SEC14-dependent Secretion inSaccharomyces cerevisiae

The SEC14 gene in Saccharomyces cerevisiae encodes a phosphatidylinositol transfer protein required for secretory protein movement from the Golgi. Mutation ofSAC1, a gene of unknown function, restores secretory flow in sec14-1 ts strains. The existing model for the bypass of the sec14-1 ts defect bysac1-22 involves stimulation of sphingolipid biosynthesis and, in particular, the synthesis of mannosyl-diinositolphosphoryl-ceramide with concomitant increases in Golgi diacylglycerol levels. To test this model, we disruptedIPT1, the mannosyl-diinositolphosphoryl-ceramide synthase of S. cerevisiae. Disruption of the IPT1 gene had no effect on the ability of sac1-22 to bypasssec14-1 ts. Furthermore, sphingolipid analysis of sec14-1 ts andsec14-1 ts sac1-22 strains showed that mannosyl-diinositolphosphoryl-ceramide synthesis was not stimulated in the bypass mutant. However, thesec14-1 ts strain had elevated mannosyl-monoinositolphosphoryl-ceramide levels, and thesec14-1 ts sac1-22 strain showed an 8-fold increase in phosphatidylinositol 4-phosphate along with a decrease in phosphatidylinositol 4,5-bisphosphate. Cellular diacylglycerol levels, measured by [14C]acetate incorporation, did not differ between thesec14-1 ts and the sec14-1 sac1-22bypass strains, although disruption of IPT1 in the bypass strain resulted in reduced levels. These data indicate that phosphatidylinositol 4-phosphate, rather than mannosyl-diinositolphosphoryl-ceramide, accumulates in thesec14-1 ts sac1-22 bypass strain, and that Golgi diacylglycerol accumulation is not required for bypass of the sec14-1 ts growth and secretory phenotypes.

The yeast Saccharomyces cerevisiae is an important model system for determining the molecular mechanisms of eukaryotic intracellular protein transport. Of particular interest is the yeast secretory pathway in which SEC genes essential for protein secretion by S. cerevisiae have been identified (1). These genes were discovered as conditional lethal mutations that interrupt secretory flow at specific points along the secretory pathway. Biochemical and functional identification of the SEC gene products has increased the understanding of mechanisms involved in the yeast secretory pathway.
The yeast SEC14 gene is required for the formation of transport vesicles from the Golgi (2). However, the mechanism by which its encoded protein, Sec14p, facilitates this process is not precisely known. When conditional lethal mutant sec14-1 ts strains are shifted to nonpermissive temperatures, protein secretion halts (1). Sec14p is a phosphatidylinositol transfer protein with the ability to transfer both phosphatidylinositol and phosphatidylcholine between membranes (2). A unique protein in S. cerevisiae, Sec14p, is set apart from other phospholipid transfer proteins by its ability to interact specifically with the Golgi (3), an association that is required for secretory competence in yeast (4). Sec14p is proposed to have a sensor function that maintains a critical phosphatidylinositol to phosphatidylcholine ratio required for vesicle formation at the Golgi (3).
To further understand the mechanism of SEC14 action in the secretory pathway, suppressor mutants of sec14-1 ts were identified (3). Characterization of the genes defective in these mutants revealed that genes of three of six gene complementation groups identified were involved in the CDP-choline pathway for phosphatidylcholine biosynthesis. Apparently, inhibition of phosphatidylcholine synthesis through the CDP-choline pathway leads to efficient bypass of sec14-1 ts , probably by altering phospholipid composition of the Golgi membrane.
Other suppressors of sec14-1 ts are mutations of the SAC1 gene (5). Originally identified as a suppressor of act1 alleles when mutated, SAC1 encodes an integral membrane protein that is located in the endoplasmic reticulum and Golgi (6). Generally, SAC1 mutations lead to a cold-sensitive phenotype and inositol auxotrophy. However, mutants with the sac1-22 allele do not display inositol auxotrophy, despite an inositol requirement for its bypass of sec14-1 ts . Sac1p has significant homology with the noncatalytic domains of the yeast and mammalian polyphosphoinositide 5-phosphatases (7), but its function is not known. Phospholipase D activity was recently shown to be essential for bypass of the sec14-1 ts phenotypes in several bypass mutants, but not for normal secretion (8,9). The formation of phosphatidic acid from phospholipase D-mediated hydrolysis of phosphatidylcholine has been implicated in this process. This is supported by results showing that phosphatidic acid promotes Golgi vesicle formation in a mammalian cell system (10).
Kearns et al. (11) suggested that sec14-1 ts bypass by sac1-22 occurs by increasing the amount of diacylglycerol (DAG) 1 in the Golgi through stimulated sphingolipid biosynthesis. This model was based on an apparent 6-fold increase in the cellular levels of mannosyl-diinositolphosphoryl-ceramide (M(IP) 2 C). M(IP) 2 C synthesis occurs in the Golgi with the transfer of phosphorylinositol from phosphatidylinositol to mannosylmonoinositolphosphoryl-ceramide (MIPC) to yield a molar equivalent of DAG (12). The sac1-22 bypass was perturbed by heterologous expression of Escherichia coli DAG kinase, pre-sumably due to the conversion of DAG to phosphatidic acid.
The recent identification of IPT1 as M(IP) 2 C synthase (13) allowed us to test the role of M(IP) 2 C production in the bypass of sec14-1 ts by sac1-22. IPT1 was also recently found to be identical to SYR4, a gene necessary for yeast growth inhibition by the bacterial metabolite syringomycin E. 2 In this work, we report that disruption of IPT1 has no effect on the bypass, indicating that M(IP) 2 C production is not important in the sac1-22 suppression of sec14-1 ts . Instead, phosphatidylinositol 4-phosphate (PtdIns(4)P) is shown to increase in the sec14-1 ts sac1-22 bypass mutants, suggesting a role for this lipid in Golgi-localized secretory pathway events.  (11), and CTY-SR (CTY165 ⌬ipt1::URA3, this work). Cells were typically grown in synthetic complete (SC) media as described by Kaiser et al. (15). For lipid analyses, cells were grown in synthetic minimal medium (16).
Disruption of IPT1 in CTY165-The IPT1 disruptant strain CTY-SR was constructed by the one-step disruption method (17). The IPT1 disruption construct described elsewhere 2 was made by replacing a 1-kilobase PvuII-PvuII fragment, which included the 5Ј portion of IPT1, with a 1.1-kilobase URA3 fragment. This construct was linearized and used to transform CTY165. Disruptants were selected on SC-ura, and IPT1 disruption was confirmed by Southern blot analysis.
Liquid Secretion Assay-Ten-ml cultures grown in SC medium at 28°C were harvested at an optical density between 0.5 and 1 at 600 nm, washed once in SC low-glucose medium (1 mM glucose), resuspended into 10 ml of low-glucose medium, and split into two 5-ml cultures. Low-glucose cultures were incubated at either 28°C or 37°C for 1 h. Secretory indices were determined by assaying the total and secreted invertase activity as described by Westphal et al. (18).
Quantification of Sphingolipids-Sphingolipid extractions were modified from the methods described by Smith and Lester (19). Steadystate measurements of sphingolipid levels were conducted by growing 20-ml cultures for 18 -24 h at 28°C in the presence of 200 Ci of H 3 32 PO 4 . Growth was terminated by the addition of trichloroacetic acid to a final concentration of 5%. Cells were washed twice with 1 ml of H 2 O and freeze-dried overnight. Dried cells were resuspended in 1 ml of H 2 O and extracted by the addition of 1.4 ml of ethanol:diethyl ether:pyridine (15:5:1) for 30 min at 57°C. Debris was pelleted by centrifugation, and the supernatant was removed to a fresh tube. After drying under N 2 , lipids were resuspended in 1 ml of solvent A (chloroform:methanol: water (16:16:5 v/v/v)), and glycerophospholipids were deacylated by the addition of 1 ml of 0.2 N NaOH in methanol and incubation at 30°C for 45 min. One ml of 5% EDTA was added, and samples were neutralized by the addition of 0.2 ml of 1 N acetic acid. Lipids were then extracted with 1 ml of chloroform and dried under N 2 . Lipids were resuspended in 0.1 ml of solvent A and analyzed by thin layer chromatography on 1-mm Silica Gel G plates (Analtech, Newark, DE) treated with 2.5% EDTA (pH 7.2) and developed in chloroform:methanol:4.2 N ammonium hydroxide (9:7:2 v/v/v). Radiolabeled lipids were identified by autoradiography and quantified by scraping relevant spots followed by liquid scintillation counting. The identities of inositolphosphoryl-ceramide, MIPC, and M(IP) 2 C were confirmed by electrospray ionization mass spectroscopy (Utah State University Biotechnology Center).
Quantification of Diacylglycerol-Steady-state and pulse analyses of DAG levels were measured by labeling cells with [ 14 C]acetate (ICN Pharmaceuticals, Costa Mesa, CA). Steady-state labeling was conducted by adding 5 Ci of [ 14 C]acetate to 20-ml cultures and growing the cells for 18 -24 h at 28°C. Pulse labeling was done by adding 40 Ci of [ 14 C]acetate to 20-ml cultures, with an optical density at 600 nm between 0.7 and 0.8, for 20 min at 28°C. DAG extraction was done as described by Buttke and Pyle (20). The total lipid fractions from 14 C pulse-labeled and steady-state-labeled cells were extracted with 2 ml of methanol under reflux for 1 h, followed by the addition of 4 ml of chloroform and incubation at room temperature for 18 h. Phases were separated by the addition of 1.2 ml of 0.1 M potassium chloride. The chloroform phase was removed to a fresh tube and dried under N 2 .
Labeled lipids were suspended in 100 l of chloroform and separated by thin layer chromatography with the solvent petroleum ether:diethyl ether:acetic acid (85:15:1 v/v/v). 14 C-labeled lipids were detected by autoradiography, scraped, and quantified by liquid scintillation counting. Nonlabeled DAG and other lipid standards were detected by charring after spraying with 50% sulfuric acid.
Phospholipid Analysis-Cells were grown at 26°C for 18 h in synthetic defined minimal media supplemented with 1 mM myo-inositol in the presence of 40 Ci/ml H 3 32 PO 4 (ICN Pharmaceuticals) as performed in the work of Kearns et al. (11). Total lipids were extracted from 20 A 590 units of each cell culture as described for sphingolipid analysis without deacylation. Dried lipids were suspended in 400 l of solvent A and separated by two-dimensional thin layer chromatography as described above and visualized by autoradiography. The relevant lipid spot was scraped from plates and extracted three times with 500 l of solvent A, dried under N 2 , and processed for high performance liquid chromatography (HPLC) analysis as described below.
Phosphoinositide Analysis-Yeast cells were grown at 26°C in synthetic minimal media supplemented with 100 M myo-inositol and 5 Ci/ml myo-[ 3 H]inositol (23 Ci/mmol; NEN Life Science Products, Boston, MA) to a mid-log phase. Lipids were extracted as described under "Quantification of Sphingolipids." Deacylation was carried out by the method of Serunian et al. (21). Dried glycerophosphoinositols were subjected to anion exchange chromatography using a Whatman Partisil 5 SAX column (25 cm ϫ 4.6 mm; Whatman Inc., Clifton, NJ) on a Beckman System Gold HPLC system. The column was pre-equilibrated with 10 mM ammonium phosphate, pH 3.8. A portion of each sample (2.5 ϫ 10 6 cpm) was applied to the column, washed with 5 ml of 10 mM ammonium phosphate (pH 3.8), and eluted with a 40-ml linear gradient from 10 mM to 0.7 M ammonium phosphate (pH 3.8) at a flow rate of 1 ml/min. [ 32 P]Glycerophosphoinositol 3-phosphate and [ 32 P]glycerophosphoinositol 4-phosphate standards were generated by in vitro phosphorylation of phosphatidylinositol with ␣-[ 32 P]ATP followed by deacylation. For the production of PtdIns(3)P, an extract enriched in PtdIns 3-kinase was prepared from a 25-30% ammonium sulfate precipitate of a yeast cytosolic fraction of strain TVY614 (22) carrying a multicopy plasmid of the VPS15 gene (pJSY324.15) (23) and the VPS34 gene (pPHY52) (24). The PtdIns 4-kinase source was a crude cell extract from a yeast strain (PHY102) that lacks the VPS34 gene (24). The glycerophosphoinositol 3,5-bisphosphate peak was identified as a species that increased 16-fold by osmotic stress (1 M NaCl for 20 min; data not shown) (25).
Phosphatidylinositol 4-Phosphate Levels Are Elevated in the sac1-22 Mutant-The above finding that M(IP) 2 C levels are unchanged in the CTY165 (sec14-1 ts sac1-22) bypass strain prompted us to revisit the effects of the sac1-22 mutation on the cellular levels of inositol-containing lipids. Strains CTY1-1A (sec14-1 ts ) and CTY165 (sec14-1 ts sac1-22) were grown at permissive temperature in the presence of [ 32 P]phosphate, and the total lipids were extracted and analyzed by two-dimensional thin layer chromatography (Fig. 4). A preferential increase in the level of a phosphate-containing lipid that resembled a diphosphoinositide (26) was observed in extracts of strain CTY165 (sec14-1 ts sac1-22). To identify this lipid, it was isolated from the silica matrix of the thin layer chromatographic plates, extracted, deacylated, and subjected to anionic exchange HPLC analyses (21) (Fig. 4). Comparisons to authentic glycerophosphoinositol standards indicated that the lipid was PtdIns(4)P. To confirm this identification, lipid extracts were prepared from cultures of strains grown in the presence of myo-[ 3 H]inositol and deacylated, and the glycerophosphoinositols were identified by anionic exchange HPLC analyses (Fig.  5). A HPLC peak that co-eluted with glycerophosphoinositol derived from PtdIns(4)P was 8-fold more abundant in the CTY165 (sec14-1 ts sac1-22) samples as compared with the CTY1-1A (sec14-1 ts ) samples. The only other significant difference detected between the two strains was a 60% reduction in the glycerophosphoinositol derived from PtdIns(4,5)P 2 in strain CTY165.
DAG Production-In the current model for sac1-22 suppression of sec14-1 ts , bulk DAG levels are hypothesized to increase as a result of elevated M(IP) 2 C synthesis (11). However, the results presented above show that M(IP) 2 C levels do not increase with sac1-22 suppression. To reconcile this discrepancy, the effects of IPT1 disruption and sac1-22 suppression of sec14-1 ts on DAG production were determined. Cultures of strains CTY182 (wild type), CTY1-1A (sec14-1 ts ), CTY165 (sec14-1 ts sac1- 22), and CTY-SR (sec14-1 ts sac1-22 ⌬ipt1) were incubated with [ 14 C]acetate continually (steady-state labeling) or for a 20-min interval (pulse labeling) before harvesting the cells. Total lipid extracts were separated by thin layer chromatography, and DAG was identified by comparison to an authentic standard. With steady-state labeling, no significant differences in DAG levels were observed between the four strains (data not shown). With pulse labeling, strains CTY182 (wild type), CTY1-1A (sec14-1 ts ), and CTY165 (sec14-1 ts sac1-22) showed similar rates of net DAG production, but CTY-SR (sec14-1 ts sac1-22 ⌬ipt1) produced approximately one-third less DAG (Fig. 6). DISCUSSION The main conclusion of our work is that M(IP) 2 C synthesis, which is coupled to DAG production, is not involved in the bypass of sec14-1 ts by sac1-22. This was determined by assessing the consequences of disrupting IPT1, the gene that encodes the M(IP) 2 C synthase, and also by examining the cellular levels of M(IP) 2 C and DAG as a function of the bypass. These conclusions differ from those of Kearns et al. (11), who previously reported increases in both M(IP) 2 C and DAG in the sec14-1 ts sac1-22 bypass mutant. It is clear that in the previous report, M(IP) 2 C was misidentified. A re-evaluation in the present work of alterations in lipid composition caused by the sac1-22 bypass mutation revealed that the amount of a single phosphoinositolcontaining lipid species, PtdIns(4)P, was preferentially elevated rather than M(IP) 2 C.
Our analyses, however, did reveal an increase in the relative level of MIPC in CTY1-1A (sec14-1 ts ) at permissive temperature (Fig. 3). This could be due to product inhibition of Ipt1p by accumulated M(IP) 2 C in the Golgi compartment caused by a defect in the secretory pathway that is nonetheless partially functional at permissive temperature. Alleviation of this defect by sac1-22 would account for the lowered MIPC levels observed in strain CTY165 (sec14-1 ts sac1- 22). However, why the MIPC levels of strain CTY165 were lower than those of strain CTY182 (wild type) (Fig. 3) remains unexplained.
It is premature to assign a specific role to PtdIns(4)P in the SEC14-dependent protein secretory pathway. The observed increased amounts of this lipid (Figs. 4 and 5) are only correlated with the sec14-1 ts sac1-22 bypass. Nevertheless, the potential importance of PtdIns(4)P in the bypass is consistent with several observations that link phosphatidylinositol transfer pro-tein function and phosphoinositide production. For example, it has been demonstrated that mammalian phosphatidylinositol transfer proteins are co-factors for phosphoinositide production by PtdIns kinases that participate in signaling and membrane traffic (27). In assays for certain signaling or trafficking processes, Sec14p can substitute for mammalian phosphatidylinositol transfer proteins (28 -31). It is worth noting that Sac1p has significant similarity to a noncatalytic domain of mammalian polyphosphoinositide 5-phosphatases (7), and Sac1p may conceivably be a novel phosphatase that regulates PtdIns(4)P levels. If so, Sac1p inactivation would lead to elevated levels of PtdIns(4)P, which, in turn, may compensate for insufficient PtdIns(4)P production due to a defective Sec14p. In numerous studies of phosphoinositide roles in membrane traffic, PtdIns(4)P is typically regarded as a precursor to PtdIns(4,5)P 2 (32). However, the present finding that PtdIns(4,5)P 2 levels are lowered in the sec14-1 ts sac1-22 bypass emphasizes the potential importance of PtdIns(4)P, rather Lipids were extracted from the spot indicated by an arrow in B, deacylated, and analyzed by HPLC as described under "Experimental Procedures" (black line). Fractions (0.5 ml each) were collected from 0 to 40 ml. Counts in the fractions from 15 to 25 ml are shown. No radioactivity above the background level was detected in the rest of the fractions. The chromatogram with 32 P-labeled deacylated phosphatidylinositol 3-phosphate and PtdIns(4)P is superimposed (gray line). gPI(3)P, glycerophosphoinositol 3-phosphate; gPI(4)P, glycerophosphoinositol 4-phosphate.
than PtdIns(4,5)P 2 , in SEC14-dependent secretion. In support of the significance of PtdIns(4)P in trafficking, Matsuoka et al. (14) reported that either PtdIns(4)P or PtdIns(4,5)P 2 promoted the binding of coat proteins to liposomes in an in vitro assay for COPII-coated vesicle formation.
Roles for phospholipase D and phosphatidic acid in the bypass have been suggested recently (8,9). How the presently observed increases in PtdIns(4)P relate to phospholipase D function and phosphatidic acid production in the Golgi secretory machinery is not clear. However, it is conceivable that all of these elements are coordinately regulated in yet unknown ways to allow efficient operation of the SEC14-dependent secretory pathway.
We also examined the effect of IPT1 disruption on DAG production after both steady-state and pulse labeling of cells with [ 14 C]acetate. The pulse-labeling experiments revealed that rates of DAG production were reduced by one-third in the strain CTY-SR (sec 14-1 ts sac1-22 ⌬ipt1) compared with the CTY182 (wild type), CTY1-1A (sec14-1 ts ), or CTY165 (sec14-1 ts sac1-22) strains. The decrease in DAG production with IPT1 disruption is plausible because elimination of M(IP) 2 C synthase, which is coupled to phosphorylinositol transfer from PtdIns, would eliminate a source of cellular DAG. Steady-state and pulse labeling of DAG did not reveal an increase in DAG levels in strain CTY165 as reported previously (11). This discrepancy between the previous work of Kearns et al. (11) and the present work remains unresolved. However, the present findings are consistent with results reported by Sreenivas et al. (9), who found no difference in the DAG levels of other bypass mutants of sec14-1 ts . Furthermore, because IPT1 disruption leads to lower DAG production levels despite the retention of the sec14-1 ts bypass phenotype, our results suggest that an increase in DAG levels is not essential for suppression by sac1- 22.
Taken together, the present findings raise concerns about certain features of the proposed mechanisms of sac1-22-mediated bypass of sec14-1 ts (11). DAG production through sphingolipid metabolism and M(IP) 2 C biosynthesis does not appear to play a role in the bypass. Instead, another lipid, PtdIns(4)P, is revealed to be of potential importance to SEC14-dependent secretion. However, insight into the mechanisms of the role of PtdIns(4)P in the secretion pathway will require further investigation.