Abstract
SEC9 and SPO20 encode SNARE proteins related to the mammalian SNAP-25 family. Sec9p associates with the SNAREs Sso1/2p and Snc1/2p to promote the fusion of vesicles with the plasma membrane. Spo20p functions with the same two partner SNAREs to mediate the fusion of vesicles with the prospore membrane during sporogenesis. A chimeric molecule, in which the helices of Sec9p that bind to Sso1/2p and Snc1/2p are replaced with the homologous regions of Spo20p, will not support vesicle fusion in vegetative cells. The phosphatidylinositol-4-phosphate-5-kinase MSS4 was isolated as a high-copy suppressor that permits this chimera to rescue the temperature-sensitive growth of a sec9-4 mutant. Suppression by MSS4 is specific to molecules that contain the Spo20p helical domains. This suppression requires an intact copy of SPO14, encoding phospholipase D. Overexpression of MSS4 leads to a recruitment of the Spo14 protein to the plasma membrane and this may be the basis for MSS4 action. Consistent with this, deletion of KES1, a gene that behaves as a negative regulator of SPO14, also promotes the function of SPO20 in vegetative cells. These results indicate that elevated levels of phosphatidic acid in the membrane may be required specifically for the function of SNARE complexes containing Spo20p.
THE endomembrane system of the eukaryotic cell consists of a series of distinct membrane-bound compartments. Lipids and proteins are shuttled between compartments by carrier vesicles. Maintenance of proper organization of the endomembrane system requires regulation of vesicle flow such that transport vesicles fuse only at the appropriate acceptor compartment. Soluble NEM-sensitive factor receptor (SNARE) complexes are thought to play a key role in maintaining the specificity of vesicle docking and may directly mediate vesicle fusion (Sollneret al. 1993; Rothman 1994; Rothman and Warren 1994; Pelham 1999; Parlatiet al. 2002; Jahnet al. 2003).
SNAREs are a family of proteins that share a related 60-amino-acid (aa) helical region (Weimbset al. 1997). Different SNAREs can interact to form oligomers at whose core is a bundle of four such helices (Suttonet al. 1998). In most cases, SNAREs are integral membrane proteins and the helical region of the SNARE protein is adjacent to a transmembrane domain. When SNAREs present in the membrane of an acceptor compartment (t-SNAREs) form complexes with SNAREs in the vesicle membrane (v-SNAREs) the transmembrane domains of the SNAREs in both membranes are brought into close proximity and this is proposed to lead, through an as yet undefined mechanism, to the fusion of the lipid bilayers (Weberet al. 1998). Specificity as to which combinations of SNAREs will form productive fusion complexes is proposed as the means by which fusion between vesicles and their appropriate target membranes is regulated (Sollneret al. 1993; Rothman and Warren 1994; Parlatiet al. 2002). However, as isolated SNAREs show only limited binding specificity in vitro (Fasshaueret al. 1999), the molecular basis of in vivo SNARE specificity remains unknown.
One SNARE subfamily, the SNAP-25-related proteins, has an architecture different from that of other SNAREs in that these proteins lack a transmembrane domain but contain two SNARE helices (Hesset al. 1992; Weimbset al. 1997). Complexes including these SNAREs are heterotrimers (Suttonet al. 1998). In the budding yeast Saccharomyces cerevisiae, the two SNAP-25 orthologs Sec9p and Spo20p interact with the t-SNAREs Sso1/2p and the v-SNAREs Snc1/2p to mediate the fusion of exocytic vesicles with the plasma membrane during vegetative growth or the prospore membrane during sporulation, respectively (Gerstet al. 1992; Protopopovet al. 1993; Brennwaldet al. 1994; Couve and Gerst 1994; Rossiet al. 1997; Neiman 1998).
Sporulation is driven by an unusual cell division event in which daughter nuclei are enveloped within plasma membranes, termed prospore membranes, which form de novo within the cytoplasm of the mother cell. These prospore membranes are formed by the redirection of post-Golgi secretory vesicles away from the cell surface and the subsequent fusion of these vesicles within the cytoplasm (Neiman 1998). This fusion event requires much of the same machinery as fusion of vesicles with the plasma membrane, with the exception that Sec9p is largely replaced by Spo20p.
The Spo20 and Sec9 proteins share 40% sequence identity through their conserved SNAP-25 domain and both partner with Ssop and Sncp to mediate the fusion of post-Golgi vesicles with a target membrane (Neiman 1998; Neimanet al. 2000). However, they are specific to their respective compartments; Spo20p functions at only the prospore membrane and cannot function at the plasma membrane and the reverse is true for Sec9p. Thus, changing only one subunit of this SNARE complex, Sec9p or Spo20p, is sufficient to change the target membrane specificity of the complex. Sec9p and Spo20p therefore provide an excellent model system to investigate the basis for SNARE specificity in vivo.
In previous work, a collection of chimeric SEC9/SPO20 genes was used to define domains important for the vegetative- or sporulation-specific activity of the two gene products (Neimanet al. 2000). Of particular relevance to the current study, an asymmetry was identified in the functionality of the helical regions that mediate SNARE assembly. Specifically, a chimeric protein in which the Spo20p helices have been replaced with those of Sec9p can rescue sporulation in a spo20Δ mutant; however, the reciprocal chimera, Sec9p with the Spo20p helices, cannot support growth of a sec9-ts mutant. In this report, we explore the basis for this inability of the Spo20p helices to support vesicle fusion at the plasma membrane. Our results suggest that the lipid composition of the membrane can affect the ability of specific SNARE complexes to mediate membrane fusion.
MATERIALS AND METHODS
Strains and media: Standard media and genetic methods were used (Rose and Fink 1990). Strain AN123-4A (MATa ura3 leu2-2,113 trp1 lys2 sec9-4) is a segregant of a cross of W303-1b and AN63-2C (Neiman 1998). AN123-4A derivatives expressing various SNARE chimeras under control of the SEC9 promoter were constructed by transformation of AN123-4A with integrating plasmids expressing the appropriate chimera, as described previously (Neimanet al. 2000). Strain AN1076 (MATa ura3 leu2-2,113 trp1 lys2 sec9-4 spo14::ΔURA3) was constructed by transformation of AN123-4A with an XbaI-ClaI fragment from plasmid pKR466 (Roseet al. 1995). Strain AN1075 (MATa ura3 leu2-2,113 trp1 lys2 sec9-4 kes1::ΔURA3) was constructed by transformation of AN123-4A with an EcoRI-BamHI fragment from pRE352 (Fanget al. 1996). Strain AC14 (MATa ura3 leu2 arg4 lys2 hoΔLYS2 kes1::ΔURA3 spo14::ΔURA3) is a segregant of a cross of strain AN1075 with Y435 (Roseet al. 1995). Yeast strains were transformed by the lithium acetate procedure (Itoet al. 1983).
Identification of the suppressor gene: The suppressor plasmid pW18 contained an ∼9-kb segment of chromosome IV carrying a portion of a TY element as well as open reading frames (ORFs) YDR210w, YDR209c, YDR208w (MSS4), and YDR207c (UME6). Deletion of ORFs YDR210w, YDR209c, and a portion of MSS4 using internal SpeI sites led to a loss of suppressor activity, indicating that one of these three ORFs encodes the suppressor. Subcloning experiments further showed that suppressor activity was contained in a 3.9-kb BamHI-XhoI fragment that contained only MSS4 and a small portion of the UME6 gene, thus identifying the suppressor as MSS4.
Plasmids: Integrating plasmids for expression of the various chimeric SNAREs have been described previously (Neimanet al. 2000). The pRS426-Sec9pr-SPO20 plasmid was constructed by cloning a 1.7-kb SacI-KpnI fragment carrying the SPO20 coding region under control of the SEC9 promoter from plasmid pRS306-Sec9pr-Spo20 (Neimanet al. 2000) into similarly digested pRS426 (Christiansonet al. 1992). The YEp-GFP-Spo14 and GFP-Spo14K533E plasmids (Sciorraet al. 2002) were provided by J. Engebrecht. To construct pRS425-MSS4, a 3.9-kb BamHI-XhoI fragment carrying the MSS4 ORF with ∼1.2 kb of upstream and 400 bp of downstream sequence was subcloned from the suppressor plasmid pW18 into BamHI-XhoI-digested pRS425 (Christiansonet al. 1992).
Microscopy: For analysis of GFP-Spo14p, strains carrying pYEp-GFP-Spo14 were grown to midlog phase in YPD medium and examined on a Zeiss Axioplan 2 microscope. Images were acquired using a Zeiss Axiocam mRM and Axiovision 3.0.6 software.
RESULTS
Identification of a high-copy plasmid that promotes the function of a chimeric SNARE: In previous work (Neimanet al. 2000), we found that a SEC9 gene in which the regions encoding the helical domains were replaced with the corresponding regions of SPO20 (designated SPSP—each domain in a chimeric gene is designated “S” if the sequence comes from SEC9 or “P” if the sequence comes from SPO20; letters indicate the origin, in order, of the amino-terminal domain, helix 1, the interhelical domain, and helix 2) could not rescue the temperature-sensitive growth defect of a sec9-4 strain. By contrast, the reciprocal case in which the helical regions of SPO20 were replaced with those of SEC9 (PSPS) was competent to support the sporulation of a spo20Δ mutant. To investigate the reason for this difference, a screen was performed to identify genes that, when overexpressed, allow the SPO20 helical regions to function during vegetative growth. A sec9-4 strain was constructed that carried the SPSP SNARE integrated into the chromosome between a duplication of the URA3 locus and expressed from the SEC9 promoter. The expression level of the SPSP protein is comparable to native SEC9 (Neimanet al. 2000).
This strain (AN123-4A-SPSP) was transformed with a high-copy genomic library (Yu and Hirsch 1995). From ∼26,000 transformants, 60 clones capable of growth at 37° were isolated, of which 49 retested as displaying growth above background at 37°. To distinguish between plasmids that might directly suppress the sec9-4 defect (e.g., SEC9) and those that act by stimulating the activity of the SPSP chimera, the candidates were replica plated to medium containing 5-fluoroorotic acid to select for cells that had lost the SPSP expression construct. After loss of the SPSP gene, the strains were again tested for growth at 37°. Of the initial candidates, 10 reverted to temperature sensitivity after loss of the SPSP gene. Library plasmids were recovered from these 10 candidates and retransformed into strain AN123-4A-SPSP. Only one plasmid, pW18, reproducibly gave good growth at high temperatures (Figure 1). Sequencing of the ends of the genomic insert in this plasmid indicated that it carried a 9-kb segment of chromosome IV including several ORFs. Further deletion and subcloning of the suppressing activity (see materials and methods) identified the suppressor as the MSS4 gene.
MSS4 encodes phosphatidylinositol-4-phosphate-5-kinase (PI4P-5-kinase). MSS4 is the only PI4P-5-kinase in S. cerevisiae and is an essential gene (Desriviereset al. 1998; Hommaet al. 1998). Although not previously implicated in secretory function, the protein is localized predominantly to the plasma membrane, as is the enzyme's product, phosphatidylinositol (4,5)-bisphosphate (PI4,5P2) (Desriviereset al. 1998; Hommaet al. 1998; Stefanet al. 2002). Importantly, it has been shown that expression of MSS4 from a 2μ plasmid elevates intracellular PI4,5P2 levels about twofold (Desriviereset al. 1998). The identification of MSS4 thus suggests that elevation of PI4,5P2 levels might promote the function of the SPSP SNARE during vegetative growth.
Suppression by MSS4 is specific to SNAREs containing the SPO20 helices: In addition to the sporulation-specific function of the Spo20p helices, the Sec9/Spo20 chimera studies identified two distinct functional regions of the Spo20p amino terminus, a domain that is inhibitory to SNARE function and a domain specifically required for function at the prospore membrane (Neimanet al. 2000). To determine if suppression by MSS4 was specific to the SPO20 helices, the ability of MSS4 to promote the rescue of sec9-4 by other chimeras, as well as by native SPO20, was tested (Figure 2). To allow for a more quantitative assessment of the suppression, serial 10-fold dilutions of the indicated strains were spotted onto selective medium and incubated at either 37° or 25°.
In this assay, overexpression of MSS4 alone gave a slight suppression of sec9-4 (Figure 2, rows 2 and 3). Expression of SPSP alone showed no effect (Figure 2, row 5), but coexpression of both MSS4 and SPSP (Figure 2, row 4) improved the plating efficiency by two to three orders of magnitude, comparable to expression of SEC9 itself.
—MSS4 rescues the growth of sec9-ts cells in the presence of the SPSP SNARE. Strains AN123-4A (sec9-4) and AN123-4A SPSP (sec9-4 ura3::Sec9pr-SPSP::URA3) were transformed with high-copy plasmids containing MSS4 (pRS425-MSS4), SEC9 (Yep352-SEC9), or an empty vector (pRS425). Cells were streaked out for single colonies and grown at the indicated temperature for 3 days.
Overexpression of MSS4 also allowed a sec9-4 strain expressing only the conserved SNAP-25 domain of SPO20 (xPPP), which contains the two SNARE helices, to grow well at 37° (Figure 2, row 6). By contrast, because it lacks the amino-terminal domain essential for function in sporulation, the SNAP-25 domain of Spo20p will not complement the spo20Δ sporulation defect (Neimanet al. 2000). Moreover, overexpression of MSS4 does not allow xPPP to complement the spo20 sporulation defect (data not shown). Thus, increased MSS4 specifically allows the Spo20p SNARE domain to function in vesicle fusion at the plasma membrane.
Fusion of the SPO20 amino-terminal domain to the SEC9 SNAP-25 domain (PSSS) interferes with the ability of this region of SEC9 to complement sec9-4 (Neimanet al. 2000). To examine if MSS4 can influence the inhibitory activity of the Spo20p amino terminus, the ability of MSS4 to promote the function of PSSS was tested. Overexpression of MSS4 does not allow rescue of sec9-4 by PSSS (Figure 2, row 8). Thus, the ability of MSS4 to promote Spo20p function appears to be specific to the Spo20p SNARE domain.
Consistent with the inability of MSS4 to overcome inhibition by the Spo20p amino terminus, MSS4 will not permit an integrated copy of SPO20 to rescue sec9-4 (Figure 2, row 10). However, when SPO20, expressed from the SEC9 promoter, is introduced into sec9-4 cells on a 2μ plasmid along with MSS4, then growth of the sec9-4 strain is seen (Figure 2, row 12). As high-copy MSS4 does not bypass the negative effect of the Spo20p amino terminus, this last result indicates that, in the presence of overexpressed MSS4, overexpression of SPO20 itself overcomes the inhibitory property of the Spo20p amino terminus.
SPO14 is required for suppression by MSS4: The isolation of MSS4 as a high-copy suppressor suggests that elevation of PI4,5P2 levels promotes the function of the SPO20 helices. This effect may be quite indirect; for example, elevation of PI4,5P2 levels might promote the function of some PI4,5P2 binding proteins. Indeed, a large number of PI4,5P2 binding proteins with diverse intracellular functions have been identified (Ostranderet al. 1995; Levine and Munro 1998; Ojalaet al. 2001; Palmgrenet al. 2001; Hallettet al. 2002; Sciorraet al. 2002). Of particular interest in this regard is the SPO14 gene product. SPO14 encodes a phosphatidylcholine-specific phospholipase D and the activity of the enzyme requires the presence of PI4,5P2 (Sciorraet al. 1999). Further, although spo14 mutants have only minimal phenotypes in vegetative cells, they are defective in sporulation due to an inability to form the prospore membrane and, in fact, Spo14p is localized to the prospore membrane during sporulation (Roseet al. 1995; Rudgeet al. 1998b).
Because SPO14 encodes a PI4,5P2-dependent enzyme required for formation of the same membrane at which Spo20p normally functions, we examined whether the action of MSS4 on SNARE function required the presence of SPO14. The SPO14 gene was deleted in our sec9-4 strain and MSS4 was introduced on a high-copy vector along with high-copy SPO20. Deletion of SPO14 caused a slight decrease in the growth rate of the sec9-4 strain (Figure 3, 25°). More importantly, in the absence of SPO14, overexpression of MSS4 was unable to promote SPO20 function (Figure 3, rows 4 and 5). Even the modest improvement of growth by overexpression of MSS4 alone was lost (Figure 3, rows 3 and 6). Thus, the suppression activity of MSS4 is dependent on SPO14. These data suggest that MSS4 might promote SPO20 function by altering the activity or localization of the SPO14 gene product.
—MSS4 specifically promotes the function of SNAREs containing the helical domains of Spo20p. Strain AN123-4A (sec9-4) and derivatives expressing the indicated SNARE were transformed with high-copy plasmids carrying MSS4 (pRS425-MSS4), SEC9 (Yep352-SEC9), or an empty vector (pRS425). Strains were grown to saturation in selective medium and 10-fold serial dilutions were plated. Plates were photographed after 3 days incubation at the indicated temperature.
Localization of Spo14-GFP is altered by MSS4 overexpression: To investigate the possible effect of the suppressor plasmids on SPO14, we first examined whether overexpression of SPO14 itself could promote SPO20 function. A high-copy SPO14 plasmid was introduced into AN123-4A along with high-copy SPO20. Unlike MSS4, overexpression of SPO14 was unable to promote rescue of sec9-4 by SPO20 (data not shown). This may indicate that the action of MSS4 requires a change in the localization of SPO14 rather than a simple increase in SPO14 activity.
—Suppression by MSS4 requires an intact SPO14 gene. Strains AN123-4A (sec9-4) and AN1076 (sec9-4 spo14Δ::URA3) were transformed with high-copy plasmids carrying MSS4 (pRS425-MSS4), SEC9 (Yep-352-SEC9), an empty vector (pRS425), or two plasmids to introduce both MSS4 (pRS425-MSS4) and SPO20 (pRS426-Sec9pr-SPO20). Cultures were grown to saturation in selective medium and 10-fold serial dilutions were plated. Plates were photographed after 3 days incubation at the indicated temperature.
—Localization patterns of GFP-Spo14p in cells overexpressing MSS4. Strain AN123-4A was transformed with plasmids pRS425-MSS4 and pYEp-GFP-SPO14, and cells were grown to log phase at 30° and examined by fluorescence microscopy. Examples of each of the classes listed in Table 1 and described in the text are shown. (A) Dots underneath the plasma membrane. (B) Plasma membrane. (C) Intracellular rings and dots. (D) Diffuse cytosolic fluorescence. Bars: (A and D) 1 μm, (B and C) 1.5 μm.
To examine the localization of Spo14p, a Spo14-green fluorescent protein (GFP) fusion was introduced into strain AN123-4A. Vegetatively growing cells expressing Spo14-GFP displayed a diffuse pattern throughout the cytosol (similar to Figure 4D) as has been reported previously (Rudgeet al. 1998b). A small but significant fraction of the cells (7.5%) displayed weak plasma membrane fluorescence. Introduction of high-copy MSS4 produced a marked alteration in GFP-Spo14p localization. Four different patterns of localization were seen. Examples of each pattern are shown in Figure 4 and the distribution of cells displaying each pattern is given in Table 1. The most striking change was an increase in the fraction of cells (44%) showing plasma membrane localization of GFP-Spo14p (Figure 4B). Additionally, two minor classes showing distinct intracellular staining, either dots underneath the plasma membrane (Figure 4A) or larger intracellular rings and dots (Figure 4C), were identified. These patterns were very rare or absent in the control cells. The precise nature of these intracellular structures is not yet clear, although they may represent localization of Spo14-GFP to an endosomal compartment, as reported previously (Liet al. 2000). Nonetheless, these data indicate that overexpression of MSS4 alters the intracellular localization of Spo14p, particularly by increasing its abundance on the plasma membrane. This relocalization of Spo14p may be the means by which MSS4 promotes SPO20 function.
The effect of MSS4 overexpression on the distribution of GFP-Spo14p
Spo14p carries two PI4,5P2 binding domains. The first is a polybasic motif required for catalytic activity and the second is a pleckstrin homology (PH) domain by which PI4,5P2 influences the localization of the protein (Sciorra et al. 1999, 2002). To examine if elevated PI4,5P2 levels generated by MSS4 overexpression are responsible for the mislocalization of Spo14, the localization of Spo14K533E carrying a mutation that inactivates PI4,5P2 binding by the PH domain was examined (Table 1). Unlike the wild-type protein, GFP-Spo14K533E was not relocalized by overexpression of MSS4. This result indicates that the elevation of PI4,5P2 levels is responsible for plasma membrane recruitment of Spo14p.
Deletion of the KES1 gene promotes SPO20 function in vegetative cells: The results presented define a novel phenotype for spo14 in vegetative cells. A known phenotype for SPO14 in vegetative growth is its role in “bypass sec14”(Sreenivaset al. 1998; Xieet al. 1998). The SEC14 gene encodes an essential protein required for transport through the Golgi apparatus. A number of second-site suppressors that suppress the lethality of a sec14 mutation have been identified (Cleves et al. 1989, 1991; Albet al. 1995; Fanget al. 1996). Although these mutations affect various aspects of lipid metabolism they share the common feature that SPO14 activity is required for the bypass phenotype.
Mutation of the KES1 gene produces a bypass sec14 phenotype (Fanget al. 1996). To test whether the bypass sec14 phenotype was related to SPO14-mediated promotion of SPO20 function, the KES1 gene was deleted from our sec9-4 strain and the ability of 2μ SPO20 to confer growth at 37° was examined. Remarkably, deletion of KES1 was sufficient to allow high-copy SPO20 to suppress sec9-4 (Figure 5, row 5). As with overexpression of MSS4, the ability of the kes1Δ mutant to promote SPO20 function required an intact copy of SPO14 (Figure 5, row 8). These observations suggest that mutation of KES1 might similarly cause a redistribution of Spo14p to the plasma membrane. In contrast to MSS4 overexpression, however, the subcellular distribution of GFP-Spo14p in the kes1Δ mutant strain was not significantly different from that of wild-type cells (data not shown). Although the precise mechanism by which deletion of kes1 leads to SPO14-dependent suppression remains unknown, these results provide further evidence that alterations in lipid metabolism enhance the ability of the Spo20p helices to promote vesicle fusion at the plasma membrane.
DISCUSSION
The results presented document a role for lipid-modifying enzymes in the function of a specific SNARE molecule. Overexpression of the PI4P-5-kinase MSS4 allows SNAREs containing the Spo20p helical regions to support fusion of vesicles at the plasma membrane. However, this is true only if the phospholipase D encoded by SPO14 is present. The exact relationship between SPO14 and MSS4 in this system is still unclear. The Spo14 protein has been shown to require binding to PI4,5P2 for enzymatic activity and for localization (Sciorra et al. 1999, 2002). Conversely, in higher cells, the activity of PI4P-5-kinase is promoted by phospholipase D activity (Hondaet al. 1999). Thus, overexpression of MSS4 might promote Spo14p activity or SPO14 might be necessary to support increased activity of Mss4p in the overexpressing cells. It should be noted, however, that stimulation of PI4P-5-kinase by phospholipase D in higher cells is mediated via the phospholipase D activator ARF (Hammondet al. 1995; Hondaet al. 1999). By contrast, Spo14p is not activated by ARF in yeast cells (Rudgeet al. 1998a). Moreover, recent studies suggest that SPO14 acts downstream of MSS4 in sporulating cells (J. Engebrecht, personal communication). Given these results and our observation that high-copy MSS4 leads to relocalization of Spo14p to the plasma membrane, the simplest interpretation of our results is that overexpression of MSS4 elevates the levels of PI4,5P2 in the plasma membrane. This elevation recruits Spo14p to the plasma membrane, leading to an elevation of the Spo14p product phosphatidic acid (PA) in the plasma membrane and, in turn, PA aids Spo20p in catalyzing membrane fusion.
—Deletion of KES1 allows rescue of sec9-4 by SPO20. Strains AN123-4A (sec9-4), AN1075 (sec9-4 kes1::ΔURA3), or AC14 (sec9-4 kes1::ΔURA3 spo14::ΔURA3) were transformed with the indicated plasmid. Cultures were grown to saturation and 10-fold serial dilutions were plated and incubated at either 37° or 25°.
Analysis of the lipids of the plasma membrane indicates that its composition is distinct from that of other intracellular membranes (Schneiteret al. 1999). Although no lipid analysis of the prospore membrane has been reported, studies of SPO14 suggest that the prospore membrane may have a lipid composition different from that of the plasma membrane (Rudgeet al. 2001). SPO14 is required for the formation of the prospore membrane and, further, Spo14p is localized to the prospore membrane during normal spore morphogenesis. These results suggest both that the production of PA by Spo14p is required for the coalescence of vesicles to form a prospore membrane and that the prospore membrane might be richer in PA than the plasma membrane. Thus, overexpression of MSS4, by recruiting Spo14p to the plasma membrane, may serve to make the plasma membrane more similar to the prospore membrane and thereby promote Spo20p function.
The proposition that alterations in lipid composition of the plasma membrane are promoting function of Spo20p is strengthened by our findings with kes1. KES1 encodes a Golgi-resident protein that is a member of the oxysterol-binding protein family; however, the lipid ligand for Kes1p is not a sterol, but rather phosphatidyl-inositol-4-phosphate (Liet al. 2002). Mutation of KES1 bypasses the need for SEC14 for Golgi function in a SPO14-dependent manner and, as with other sec14-bypass mutations, is thought to do so by elevating PA or diacylglycerol (DAG) levels in the Golgi. Further, overexpression of KES1 phenocopies a deletion of the SPO14 gene (Fanget al. 1996; Xieet al. 1998; Liet al. 2002). Thus, KES1 behaves genetically as if it is antagonistic to SPO14, although no direct effect of KES1 on Spo14p activity has been reported. Our observations strengthen the proposal that KES1 acts in opposition to SPO14 (Rivaset al. 1999).
One important unresolved issue is how PA acts to enhance Spo20p activity. For example, the critical lipid might be PA or some derivative thereof such as DAG or lyso-PA. More generally, it remains to be determined whether the effects we see are mediated through a protein intermediate or by the lipid directly. The two most straightforward possibilities are that (1) PA stimulates the activity of a PA-binding protein, which in turn aids in fusion of vesicles, or (2) the increased concentration of PA itself, by altering the lipid composition of the membrane, is directly responsible for rescuing vesicle fusion.
How might membrane composition promote or prohibit SNARE function? In current models of SNARE-mediated membrane fusion, the oligomerization of the v- and t-SNARE helices drives the two membranes together (Jahnet al. 2003). To drive the two membranes together, the energy provided by assembly must be sufficient to overcome the potential energy barrier to fusion. This barrier is produced by both repulsion of the negatively charged membrane surfaces and the stress of membrane bending in the fusion intermediates (Kuzminet al. 2001; Kozlovsky and Kozlov 2002; Jahnet al. 2003). Perhaps the energy made available by assembly of SNARE complexes consisting of Spo20-Sso-Snc is insufficient to drive fusion at the yeast plasma membrane. Elevation of PA levels in the membrane, we suggest, might lower the potential energy barrier to fusion, permitting the Spo20p helices to work.
Previous chimera studies (Neimanet al. 2000) indicated that the inability of SPO20 to function in vegetative cells is due to two factors: (1) inhibition of SNARE function by an amino-terminal domain and (2) the inability of the Spo20p helices to promote vesicle fusion at the plasma membrane. The results presented here suggest that this latter deficiency is related to the lipid composition of the plasma membrane. If the barrier to fusion at the plasma membrane is intrinsically higher than that at the prospore membrane, then the plasma membrane SNARE Sec9p might be expected to function well at the prospore membrane. Indeed, chimeras containing the Sec9p helices can promote fusion at the prospore membrane. However, intact Sec9p cannot. This failure is largely attributable to the fact that Spo20p contains, in its unique amino-terminal domain, a short region essential for SNARE activity at the prospore membrane (Neimanet al. 2000). This short region binds to PA in vitro (H. Nakanishi and A. M. Neiman, unpublished observations), suggesting that differences in lipid composition of the prospore membrane might also account for the inability of Sec9p to function at that compartment.
In a reconstituted liposome fusion assay, SNARE proteins display high selectivity in complex assembly (Parlatiet al. 2002). However, additional factors appear to contribute to specificity in vivo (Jahnet al. 2003). Our results suggest that a specific lipid environment is required for the function of Spo20p. Some evidence suggests that other SNAREs also prefer specific lipid environments. In higher cells, the plasma membrane SNAREs have been shown to be present in lipid rafts or cholesterol-dependent clusters (Lafontet al. 1999; Chamberlainet al. 2001; Langet al. 2001). Lipid rafts are enriched in cholesterol and, at synapses, this may enhance exocytosis by aiding in the recruitment of the cholesterol-binding SNARE cofactor synaptophysin (Mitteret al. 2003). However, plasma membrane SNAREs are also found in cholesterol-rich domains in nonneuronal cells that lack synaptophysin (Lafontet al. 1999; Chamberlainet al. 2001). In these cases, whether the lipids of the raft similarly serve to recruit necessary cofactors or, perhaps, directly influence the fusion of lipid bilayers remains to be determined.
Acknowledgments
The authors thank J. Engebrecht (University of California, Davis) for comments on the manuscript and communication of results prior to publication. We also thank J. Engebrecht and V. Bankaitis (University of North Carolina, Chapel Hill) for suggestions and plasmids. This work was supported by National Institutes of Health grant GM62184 to A.M.N.
Footnotes
-
Communicating editor: M. D. Rose
- Received June 23, 2003.
- Accepted October 9, 2003.
- Copyright © 2004 by the Genetics Society of America
