The type 1 phosphatase Reg1p-Glc7p is required for the glucose-induced degradation of fructose-1,6-bisphosphatase in the vacuole.

Protein phosphatases play an important role in vesicular trafficking and membrane fusion processes. The type 1 phosphatase Glc7p and its regulatory subunit Reg1p were identified as required components in the glucose-induced targeting of the key gluconeogenic enzyme fructose-1,6-bisphosphatase (FBPase) to the vacuole for degradation. The interaction of Reg1p with Glc7p was important for the transport of FBPase from intermediate vacuole import and degradation (Vid) vesicles to vacuoles. The glc7-T152K mutant strain exhibited a reduced Reg1p binding along with defects in FBPase degradation and Vid vesicle trafficking to the vacuole. In this mutant, Vid vesicles were the most defective components, whereas the vacuole was also defective. Shp1p and Glc8p regulate Glc7p phosphatase activity and are required for FBPase degradation. In the Deltashp1 and Deltaglc8 strains, Reg1p-Glc7p interaction was not affected, suggesting that phosphatase activity is also necessary for FBPase degradation. Similar to those seen in the glc7-T152K mutant, the Deltashp1 and Deltaglc8 mutants exhibited severely defective Vid vesicles, but partially defective vacuoles. Taken together, our results suggest that Reg1p-Glc7p interaction and Glc7p phosphatase activity play a required role in the Vid vesicle to vacuole-trafficking step along the FBPase degradation pathway.

Regulation of the protein degradation process plays an essential role in cells. Although each cellular protein has its own intrinsic rate of turnover, this can be altered dramatically depending upon changes in the cellular environment. For example, in Saccharomyces cerevisiae, proteins are targeted to the vacuole (the yeast homologue of the lysosome) following periods of nitrogen starvation by autophagy (1)(2)(3). This process appears to be nonspecific, because cytosolic proteins are engulfed non-selectively in autophagosomes and delivered to the vacuole (2)(3)(4)(5). Serum starvation also enhances the rate of turnover of cellular proteins in mammalian cells, by targeting proteins for degradation in the lysosome (6,7). This is a more selective process, in that proteins with a KFERQ motif are delivered to the vacuole in a molecular chaperone-mediated manner (8).
The process of protein degradation also removes misfolded proteins, or proteins that are no longer required for a specific function. Along these lines, certain gluconeogenic enzymes are essential when yeast are grown in poor carbon sources, but they are rapidly inactivated when cells are replenished with fresh glucose (9,10). This process is critical for the cell, because it helps to prevent energy futile cycles. At present, the mechanisms of inactivation and degradation are not well established for most gluconeogenic enzymes. However, both vacuolar and proteasome-dependent degradation pathways have been described for fructose-1,6-bisphosphatase (FBPase) 1 (11)(12)(13). The vacuolar degradation pathway has been shown to require a vesicle-dependent trafficking step (14 -16). FBPase is first targeted to intermediate vesicles and then to the vacuole for degradation. These vacuole import and degradation (Vid) vesicles have been purified to near homogeneity and partially characterized (14). The formation of Vid vesicles requires the ubiquitin-conjugating enzyme Ubc1p (17). FBPase import into Vid vesicles is dependent on the heat shock protein Ssa2p (18), cyclophilin A (19), and Vid22p (20). Following FBPase import into Vid vesicles, this protein is then delivered to the vacuole for degradation via a Vid24p-, Ypt7p-, and SNARE-mediated process (21). Membrane trafficking events require the participation of various kinases and phosphatases (1)(2)(3)(22)(23)(24)(25). For example, Glc7p is the catalytic subunit of the protein phosphatase type 1 (PP1) and plays a critical role in several cellular trafficking events, including homotypic vacuole fusion, endoplasmic reticulum-to-Golgi transport and endocytic transport (22). Specificity of Glc7p is dictated by regulatory subunits that target the catalytic subunit to various substrates for different functions under different growth conditions. As an example, Gac1p and Pig1p target Glc7p to glycogen accumulation (26,27). Reg2p targets Glc7p to growth and cell cycle progression (28), whereas the formation of the Bni4p-Glc7p complex is required for cytokinesis (29). Bud14p interacts with Glc7p and is involved in cellular morphogenesis during vegetative growth (30). Gip1p targets Glc7p to participate in meiotic and sporulation processes (31). The regulatory subunit Reg1p has been shown to strongly interact with Glc7p in the presence or absence of glucose (36), and this interaction directs Glc7p to a number of cellular processes, including glucose repression, growth, and glycogen accumulation (28,(32)(33)(34)(35)(36)(37). In addition to proteins that regulate Glc7p specificity, Glc7p-PP1 activity can also be modulated by Shp1p and Glc8p (38,39). SHP1 (suppressor of high copy PP1) was identified as a mutation that suppresses the lethality caused by Glc7p overexpression (38). Mutations in the SHP1 gene lead to a marked reduction of the Glc7p-PP1 activity (38). Likewise, deletion of the GLC8 gene also significantly decreases Glc7p-PP1 activity (39). Under the same conditions, deletion of other regulatory subunits such as Reg1p, Reg2p, and Gac1p had minimal effects on Glc7p-PP1 activity (39). At present, no role has been established for any phosphatase or phosphatase regulators in the trafficking of FBPase to the vacuole. However, Reg1p has been reported to play a role in the proteasome-dependent degradation of FBPase, presumably via its action as a signaling molecule (40).
The interaction of Glc7p with its regulatory subunits has been studied in detail. Analysis of the crystal structure revealed a hydrophobic grove on the surface of PP1c that is important for PP1c interaction with the PP1c-binding motif ((R/K)(V/I)XF) found in many targeting subunits of PP1c (41).
Site-directed mutagenesis studies demonstrated that this hydrophobic groove is indeed necessary for Glc7p binding to (R/ K)(V/I)XF-containing regulatory subunits, including those in Reg1p and Gac1p. This hydrophobic groove is not only important for Glc7p binding to the motif, but it is also necessary for Glc7p biological activity and for its proper subcellular localization (42). In addition to the hydrophobic groove, charged residues on the surface of Glc7p are important for its functions. For glucose derepression and cell cycle control, critical residues have been mapped to specific regions of the protein. For glycogen synthesis and sporulation, these charged residues are more widely distributed over the protein surface (43).
We have identified a number of proteins that participate in the degradation of FBPase. These include the molecular chaperone Ssa2p (18), the immunophilin cyclophilin A (19), and a plasma membrane protein Vid22p (20). Likewise, Vid24p (16) and various members of the SNARE and homotypic fusion vacuole protein sorting families of proteins (21) play roles in this process. To identify additional proteins that are involved in the FBPase degradation pathway, we screened a yeast GST library in search of proteins that bind to FBPase. The phosphatase regulatory subunit Reg1p was identified as a putative FBPase-binding protein. FBPase degradation was defective in a ⌬reg1 strain, suggesting that Reg1p plays some role in the FBPase degradation. An interaction between Reg1p and Glc7p appears necessary for FBPase trafficking. A yeast strain harboring a mutant form of Glc7p that inhibits the Reg1p interaction was defective in both FBPase trafficking and degradation. The ⌬glc8 and ⌬shp1 strains that affect the catalytic activity of Glc7p also exhibited impaired degradation of FB-Pase. The functions of Vid vesicles and vacuoles were affected by mutations in the GLC7, SHP1, and GLC8 genes. Taken together, our results indicate that Reg1p-Glc7p plays an essential role in the trafficking of Vid vesicles to the vacuole.

EXPERIMENTAL PROCEDURES
Yeast Strains, Primers, and Antibodies-S. cerevisiae strains used in this study are listed in Table I. The deletion strains derived from BY4742 were from Euroscarf (Euroscarf, Germany), and the yeast GST strains in this study were from Dr. Eric Phizicky (University of Roch- Mouse monoclonal anti-HA and anti-myc were purchased from Covance. Mouse monoclonal anti-V5 antibody was purchased from Invitrogen. Goat polyclonal antibodies directed against GST were purchased from Amersham Bioscience, as were horseradish peroxidase (HRP)conjugated donkey anti-goat antibodies, HRP goat anti-rabbit, and HRP goat anti-mouse antibodies. Anti-Glc7 sera were obtained from Dr. A. Mayer (Friedrich-Miescher-Labor). The enhanced chemiluminescence kit was purchased from PerkinElmer Life Sciences. A forward primer, CATCTTTGAAAAATAAAGTCGAGTCCAGTGATTGTTCTTTTGAGT-TTGCTCGGATCCCCGGGTTAATTAA, and a reverse primer, TAGAC-ATAGACATGCTGTTATCATACCAAATAGAAAAGTGTACAGTCTTT-GAATTCGAGCTCGTTTAAAC, were used for Vid24p-HA integration. A forward primer, GTCGACATGGGTTCAACAAATCTAGCAAAT, and a reverse primer, GTCGACACTGCTGTCATTTCCATTTTC, were used for the Reg1p-V5-His6 expression construct. A forward primer, CTTG-GAAAAAGAGTGACGTCAAGCCACAAGAAAATGGAAATGACAGCA-GTCGGATCCCCGGGTTAATTAA, and a reverse primer, TGACAATG-CCAGTCGATTACAGCTTACTTGGATCCTAAAGACGGCACTGAGA-ATTCGAGCTCGTTTAAAC, were used for Reg1p-myc tagging. The PCR products were purified with the Wizard SV PCR Clean-Up system (Promega) and integrated as described previously (44).
Screening for FBPase Interacting Proteins-Yeast strains containing a GST tag on selected open reading frames, or pools of these strains, were grown for 2 days in YPKG (10 g/l bacto-yeast extract, 20 g/l bacto-peptone, 10 g/l potassium acetate, 5 g/l dextrose) media (5 ml) and shifted to YPD (10 g/l bacto-yeast extract, 20 g/l bacto-peptone, 20 g/l dextrose) for 30 min as described (16,21). Cells were lysed in GST binding buffer (140 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 ) supplemented with 1 mM phenylmethylsulfonyl fluoride. Lysates were clarified by centrifugation at 1,000 ϫ g for 5 min at 4°C. The supernatant was then incubated with 2% Triton X-100 for 30 min at room temperature to solubilize membranes. Insoluble cell debris was removed by centrifugation (13,000 ϫ g for 20 min at 4°C). Supernatants were then incubated with pre-washed glutathione-Sepharose 4B beads (Amersham Biosciences) for 3 h at 4°C. The beads were collected by centrifugation at 500 ϫ g for 1 min, and then washed five times with GST binding buffer. The washed beads were separated by SDS-PAGE. FBPase bound to GST fusion proteins were detected via Western blotting procedures using anti-FBPase antibodies.
Immunoprecipitation-Yeast cells were grown in YPKG (10 ml) for 2 days and shifted to YPD for 30 min. Cells were harvested, resuspended in IP buffer (200 mM NaCl, 5 mM EDTA, 50 mM Tris, pH 7.4, 1 mM phenylmethylsulfonyl fluoride), and lysed. Cell lysates were clarified by centrifugation (1,000 ϫ g for 5 min at 4°C), followed by Triton X-100 (2%) solubilization for 30 min at room temperature. After centrifugation (13,000 ϫ g for 20 min at 4°C), the supernatant was then incubated with anti-FBPase, anti-V5, or anti-myc antibodies at 4°C for 2 h. The protein-antibody complex was precipitated using 100 l of 50% slurry of protein A or protein G beads (Amersham Biosciences), followed by incubation at 4°C for 1 h. Beads were then washed three times with IP buffer, and proteins were solubilized in SDS buffer and separated by SDS-PAGE. Proteins bound to the beads were detected by Western blotting. For Reg1p-V5-His6 pull down experiments, 10 ml of cultures was grown and lysed in buffer containing 50 mM NaH 2 PO 4 , 300 mM NaCl, and 10 mM imidazole, pH 8. Lysates were solubilized with 2% Triton X-100 and centrifuged at 13,000 ϫ g, and the supernatants were incubated with 50-l nickel beads. The bound material was washed three times in washing buffer containing 50 mM NaH 2 PO 4 , 300 mM NaCl, and 30 mM imidazole, pH 8. Proteins were solubilized in SDS buffer, separated by SDS-PAGE, and blotted with anti-Glc7p or anti-V5 antibodies.
FBPase phosphorylation labeling experiments were performed as described by Jiang and Broach (45). Briefly, cells were grown in synthetic complete (SC) medium containing 3% glycerol for 2 days, followed by a shift to low phosphate medium (0.2 mM potassium phosphate) containing 3% glycerol for 12 h. After incubation with 32 PO 4 (3 mCi, Amersham Biosciences) for 15 min, 2% glucose was added. Cells were removed at indicated times and fixed via the addition of trichloroacetic acid. Cells were washed and lysed, and the cell lysate was subjected to immunoprecipitation as described above. The 32 P-labeled proteins were visualized by autoradiography, and incorporated 32 P was quantitated using a PhosphorImager (Amersham Biosciences). To quantitate the overall levels of FBPase, 1 ml of cells was removed just prior to the addition of 32 PO 4 and examined by Western blot. Cells were lysed, and samples were separated by SDS-PAGE. FBPase was detected using anti-FBPase antibody.
For FBPase pulse-chase experiments, we utilized a modification of the protocol described by Horak and Wolf (40). Cells were grown in SC media without L-methionine to A 600 ϭ 5. Cells then were washed twice with pulse-chase media (SC media without L-methionine, but containing 2% ethanol) and then incubated in the same media for 4 -5 h. [ 35 S]Met and [ 35 S]Cys were added to the cells to the final concentration of 45 Ci/ml. After 5-6 h of labeling, the cells were collected by centrifugation and resuspended in SC media containing 10 mM L-methionine, 10 mM L-cysteine, and 2% glucose. At the indicated time points, 1-ml aliquots of cells were collected and fixed by the addition of trichloroacetic acid (5%). After washing the cells three times with cold water, cells were lysed and subjected to immunoprecipitation as described above. The 35 S-labeled proteins were quantified with a PhosphorImager.
In Vitro Vid Vesicle-Vacuole Assay-This assay was performed as described previously (21) with minor modifications. Cells were grown in YPKG media for 2 days at 30°C and then shifted to glucose containing media for 30 min. Vid vesicles were isolated by differential centrifugation followed by a two-step 20 and 40% sucrose density gradients. Fractions containing fusion competent vesicles were collected from the interface and used in the in vitro assays. Vid vesicles were isolated either from wild type cells or the ⌬vam3 strain. Indistinguishable results were obtained with these vesicles (21). Vacuoles and cytosol were isolated as described previously (21). In a typical assay, 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 then terminated by centrifugation 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 examined for alkaline phosphatase activity using ␣-naphthyl phosphate as a substrate.

REG1
Is Required for FBPase Degradation-We have established a model for the FBPase degradation pathway that consists of at least two steps. The first step involves the targeting of FBPase to Vid vesicles, whereas in the second step, Vid vesicles traffic to the vacuole for degradation (16). In previous studies, we have identified proteins that are required for each step in the pathway (18 -21). However, none of these proteins appear to exert their functions via a direct interaction with FBPase. To identify FBPase-interacting proteins that may play a role in the FBPase degradation pathway, we screened 64 pools of yeast strains in which each pool contained 96 cells expressing individual open reading frame fused with a GST tag. These cells were grown under glucose starvation conditions and then shifted to glucose containing media for 30 min. The GST fusion proteins were pulled down and examined for the presence of bound FBPase by Western blotting. The strongest positive pools were subjected to further analysis in which individual strains were cultured and processed as described above. Following the screening of pools containing more than 1000 strains, we found that Reg1p, Met25p, and an unknown protein coded by YNR065C showed the strongest FBPase interactions.
To determine whether these putative FBPase-binding proteins play some role in the FBPase degradation pathway, we obtained yeast deletion strains that lacked the genes identified in the above screen. The selected deletion strains, as well as wild type and ⌬vid24 strains, were grown under glucose starvation conditions and examined for their ability to degrade FBPase following a glucose shift (Fig. 1A). As expected, FBPase was degraded in the wild type strain within 2 h of the addition of glucose containing media. In contrast, FBPase degradation was defective in the ⌬vid24 strain. Interestingly, ⌬reg1 exhib-ited a defect in FBPase degradation. By contrast, neither ⌬met25 nor ⌬ynr065C strains was defective in FBPase degradation. We next performed pulse-chase labeling and immunoprecipitation analysis to confirm that FBPase degradation was impaired in the ⌬reg1 mutant. In a wild type strain, FBPase was degraded within a 2 h chase period, whereas FBPase degradation was significantly retarded in the ⌬reg1 strain (Fig. 1B).
REG1 Does Not Prevent FBPase Phosphorylation-Reg1p is known to play an important role in signal transduction pathways (36,37,46). In wild type cells, FBPase is phosphorylated following a shift from low glucose to high glucose containing media via a Ras2p and cAMP-mediated signaling pathway (45). At present, however, it is unknown whether phosphorylation plays a direct role in FBPase trafficking through the Vid vesicle pathway. To test whether the absence of the REG1 gene affects the ability of cells to phosphorylate FBPase in response to glucose, ⌬reg1 cells were metabolically labeled with 32 P and then shifted to glucose-containing media for various periods of time. Cell lysates were subjected to immunoprecipitation analysis to determine the level of FBPase phosphorylation. As controls, we utilized a RAS2 strain and a ras2 318S strain, the later of which has been shown to exhibit a drastic reduction in FBPase phosphorylation (45). As shown in Fig. 2A, the RAS2 strain exhibited a high level of FBPase phosphorylation following a glucose shift, whereas the ras2 318S strain had a much lower level of phosphorylated protein. The ⌬reg1 strain, on the other hand, had levels of phosphorylation that were as high, or somewhat higher than wild type cells (Fig. 2B). Note that the overall levels of FBPase were similar for each of these strains, indicating the differences were not due to changes in the levels of this protein. Because FBPase phosphorylation was not prevented in the absence of Reg1p, Reg1p is unlikely to be involved in the signaling pathway leading to FBPase phosphorylation.
Reg1p Is Not Required for FBPase Import to Vid Vesicles-The trafficking of FBPase to the vacuole requires the presence of functional Vid vesicles. Although Vid vesicle formation is compromised in the absence of the UBC1 gene (17), this is the only gene that is known to be essential for this process. To determine whether Reg1p plays a role in Vid vesicle formation, we performed a differential centrifugation analysis using Vid24p as a specific marker of Vid vesicles. Cells were starved and shifted to glucose-rich media for 30 min, and cell extracts were subjected to differential centrifugation as described previously (16,18). As is shown in Fig. 3A, the levels of Vid24p in the Vid vesicle fraction were reduced in the ⌬ubc1 strain. In contrast, similar amounts of Vid24p were detected in the Vid vesicle fractions in both wild type and ⌬reg1 strains, suggesting that REG1 does not affect Vid vesicle formation.
In wild type cells, FBPase is imported into Vid vesicles following a glucose shift (14 -16). Therefore, we examined whether FBPase import was compromised in the absence of the REG1 gene. Differential centrifugation experiments were performed using lysates obtained from ⌬reg1, ⌬vid24, and ⌬ubc1 cells that had been shifted to glucose (Fig. 3B). A substantial portion of FBPase was found in the Vid vesicle-containing FIG. 1. FBPase degradation requires REG1. A, Reg1p, Met25p, and Ynr065p were identified as potential FBPase binding proteins using a GST library as described under "Experimental Procedures." Wild type, ⌬vid24, ⌬reg1, ⌬met25, and ⌬ynr065c deletion strains were then examined for the degradation of FBPase. B, wild type and ⌬reg1 cells were pulse labeled in poor carbon sources and chased in the presence of glucose for 0 and 2 h as described under "Experimental Procedures." Cell lysates were immunoprecipitated with FBPase antibodies and separated by SDS-PAGE, and the FBPase levels were visualized using a PhosphorImager. fraction of ⌬reg1 and ⌬vid24 cells. By contrast, low levels of FBPase were found in the Vid vesicle fraction of the ⌬ubc1 mutant. Thus, this suggests that FBPase can be imported into Vid vesicles in the absence of Reg1p.
Reg1p Is Involved in the Trafficking of Vid Vesicles to the Vacuole-To determine whether Reg1p plays a role in the second step of the FBPase pathway, we used an in vitro assay to quantitate this process (21). For this assay, FBPase was fused with a truncated form of alkaline phosphatase that lacks the N-terminal 60 amino acids. Vid vesicles containing FBPase-⌬60Pho8p can be isolated and combined with isolated vacuoles and cytosol from cells lacking the PHO8 gene. After incubation, the activation of alkaline phosphatase can be used to quantitate Vid vesicle-vacuole fusion in vitro. Furthermore, this assay can be used to identify the function site of a particular protein, because various combinations of mutant or wild type Vid vesicles, vacuoles, or cytosol can be examined.
The ⌬reg1 strain exhibited a very low level of FBPase-⌬60Pho8p expression following transformation. Because the in vitro assay relies upon high level expression of the fusion protein, we were unable to accurately assess Vid vesicle function in the ⌬reg1 strain. However, we were able to examine the function of ⌬reg1 vacuoles and cytosolic components. For these assays, Vid vesicles were isolated from a wild type strain expressing high levels of FBPase-⌬60Pho8p. When wild type Vid vesicles, vacuoles, and cytosol were tested in the in vitro assay, alkaline phosphatase activity was high, indicating that the fusion protein had been processed by the vacuole Pep4p (Fig.  4A). In contrast, when ⌬reg1 vacuoles were used, alkaline phosphatase activity was reduced, suggesting Reg1p plays a role on the vacuole. However, cytosol from the ⌬reg1 mutant did not reduce the alkaline phosphatase activity significantly.
Because Reg1p exerts its function on the vacuole, we tested whether this protein could be found in the vacuole fraction. Initially, we used Reg1p-GFP and found that the GFP signal was primarily in the cytosol, consistent with a previous report (37). Because cytosolic distribution could mask the fluorescence signal in the vacuole, we used differential centrifugation techniques to separate these compartments (18). Using this procedure, the vacuole marker Vph1p is enriched in the 13,000 ϫ g pellet fraction, whereas the cytosol marker enolase is enriched in the 200,000 ϫ g supernatant fraction. The Vid vesicle marker Vid24p is found primarily in the 100,000 ϫ g pellet with a smaller amount in the 200,000 ϫ g pellet fractions (18). As is shown in Fig. 4B, Reg1p was present in multiple fractions. The soluble fraction S200 contained the highest amounts of Reg1p, whereas a significant amount of Reg1p was found in the vacuole-enriched fraction (P13). A small amount of Reg1p was also present in the Vid vesicle-enriched fraction (P100 and P200). Vid Vesicles and Vacuole Functions Require Glc7p-Reg1p is known to interact with the type I phosphatase Glc7p both genetically and biochemically (34,36). GLC7 is an essential gene that is involved in a number of cellular processes, including glucose repression, glycogen metabolism, translation, sporulation, chromosome segregation, and cell cycle progression (26 -31, 34 -37, 47-54). Furthermore, different Glc7p functions are mediated through its interaction with distinct regulatory subunits (26 -31, 34 -37). To assess whether the interaction of Reg1p with Glc7p plays a role in the trafficking and degradation of FBPase, we used a strain with the glc7-T152K mutation. This mutation has been shown to interfere with the Glc7p-Reg1p interaction, as determined by a yeast two hybrid analysis (36). However, the interaction characteristics under our in vivo conditions have not been examined. To verify this defect, wild type and glc7-T152K strains expressing Reg1p-V5-His6 were glucose-starved and shifted to fresh glucose for 30 min. Reg1-V5-His6 was pulled down with nickel beads, and bound proteins were immunoblotted with anti-V5 and Glc7p antibodies (Fig. 5A). When a wild type strain was examined, we observed the interaction of Reg1p and Glc7p, as reported previously (36). In contrast, the level of this complex was substantially reduced when the same experiments were performed using the glc7-T152K mutant strain. The reduced interaction was not due to a decrease in expression of Reg1p and Glc7p, because similar levels of these proteins were found in total lysates from these strains (Fig. 5B).
We next tested FBPase degradation in the glc7-T152K strain. This strain exhibited a retarded degradation (Fig. 5C), suggesting that the Glc7p-Reg1p complex plays some role in this process. Because different mutations of GLC7 have been shown to affect distinct biological processes (36,43,54), we attempted to determine whether the FBPase degradation defect is allelic specific. For instance, the glc7-T152K mutation inhibits the glucose repression function of the protein, but it does not impair glycogen synthesis (36,54). In contrast, the glc7-1 mutation impairs the binding to Gac1p leading to a defective glycogen synthesis (36,46,54). We found that the glc7-1 strain had minimal defect in FBPase degradation (Fig. 5C).
Reg1p is known to interact with a number of proteins in addition to Glc7p. For example, Snf1p is the yeast homologue of the catalytic subunit of AMP-activated protein kinase and is shown to phosphorylate Reg1p (55). Reg1p also interacts with Snf4p, the activating subunit of the Snf1p kinase complex (55). In this study, neither the ⌬snf1, nor the ⌬snf4 mutation inhibited FBPase degradation to a significant degree (Fig. 5D), suggesting that Snf1p, Snf4p, and/or phosphorylation of Reg1p by the Snf1p complex are not critical for FBPase degradation. Reg1p can also interact with Bmh1p and Sip5 (56). However, FBPase was degraded in strains containing the deletion of these genes (Fig. 5D), suggesting that they are not involved in the FBPase degradation pathway.
If the Glc7p-Reg1 complex is required for FBPase degradation, then Glc7p and Reg1p most likely function in the same step of the FBPase degradation pathway. Therefore, the role of Glc7p in Vid vesicle trafficking was tested using our in vitro fusion assay. The glc7-T152K mutant expressed high levels of FBPase-⌬60Phop8 and allowed us to examine Vid vesicles in this strain. As is shown in Fig. 6A, alkaline phosphatase activity was low, when Vid vesicles or vacuoles from the glc7-T152K mutants were used. In contrast, alkaline phosphatase activity remained high when mutant cytosol was used. This suggests that Glc7p functions predominantly on Vid vesicles, although it is also required for vacuole function to a lesser extent.
If Glc7p functions on the vacuole or Vid vesicles, this protein may associate with these organelles. As is shown in Fig. 6B, the highest amount of Glc7p was found in the cytosolic enriched fraction. Lower amounts of Glc7p were also found in the vacuole-and Vid vesicle-enriched fractions. Because the majority of Reg1p and Glc7p is cytosolic, the localization of Reg1p (Fig. 4B)

FIG. 3. REG1 is not required for Vid vesicle formation and FBPase import into Vid vesicles.
A, wild type, ⌬ubc1, and ⌬reg1 cells were transformed to express Vid24p-HA. Transformed cells were glucose-starved and shifted to glucose for 30 min. Cells were homogenized and subjected to differential centrifugation as described previously (16,18). Proteins from the high speed supernatants (S) and high speed pellets (P) were solubilized in SDS buffer and resolved by SDS-PAGE. The distribution of Vid24p-HA in the S and P fractions was determined using anti-HA antibodies. The % of total Vid24p-HA in each fraction is indicated. B, ⌬vid24, ⌬ubc1, and ⌬reg1 strains were glucose-starved and then shifted to high glucose media for 1 h. Cell lysates were subjected to differential centrifugation and the resultant high speed supernatants and pellets were examined for the distribution of FBPase. and Glc7p in the vacuole and Vid vesicle fraction may result from nonspecific binding of these proteins to these organelles. However, the presence of Glc7p in the vacuolar enriched fraction was consistent with a previous report in which Glc7p was found in this organelle and plays a role in homotypic vacuole fusion (22).
Shp1p and Glc8p Are Necessary for Vid Vesicle and Vacuole Functions-Glc7p is known to regulate a diverse variety of processes, and this is mediated via binding to different regulatory subunits. As mentioned above, Glc7p is targeted to glycogen accumulation processes by interacting with Gac1p and Pig1p (26,27). Reg2p, on the other hand, targets Glc7p to growth and cell cycle progression (28), whereas Gip1p regulates meiosis and sporulation (31). In addition, Bud14p directs Glc7p to cellular morphogenesis (30), whereas Bni4p targets Glc7p to cytokinesis (29). When genes encoding these proteins were deleted, FBPase degradation was not impaired (Fig. 7A). Thus, these Glc7p regulatory proteins are not required for FBPase degradation. In addition to the aforementioned Glc7p-interacting proteins, Shp1p and Glc8p are known to regulate the phosphatase activity of Glc7p (38, 39). Accordingly, mutations or deletions of these genes reduce the phosphatase activity of Glc7p (38,39). When FBPase degradation was examined in the ⌬glc8 and ⌬shp1 strains, it was retarded (Fig. 7B).
The FBPase degradation defects in the ⌬glc8 and ⌬shp1 strains were not due to a decrease in the Glc7p-Reg1p interaction. When Reg1p was immunoprecipitated from wild type, ⌬glc8, and ⌬shp1 strains, the amounts of bound Glc7p were comparable (Fig. 8). Likewise, changes in the subcellular distribution or expression of Reg1p or Glc7p were not responsible for this defect, because these parameters were similar to those seen in wild type cells (data not shown). Therefore, in these mutants, reduced Glc7p phosphatase activity appears to be the cause of the FBPase degradation defect.
We next examined whether the ⌬glc8 and ⌬shp1 mutants affected the same step of the FBPase degradation pathway as Reg1p and Glc7p. Via the use of our in vitro fusion assay, we found that Vid vesicles were severely defective, whereas the vacuole was partially defective in these mutants (Fig. 8, C and  D). These results further confirm that Glc7p and its regulators play an essential role in the trafficking of Vid vesicles to the vacuole. Furthermore, they all exert their functions primarily on Vid vesicles, and to a lesser degree on the vacuole. Vacuoles and cytosol were isolated from either wild type or ⌬reg1 strains lacking the PHO8 gene. Isolated vesicles, vacuoles, and cytosol were incubated in various combinations in the presence of an ATP-regenerating system. The reactions were then terminated and examined for alkaline phosphatase activity. B, wild type cells transformed with a myc-tagged form of Reg1p were subjected to glucose starvation and re-feeding procedures. Cell lysates were fractionated using differential centrifugation techniques described previously (16,18). In these experiments, one-tenth of the total S200 proteins were loaded onto SDS-PAGE. The distribution of Reg1p-myc in S200, P13, P100, and P200 was detected with anti-myc antibodies.
FIG. 5. The glc7-T152K mutation reduces Reg1p-Glc7p interaction and inhibits FBPase degradation. A, the interaction of Reg1p with Glc7p was examined in wild type and glc7-T152K strains expressing Reg1p-V5-His6. Reg1p-V5-His6 was pulled down by nickel beads from both transformed and untransformed cells, and the bound Glc7p was examined by Western blotting with Glc7p antibodies. B, total lysates from these cells were examined for the expression of Reg1p and Glc7p by Western blot. C, the wild type, glc7-T152K, and glc7-1 mutant strains were shifted from low to high glucose media and examined for FBPase degradation. D, the degradation of FBPase was tested in ⌬snf1, ⌬snf4, ⌬bmh1, and ⌬sip5 strains.

DISCUSSION
We have utilized a variety of methods to identify proteins that play a role in the degradation of FBPase. Genetic screens have yielded a number of VID genes that are required for the targeting of FBPase to the vacuole following a shift to glucose rich media (57). Furthermore, via the use of in vivo and in vitro assays, we have been able to identify the site of action of the VID gene protein products (14 -21). However, we have not identified any FBPase-interacting proteins that may play a role in early steps of the FBPase degradation pathway. Therefore, to address this issue, we screened for GST-tagged proteins that might exhibit an interaction with FBPase. Here we hypothesized that FBPase-interacting proteins may regulate FBPase recognition or FBPase sequestration into Vid vesicles. Unfortunately, we were unable to identify any FBPase-binding proteins that carried out these functions. However, we did discover a previously undefined role for the phosphatase Reg1p-Glc7p in the Vid vesicle trafficking pathway.
Although FBPase degradation was compromised in the absence of the REG1 gene, we showed that Reg1p was not involved in any of the early steps of the FBPase degradation pathway. For instance, the absence of REG1 did not affect the ability of cell to phosphorylate FBPase in response to glucose. In addition, Vid vesicle biogenesis was not compromised in cells lacking the REG1 gene, and Vid vesicles were functional in terms of importing FBPase. These results appear to be inconsistent with the idea that there is a functional FBPase-Reg1p interaction in cells. Accordingly, when we attempted to verify this interaction under our in vivo conditions, we did not observe a significant interaction between Reg1p and FBPase in wild type, glc7-T152K, ⌬shp1, or ⌬glc8 strains (data not shown). Little interaction was observed when Reg1p was immunoprecipitated and then blotted with FBPase antibodies or when FBPase was immunoprecipitated and examined for the presence of bound Reg1p. One possibility for this discrepancy may be that the GST-Reg1p fusion protein has an altered conformation that allows artificial binding of FBPase that does not normally occur under in vivo conditions. Alternatively, this may indicate that the interaction is too weak to capture using our standard immunoprecipitation conditions. However, under the same conditions, an interaction between Reg1p and Glc7p was observed.
Reg1p is part of the Glc7p protein phosphatase complex, and this complex plays a required role in a number of cellular processes (28, 32-37, 47, 48, 54). Glc7p has important functions in homotypic vacuole fusion, endoplasmic reticulum-to-Golgi transport, and endocytic vesicular trafficking (22). For homotypic fusion, Glc7p is required for the last fusion step, although the PP1 substrate proteins have not been identified. Following docking of vacuoles, calcium is released from the vacuole and this calcium efflux stimulates calmodulin binding to the vacuole H ϩ -ATPase. This then triggers the formation of a transcomplex on opposing vacuoles (22,58). In this present study, we have demonstrated a requirement for GLC7 in what is most likely a membrane fusion event: Vid vesicle to vacuole trafficking. This conclusion was based upon the use of mutant strains that interfered with Reg1p interactions. However, corroborative results were also obtained with strains in which the Glc7p FIG. 6. The glc7-T152K strain impairs the functions of Vid vesicles and vacuoles. A, Vid vesicles were isolated from wild type or glc7-T152K strains expressing the FBPase-⌬60Pho8p fusion protein, whereas vacuoles and cytosol were isolated from wild type or glc7-T152K strains lacking the PHO8 gene. Isolated Vid vesicles, vacuoles and cytosol were incubated and the reactions were examined for alkaline phosphatase activity. B, the distribution of Glc7p was determined by differential centrifugation and Western blotting. In these experiments, one-tenth of the total S200 proteins were loaded onto SDS-PAGE. regulator genes SHP1 and GLC8 were deleted. The absence of these genes resulted in reduced phosphatase activity of Glc7p and a defect in FBPase degradation and trafficking.
For homotypic vacuole fusion, Glc7p is found in a multisubunit complex containing calmodulin in the vacuole-enriched fraction (22). Furthermore, Glc7p is also present in the vertex that contains Vam3p, Ypt7p, homotypic fusion vacuole protein sorting, and Vac8p (58). Consistent with these results, we observed a portion of Glc7p in the vacuole fraction. Likewise, a small, but detectable amount of Glc7p and Reg1p can also be found in the Vid vesicle fraction. Because the majority of these proteins are in the cytosol, we cannot rule out the possibility that nonspecific binding of Reg1p-Glc7p to these organelles occurred under our conditions. It will be important to address the questions whether Reg1p-Glc7p binding to these compartments is specific and whether stable or transient binding is necessary for Vid vesicle/vacuole fusion. Mutations that interfere with Reg1p or Glc7p binding to these organelles will be useful for these studies in the future.
Whether these mutations affect Glc7p binding to different regulatory subunits has not been established. For this reason, we chose two mutations that are known to interfere with the interactions between Glc7p and two well characterized regulatory subunits, Gac1p and Reg1p. The glc7-1 mutant interferes with Gac1p binding but not Reg1p binding. Interestingly, this mutant strain was not defective in FBPase degradation. In contrast, the glc7-T152K point mutation that reduces the interaction between Glc7p and Reg1p did adversely affect FBPase degradation. Thus, we suspect that Reg1p specifically targets Glc7p for function in the Vid vesicle-vacuole trafficking pathway. When we examined deletion mutant strains of several other Glc7p regulatory subunits, they were not defective in FBPase degradation, suggesting that they are not involved in the FBPase degradation pathway. These results and our previous observations suggest that a specific interaction between Reg1p and Glc7p is important for FBPase degradation to occur.
Vid vesicles are thought to fuse with vacuoles via mechanisms that are similar to other vesicular fusion events. Vid vesicle-vacuole fusion requires the participation of a number of SNARE proteins that are also necessary for homotypic vacuolar fusion (21). However, Vid vesicle/vacuole fusion appears to more closely resemble heterotypic rather than homotypic fusion. A number of v-SNAREs and t-SNAREs are present on both Vid vesicles and vacuoles. However, the v-SNAREs (Ykt6p, Nyv1p, and Vti1p) are required on the vesicles, whereas the t-SNAREs Vam3p only functions on the vacuole. A, strains with deletions of Glc7p regulatory proteins were shifted from low to high glucose media and FBPase degradation was determined. B, wild type, ⌬glc8 and ⌬shp1 strains were examined for their ability to degrade FBPase following a shift from low to high glucose media.