The Sec1/Munc18 Protein Vps45 Regulates Cellular Levels of Its SNARE Binding Partners Tlg2 and Snc2 in Saccharomyces cerevisiae

Intracellular membrane trafficking pathways must be tightly regulated to ensure proper functioning of all eukaryotic cells. Central to membrane trafficking is the formation of specific SNARE (soluble N-ethylmeleimide-sensitive factor attachment protein receptor) complexes between proteins on opposing lipid bilayers. The Sec1/Munc18 (SM) family of proteins play an essential role in SNARE-mediated membrane fusion, and like the SNAREs are conserved through evolution from yeast to humans. The SM protein Vps45 is required for the formation of yeast endosomal SNARE complexes and is thus essential for traffic through the endosomal system. Here we report that, in addition to its role in regulating SNARE complex assembly, Vps45 regulates cellular levels of its SNARE binding partners: the syntaxin Tlg2 and the v-SNARE Snc2: Cells lacking Vps45 have reduced cellular levels of Tlg2 and Snc2; and elevation of Vps45 levels results in concomitant increases in the levels of both Tlg2 and Snc2. As well as regulating traffic through the endosomal system, the Snc v-SNAREs are also required for exocytosis. Unlike most vps mutants, cells lacking Vps45 display multiple growth phenotypes. Here we report that these can be reversed by selectively restoring Snc2 levels in vps45 mutant cells. Our data indicate that as well as functioning as part of the machinery that controls SNARE complex assembly, Vps45 also plays a key role in determining the levels of its cognate SNARE proteins; another key factor in regulation of membrane traffic.


Introduction
Spatial and temporal regulation of membrane traffic is essential for maintenance of one of the defining features of eukaryotic cells; intracellular compartmentalisation into discrete membrane bound organelles [1]. The SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) proteins are central to membrane traffic in all eukaryotes, with formation of SNARE complexes between SNAREs on opposing membranes being defined as the minimal requirement for membrane fusion [2]. SNARE complexes consist of 4 helical SNARE motifs contributed by members of the Qa (syntaxin), Qb, Qc (also called t-SNAREs) and R-(also called v-SNAREs) subfamilies [3].
Regulating SNARE complex assembly allows the cell to regulate membrane traffic. One family of proteinsthat regulate this process are the Sec1/Munc18 (SM) proteins [4]. Although their mechanism of action is not completely understood, it is clear that SM proteins are universally required for functional SNARE complex assembly [4]. We have previously characterised a role for the SM protein Vps45 in regulating SNARE-mediated membrane traffic through the endosomal system of the yeast Saccharomyces cerevisiae [5,6,7]. Vps45 activates the syntaxin Tlg2 for entry into functional SNARE complexes with its partner t-SNAREs Tlg1, Vti1 and the v-SNARE Snc2 likely by facilitating a switch of Tlg2 from a closed (inactive) conformation to an open (active) state [5]. Vps45 binds directly to Tlg2, as well as to the v-SNARE, Snc2 and the assembled SNARE complex [6,7,8]. Of these interactions, Vps45's binding to Tlg2 is the most thoroughly characterised. Like other SM proteins Vps45 has two distinct binding sites for its cognate syntaxin [8]. The higheraffinity binding site involves a conserved leucine residue in Vps45 that is predicted to form a hydrophobic pocket on the outer surface of the arch shaped SM protein [7]. This facilitates 'mode 2' binding to Tlg2 which requires the N-terminal 36 residues of the syntaxin, and is akin to the interaction captured by the Sly1/Sed5 crystal structure [7,9]. In addition, Vps45 binds Tlg2 through a second mechanism that does not involve the N-terminal peptide of the syntaxin, and while the structure of this interaction has not been solved, biochemical and biophysical studies indicate that it is akin to the interaction captured in the crystal structure of the neuronal Munc18a bound to Syntaxin1a; mode 1 binding [8,10]. The role that these two distinct binding modes between Vps45 and Tlg2 and indeed, between other SM proteins and their cognate syntaxins is not understood.
Vps45 was originally identified through genetic screens for yeast defective in vacuolar protein sorting (vps mutants) [11,12], and has been shown to be required for delivery of proteins from the trans Golgi network (TGN) into the endosomal system [13,14]. Classification of the vps mutants into 6 groupings (A-F) based on vacuolar morphology suggested the existence of 6 separate processes whose individual perturbation result in missorting of the vacuolar hydrolase carboxy peptidase Y (CPY) [15]. This suggested that there are multiple trafficking steps in the Vps pathway from the TGN to the vacuole, and that gene products affected in mutants of each of the classes act in the same process or membrane trafficking step. Further studies have revealed that this in indeed the case, and in many cases proteins encoded by genes in the same class function as part of a multi-protein complex. This is perhaps best exemplified by the class E vps mutants [15]. Class E vps mutants accumulate an exaggerated form of the endosomal compartment through which biosynthetic and endocytosed material transits en route to the vacuole [15]. Biochemical, morphological and genetic studies have demonstrated that the gene products represented by Class E vps mutants all function in sorting ubiquitinated proteins into multivesicular bodies as part of the ESCRT (endosomal sorting complex required for transport machinery [16,17]. Similarly, the 4 class C proteins (Vps11, Vps16, Vps18 and Vps33) form the core of the CORVET and HOPS tethering complexes that function at endosomal and vacuolar membranes respectively [18]. VPS45 is a class D gene [15]. Class D Vps proteins, including Vps45 and the regulatory and catalytic subunits of PI-3 kinase (Vps15 and Vps34), are required for traffic between the TGN and endosomes [19].
Interaction between Vps45 and Tlg2 is not only important for regulating the function of Tlg2, but also stabilises the syntaxin, with cellular levels of Tlg2 being dramatically reduced in cells lacking Vps45 [5,20]. This feature appears to be conserved across species, and also across different membrane trafficking steps, as several SM proteins have been shown to stabilise their syntaxins [21,22,23,24]. Controlling cellular levels of SNARE proteins represents another mechanism by which membrane fusion may be regulated. Like many SM proteins, Vps45 also binds directly to its cognate v-SNARE [7], and we therefore set out to investigate whether v-SNARE levels are also regulated by binding to the SM protein. We now report that, like those of Tlg2, levels of the Snc v-SNARE are regulated by Vps45.
As well as assembling into Tlg2p-containing SNARE complexes to regulate endosomal traffic, the Snc2 v-SNAREs also participate in exocytic Sso-containing SNARE complexes required for fusion of secretory vesicles with the plasma membrane; a trafficking pathway required for growth [25,26]. These data explain why vps45D cells are unique amongst the Class D vps mutants in that they display multiple growth phenotypes and reveal the importance of regulating cellular levels of SNARE proteins.

Vps45p Regulates Levels of its Cognate SNARE Binding Partners
Cells lacking the SM protein Vps45 have reduced levels of its cognate syntaxin Tlg2 [5,7,20,27]. In addition to binding Tlg2, Vps45 also binds the v-SNARE Snc2 [7]. To address whether absence of Vps45 affects cellular levels of Snc2, we used immunoblot analysis to assess the levels of an HA-tagged version of Snc2 in wild-type and vps45D mutant cells. Figure 1A shows that cells lacking Vps45 (vps45D) contain substantially less HA-Snc2 than wild-type cells harbouring the same HA-Snc2 expression construct (pCOG054). Analysis of two different strain backgrounds demonstrates that cellular levels of HA-Snc2 produced from the 2m-plasmid in vps45D mutant cells are approximately 33% of those found in their congenic wild-type strains (Fig. 1A). To extend these studies and look at levels of the endogenous v-SNARE, we used polyclonal antiserum raised against residues 72-86 (GFKRGANRVRKQMWW) of Snc2 [28]. This region of the protein shares 87% identity with the analogous region of Snc1 (residues 73-87; GFKRGANRVRKAMWY), whose functions are redundant with Snc2 for cellular processes including exo-and endocytosis [26,29]. Immunoblot analysis using this antiserum demonstrates that cellular levels of endogenous Snc v-SNAREs in cells lacking Vps45 (vps45D) are approximately 65% of those in congenic wild-type cells (Fig. 1B).
The data presented in Figure 1 indicate that cellular levels of the v-SNARE Snc proteins and the syntaxin Tlg2 are regulated by their cognate SM protein Vps45p. To test this hypothesis we set out to increase the cellular levels of Vps45 and ask whether this leads to Vps45 Regulates Tlg2 and Snc Levels PLOS ONE | www.plosone.org a concomitant increase in Tlg2 and Snc levels. This was achieved by expressing VPS45 from a multicopy plasmid in wild-type cells ( Fig. 2A). Cells harbouring pCOG070, encoding HA-tagged Vps45 [7], contain approximately 2.5-fold more Vps45 than those carrying the parental vector YEplac195 [30], and also contain increased levels of Tlg2 and the Snc v-SNAREs (both $2-fold).

Selective Stabilisation of Snc by a Mutant Version of Vps45 Abrogated for Tlg2-Binding
We have previously reported that a version of Vps45 (Vps45 L117R ) harbouring a mutation that abrogates binding of the SM protein to the N-terminal peptide of Tlg2 [7,8] does not stabilise Tlg2, demonstrating that this mode of binding between Vps45 and Tlg2 stabilises the syntaxin [27]. The L117R mutation disrupts the hydrophobic pocket-binding site situated on the outer surface of domain-1 of the arch-shaped SM protein [7,8]. The binding site for Snc2 is distinct from this hydrophobic pocket, demonstrated by the observation that the Vps45 L117R mutant binds the cytosolic domain of Snc2 in vitro in a manner indistinguishable from wild-type Vps45 [7]. Figure 2B shows that, as previously reported [27], vps45D cells expressing Vps45 L117R have reduced levels of Tlg2 compared to the same cells expressing wild-type Vps45. In contrast, Snc levels are comparable in vps45D cells expressing either Vps45 L117R or wild-type Vps45 (both substantially higher than in the same cells carrying empty vector; Fig. 2B). These data indicate that binding to Vps45 is important for the regulation of cellular levels of its cognate SNARE binding partners. This is an important observation as it provides the cell with a mechanism to ensure that the SNAREs are not present in excess compared to their regulatory SM protein. Were such a situation to arise it could prove catastrophic for the cell allowing SNARE complexes to form in an unregulated manner.

Growth Phenotypes of vps45 Mutant Cells Correlate with Reduced Cellular Levels of Snc v-SNAREs
VPS45 is a class D VPS gene [15]. Class D VPS gene products, including Vps45, and the regulatory and catalytic subunits of PI-3 kinase (Vps15 and Vps34) are required for traffic between the TGN and endosomes [19]. The original VPS genes, identified through genetic screens for mutants that missort CPY [11,12] are not essential due to the existence of multiple pathways for delivery of macromolecules to the vacuole [31,32]. Consistent with this, most vps mutants do not show growth defects under normal laboratory conditions. One exception to this are vps45 mutants [13,14]. Figure 3A shows that cells lacking Vps45p have significantly longer doubling time, and slower growth rates, than congenic wild-type cells, whereas other class D vps mutants do not. This difference in growth phenotype between vps45 mutant cells and other class D vps mutants is also illustrated by the observation that vps45 mutants display a marked growth defect at 39uC whereas other class D mutants tested do not (Fig. 3B). These observations are surprising, as removal of different components involved in the same transport step would be expected to result in similar phenotypes.
Although not identified in the original vps genetic screens, tlg2 mutants are vps mutants in that they missort the vacuolar hydrolase CPY [33]. Like other VPS genes, TLG2 is not essential. Similarly, Vam3, the syntaxin that regulates traffic at the vacuolar membrane is also not essential [34,35]. In contrast SNAREs required for membrane traffic to the cell surface are essential [36]. As well as participating in endosomal Tlg2-containing SNARE complexes [33], Snc2 is also required for the fusion of secretory vesicles with the plasma membrane for exocytosis [26]. Exocytosis is required for growth, and as such snc mutants are inviable [26]. We hypothesized that the growth phenotype of vps45D cells might be due to their reduced levels of Snc v-SNAREs and took advantage of our finding that expression of the Vps45 L117R mutant in vps45D cells selectively restores the levels of Snc2 (Fig. 2) to test this. Figures 4A and 4B show that expression of Vps45 L117R We have also previously characterised a second mutant of Vps45p, Vps45 W244R , which gives a dominant negative phenotype for sorting of vacuolar hydrolases [7,27]. The CPY missorting phenotype of this dominant negative mutant is dose-dependent and can be abrogated by co-expression of either wild-type Vps45 or Snc2, but not Tlg2 [7]. This suggests that Vps45 W244R disrupts endosomal trafficking by titrating out the v-SNARE Snc2. Expression of Vps45 W244R in vps45D cells does not complement the CPY missorting phenotype observed in vps45D cells [7,27]. Figure 4B shows that expression of the mutant Vps45 W244R in vps45D cells partially complements the temperature-sensitive growth phenotype of vps45D cells, restoring growth at 39uC. Moreover, Figure 4C demonstrates that coexpression of HA-Snc2 from a multicopy plasmid with Vps45 W244R fully restores growth to levels seen with Vps45p and Vps45p L117R mutants. These data indicate that the growth defects of vps45D cells are independent of the CPY missorting phenotype as they can be overcome by increasing Snc2 levels in cells containing a nonfunctional version of Vps45p.
It is important to remember that SM protein function in SNAREmediated membrane traffic extends beyond a role in maintaining cellular levels of its binding partners. This is demonstrated by our previous work showing that restoration of Tlg2 levels in vps45D cells is not sufficient to complement the vacuolar trafficking defects of these cells, and has also been reported for other SM/Syntaxin pairs including syntaxin-1/Munc18-1 [5]. The multiple functions of SM proteins complicates the challenge of understanding the role(s) of this protein family in vivo [2][3][4]. The observation that increasing levels of Snc2 complements growth phenotypes of vps45D cells likely reflects the role of the exocytic SM protein, Sec1p, in regulating Snccontaining SNARE complexes at the plasma membrane, as oppose to Vps45 in the endosomal system v-SNAREs [4]. Further studies will be required to ascertain whether this is the case as alternative/ compensatory mechanisms or pathways may complicate genetic analyses in vivo. The data presented here indicate that the growth phenotypes of cells lacking Vps45 are not related to the SM proteins  function in SNARE complex assembly, but are a secondary consequence of the reduced levels of Snc2 in these cells. Work is currently underway to identify mutant versions of Vps45 that stabilise Tlg2, but not Snc2, and also versions of Snc2 that are not subject to downregulation in the absence of Vps45. Such tools will allow us to address many important questions that arise from the present study, such as what factors determine whether a particular Snc participates in an exocytic or an endosomal SNARE complex.
Here, we have demonstrated that Vps45 plays a role in regulating the cellular levels of, not only its syntaxin (Tlg2) but also its v-SNARE (Snc) binding partners. Furthermore, we have shown that this reduction in cellular Snc levels correlates with the impaired growth phenotypes seen for vps45 mutant cells. Prior to this study, this had been a puzzling observation since these growth phenotypes are not displayed by cells lacking Tlg2. Growth phenotypes are associated with loss of Vps45 orthologues in other organisms: in addition to endocytic defects, vps-45 mutation in Caenorhabditis elegans results in temperature sensitive lethality [37], and RNAi knockdown of AtVPS45 in Arabidopsis thaliana results in stunted growth [23]. Further work is required to investigate whether these are due to reduced levels of exocytic v-SNAREs. In addition, this study describes how cells may regulate levels of SNARE proteins as a mechanism to control membrane trafficking pathways.

Immunoblot Analysis of Yeast Cell Lysates
Yeast cells were grown in standard minimal media (SD, lacking amino acids where appropriate for plasmid selection) [40]. 10 OD 600 equivalents were harvested from cultures in early log phase and vortexed in 200 ml of Laemmli sample buffer in the presence of glass beads (425-600 mm, acid washed). Glass beads were allowed to settle during a 10 min incubation at 65uC and 10 ml of the lysate was loaded per lane of a 10% SDS-polyacrylamide gel. Proteins were next transferred to nitrocellulose for immunoblot analysis with antibodies as indicated. Immunoreactive proteins were visualized by detection with an anti-rabbit/mouse IgG horse radish peroxidase linked Table 1. Yeast strains used in this study.

Strain
Genotype Reference

Growth Rate Determination
Yeast were grown for 16 h at 30uC in a rotary shaker (250 rpm), then diluted into fresh medium at OD 600 = 0.2. OD 600 was measured every hour until the stationary phase of the growth curve was reached. The doubling time (DT) during logarithmic growth was determined as described by Willett [43]: DT = ln 26t/ (ln b -ln a) (t is the time period in hours; a is the optical density at the beginning of time period; b is the optical density at the end of time period). The specific growth rate (m) was calculated as described by Harvey [44]: m = (ln b -ln a)62.303/t.