Biochemical analysis of fructose-1,6-bisphosphatase import into vacuole import and degradation vesicles reveals a role for UBC1 in vesicle biogenesis.

When Saccharomyces cerevisiae are shifted from medium containing poor carbon sources to medium containing fresh glucose, the key gluconeogenic enzyme fructose-1,6-bisphosphatase (FBPase) is imported into Vid (vacuole import and degradation) vesicles and then to the vacuole for degradation. Here, we show that FBPase import is independent of vacuole functions and proteasome degradation. However, FBPase import required the ubiquitin-conjugating enzyme Ubc1p. A strain containing a deletion of the UBC1 gene exhibited defective FBPase import. Furthermore, FBPase import was inhibited when cells overexpressed the K48R/K63R ubiquitin mutant that fails to form multiubiquitin chains. The defects in FBPase import seen for the Deltaubc1 and the K48R/K63R mutants were attributed to the Vid vesicle fraction. In the Deltaubc1 mutant, the level of the Vid vesicle-specific marker Vid24p was reduced in the vesicle fraction, suggesting that UBC1 is required for either Vid vesicle production or Vid24p binding to Vid vesicles. However, the K48R/K63R mutant did not prevent Vid24p binding to Vid vesicles, indicating that ubiquitin chain formation is dispensable for Vid24p binding to these structures. Our results support the findings that ubiquitin conjugation and ubiquitin chain formation play important roles in a number of cellular processes including organelle biogenesis.

The vacuole of the yeast Saccharomyces cerevisiae is homologous to the lysosome of higher eucaryotes and as such, plays an important role in protein degradation (1)(2)(3)(4). The function of the vacuole requires the targeting of a number of vacuole resident proteins into this organelle. These proteins are sorted to this organelle by several mechanisms and require the assistance of numerous genes. For example, targeting of the vacuole lumenal protein carboxypeptidase Y (CPY) 1 from the late Golgi requires more than 40 VPS genes (1)(2)(3)(4).
Proteins and organelles can be delivered to the vacuole from the cytoplasm by the microautophagy or macroautophagy path-ways (5)(6)(7)(8)(9)(10)(11). Regulation of the autophagic process can have important consequences on cellular physiology. For example, the tumor suppresser gene beclin-1 is homologous to APG6/ VPS30 and induces autophagy in yeast and mammalian cells. Therefore, a decrease in autophagic protein degradation may contribute to the development or progression of human malignancy (13).
Fructose-1,6-bisphosphatase (FBPase), the key regulatory enzyme in gluconeogenesis in S. cerevisiae, is induced when yeast cells are grown in medium containing poor carbon sources (18). When fresh glucose is added to the medium, however, FBPase is targeted to the vacuole and degraded (19,20). This redistribution of FBPase to the vacuole has been observed by immunofluorescence microscopy, cell fractionation, and electron microscopy (19,20). More recently, FBPase targeting to the vacuole has been reconstituted in vitro using permeabilized yeast cells incubated with purified radiolabeled FBPase in the presence of ATP, an ATP regenerating system and cytosolic proteins (21).
FBPase is imported into a novel type of Vid (vacuole import and degradation) vesicle prior to its uptake by the vacuole (22). These vesicles have been purified to near homogeneity from wild-type cells (22). The identification of Vid vesicles in the FBPase degradation pathway suggests that this pathway can be divided into at least two steps. The first step is the targeting and sequestration of FBPase into Vid vesicles. The second step is the delivery of FBPase from Vid vesicles to the vacuole for degradation.
Since Vid vesicles do not contain markers from known organelles, they may represent a novel transport structure, although it is possible that Vid vesicles are derived from existing structures. Thus far, the heat shock protein Ssa2p is the only molecule that has been shown to play a role in the import of FBPase into Vid vesicles (23). To identify more molecules involved in this process, we analyzed the import of FBPase into Vid vesicles using various inhibitors and mutants. We found that FBPase import was not affected by inhibitors or mutants that block vacuole acidification, vacuole proteolysis, or proteasome degradation. However, FBPase import did require ubiquitin chain formation and the ubiquitin conjugation enzyme Ubc1p. The ⌬ubc1 mutant contained defective vesicles, but competent cytosol. Furthermore, FBPase import was inhibited when cells overexpressed a ubiquitin mutant (K48R/K63R) that prevents the formation of multiubiquitin chains. The defect of the K48R/K63R mutant was associated with Vid vesicles, indicating that ubiquitin chain formation is required to produce competent Vid vesicles.
In the absence of the UBC1 gene, the level of the Vid vesiclespecific marker Vid24p was reduced in the Vid vesicle pellet fraction, suggesting that UBC1 is required for Vid vesicle production. Alternatively, Vid24p binding to Vid vesicles may be compromised in the absence of ubiquitination. However, overproduction of the K48R/K63R mutant did not prevent Vid24p binding to Vid vesicles. Since ubiquitin chain formation is necessary for Vid vesicle function, but is dispensable for Vid24p binding to Vid vesicles, these results are consistent with the hypothesis that Vid vesicle formation is regulated by ubiquitin conjugation and ubiquitin chain formation. Thus, our work complements previous studies in which ubiquitin conjugation is important for peroxisome biogenesis (24), mitochondrial inheritance (25), mitochondrial targeting (26), and receptor-mediated endocytosis (27)(28)(29)(30)(31).

EXPERIMENTAL PROCEDURES
Yeast Strains, Chemicals, and Antibodies-S. cerevisiae strains used in this study are listed in Table I. For the in vitro experiments, the endogenous FBP1 gene was deleted and a known quantity of purified FBPase was added to the reaction. To produce the fbp1 null strain, the FBP1 gene was cloned into pBR322 to yield the plasmid pJS31. The fbp1 deletion construct was generated by removing 90% of the FBP1 gene from pJS31 with StuI and religating with a LEU2 containing fragment which was produced by digestion of the YEP13 plasmid with BglII. The deletion construct was then digested with BamHI and Hin-dIII and transformed into yeast strains using the standard lithium acetate method. The deletion of FBP1 was confirmed by Western blotting with anti-FBPase antibodies.
A pep4 null mutation was produced using the pTS15 plasmid pro-vided by Dr. T. Stevens (University of Oregon). This plasmid was digested with EcoRI and XhoI to disrupt the PEP4 locus (32). The defect in the pep4 null strains was confirmed by the accumulation of the p2 form of CPY intracellularly. A strain with a null mutation of the VID1 gene was also utilized. The VID1 gene is identical to the ISE1 or ERG6 gene. 2 The gene was amplified by polymerase chain reaction and cloned into a TA cloning vector (Invitrogen) using a 5Ј primer AGCGGC-CGCGGGATGGGGAGTGAAACAGAATTGAGAAAA and a 3Ј primer TGAGGCGGCCGCCTTGAGTTGCTTCTTGGGAAGTTTGGG. A deletion construct was produced by removing 80% of the gene via KpnI and PflMI digestion, and religating with a URA3 containing fragment produced by digesting the YIP352 plasmid with SmaI and HpaI. The resultant construct was linearized with NotI and transformed into a wild type strain. The deletion was confirmed by polymerase chain reaction analysis. The pUB141 plasmid containing the wild type Myctagged ubiquitin, the pUB223 plasmid containing the Myc-tagged K48R/K63R ubiquitin mutant and the Ub-Pro-␤gal plasmid (33,34) were obtained from Dr. D. Finley (Harvard Medical School). YPD is a complete medium (10 g/liter of Bacto-yeast extract, 20 g/liter of Bacto-peptone, Difco Labs Inc.) supplemented with 20 g/liter dextrose (Fisher Scientific). YPKG contained 10 g/liter Bacto-yeast extract, 20 g/liter Bacto-peptone, 10 g/liter potassium acetate, and 5 g/liter dextrose. Synthetic minimal medium consisted of 6.7 g/liter yeast nitrogen base without amino acids, supplemented with 5 g/liter casamino acids, 40 mg/liter adenine, 60 mg/liter leucine, and 20 g/liter dextrose. Inhibitors used in this study included ATP␥S, N-ethylmaleimide, brefeldin A, bafilomycin A, and concanamycin A and were purchased from Sigma. MG132 (carbobenzoxyl-leucinyl-leucinyl-leucinal) and ␤-lactone were gifts from Dr. A. Goldberg (Harvard Medical School). Tran 35 S-label (10 mCi/mmol) was obtained from ICN. Rabbit anti-FBPase and rabbit anti-CPY polyclonal antibodies were raised by Berkeley Antibody Co. (Berkeley, CA) using purified FBPase and CPY (Sigma). Mouse and rabbit anti-Myc antibodies were purchased from Berkeley Antibody Co. Mouse anti-␤-galactosidase antibodies were purchased from Promega.
The FBPase Import Assay-The FBPase import assay was performed according to Shieh and Chiang (21). In a typical experiment, the reaction mixture (100 l) contained 3 A 600 nm units of semi-intact cells, 11 g of 35 S-FBPase, an ATP regenerating system (0.5 mM ATP, 0.2 mg/ml creatine phosphokinase, 40 mM creatine phosphate), and 0.5 mg/ml cytosolic proteins. The mixture was incubated at 30°C for the indicated times, after which 0.8 mg/ml proteinase K was added to identify the Miscellaneous Assays-Isolation of Vid vesicles by differential centrifugation was performed as described (23). Briefly, total lysates were subjected to differential centrifugation at 13,000 ϫ g for 20 min and the supernatant was further centrifuged at 200,000 ϫ g for 2 h. The distribution of Vid24p in the high speed pellet (200,000 ϫ g pellet) and the high speed supernatant (200,000 ϫ g supernatant) was determined by Western blotting with anti-Vid24p antibodies. The biosynthesis of CPY was studied using the protocol described by Graham et al. (35). The exponentially grown ⌬ise1, ⌬ise1⌬pep4, ⌬vid24, and ⌬vid24⌬pep4 strains were labeled with Tran 35 S-label for 10 min at 30°C and then chased for 40 min at 30°C. To examine the effect of brefeldin A on CPY processing, an ise1 strain was preincubated in the presence or absence of brefeldin A (75 g/ml) at 22°C for 10 min. Cells were pulsed for 10 min, chased for 40 min, and then harvested. Total lysates were immunoprecipitated with CPY antiserum, subjected to SDS-PAGE using 7.5% polyacrylamide gels, and analyzed with a Fuji Bioimaging Analyzer. The degradation of short-lived and long-lived proteins was examined using the protocols described by Lee and Goldberg (36).

RESULTS
FBPase Import in Vitro-To biochemically analyze FBPase import into Vid vesicles, we used an in vitro system that reproduces the defects seen for mutants affecting the FBPase degradation pathway. For example, both the ⌬vid1 (⌬ise1) and ⌬vid24 mutants inhibit the degradation of FBPase in vivo (Fig.  1A). However, these mutations affect different steps in the FBPase degradation pathway. The ⌬vid24 mutant strain imports FBPase into Vid vesicles normally, but this mutation blocks the trafficking of Vid vesicles to the vacuole. As such, this mutation results in the accumulation of FBPase in Vid vesicles (37). On the other hand, a mutation of the VID1 gene (a gene that is identical to the ISE1 or ERG6 gene) 2 blocks FBPase import into Vid vesicles (38) and serves as a negative control for in vitro import.
To examine FBPase import in the ⌬ise1 and ⌬vid24 strains, the endogenous FBP1 gene was deleted so that a known quantity of radiolabeled, purified FBPase could be added and followed in the in vitro system. Each strain was glucose starved and then shifted to glucose containing medium prior to their conversion to semi-intact cells. Purified FBPase was incubated with semi-intact cells in the absence or presence of ATP, an ATP regenerating system and cytosol. At selected times, proteinase K was added to digest the FBPase that was not protected in a membrane-sealed compartment. In the absence of both ATP and cytosol, FBPase import into the ⌬vid24 semiintact cells was minimal (Fig. 1B). In the presence of ATP and cytosol, however, FBPase import increased in a time-dependent manner. When quantitated, ϳ25-35% of the total added FB-Pase was proteinase K protected after 30 min of import. In contrast, the ⌬ise1 mutant had background levels of FBPase import either in the presence or in the absence of ATP and cytosol (Fig. 1B).
The ⌬ise1 Mutant Contains Defective Vesicles-The defect of FBPase import seen for the ⌬ise1 mutant could result from an inability of cytosol to stimulate FBPase import or an inability of Vid vesicles to take up FBPase. To determine the site of this defect, we performed an in vitro assay using various combinations of semi-intact cells and cytosol from the ⌬ise1 and ⌬vid24 mutants. When the ⌬ise1 semi-intact cells were used, FBPase import was defective regardless of whether the cytosol was isolated from the ⌬ise1 (Fig. 2, lane 1) or the ⌬vid24 mutants (lane 3). By contrast, FBPase import into the ⌬vid24 semiintact cells was observed when cytosol was prepared from either the ⌬vid24 mutant (lane 2) or the ⌬ise1 mutant (lane 4). This experiment suggests that the ⌬ise1 mutant strain has competent cytosol that can stimulate FBPase import into competent Vid vesicles. However, the ⌬ise1 mutant contains defective vesicles that cannot support FBPase import, even when combined with import-competent cytosol.
FBPase Import Is Independent of Vacuole Proteolysis and Vacuole Acidification-Next, we utilized our in vitro assay to investigate whether FBPase import into Vid vesicles was dependent on other cellular processes such as vacuole proteolysis or vacuole acidification. The PEP4 gene is required for the maturation of several major vacuolar proteinases including CPY. Hence, the deletion of the PEP4 gene renders cells defective in vacuolar proteolysis (1,3). In wild type cells, CPY is synthesized as prepro-CPY and then translocated into the endoplasmic reticulum where it is glycosylated to p1-CPY in the endoplasmic reticulum (1)(2)(3)(4). CPY is further modified in the Golgi to p2-CPY and finally processed to the mature form in the vacuole (1)(2)(3)(4). Therefore, the deletion of the PEP4 gene resulted in the accumulation of p2-CPY in the ⌬ise1⌬pep4 and ⌬vid24⌬pep4 strains (Fig. 3A). When FBPase import was measured, the level was low in the ⌬ise1 single mutant (Fig.  3B, lane 1) and there was no significant increase in the FBPase import in the ⌬ise1⌬pep4 double mutant (lane 2). Likewise, there was no significant change in FBPase import in the FIG. 1. The kinetics of FBPase import into the ⌬ise1 and ⌬vid24 semi-intact cells. A, wild type (HLY223), ⌬ise1 (HLY001), and ⌬vid24 (HLY227) were grown in YPKG to induce FBPase. Cells were shifted to glucose for 0, 60, and 120 min. Total lysates from these cells were solubilized in SDS buffer, separated by SDS-PAGE and FBPase degradation was followed in these cells. B, both ⌬ise1 (HLY208) and ⌬vid24 (HLY232) mutants were shifted to glucose for 20 min. Semiintact cells and cytosol were prepared as described (21). FBPase import was measured for 0, 10, 20, and 30 min in the absence or presence of ATP and cytosol. The % FBPase import is indicated.
⌬vid24⌬pep4 double mutant (lane 4) as compared with the ⌬vid24 single mutant (lane 3). Since uptake of FBPase by Vid vesicles is independent of the PEP4 gene, this supports our model that FBPase import into Vid vesicles occurs prior to trafficking to the vacuole.
As is shown in Fig. 1, the addition of ATP and cytosol stimulates FBPase import into Vid vesicles. This suggests that ATPases and/or ATP hydrolysis (see below) may play some role in FBPase import. The VMA3 gene, which encodes the 16-kDa proteolipid subunit of the membrane sector of the V-ATPase (1,39), has previously been shown to play a role in autophagy (40). However, when FBPase import was measured in the vma3 deletion mutant, there was no significant defect (Fig. 3B, lane  5). Therefore, V-ATPase is not essential for FBPase import into Vid vesicles.
FBPase Import Requires the UBC1 Gene-Ubiquitination plays an important role in distinct biological functions including DNA repair, protein degradation, organelle biogenesis, and protein trafficking (41,42). For example, the ubiquitin protein ligase Rsp5p is essential for mitochondrial inheritance and mitochondrial import (25,26). Rsp5p is also involved in receptor-mediated internalization of Ste2p, Ste3p, and other cell surface proteins (31). In addition, the ubiquitin-conjugating enzyme Ubc10p plays a critical role in peroxisomal biogenesis (24). Ubc10p is one of 13 ubiquitin-conjugating enzymes found in yeast (41,42). UBC1, UBC4, and UBC5 are functionally overlapping and are involved in degrading abnormal or shortlived proteins (43,44). As expected, the ⌬ubc1 strain displayed a reduced rate of degradation of short-lived proteins as compared with the wild type control (Fig. 4A). In contrast, UBC6 and UBC7 are involved in the ubiquitination of misfolded or unassembled proteins in the endoplasmic reticulum degradation pathway (45,46). Therefore, ⌬ubc6 and ⌬ubc7 strains did not inhibit the degradation of short-lived proteins (Fig. 4A). When the ⌬ubc1, ⌬ubc6, and ⌬ubc7 strains were tested for FBPase import, a reduced level of import was observed for ⌬ubc1 (Fig. 4B, lane 1), but not for the ⌬ubc6 and ⌬ubc7 strains (lanes 2 and 3), suggesting a specific role for UBC1 in the import process.
One of the major functions of ubiquitin conjugation is to target proteins for degradation by the proteasome (41,42). However, ubiquitin conjugation can also have other important functions unrelated to protein degradation (24 -31, 41, 42). We investigated whether the proteasome plays a role in FBPase import using the pre1-1pre2-1 proteasome mutant. PRE1 and PRE2 encode subunits of the 20 S core particle of the proteasome and an interaction between Pre1p and Pre2p is necessary for formation of the chymotrypsin-like active site in the proteasome (47,48). A decrease in the degradation rate of short-lived proteins was observed for the pre1-1pre2-1 mutant strain (Fig.  4A). However, the import of FBPase in the pre1-1pre2-1 mutant was not altered (Fig. 4B, lane 4). Thus, the proteasome is unlikely to be involved in the import process.
Inhibitor Studies-We next investigated whether FBPase import was dependent upon vacuole acidification or proteasome degradation using inhibitors that block these processes (Fig. 5). For these experiments, ⌬vid24 semi-intact cells and cytosol were preincubated with various concentrations of inhibitors. These concentrations were chosen based upon previous studies demonstrating maximal inhibition in the yeast system (35, 36, 49 -51). FBPase, ATP, and an ATP regenerating system were then added to the reaction mixture to commence the import process. The in vitro import of FBPase was inhibited by nonhydrolyzable ATP␥S (Fig. 5A, lane 3). However, N-ethylmaleimide, which inhibits V-ATPase (1) did not affect FBPase import in vitro (lane 4). Likewise, brefeldin A had no effect on in vitro FBPase import (lane 5), even though this inhibitor caused accumulation of p1-CPY in the ise1 (brefeldin A permeable) strain (Fig. 5B, lane 2). FBPase import was also unaffected by the proteasome inhibitors MG132 or ␤-lactone (Fig.  5A, lanes 6 and 7), although these inhibitors did reduce the degradation of short-lived proteins in vivo (Fig. 5C). Inhibitors that perturb vacuole acidification such as bafilomycin A and concanamycin A (1, 50, 51) also had no effect on FBPase import (Fig. 5A, lanes 8 and 9), but they did reduce the degradation of long-lived proteins in vivo (Fig. 5D). Taken together, the mutant analyses and the inhibitor studies suggest that FBPase import into Vid vesicles is independent of vacuole proteolysis, vacuole acidification, and proteasome degradation. However,

FIG. 4. FBPase import into Vid vesicles is defective in the
UBC1 Is Necessary for FBPase Import-Since the ⌬ubc1 strain displayed defective FBPase import in vitro, we next determined whether this strain was also defective in FBPase degradation in vivo. As is shown in Fig. 6A, wild type cells degraded FBPase after a shift to glucose for 180 min. In contrast, FBPase degradation was significantly retarded in the ⌬ubc1 mutant, but was normal in the ⌬ubc6 mutant. Therefore, UBC1 is required for FBPase degradation, whereas UBC6 is not.
We next examined whether the defect in FBPase import observed for the ⌬ubc1 mutant resulted from an inability of cytosol to support FBPase import or an inability of Vid vesicles to take up FBPase. As is shown in Fig. 6B, when cytosol and semi-intact cells from the ⌬ubc1 strain were used, FBPase import was impaired (lane 1). By contrast, when cytosol and semi-intact cells from the ⌬vid24 strain were combined, a high level of FBPase import was observed (lane 2). FBPase import decreased when ⌬ubc1 semi-intact cells were incubated with cytosol from the ⌬vid24 strain (lane 3). Since the ⌬vid24 strain contained import competent cytosol, this result indicates that the ⌬ubc1 mutant had defective vesicles. In contrast, the ⌬ubc1 strain appears to contain competent cytosol, because cytosol from the ⌬ubc1 strain supported FBPase import into import competent Vid vesicles in ⌬vid24 semi-intact cells (lane 4).
The impaired ability of the ⌬ubc1 semi-intact cells to import FBPase could be due to a decrease in Vid vesicle production. Alternatively, the reduced import could result from a defect in the import machinery. In initial experiments, we examined the levels of the Vid vesicle specific marker, Vid24p. Vid24p is induced in response to glucose and a significant portion of this protein is associated with Vid vesicles as a peripheral protein (37). When cells were maintained in low glucose medium (t ϭ 0 min), Vid24p was undetectable in total lysates. However, this protein was induced to a similar level after wild type, ⌬ubc1, and ⌬ubc6 strains were shifted to glucose for 20 min (Fig. 7A). Therefore, Vid24p production is not altered in the ⌬ubc1 mutant.
If UBC1 is required for Vid vesicle formation, the number of Vid vesicles should be reduced in the ⌬ubc1 mutant. This would be reflected as a decreased level of Vid24p within fractions that contain Vid vesicles. Conversely, if UBC1 is required for the function of the import machinery, the level of Vid24p would not be altered in the Vid vesicle containing fractions. To test these possibilities, the wild type, ⌬ubc1, and ⌬ubc6 strains were shifted to glucose and cell extracts were subjected to differential centrifugation using the protocol described previously (23). In wild type and ⌬ubc6 mutant cells, most of the Vid24p was in the Vid vesicle containing pellet fraction (Fig. 7B). By contrast, the ⌬ubc1 mutant exhibited a significantly decreased level of Vid24p in the pellet fraction, but a greater concentration of Vid24p in the soluble fraction (Fig. 7B). The decreased level of Vid24p in the pellet fraction most likely represents a reduced production of Vid vesicles, since Vid24p induction is not altered in the ⌬ubc1 strain. However, a decreased binding of Vid24p to Vid vesicles in the ⌬ubc1 strain could also account for this observation.
The K48R/K63R Ubiquitin Mutant Inhibits FBPase Degradation-Ubiquitin molecules are most often linked to one another by isopeptide bonds between the carboxyl terminus of one ubiquitin and the ⑀-amino group of lysine 48 of the next ubiquitin (41,42). However, ubiquitin chains can also be formed at lysine 63 (41,42). Therefore, when both lysine 48 and lysine 63 are replaced with arginine (K48R/K63R), the formation of multiubiquitin chains is inhibited. To study the effect of ubiquitin chain formation on FBPase degradation, a strain overexpressing the K48R/K63R mutation was used. When wild type ubiquitin was overproduced, FBPase was degraded in response to glucose in vivo (Fig. 8A). However, when the K48R/K63R ubiquitin mutant was overexpressed, FBPase degradation was impaired (Fig. 8A). Therefore, the degradation of FBPase requires the formation of multiubiquitin chains.
FBPase Import into Vid Vesicles Is Inhibited by the K48R/ K63R Ubiquitin Mutant-We investigated whether ubiquitin chain formation is necessary for FBPase import in vitro. FB-Pase was imported when cytosol and semi-intact cells were were added to semi-intact cells and cytosol for 20 min before the addition of FBPase, ATP, and an ATP regenerating system. FBPase import was measured as described. The percentage of FBPase import in semi-intact cells treated with various inhibitors is indicated. B, the addition of brefeldin A caused p1-CPY to accumulate in the ise1 strain (lane 2). C, the degradation of short-lived proteins was inhibited by MG132 and ␤-lactone. D, inhibitors that perturb the acidification of the vacuole reduced the degradation of long-lived proteins.
prepared from the wild type strain overexpressing wild type ubiquitin (Fig. 8B, lane 1). By contrast, in vitro FBPase import was significantly reduced when both cytosol and semi-intact cells were prepared from the strain that overproduced the K48R/K63R mutant (lane 2). When cytosol from the K48R/ K63R strain was incubated with semi-intact cells from the strain overexpressing wild type ubiquitin, a high level of FB-Pase was imported (Fig. 8B, lane 3). By contrast, FBPase import decreased when cytosol from the strain overexpressing wild type ubiquitin was incubated with semi-intact cells from the K48R/K63R strain (lane 4). Therefore, the K48R/K63R mutant inhibits the function of Vid vesicles to import FBPase, but does not affect the ability of cytosol to stimulate FBPase import into competent Vid vesicles.
The K48R/K63R Mutant Does Not Prevent Vid24p Binding to Vid Vesicles-As mentioned above, the decreased level of Vid24p in the ⌬ubc1 high speed pellet may result from a reduced number of Vid vesicles, or it may be due to a decreased binding of this protein to Vid vesicles. Accordingly, if ubiquitin chain formation is necessary for Vid24p binding to Vid vesicles, the distribution of Vid24p might be altered when the K48R/ K63R mutant was overproduced. When Vid24p was induced in cells overexpressing wild type ubiquitin, most of the Vid24p was in the pellet fraction and very little was in the supernatant fraction (Fig. 8C). However, in cells overproducing the K48R/ K63R mutant, the level of Vid24p decreased to one-third of that observed in cells overexpressing wild type ubiquitin (Fig. 8C). It is unknown why the K48R/K63R mutant reduced total amounts of Vid24p. However, this was not due to an overall decrease in protein concentration, because both wild type and K48R/K63R strains had similar protein concentrations in total lysates as well as in individual supernatant (9.92 versus 8.88 mg/ml) and pellet (4.16 versus 4.61 mg/ml) fractions. When Vid24p distribution was quantitated in the K48R/K63R mutant, more than 90% of the Vid24p was in the pellet fraction and less than 10% was in the soluble fraction. Thus, the ratio of bound versus unbound Vid24p was not altered when the K48R/ K63R mutant was overproduced. Given that the association of Vid24p with Vid vesicles was not prevented by the K48R/K63R mutant, polyubiquitination is not required for Vid24p binding to the Vid vesicles. Therefore, these data are consistent with the hypothesis that the ⌬ubc1 and K48R/K63R mutations result in a decreased production of Vid vesicles.
We next examined whether Vid24p was ubiquitinated by transforming wild type and ⌬vid24 strain with or without the Myc-tagged wild type ubiquitin plasmid. These strains were incubated in glucose poor medium containing copper to induce Myc ubiquitin and FBPase. Cells were then shifted to glucose for 20 min to induce Vid24p. Ub-Pro-␤-galactosidase was used as a positive control since Ub-Pro-␤ -galactosidase is known to be polyubiquitinated constitutively (34). Immunoblotting experiments indicate that high levels of Myc ubiquitin were expressed in cells transformed with the Myc ubiquitin plasmid, but not in cells that did not harbor the Myc ubiquitin plasmid (Fig. 9A, lanes 1-4). As shown by immunoblotting and immunoprecipitation experiments, Ub-Pro-␤-galactosidase was present as multiple bands in cells transformed with the Ub-Pro-␤galactosidase plasmid (lanes 5, 6, 9, and 10). However, these bands were not observed in control cells that did not contain the Ub-Pro-␤-galactosidase plasmid (lanes 7, 8, 11, and 12). In cells transformed with both Ub-Pro-␤-galactosidase and Myc ubiquitin plasmids, multiple Ub-Pro-␤-galactosidase bands were detected by anti-Myc antibodies, suggesting that these bands were polyubiqutinated forms of Ub-Pro-␤-galactosidase (lane 14). By contrast, no Myc signal could be found in cells that did not harbor the Myc ubiquitin plasmid (lane 13) or in cells FIG. 6. The ⌬ubc1 mutant contains defective vesicles, but normal cytosol. A, FBPase degradation was followed in wild type, ⌬ubc1 and ⌬ubc6 cells for 0, 45, 90, 120, and 180 min. B, both ⌬vid24 (HLY232) and ⌬ubc1 (HLY212) were shifted to glucose for 20 min. Semiintact cells and cytosol were prepared from the glucose-shifted ⌬vid24 and ⌬ubc1 strains. Lane 1, FBPase import into ⌬ubc1 semi-intact cells with cytosol from the ⌬ubc1 strain. Lane 2, FBPase import into ⌬vid24 semi-intact cells with ⌬vid24 cytosol. Lane 3, FBPase import into ⌬ubc1 semi-intact cells with ⌬vid24 cytosol. Lane 4, FBPase import into ⌬vid24 semi-intact cells with ⌬ubc1 cytosol.
FIG. 7. Vid vesicle function is impaired in the ⌬ubc1 mutant. A, wild type (HLY223), ⌬ubc1, and ⌬ubc6 strains were glucose starved (t ϭ 0), or glucose starved and then shifted to glucose for 20 min (t ϭ 20). Total lysates from t ϭ 0 and t ϭ 20 were separated by SDS-PAGE and Vid24p was detected by Western blotting with Vid24p antibodies. B, wild type, ⌬ubc1, and ⌬ubc6 strains were shifted to glucose for 20 min. Cells were homogenized and subjected to differential centrifugation. Proteins from the high speed supernatant (S) and high speed pellet (P) were solubilized in SDS buffer and resolved by SDS-PAGE. The distribution of Vid24p in the S and P fractions was detected by anti-Vid24p antibodies. The lower panel indicates the % recovery of Vid24p in each fraction from these strains. that did not contain the Ub-Pro-␤-galactosidase plasmid ( lanes  15 and 16).
To determine whether Vid24p was ubiquitinated, this protein was immunoprecipitated from total lysates of wild type cells and then immunoblotted with anti-Myc antibodies. Vid24p was expressed in wild type cells, but was absent in the ⌬vid24 strain, as indicated by immunoblotting (Fig. 9B, lanes [5][6][7][8] and immunoprecipitation experiments (lanes 9 -12). When the precipitated Vid24p was immunoblotted with anti-Myc antibodies, there was no detectable Myc signal (lane 14). Likewise, no ubiquitination of Vid24p could be found in cells that did not contain the Myc ubiquitin plasmid (lane 13), or in the ⌬vid24 strain (lanes 15 and 16). Furthermore, no ubiquitination of Vid24p could be detected using pulse-chase experiments followed by immunoprecipitation with anti-Vid24p antibodies (data not shown). Therefore, Vid24p is unlikely to be ubiquitinated. This supports our contention that ubiquitination is not required for the function of Vid24p.
Although the site of FBPase degradation has been a matter of debate (52,53), a PEP4-dependent degradation of FBPase was confirmed by an independent research group (54). To examine whether FBPase was polyubiquitinated, wild type and ⌬fbp1 strains were transformed with or without the Myctagged ubiquitin plasmid using the protocol described by the Wolf group (52,53). FBPase was reported to be polyubiquitinated under these conditions (52,53). As shown by both immunoblotting (Fig. 9C, lanes 5-8) and immunoprecipitation (lanes 9 -12) experiments, FBPase was detected in wild type cells, but not in the ⌬fbp1 strain. When the precipitated FBPase was immunoblotted with anti-Myc antibodies, some faint bands migrating below the IgG band were detected in wild type cells transformed with the Myc ubiquitin plasmid (lane 14). However, these bands were also seen in cells that did not harbor the Myc ubiquitin plasmid (lane 13) as well as in the ⌬fbp1 strain that did not have the FBP1 gene (lanes 15 and 16). Thus, these bands were unlikely to represent polyubiquitinated FBPase. Similarly, no polyubiquitination of FBPase was detected when wild type cells were pulsed-chased and then immunoprecipitated with anti-FBPase antibodies (data not shown). Since purified FBPase without ubiquitination was imported into Vid vesicles in vitro, polyubiquitination of FBPase is unlikely to be required for the import process. DISCUSSION In this study, we analyzed FBPase import into Vid vesicles to identify molecules involved in early stages of the FBPase degradation pathway. Our results suggest that vacuole proteolysis, vacuole acidification, and proteasome degradation are unlikely to be involved in FBPase import. The ⌬ise1⌬pep4 or ⌬vid24⌬pep4 double mutants did not alter FBPase import as compared with the ⌬ise1 or ⌬vid24 single mutants, suggesting that FBPase import is independent of the PEP4 gene. Furthermore, the vma3 deletion mutant and compounds such as bafilomycin A or concanamycin A that block acidification of the vacuole did not inhibit FBPase import. Therefore, FBPase import is independent of the two major vacuole functions, vacuole proteolysis and vacuole acidification. This further supports our model that FBPase import into Vid vesicles occurs prior to the trafficking to the vacuole.
Our results indicate that the cytosolic ubiquitin-conjugating enzyme Ubc1p is an important regulator of the FBPase import process. FBPase import into Vid vesicles is defective in the ⌬ubc1 mutant, but not in the ⌬ubc6 or ⌬ubc7 mutants, suggesting a specific role for UBC1 in the import process. However, this requirement is not linked to proteasome degradation. The pre1-1pre2-1 proteasome mutant showed normal FBPase import and proteasome inhibitors such as MG132 and ␤-lactone had no effect on FBPase degradation.

FIG. 8. FBPase import is impaired when ubiquitin chain formation is inhibited.
A, wild type cells were transformed with multicopy plasmids containing either the wild type ubiquitin (HLY824) or the K48R/K63R ubiquitin mutant (HLY823) under an inducible copper promoter. The transformants were grown in synthetic medium and ubiquitin was induced by 100 mM CuSO 4 using the protocol described by Schork et al. (52). These cells were then shifted to glucose for the indicated times and FBPase degradation was examined. B, wild type cells overexpressing either wild type ubiquitin (HLY820) or the K48R/K63R ubiquitin mutant (HLY819) were shifted to glucose for 20 min. Cytosol and semi-intact cells from these strains were combined as indicated and in vitro import of FBPase was performed as described under "Experimental Procedures." C, total lysates from wild type cells over-expressing either wild type ubiquitin (HLY824) or the K48R/ K63R ubiquitin mutant (HLY823) were fractionated by differential centrifugation. The distribution of Vid24p in total (T), high speed pellet (P), and high speed supernatant (S) fractions was examined by Western blotting with anti-Vid24p antibodies.
Our data show that UBC1 is required for the proper function of Vid vesicles. The ⌬ubc1 mutant contained defective vesicles, but normal cytosol. In the absence of the UBC1 gene, cells may decrease the production of Vid vesicles or reduce the efficiency of the import machinery. In the control wild type and ⌬ubc6 strains, most of the Vid vesicle marker Vid24p was found in fractions containing Vid vesicles. When quantitated, ϳ90% of the Vid24p was recovered in the Vid vesicle containing pellet fraction in wild type cells. However, in the ⌬ubc1 strain, about 25% of the Vid24p was in the pellet fraction, while most of the Vid24p was in the soluble fraction. The reduced levels of Vid24p in the pellet fraction could result from decreased Vid vesicle production or a decreased binding of Vid24p to Vid vesicles. However, the K48R/K63R mutant did not prevent Vid24p binding to Vid vesicles, even though it inhibited vesicle import. Therefore, polyubiquitination is necessary for FBPase import into Vid vesicles, but does not play an important role in Vid24p binding to Vid vesicles.
Based upon results from this study and from previous studies (23,37), we have proposed a model for the FBPase degradation pathway (Fig. 10). In the initial step, FBPase is imported into Vid vesicles through a process that requires the presence of the heat shock protein Ssa2p. Following FBPase sequestration inside these structures, the loaded vesicles then traffic to the vacuole via a process controlled by Vid24p. At   FIG. 9. Vid24p is not ubiquitinated. A, wild type cells were transformed with or without a multicopy Ub-Pro-␤-galactosidase plasmid. These cells were then transformed with or without a multicopy Myc ubiquitin plasmid. Cells were grown in synthetic medium and then shifted to synthetic medium containing 2% ethanol and 100 mM CuSO 4 for 5 h to induce FBPase and ubiquitin. Cells were transferred to synthetic medium containing fresh 2% glucose for 20 min. Total lysates from these cells were aliquoted into two parts. One-half of the lysates were immunoblotted with anti-Myc antibodies (lanes 1-4), or anti-␤-galactosidase antibodies (lanes [5][6][7][8]. Another half of the total lysates were immunoprecipitated first with anti-␤-galactosidase antibodies and then immunoblotted with either anti-␤-galactosidase antibodies (lanes 9 -12) or with anti-Myc antibodies (lanes 13-16). B, wild type or ⌬vid24 strains were transformed with or without a multicopy Myc ubiquitin plasmid under a copper inducible promoter. Total lysates were aliquoted into two portions. Half of the lysates were subjected to immunoblotting with anti-Myc antibodies (lanes 1-4) or anti-Vid24p antibodies (lanes [5][6][7][8]. The other half of the total lysates were subjected to immunoprecipitation with anti-Vid24p antibodies and then immunoblotted with either Vid24p antibodies (lanes 9 -12) or anti-Myc antibodies (lanes [13][14][15][16]. C, wild type and ⌬fbp1 strains were transformed with or without a multicopy Myc ubiquitin plasmid. Half of the total lysates were immunoblotted with anti-Myc antibodies (lanes [1][2][3][4] or FBPase antibodies (lanes [5][6][7][8]. Another half of the lysates were immunoprecipitated with FBPase antibodies and then immunoblotted with anti-FBPase antibodies (lanes 9 -12) or anti-Myc antibodies (lanes 13-16).
FIG. 10. The FBPase degradation pathway. When glucosestarved cells are shifted to medium containing fresh glucose, FBPase is imported into Vid vesicles and then to the vacuole for degradation. The cytosolic heat shock protein Ssa2p is required for FBPase import into Vid vesicles. After FBPase is sequestered inside the vesicles, Vid vesicles then carry FBPase to the vacuole in a process that is dependent upon Vid24p. Although the origin of Vid vesicles is not known, the formation of Vid vesicles is regulated by the cytosolic ubiquitin conjugating enzyme Ubc1p through unidentified factors that are likely to be polyubiquinated. present, the site of origin for Vid vesicles is unknown, although the formation of these organelles appears to be regulated by the cytosolic ubiquitin-conjugating enzyme Ubc1p. In the absence of this enzyme, levels of Vid vesicles are reduced and FBPase degradation is compromised. Therefore, we propose that ubiquitination plays an important role in the degradation of FB-Pase through its effect on the machinery (Vid vesicles) that transports FBPase to the vacuole. Identification of the factors that are polyubiquitinated by Ubc1p may ultimately help identify their sites of action as well as to elucidate how FBPase is imported into Vid vesicles.