Phospholipid Transfer Protein Sec14 Is Required for Trafficking from Endosomes and Regulates Distinct trans-Golgi Export Pathways*

A protein known to regulate both lipid metabolism and vesicular transport is the phosphatidylcholine/phosphatidylinositol transfer protein Sec14 of Saccharomyces cerevisiae. Sec14 is thought to globally affect secretion from the trans-Golgi. The results from a synthetic genetic array screen for genes whose inactivation impaired growth of cells with a temperature-sensitive SEC14 allele implied Sec14 regulates transport into and out of the Golgi. This prompted us to examine the role of Sec14 in various vesicular transport pathways. We determined that Sec14 function was required for the route followed by Bgl2, whereas trafficking of other secreted proteins, including Hsp150, Cts1, Scw4, Scw10, Exg1, Cis3, and Ygp1, still occurred, indicating Sec14 regulates specific trans-Golgi export pathways. Upon diminution of Sec14 function, the v-SNARE Snc1 accumulated in endosomes and the trans-Golgi. Its accumulation in endosomes is consistent with Sec14 being required for transport from endosomes to the trans-Golgi. Sec14 was also required for trafficking of Ste3 and the lipophilic dye FM4-64 from the plasma membrane to the vacuole at the level of the endosome. The combined genetic and cell biology data are consistent with regulation of endosome trafficking being a major role for Sec14. We further determined that lipid ligand occupancy differentially regulates Sec14 functions.

A protein known to regulate both lipid metabolism and vesicular transport is the phosphatidylcholine/phosphatidylinositol transfer protein Sec14 of Saccharomyces cerevisiae. Sec14 is thought to globally affect secretion from the trans-Golgi. The results from a synthetic genetic array screen for genes whose inactivation impaired growth of cells with a temperature-sensitive SEC14 allele implied Sec14 regulates transport into and out of the Golgi. This prompted us to examine the role of Sec14 in various vesicular transport pathways. We determined that Sec14 function was required for the route followed by Bgl2, whereas trafficking of other secreted proteins, including Hsp150, Cts1, Scw4, Scw10, Exg1, Cis3, and Ygp1, still occurred, indicating Sec14 regulates specific trans-Golgi export pathways. Upon diminution of Sec14 function, the v-SNARE Snc1 accumulated in endosomes and the trans-Golgi. Its accumulation in endosomes is consistent with Sec14 being required for transport from endosomes to the trans-Golgi. Sec14 was also required for trafficking of Ste3 and the lipophilic dye FM4-64 from the plasma membrane to the vacuole at the level of the endosome. The combined genetic and cell biology data are consistent with regulation of endosome trafficking being a major role for Sec14. We further determined that lipid ligand occupancy differentially regulates Sec14 functions.
The composition of lipids within a membrane affects vesicle fission, transport, and fusion. Information on proteins that integrate lipid metabolism with vesicular transport is sparse. One protein that does so is Sec14 from Saccharomyces cerevisiae. Sec14 is an essential protein that extracts phosphatidylcholine (PC) 2 and phosphatidylinositol (PI) from membranes in vitro and regulates PC and phosphoinositide metabolism in cells (1). Decreasing Sec14 function results in a reduction in Golgi PI 4-phosphate levels, an increase in the rate of PC syn-thesis, increased turnover of PC via the PC-phospholipase D Spo14, and decreased PC turnover by the PC-phospholipase B Nte1 (Fig. 1A) (2)(3)(4)(5). The mechanisms by which Sec14 regulates PI 4-phosphate and PC metabolism are not known; however, this appears to be a major function in its regulation of cell biology as inactivation of numerous genes that mediate PC and phosphoinositide metabolism enhances or impairs growth of cells with defective Sec14 function (2, 4 -6, 8 -11).
Sec14 is known to be required for the transport of vesicles from the trans-Golgi to the plasma membrane and vacuole (equivalent to the lysosome in mammalian cells) as reduced Sec14 function results in a rapid decrease in Golgi-derived transport of invertase and acid phosphatase out of the cell and carboxypeptidase Y to the vacuole (2-6, 8 -10, 12-15). Regulation of PC and phosphoinositide metabolism by Sec14 is hypothesized to provide the appropriate membrane environment for fission of vesicles from the trans-Golgi, although the precise step(s) where Sec14 regulates vesicular transport has yet to be determined.
The human genome contains at least 29 Sec14 domain-containing genes that code for more than 45 proteins (16). Mutations in several human Sec14 domain-containing proteins result in the onset of human diseases, including neurodegeneration, blindness, and cancer (17)(18)(19)(20)(21)(22). Sec14 domains from higher eukaryotes are often embedded as part of a larger protein, many of which are guanine exchange factors (GEFs) and GTPase-activating proteins (GAPs), indicating that regulation of small G protein activity is a unifying theme for several members of the Sec14 superfamily. Human Sec14 domains bind hydrophobic ligands, including phospholipids (16).
In this study, we used synthetic genetic array (SGA) analysis to determine more precisely how S. cerevisiae Sec14 regulates vesicular transport. Genetic interactions were observed between a SEC14 allele with reduced function and GTPases that regulate vesicular transport in and out of the Golgi via endosomes. This is the first instance of Sec14 regulating import into the Golgi.
A main route for fusion of endosome-derived vesicles with the trans-Golgi is facilitated by the t-SNARE Tlg2 (23). Tlg2, along with Vps45 (a protein of the Sec1/Munc18 family), tethers endosome-derived vesicles to the trans-Golgi to facilitate membrane fusion (23) (Fig. 1B). The tethering complex that directly interacts with Tlg2 is the Golgi-associated retrograde protein (GARP) complex. The GARP complex is composed of four subunits Vps51-54 (23). The GARP complex directly interacts with Ypt6 (a Rab GTPase) and Arl1 (an Arf-like GTPase) that together comprise the two main routes for endosome fusion with the trans-Golgi. Ypt6 is activated by a heterodimeric GEF comprised of Ric1 and Rgp1. The GEF for Arl1 is unclear; however, the localization of Arl1 to the Golgi is dependent on Arl3 and Sys1 (24,25). Arl3 is also an Arf-like GTPase, whereas Sys1 is an integral membrane protein that targets Arl3 to the Golgi.
A second Golgi-associated tethering complex is the transport protein particle (TRAPP). TRAPP I regulates early Golgi events in part by acting as a GEF for the Rab Ypt1. TRAPP II is specific for transport in and out of the trans-Golgi and is a GEF for the Rab pair Ypt31/32 ( Fig. 1B) (26). TRAPP I and II share seven subunits (Bet3, Bet5, Trs20, Trs23, Trs31, Trs33, and Trs35) with three subunits (Trs65, Trs120, and Trs130) added to form TRAPP II.
Sec14 has previously been implied to globally regulate vesicular transport out of the Golgi. Our genetic screen dramatically expanded the network of genes required for vesicular transport when Sec14 function is impaired, including those comprising GARP and TRAPP complexes. The results from the genetic screen prompted us to examine the role of Sec14 in multiple vesicular transport routes. We report that Sec14 regulates a subset of vesicular transport pathways out of the trans-Golgi. The genetic analysis, coupled with assessments of vesicular transport pathways, go on to indicate that a primary function of Sec14 is regulation of vesicular transport through endosomes. We also demonstrate that ligand binding by Sec14 differentially regulates function.

MATERIALS AND METHODS
Synthetic Genetic Array Screen-SGA analysis was performed essentially as described (27,28) with the below modifications. CMY503 (contains a temperature-sensitive SEC14 allele, sec14 ts ) was mated with 4,800 S. cerevisiae single gene deletion strains at 25°C; diploids were selected, and cells were sporulated for 5 days at 25°C. To ensure that the haploid cells obtained were from mated diploids, cells were selected for histidine prototrophy followed by growth on medium containing G418 and nourseothricin. The resulting haploids were incubated at 25 or 35°C. Three independent screens were performed. Mutants whose inactivation resulted in decreased growth when in combination with the sec14 ts allele in at least two of three screens were subjected to random spore analysis to determine whether the genetic interactions observed were true.
Yeast Strain Construction-The sec14 ts allele linked to the nourseothricin drug resistance cassette was constructed as follows (mutated sites in primer sequences are underlined). Plasmid-borne SEC14 was subjected to site-directed mutagenesis to convert Gly-266 to Asp using primers 5Ј-CTTACCA-GTCAAATTTGGCGATAAGTCTGAAGTTGATGATCA-3Ј and 5Ј-GATTCATCAACTTCAGACTTATCGCCAAATTT-GACTGGTAAG-3Ј. This mutation confers temperature sensitivity to Sec14. To insert the natMX4 cassette next to the sec14 ts gene, an SphI site was generated 300 bp downstream of the SEC14 stop codon by site-directed mutagenesis with the primers 5Ј-GGCAGTATATACATTTATCTTATAGAATA-CGCATGCACGTATTAAACAATAAAAAGAACATAAG-3Ј and 5Ј-CTTATGTTCTTTTTATTGTTTAATACGTGCAT-GCGTATTCTATAAGATAAATGTATATACTGCC-3Ј. The natMX4 cassette was amplified with SphI sites at each end using primers 5Ј-CTGAGAGCATGCACATGGAGGCCCAG-AATACCC-3Ј and 5Ј-CTGAGAGCATGCCAGTATAGCGA-CCAGCATTCAC-3Ј. An SphI site in the natMX4 DNA sequence was removed by site-directed mutagenesis (without changing the amino acid sequence) using primers 5Ј-CGCTC-FIGURE 1. Sec14 regulation of phospholipid metabolism. A, regulation of phospholipid metabolism by Sec14 is thought to be lipid ligand-dependent. Sec14 bound to PC regulates PC metabolism, whereas Sec14 bound to PI controls phosphoinositide metabolism. PC-bound Sec14 is thought to inhibit PC synthesis at the rate-limiting step of Pct1 while at the same time inhibiting PC turnover by the phospholipase D Spo14 (to decrease the production of choline to prevent supply of PC precursor). At the same time Sec14-PC shunts PC turnover toward the phospholipase B Nte1, away from supplying choline for PC synthesis. In this way, Sec14 acts as a homeostat for regulation of PC levels. Sec14 bound to PI maintains PI-4P levels, although whether this is by activating the PI 4-kinase Pik1 is not known. The solid arrows indicate metabolic pathways; the dotted arrows activation, and the dotted lines inhibition. GPC, glycerophosphocholine. B, pathways in and out of the trans-Golgi. A subset of vesicles destined to fuse with the trans-Golgi do so via the GARP complex (contains Vps51, Vps52, Vps53, and Vps54). The GARP complex binds Ypt6, a Rab regulated by the GEF pair Ric1/Rgp1. The GARP complex also directly interacts with the Arf-like GTPase Arl1 that is anchored to the trans-Golgi via Arl3 and Sys1. Fusion to the trans-Golgi is mediated in part by the t-SNARE Tlg2 and its accessory protein Vps45. A second Golgi-associated complex, the transport protein particle II (TRAPP II) complex (contains the essential subunits Bet3, Bet5, Trs20, Trs23, Trs31, Trs120, and Trs130 and the nonessential subunits Trs33, Trs65, and Trs85), acts as a GEF for the Rab pair Ypt31/32 and mediates transport both in and out of the trans-Golgi.
TACATGAGTATGCCCTGCCCCTA-3Ј and 5Ј-TAGGGGC-AGGGCATACTCATGTAGAGCG-3Ј, and this DNA fragment was subcloned into the SphI site downstream of the sec14 ts open reading frame. The sec14 ts :natMX4 disruption cassette was PCR-amplified from the resulting vector using primers 5Ј-CCTTTCTTGGATCCAGTCACTG-3Ј and 5Ј-GGCA-TGGACTTGATATCCTAG-3Ј to give a fragment containing both the sec14 ts mutation and the natMX4 cassette along with 5Ј and 3Ј SEC14-flanking DNA. This DNA fragment was transformed into the Y2454 yeast strain to replace the SEC14 gene with the sec14 ts allele linked to the natMX4 cassette to generate strain CMY503 (MAT␣ mfa1⌬::MFA1pr-HIS3 can1⌬0 ura3⌬0 leu2⌬0 his3⌬1 lys2⌬0 sec14 ts :natMX4). The sec14 ts :natMX4 cassette was transformed into the indicated strains for the construction of isogenic sets of sec14 ts containing yeast ( Table 1). Strain genotypes were confirmed by genomic PCR and/or genomic Southern blot using the digoxygenin High Prime DNA Labeling and Detection Kit II (Roche Applied Science).
Monitoring Bulk Protein Secretion and Mass Spectrometry Identification of Secreted Proteins-Cells were grown to midlog phase at 25°C, and an equal cell number was shifted to 37°C for 10 min and incubated with [ 35 S]methionine/cysteine for 10 min and then with 10 mM unlabeled methionine/cysteine for 15 or 30 min. Proteins in the medium were precipitated with 10% trichloroacetic acid and separated by SDS-PAGE, and the gel was exposed to x-ray film. Similarly, equal numbers of log phase cells were shifted to 37°C for 1 h in fresh YPD medium and proteins precipitated with 10% trichloroacetic acid, separated by SDS-PAGE, and stained with Gelcode (Thermo Scientific). Proteins contained in gel slices were reduced with dithiothreitol, carboxamidomethylated with iodoacetamide, and digested with trypsin. Peptides were extracted with 70% acetonitrile, 1% formic acid. The extraction solvent was removed under vacuum, and the tryptic peptides were resuspended in 5% methanol, 0.5% formic acid.
Liquid chromatography MS/MS was performed using an Ultimate pump and Famos auto-sampler (LC Packings, Amsterdam, Netherlands) interfaced to the nanoflow electrospray ionization source of a hybrid triple quadrupole linear ion trap mass spectrometer (QTrap, Applied Biosystems). Samples were injected onto a capillary column (0.10 ϫ 150 mm Chromolith C18, monolithic, Merck) at a flow rate of 1.2 l per min. Solvent A consisted of 98% water, 2% acetonitrile, 0.1% formic acid, and Solvent B consisted of 2% water, 98% acetonitrile, 0.1% formic acid, and the linear gradient was as follows: 5% solvent B to 35% solvent B over 35 min then 90% B for 6 min before re-equilibration at 5% solvent B. The sample was sprayed through distal coated fused silica emitter tips, 75-m inner diameter with 15-m inner diameter tip (New Objectives PicoTip). The capillary voltage was 2.10 kV with a declustering potential of 60 V, and the curtain gas was set to 15 (arbitrary units). Spectra were acquired using the information-dependent acquisition mode. The two most intense ions from the survey scan (375-1200 m/z) were selected for tandem MS, and the collision energy was set based on the mass of the precursor as determined by the m/z and charge. The raw MS/MS data were searched against NCBI yeast entries and against all SwissProt entries using the MASCOT algorithm (Matrix Science). Search parameters were peptide mass and fragment mass tolerance 0.8 and 0.5 Da, respectively, with one missed cleavage allowed. Oxidized methionines and carboxamidomethylated cysteines were chosen as variable modifications.
Invertase and Bgl2 Secretion Assays-The invertase secretion assay was performed as described (29) with modifications (13). Bgl2 secretion was determined as described (30). Mid-log phase cells were incubated at 25°C or shifted to 37°C for 15-60 min. Cells were then harvested by centrifugation at 1000 ϫ g for 5 min. Cell pellets were resuspended in 10 mM NaN 3 , 10 mM KF solution, incubated on ice for 10 min, and then transferred to microcentrifuge tubes. Cells were centrifuged at 10,000 ϫ g for 1 min, and pellets were resuspended in fresh pre-spheroplasting buffer (100 mM Tris-H 2 SO 4 , pH 9.4; 50 mM ␤-mercaptoethanol; 10 mM NaN 3 ; 10 mM KF), incubated on ice for 15 min, centrifuged as before, washed with 0.5 ml of spheroplast buffer (50 mM KH 2 PO 4 -KOH, pH 7; 1.4 M sorbitol; 10 mM NaN 3 ), and pelleted. Cells were resuspended in spheroplast buffer containing 167 g/ml Zymolyase 100T and incubated for 30 min at 25°C. Spheroplasts were then pelleted at 5,000 ϫ g for 10 min and resuspended in 2ϫ SDS-PAGE sample buffer. Proteins were separated using a 10% SDS-polyacrylamide gel, and Bgl2 was detected by Western blot using a rabbit polyclonal antibody against Bgl2. Bgl2 antibodies were the kind gifts of Randy Schekman (University of California, Berkeley) and Wei Guo (University of Pennsylvania).
Ste3 Internalization-MAT␣ cells were grown to mid-log phase at 25°C and shifted to 37°C for 2 h. Proteins were extracted, separated by SDS-PAGE, transferred to polyvinylidene difluoride membrane, and probed with antibodies for Ste3 and Pgk1 (Molecular Probes). Ste3 antibodies were the kind gift of George Sprague (University of Oregon).
Microscopy-Live cells were observed using a Zeiss Axiovert 200 M microscope fitted with a plan-neofluor 100ϫ oil immersion lens. Images were captured using a Zeiss Axio Cam HR using Axiovision 4.5 software. Cells were visualized using differential interference contrast (DIC), green fluorescent protein, or rhodamine filters as required. GFP-Snc1 subcloned into the plasmid pRS416 (pMJL1) was the kind gift of Hugh Pelham (MRC Laboratory of Molecular Biology, Cambridge, UK). The plasmid expressing the Vps27-GFP fusion was from Christopher Stefan (Cornell University, Ithaca, NY). The yeast strains expressing red fluorescent protein-tagged organelle marker proteins were from Erin O'Shea (University of California, San Francisco).

SGA Analysis of sec14 ts Cells Uncovers Novel Genetic
Interactions-S. cerevisiae contains 6607 genes of which ϳ5000 are not essential. A haploid query strain containing a temperature-sensitive sec14 ts allele was mated to ϳ4800 viable single gene deletion haploid yeast strains of the opposite mating type using SGA technology (28). The resulting diploids were isolated and forced to undergo meiosis, and haploid cells were selected that contained the sec14 ts allele in combination with each single gene deletion. The resulting strains, obtained from three sepa-rate screens, were incubated at permissive (25°C) and semipermissive (35°C) temperatures for growth of cells containing the sec14 ts allele alone. We considered only strains with the sec14 ts gene in combination with each single gene deletion able to grow at 25°C, but not 35°C, to eliminate genes unable to grow because of defective mating, sporulation, or restoration of vegetative growth from spores.
Forty genes were isolated from the SGA screen whose inactivation aggravated growth of sec14 ts -containing cells ( Fig. 2A and Table 2). Inactivation of SPO14 (PC phospholipase D) had been previously determined to diminish growth of sec14 ts cells (1,5,9), and it was isolated from our screen. Of the 40 genes isolated, the largest group contained genes known to regulate Golgi-dependent vesicular transport processes consistent with Sec14 acting as an essential regulator of this process ( Fig. 2A). As genetic interactions can be representative of direct regulation of a protein/process, or may indicate processes whose functions are linked but by several degrees of separation, we focused on the genes isolated from the SGA screen with known roles in vesicular transport.
Secretion Pathways Are Differentially Affected in Cells Defective in Sec14 Function-Sec14 regulates fission of vesicles from the trans-Golgi. A hallmark of Sec14 deficiency is reduced invertase secretion from the cell. Invertase follows a pathway from the trans-Golgi via endosomes on its route to the plasma membrane (35,36). We monitored invertase secretion in sec14 ts cells and compared it with sec14 ts cells containing inac-tivated genes isolated from our SGA screen. In sec14 ts cells invertase secretion was reduced to 66% at 35°C (Fig. 2B) (compared with ϳ30% at 37°C, which is the nonpermissive temperature for the sec14 ts allele; data not shown). In sec14 ts cells that also contained inactivated SPO14, YPT31, TRS33, GYP1, or TLG2 invertase secretion was reduced to 22-48% of wild type at 35°C (Fig. 2B). All strains secreted invertase similar to wild type at 25°C and in addition all single gene deletions secreted invertase at wild type levels at 35 and 37°C (data not shown). Therefore, inactivation of SPO14, YPT31, TRS33, GYP1, or TLG2 decreases Sec14-dependent vesicular transport of invertase from the trans-Golgi.
Two types of secretory vesicles have been isolated from cells based on differences in density (35,36). The endosome-dependent route fractionates with higher density vesicles that contain invertase and acid phosphatase. These are the only two proteins that have been shown to be inefficiently secreted in sec14 ts cells.
A second route to the plasma membrane is found in lower density vesicles containing Bgl2 and Pma1 and represent an endosome-independent route from the trans-Golgi to the plasma membrane. To assess if a deficiency in Sec14 function also affects secretion through the endosome-independent route, we monitored Bgl2 secretion. Bgl2 accumulated in sec14 ts cells compared with wild type after incubation at the non-permissive temperature for the sec14 ts allele. We observed a build up of Bgl2 in cells with defective Sec14 function as early as 15 min after temperature shift, and this was obvious within 1 h (Fig. 3A). The osh⌬ osh4 ts strain is known to accumulate Bgl2 and served as a positive control (30). Bgl2 observed outside the cell in all strains is standard for this assay as Bgl2 is present in the periplasmic space prior to onset of the secretory defect.
The high and low density vesicles that contain invertase/acid phosphatase and Bgl2/Pma1, respectively, were isolated from cells with defects in fusion of Golgi-derived vesicles with the plasma membrane (36). These experiments did not rule out other routes to the plasma membrane from the trans-Golgi. To determine the extent and specificity of the protein secretion defect in sec14 ts cells, wild type and sec14 ts cells (from two different genetic backgrounds) were grown at 37°C for 10 min, incubated with [ 35 S]methionine/cysteine for 10 min to radiolabel cellular protein, and chased with 10 mM unlabeled methionine/cysteine for various times, and radiolabeled proteins secreted into the medium were separated by SDS-PAGE. Wild . Genes whose inactivation aggravates growth of sec14 ts cells. A, pie chart of the functions of the 40 genes identified by the SGA screen whose inactivation decreased growth of sec14 ts cells. A subset of these genes (SPO14, YPT31, TRS33, TLG2, GYP1, and VPS1) had been identified previously as genetic interactors in cells with decreased Sec14 function (2). B, isogenic mutant strains with the indicated genotypes were grown to mid-log phase in synthetic complete medium at 25°C; identical numbers of cells were serial diluted, spotted onto agar plates, and grown for 3 days at the indicated temperatures. Invertase secretion is the ratio of external invertase enzyme activity to total (internal ϩ external) invertase activity. Invertase data are the mean of three individual experiments performed in triplicate. Standard errors of the mean are indicated in the figure.

Sec14-mediated Vesicular Transport
type and sec14 ts cells secreted a similar amount of protein. The pattern of proteins secreted was similar but not completely identical (Fig. 3B). As a control we monitored the amount of radiolabeled protein secreted from sec18 ts cells that possess a well characterized defect in secretion (Sec18 encodes the ATPase-N-ethylmaleimide-sensitive fusion protein that primarily regulates endoplasmic reticulum to Golgi transport). The sec18 ts cells secreted significantly less protein.
To determine whether the radiolabeled bands could be due to protein breakdown, or whether they represented a series of proteins, proteins secreted into the medium were acid-precip-itated and separated by SDS-PAGE, and gels were stained with Coomassie Blue and silver nitrate. Like the secretion of [ 35 S]methionine/cysteine-labeled proteins, the banding pattern and amount were similar between wild type and sec14 ts cells (data not shown). Identity was determined by mass spectrometry and included Hsp150, Cts1, Scw4, Scw10, Exg1, Cis3, and Ygp1, proteins known to be secreted by S. cerevisiae.
Decreasing Sec14 function does not appear to diminish all vesicular transport routes from the trans-Golgi, but it affects specific vesicular transport pathways emanating from this organelle.

GFP-Snc1 Trafficking Is Defective in Cells with Diminished Sec14
Function-Invertase transits through early endosomes on its way to the cell surface (35,36). A possible explanation for the defect in delivery of invertase to the cell surface is that Sec14 is required for trafficking via endosomes. Consistent with this notion, genes isolated from the SGA screen (YPT31, TRS33, TLG2, and VPS1) regulate vesicular transport pathways through endosomes ( Table 2). The endosome is a central organelle required for transit of a subset of proteins from the trans-Golgi to the plasma membrane, as well as trafficking of proteins from the plasma membrane back to the Golgi as well as the vacuole. To further assess if Sec14 regulated endosome requiring pathways we first monitored trafficking of the v-SNARE Snc1.
Snc1, the S. cerevisiae homologue of synaptobrevin, continually cycles between the plasma membrane and trans-Golgi. Snc1 is found in both low and high density Golgi-derived vesicles during transit to the plasma membrane (37), and thus the SNARE machinery is thought to be shared among vesicular transport pathways emanating from the trans-Golgi. Subsequently, Snc1 is transported from the plasma membrane back to the trans-Golgi via endosomes (36,38). At steady state Snc1 is primarily found at the plasma membrane (34,38), and consistent with prior studies GFP-Snc1 was primarily localized at the plasma membrane as well as a few punctate structures in the cytoplasm (endosomes and trans-Golgi) in wild type cells (Fig.   4A). A similar localization was observed in sec14 ts cells grown at the permissive temperature for function of this allele (25°C); however, when sec14 ts cells were shifted to the non-permissive temperature (37°C), GFP-Snc1 rapidly localized to large intracellular punctate spots. Localization of GFP-Snc1 with organellar marker proteins revealed that in the sec14 ts cells shifted to 37°C for 30 min GFP-Snc1 largely colocalized with the trans- FIGURE 3. Cells with reduced Sec14 function do not have a major defect in bulk protein secretion. A, mid-log phase cells were incubated at 25°C or shifted to 37°C for 1 h. Cells were harvested and resuspended in 10 mM NaN 3 , 10 mM KF and incubated on ice. Cells were washed and resuspended in buffer containing 1.4 M sorbitol and zymolyase. Resulting spheroplasts were pelleted and proteins separated by SDS-PAGE. Bgl2 was detected by Western blot. B, wild type, sec14 ts , and sec18 ts cells (strains used were BY4741 (wild type), CMY505 (sec14 ts ), X2180 (wild type), SF292-1A (sec14 ts ), and SF282-1D (sec18 ts )) were grown to mid-log phase at 25°C, shifted to 37°C for 10 min, incubated with [ 35 S]methionine/cysteine for 10 min, and then with 10 mM unlabeled methionine/cysteine for 15 and 30 min at 37°C. Proteins in the medium were separated by SDS-PAGE, and the gel was exposed to x-ray film. FIGURE 4. Sec14 regulates Snc1 cycling. A, wild type and sec14 ts cells expressing GFP-Snc1 were grown to early log phase at 25°C; a portion was shifted to 37°C, and live cells were imaged using DIC and fluorescence microscopy at the indicated time points. B, wild type SEC14 gene was replaced with the sec14 ts allele in yeast strains where RFP was fused to the organellar markers Chc1 (trans-Golgi) and Snf7 (endosomes). These strains were transformed with the plasmid expressing GFP-Snc1, and colocalization of GFP-Snc1 was determined in sec14 ts cells after cells were shifted to 37°C for 30 min. C, wild type and sec14 ts cells (strains BY4741 and CMY505) were shifted to 37°C for 30 min resulting in GFP-Snc1 localization to the trans-Golgi in sec14 ts cells. Lat-A was added for 10 min to prevent GFP-Snc1 cycling back into the cell from the plasma membrane, and cells were grown for a further 15-30 min at 37°C. Live cells were visualized using DIC and fluorescence microscopy. The number of cells from random fields (four fields from two separate experiments totaling at least 100 cells) was visually assessed for GFP-Snc1 at the plasma membrane using a double-blinded protocol.
Accumulation of GFP-Snc1 at the trans-Golgi could be due to either reduced fusion to or reduced transit from this organelle. To assess whether the GFP-Snc1 accumulating in the trans-Golgi upon reduction of function of Sec14 can exit this organelle, we treated cells with latrunculin-A (Lat-A). Lat-A blocks actin polymerization and blocks endocytosis from the plasma membrane to endosomes in S. cerevisiae. These conditions allowed us to monitor trans-Golgi to plasma membrane transport in the absence of recycling of GFP-Snc1 out of the plasma membrane. Cells were grown at 37°C for 30 min to localize GFP-Snc1 to the trans-Golgi in sec14 ts cells, and Lat-A was added to inhibit endocytosis, and cells were grown for a further 15-30 min at 37°C to determine whether GFP-Snc1 could exit the trans-Golgi and traffic to the plasma membrane. As expected, wild type cells exhibit only plasma membrane GFP-Snc1 after treatment with Lat-A. A significant number of the Lat-A-treated sec14 ts cells (58 -67%) were able to transport GFP-Snc1 from the trans-Golgi to the plasma membrane (Fig. 4C). This is consistent with a Sec14independent vesicular transport route from the trans-Golgi to the plasma membrane.
Endosome-dependent Routes from the Plasma Membrane Are Defective in Cells with Reduced Sec14 Function-Endosomes are also required for transport of vesicles from the plasma membrane to the vacuole. Ste3 is a mating factor receptor that, in the absence of mating factor, is transported from the plasma membrane via endosomes to the vacuole for degradation (39). Ste3 was observed in wild type cells at 25°C with reduced levels at 37°C (Fig. 5A). At increased temperature the rate of endocytosis is increased (40) resulting in the lower steady-state levels of Ste3 observed in wild type cells at 37°C. In cells with defective Sec14, Ste3 was stabilized at 37°C. Sec14 functions in the transport of Ste3 from the plasma membrane to the vacuole.
We also monitored internalization of the lipophilic dye FM4-64 as a separate assay to determine whether there were defects in endosome to vacuole trafficking upon reduction of function of Sec14. FM4-64 incorporates into the plasma membrane and traffics to the vacuole via early and late endosomes (40). In wild type cells FM4-64 was transported to the vacuole in a matter of minutes at 37°C, whereas in sec14 ts cells shifted to 37°C, FM4-64 accumulated in punctate spots in the cytoplasm and was unable to transit out of these punctate regions (Fig. 5B). These punctate structures colocalized with the endosomal marker Vps27-GFP (Fig. 5C). The inhibition of FM4-64 trafficking and stabilization of Ste3 upon inactivation of Sec14 is consistent with Sec14 regulating plasma membrane-derived endosome trafficking pathways.
Genetic Interactions of sec14 ts with Genes That Regulate Endosome Trafficking-Our data are consistent with Sec14 regulating endosome function, and genes isolated from our SGA screen participate in this process. A main strength of the SGA strategy is comprehensiveness; however, SGA screens can miss genetic interactions depending on several factors, including the allele used in the screen, the nature of the screen, the fitness of each single gene deletion strain, and the effect of the gene(s) on mating, sporulation, and the transition from spore to growing cell (41). Therefore, we tested all nonessential genes involved in regulation of tethering and fusion of endosome-derived vesicles at the trans-Golgi for aggravating effects on growth of sec14 ts cells (see Fig. 1B).
The wild type SEC14 gene was replaced with the sec14 ts allele in an isogenic set of strains containing inactivated genes that regulate GARP-mediated fusion from endosome to the trans-Golgi, including VPS51, VPS52, VPS53, VPS54, VPS45, YPT6, RIC1, RGP1, ARL1, ARL3, or SYS1. Inactivation of any of these genes in sec14 ts cells decreased growth at 35 and 33°C compared with cells containing the sec14 ts allele alone, except for cells containing an inactivated RGP1 gene that was inviable at any temperature (Fig. 6).
Inactivation of YPT31 was found by our SGA screen to aggravate growth of sec14 ts cells. Ypt31/Ypt32 and the TRAPP II complex regulate both export from the trans-Golgi to the plasma membrane and endosome trafficking to the trans-Golgi (26,31,32). Inactivation of any of the nonessential subunits of the TRAPP complexes, including FIGURE 5. Sec14 is required for trafficking of endosomes to the vacuole. A, wild type and sec14 ts MAT ␣ cells (BY4742 and CMY506) were grown to mid-log phase at 25°C and shifted to 37°C for 2 h. Proteins were extracted, separated by SDS-PAGE, transferred to polyvinylidene difluoride membrane, and probed with antibodies for Ste3 and Pgk1. B, wild type (BY4741) and sec14 ts (CMY505) cells were grown to mid-logarithmic phase at 25°C, shifted to 37°C for 15 min, labeled for 2 min with 40 M FM4-64, and incubated post-label in fresh pre-warmed 37°C medium for the indicated times. Live cells were visualized using DIC and fluorescence microscopy. C, FM4-64 was colocalized with the endosome marker Vps27-GFP at various time points after the shift to 37°C. MARCH 13, 2009 • VOLUME 284 • NUMBER 11

JOURNAL OF BIOLOGICAL CHEMISTRY 7371
TRS65 which is specific to the TRAPP II complex, aggravated growth of sec14 ts cells (Fig. 6).
Inactivation of YPT32 did not aggravate growth of sec14 ts cells. Ypt32 is found in cells at lower levels than Ypt31 (42), and thus inactivation of the YPT32 gene may have its phenotype masked by the Ypt31 still present in the cell. The converse would not be true as the low level of Ypt32 could not buffer the effect of inactivation of the YPT31 gene.
For all of the genes tested, growth of isogenic strains containing inactivating mutations for each gene in the presence of the wild type SEC14 gene was determined at 35, 33, and 25°C. All grew at rates similar to wild type except strains with inactivated TRS85, YPT6, or RIC1 genes that displayed moderately reduced growth at 35°C. However, the combination of the sec14 ts allele along with inactivation of TRS85, YPT6, or RIC1 resulted in much more severe growth defects than inactivation of any one gene alone (data not shown). Combined, the cell biology and genetic evidence clearly shows that Sec14 regulates endosome trafficking from the plasma membrane to both the trans-Golgi and the vacuole.
Specific Lipid Ligands Regulate Discrete Vesicular Transport Processes-Sec14 extracts PC and PI from membranes, and this activity is essential for its function in cells (4,10,43). A mutant version of Sec14 (Sec14 K66A,K239A known as Sec14-PC) that transfers PC with much higher affinity than PI in vitro was found to restore growth and secretion to cells lacking Sec14 function (7,43). Yeast contain five other Sec14 homologues (Sfh1-5) capable of PI (but not PC) transfer, and overexpression of two of these, Sfh2 and Sfh4, restored growth and secretion to cells lacking Sec14 function.
To facilitate analysis of roles for PC versus PI binding by Sec14, we transformed plasmids expressing wild type Sec14, the Sec14-PC allele, or Sfh4 into the strains containing the sec14 ts allele in combination with every gene deletion determined in this study to diminish Sec14-dependent growth. If there are separate functions for PC and PI binding by Sec14, then a subset may be rescued by only one or more of wild type Sec14, Sec14-PC and Sfh4.
Growth of sec14 ts cells lacking VPS51-53, three of the four nonessential components of the endosome to Golgi tethering complex GARP, required Sec14 or Sfh4, whereas Sec14-PC was insufficient (Fig. 7). Growth of cells inactivated for the fourth member of the GARP complex, VPS54, was only restored by wild type Sec14. Inactivation of the GARP complex as a whole requires Sec14 that can exchange PC and PI in cis within the same Sec14 molecule.
Endosome to Golgi fusion by the GARP complex is facilitated by the Rab Ypt6, and sec14 ts cells lacking YPT6 were similar to those lacking VPS54 in that growth was supported only by wild type Sec14 but not by Sec14-PC or Sfh4 (Fig. 7). Ypt6 is activated by the GEF Ric1, and growth of sec14 ts cells with inactivated RIC1 was supported by Sfh4 but not by wild type Sec14 or Sec14-PC. A similar sec14 ts rescue phenotype associated with loss of RIC1 was observed for loss of function of VPS45. Vps45 is required for the t-SNARE Tlg2 to tether the GARP complex to the Golgi. The sec14 ts cells with inactivated TLG2 could grow normally when expressing wild type Sec14 or Sfh4, with some growth observed if Sec14-PC was expressed.
The second complex capable of activating GARP for fusion of endosomes with the Golgi is mediated by Arl1, Arl3, and Sys1. Growth was restored upon inactivation of ARL3 or SYS1 in sec14 ts cells by Sec14, Sec14-PC, and Sfh4 (Fig. 7). ARL1 was unique in that its inactivation in sec14 ts cells was not restored by wild type Sec14, but it was restored by either Sec14-PC or Sfh4. The two pathways that mediate GARP complex fusion of endosomes with the Golgi have specific lipid ligand occupancy requirements by Sec14. Growth of sec14 ts cells lacking YPT31 or TRS33 was restored by Sec14, Sec14-PC, or Sfh4, whereas growth of sec14 ts cells with inactivated TRS85 required wild type Sec14 or Sec14-PC (although growth was restored to a lesser extent).
Growth of sec14 ts cells lacking the GAP Gyp1 (has specificity for several small G proteins in vitro, and the essential Rab Ypt1 in vivo; with Ypt1 activity activated by the TRAPP complex) was restored by wild type Sec14, Sfh4, and to a lesser extent Sec14-PC. Processes influenced by Sec14 function are differentially affected depending on Sec14 lipid ligand occupancy.

DISCUSSION
Sec14 Regulates Specific trans-Golgi Export Pathways-Sec14 was known to be required for secretion through the endosome-FIGURE 6. Tethering complex defects aggravate Sec14 function. Isogenic mutant strains with the indicated genotypes were grown to mid-log phase in synthetic complete medium at 25°C, and identical numbers of cells were serial diluted, spotted onto agar plates, and grown for 3 days at the indicated temperatures.

Sec14-mediated Vesicular Transport
dependent route followed by invertase and acid phosphatase. In this study we demonstrate that Sec14 also regulates the endosome-independent route from the trans-Golgi to the plasma membrane used by Bgl2. Moreover, not all secretion was inhibited upon diminution of Sec14 function as many proteins were still secreted at a similar level to wild type, including Hsp150, Cts1, Scw4, Scw10, Exg1, Cis3, and Ygp1. This implies that there is at least one more route from the trans-Golgi to the plasma membrane, and it is independent of Sec14 (Fig. 8).
The v-SNARE Snc1 traffics from the trans-Golgi to the plasma membrane via both the endosome-dependent and -independent routes (35,36). Decreasing Sec14 function resulted in ϳ50% of Snc1 still being able to traffic from the trans-Golgi to the plasma membrane, implying Snc1 may also transit to the plasma membrane by the route used by Hsp150, Cts1, Scw4, Scw10, Exg1, Cis3, and Ygp1. This is consistent with the idea that the SNARE machinery required for secretion is shared by the trans-Golgi to plasma membrane routes.
Regulation of Endosome Trafficking by Sec14-We also determined in this study that Sec14 was required for trafficking from the plasma membrane to the vacuole via endosomes. FM4-64 internalization was inhibited at the endosome stage of transport, and consistent with this Ste3 was stabilized in cells with diminished Sec14 function. Indeed, trafficking of FM4-64 was dramatically reduced within 5 min of reduction of Sec14 function. These findings imply a role for Sec14 is regulation of endosome trafficking from the plasma membrane to the vacuole.
We observed that inactivation of Sec14 resulted in the accumulation of a portion of GFP-Snc1 in endosomes implying a defect in endosome to trans-Golgi transport. We also present substantive genetic evidence pointing to an important role for Sec14 in this pathway. Cells with defective Sec14 function were further impaired for growth by inactivation of nonessential components of the GARP or TRAPP II complexes and their regulators, including Rabs, Arf-like proteins, and their GEFs. In total, this study has revealed that endosome trafficking is likely a major role for Sec14 (Fig. 8).
Sec14 Affects Phospholipid-mediated Regulation of Vesicular Trafficking-Sec14 binds PC and PI and affects the metabolism of both these ligands, although precisely how is not known (1-6, 8, 9, 11, 44, 45). It is believed that occupancy of Sec14 by PC regulates PC synthesis and turnover, whereas occupancy by PI regulates PI-4P levels (2,4,16,45,46). Two separate SGA screens using different temperature-sensitive alleles of the Golgi PI 4-kinase PIK1 determined that inactivation of many of the genes in the current study (as aggravating growth of sec14 ts cells) exacerbated Pik1-dependent vesicular trafficking from the Golgi, including YPT31 (but not YPT32), TRS33, TRS65, ARL1, SYS1, VPS51, VPS54, YPT6, RIC1, and VPS1 (47,48). As inactivation of a similar set of genes was found to inhibit growth of pik1 ts and sec14 ts cells (this study), and the fact that Golgi PI-4P levels are reduced in cells with diminished Pik1 and Sec14 function (2,11,44), it is likely that inactivation of these genes inhibits growth when Golgi PI-4P levels are reduced. Indeed, FIGURE 7. Regulation of specific vesicular transport processes based on Sec14 lipid ligand occupancy. A, isogenic mutant strains with the indicated genotypes transformed with low copy plasmids expressing Sec14, Sec14-PC, or a high copy plasmid expressing Sfh4 were grown to mid-log phase in synthetic complete medium at 25°C, and identical numbers of cells were serial diluted, spotted onto agar plates, and grown for 3 days at the indicated temperatures. B, tabulation of the ability of Sec14, Sec14-PC, and Sfh4 to rescue cell growth of sec14 ts cells with the indicated inactivated genes. resulted in mislocalization of Rab11 implying a conserved role for PI-4P in regulation of Rab signaling as inactivation of YPT31 affected growth and vesicular transport in S. cerevisiae cells with reduced Pik1 or Sec14 function (7).
Indeed, a similar coordination in Rab signaling appears to operate in S. cerevisiae. Inactivation of the S. cerevisiae Rab YPT6 was also aggravating for growth of sec14 ts cells. GTPbound Ypt6 physically interacts with the GAP for Ypt31, Gyp2, while exhibiting no GAP activity toward Ypt6 (47). Regulation of timing of the Ypt31 and Ypt6 Rab cascades has been proposed to facilitate transport between endosomes and the trans-Golgi, similar to coordination of mammalian Rab and Rho pathways. There may be a conserved role for the Sec14 domain in regulation of Rab/Rho signaling cascades.