The Na⁺/H⁺ exchanger Nhx1p regulates the initiation of Saccharomyces cerevisiae vacuole fusion

Membrane fusion of endosomal vesicles is crucial for maintaining cellular homeostasis in all eukaryotes. The machinery that controls membrane fusion is conserved from yeast to mammals (Jahn and Sudhof, 1999). We used vacuoles (lysosomes) from the yeast Saccharomyces cerevisiae to examine homotypic membrane fusion. Homotypic vacuole fusion requires soluble N-ethylmaleimide-sensitive factor (NSF) attachment protein receptors (SNAREs) and their chaperones Sec18p (NSF) and Sec17p (α-SNAP), the Rab GTPase Ypt7p and its effector homotypic fusion and vacuole protein sorting (HOPS) complex, Rho GTPase-regulated actin remodeling (Isgandarova et al., 2007) and a set of regulatory lipids that include ergosterol, diacylglycerol and phosphoinositides (Fratti et al., 2004; Jun et al., 2004; Kato and Wickner, 2001).

Membrane fusion occurs through distinct experimentally defined phases. During priming, Sec17p-bound cis-SNARE complexes are disassembled by Sec18p, releasing Sec17p (Mayer et al., 1996) and the soluble SNARE Vam7p (Boeddinghaus et al., 2002). Vam7p re-associates with vacuoles through its interactions with HOPS and phosphatidylinositol (PI) 3-phosphate (PI3P). After priming, vacuoles tether and dock in an Ypt7p-dependent cascade (Haas et al., 1995; Mayer and Wickner, 1997). During docking, SNARE complexes form in trans (between two membranes) (Ungermann et al., 1998), leading to the release of luminal calcium stores (Merz and Wickner, 2004a). Upon tethering and/or docking, vacuoles form three morphologically distinct domains. When vacuoles dock, they become tightly associated forming two flattened discs of apposed membranes termed the boundary. The perimeter of the boundary is the vertex ring and is enriched in proteins and lipids that catalyze fusion (Fratti et al., 2004; Wang et al., 2003; Wang et al., 2002). Fusion occurs at the vertex ring, resulting in the internalization of the boundary membranes as an inverted lumenal vesicle.

Ion homeostasis plays an important function in membrane trafficking. A wealth of research has been performed on understanding the roles of Ca²⁺ transport, as well as pH regulation by V-ATPases. However, the role of Na⁺(K⁺) transporters in membrane traffic is not well understood. The yeast Na⁺(K⁺)/H⁺ antiporter Nhx1p, originally identified as Vps44p, is homologous to the NHE family of transporters in mammalian cells. Nhx1p regulates vacuolar pH by transporting Na⁺ and K⁺ ions into the vacuole lumen while exporting H⁺ into the cytoplasm (Brett et al., 2005). The exchange of Na⁺(K⁺) and H⁺ is associated with regulating the trafficking of cargo from late-endosomal compartments to the vacuole (Bowers et al., 2000). Nhx1p is a polytopic protein with 12 transmembrane domains that comprise the ion transporter, and an extended cytoplasmic C-terminal domain. Whereas the ion transporter domain is well conserved throughout the NHE family, the cytoplasmic tail is highly divergent and serves a multitude of functions that include interactions with calmodulin (Yamaguchi et al., 2005), ezrin (Wu et al., 2004) and phosphoinositides (Aharonovitz et al., 2000). The cytoplasmic tail of Nhx1p is glycosylated (Wells and Rao, 2001) and interacts with the GTPase activating protein (GAP) Gyp6p that can activate both Ypt6p and Ypt7p (Ali et al., 2004; Will and Gallwitz, 2001). Deletion of NHX1 in S. cerevisiae leads to changes in vacuolar morphology, a vacuolar protein-sorting defect and the hyperacidification of the vacuole lumen (Bonangelino et al., 2002; Brett et al., 2005; Seeley et al., 2002).

In this study, we examined the role of Nhx1p in vacuole homotypic fusion. We found that vacuoles from nhx1Δ yeast are attenuated for fusion and that Nhx1p regulates the initiation of fusion. Furthermore, we found that the Nhx1p regulates fusion via ion transport and maintenance of vacuole pH.

Nhx1p has been reported to operate in vesicle trafficking between pre-vacuolar late endosomes and the vacuole (Bowers et al., 2000; Nass and Rao, 1998). Owing to the reported altered vacuole morphology and defective protein sorting, we examined whether Nhx1p might play an important role in regulating vacuole fusion. To examine the vacuolar phenotype of our deletion strains, we incubated wild-type (WT) and nhx1Δ strains with the vital dye FM4-64, which collects in the vacuole (Vida and Emr, 1995). Previously, nhx1Δ strains were shown to have a Class E vacuolar compartment, meaning that cells harbored a single large vacuole, flanked by enlarged prevacuolar compartments (PVC) (Bowers et al., 2000; Seeley et al., 2002). As expected, Class E compartments were observed; however, we also observed an increase in cells harboring multiple vacuoles (Fig. 1A). When quantitated, there was a 50% reduction in cells harboring a single vacuole (Fig. 1B). Concomitantly, there was a several-fold increase in cells with multiple vacuoles, suggesting that vacuole fusion was probably impaired in vivo. To determine whether vacuoles from nhx1Δ yeast contained the necessary components for homotypic fusion, western blot analysis was performed. No significant differences were observed in the composition of the fusion machinery on mutant vacuoles relative to their isogenic WT counterparts (Fig. 1C). Vam3p and Vam7p levels were slightly reduced on nhx1Δ vacuoles relative to WT. Vph1p levels were also reduced on nhx1Δ vacuoles (~50%) relative to WT vacuoles (Fig. 1C). This is consistent with a previous finding that showed reduced vacuole-localized Vph1p in nhx1Δ cells (Bowers et al., 2000). To document the presence of Nhx1p on purified vacuoles, we constructed a strain harboring Nhx1p-GST. Vacuoles were isolated from the WT and Nhx1p-GST strains and examined by western blot. We found that Nhx1p-GST was indeed enriched on vacuoles relative to the microsomal fraction (Fig. 2A). We also blotted for endosomal proteins that do not cycle between endosomes and the vacuole. In Fig. 2A, we observed that the ESCRT I protein Vps23p (Katzmann et al., 2001) and the ESCRT III component Snf7p (Babst et al., 2002) were enriched in the microsomal fraction and only lightly detectable in the vacuole fraction. By contrast, the vacuolar GTPase Ypt7p was highly enriched on vacuoles relative to bulk microsomes. This suggests that although vacuoles do co-purify with trace levels of endosomal contaminants, the high levels of Nhx1-GST on vacuoles was not due to contamination. The detection of robust Nhx1p-GST levels in vacuole preparations suggests that a significant fraction of the antiporter is always present on the vacuole membrane. To further document the distribution of Nhx1p, we constructed a strain of yeast harboring an Nhx1p-GFP fusion protein. In Fig. 2B, we show a single slice from a deconvolved z-stack of images of a vacuole harboring Nhx1p-GFP stained with PSS-380 to label phosphatidylserine. Here we found that Nhx1p-GFP was indeed on the limiting membrane of the vacuole and not on a neighboring endosomal vesicle. Although PVC membranes contain high concentrations of Nhx1p, these results illustrate the unambiguous presence of Nhx1p-GFP on the vacuole membrane. In whole cells, PVC-localized Nhx1-GFP was sufficiently bright to obscure vacuole labeling and quantitation. However, when isolated vacuoles were quantitated, we found that 94.1% (±3.4 s.e.m., n=541 vacuoles) harbored visible Nhx1-GFP. This illustrates that vacuole localization is not atypical and that Nhx1 is normally part of the organelle composition.

We next examined the effect of Nhx1p on vacuole homotypic fusion. Fusion is measured by the maturation of proPho8p (pro-alkaline phosphatase) by the protease Pep4p. These experiments were carried out using the reciprocally deleted strains DKY6281 (PEP4 pho8Δ) and BJ3505 (pep4Δ PHO8) (Haas et al., 1994). We deleted NHX1 from these strains to generate RFY5 and RFY6, and found that the fusion of nhx1Δ vacuoles was attenuated and showed an approximate 35% reduction in fusion relative to WT vacuoles (Fig. 3A). This was not due to altered trafficking of Pho8p or Pep4p (Fig. 1C). To investigate whether the difference was due to a kinetic defect, we employed a method shown to differentiate the first round of fusion from later ongoing rounds (Merz and Wickner, 2004b). Here we used an excess of reporter vacuoles (PHO8 pep4Δ) to report late rounds of fusion (Fig. 3B), or an excess of effector vacuoles (pho8Δ PEP4) to report early fusion (Fig. 3C). We found that nhx1Δ vacuoles were only defective in early rounds of fusion (Fig. 3C) and not later ongoing rounds (Fig. 3B). This suggests that kinetics of early fusion stages are regulated in part by Nhx1p. Mutant nhx1Δ strains were complemented with WT or mutant NHX1 encoded on pRS413 or pRS414. We found that strains complemented with WT NXH1 were rescued and fused at WT levels, demonstrating that the Nhx1p is directly involved in fusion regulation (Fig. 4A). We also used mutations that have been previously characterized by Stevens and colleagues (Bowers et al., 2000). These mutations target conserved acidic residues required for ion transport in the NHE family (Dibrov et al., 1998; Fafournoux et al., 1994; Inoue et al., 1995). We found that Nhx1p^D201N did not restore fusion (Fig. 4B), which was consistent with previous findings showing that FM4-64 and CPY trafficking were not restored with Nhx1p^D201N. Interestingly, Nhx1p^E355Q was previously shown to fully complement FM4-64 and CPY trafficking (Bowers et al., 2000), yet this mutant only partially rescued fusion in our experiments (Fig. 4C). This is consistent with a previous report that showed a partial complementation of CPY trafficking and vacuolar pH by Nhx1p^E355Q (Mukherjee et al., 2006). These data indicate that ion transport activity is required for Nhx1p-mediated fusion regulation. To determine whether the acidified lumenal pH of nhx1Δ vacuoles caused the attenuated fusion, we treated WT and mutant vacuoles with the weak base chloroquine, a lysosomotropic agent that preferentially accumulates in lysosomes. We found that treating vacuoles with chloroquine restored nhx1Δ vacuole fusion to WT levels (Fig. 4D), further illustrating that Nhx1p regulates vacuole fusion by controlling lumenal pH.

Because nhx1Δ cells contain wild-type levels of the fusion machinery, we investigated whether we could bypass the fusion defect. To stimulate nhx1Δ vacuoles, we used recombinant Vam7p, which bypasses blocks in early fusion stages such as priming and Ypt7p-dependent tethering (Fratti et al., 2007; Thorngren et al., 2004). Here, WT and nhx1Δ vacuoles were treated with anti-Sec17p IgG to inhibit priming. Reactions were rescued with GST-Vam7p or GST-Vam7p^Q283R, which disrupts the 3Q:1R SNARE paradigm and does not support fusion under native membrane conditions (Fratti et al., 2007). Both WT and nhx1Δ vacuoles were rescued by Vam7p (Fig. 5A, black symbols). As one would predict, Vam7p^Q283R did not stimulate either set of vacuoles (Fig. 5A, white symbols). These data illustrate that mutant vacuoles have all the necessary components for full fusion, and that mis-sorted cargo (e.g. CPY) has no deleterious effect. Furthermore, this demonstrates that the Pho8p reporter system is not affected by the acidified lumen of nhx1Δ vacuoles. In separate experiments, we examined whether Vam7p could stimulate fusion when vacuoles were allowed to undergo priming. Uninhibited reactions were treated with Vam7p as described above and tested for fusion. We found that exogenous Vam7p did not increase the fusion of WT or nhx1Δ vacuoles (Fig. 5B). This is consistent with our previous report showing that exogenous Vam7p does not alter uninhibited fusion reactions (Fratti et al., 2007).

Exogenous Vam7p readily enters into complexes with endogenous SNAREs and HOPS. To verify whether Vam7p-dependent rescue of nhx1Δ vacuoles also requires these complexes, we performed GST-Vam7p pull-down assays. SNARE priming was inhibited with anti-Sec17 IgG. After 15 minutes, anti-Vam3p was added to the indicated reactions and allowed to act for 5 minutes. Fusion reactions were then rescued with 400 nM GST-Vam7p and incubated for an additional 70 minutes (Fig. 6A). In agreement with the findings in Fig. 5A, fusion was restored when either WT or mutant vacuoles were treated with Vam7p (Fig. 6B). GST-Vam7p complexes were isolated and examined by western blotting. GST-Vam7p complexes included the SNAREs Vam3p and Nyv1p, and the HOPS subunits Vps16p and Vps33p (Fig. 6C). SNARE complex formation was blocked when reactions were treated with anti-Vam3p. GST-Vam7p complexes were indistinguishable between WT and nhx1Δ vacuoles, suggesting that SNARE-HOPS complex formation was not inhibited by the lack of Nhx1p. The interaction between Vam7p and HOPS was not disrupted by anti-Vam3p, possibly owing to the direct interaction of the Vam7p Phox homology (PX) domain with HOPS (Stroupe et al., 2006).

As seen in Fig. 6, exogenous GST-Vam7p was able to form complexes with free SNAREs and HOPS on nhx1Δ vacuoles when priming was blocked. However, it was unclear whether endogenous SNAREs could form trans-complexes when priming was allowed. To address this, NHX1 was deleted from the trans-SNARE reporter strain BJ3505 CBP-Vam3p nvy1Δ to create RFY7. Vacuoles from RFY6 (pho8Δ VAM3 NYV1) and RFY7 were incubated together and trans-SNARE complexes were isolated through the calmodulin binding peptide (CBP) integrated into Vam3p. CBP-Vam3p complexes were isolated using calmodulin sepharose beads and resolved by SDS-PAGE. As shown in previous figures, nhx1Δ vacuoles were attenuated for fusion (Fig. 7A); however, the addition of excess recombinant Vam7p did not rescue fusion. Although there is a modest increase in fusion from the presence of Vam7p, the nhx1Δ vacuoles still fused poorly. This might be attributable to the reduction in available Nyv1p, because RFY7 (but not RFY6) lacks this protein. Surprisingly, we found that nhx1Δ vacuoles were able to form trans-SNARE complexes as well as WT vacuoles (Fig. 7B). This suggests that nhx1Δ vacuoles might also be defective in fusion stages that occur downstream of trans-SNARE pairing.

Because of the possibility that the inability of Vam7p to rescue fusion in Fig. 7A was due to a shift in its K_m, we performed a dose-response assay of Vam7p using BJ3505 nhx1Δ or nhx1Δ nyv1Δ double deletions. These were incubated with DKY6281 nhx1Δ vacuoles and priming was blocked with anti-Sec17p. We found that, in the absence of Nyv1p (RFY7 vacuoles), Vam7p was unable to fully restore fusion at any concentration (Fig. 7C). In the presence of Nyv1p, Vam7p restored fusion to wild-type levels, as seen in Fig. 5A. In Fig. 7A, we performed fusion assays in the absence of a priming block, thus a better comparison would be to equate it with dose curves performed in the absence of a priming block. As seen in Fig. 5B, Vam7p does not affect fusion when priming was allowed and is in keeping with previous findings (Fratti et al., 2007).

The data thus far suggested that the deletion of NHX1 did not affect SNARE pairing but did affect the first round of fusion. To get a better sense of the latter, we employed a real-time lipid-dequenching assay (Reese et al., 2005; Reese and Mayer, 2005). Here, vacuoles were labeled with octadecyl rhodamine (R18), which is self-quenching at high concentrations. R18 was incorporated to the outer leaflets of vacuoles to serve as a reporter system. When labeled donor vacuoles were incubated with an excess (8-fold) of unlabeled acceptor vacuoles, rhodamine dequenching occurred immediately after the first round of fusion. Fusion was initiated by adding ATP after 5 minutes of incubation and reactions were monitored for rhodamine dequenching. WT vacuoles (white circles) showed characteristic rapid dequenching whereas lipid mixing was severely inhibited in nhx1Δ vacuoles (white squares), further suggesting that Nhx1p regulates early kinetics of membrane fusion (Fig. 8A). Because Vam7p rescued content mixing, we next examined whether Vam7p could restore lipid mixing. When nhx1Δ vacuoles were incubated with 400 nM Vam7p (gray squares), lipid mixing was not only rescued, but the rate of the reaction surpassed that of untreated WT reactions. Vam7p also enhanced the rate WT lipid mixing (gray circles). Interestingly, Vam7p did not significantly enhance the V_max of the reactions relative to untreated WT vacuoles, suggesting that Vam7p primarily activated the first round of fusion and not ongoing rounds. Because we observed disparate effects of Vam7p that were dependent on the state of priming, we examined lipid mixing using vacuoles treated with anti-Sec17p. The addition of anti-Sec17p did not affect the basal level of fluorescence of nhx1Δ vacuoles (Fig. 8B). When Vam7p was added, nhx1Δ vacuoles were rescued and underwent accelerated lipid mixing. The initial rate of rescued fusion was similar to untreated WT. When compared with content mixing assays in Fig. 5B, the Vam7p rescue data illustrated that exogenous Vam7p bypassed the first round of fusion only, leading us to propose that once fusion was activated, subsequent rounds of fusion no longer required exogenous Vam7p.

These experiments have led us to conclude that nhx1Δ vacuoles are primarily attenuated at the initiation stage of the fusion cascade. To begin examining early events in membrane fusion, we examined the release of Vam7p after priming. Because Vam7p lacks a transmembrane domain, it can be separated from the membrane after it is released from the cis-SNARE complex during priming (Boeddinghaus et al., 2002). After incubation, reactions were sedimented by centrifugation and supernatants discarded. The remaining membrane-bound Vam7p was detected by western blotting (Fig. 9A). As seen previously, WT vacuoles retain a constant level of Vam7p when reactions are uninhibited (Fig. 9B). By contrast, nhx1Δ vacuoles gradually accumulated Vam7p in the membrane fraction. Although this did not directly indicate that priming machinery is defective, these data do suggest that nhx1Δ vacuoles are altered in Vam7p binding. As a control for priming-dependent Vam7p release, we also performed parallel experiments in the presence of anti-Sec17p. Inhibiting priming prevents the release of Vam7p during incubations (Boeddinghaus et al., 2002; Fratti et al., 2004). In the presence of anti-Sec17p, both WT and nhx1Δ vacuoles maintained level concentrations of membrane-bound Vam7p throughout the experiment (Fig. 9C), suggesting that Vam7p accumulation on mutant vacuoles might be due to a priming defect. Any dysfunction in priming that occurred on nhx1Δ vacuoles was not mirrored when inhibition curves were performed using antibodies against Sec17p or Sec18p. WT and nhx1Δ vacuoles were incubated with increasing concentrations of anti-Sec17p (Fig. 9D) or anti-Sec18p (Fig. 9E). Although both antibodies potently blocked fusion, we did not observe a shift in sensitivity when using nhx1Δ vacuoles. This suggests that the priming machinery itself is not affected and that Vam7p retention on mutant vacuoles might be due to an indirect mechanism.

After priming, vacuoles undergo Ypt7p-dependent tethering that is sensitive to GAP activity. Because Nhx1p interacts with the GAP Gyp6p, which is reported to activate Ypt7p, we examined whether eliminating this interaction would influence the effectiveness of Gyp1-46p on Ypt7p (Wang et al., 2003). When WT and nhx1Δ vacuoles were treated with increasing concentrations of Gyp1-46, we did not observe a shift in sensitivity (Fig. 10). In keeping with these results, Ypt7p levels were not affected on mutant vacuoles (Fig. 1C). Together, these data suggest that tethering was not affected on nhx1Δ vacuoles.

Nhx1p regulates the trafficking of cargo from late endosomes to the vacuole and includes the transport of the protease CPY and Vph1p, a V₀ subunit of the vacuolar V-ATPase (Bowers et al., 2000). However, the exact role of Nhx1p in regulation of the core fusion machinery had not been addressed. In this study, we showed that Nhx1p indirectly regulates early kinetics of fusion through mechanisms dependent on ion transport and lumenal pH maintenance. However, the direct effect of dysregulated pH maintenance remains unclear. Defective fusion of nhx1Δ vacuoles was not irreversible, as the addition of exogenous Vam7p restored fusion to wild-type levels. Fusion was only rescued when priming was blocked with anti-Sec17p antibody. Vam7p was not able to restore fusion when priming was allowed. This was observed in both standard content mixing assays in the presence of a full complement of SNAREs, as well as trans-SNARE complex isolation experiments performed in the absence of Nyv1p on BJ3505 vacuoles. Although fusion was altered under these conditions, SNARE complex formation and SNARE-HOPS interactions were not affected on nhx1Δ membranes, suggesting that Nhx1p might also regulate fusion after trans-SNARE pairing. When examining the first round of fusion, as measured by lipid mixing, we found that Vam7p stimulated lipid mixing whether or not priming was blocked, indicating that Vam7p can only initiate the first round of fusion. Taken together, these data imply that Nhx1p might regulate fusion at multiple stages and that nhx1Δ vacuoles are defective in the initial activation of the fusion pathway, as well as post-docking events downstream of trans-SNARE complex formation.

We were able to determine by two criteria that Nhx1p regulates the initiation of the first round of fusion. Initially, we used content mixing assays that distinguish between the first round of fusion from late ongoing rounds. This approach illustrated that nhx1Δ vacuoles were only defective during the first round of fusion. The second measure came from a well-established lipid-mixing assay that relies on the dequenching of R18 during bilayer fusion. This experimental system utilizes an 8-fold excess of unlabeled vacuoles in the presence of R18-labeled membranes and, similar to the aforementioned content mixing assay, primarily shows the first round of fusion. Previously, this assay has been used as a measure of hemifusion that occurs between docking and full content mixing (Jun and Wickner, 2007; Reese et al., 2005; Reese and Mayer, 2005). However, we observed that nhx1Δ vacuoles exhibited a greater attenuation in lipid mixing relative to full content mixing. These results are inconsistent with the interpretation that lipid mixing equates with hemifusion. Currently, there is no evidence indicating that content mixing occurs without lipid mixing. Robust lipid mixing accompanies all fusion reactions, including transient reversible ‘kiss-and-run’ fusion (Giraudo et al., 2005). Thus, we posit that lipid-dequenching experiments report the first round of fusion and not hemifusion. This recasting reconciles our data in a manner consistent with dysfunctional early kinetics of membrane fusion.

The mechanisms by which Nhx1p directly regulates fusion remain unknown. One possibility is that fusion requires Nhx1p-dependent trafficking of Vph1p subunits to the vacuole. In the absence of Nhx1p, protein sorting to the vacuole is defective and cargo such as CPY and Vph1p are mis-sorted. Whereas CPY is secreted, Vph1p accumulates in the prevacuolar compartment, leading to dysfunctional pH maintenance, and potentially inhibits pH-independent function(s). Although Vph1p accumulates in the PVC of nhx1Δ cells (Bowers et al., 2000), approximately 50% reaches the vacuole (Fig. 1C). This is of particular interest because of the proposed role of the V₀ complex in homotypic fusion (Peters et al., 2001). Previous studies have suggested that the V₀ subunit of the V-ATPase is crucial for membrane fusion in yeast and other organisms (Bayer et al., 2003; Hiesinger et al., 2005; Peri and Nusslein-Volhard, 2008; Peters et al., 2001; Wassmer et al., 2005). The function of the V₀ complex in membrane fusion is independent of proton pumping and is thought to form a trans-V₀ complex that functions downstream of trans-SNARE complexes (Peters et al., 2001). Here, we found that nhx1Δ vacuoles, which are reduced in Vph1p, were able to fuse at WT levels when stimulated with Vam7p. This suggests that sufficient Vph1p reaches the vacuole for fusion to occur. However, this does not exclude the possibility that exogenous Vam7p (in the presence of a anti-Sec17 block) bypasses the need for full V₀ function.

Another possible regulatory target of Nhx1p is actin remodeling. Vacuole fusion requires actin remodeling at two different stages of fusion (Eitzen et al., 2002). Early fusion stages that precede docking are accompanied by the depolymerization of actin filaments. Following the docking stage, actin monomers are repolymerized prior to fusion. Although the role for actin remodeling remains unclear, inhibition of either mechanism abolishes vacuole fusion. The mammalian homolog of Nhx1p, Nhe1, is able to partially regulate actin remodeling through its interactions with Rho1, Cdc42 and ezrin (Denker et al., 2000; Frantz et al., 2007; Paradiso et al., 2004; Rasmussen et al., 2008; Wu et al., 2004). This raises the question of whether Nhx1p regulates actin remodeling during vacuole fusion. Thus far, interactions between Nhx1p and the actin network have not been explored.

Our study has revealed an unexpected role for the function of a Na⁺(K⁺)/H⁺ exchanger in vacuole fusion. This is the first report to show the direct regulation of membrane fusion by a member of the NHE family. These findings add another layer to the exquisite regulation of membrane fusion and open new avenues for future investigations.

Reagents were dissolved in PS buffer (20 mM PIPES-KOH, pH 6.8, 200 mM sorbitol). Antibodies against actin, Nyv1p, Pep4p, Pho8p, Sec17p, Sec18p, Vam3p, Vam7p, Vps11p, Vps16p, Vps33p, Vps39p and Vps41p were described previously (Haas and Wickner, 1996; Mayer and Wickner, 1997; Nichols et al., 1997; Price et al., 2000; Seals et al., 2000; Haas et al., 1995; Tedrick et al., 2004; Ungermann and Wickner, 1998). Anti-Vph1p monoclonal antibody was purchased from MitoSciences (Eugene, OR, USA). Antibodies against Vps23p and Snf7p were a gift from M. Babst (University of Utah, USA). Recombinant GST-Vam7p (wild-type and Q283R) and His6-Gyp1-46p were produced as described and dialyzed into PS buffer with 125 mM KCl (Fratti et al., 2007; Wang et al., 2003). Monoclonal anti-GST and chloroquine were from Sigma.

BJ3505 [MATα pep4::HIS3 prb1-Δ1.6R his3–200 lys2–801 trp1Δ101 (gal3) ura3–52 gal2 can1] and DKY6281 (MATα leu2–3 leu2–112 ura3–52 his3-Δ200 trp1-Δ901 lys2–801) were used for fusion assays (Haas et al., 1994) (Table 1). BJ3505 CBP-Vam3p nyv1Δ was used for trans-SNARE complex isolation (Collins and Wickner, 2007). NHX1 was deleted from BJ3505, DKY6281 and BJ3505 CBP-Vam3p nyv1Δ by homologous recombination. NHX1 was deleted with the URA3 cassette using PCR products amplified from pRS406 (Brachmann et al., 1998) with homology flanking the NHX1 coding sequence with forward primer 5′-GCACCCAAGATTGACAAATAGTTGTAGATTAAACATAGATAGATTGTACTGAGAGTGCAC-3′ and reverse primer 5′-GTATTCCTTCGGCGTTGAGTAAGAGAGAATGTATAAAGACCTGTGCGGTATTTCACACCG-3′. The PCR product was transformed into BJ3505, DKY6281 and BJ3505 CBP-Vam3p nyv1Δ by standard lithium acetate methods and plated on complete synthetic media lacking uracil to generate BJ3505 nhx1Δ::URA3 (RFY5), DKY6281 nhx1Δ::URA3 (RFY6), BJ3505 CBP-Vam3p nyv1Δ nhx1Δ::URA3 (RFY7) and DKY6281 nyv1Δ nhx1Δ::URA3 (RFY8). For vacuole localization studies, NHXI was fused inframe to GST or GFP by homologous recombination. DKY6281 was transformed with PCR products amplified from pFA6a-kanMX6-GST with forward primer 5′-GGCTACGCAATCACCTGCAGATTTCTCTTCCCAAAACCACCGGATCCCCGGGTTAATTAA-3′ and reverse primer 5′-ATATTTATATTAGAAACAAGGAAACCATACACTTTAAAGTGAATTCGAGCTCGTTTAAAC-3′ (Longtine et al., 1998) with homology flanking the stop codon of the NHX1 gene to make RFY9 (DKY6281 NHX1::GST). Similarly, DKY6281 was transformed with PCR products amplified from pFA6a-GFP(S65T)-kanMX6 with forward primer 5′-TCTCCATCCTTACAAGATTCGGCTACGCAATCACCTGCAGATTTCTCTTCCCAAAACCACCGGATCCCCGGGTTAATTAA-3′ and reverse primer 5′-TGTATAAAGACTTAATTAATATATTTATATTAGAAACAAGGAAACCATACACTTTAAAGTGAATTCGAGCTCGTTTAAAC-3′ with homology flanking the stop codon of the NHX1 gene to make RFY16 (DKY6281 NHX1::GFP). RFY5 and RFY6 were complemented with WT and mutant NHX1-HA as described (Bowers et al., 2000). WT NHX1-HA and NHX1-HA mutants (D201N and E355Q) were subcloned from the pRS316 CEN-URA3 plasmids (Bowers et al., 2000) (a gift from T. Stevens, University of Oregon, USA) into pRS413 (HIS3) and pRS414 (TRP1) using SalI and NotI. Transformants were selected using complete synthetic media lacking His or Trp.

Vacuoles were isolated by floatation as described (Haas et al., 1994). Standard in vitro fusion reactions (30 μl) contained 3 μg each of vacuoles from BJ3505 and DKY6281 backgrounds, fusion reaction buffer (20 mM PIPES-KOH pH 6.8, 200 mM sorbitol, 125 mM KCl, 5 mM MgCl₂), ATP regenerating system (1 mM ATP, 0.1 mg/ml creatine kinase, 29 mM creatine phosphate), 10 μM CoA and 283 nM IB₂. Reactions were incubated at 27°C and Pho8p activity was assayed in 250 mM Tris-Cl, pH 8.5, 0.4% Triton X-100, 10 mM MgCl₂, 1 mM p-nitrophenyl phosphate. Fusion units were measured by determining the p-nitrophenolate produced minute⁻¹ μg⁻¹pep4Δ vacuole. p-nitrophenolate absorbance was detected at 400 nm.

Microsomes were prepared as described (Sorin et al., 1997). WT and nhx1Δ cultures were grown in YPD, washed three times with PBS and suspended in lysis buffer (10 mM Tris-CL, pH 7.4, 300 mM sorbitol, 100 mM NaCl, 5 mM MgCl₂) with protease inhibitors (1 mM PMSF, 10 μM Pefabloc-SC, 5 μM pepstatin A and 1 μM leupeptin). Cells were disrupted with glass beads by cycles of vortexing and cooling on ice. Lysates were centrifuged to pellet large debris and supernatants were transferred to ultracentrifuge tubes. A total membrane fraction was collected by centrifugation (100,000 g, 1 hour, 4°C) and resuspended in buffer (20 mM HEPES-KOH, pH 7.4 with protease inhibitors).

SNARE complex isolation was performed as described previously using GST-Vam7p (Fratti et al., 2007; Fratti and Wickner, 2007). Briefly, large-scale 6× fusion reactions (180 μl) were incubated with 85 μg/ml anti-Sec17p IgG to block priming. After 15 minutes, 4.3 μg/ml affinity-purified anti-Vam3p IgG was added to selected reactions and incubated for an additional 5 minutes before adding 400 nM GST-Vam7p. After a total of 90 minutes, reactions were placed on ice for 5 minutes and 30 μl aliquots were removed to measure fusion activity. The remaining 150 μl reactions were sedimented (11,000 g, 10 minutes, 4°C) and the supernatants removed before extracting vacuoles with solubilization buffer (SB; 20 mM HEPES-KOH, pH 7.4, 100 mM NaCl, 2 mM EDTA, 20% glycerol, 0.5% Triton X-100, 1 mM DTT) with protease inhibitors (1 mM PMSF, 10 μM Pefabloc-SC, 5 μM pepstatin A and 1 μM leupeptin). Vacuole pellets were overlaid with 100 μl SB and resuspended gently. An additional 100 μl SB was added, gently mixed and incubated on ice for 20 minutes. Insoluble debris was sedimented (16,000 g, 10 minutes, 4°C) and 176 μl of supernatants were removed and placed in chilled tubes. Next, 16 μl was removed from each reaction as 10% total samples, mixed with 8 μl of 3× SDS loading buffer and heated (95°C, 5 minutes). The remaining extracts were incubated with 30 μl equilibrated glutathione sepharose 4B beads (GE Healthcare; 15 hours, 4°C, nutating). Beads were sedimented and washed 5× with 1 ml SB (735 g, 2 minutes, 4°C) and bound material was eluted with 40 μl 1× SDS loading buffer. Protein complexes were examined by western blotting. Secondary antibodies conjugated to alkaline phosphatase (Pierce) were used with ECF reagent (GE Healthcare). Images were acquired using a Storm Phosphorimager 840 (Molecular Dynamics).

Analysis of trans-SNARE complex formation was conducted as described with some modifications (Collins and Wickner, 2007; Jun and Wickner, 2007). Complex formation was compared between reactions containing vacuoles from RFY6 and RFY7 relative to those with BJ3505-CBP-Vam3p nyv1Δ and DKY6281 vacuoles. The trans-SNARE assays were performed using 16× large-scale reactions (480 μl) containing 48 μg of vacuoles each from BJ3505 CBP-Vam3p nyv1Δ and DKY6281 backgrounds and incubated at 27°C for 60 minutes. After incubating, reactions were placed on ice for 5 minutes and 30 μl was withdrawn from each sample to assay Pho8p activity. The remaining 450 μl samples were centrifuged (13,000 g, 15 minutes, 4°C) and the supernatants decanted. Vacuole pellets were overlaid with 200 μl ice-cold SB (20 mM Tris-Cl pH 7.5, 150 mM NaCl, 1 mM MgCl₂ 0.5% Nonidet P-40 alternative, 10% glycerol) with protease inhibitors (0.46 μg/ml leupeptin, 3.5 μg/ml pepstatin, 2.4 μg/ml pefabloc-SC, 1 mM PMSF) and gently resuspended. SB was added to a final volume of 600 μl and extracts were mixed (20 minutes, 4°C, nutating). Detergent-insoluble debris was removed by centrifugation (16,000 g, 20 minutes, 4°C). Supernatants were transferred to fresh tubes and 10% of the extract was removed for input samples. The remaining extracts were brought to 2 mM CaCl₂. CBP-Vam3p complexes were incubated with 50 μl equilibrated calmodulin sepharose 4B (GE Healthcare; 4°C, 12 hours, nutating). Beads were collected by centrifugation (4000 g, 2 minutes, 4°C) and suspended four times with the SB, followed by bead sedimentation. Bound proteins were eluted with SDS sample buffer containing 5 mM EGTA and heated at 95°C for 5 minutes. The samples were used for SDS-PAGE analysis and immunoblotting.

Lipid mixing assays were conducted using octadecyl rhodamine B (R18; Invitrogen) as described with minor modifications (Jun and Wickner, 2007). Briefly, BJ3505 background vacuoles (300 μg) were incubated in 400 μl of PS buffer containing 150 μM R18 (10 minutes, 4°C, nutating). Samples were mixed with 15% Ficoll in PS buffer (wt/vol) and transferred to a polyallomer tube (11 × 60 mm). Samples were overlaid with 1.0 ml each of 8%, 4% and 0% Ficoll. Labeled vacuoles were re-isolated by centrifugation using a SW60 Ti rotor (105,200 g, 30 minutes, 4°C) and recovered from the 0–4% Ficoll interface. Lipid mixing assays (90 μl) contained 2 μg of R18-labeled vacuoles and 16 μg of unlabeled vacuoles in fusion buffer. Reaction mixtures were transferred to a black, half-volume 96-well flat-bottom microtiter plate with nonbinding surface (Corning). Rhodamine fluorescence (λ_ex=544 nm; λ_em=590 nm) was measured using a POLARstar Omega plate reader (BMG Labtech) at 27°C. Measurements were taken every minute for 75 minutes, yielding fluorescence values at the onset (F₀) and during the reaction (F_t). The final ten measurements of a sample after adding 0.33% (vol/vol) Triton X-100 were averaged and used as a value for the fluorescence after infinite dilution (F_TX100). The relative total fluorescence change ΔF_t/F_TX100=(F_t–F₀)/F_TX100 was calculated.

Vacuole morphology was monitored by incubating yeast cells with YPD (yeast-peptone-dextrose) broth containing 5 μM FM4-64. Cultures were grown overnight, washed with PS buffer, concentrated by centrifugation, mixed with 0.6% agarose and mounted on glass slides for observation. Images were acquired using a Zeiss Axio Observer Z1 inverted microscope equipped with an X-Cite 120XL light source, Plan Apochromat 63× oil objective (NA 1.4) and an AxioCam CCD camera. FM4-64 images were acquired using a 43 HE CY 3 shift-free filter set. Image deconvolution of z-stack images was performed using the AxioVison 3D deconvolution software module and fast iterative algorithm.

We thank Tom Stevens for NHX1 plasmids, and William Wickner and Markus Babst for generous gifts of antisera. We also thank Kevin Collins, Vincent Starai, Robert Switzer, Emad Tajkhorshid and members of the Fratti laboratory for helpful discussions. This work was supported by startup funds provided to R.A.F. by the University of Illinois at Urbana-Champaign.