Mso1 Is a Novel Component of the Yeast Exocytic SNARE Complex*

The yeast exocytic SNARE complex consists of one molecule each of the Sso1/2 target SNAREs, Snc1/2 vesicular SNAREs, and the Sec9 target SNARE, which form a fusion complex that is conserved in evolution. Another protein, Sec1, binds to the SNARE complex to facilitate assembly. We show that Mso1, a Sec1-interacting protein, also binds to the SNARE complex and plays a role in mediating Sec1 functions. Like Sec1, Mso1 bound to SNAREs in cells containing SNARE complexes (i.e. wild-type, sec1-1, and sec18-1 cells), but not in cells in which complex formation is inhibited (i.e. sec4-8 cells). Nevertheless, Mso1 remained associated with Sec1 even in sec4-8 cells, indicating that they act as a pair. Mso1 localized primarily to the plasma membrane of the bud when SNARE complex formation was not impaired but was mostly in the cytoplasm when assembly was prevented. Genetic studies suggest that Mso1 enhances Sec1 function while attenuating Sec4 GTPase function. This dual action may impart temporal regulation between Sec4 turnoff and Sec1-mediated SNARE assembly. Notably, a small region at the C terminus of Mso1 is conserved in the mammalian Munc13/Mint proteins and is necessary for proper membrane localization. Overexpression of Mso1 lacking this domain (Mso1-(1–193)) inhibited the growth of cells bearing an attenuated Sec4 GTPase. These results suggest that Mso1 is a component of the exocytic SNARE complex and a possible ortholog of the Munc13/Mint proteins.

The membrane fusion apparatus in eukaryotes relies upon SNAREs 2 to mediate intracellular membrane docking and fusion events. SNAREs are membrane-associated proteins bearing ␣-helical domains (termed SNARE motifs) that can assemble into intermolecular four-helix bundles (1)(2)(3). Assembly of the SNARE complex, which is composed of vesicular (v) and target (t) SNAREs (or R-and Q-SNAREs, depending upon the nomenclature followed), bridges apposed bilayers and allows for the extrusion of interposed water molecules, resulting in membrane fusion. Although shown to be sufficient to confer membrane fusion in vitro or under artificial conditions (4 -6), a number of other molecules (termed SNARE regulators) act upstream of the SNAREs and play essential roles in the temporal and spatial control of SNARE assembly in vivo (7)(8)(9). At the level of exocytosis, these include the Rab family of small GTPases; the vesicle-tethering exocyst complex; the Sec1/ Munc18 family of SNARE assembly factors, which are conserved from yeast to mammals; and the complexin, Mint, Munc13, and synaptotagmin families of SNARE regulators, which appear to act only upon stimulus-coupled exocytic processes (7)(8)(9)(10).
In yeast, the exocytic SNARE complex consists of one representative molecule each of the Sso1/2 t-SNAREs (Q-SNARE) (11) and Snc1/2 v-SNAREs (R-SNARE) (12) as well as one molecule of the Sec9 t-SNARE (Q-SNARE) (13). However, all Q-SNARE complexes composed either of a mutated Snc v-SNARE and one molecule each of Sso and Sec9 (14,15) or of two molecules of Sso and one of Sec9 (16) have also been shown to be functional and to confer exocytosis. In addition, Sec1, a conserved regulator of the Sso/syntaxin t-SNAREs, is known to bind to the assembled SNARE complex and may play a definitive role in facilitating assembly (17,18). Interestingly, Sec1 in yeast is known to interact with Mso1, a protein of unknown function that is not essential for membrane fusion, but in its absence, has defects in secretion and accumulates secretory vesicles (19). Although the functions of neither Munc13 (or the related Mint proteins) in mammals nor Mso1 in yeast have been fully resolved, these proteins are known to interact with Sec1/ Munc18 (20 -23).
To determine whether Mso1 is functionally related to Munc13/Mint, we examined the genetic and physical interactions of Mso1 with components of the yeast exocytic SNARE complex. We found that Mso1 bound to the SNARE complex in cells that can assemble or accumulate such complexes (i.e. wild-type, sec1-1, sec9-7, and sec18-1 cells), but not in cells in which complex formation is inhibited (i.e. sec4-8 cells). In contrast, Mso1 binding to Sec1 was not affected under these conditions, indicating that it is a partner of Sec1 whether or not Sec1 is bound to the SNARE complex. Thus, Sec1 and Mso1 constitute a functional pair that confers exocytosis in yeast. Mso1 localizes to the plasma membrane in a manner sensitive to mutations that affect SNARE complex assembly and requires a domain from the C terminus that is conserved in the Munc13 and Mint proteins from higher organisms (22,24). These and other results suggest that Mso1 is functionally homologous to the Munc13/Mint proteins. In addition, we propose that it helps bridge the connection between Sec4 function and SNARE assembly via the recruitment and release of Sec1 from the SNARE complex.

MATERIALS AND METHODS
Media, DNA, and Genetic Manipulations-Yeast cells were grown in standard growth medium containing either 2% glucose or 3.5% galactose. The preparation of synthetic complete and dropout media was similar to that described previously (25). Standard methods were used for the introduction of DNA into yeast and the preparation of genomic DNA (25).
Growth Tests-Yeast cells were grown on synthetic and rich growth media (25). For growth tests on plates, yeast cells were grown to mid-log phase, normalized for absorbance at 600 nm, diluted serially, and plated by drops onto solid medium preincubated at different temperatures.
Yeast Strains-The yeast strains used are listed in TABLE ONE.

Plasmids-Yeast
expression vectors included pUG36 (MET25::yEGFP, CEN, URA3) and pAD54 and pAD6 (2, LEU2, ADH1 promoter), which contain sequences encoding the hemagglutinin (HA) or Myc epitope, respectively, downstream of the ADH1 promoter. A plasmid expressing HA-tagged Sso1 was described previously (16). A plasmid expressing HA-tagged Mso1 was created by amplifying MSO1 by PCR from genomic DNA using oligonucleotides bearing SalI and SmaI sites, respectively, and cloning into the same sites of pAD54 to yield pADH-HAMSO1. This construct lacks the initial two methionines of Mso1 and utilizes the initiator methionine preceding the HA epitope. A plasmid expressing HA-tagged Sec1 was created in a similar fashion to yield pADH-HASEC1. A single-copy plasmid expressing green fluorescent protein (GFP)-tagged Mso1 under the control of the inducible MET25 promoter was created by amplifying MSO1 plus 536 bp of its 3Ј-untranslated region by PCR using oligonucleotides bearing SpeI and ClaI sites, respectively, and cloning into the same sites in pUG36 (26) to yield pMET-GFPMSO1. C-terminal truncation mutants of Mso1 (Mso1- ) were created in pADH-HAMSO1 and pMET-GFPMSO1 by mutagenesis with Pfu polymerase using the appropriate oligonucleotides to yield pADH-HAMSO1-  and pMET-GFPMSO1- . A plasmid containing monomeric red fluorescent protein (mRFP) in pGEM (Clontech), pmRFP-SalI, was created by PCR using oligonucleotides bearing SalI sites and an mRFP template obtained from R. Tsien (University of California, San Diego, CA). Next, a SalI-SalI mRFP fragment (lacking the start and stop codons) was cloned in-frame into the SalI site of pADH-HAcSNC1, yielding pADH-HAmRFPcSNC1. pADH-HAcSNC1 was created by PCR using pADH-cSNC1 (27) as a template and oligonucleotides bearing SalI and SacI sites and cloning the PCR product into the SalI and SacI sites of pAD54. All constructs were verified by DNA sequencing. The oligonucleotides used are listed in TABLE TWO. Microscopy-GFP fluorescence in strains expressing the GFP-tagged fusion proteins was visualized by confocal fluorescence microscopy. Yeast cells were grown to log phase on selective synthetic medium containing methionine and then transferred to medium lacking methionine for 2 h prior to visualization to induce expression from the MET25 promoter. For temperature-sensitive strains, cells were either maintained at 26°C or shifted to 37°C for an additional 45 min.

Mso1 and the Exocytic SNARE Complex
inhibit SNARE complex dissociation and degradation (16

MSO1
Overexpression Rescues sec1-1 Cells but Inhibits the Growth of Late sec Mutants-Yeast MSO1 was characterized earlier as a multicopy suppressor of defects in the yeast SEC1 gene (i.e. sec1-1 cells) (19). Overexpression of MSO1 rescues cells bearing mutations in SEC1 at restrictive temperatures, whereas its deletion in the sec1-1, sec2-41, and sec4-8 mutant backgrounds leads to synthetic lethality. In other late-acting sec mutants, the deletion of MSO1 leads to reduction in cell growth (19). This suggests that Mso1 acts to confer exocytosis and, in particular, may enhance the functioning of Sec1. Indeed, Mso1 was found to bind to Sec1 using the yeast two-hybrid assay, suggesting a physical interaction that may be meaningful in regulating SNARE assembly (19).
Because the function of Mso1 is still unknown, we further examined its ability to interact with components of the SNARE complex. First, we performed multicopy suppression assays with other late-acting sec mutants (Fig. 1). We found that overexpression of HA-MSO1 rescued sec1-1 cells at restrictive temperatures ( Fig. 1A), as previously demonstrated (19). In contrast, we found that HA-MSO1 overexpression inhibited growth of the sec4-8, sec8-9, and sec15-2 mutants, which act upstream of SNARE assembly, at both permissive and semirestrictive temperatures (Fig. 1B). SEC4 encodes the Rab GTPase necessary to activate vesicle tethering and docking, and both SEC8 and SEC15 encode components of the exocyst that acts as the downstream effector for Sec4 (19, 29 -33). This interesting dichotomy suggests that Mso1 may have two functions, one to enhance the functioning of Sec1 and a second perhaps to attenuate Sec4. MSO1 overexpression also inhibited the growth of yeast with mutations in other proteins that function in the late secretory pathway (Fig. 1B), such as sec18-1 cells, which are defective in SNARE complex disassembly (34); sso2-1 cells, which bear a temperature-sensitive Sso2 t-SNARE (11); and snc⌬ cells, which lack the Snc1/2 exocytic and endocytic v-SNAREs (12,35). In the case of snc⌬ cells, growth inhibition upon MSO1 overexpression was observed only on glucose-containing medium, which prevents expression of a plasmid-borne copy of SNC1 that is under the control of a galactoseinducible promoter. When SNC1 was induced in these cells, however, no deleterious effect on growth was seen (Fig. 1B). We note that the inhibitory effects of MSO1 overexpression were seen in these late sec mutants primarily at permissive or subpermissive temperatures, although sec9-7 cells, which may be blocked in the trans-SNARE complex (36), were found to be inhibited at elevated temperatures. In contrast to the effects on the late sec mutants, MSO1 overexpression had little to no effect on wild-type cells or early mutants of the secretory pathway (i.e. sec23-2 and sec31-1 cells) under the same conditions ( Fig. 1, A and C). Thus, MSO1 overexpression preferentially inhibits the growth of late-acting mutants of the secretory pathway.
Similar to the effects seen with MSO1 overexpression in the different sec mutants, we found that overexpression of Myc-SEC1 also had deleterious effects on the growth of sec4-8, sec8-9, and sec15-2 cells, but not the other mutants tested (Fig. 1B). In contrast, we note that myc-SEC1 overexpression mildly rescued the growth of sso2-1 cells at elevated temperatures ( Fig. 1B), implying a positive role in restoring SNARE FIGURE 1. MSO1 and SEC1 overexpression inhibits mutants in Sec4 and Sec4 effectors. A, HA-MSO1 overexpression restores the growth of sec1-1 cells. Wild-type (WT) or sec1-1 cells bearing either a control plasmid (vector) or a plasmid overexpressing HA-MSO1 were grown to mid-log phase on selective synthetic medium at 26°C, diluted serially, plated onto prewarmed solid medium, and incubated at various temperatures for 48 h. B, HA-MSO1 and myc-SEC1 overexpression inhibits the growth of sec4-8, sec8-9, and sec15-2 cells at elevated temperatures. Yeast cells (e.g. sec4-8, sec8-9, sec15-2, sec18-1, sso2-1, and sec9-7 cells) bearing a control plasmid (vector) or a plasmid expressing HA-MSO1 or myc-SEC1 were grown and plated as described for A. snc⌬ null cells were grown on galactose-containing medium to induce SNC1 gene expression and then transformed with a control plasmid or a plasmid expressing HA-MSO1 or myc-SEC1. Cells were then grown to mid-log phase on galactose-containing medium prior to serial dilution and plating onto galactose or glucose-containing medium. Use of the latter turns off SNC1 gene expression and results in the snc⌬ phenotype. C, HA-MSO1 overexpression does not affect the growth of early sec mutants. sec23-2 and sec31-1 cells bearing either a control plasmid (vector) or a plasmid overexpressing HA-MSO1 were grown to mid-log phase on selective synthetic medium at 26°C, diluted serially, plated onto prewarmed solid medium, and incubated at various temperatures for 48 h. OCTOBER 7, 2005 • VOLUME 280 • NUMBER 40 assembly and exocytosis therein. The parallel effect of MSO1 and SEC1 overexpression on Sec4 and the Sec4 effector pathway suggests a common mode of action for Mso1 and Sec1 in exocytosis. As Sec1 is likely to act downstream of Sec4 (34), deficiencies in Sec4 GTPase function appear to hinder Sec1.

Mso1 and the Exocytic SNARE Complex
Mso1 Is a Component of the Exocytic SNARE Complex-Recent studies have suggested that proteins of the Sec1/Munc18 family facilitate SNARE assembly at the level of the Golgi (37,38), endosomes (39), and plasma membrane (17,18) in yeast and are components of the assembled SNARE complex. This suggested to us that Mso1 might be a component of the complex, particularly if it acts to stimulate Sec1 function, as might be predicted from the genetic studies (Fig. 1A) (19). To test this idea, we expressed HA-tagged Mso1 in wild-type cells and various secretion mutants (i.e. sec1-1, sec4-8, sec9-7, and sec18-1 cells) and examined its ability to precipitate Sec1 and components of the exocytic apparatus (Fig. 2, A and B). We found that Sec1 was able to coprecipitate with HA-Mso1 from lysates derived from wild-type cells grown at permissive and elevated temperatures ( Fig. 2A), as shown previously (19). Notable, we found that v-and t-SNAREs of the exocytic complex (i.e. Sso1/2, Sec9, and Snc1/2) also co-immunoprecipitated with Mso1 in a specific fashion ( Fig. 2A). No precipitation of Sec4 was observed in these immunoblots, suggesting that Mso1 does not stably associate with this small GTPase in wild-type yeast. Nevertheless, these results show that Mso1 is a likely component of the exocytic SNARE complex.
Mutations in components of the secretory apparatus often increase or decrease the stability of SNARE complexes at non-permissive temperatures. To verify that Mso1 is a component of the exocytic SNARE complex, we examined whether Sec1 and SNAREs interact with HA-Mso1 in yeast secretion mutants by co-immunoprecipitation. At permissive temperatures, we found that both Sec1 and the exocytic t-SNAREs co-immunoprecipitated with HA-Mso1 from cells known to accumulate SNARE complexes (Fig. 2B, upper panel), such as sec1-1 and sec9-7 cells, which are defective in fusion (17,36), and sec18-1 cells, which are defective in the AAA-ATPase that disassembles cis-SNARE complexes (40). Interestingly, these SNAREs could not be co-immunoprecipitated with HA-Mso1 from lysates derived from sec4-8 cells, which are defective in their ability to assemble SNARE complexes (34). Despite this, the Mso1-Sec1 interaction remained intact in these cells. Thus, even at temperatures permissive for growth, the Mso1-SNARE interaction, but not the Mso1-Sec1 interaction, is inhibited in cells bearing an attenuated allele of Sec4.
We found similar results when examining sec mutants that were shifted to the restrictive temperature (Fig. 2B, lower panel). The exocytic t-SNAREs were still complexed with HA-Mso1 in lysates derived from sec1-1, sec9-7, and sec18-1 cells. In contrast, HA-Mso1 was unable to precipitate these SNAREs from lysates derived from sec4-8 cells (Fig.  2B, lower panel). Thus, normal Sec4 function is necessary to recruit Mso1-Sec1 to the SNARE complex.
We note that the amount of Sso1 that coprecipitated with Mso1 was consistently reduced in the sec9-7 background (Fig. 2, A and B), indicating a possible defect in SNARE complex assembly or stability in this background. We also note that some Sso was observed to precipitate with Mso1 in sec4-8 cells at 26°C (Fig. 2B, upper panel); however, this was not consistently observed (lower panel). Thus, although Sec1 has also been shown to bind to monomeric Sso in vitro (18), Mso1 does not seem able to readily precipitate uncomplexed Sso from cell lysates (this study and Ref. 17).
Given that Mso1 and Sec1 form a stable complex in wild-type cells and late sec mutants (Fig. 2, A and B) (19) and that Sec1 function is required to facilitate assembly of the exocytic SNARE complex (17, 18), we examined whether Sec1 can interact with the Sec9 and Sso t-SNAREs in cells lacking MSO1 (Fig. 2C). We immunoprecipitated HA-Sso1 from cells lacking MSO1 (e.g. mso1⌬ cells) and found that both Sec9 and Sec1 could still co-immunoprecipitate in a specific fashion. Thus, Mso1 is not absolutely required for SNARE assembly or for the association of Sec1 with the SNARE complex.
Finally, we examined whether the interaction of Mso1-Sec1 with the v-and t-SNAREs of the exocytic complex is specific. We immunoprecipitated HA-Mso1 from wild-type cells and probed the precipitates with antibodies against the Sed5 Golgi t-SNARE (41), the Vam3 vacuolar t-SNARE (42), and the Tlg1 early endosomal t-SNARE (43). We found that none of these SNAREs could interact with Mso1 (Fig. 2D). Thus, Mso1 interacts specifically with SNAREs of the exocytic complex.
Mso1 Localizes to the Plasma Membrane under Conditions That Permit SNARE Assembly-Because Mso1 interacts with Sec1 and SNAREs of the exocytic apparatus, we examined its localization in yeast. We expressed N-terminally GFP-tagged Mso1 (including 536 bp of the 3Ј-untranslated region) from a single-copy plasmid and examined its localization by confocal microscopy (Fig. 3A). This construct was capable of improving the growth of sec1-1 cells at semirestrictive temperatures, indicating functionality (data not shown). We found that GFP-Mso1 labeled the plasma membrane in wild-type cells, which is typical for the localization of the Sso and Snc SNAREs (12,13,18). The localization of Sec1 was originally proposed to be at the bud tip and bud neck (17); however, a more recent study also demonstrated co-localization with the Sso t-SNAREs (18). This localization pattern more accurately reflects the physical association of Sec1 with its known partner proteins, the SNAREs, which are distributed all over the plasma membrane (this study and Refs. 12 and 13).
When examining yeast mutants defective in SNARE assembly and disassembly (Fig. 3A), we found that GFP-Mso1 association with the plasma membrane was unaffected in sec1-1 cells, which were rescued by MSO1 expression (Fig. 1A), even at 37°C. In contrast, bud-enriched plasma membrane labeling of GFP-Mso1 was absent from the surface of sec4-8 and sec9-7 cells at both permissive and restrictive temperatures. Similar results were observed in sec18-1 cells, wherein cis-SNARE complexes accumulate. The nature of the punctate structures observed particularly in the sec18-1 cells is currently unknown. Thus, alterations in the ability of SNAREs to assemble into trans-complexes affect the ability of Mso1 to stably associate with the plasma membrane.
Co-localization studies performed with GFP-Mso1 and RFP-Snc1 in wild-type cells showed that both proteins overlapped at the level of the plasma membrane (Fig. 3B), as expected. RFP-Snc1 also showed some localization to the vacuolar membrane, but this is probably a result of overexpression.
Mso1 Shares a Domain Found in the Munc13 and Mint SNARE Regulators-In addition to constitutive secretion, higher organisms also have a stimulus-coupled pathway. This utilizes SNARE family members that are direct orthologs of the yeast SNAREs, but also has an additional level of regulation tied to calcium influx during membrane depolarization. Another SNARE regulator, called Munc13, has been suggested to play an important role in regulating SNARE assembly (7,44), although its molecular function has yet to be fully elucidated. In Caenorhabditis elegans, UNC-13 has been shown to displace UNC-18/Sec1 from the SNARE complex, leading to exocytosis (23). As both Mso1 and the Munc13 family members are (a) connected with exocytosis, (b) interact functionally with Sec1/Munc18, and (c) are components of the exocytic SNARE complex, we examined whether there is sequence homology between these proteins. By performing local sequence alignments, we found that a small region (residues 194 -210) at the C terminus of Mso1 Blots of the total cell lysates (TCL) revealed that all proteins were present at similar levels in wild-type cells. B, HA-Mso1 does not precipitate the exocytic SNARE complex from sec4-8 mutants at permissive and elevated temperatures. Yeast cells (e.g. sec1-1, sec4-8, sec9-7, sec18-1, and wild-type W303-1a cells) bearing either a control plasmid (Ϫ) or a plasmid expressing HA-MSO1 were grown to mid-log phase on selective synthetic medium at either 26°C (upper panel) or shifted to 37°C for 45 min (lower panel) prior to processing for immunoprecipitation and immunoblotting as described above for A and under "Experimental Procedures." C, the exocytic SNARE complex forms in cells lacking MSO1. mso1⌬ cells bearing either a control plasmid (Ϫ) or a plasmid expressing HA-SSO1 were grown to mid-log phase on selective synthetic medium at 26°C prior to processing for immunoprecipitation and immunoblotting for Sec9, Sec1, Sso1, and Sec4 (see A). D, Mso1 does not bind to other SNAREs. Wild-type cells (BY4741) bearing either a control plasmid (Ϫ) or a plasmid expressing HA-MSO1 were grown to log phase and processed for immunoprecipitation. Blots were probed with antibodies against Sed5 (1:3000 dilution), Vam3 (1:5000 dilution), Tlg1 (1:3000 dilution), and Mso1 (1:8000 dilution).
is highly homologous to residues 561-579 of Munc13-2 from rats (Fig.  4A). This region of 19 amino acids is ϳ50% identical and ϳ75% similar and is not found in any other yeast protein (data not shown). Furthermore, introduction of a gap into the Munc13-2 sequence allows for the inclusion of an additional region of homology, beginning at residue 184 of Mso1 and residue 407 of Munc13-2. Because this C-terminal region appears to be conserved between these proteins, it suggests either a possible common mode of action or a domain necessary for proteinprotein interactions. Interestingly, the Munc13-related Mint proteins (22,24) also have homology over this small region (i.e. human MINT1 is 27% identical and 60% homologous to Mso1 over residues 519 -532, LFILTQRIKVLNAD). Moreover, alignment of Mso1 and human MINT1 revealed that Mso1 has low overall homology (22% identity and 34% similarity) to the phosphoinositide-binding domain of the Mint proteins (Fig. 4B). Thus, Munc13 and Munc13-related proteins may have domains structurally conserved in Mso1 and could represent functional orthologs.
We next examined the role of the highly conserved C-terminal domain in Mso1 function. We expressed a deletion mutant (Mso1- ) in the sec mutants and examined them for either the enhancement or inhibition of growth, as seen for full-length Mso1 (Fig. 1, A and B). We found that overexpression of Mso1-(1-193) did not alter the ability of the protein to rescue sec1-1 cells (Fig. 5A). However, we noticed that it inhibited the growth of sec4-8 cells better than native Mso1 (Fig. 5A). Thus, removal of the C terminus of Mso1, which does not alter its association with Sec1 (19), enhances its ability to attenuate the function of Sec4.
Next, we examined the effect of the removal of this domain on GFP-Mso1 localization (Fig. 5B). We found that, unlike GFP-Mso1, GFP-Mso1-(1-193) mislocalized primarily to the cytosol at both 26 and 37°C in wild-type cells. Thus, this conserved region plays a role in Mso1 localization to the plasma membrane.
Finally, we examined whether the truncated form of Mso1 could interact with components of the SNARE complex either in wild-type or mso1⌬ cells by immunoprecipitation (Fig. 5C). We found that both Sec1 and Sso proteins were detectable to a similar extent in precipitates from either cell type. This indicates that Mso1-  interacts as well as full-length Mso1 in binding to these proteins.

DISCUSSION
Numerous proteins are suggested to have roles in modulating SNARE assembly, leading to exocytosis (7)(8)(9)45). However, only those conserved between constitutive and stimulus-coupled systems (i.e. Sec1/ Munc18 and Sec4/Rab3) are likely to encompass the core machinery that regulates assembly and membrane fusion. As Mso1 was shown to interact with Sec1 in yeast and to lead to the accumulation of secretory vesicles (19), we examined whether it is a bona fide component of the exocytic SNARE complex. Our results show that Mso1 is a likely component thereof, being able to coprecipitate the SNARE complex from cells capable of assembling and/or accumulating such complexes (i.e. wild-type, sec1-1, and sec18 cells) (Fig. 2, A and B). In contrast, Mso1 did not precipitate SNAREs from cells unable to assemble such fusion complexes (i.e. sec4-8 cells) (Fig. 2, A and B). Yet, under all conditions tested, Mso1 was found to remain in a complex with Sec1, suggesting that these proteins act together as a pair in regulating SNARE functions. This suggests that, along with Sec1, Mso1 is a component of the exocytic SNARE complex in yeast. This idea is further supported by experiments demonstrating a cellular distribution for Mso1 (Figs. 3 and 5B) identical to that of the exocytic v-and t-SNAREs (12,13) as well as that recently proposed for Sec1 (18).
Although Sec1 and Mso1 appear to interact as a functional pair, Mso1 is not necessary for Sec1 association with the SNARE complex per se, as  sec1-1, sec4-8, sec9-7, and sec18-1 cells) bearing a single-copy plasmid expressing GFP-Mso1 were grown to mid-log phase on selective synthetic medium at 26°C, shifted to medium lacking methionine for 2 h, and then either maintained at 26°C or shifted to 37°C for 45 min prior to visualization. Scale bars ϭ 5 m. B, GFP-Mso1 co-localizes to the plasma membrane with mRFP-Snc1. mso1⌬ cells (EUROSCARF) transformed with plasmids expressing GFP-Mso1 and mRFP-Snc1 were grown to log phase on selective synthetic medium at 26°C, shifted to medium lacking methionine for 2 h, and visualized. shown here (Fig. 2C) and in in vitro studies (18). Thus, although the precise function of Mso1 is unknown, it may facilitate Sec1 binding to the SNARE complex and subsequent membrane fusion. This would account for the secretion defects seen in the absence of MSO1 (19). That said, it is also possible that Mso1 has other binding partners and perhaps additional functions that have not been characterized. Thus, Mso1 and Sec1 may not necessarily be associated in an obligatory fashion.
Interestingly, Mso1 and Munc13 (and less so Mint1) share a common region (Fig. 4A) that is necessary for proper Mso1 localization to the plasma membrane (Fig. 5B). Its removal resulted in a protein with more inhibitory potential directed toward an attenuated allele of Sec4 (Fig.  5A), a GTPase involved in SNARE assembly. Although the precise role for Sec4 in SNARE assembly also remains somewhat elusive, Sec4 has been proposed to act upon vesicle tethering and exocyst function prior to SNARE assembly (46). The connection between Sec4 and Mso1-Sec1 action seen here is striking. We found that a mutation in SEC4 blocked both assembly of the exocytic SNARE complex and association of Mso1-Sec1 with SNAREs (Fig. 2), which is linked to the assembly process (7,17,18). We further note that mso1⌬ deletions are synthetically lethal with either sec4-8 or sec1-1 cells (19) and that both MSO1 and SEC1 inhibited the growth of sec4-8, sec8-9, and sec15-2 cells (Fig. 1B). One explanation for this is that Mso1 activates the GTPase of Sec4 to stimulate the recruitment of Mso1-Sec1 and to facilitate SNARE assembly, a role proposed for Sec1 and other Sec1/Munc18 proteins (7,18,47). Another explanation is that Mso1 turns off Sec4 perhaps after Sec1 facilitation of SNARE assembly. To test the latter possibility, we examined whether MSO1 overexpression could alleviate the synthetic lethality observed between an activating Sec4 mutation (Q61L) and sec15Ϫ1 (48). If Mso1 has intrinsic GTPase-activating protein activity or an ability to activate the Sec4 GTPase-activating proteins Msb3 and Msb4 (48), we expected that its overexpression would relieve the stress exerted by overactive Sec4 on an underactive effector, Sec15-1. However, we were unable to observe rescue of the synthetic lethality seen in sec4-Q61L sec15-1 cells by MSO1 overexpression, 3 as shown previously for both Msb3 and Msb4 (48). Thus, Mso1 may play a different role in regulating Sec4, perhaps in mediating GDP-GTP exchange or protein-protein interactions between Sec4 and Sec1. That said, we were unable to copre-3 A. Castillo-Flores and J. E. Gerst, unpublished data.  . Cells were grown to log phase on selective synthetic medium at 26°C, diluted serially, plated onto prewarmed solid medium, and incubated at various temperatures for 48 h. B, removal of the C-terminal conserved domain of Mso1 alters its cellular localization. Yeast cells bearing a single-copy plasmid expressing either GFP-Mso1 (yEGFP-Mso1) or GFP-Mso1-(1-193) (yEGFP-Mso1  ) were grown to log phase on selective synthetic medium at 26°C, shifted to medium lacking methionine for 2 h, and visualized by confocal microscopy. C, removal of the C-terminal conserved domain of Mso1 does not alter its ability to bind to Sec1 or Sso1. Wildtype and mso1⌬ cells (EUROSCARF) were transformed with a control plasmid (Ϫ) or a plasmid expressing HA-MSO1 or HA-MSO1-  from multicopy plasmids. Cells were grown to log phase and processed for immunoprecipitation (IP) with anti-HA antibodies. Detection was performed with anti-Sec1 (1:4000 dilution), anti-Sso (1:7000 dilution), anti-Sec4 (1:3000 dilution), and anti-HA (1:8000 dilution) antibodies. TCL, total cell lysates. OCTOBER 7, 2005 • VOLUME 280 • NUMBER 40 cipitate Sec4 though an interaction with either Mso1 or Sso1, indicating that the native proteins are not stably associated.

Mso1 and the Exocytic SNARE Complex
Because a conserved region shared between Mso1 and Munc13/Mint is necessary for the proper localization and function of the former, we suggest that Mso1 may be the functional progenitor for Munc13 or Mint functions. The latter are only partly understood, but are known to play important roles in the priming of SNARE assembly, leading to fusion. For example, Munc13 has been proposed to mediate the dissociation of Munc18 from syntaxin-1 to allow for full SNARE assembly, leading to fusion (23,44,49,50). Moreover, the Mint proteins directly interact with Munc18 and thus regulate its ability to associate with syntaxin (21,22). The association of Mint and Munc18 with the SNAREs in mammalian cells corresponds well with the binding of Mso1 to Sec1 and to Sso and the SNARE complex in yeast (this study and Refs. 18 and 19).
As suggested for C. elegans and Drosophila, defects in either Rab3 or UNC-13 function are expected to lead to alterations in SNARE assembly that result in a loss of secretion competence (23, 49 -51). Likewise, the cellular defects in sec4-8 cells parallel those seen with rab mutations in higher organisms (30,34). Thus, for all intents and purposes, the basic machinery regulating SNARE assembly, from Rab proteins to SNARE regulators (e.g. Sec1/Munc18/UNC-18 and Mso1/Munc13/UNC-13) to the SNAREs, has remained intact in evolution. We note, however, that, unlike Sec4, Sec1, and t-SNARE proteins, Mso1 is not essential for vesicle docking and fusion in vegetatively growing cells (19). This suggests that Mso1 plays an auxiliary role in facilitating SNARE assembly and explains why mso1⌬ cells show only defects in exocytosis (19). However, MSO1 was shown to be required for the sporulation of yeast (52), indicating a possible essential role in spore wall formation after meiosis. Why Mso1 plays an essential role in one aspect of cellular development and not another remains unclear. Notable, the knockout of Mint1 in mice is also without a strong phenotype, suggesting either functional redundancy, perhaps with Munc13, or a lesser role in exocytosis (53). Munc13 proteins might play more predominant roles than the Mint proteins. Clearly, a number of studies have shown the importance of Munc13 function in a variety of exocytic programs apart from neurotransmitter release (54 -56), such as insulin secretion (57), the processing of amyloid precursor protein (58), and the fusion of dense core granules in lymphocytes and platelets (59,60). Interestingly, mutations in Munc13-4, the isoform with the most overall homology (21% identity and 37% homology) to Mso1 over 210 amino acids (ClustalW align-ment), 3 result in the genetic secretory disorder familial hemophagocytic lymphohistiocytosis (61).
The strong parallels seen between the Rab and Munc13 functions in C. elegans, Drosophila, and mammals (23, 44, 49 -51) and between Sec4 and Mso1 in yeast (this study) suggest a model by which Mso1 may function in yeast (Fig. 6). Although speculative, the model suggests that, after Sec18/N-ethylmaleimide-sensitive factor-mediated dissociation of cis-SNARE complexes, resulting in the priming of the SNAREs prior to assembly (1,8,45), Sec4 in its activated state recruits the Mso1-Sec1 pair (Fig. 6, step 1). Because Sec1 binds the SNARE complex and is less disposed toward uncomplexed t-SNAREs (17,18), unlike Sly1 or Vps45 (37-39), we predict that Mso1-Sec1 attaches to assembling v-SNARE⅐t-SNARE complexes formed in trans between tethered membranes, resulting in partial assembly of the SNARE complex containing Sec1 (Fig. 6, step 2). This does not preclude the idea that Sec1 binds to uncomplexed Sso to initiate t-SNARE⅐t-SNARE complex assembly (18), although our data and those of Carr et al. (17) do not go far enough to support it. Next, GTP hydrolysis on Sec4 results in the dissociation of Mso1-Sec1, leading to full SNARE assembly (Fig. 6, step 3) and ultimately fusion (step 4). According to this model, Mso1 acts to coordinate the recruitment and subsequent removal of Sec1 in response to the GTPase cycle of the Rab protein. This explains why Mso1-Sec1 does not bind to the SNAREs in sec4-8 cells, but remains associated in sec1-1 and sec18-1 cells, which contain assembled SNAREs. Defects in secretion in the latter two cell types could therefore result from the failure to remove Sec1 from the SNARE complex. Thus, a connection between Sec18/N-ethylmaleimide-sensitive factor action and Sec1 removal should not be ruled out. In addition, this model predicts that physical interactions between Mso1 and Sec4 might occur under circumstances whereby the GTPase is activated. Interestingly, a recent parallel study by Knop et al. (62) also demonstrated Mso1 binding to the SNARE complex. The N terminus of Mso1 was shown to be necessary for binding to Sec1 and presumably the SNAREs. Moreover, binding experiments performed with locked forms of Sec4 (e.g. Sec4-N133I) showed that they could also coprecipitate with Mso1 and the SNARE complex (62). This lends support to the idea that Mso1 and Sec4 interact during SNARE assembly. Thus, it is likely that Mso1 bridges the connection between Rab and Sec1/Munc18 functions, leading to SNARE assembly and fusion. Step 1, upon encroachment and tethering of a secretory vesicle (SV) to the plasma membrane (PM), Sec4 GTPase activation (Sec4*) and ATP-dependent Sec18 priming (to dissociate cis-SNARE complexes) occur (details not shown).
Step 2, Mso1 and Sec1 are recruited as a heteromeric pair to the assembling SNARE complex. This may be initiated by binding to Sso to allow the other SNAREs to assemble in sequence: first Sec9 to form the t-SNARE complex and then Snc to form the v-SNARE⅐t-SNARE complex between the apposed membranes.
Step 3, upon GTP hydrolysis on Sec4, the Mso1-Sec1 pair is removed, and the SNARE complex fully assembles, leading to fusion (step 4). Thus, as seen in this study, inactivating mutations in SEC4 (i.e. sec4-8) prevent the Mso1-Sec1 recruitment and SNARE assembly, whereas mutations in SEC1 (i.e. sec1-1) stabilize the interaction, resulting in a block in fusion. Overexpression of either MSO1 or SEC1 further inhibits sec4-8 cells, probably due to defects in the GTPase cycle of the mutant, which is necessary to remove Mso1-Sec1.