The Vid vesicle to vacuole trafficking event requires components of the SNARE membrane fusion machinery.

The key gluconeogenic enzyme fructose-1,6-bisphosphatase (FBPase) is targeted to Vid vesicles when glucose-starved cells are replenished with glucose. Vid vesicles then deliver FBPase to the vacuole for degradation. A modified alkaline phosphatase assay was developed to study the trafficking of Vid vesicles to the vacuole. For this assay, FBPase was fused with a truncated form of alkaline phosphatase. Under in vivo conditions, FBPase-delta60Pho8p was targeted to the vacuole via Vid vesicles, and it exhibited Pep4p- and Vid24p-dependent alkaline phosphatase activation. Vid vesicle-vacuole targeting was reconstituted using Vid vesicles that contained FBPase-delta60Pho8p. These vesicles were incubated with vacuoles in the presence of cytosol and an ATP-regenerating system. Under in vitro conditions, alkaline phosphatase was also activated in a Pep4p- and Vid24p-dependent manner. The GTPase Ypt7p was identified as an essential component in Vid vesicle-vacuole trafficking. Likewise, a number of v-SNAREs (Ykt6p, Nyv1p, Vti1p) and homotypic fusion vacuole protein sorting complex family members (Vps39p and Vps41p) were required for the proper function of Vid vesicles. In contrast, the t-SNARE Vam3p was a necessary vacuolar component. Vid vesicle-vacuole trafficking exhibits characteristics similar to heterotypic membrane fusion events.


From the Department of Cellular and Molecular Physiology, Penn State College of Medicine, Hershey, Pennsylvania 17033
The key gluconeogenic enzyme fructose-1,6-bisphosphatase (FBPase) is targeted to Vid vesicles when glucose-starved cells are replenished with glucose. Vid vesicles then deliver FBPase to the vacuole for degradation. A modified alkaline phosphatase assay was developed to study the trafficking of Vid vesicles to the vacuole. For this assay, FBPase was fused with a truncated form of alkaline phosphatase. Under in vivo conditions, FBPase-⌬60Pho8p was targeted to the vacuole via Vid vesicles, and it exhibited Pep4p-and Vid24p-dependent alkaline phosphatase activation. Vid vesicle-vacuole targeting was reconstituted using Vid vesicles that contained FBPase-⌬60Pho8p. These vesicles were incubated with vacuoles in the presence of cytosol and an ATP-regenerating system. Under in vitro conditions, alkaline phosphatase was also activated in a Pep4p-and Vid24p-dependent manner. The GTPase Ypt7p was identified as an essential component in Vid vesicle-vacuole trafficking. Likewise, a number of v-SNAREs (Ykt6p, Nyv1p, Vti1p) and homotypic fusion vacuole protein sorting complex family members (Vps39p and Vps41p) were required for the proper function of Vid vesicles. In contrast, the t-SNARE Vam3p was a necessary vacuolar component. Vid vesicle-vacuole trafficking exhibits characteristics similar to heterotypic membrane fusion events.
Lysosomal protein degradation is induced when cultured cells are starved of nutrients or serum. This process recycles amino acids and allows for the synthesis of proteins that are crucial for cell survival during starvation. Several mechanisms are responsible for increased lysosomal degradation during starvation (1)(2)(3)(4)(5). The macroautophagy pathway utilizes autophagosomes to deliver cytosolic proteins and organelles to lysosomes for degradation. The chaperone-mediated autophagy pathway targets cytosolic proteins containing a KFERQ motif for degradation. These proteins are recognized by the heat shock protein hsc73 and a lysosomal receptor LGP96 (1, 6 -9). Overproduction of LGP96 leads to an increased lysosomal degradation of substrate proteins both in vivo and in vitro (2,3).
Protein degradation by the macroautophagy pathway is also induced when Saccharomyces cerevisiae are starved of nitrogen (10 -16). Interestingly, the macroautophagy pathway overlaps with the cytoplasm to vacuole (Cvt) 1 pathway that directs aminopeptidase I from the cytoplasm to the vacuole (17)(18)(19)(20)(21)(22)(23)(24)(25)(26). Aminopeptidase I is targeted to the vacuole by Cvt vesicles during growth of cells in rich medium. However, when cells are starved of nitrogen, aminopeptidase I is targeted to the vacuole by an autophagosome-mediated process. The Cvt pathway also shares components with the pexophagy pathway that degrades peroxisomes in response to glucose (27)(28)(29)(30).
The key gluconeogenic enzyme fructose-1,6-bisphosphatase (FBPase) is degraded rapidly when glucose-starved S. cerevisiae are replenished with fresh glucose (31)(32)(33)(34)(35). After a glucose shift, FBPase is first targeted into Vid vesicles, and these vesicles then traffic to the vacuole. Vid vesicles appear to be distinct from other types of vesicle (34). Although the origin of Vid vesicles has not been established, their formation requires the ubiquitin-conjugating enzyme Ubc1p (36).
The import of FBPase into Vid vesicles has been reconstituted utilizing both a semi-intact cellular assay (35) and an isolated Vid vesicle assay (37). Via these approaches, we have identified several proteins that participate in this first trafficking step. These include the molecular chaperone Ssa2p (37), cyclophilin A (38), and Vid22p (39). The molecular mechanisms for the trafficking of Vid vesicle to the vacuole are less well characterized. Although FBPase delivery to the vacuole has been reconstituted in a semi-intact cellular assay, this assay has not allowed us to identify specific molecules involved in the trafficking of Vid vesicles to the vacuole. To address this issue, we have developed an assay to quantitate the Vid vesicle to vacuole trafficking process. FBPase was fused with a form of alkaline phosphatase that lacks the N-terminal 60 amino acids. Under in vivo conditions, the resultant fusion protein was delivered to the vacuole after a glucose shift, and alkaline phosphatase was activated in a Pep4p-and Vid24p-dependent manner. An in vitro Vid vesicle-vacuole assay was developed which reproduces the activation of alkaline phosphatase, and this activation was also dependent on Pep4p and Vid24p. Using this assay, we have identified the small GTPase Ypt7p as being required for Vid vesicle trafficking. Likewise, members of the SNARE and homotypic fusion vacuole protein sorting (HOPS) families of proteins are necessary components in this step of the FBPase trafficking pathway.

EXPERIMENTAL PROCEDURES
Yeast Strains, Plasmids, and Antibodies-S. cerevisiae strains used in this study are listed in Table I. Rabbit polyclonal antibodies directed against FBPase, Vid24p, and CPY were raised by Berkeley Antibody Company (Berkeley, CA) using purified proteins. A pho8::TRP1 knock-out plasmid and rabbit alkaline phosphatase polyclonal serum were obtained from Dr. D. Klionsky (University of Michigan). A rabbit polyclonal Ypt7p antiserum and the deletion strains ⌬vam3, ⌬vam7, ⌬vps39, ⌬vps41, and ⌬ypt7 were obtained from Dr. S. Emr (University of California San Diego). Anti-Vti1p antibody and the ⌬nyv1 deletion strain were obtained from Dr. T. Stevens (University of Oregon). A rabbit polyclonal Ykt6p antiserum was obtained from Dr. W. Wickner (Dartmouth Medical College).
The FBPase-⌬60Pho8p Fusion Protein-To produce ⌬60Pho8p, the PHO8 gene was amplified by PCR using a forward primer GCGGCC-GCACGTTCTGCATCACACAAG and a reverse primer CCGCGGGTT-GGTCAACTCATGGTA. The FBP1 gene was PCR amplified with a forward primer GGTACCATGGGTCCAACTCTAGTAAATGGA and a reverse primer GCGGCCGCCTGTGACTTGCCAATATG. These PCR products were cloned into a pYES2.1 TOPO TA plasmid (Invitrogen), and the orientation of the inserts was confirmed by PCRs. The PHO8 gene was linearized with NotI and XbaI and ligated into NotI and XbaI sites of the FBP1 plasmid. The fusion construct was either transformed into ⌬fbp1⌬pho8 strains or subcloned into an integration vector and integrated into the FBP1 locus. The expression of FBPase-⌬60Pho8p was examined by Western blotting with anti-FBPase antibodies. Similar results were obtained whether the fusion construct was integrated or expressed on a plasmid. The fbp1 deletion construct was produced as described previously (36). The deletion of fbp1 and pho8 was confirmed by Western blotting with anti-FBPase antibodies and anti-alkaline phosphatase antibodies.
To examine the degradation of FBPase-⌬60Pho8p, strains were grown in medium containing low glucose. Cells were shifted to glucosecontaining medium for the indicated times, and the levels of FBPase-⌬60Pho8p were determined. To examine the cellular distribution of FBPase-⌬60Pho8p, cell lysates were subjected to differential centrifugation techniques as described previously (37). Briefly, yeast strains (25 ml) were grown in low glucose medium and then shifted to medium containing 2% glucose for 0 and 60 min at 30°C. Total lysates were subjected to differential centrifugation conditions of 1,000 ϫ g for 10 min, 13,000 ϫ g for 20 min, and 200,000 ϫ g for 2 h. The 200,000 ϫ g pellet was resuspended in 100 l of SDS sample buffer, whereas the final 200,000 ϫ g supernatant (S200) was precipitated with 10% trichloroacetic acid, washed three times in ice-cold acetone, and solubilized in 100 l of SDS sample buffer. Proteins were resolved by SDS-PAGE, and the distribution of FBPase-⌬60Pho8p was determined by Western blotting with anti-FBPase antibodies.
Preparation of Vid Vesicles, Vacuoles, and Cytosol-Cells (25 ml) were grown to stationary phase and shifted to glucose for 30 min. Vid vesicles and cytosolic materials were obtained using differential centrifugation techniques described above. The final 200,000 ϫ g pellets (Vid vesicles) and supernatant (cytosol) were aliquoted and frozen at Ϫ70°C until further use. As an additional purification step, the 200,000 ϫ g vesicle-containing material was fractionated further on a 20 -50% sucrose density gradient. The location of vesicles within this gradient was verified by electron microscopy. Samples (5 l) were adsorbed to a Formar-coated grid for 2-3 min. Excess material was blotted away, and the samples were negatively stained with 2% phosphotungstic acid, pH 7.0. The grids were examined using a Phillips TEM400 transmission electron microscope, and micrographs were recorded at a magnification of 46,000.
Fractions containing fusion competent vesicles (fractions 3-5) were collected and used in the in vitro assays. Vacuoles were isolated from ⌬fbp1⌬pho8 strains using a protocol described previously (40). Cells were subjected to glucose starvation and a 30-min glucose shift. Spheroplasts were washed and resuspended in freezing buffer (1.2 M sorbitol, 5 mM HEPES, pH 6.8, 50 mM KOAc, and 5 mM MgOAc) and then stored at Ϫ70°C. Cells were thawed, washed once in freezing buffer, and then broken via passage through a Millipore 3.0-m filter. Samples were centrifuged at 1,000 ϫ g for 20 min and then 13,000 ϫ g for 20 min. The resultant 13,000 ϫ g pellet was resuspended in 75 l of freezing buffer, and this enriched vacuole material was used for the in vitro assay. For homotypic vacuole fusion experiments, vacuoles were also isolated as described previously (41). Protein concentrations were determined by Bio-Rad Dc protein assays.
In Vitro Vid Vesicle-Vacuole Assay-In a typical experiment, the reaction mixture (100 l) contained 20 g of vesicle material, 20 g of vacuolar material, 30 g of cytosolic proteins, and an ATP-regenerating system (0.5 mM ATP, 0.2 mg/ml creatine phosphokinase, and 40 mM creatine phosphate). The reaction mixture was incubated at 30°C for 60 min, and the reaction was terminated by a 20-min 13,000 ϫ g centrifugation step at 4°C. The 13,000 ϫ g pellet was resuspended in 500 l of alkaline phosphatase assay buffer (250 mM Tris-HCl, pH 9.0, 10 mM MgSO 4 , and 10 mM ZnSO 4 ) and then examined for alkaline phosphatase activity using 55 mM ␣-naphthyl phosphate as a substrate. Samples were incubated at 30°C for 20 min, and the reaction was quenched by the addition of 500 l of 2 M glycine, pH 11. Alkaline phosphatase activity was measured using a fluorometer (Fluoro IV, Gilford) with an excitation at 345 nm and an emission at 472 nm. Homotypic vacuole fusion experiments were conducted as described previously (41).
For antibody blocking experiments, various amounts (5-10 l) of polyclonal antibodies were added to the individual vesicle and vacuole components, and these were then incubated for 30 min at 25°C. To Mat␣ his3-⌬200 ura3-52 leu2, 3-112 lys2⌬800 trp1-⌬901 nyv1::HIS5 fbp1::FBP1-⌬60PHO8 pho8::TRP remove unbound antibody, the Vid vesicles were reisolated using the sucrose gradient protocol described above, whereas vacuoles were reisolated by a 20-min 13,000 ϫ g centrifugation procedure. The antibodytreated components were then mixed with other treated or untreated components and tested in the in vitro assay. Statistical Analysis-Statistical significance was determined using one-way analysis of variance followed by a Student-Newman-Keuls test. Data are presented as the mean Ϯ S.E. * indicates mean values that are significantly different from controls at p Ͻ 0.05.

A Fusion Protein for the Study of Vid Vesicle Trafficking-To
identify molecules involved in the trafficking of Vid vesicle to the vacuole, we used an alkaline phosphatase assay that was originally developed by the Wickner laboratory to study homotypic vacuole-vacuole fusion (42,(47)(48)(49)(50)(51)(52) and modified by the Ohsumi laboratory to study the starvation-induced autophagy pathway (12)(13)(14)(15)(16). These assays take advantage of two key characteristics of alkaline phosphatase. First, alkaline phosphatase is translocated into the endoplasmic reticulum and then transported to the Golgi where it is sorted to the vacuole (43)(44)(45)(46). Once inside the vacuole, the C-terminal pro sequence of the protein is cleaved by Pep4p, and alkaline phosphatase is activated. Second, endoplasmic reticulum translocation of alkaline phosphatase is dependent upon the N-terminal 60 amino acids of the protein, and if these amino acids are removed, the resulting ⌬60Pho8p remains in the cytosol in an inactive state.
FBPase was fused with ⌬60Pho8p and expressed in a strain in which the endogenous PHO8 gene was deleted. The 100-kDa fusion protein was recognized by both FBPase-and alkaline phosphatase-specific antibodies, but it was not present in untransformed cells (data not shown). In wild type cells, FBPase is degraded rapidly in the vacuole in a Pep4p-and Vid24p-dependent manner after the addition of glucose. Therefore, we first tested whether FBPase-⌬60Pho8p was directed to the vacuole and degraded in a similar manner. FBPase-⌬60Pho8p was induced by glucose starvation, and cells were then shifted to glucose-containing medium for various periods of time. As is shown in Fig. 1A, the addition of glucose to wild type cells resulted in a significant reduction in the levels of FBPase-⌬60Pho8p over time. In contrast, FBPase-⌬60Pho8p remained at high levels in a ⌬pep4 strain after a glucose shift. The levels of FBPase-⌬60Pho8p also remained high in the ⌬vid24 strain, a strain that blocks the trafficking of Vid vesicles to the vacuole (53). Therefore, these results suggest that FBPase-⌬60Pho8p is degraded in a manner similar to the wild type FBPase protein.
Our FBPase trafficking model predicts that FBPase- A, a FBPase-⌬60Pho8p fusion protein was expressed in wild type (WT), ⌬pep4, and ⌬vid24 cells that had the endogenous FBP1 and PHO8 genes deleted. Cells were starved of glucose and then shifted to medium containing high glucose for various periods of time. Cell lysates were examined for the levels of FBPase-⌬60Pho8p after a glucose shift using anti-FBPase antibodies (B). Wild type cells expressing FBPase-⌬60Pho8p were glucose starved and refed for 0 or 30 min. Cell lysates were then fractionated using differential centrifugation protocols, and the distribution of FBPase-⌬60Pho8p was examined by Western blot analysis. P, pellets; S, supernatant. C, cells expressing FBPase-⌬60Pho8p were either glucose starved (0 min) or shifted to glucose for 30 min. Total lysates were subjected to proteinase K (PK) treatments in the presence or absence of Triton X-100 (TX-100) and examined for the levels of FBPase-⌬60Pho8p. D, wild type, ⌬pep4, and ⌬vid24 cells expressing FBPase-⌬60Pho8p were glucose starved and then refed with fresh glucose-containing medium for 0 or 3 h. Cell lysates were examined for alkaline phosphatase activity. E, wild type cells expressing either ⌬60Pho8p or FBPase-⌬60Pho8p were subjected to nitrogen starvation conditions for 0 or 4 h and then examined for alkaline phosphatase activity. For D and E, the data shown are the averages of three experiments. * indicates values that are significantly different from controls at p Ͻ 0.05.
⌬60Pho8p should reside within Vid vesicles prior to its transport to the vacuole. To verify this localization, cells were glucose starved and then subjected to a glucose shift for 0 or 30 min. The distribution of the FBPase-Pho8⌬60p fusion protein in the high speed supernatant (enriched for cytosol) and high speed pellet (enriched for Vid vesicles) was then determined via differential centrifugation and Western blot analysis (Fig. 1B). FBPase-⌬60Pho8p was observed in the high speed supernatant fraction when cells were glucose starved. However, after a 30-min glucose shift, FBPase-⌬60Pho8p was found in the Vid vesicle pellet fraction (Fig. 1B), suggesting that it was transported into Vid vesicles. To examine whether FBPase-⌬60Pho8p is imported into a membrane bound organelle, lysates were treated with 0.8 mg/ml proteinase K for 15 min in the presence or absence of 2% Triton X-100. Proteinase K was inactivated via the addition of 1 ml of 15% w/v trichloroacetic acid, and trichloroacetic acid precipitants were examined for the presence of FBPase-⌬60Pho8p. FBPase-⌬60Pho8p was sensitive to proteinase K digestion when total cell lysates were obtained from cells that were glucose-starved (Fig. 1C). However, FBPase-⌬60Pho8p was resistant to proteinase K digestion when cells were shifted to glucose for 30 min. FBPase-⌬60Pho8p was imported into the lumen of a membrane-bound organelle because FBPase-⌬60Pho8p was degraded by proteinase K when cell lysates were treated with Triton X-100.
If the FBPase-⌬60Pho8p fusion protein behaves in a manner similar to FBPase, it should remain in the cytosol during periods of glucose starvation and not exhibit alkaline phosphatase A, a FBPase-⌬60Pho8p fusion protein was expressed in ⌬fbp1⌬pho8 strains. FB-Pase-⌬60Pho8p was imported into Vid vesicles after a shift of cells to glucose for 30 min. Vid vesicles were then isolated and incubated with vacuoles in the presence of cytosol and an ATP-regenerating system. Vacuoles and cytosol were isolated from ⌬fbp1⌬pho8 strains that contain the PEP4 gene. After the incubation, vacuoles were then reisolated and examined for alkaline phosphatase activity using a fluorometer. B, wild type cells expressing FBPase-⌬60Pho8p were shifted to glucose for 30 min. Cell lysates were subjected to differential centrifugation, and the FBPase-⌬60Pho8p containing fractions were centrifuged further using a sucrose density gradient. Each fraction was examined for the distribution of FB-Pase-⌬60Pho8p. C, fractions 4 and 10 were subjected to electron microscopy analysis. D, fractions from the sucrose gradient were tested for their ability to activate alkaline phosphatase when combined with vacuoles.
activity. However, after the addition of fresh glucose, FBPase-⌬60Pho8p should be targeted to the vacuole, and alkaline phosphatase should be activated by Pep4p. Wild type, ⌬pep4, and ⌬vid24 strains were transformed to express FBPase-⌬60Pho8p. These strains were shifted to glucose, and the activation of alkaline phosphatase was measured (Fig. 1D). When wild type cells were maintained in low glucose medium, alkaline phosphatase was not activated. However, when these cells were shifted to glucose-containing medium for 3 h, alkaline phosphatase activity increased. Therefore, FBPase was able to direct ⌬60Pho8p to the vacuole where it was enzymatically activated. In contrast, alkaline phosphatase activity was low when either the ⌬pep4 or ⌬vid24 strain was shifted to glucose for 3 h. These results further confirm that FBPase-⌬60Pho8p is targeted to the vacuole by the Vid vesicle-mediated pathway.
Autophagic processes often result in the nonselective transport of cytosolic proteins to the vacuole for degradation (23). To determine whether FBPase-⌬60Pho8p could be targeted to vacuoles via the autophagy pathway, we subjected cells to a period of nitrogen starvation and then examined for the activation of alkaline phosphatase. A wild type strain expressing ⌬60Pho8p was used as a control because ⌬60Pho8p is targeted to the vacuole via the autophagy pathway after nitrogen starvation (13, 15). As expected, low levels of alkaline phosphatase activity were observed when the ⌬60Pho8p strain was grown under normal conditions. However, a high level of alkaline phosphatase activity was observed when these same cells were subjected to nitrogen starvation (Fig. 1E). In contrast, when FBPase was fused with ⌬60Pho8p, alkaline phosphatase activity was low either before or after nitrogen starvation. Therefore, it appears that the FBPase-⌬60Pho8p fusion protein is targeted to the vacuole via the Vid-dependent pathway, but not via the autophagy pathway.
Development of an In Vitro Assay-Alkaline phosphatase activity has been used as an indicator of in vitro homotypic vacuole fusion. In this assay, donor vacuoles contain an inactive precursor form of alkaline phosphatase in a ⌬pep4 genetic background, whereas acceptor vacuoles are derived from a strain that contains the PEP4 gene but lacks the PHO8 gene. The fusion of donor and acceptor vacuoles results in the conversion of inactive proalkaline phosphatase to active alkaline phosphatase after the cleavage of the pro sequence by Pep4p. Using this as our model, we devised a strategy to examine Vid vesicle to vacuole content mixing under in vitro conditions ( Fig.  2A). A shift to glucose-containing medium initiates the import of FBPase-⌬60Pho8p into Vid vesicles (see Fig. 1). These loaded vesicles can be isolated and incubated with vacuoles in the presence of cytosol and an ATP-regenerating system. If the content mixing of Vid vesicles with vacuoles occurs under in vitro conditions, alkaline phosphatase should be activated. To enrich for Vid vesicles that were competent for our in vitro assay, an additional fractionation step was performed. The high speed pellet material was further resolved via the use of a 20 -50% sucrose density gradient. Fractions were collected from the sucrose gradient and examined for the distribution of FBPase-⌬60Pho8p (Fig. 2B). The FBPase-⌬60Pho8p fusion protein was found in both light (fractions 3-5) and heavy (fractions 9 -11) density peaks. Vesicular material was observed as the primary component of samples taken from each of these peaks (Fig. 2C).
To determine whether these two Vid vesicle populations were functional in our in vitro assay, small aliquots of the sucrose gradient fractions were combined with vacuoles in the presence of ATP and cytosol and examined for alkaline phosphatase activity (Fig. 2D). Interestingly, the highest alkaline phosphatase activity was found when the light density fractions were mixed with vacuoles. In contrast, fractions 9 -11 failed to activate alkaline phosphatase when incubated with vacuoles. Consequently, all subsequent assays were conducted using the light density Vid vesicle fractions.
To determine the optimal time required for alkaline phosphatase activation, the in vitro components were mixed and incubated for various periods of time. Alkaline phosphatase activity was low at the 0 incubation time (Fig. 3A). However, this activity increased in a time-dependent manner, reaching peak levels after a 60-min incubation. Alkaline phosphatase activity also increased in a dose-dependent manner. The optimal concentrations for Vid vesicles (Fig. 3B), vacuoles (Fig. 3C), and cytosol (Fig. 3D) were 20, 12, and 32 g, respectively. Note that vesicles, vacuoles, and cytosol are all required for in vitro activation of alkaline phosphatase activity (Fig. 3E).
Pep4p and Vid24p Are Required for In Vitro Activation of Alkaline Phosphatase-A prime reason for the development of the in vitro assay was to identify proteins that are involved in the Vid vesicle to vacuole trafficking step. To verify the efficacy of this assay, we first tested ⌬pep4 and ⌬vid24 mutant strains because these strains have site-specific defects. Consistent with previous data, wild type vesicles and vacuoles exhibited high alkaline phosphatase activity when they were combined in our in vitro assay. By contrast, alkaline phosphatase activity was low when these components were isolated from the ⌬pep4 strain (Fig. 4A). When Vid vesicles from the ⌬pep4 strain were combined with wild type vacuoles, alkaline phosphatase was activated. By contrast, when vacuoles from the ⌬pep4 strain were incubated with Vid vesicles from the wild type strain, alkaline phosphatase activity was low. Therefore, the ⌬pep4 strain contains competent Vid vesicles but defective vacuoles that failed to activate alkaline phosphatase.
The ⌬vid24 strain exhibits a defect in the trafficking of Vid vesicles to the vacuole under in vivo conditions (53). Likewise, the combination of ⌬vid24 vesicles and vacuoles was also defective in the in vitro assay (Fig. 4B). When wild type vesicles were incubated with vacuoles from the ⌬vid24 strain, alkaline phosphatase activity was high (Fig. 4B). By contrast, alkaline phosphatase activity was low when ⌬vid24 Vid vesicles were incubated with wild type vacuoles. Therefore, the ⌬vid24 strain contains defective Vid vesicles but competent vacuoles. Because Vid24p normally resides on the surface of Vid vesicles, these results further support the notion that Vid24p is a required component on the Vid vesicle membrane.
The GTPase Ypt7p Is Required for the Trafficking of Vid Vesicles to the Vacuole-GTPases are known to play an important role in a number of protein trafficking events. Therefore, we suspected that GTPases may be involved in the trafficking of Vid vesicles to the vacuole. To test this idea, an in vitro Vid vesicle-vacuole reaction experiment was conducted in which wild type components were incubated in the presence or absence of the GTPase inhibitor GTP␥S. In the absence of GTP␥S, high alkaline phosphatase activity was observed when Vid vesicles were incubated with vacuoles under the in vitro conditions (Fig. 5A). However, alkaline phosphatase activity was significantly reduced when GTP␥S was added to the wild type reaction mixture, indicating that the in vitro Vid vesicle-vacuole reaction requires the hydrolysis of GTP.
Ypt7p is a GTPase that is involved in the trafficking of late endosomes to the vacuole (54, 55), targeting of Cvt vesicles or autophagosomes to the vacuole (22,23) and homotypic vacuole fusion (56,57,59). Ypt7p binds and hydrolyzes GTP and thus cycles between active (GTP-bound) and inactive (GDP-bound) states (56 -59). To determine whether Ypt7p is required for FBPase trafficking, we first examined the degradation of FB-Pase in a ⌬ypt7 strain. The ⌬ypt7 strain exhibited a defect in the degradation of FBPase compared with a wild type strain (Fig. 5B). This suggests that Ypt7p plays some role in the FBPase degradation pathway and that it is necessary for the delivery of FBPase to the vacuole. To verify this, we next performed our in vivo alkaline phosphatase assay using a ⌬ypt7 strain expressing FBPase-⌬60Pho8p. Wild type and ⌬ypt7 strains were subjected to glucose starvation and refeeding protocols, and cellular lysates were examined for the in vivo activation of alkaline phosphatase. Wild type cells exhibited an increased alkaline phosphatase activity in response to glucose. In contrast, the ⌬ypt7 strain failed to activate alkaline phosphatase after a glucose shift (Fig. 5C). Thus, Ypt7p is clearly required for one of the steps in the FBPase trafficking pathway. To identify the site of the ⌬ypt7 defect, cells were subjected to a glucose shift at 0 or 60 min, and the distribution of FBPase in the cytosol (high speed supernatant) and Vid vesicles (high speed pellet) was examined. As controls, we also examined the localization of FBPase in ⌬vid22 and ⌬vid24 strains. For the ⌬ypt7 strain, a significant portion of FBPase was observed in the Vid vesicle containing fraction following a glucose shift at 60 min (Fig. 5D). This was similar to the results obtained with the ⌬vid24 strain, a strain that is defective in the Vid vesicle to vacuole trafficking step. By contrast, the ⌬vid22 strain exhibits a defect in FBPase import into Vid vesicles (39) and therefore accumulated FBPase in the cytosol. These results suggest that the ⌬ypt7 and ⌬vid24 strains block FBPase degradation in the same trafficking step.
We next performed in vitro experiments to clarify further the function of Ypt7p in Vid vesicle-vacuole trafficking. When ⌬ypt7 Vid vesicles and vacuoles were combined in the in vitro assay, alkaline phosphatase activity was low (Fig. 6A). When ⌬ypt7 vesicles were combined with wild type vacuoles, alkaline phosphatase was not activated. However, when wild type vesicles were combined with ⌬ypt7 vacuoles, high levels of alkaline phosphatase activity were observed (Fig. 6A). Thus, it appears that Ypt7p is a required component for the proper function of Vid vesicles.
The preceding data suggest that Ypt7p exerts its effect at the Vid vesicle, and thus it may be a Vid vesicle resident protein.
To test this possibility, the distribution of Ypt7p was examined in various cellular fractions (Fig. 6B). As a control, we also examined the localization of FBPase-⌬60Pho8p and CPY because these proteins reside in the Vid vesicles and vacuoles, respectively. In cells that were shifted to glucose for 30 min, the majority of FBPase-⌬60Pho8p was detected in the Vid vesicle fraction, whereas CPY was found primarily in the vacuole fraction. In contrast, Ypt7p was distributed in both the vacuoleand Vid vesicle-containing fractions (Fig. 6B). Taken together, these results indicate that a portion of Ypt7p resides on Vid vesicles. Therefore, Ypt7p appears to play a direct role in the trafficking of Vid vesicles to the vacuole.

t-SNAREs, v-SNAREs, and the HOPS Complex Are Required for the Activation of the FBPase-⌬60Pho8p Fusion Protein-
The transport of cargo from one cellular compartment to another often requires the participation of SNARE proteins. v-SNARES are on the donor membranes, whereas t-SNAREs are required on the target membranes. Unique combinations of SNAREs are often utilized in particular trafficking events (60). For example, v-SNAREs (Nyv1p, Vti1p,Ykt6p), t-SNAREs (Vam3p, Vam7p), and HOPS complex family members (Vps39p, Vps41p) participate in homotypic vacuolar fusion (42). However, the function of individual SNAREs is not restricted to one cellular compartment, and there can be consid-

FIG. 5. Ypt7p is necessary for the degradation of FBPase and the activation of alkaline phosphatase in vivo.
A, Vid vesicles and vacuoles were isolated from a wild type strain and used in the in vitro reaction. GTP␥S was added to the in vitro reaction mixture at 2 and 4 M, and alkaline phosphatase activity was measured. B, wild type (WT) and ⌬ypt7 strains were glucose starved and then refed with glucose-containing medium. Cells were harvested at various times, and cell lysates were examined for the degradation of FBPase. C, wild type and ⌬ypt7 cells expressing FBPase-⌬60Pho8p were glucose starved and then refed for 0 or 3 h. Cell lysates were examined for alkaline phosphatase activity. For A and C, the data shown are the average of three experiments, where * indicates values that are significantly different from controls at p Ͻ 0.05. D, ⌬ypt7, ⌬vid22, and ⌬vid24 strains were glucose starved and then fed with glucose-containing medium for 0 or 60 min. Cell lysates were subjected to differential centrifugation, and the resultant 200,000 ϫ g supernatant (S) and pellets (P) were examined for the distribution of FBPase. erable overlap between different trafficking pathways. For example, many of these same proteins are also involved in heterotypic fusion events in which two dissimilar membranebound compartments (e.g. vesicles and vacuoles) fuse together. Because FBPase traffics to the vacuole via Vid vesicles, it is possible that the final step in this transport also involves the fusion of Vid vesicles with the vacuole. If so, then one would also predict that a particular subset of SNARE proteins may participate in this process.
In preliminary experiments, we determined that each of the aforementioned SNARE proteins associates with Vid vesicle membranes (data not shown). To determine whether these SNAREs play an essential role in FBPase trafficking, we tested various deletion strains for their ability to degrade FBPase. Nyv1p is a v-SNARE that is required for heterotypic vesicle fusion (47,61) and homotypic vacuolar fusion (61,62). The ⌬nyv1 strain exhibited a defect in FBPase degradation compared with a wild type strain (Fig. 7A), indicating that some step in the FBPase degradation pathway was compromised in this strain. To identify the site of the defect, we performed fractionation experiments and determined that FBPase was imported into Vid vesicles in the wild type strain (data not shown). This suggested that the Vid vesicle to vacuole trafficking step might be defective in this strain. To verify this, we next performed our in vitro alkaline phosphatase activity assay. As expected, there were low levels of alkaline phosphatase activity when Vid vesicles and vacuoles were obtained from the ⌬nyv1 strain (Fig. 7B). In further experiments, we were able to identify the Vid vesicles as the defective in vitro component because alkaline phosphatase activity was observed when wild type vesicles were mixed with ⌬nyv1 vacuoles but not when ⌬nyv1 vesicles were mixed with wild type vacuoles (Fig. 7B).
The v-SNAREs Vti1p and Ykt6p play important roles in multiple cellular trafficking events. For example, Vti1p is involved in Golgi to vacuole trafficking (62,63), whereas Ykt6p is a v-SNARE that plays a role in trafficking through the Golgi (64,65). Both of these proteins also play a role in homotypic vacuolar fusion (51). Deletion of either the VTI1 or YKT6 genes is lethal in yeast. Furthermore, strains with temperature-sensitive mutations of these genes were problematic in terms of FBPase and fusion protein expression. Therefore, we performed antibody-blocking experiments to examine the role that these proteins play in Vid vesicle trafficking. When anti-Vti1p or anti-Ykt6p antibodies were added both to Vid vesicles and vacuoles, there was a statistically significant decrease in the levels of alkaline phosphatase activity (Fig. 8). In contrast, the addition of equivalent amounts of preimmune serum had no effect on alkaline phosphatase activity. Antibodies were also added to the individual Vid vesicle or vacuole components, and these components were reisolated prior to their use in the in vitro assay. The addition of anti-Vti1p antibody to the Vid vesicle fraction resulted in a significant decrease in alkaline phosphatase activity (Fig. 8A). In contrast, the addition of anti-Vti1p antibody to the vacuole fraction did not have an inhibitory effect. Thus, it appears that Vid vesicles require the presence of functional Vti1p for the in vitro activation of the fusion protein, but vacuoles do not. Similar results were ob- tained when antibody-blocking experiments were performed using the anti-Ykt6p antibody. The anti-Ykt6p antibody blocked the function of Vid vesicles, but it had no effect on the function of vacuoles (Fig. 8B).
Vam3p and Vam7p are vacuolar t-SNAREs that are part of a cis-SNARE complex on vacuoles (51). Vam3p plays a role in the docking process (61), whereas Vam7p stabilizes the interactions of other SNARE components (66). Deletion of either of the VAM3 or VAM7 genes resulted in an FBPase degradation defect (Fig. 9A). Furthermore, both of these strains exhibited a defect in alkaline phosphatase activation, when Vid vesicles and vacuoles were obtained from the deletion strains (Fig. 9). However, in contrast to our results with v-SNAREs, the presence of Vam3p appeared to be required on the vacuole fraction because vacuoles obtained from the deletion strain were defective (Fig. 9B). Interestingly, the combination of ⌬vam7 vesicles and wild type vacuoles, or wild type vesicles and ⌬vam7 vacu-oles, resulted in intermediate alkaline phosphatase activities (Fig. 9C). Thus, the presence of this protein on either compartment may allow a partial activation of alkaline phosphatase.
The HOPS complex is known to be required for homotypic vacuole fusion and endosome to vacuole fusion (42). Two members of this complex, Vps39p and Vps41p, play an essential role in the docking step of homotypic vacuole fusion (67). Therefore, we tested whether these proteins also play some role in the FBPase trafficking pathway. Both ⌬vps39 and ⌬vps41 deletion strains exhibited a defect in FBPase degradation (Fig. 10A). Likewise, both of these strains contained defective in vitro components that exhibited low alkaline phosphatase activities (Fig. 10, B and C). Here again, the vesicles were identified as the site of this defect because there was low alkaline phosphatase activity when mutant vesicles were combined with wild type vacuoles. However, neither Vps39p nor Vps41p appears to be required on the vacuole fraction. This is in contrast to the results observed when homotypic fusion experiments were performed with these same strains. Isolated vacuoles from these strains exhibited low levels of in vitro homotypic fusion activity (Fig. 10D). DISCUSSION The FBPase degradation pathway can be divided into at least two steps: FBPase import into Vid vesicles and Vid vesicle trafficking to the vacuole. At present, we have identified several molecules that participate in the sequestration of FBPase into Vid vesicles. However, the Vid vesicle to vacuole trafficking step has not been well characterized. The investigation of the Vid vesicle/vacuole step has been limited to in vivo analyses utilizing various mutants that accumulate FBPase in Vid vesicles. However, some of these mutants may have indirect effects. Therefore, to identify molecules that are directly involved in this step of the FBPase degradation pathway, we have developed an in vitro alkaline phosphatase assay that can recapitulate Vid vesicle to vacuole trafficking or fusion. First, we demonstrated that FBPase-⌬60Pho8p could be delivered to Vid vesicles and vacuoles under in vivo conditions. Targeting of FBPase-⌬60Pho8p to the vacuole led to a Pep4p-dependent activation of alkaline phosphatase activity. This Pep4p-and Vid24p-dependent activation of alkaline phosphatase was re-produced using our in vitro conditions. Hence, the activation of alkaline phosphatase can be used to quantitate the trafficking of Vid vesicles to the vacuole under both in vivo and in vitro conditions.
Ypt7p was found to play an important role in the Vid vesicle/ vacuole trafficking step. Ypt7p is a GTPase that is essential for a number of cellular trafficking events. Endosomes require the presence of Ypt7p to be targeted to the vacuole (54,55), as do Cvt vesicles or autophagosomes (22,23). Likewise, Ypt7p is required for the process of homotypic vacuolar fusion (56,57,59) where it mediates tethering (48). In our studies, FBPase accumulated in Vid vesicles in the ⌬ypt7 mutant. In vitro activation of alkaline phosphatase was defective in this mutant, and Vid vesicles were identified as the defective component. Therefore, Ypt7p is necessary for the proper function of Vid vesicles. Interestingly, Ypt7p was not required for the function of vacuoles in our assay. This is in contrast to the situation in homotypic vacuole fusion, where Ypt7 is required on both the donor and acceptor vacuoles (56).
We have also identified a role for various SNARE proteins in the trafficking and degradation of FBPase. v-SNAREs and t-SNAREs are generally found on opposing vesicle and target membranes, respectively. Under the proper circumstances, A, wild type (WT), ⌬vps39, and ⌬vps41 strains were glucose starved and then refed with glucose-containing medium. Cells were harvested at various times, and cell lysates were examined for the degradation of FBPase. B, Vid vesicles and vacuoles were isolated from wild type and ⌬vps39 strains. C, Vid vesicles and vacuoles were isolated from wild type and ⌬vps41 strains and used in the in vitro assay as indicated. D, homotypic vacuolar fusion experiments were performed using vacuoles that were isolated from wild type ⌬pho8, ⌬vps39⌬pho8, and ⌬vps41⌬pho8 strains. These vacuoles were incubated with vacuoles from a ⌬pep4 strain and examined for alkaline phosphatase activity. The data shown in B, C, and D are the average of three experiments. * indicates values that are significantly different (p Ͻ 0.05) from controls. these SNAREs can assemble to form a SNARE complex, which promotes the fusion of the membranes. In our studies, the v-SNAREs Nyv1p, Vti1p, and Ykt6p were required for the trafficking of Vid vesicles to the vacuole, and they exerted their effects on the Vid vesicle fraction. The requirement for multiple v-SNAREs on Vid vesicles is similar to that reported for the homotypic fusion of vacuoles. Vacuoles utilize these same three v-SNAREs (42). The t-SNARE Vam3p was also necessary for in vitro activation in our assay, but it was functional only on the vacuoles. Interestingly, when examined under in vitro conditions, the SNAP25 homolog Vam7p exhibited a partially defective phenotype. When Vam7p was absent from both the Vid vesicles and vacuoles, there were low levels of alkaline phosphatase activity. However, if Vam7p was present on either of these fractions, there was an intermediate level of alkaline phosphatase activity observed. The explanation for this effect is not clear at the present time. However, Vam7p does not directly mediate fusion, but instead it stabilizes the interaction of other SNARE components in homotypic vacuolar fusion (66). Therefore, there may be a dosage effect whereby the presence of Vam7p on one compartment allows partial formation of these complexes.
Many of the proteins required for homotypic vacuolar fusion are present on the surface of the vacuole in the form of oligomeric complexes (42). These include Vam3p, Vam7p, Nyv1p, Vti1p, and Ykt6p. Likewise, the HOPS complex is also part of a hexameric complex that localizes to this organelle (48). We have not determined whether similar complexes exist on the Vid vesicles. However, the presence of these proteins on Vid vesicles may indicate that they function in a similar manner on this organelle. Because members of the HOPS complex were required on the Vid vesicle fraction, this suggests that the association of Vid vesicles with vacuoles appears to resemble more closely a heterotypic rather than a homotypic fusion event. This was further confirmed by our results with the t-SNARE Vam3p. Even though this protein was present on both Vid vesicles and vacuoles (data not shown), it was only required on the vacuolar fraction. This observation is in agreement with the established role of Vam3p as a t-SNARE. In homotypic vacuole fusion, transient formation of a Vam3p and Nyv1p SNARE pair between opposite membranes stabilizes vacuole docking (50). Therefore, it is possible that a similar situation exists during the association of Vid vesicles with vacuoles.
Recent data regarding the structural characteristics of Vam3p and Vti1p may shed some light on the specificity of Vid vesicle trafficking (68,69). Conformational changes in the target membrane syntaxin may play an important role in SNARE complex formation. It has been proposed that the functions of many SNAREs can be regulated by their adoption of a closed versus an open conformation of their four-helix bundle structure. SNAREs in the closed conformation are inactive, whereas SNAREs in the open conformation are active. Both Vam3p and Vti1p, however, appear to remain in the open conformation (68,69), suggesting that there must be some other method of regulation. This may include the presence of other proteins that regulate the conformation of the SNARE complex. One likely candidate is the protein Vps33, a member of the HOPS complex which has been shown to form an interaction with Vam3p (68).
The FBPase trafficking and degradation pathway is distinct from other known protein trafficking pathways. As such, it requires the participation of at least one protein, Vid24p, that has no other known function. Here, for the first time, we have identified components that are required both for Vid vesicle trafficking and for other well characterized protein trafficking pathways. Thus, the Vid vesicle-vacuole trafficking step re-quires proteins that are common to other trafficking pathways, as well as molecules specifically involved in FBPase degradation. Further experiments will be required to identify additional components that are essential for this step of the FBPase degradation pathway.