild-type background was created by transforming the SphI-SalI fragment containing the vsm1::URA3 disruption into diploid yeast strain W303. A disruption of one of the VSM1 loci was verified by Southern analysis; this diploid (VL1) was sporulated, and the resulting tetrads were dissected to yield haploid vsm1 cells (VL2 and VL3). VL2 and VL3 cells were then crossed to a variety of early and late sec mutants, end4 yeast, and myo2 yeast to give diploid strains which were sporulated and dissected to yield haploid yeast bearing both mutations (Table 1).

Antibodies, immunoprecipitation, and immunoblot analysis. A polyclonal antiserum to Vsm1 was created by expressing recombinant Vsm1 in bacteria and injecting the affinity-purified protein into rabbits. A bacterial expression construct, pHISVSM1, was created so as to allow for the expression of a His₆-tagged Vsm1 protein (molecular mass  64 kDa) in Escherichia coli TOP10 (see Fig. 3). Bacterial extracts were purified over Ni²⁺ chelation resin (Ni²⁺-nitrilotriacetic acid; Qiagen) to yield protein that was used for injection into rabbits. Polyclonal antisera 3609 and 3610 were obtained in this fashion and were used to detect both native and tagged Vsm1 in cell extracts.

Plasmids. Vectors included YEp13M4, a 2µm plasmid bearing the LEU2 marker; pAD4, a similar plasmid which bears the ADH1 constitutive promoter; pAD54, a plasmid identical to pAD4 but containing an oligonucleotide encoding a peptide derived from HA; and pAD6, a similar plasmid which bears an oligonucleotide encoding a peptide derived from the Myc protein. Directional subcloning into pAD54 or pAD6 allows for the in-frame fusion between the sequence encoding the epitope and the coding region of the subcloned gene of interest. Centromeric vectors included pRS315, which bears a LEU2 marker, and pSE358, which bears a TRP1 marker.

Cellular fractionation and vesicle preparations. Cellular fractionation of yeast cells was performed by standard techniques. Briefly, cells (25 OD₆₀₀ [optical density at 600 nm] units) were harvested during log phase and lysed using glass beads in 300 µl of lysis buffer containing 25 mM potassium phosphate (pH 7.0), 100 mM NaCl, 2 mM EDTA, and the following protease inhibitors: aprotinin (1 µg/ml), leupeptin (2 µg/ml), pepstatin (1 µg/ml), soybean trypsin inhibitor (10 µg/ml), and phenylmethylsulfonyl fluoride (1 mM). After lysis, extracts were spun at 600 × g for 4 min to remove intact cells and cell wall debris to yield the total cell lysate; the latter was then centrifuged at 10,000 × g for 10 min to yield the S10 supernatant and P10 pellet fractions. The S10 was next centrifuged at 100,000 × g for 1 h to yield the S100 supernatant and P100 pellet fractions. Protein determination was performed by using the micro-bicinchoninic acid technique (Pierce). For experiments involving the treatment of membranes with either high salt (0.5 to 1.5 M NaCl) or low pH (200 mM glycine [pH 2.5]), aliquots of the S10 fraction were treated with the above reagents for 15 min on ice prior to centrifugation at 100,000 × g for 1 h. Quantitation of proteins present in the various fractions was performed by using specific antisera and quantitative Western analysis (described above).

Enzymatic assays. Acid phosphatase and exoglucanase were assayed by the methods of Van Rijn et al. (66) and Santos et al. (59), respectively, as described by Harsay and Bretscher (33). Activity is expressed in arbitrary units based on absorbance measured at 415 nm. ATPase activity was assayed by the method of Bowman and Slayman (8), with some modifications of its application described by Harsay and Bretscher (33). First, substrate concentration was modified to 2 mM ATP, as higher concentrations result in an elevated background. Second, blanks were measured to determine the rate of endogenous ATP hydrolysis and were subtracted from values obtained with samples from the gradient. Inorganic phosphate was assayed by the method of Ames (3). Optical density was measured after 10 min, and activity was expressed in arbitrary units, based on absorption at 820 nm. Invertase activity was assayed by the method of Goldstein and Lampen (31) and is expressed either in arbitrary units based on absorption at 540 nm or in units, where 1 U = 1 µmol of glucose released/min/100 mg of dry cells.

Immunofluorescence and electron microscopy. For immunofluorescence, cells containing the appropriate plasmids were grown to log phase at 25°C and fixed by the direct addition of 1 M KPO₄ (pH 6.5) and 37% formaldehyde to reach final concentrations of 0.1 M KPO₄ and 3.7% formaldehyde. After 30 min, 5 to 10 OD₆₀₀ units of cells was harvested by centrifugation for 5 min at 2,000 rpm and resuspended in 5 ml of a solution containing 0.1 M KPO₄ (pH 6.5) and 3.7% formaldehyde. Cells were incubated at room temperature for 1.5 h with gentle rocking, washed once with 5 ml of 0.1 M KPO₄ (pH 7.5), and stored overnight at 4°C in 5 ml of a solution containing 0.1 M KPO₄ (pH 7.5) and 1.2 M sorbitol. Cells were harvested by centrifugation, and spheroplasts were prepared by incubating half of the cells for 30 min at 30°C in 1 ml of a solution containing 0.1 M KPO₄ (pH 7.5), 1.2 M sorbitol, 0.25 mg of zymolase per ml, and 0.5% -mercaptoethanol. Spheroplasts were pelleted at 2,000 rpm for 2 min and washed twice with a 1 ml of a solution of 0.1 M HEPES (pH 7.4) and 1.0 M sorbitol (HS buffer). Spheroplasts were then permeabilized with HS buffer containing 0.5% sodium dodecyl sulfate (SDS) at room temperature for 5 min. The cells were washed three times with 1 ml of HS buffer and resuspended in the same. About 20 µl of fixed spheroplasts was placed onto coverslips precoated with 0.1% polylysine for 20 min. Coverslips were washed with phosphate-buffered saline containing 1 mg of bovine serum albumin per ml and 0.02% Tween 20 (PBT), incubated for 1.5 h with primary antibody, and washed extensively with PBT. Next, secondary antibody (fluorescein isothiocyanate [FITC]-conjugated anti-mouse or anti-rabbit) along with normal goat serum (1:100 dilution) was added for 1.5 h, followed by extensive washing with PBT. For visualization of HA-tagged protein, preabsorbed affinity-purified anti-HA monoclonal antibody (1:1,000 dilution) was used. For the visualization of Dpm1, a commercial affinity-purified mouse monoclonal anti-Dpm1 antibody was used (3 µg/ml). For visualization of Sso, an affinity-purified polyclonal anti-Sso antibody (1:500) was used. For labeling with secondary antibody, FITC-labeled goat anti-mouse or anti-rabbit antibody (Jackson Laboratories) at a dilution of 1:200 was used. After extensive washing with PBT, coverslips were labeled with propidium iodide (1 µg/ml) in PBT for 20 min. Next, coverslips were washed extensively with PBT, aspirated, and, after addition of mounting medium, placed onto slides. Proteins were visualized by confocal microscopy.

Metabolic labeling studies. Pulse-chase studies using [³⁵S]methionine (Amersham) were performed as described previously (15). For medium protein secretion experiments, cells were labeled for 30 min and chased with medium containing excess methionine and cysteine for 30 min. Next, cells were centrifuged at 10,000 × g for 2 min, following which the medium was transferred to a fresh tube and the proteins were precipitated by using Strataclean resin (Stratagene). Precipitated proteins were washed and counted in a scintillation counter. For autoradiography, precipitates were resolved on SDS-polyacrylamide gels, which were then fixed, incubated with a fluorescence enhancer, dried, and exposed to X-ray film.

Nucleotide sequence accession number. The GenBank accession number for VSM1 is AF034895.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of a v-SNARE-interacting protein. In an effort to isolate proteins which interact directly with v-SNAREs of the late secretory pathway in yeast, we used the two-hybrid system (18) with a C-terminally truncated form of Snc2 (Snc2^4-93) as the bait. The truncated protein lacks the membrane-spanning domain of the v-SNARE, which is essential for conferring exocytosis in vivo but is not likely to be essential for forming complexes with the t-SNAREs (29, 53). In screening a yeast cDNA library fused to the transactivation domain of Gal4, we isolated one clone out of 400,000 colonies screened which conferred robust -galactosidase activity and growth in the presence of 3-aminotriazole in a plasmid-dependent fashion (Fig. 1 and data not shown). 3-Aminotriazole is a metabolic inhibitor of the HIS3 gene product which serves as a selection marker in the two-hybrid screen (18). The cDNA insert contained within the plasmid was cloned and was found to encode an open reading frame (ORF) from the yeast genome (YER143w; GenBank accession no. U18917 and AF034895). The protein encoded by this ORF consists of 428 amino acids and has a calculated molecular mass of about 48 kDa (Fig. 2). The translated sequence has no potential membrane-spanning domains and, despite bearing four possible -helical segments, has little propensity for forming coiled coils, as determined by using secondary structure algorithms (data not shown). YER143w has high homology to a smaller ORF in the fission yeast Schizosaccharomyces pombe, which encodes a putative 36-kDa protein (accession no. Q10256). The translated proteins were found to be 41% identical, suggesting that they are probable homologs. In addition, sequences with significant homology to portions of YER143w are found in Caenorhabditis elegans (accession no. U50068) and Leishmania (accession no. AE001274). These sequence blocks are 40 and 49% identical over 174 and 155 amino acid residues, respectively.

View larger version (43K): [in this window] [in a new window]   FIG. 1.   The gene product of YER143w (Vsm1) interacts with Snc24-93 in the yeast two-hybrid system. Yeast Y153 cells transformed with various plasmids were tested for -galactosidase on nitrocellulose filters. Yeast expressing the Gal4 DNA binding domain alone (DB) or fused to Snc24-93 (DB-SNC24-93) were transformed with a plasmid that expresses the transactivating domain of Gal4 (TA) fused to the YER143w (VSM1) gene product (TA-VSM1), or with a control plasmid, and were patched onto selective medium. After 2 days, the cells were replica plated onto nitrocellulose filters, freeze-fractured in liquid nitrogen, and assayed visually for -galactosidase activity (see Materials and Methods). A positive (+) control consisted of yeast expressing the Gal4 DB fused to the retinoblastoma gene product Rb and the Gal4 TA fused to the PP1 phosphatase.

View larger version (47K): [in this window] [in a new window]   FIG. 2.   The amino acid sequence of Vsm1 (YER143w). Numbers correspond to amino acid residues.

Overproduction of Vsm1 inhibits the growth of cells bearing a mutant Sec9 t-SNARE and blocks rescue by Snc1. Since YER143w encodes a potential partner for the Snc v-SNAREs, we tested for its ability, when overexpressed from multicopy plasmids, to suppress growth defects seen in ts mutants of the yeast secretory pathway. Interestingly, overexpression of YER143w was found to inhibit the growth of sec9-4 cells at temperatures normally permissive for growth (30°C) (Fig. 3) and led to an accumulation of secretory vesicles within the cells (Fig. 4b). In contrast, overproduction was not found to have effects in any of the other early or late sec mutants tested (sec1, sec2, sec4, sec5, sec6, sec7, sec8, sec10, sec15, sec18, and sec22) as well as in snc null cells (49) or in cells bearing a ts mutation in the gene encoding the Sso2 t-SNARE (sso1 sso2-1) (data not shown). Similarly, overexpression of this ORF had no effect on the growth of cells bearing mutations in other genes which have been shown to regulate protein trafficking along different stages of the secretory pathway, including the end4, myo2, vps4, vps5, vps33, vps45, and pep12 mutants (data not shown).

View larger version (19K): [in this window] [in a new window]   FIG. 3.   Overexpression of VSM1 specifically inhibits the growth of sec9-4 cells and blocks rescue by SNC1 overexpression. (A) sec9-4 cells were transformed with plasmids expressing VSM1 in single copy [VSM1(CEN)] and multicopy [VSM1(2µm)] or with a control plasmid (Control). Cells were patched onto selective medium prior to being replica plated onto fresh plates and allowed to grow for 2 days at 30°C. (B) sec9-4 cells bearing plasmids expressing VSM1 in single copy [VSM1(CEN)] and multicopy [VSM1(2µm)] or a control plasmid were transformed with a plasmid expressing SNC1 under the control of a constitutive promoter [SNC1(2µm)]. Control cells bearing an empty vector were transformed either with the SNC1 plasmid [SNC1(2µm)] or with a second empty vector (Control). Cells were patched onto selective medium prior to being replica plated onto fresh plates and allowed to grow for 2 days at 30 and 34°C. (C) sec9-4 and sec9-7 yeast strains were transformed either with a control plasmid (YEp13M4) or with a plasmid expressing VSM1 in multicopy [VSM1(2µm)]. Cells were grown to saturation under selective conditions, seeded at a density of 0.02 OD600 units/ml in fresh medium, and allowed to grow till saturation at 26°C. OD600 was measured at different times.

View larger version (161K): [in this window] [in a new window]   FIG. 4.   Overexpression of VSM1 in sec9-4 cells results in the accumulation of secretory vesicles at the bud tip. sec9-4 cells bearing a control plasmid (A) or a multicopy plasmid which expresses VSM1 (B) were grown to log phase and fixed for thin sectioning and electron microscopy. Thin sections of wild-type cells (W303-1a) (C) and vsm1 cells (VL2) (D) are also shown for comparison. Bars, 1 µm.

Analysis of the disruption of VSM1. Disruption of the VSM1 gene was performed by homologous recombination. A URA3 selectable marker was cloned into the VSM1 locus carried on a genomic fragment of DNA and was transformed into a diploid wild-type strain. After sporulation and tetrad dissection, all spores were found to have undergone germination, and analysis of the resulting meiotic segregants showed no obvious defects in vsm1::URA3 cells. Cells bearing a disruption in VSM1 were found to grow normally, were not ts, and did not result in the accumulation of secretory vesicles or other membrane, as revealed by electron microscopy (data not shown and Fig. 4d). In addition, no defects in either the kinetics or extent of the intracellular processing of invertase, a secreted enzyme, were detected in cells lacking VSM1 (data not shown). Likewise, no defects in the uptake and internalization of an endocytic marker, FM4-64, were detected (data not shown).

sso1 sso2-1

Vsm1 is membrane associated and may localize to the plasma membrane. We examined the expression and localization of Vsm1 in yeast cells by Western analysis and confocal microscopy. Tools created to perform these experiments included an epitope-tagged form of Vsm1, which contains the HA epitope located at the amino terminus of the protein, as well as a polyclonal antiserum generated against bacterially expressed His₆-tagged Vsm1 (see Materials and Methods).

View larger version (67K): [in this window] [in a new window]   FIG. 5.   Vsm1 is a membrane-associated protein that localizes to the plasma membrane. (A) Vsm1 exists as a protein doublet in yeast extracts. Cell lysates were prepared from wild-type yeast (lane 1), vsm1 yeast expressing HA-tagged Vsm1 from single-copy (CEN) (lane 2) or multicopy (2µm) (lanes 3 and 4) plasmids, and vsm1 yeast lacking plasmids (lane 5) and were electrophoresed on SDS-polyacrylamide gels. In lanes 1 to 3 and 5, 50 µg of protein was added; 25 µg was added in lane 4. In addition, lane 6 contained 1 µg of affinity-purified recombinant His6-tagged Vsm1. Detection was performed with a polyclonal anti-Vsm1 antiserum (1:5,000). (B) Vsm1 associates with membranes and is enriched in the P100 fraction. sec6-4 cells maintained at 26°C or temperature shifted (2 h) to 37°C were fractionated as described in Materials and Methods. Aliquots of the various fractions were electrophoresed on SDS-10% polyacrylamide gels and blotted onto nitrocellulose membranes. Quantitative Western analysis was performed with specific antibodies (Ab) (anti-Gas1 [1:10,000], anti-Vsm1 [1:5,000], anti-Sso [1:5,000], and anti-Snc [1:1,000]) and 125I-labeled protein A (1 µCi/blot). In the left panel, 50-µg protein aliquots from the different fractions obtained from cells maintained at 26°C were electrophoresed; 25-µg aliquots of protein were used in the experiment shown in the middle panel. In the right panel, aliquots of the S10 fraction were treated with 1.5 M NaCl-200 mM glycine buffer (pH 2.5) or with additional lysis buffer (lacking salt or glycine) prior to centrifugation at 100,000 × g to yield the supernatant (sup.) and pellet fractions. Samples (15 µg) were electrophoresed, blotted, and detected for protein as described above. (C) Levels of Vsm1 are elevated in cells lacking vacuolar hydrolase, but not proteosome, activities. Total cell lysates (TCL) were prepared from sec6-4 yeast and strains which bear mutations in the protein degradative pathways (e.g., pre1-1 cells, which are deficient in the 20S proteosome activity, and pep4-3 prb1-1 and pep4-3 prb1 cells, which are deficient in the vacuolar hydrolase activity). Aliquots of total cell lysates (50 µg) were electrophoresed, blotted, and probed with anti-Vsm1 antibody (1:5,000). Detection was performed quantitatively, using 125I-labeled protein A (1 µCi). (D) Vsm1 localizes to the plasma membrane. Localization of Vsm1 was performed in wild-type cells by indirect immunofluorescence and confocal microscopy. a, cells expressing HA-tagged Vsm1 (expressed from a single-copy plasmid), detected with an affinity-purified anti-HA antibody (1:1,000) and FITC-labeled second antibody; b, staining of Dpm1 with an affinity-purified anti-Dpm1 antibody (3 µg/ml) and FITC-labeled second antibody; c, staining of HA-tagged Snc1 (expressed from a single-copy plasmid) with an affinity-purified anti-HA antibody (1:1,000) and FITC-labeled second antibody; d, staining of Sso protein with an affinity-purified anti-Sso antibody (1:100) and an FITC-labeled second antibody. Staining of the cell nucleus was performed with propidium iodide (1 µg/ml) and visualized via the rhodamine channel.

View larger version (61K): [in this window] [in a new window]   FIG. 6.   Vsm1 does not localize to secretory vesicles. (A) Two types of secretory vesicles accumulate in sec6-4 and sec9-4 cells. sec6-4 cells expressing HA-VSM1 from a multicopy plasmid and MYC-SNC2 from a single-copy plasmid were grown to log phase, shifted to low-phosphate-containing medium, and then either maintained at permissive conditions (26°C) or shifted to 37°C to induce vesicle accumulation. sec9-4 cells expressing HA-VSM1 and MYC-SNC2 from single-copy plasmids were grown to log phase, shifted to low-glucose-containing medium, and then either maintained at permissive conditions or shifted to 37°C to induce vesicle accumulation. Secretory vesicles from both strains were resolved by differential centrifugation and separation on 15 to 30% Nycodenz gradients. Aliquots of the fractions obtained by density gradient centrifugation were analyzed for density, protein concentration, and the following enzymatic activities: H+-ATPase, acid phosphatase (Acid Pho.), invertase, and exoglucanase (see Materials and Methods). Enzyme activities are expressed in arbitrary units based on absorbance; acid phosphatase and exoglucanase were measured at 415 nm, ATPase was measured at 820 nm, and invertase was measured at 540 nm. (B and C) HA-Vsm1 does not localize to the secretory vesicle fraction. Aliquots of fractions from the density gradients were electrophoresed, blotted, probed with anti-HA antibody (1:5,000), and detected by chemiluminescence. Samples of total cell lysates (50 µg) from sec6-4 (B) and sec9-4 (C) cells shifted to 37°C or maintained at 26°C were run along with 40-µl aliquots from the gradients (TCL). The solid arrow indicates the position of the low-density peak of vesicles (present at 37°C), while the hatched arrow indicates the high-density peak (present at 37°C). (D) Vsm1 does not localize to the LDSV population that accumulates in snc vbm cells. Aliquots (40 µl) of fractions from the density gradients shown in Fig. 6 of reference 17 were electrophoresed, blotted, probed with anti-Vsm1 antibody (1:5,000), and detected by chemiluminescence. Samples of total cell lysates (50 µg) from snc vbm1 and snc vbm2 cells were run in parallel (TCL). The solid arrow indicates the position of the single peak of vesicles. (E) Western analysis of ER, Golgi, and post-Golgi markers in the Nycodenz gradient fractions of sec9-4 cells. Aliquots (40 µl) of the fractions were electrophoresed, blotted, and probed with the following antisera: anti-Wbp1 (1:6,000), anti-Mnn1 (1:1,500), anti-Sec22 (1:2,500), anti-Emp47 (1:3,000), anti-Snc (1:500), anti-Sec4 (1:1,000), anti-Sso (1:5,000), anti-Gas1 (1:5,000), and anti-Hsp150 (1:1,000). Detection was performed by chemiluminescence. Molecular masses are indicated on the left.

Vsm1 does not localize to secretory vesicles. To determine whether Vsm1 also localizes to secretory vesicles, we determined whether the protein is present on vesicles that accumulate in late-acting sec mutants shifted to nonpermissive temperatures. Fractions enriched in the two different secretory vesicle populations, LDSVs and high-density secretory vesicles (HDSVs), which differ in both density and protein cargo content (17, 33), were separated by differential Nycodenz gradient centrifugation (Fig. 6A). In these experiments, we used both the sec6-4 and sec9-4 late-acting sec mutants to identify the LDSV and HDSV populations. We found that HDSVs fractionated at a density of 1.18 g/ml of Nycodenz and contain the soluble-secreted enzymes acid phosphatase, invertase, and exoglucanase. In contrast, LDSVs fractionated at a density of 1.156 g/ml of Nycodenz and are known to contain the Pma1 plasma membrane H⁺-ATPase activity. These values correspond well with results described by Harsay and Bretscher (33) and later by us (17). At permissive conditions, no significant amounts of enzymatic activity are found at these densities, indicating that vesicles do not accumulate (Fig. 6A, upper left and middle left panels; references 17 and 33). However, we did note that the membrane preparations from sec9-4 cells expressing VSM1 from a single-copy plasmid had elevated levels of ATPase activity in the early fractions of the gradient (see below).

LDSVs accumulate in sec9-4 cells which overexpress Vsm1. Several results indicated that Vsm1 might act upon a specific class of vesicles. As shown in Fig. 3 and 4, sec9-4 cells overexpressing VSM1 grow slowly and accumulate vesicles in the buds of dividing cells under normally permissive conditions (26°C). Yet, invertase secretion from these cells was shown to be unaffected (Table 3). Moreover, elevated levels of ATPase activity were found in the fractions of lower density obtained from sec9-4 cells that overexpress VSM1 from a single-copy plasmid (Fig. 6A, upper right panel). This result was not observed in sec6-4 cells which overexpress VSM1.

View larger version (83K): [in this window] [in a new window]

View larger version (18K): [in this window] [in a new window]   FIG. 7.   LDSVs accumulate in sec9-4 cells overexpressing VSM1. (A) sec9-4 cells expressing HA-VSM1 from a multicopy plasmid were grown to log phase, shifted to low-glucose-containing medium, and then processed to yield secretory vesicles as described in Materials and Methods. Vesicle-containing membrane fractions were resolved by differential centrifugation and separation on 15 to 30% Nycodenz gradients. Aliquots of the fractions obtained by density gradient centrifugation were analyzed for density, protein concentration, and the following enzymatic activities: H+-ATPase, invertase, and exoglucanase. Enzyme activities are expressed in arbitrary units based on absorbance; exoglucanase was measured at 415 nm, ATPase was measured at 820 nm, and invertase was measured at 540 nm. (B and C) Uranyl acetate-stained membranes from the early (B) and later (C) fractions of the gradient (fractions 4 and 5 and fractions 14 and 15, respectively). Bars represent 100 nm.

Vsm1 coprecipitates with Snc1 and Snc2 from detergent-solubilized cell extracts. To verify the results obtained from the two-hybrid assay, we determined whether Vsm1 coimmunoprecipitates with Snc protein. We immunoprecipitated an epitope-tagged form of Snc2 (HA-Snc2) which was expressed from a single-copy plasmid. Analysis of immunoprecipitates probed with anti-Vsm1 antibody showed that the Vsm1 doublet can be coprecipitated with Snc2 protein (Fig. 8A and data not shown). This interaction appears specific and could be blocked when HA peptide was added in excess to the immunoprecipitation reaction. Thus, the protein-protein interaction identified previously between Vsm1 and Snc2 appears genuine. In contrast, in the representative experiment shown we were unable to demonstrate a similar interaction between Vsm1 and HA-tagged Snc1 when HA-Snc1 was expressed via a single-copy plasmid. This finding suggested either that Vsm1 binding is specific to Snc2 or that the affinity of the Snc1-Vsm1 interaction is substantially lower.

View larger version (26K): [in this window] [in a new window]   FIG. 8.   Vsm1 coimmunoprecipitates with Snc1 and Snc2. Cell lysates prepared from wild-type yeast expressing HA-tagged Snc1 or Snc2 from either single-copy (A) or multicopy (B) expression plasmids were subjected to immunoprecipitation (IP) with anti-HA antibody (ab). Duplicate samples were immunoprecipitated with excess HA peptide (pep) (75 µg). Immunoprecipitates were electrophoresed, blotted, and probed with either anti-Vsm1 (1:5,000) or anti-HA (1:5,000) antibody. Detection was performed by chemiluminescence (A) and by 125I-protein A labeling and autoradiography (B). For panel A, cell lysates were prepared from sec18-1 cells expressing HA-Snc2 that were either shifted to 37°C or maintained at 26°C (room temperature [RT]). Immunoprecipitation and detection were performed in a similar manner. Molecular masses are indicated on the left.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In screening for interacting partners for the Snc v-SNAREs, we isolated a gene, designated VSM1, which encodes a novel SNARE-binding protein that may act at the level of secretory vesicle docking and fusion. The protein encoded by VSM1 binds tightly to the Snc2 v-SNARE and less so to Snc1, as evidenced by direct immunoprecipitation experiments (Fig. 8). This finding suggests that there may be a preference in the specificity of action of this SNARE-binding protein. Specificity in other SNARE-binding proteins has been described for the mammalian Sec1 protein, which interacts solely with syntaxins and not with either VAMP or SNAP-25 (24, 34, 48), as well as for synaptophysin, which interacts solely with synaptobrevin/VAMP (19, 69).

Genetic studies reveal important clues regarding the function of Vsm1 in yeast. First, overexpression of VSM1 in cells bearing a ts sec9-4 allele results in an inhibition of cell growth (Fig. 3) and an accumulation of vesicles (Fig. 4) at permissive temperatures and, at nonpermissive temperatures, blocks rescue conferred by overproduction of either Snc v-SNARE (Fig. 3 and data not shown). This phenotype is allele specific, as VSM1 overexpression does not inhibit the growth of cells bearing a sec9-7 allele (Fig. 3C), which encodes a mutant Sec9 t-SNARE that is capable of forming a ternary complex in vitro but does not confer secretory functions in vivo at restrictive temperatures (53). Second, we found that VSM1 overexpression specifically results in the accumulation of LDSVs which possess H⁺-ATPase activity (Fig. 7A) and a modest decrease in protein secretion into the medium (Table 2). In contrast, its deletion results in an increase in medium protein secretion (Table 3). Third, neither overexpression nor deletion of VSM1 affects the synthesis, processing, and secretion of invertase (Table 2 and data not shown). Together, these results imply that the inhibitory effect of Vsm1 overproduction may be conferred at the level of SNARE assembly, resulting in a partial block in the docking and fusion of the LDSVs. The results suggest that Vsm1 acts in a restrictive capacity prior to SNARE complex formation, although this remains to be formally proven. However, we cannot rule out an additional role for Vsm1 after membrane fusion has occurred.

Earlier results suggested that the two Snc v-SNAREs may fulfill slightly different functions although, individually, each confers normal cellular viability and the secretion of invertase and fully couples vesicle transport in the absence of the other isoform (29, 49). Previously, we had found that only SNC1 overexpression could suppress ts defects exhibited by the sec9-4 and sso2-1 alleles, whereas Snc2 either could not or did so weakly (14, 29). The basis for this finding has remained unclear, although it could reflect alternative modes of SNARE regulation, perhaps by specific SNARE binding proteins. This possibility is partially supported by the experiments performed here, which demonstrate a weaker interaction between Vsm1 and Snc1 than with Snc2 (Fig. 8). Since these v-SNAREs differ significantly only in the first 30 amino acids, the so-called variable region that is not essential for exocytic function (29), we suggest that it may be important for Vsm1 binding. The ability of Vsm1 to bind Snc proteins is expected to block productive interactions with the t-SNARE partners and, thus, prevent rescue of the ts defects in t-SNARE mutants (29), as shown here (Fig. 3B and data not shown). Yet, the deletion of VSM1 in sec9-4 cells did not render them less sensitive to temperature (data not shown). Thus, it appears unlikely that Vsm1 is the sole regulator of the Snc v-SNAREs. Deletion of VSM1 does not result in any deleterious phenotype or yield synthetic lethal interactions with known components of the secretory pathway, which indicates that Vsm1 does not fulfill an essential role in mediating cellular secretion. Alternatively, it may share redundant functions with another protein that could act more specifically upon Snc1. As no other structural homolog of Vsm1 is encoded by the yeast genome, such a regulator will have to be identified by using other techniques.

The results shown in this study suggest that Vsm1 plays a distinct role in the trafficking of one of two classes of secretory vesicles found in yeast. Vsm1 overproduction selectively results in the accumulation of the low-density class of vesicle, which indicates that LDSV docking and fusion may be more strongly influenced by the levels of this protein. What distinguishes between these two classes and renders LDSVs more sensitive to Vsm1 regulation is unclear. The results of this study suggest that the Vsm1-regulated class of vesicles, the LDSVs, are directly responsible for the delivery of medium proteins (e.g., HSP150) to the cell surface, in contrast with the HDSV class, which delivers soluble secreted periplasmic enzymes such as acid phosphatase, invertase, and exoglucanase to the surface (17, 33). Vsm1 could also play a role in the trafficking of other vesicle types in yeast (32, 38, 58), but this has not yet been determined.

Immunofluorescence (Fig. 5D) and membrane fractionation (Fig. 5B and C; Fig. 6) studies support the idea that Vsm1 is associated with the plasma membrane and not directly with secretory vesicles. This observation contrasts with our initial prediction that Vsm1 acts throughout the secretory pathway to restrict Snc v-SNARE function and suggests that their association may occur solely at the level of the plasma membrane. Thus, Vsm1 is likely to be a late-acting factor in the docking and fusion steps, in addition to its predicted inhibitory role. Alternatively, it could play a role in the retrieval of Snc proteins from the plasma membrane, although there is no pressing reason to believe that there is a deficiency in Snc protein available during LDSV biogenesis. Its precise physiological role remains, therefore, unclear, and in vitro binding studies may shed light on the nature and timing of the Vsm1-Snc interactions. Quantitative analysis of the immunoprecipitation experiments revealed that only 20% of cellular Vsm1 can be precipitated with Snc protein at native levels of expression. Vsm1 is mostly membrane associated, even in cells lacking Snc protein (Fig. 6D and data not shown), which suggests that other sites of anchorage are operant at the level of the membrane.

Here, we have cloned and initially characterized a novel SNARE binding protein that may modulate constitutive secretion from yeast. If Vsm1 is indeed a negative regulator of constitutive exocytosis, there may a temporal and spatial need to regulate the trafficking of the different vesicle populations, presumably in coordinating bud growth and septation once establishment of a secretion landmark and rearrangement of the cortical actin cytoskeleton have occurred. Interestingly, we note that VSM1 was identified independently as a gene adjacent to MAG1 and, like MAG1, is induced upon DNA damage and in the presence of cycloheximide (41). This finding suggests that under conditions in which growth arrest is paramount (i.e., during environmental stress), genes like VSM1 may be up-regulated in order to block cellular secretion and growth. Whether Vsm1 has a role in gene regulation, as suggested by Liu et al. (41), we cannot say, but such a role would seem less likely in the context of the results shown here.