An essential domain within Cdc34p is required for binding to a complex containing Cdc4p and Cdc53p in Saccharomyces cerevisiae.

The CDC34 gene of the yeast Saccharomyces cerevisiae encodes a ubiquitin-conjugating protein that transfers ubiquitin onto substrates to signal rapid degradation via the proteasome. Cdc34p has been implicated in signaling the destruction of a variety of substrates including the cyclin-dependent kinase inhibitor, Sic1p, which must be degraded for cells to enter S-phase. Mutants lacking CDC34 activity fail to degrade Sic1p and fail to enter S-phase, a phenotype that is also shared with cells lacking CDC4 and CDC53 activity. Here we demonstrate that Cdc4p, Cdc34p, and Cdc53p interact in vivo. We have mapped a Cdc4p/Cdc53p-binding region on Cdc34p; this region is essential for S-phase entry and thus the association of these three proteins is required for Sic1p degradation. All three proteins migrate in gel filtration to sizes that greatly exceed their actual size suggesting that they form stable associations with other proteins and we observe Cdc4p, Cdc34p, and Cdc53p fractionating into overlapping families of high molecular weight complexes. Finally, we demonstrate that Cdc4p, Cdc34p, and Cdc53p are stable throughout the cell cycle and that Cdc34p permanently resides as part of a complex throughout the cell cycle. This suggests that all Cdc34p substrates are ubiquitinated by a similar high molecular weight complex.

Ubiquitin (Ub) 1 is a highly conserved 76-amino acid protein that is found as a post-translational modification of other proteins. The best characterized role of Ub modification is as a means of regulating protein abundance. When a Ub chain forms on a particular substrate protein, that protein is rapidly degraded by the proteasome (for reviews, see Refs. [1][2][3][4][5][6]. Poly-Ub chain formation on a substrate is catalyzed by a member of the ubiquitin conjugating (Ubc) family of proteins, also known as E2 enzymes. Prior to poly-Ub chain formation, Ubc (E2) enzymes are charged with ubiquitin by a ubiquitin activating, or E1, enzyme which has the ability to capture free Ub. In some cases, the transfer of ubiquitin from an Ubc (E2) enzyme to a substrate proceeds via a third protein known as a ubiquitin ligase, or E3 (7,8). Based on sequence similarity, there are 13 members of the Ubc (E2) family in yeast and these proteins have diverse roles including cell cycle progression, heavy metal resistance, peroxisome biogenesis, and DNA repair. Members of this family have highly related catalytic domains which receive ubiquitin from an E1 and pass it directly, or via an E3, on to a substrate. The noncatalytic carboxyl termini of Ubc enzymes are more variable and are found on only some members of the family. The function of this region is unknown, however, it may be required for substrate recognition or regulation of Ubc activity.
Cdc34p, a member of the Ubc family of proteins, is essential for cell cycle progression and primarily acts at the G 1 to Sphase transition (9). Cell cycle progression in yeast is coordinated by the protein kinase, Cdc28p, and the activity of this kinase requires its association with another protein termed cyclin (for review, see Refs. 10 -12). It is thought that the cyclin will determine substrate specificity for the kinase or direct the kinase to particular locations in the cell. In yeast nine cyclins have been characterized for Cdc28p and different cyclins associate with the Cdc28p kinase during different phases of the cell cycle. For example, G 1 progression is coordinated by the cyclins Cln1p, Cln2p, and Cln3p that bind and activate Cdc28p. To initiate DNA replication, or S-phase, Cdc28p associates with cyclins Clb5p or Clb6p. Cyclin kinase activity can be negatively regulated by the physical association of an inhibitor (13)(14)(15). One such inhibitor in yeast is Sic1p and this protein negatively regulates the activity of Cdc28p-Clb5p and Cdc28p-Clb6p complexes (16 -19). Only upon Sic1p degradation are these complexes active and S-phase initiated.
Sic1p is one candidate substrate for Cdc34p-mediated ubiquitin degradation (19). Cells that lack CDC34 activity are unable to remove Sic1p and thus arrest without initiating DNA replication. This phenotype is shared by cells that lack CDC4 and CDC53 activity. Although their activity is required for Sic1p ubiquitination (19), the precise function of these proteins is unknown. CDC53 encodes a protein that defines a family of evolutionary conserved proteins which may serve to recognize substrates that have been targeted for ubiquitination (20 -22). CDC4 encodes a protein containing 7 WD-40 repeats, a motif that has been implicated in protein-protein interactions (23,24). Furthermore, Cdc4p also contains an F-Box, a domain that is the binding site for Skp1p (25). Skp1p is required for the degradation of Sic1p, Cln2p, and Clb5p (25). It has been speculated that a complex containing these proteins is required for the degradation of Sic1p and initiation of S-phase (26). Here, we define a region on Cdc34p that is necessary and sufficient for its binding Cdc4p and Cdc53p. We report that Cdc4p, Cdc34p, and Cdc53p form a high molecular weight complex and that this association is required for Cdc34p function. We also show that these three proteins are not only present throughout the cell cycle but that Cdc34p is part of a complex during this process.

EXPERIMENTAL PROCEDURES
Yeast Strains and Growth Conditions-Yeast strains used in this study are as follows; YL10 -1 MATa cdc34 -2 ura3-52 leu2⌬1 trp1⌬63 his3⌬ (27) and Y382 MAT␣ ade2 ade3 ura3 leu2 trp1 (kindly provided by A. Bender). Standard rich (YPD) and defined minimal (SD) were prepared as described previously (28). Transformations were carried out as described previously (29). For plasmid selection, yeast cells were grown in defined minimal medium supplemented with the appropriate amino acids. For galactose induction, cells were first grown to early logarithmic phase in minimal media containing sucrose instead of dextrose and then transferred to minimal media containing galactose as the sole carbon source and grown until late logarithmic phase. For complementation experiments, patches derived from single transformants were grown under permissive conditions of 23°C and then replica plated and incubated under nonpermissive conditions of 37°C. For chemical arrests, cells in early logarithmic phase were treated for 3-4 h with either 5 M ␣-factor or 100 mM hydroxyurea or 10 g/ml nocodazole. Arrest was confirmed by microscopic examination. To obtain temperature sensitive-dependent arrest, cells were grown to early logarithmic phase at the permissive temperature of 23°C and shifted to the nonpermissive temperature of 37°C for 3-4 h. Again, cell arrest was confirmed by microscopic examination.
The vector pEG(KG) was used for the expression of glutathione S-transferase (GST) fusion proteins (31). Expression of the GST fusion proteins were from the GAL1-10 promoter. All GST-CDC34 fusion constructs were created by cloning polymerase chain reaction-generated DNA fragments using plasmid-borne CDC34 DNA as template using methods as described previously (27). The use of the primers annealing at the 5Ј-end of CDC34 were designed to incorporate a XbaI restriction site and the use of primers annealing at the 3Ј-end of CDC34 were designed to incorporate a SalI restriction site. The polymerase chain reaction products were cleaved with these enzymes and ligated into pEG(KG) that had previously been digested with the same enzymes. The primers annealing to the 3Ј-end of CDC34 also contained two successive stop codons. To generate full-length CDC34 we used primers 5Ј-ATTCTAGACTCCATGAGTAGTCGCAAAAGCACCGC-3Ј and 5Ј-GGGGTCGACTTATTACTTTGAAACTCTTTCTACATCCTCCAC-3Ј. To generate a fragment encoding residues 1-170 we used primers 5Ј-ATTCTAGACTCCATGAGTAGTCGCAAAAGCACCGC-3Ј and 5Ј-GGG-GTCGACTTATTAATCTTGTTTCGATCTTTCCAC-3Ј. To generate a fragment encoding residues 171-295 we used primers 5Ј-ATTCTA-GACTCCGATATCCCTAAAGGTTTCATAATGC-3Ј and 5Ј-GGGGTC-GACTTATTACTTTGAAACTCTTTCTACATCCTCCAC-3Ј. To generate a fragment encoding residues 171-209 we used primers 5Ј-ATTCTA-GACTCCGATATCCCTAAAGGTTTCATAATGC-3Ј and 5Ј-GGGGTC-GACTTATTACAAATCACTATCATACCAAAAATTATCCG-3Ј. To generate a fragment encoding residues 1-209 we used a combination of primers mentioned above.
Protein Preparation and Western Immunoblot Techniques-Yeast lysate preparations and Western immunoblot techniques were carried out as described previously (21), except that cells were lysed in the presence of 50 mM NaCl, 50 mM Tris-HCl (pH 7.5), 0.5 mM EDTA, 0.1% Triton X-100. Antibody production against Cdc4p, Cdc34p, and Cdc53p and antibody purification have been previously described (21,37). Antibodies raised against GST were purchased from Sigma.
Purification of GST-tagged Cdc34p-Yeast cell lysates containing 10 mg of protein were incubated for 1 h with 100 l of 50% slurry of glutathione-agarose beads (Sigma) at 4°C that had been previously equilibrated with lysis buffer. The beads were then washed repeatedly with 1 ml of lysis buffer. The protein bound to the beads was then subjected to Western analysis.
Gel Filtration Chromatography-Lysate from cells grown to midlogarithmic phase or arrested with nocodazole was prepared as described previously (21). 0.5 mg of protein were applied at a flow rate of 0.2 ml/min onto a Superose 12 HR10/30 column (FPLC, Pharmacia Biotech Inc.) that had been previously equilibrated in 50 mM NaCl, 50 mM Tris-HCl (pH 7.5), 0.5 mM EDTA, 0.1% Triton X-100. Protein from 400-l fractions was precipitated with 10% trichloroacetic acid. The precipitate was washed with 80% acetone, resolubilized in 25 l of 0.1 M NaOH and 5 l of SDS-PAGE sample buffer solubilized in 30 l of SDS sample buffer, separated by SDS-PAGE, and subjected to Western immunoblot analysis. The column had been previously calibrated and a standard curve was produced. The proteins used to calibrate the column were thyroglobulin (669 kDa), apoferritin (440 kDa), catalase (232 kDa), aldolase (158 kDa), bovine serum albumin (66 kDa), ovalbumin (43 kDa), chymotrypsin (25 kDa), and ribonuclease (13.7 kDa).

The Carboxyl Terminus of Cdc34p Is Required for Its in Vivo
Activity and Binding to Cdc4p and Cdc53p-We previously demonstrated that CDC4, CDC34, and CDC53 genetically interact and that Cdc34p physically associates with Cdc4p and Cdc53p (21). We set out to determine a binding domain in Cdc34p required for its association with Cdc4p and Cdc53p. The Ubc proteins each contain a highly conserved catalytic domain. However, the catalytic domain of Cdc34p contains features that define a subclass of ubiquitin-conjugating catalytic domains (27). In some Ubc proteins a carboxyl-terminal extension exists and is distinct among the different members of the family. The highly acidic carboxyl-terminal extension of Cdc34p is also unique in its sequence composition and in that its length is greater than in other members of the family. To determine whether the catalytic domain or the carboxyl-terminal extension on Cdc34p is required for its association with Cdc4p and Cdc53p, GST fusions were constructed encoding either full-length Cdc34p or domains of Cdc34p composed of the catalytic domain (amino acid residues 1-170) and the carboxyl terminus (amino acid residues 171-295) (Fig. 1A). To assess the function of these fusion proteins, and fusion proteins subsequently used in this study, each construct as well as the empty GST vector was transformed into strain YL10-1 which contains the temperature-sensitive cdc34-2 allele. Isolated transformants were transferred onto medium lacking uracil, which selected for the presence of the plasmid, and then incubated at 23°C, the permissive temperature for cells containing cdc34-2. These patches were replica plated to the same media except that galactose was included to induce expression of each GST fusion and were incubated at either 36°C. The ability of each GST fusion to rescue the temperature-sensitive phenotype of YL10-1 was assessed by the growth of each patch at 36°C (Fig. 1B). The plasmid encoding full-length Cdc34p fused to GST was able to complement the cdc34-2 mutation and allowed YL10-1 cells to grow at the nonpermissive temperature. The tagged Cdc34p had to be overproduced to complement the cdc34-2 allele, however, indicating that the NH 2 -terminal tag may affect Cdc34p function. Low level expression of this construct, achieved when cells are grown in the presence of dextrose, did not complement the cdc34-2 allele whereas low level expression of untagged Cdc34p is able to complement the cdc34-2 allele (Ref. 27 and data not shown). Neither GST alone nor the GST fusion constructs containing Cdc34p residues 1-170 or 171-295 were able to complement the cdc34-2 mutation and failed to allow growth of YL10-1 cells at the nonpermissive temperature. Microscopic examination showed that these cells displayed the multielongated bud phenotype typical of loss of CDC34 activity. These results are in agreement with previous studies which have demonstrated that Cdc34p residues 1-170 are necessary but not sufficient to maintain growth in a cdc34 temperature-sensitive strain at 36°C (32)(33)(34).
To assess whether the GST fusion proteins could associate with Cdc4p and Cdc53p in vivo, lysate prepared from cells expressing each construct was incubated with glutathione-Sepharose beads. Proteins bound to beads were analyzed by SDS-PAGE followed by Western immunoblot analysis with antibodies raised against GST, Cdc4p, or Cdc53p ( Fig. 2A). The fusion of GST with full-length Cdc34p allowed co-purification of both Cdc4p and Cdc53p whereas GST alone did not. The fusion of GST with the Cdc34p catalytic domain (residues 1-170) did not co-purify with either Cdc4p or Cdc53p, whereas the fusion of GST with the Cdc34p carboxyl-terminal extension (residues 171-295) retained the ability to co-purify with both Cdc4p and Cdc53p. Thus the binding domain required for Cdc34p to interact with Cdc4p and Cdc53p in vivo is located within the carboxyl-terminal portion of the protein. Since the region delineated by residues 171-295 is required for the cell cycle function of Cdc34p (Fig. 1B), our results suggest that the association of Cdc4p/Cdc34p/Cdc53p is essential for Cdc34p activity.

Residues 171-209 of Cdc34p Are Necessary and Sufficient for
Binding Both Cdc4p and Cdc53p-We next set out to determine more precisely the binding domain on Cdc34p required for its association with Cdc4p and Cdc53p. Previous work has demonstrated that the portion of CDC34 that is necessary and sufficient for its cell cycle function is that portion encoding the catalytic domain and a domain of the carboxyl-terminal extension, residues 1-209 (34). We wanted to test whether these Cdc34p residues are sufficient to bind Cdc4p and Cdc53p. We therefore constructed a GST fusion protein that contained residues 1-209 of the Cdc34p protein (Fig. 1A). When this protein is overproduced in YL10-1 cells, it is able to complement the cdc34-2 allele (Fig. 1B) and allows growth of YL10-1 cells at 36°C, demonstrating that residues 1-209 are sufficient for Cdc34p activity in vivo. Furthermore, the GST-Cdc34(1-209)p fusion was able to co-precipitate Cdc4p and Cdc53p (Fig. 2B). Therefore, the addition of 39 residues to the catalytic domain of Cdc34p (residues 1-170), creating Cdc34(1-209)p, restores in vivo activity and the ability to associate with Cdc4p and Cdc53p.
As described above, we have demonstrated that residues 1-170 of Cdc34p have no in vivo activity and cannot bind Cdc4p and Cdc53p. We wanted to test whether the Cdc34p residues 171-209 are sufficient by themselves to bind Cdc4p and Cdc53p. A GST fusion construct encoding these residues was made (Fig. 1A) and as shown in Fig. 2C this GST fusion protein is sufficient to bind Cdc4p and Cdc53p selectively in vivo. Thus we have defined a 39-amino acid region on Cdc34p that is necessary for CDC34 activity and sufficient for associating with Cdc4p and Cdc53p.
Cdc4p, Cdc34p, and Cdc53p Are Present within a High Molecular Weight Species-The above data and our previous work demonstrating genetic interactions between CDC4, CDC34, and CDC53 strongly suggest that Cdc34p forms a complex with Cdc4p and Cdc53p in vivo. To test this model and to obtain approximate size for this complex, yeast lysate was subjected to gel filtration column chromatography. To allow separation of protein complexes with high molecular mass, lysate was applied to a Superose-12 column that was calibrated up to 670 kDa. Fractions were collected and the presence of Cdc4p, Cdc34p, and Cdc53p was detected by Western blotting with the appropriate antibody. Densitometric scanning of these results revealed that Cdc4p, Cdc34p, and Cdc53p were absent in fractions corresponding to their monomeric size of 88, 40, and 96 kDa, respectively. Instead, the detectable cellular pool of these proteins was present in fractions corresponding to high molec- ular masses, indicating that these proteins are part of multiprotein complexes (Fig. 3). However, Cdc4p, Cdc34p, and Cdc53p did not co-elute precisely into the same fractions. Instead we observed that Cdc4p, Cdc34p, and Cdc53p exist as overlapping families of high molecular weight complexes and, by comparison to protein standards, these fractions corresponded to molecular sizes of between 220 and 670 kDa peaking at about 400 kDa. Approximately 10% of the total amount of Cdc4p and Cdc53 were detected in fractions that correspond to a apparent molecular size of approximately 1,200 kDa. Since this value was obtained by linear extrapolation from the size markers, 1,200 kDa is an underestimate of its size. Furthermore, about 30% of Cdc4p was detected alone in fractions that correspond to a molecular size of approximately 180 kDa. These data suggest that Cdc4p and Cdc53p associate with other as yet uncharacterized proteins.
Cellular Cdc4p Levels Are Increased by Increased Amounts of Cdc34p-A number of studies have shown that the steady state level of a protein that is a component of a multiprotein complex may be enhanced by increasing the abundance of other members of the complex (35,36). We wanted to observe whether the steady state levels of Cdc4p, Cdc34p, or Cdc53p were enhanced by increasing the level of one of the other proteins. Plasmids that overexpress CDC4, CDC34, and CDC53 have been previously described (21). Overexpression of CDC4 was achieved by replacing its own promoter with a galactose inducible promoter; the promoter of CDC53 was replaced by the ADH promoter and finally CDC34 overexpression was obtained from a high copy plasmid. Y382 and Y382 transformants containing these plasmids were grown under conditions to allow for CDC4 induction, protein lysates were prepared and subjected to Western immunoblot analysis using antibodies raised against Cdc4p, Cdc34p, or Cdc53p. Although overproduction of each protein was achieved, the overproduction of one protein did not result in an increase in the amount of the other proteins (data not shown).
We next carried out a similar experiment but overproduced Cdc4p, Cdc34p, and Cdc53p in different pairwise combinations and also all three proteins together. We could control the timing of CDC4 overexpression since this gene was under the control of a galactose-inducible promoter and is only expressed when cells are grown in the presence of galactose. This allowed us to follow Cdc4p accumulation as a time course with galactose being added at time point 0 and cells were harvested every 30 min. The abundance of Cdc34p and Cdc53p was unaffected when these proteins were co-overproduced with either of the other two proteins or all three proteins co-overproduced together (data not shown). However, we observed that Cdc4p abundance was increased when CDC4 was co-overexpressed with CDC34 and increased even further when CDC4 was cooverexpressed with both CDC34 and CDC53 (Fig. 4). The levels of Cdc4p were significantly greater under these conditions than that observed when CDC4 was either overexpressed alone or co-overexpressed with CDC53. One interpretation of these results is that Cdc4p may directly associate with Cdc34p and this association allows greater stability of Cdc4p than when Cdc4p is not associated with Cdc34p. The Cdc4p⅐Cdc34p heterodimer may associate with Cdc53p and this complex results in the greatest stability of Cdc4p. This data also suggests that perhaps Cdc4p does not associate with Cdc53p in the absence of Cdc34p since we observed no increase in Cdc4p stability when CDC4 was co-overexpressed with CDC53.
Cdc4p, Cdc34p, and Cdc53p Complexes Are Present throughout the Cell Cycle-A feature of some cell cycle regulators is that their abundance oscillates during cell cycle progression. For example, the cyclins Cln1p and Cln2p are present late in G 1 and the CDK inhibitor Sic1p is present throughout G 1 , yet both cyclins and Sic1p are absent for the remainder of the cell cycle (11). Since these proteins are strong candidates for Cdc34p substrates, we wanted to test whether Cdc34p containing complexes are present only during G 1 or throughout the cell cycle.
First we set out to determine whether the levels of Cdc4p, Cdc34p, or Cdc53p fluctuate throughout the cell cycle. Lysate from cells that had been arrested at specific points of the cell cycle was prepared. We employed the use of common cell cycle inhibitors for this process. Logarithmically growing Y382 cells were incubated in the presence of either ␣-mating factor, which arrests cells during G 1 ; hydroxyurea, which arrests cells during S-phase; or nocodazole which arrests cells during mitosis. The presence of Cdc4p, Cdc34p, and Cdc53p was tested by Western   FIG. 3. Cdc4p, Cdc34p, and Cdc53p occur within overlapping high molecular weight complexes upon Superose 12 gel-filtration column chromatography. Densitometric scanning profiles of immunoblots of Cdc4p, Cdc34p, and Cdc53p obtained after fractionation of yeast whole cell lysate on a Superose 12 column. The subsequent fractions were concentrated and then subjected to Western immunoblot analysis using antibodies raised against Cdc4p, Cdc34p and Cdc53p. Size estimates, in kDa, were calculated based on the standard curve produced from the molecular markers described under "Experimental Procedures." immunoblot analysis which showed that the level of each protein was invariant throughout the cell cycle (Fig. 5). Both Cdc34p and Cdc53p are modified by ubiquitin (22,37) and in the case of Cdc53p, both the modified and unmodified species are equally prevalent throughout the cell cycle (Fig. 5). However, Cdc34p ubiquitination, which has been mapped to the carboxyl-terminal (residues 254 -295 (27)), appeared to be cell cycle-specific (Fig. 5). Cdc34p is least ubiquitinated in cells arrested with ␣-factor and most ubiquitinated in cells arrested with nocadozole, although this modification appears to affect only a small portion of the Cdc34p population.
To test whether Cdc34p containing complexes are present throughout the cell cycle, lysate prepared from cells arrested with nocodazole was applied to the same gel filtration procedure as described above. The resulting fractions were analyzed using antibodies raised against Cdc34p. The profile obtained from cells arrested with nocodazole was similar to that for unarrested cells (Fig. 6) for the majority of the Cdc34p. A small amount of Cdc34p, approximately 10%, did appear to be of very high molecular mass, in excess of 669 kDa. A similar analysis of Cdc53p also demonstrated that the gel filtration profiles for this protein was unchanged (data not shown). These data suggest that Cdc34p is present as a high molecular weight complex throughout the cell cycle. DISCUSSION One essential function of the ubiquitin-conjugating enzyme Cdc34p is to attach Ub onto the cyclin-dependent kinase inhibitor Sic1p and thus signal Sic1p for proteolysis. The precise mechanism by which Cdc34p transfers Ub onto Sic1p is not known, however, this process is dependent on at least two other proteins, Cdc4p and Cdc53p. In this report we map a Cdc4pand Cdc53p-binding domain on Cdc34p and show the importance of in vivo interactions between these three proteins, strengthening the hypothesis that Sic1p ubiquitination is me-diated by a multi-protein complex.
The Association of Cdc34p with Cdc4p and Cdc53p Is Essential for Cdc34p Activity-We have defined a Cdc4p/Cdc53p- FIG. 4. Steady state levels of Cdc4p are increased when CDC4 is co-overexpressed with CDC34 and CDC53. Y382 cells were transformed with different combinations of plasmids that were capable of overexpressing CDC4, CDC34, and CDC53, as indicated. Plasmid borne expression of CDC4 was induced in the presence of galactose. Lysate prepared from cells at the times indicated were subjected to Western immunoblot analysis using anti-Cdc4p antibodies. The loading control is a protein that cross-reacts with the anti-Cdc4p antibodies and its abundance is not affect by overexpressing CDC4, CDC34, or CDC53. binding domain on Cdc34p (Fig. 7). This region is not found in other Ubc proteins which suggests that these interactions are unique to Cdc34p. However, the efficiency of Cdc4p and Cdc53p binding to a fusion containing Cdc34p residues 1-209 or residues 171-209 is less than that of a fusion containing full-length Cdc34p. Therefore the remainder of the Cdc34p protein may act to stabilize its association with Cdc4p and Cdc53p. Ptak et al. (34) have shown that this region is required for Cdc34p dimerization in vitro and it is possible that Cdc4p and Cdc53p bind to a Cdc34p dimer in vivo. Absence of this binding region results in loss of Cdc34p activity and cells are prevented from entering S-phase due to failure to degrade the CDK inhibitor Sic1p. Thus the association of Cdc34p with Cdc4p and Cdc53p is essential for Sic1p ubiquitination. What role does Cdc4p and Cdc53p play in Sic1p ubiquitination? Cdc34p residues 1-170 contain the catalytic domain of Cdc34p and its active site cysteine (Fig. 7). This region has been shown to be charged with ubiquitin by the ubiquitin-activating enzyme, E1 (9), an event that does not require Cdc4p or Cdc53p association. Therefore Cdc4p and Cdc53p most likely play a role in transferring Ub from Cdc34p onto its substrates either by substrate recognition or by having a Ub-ligase (E3) activity. Ub-ligases (E3s) have been described (7,8) yet the sequence comparison of these proteins do not resemble either Cdc4p or Cdc53p. Possibly Cdc53p is required to present the substrate to Cdc34p. Cdc53p has been found to associate with Cln2p, a G 1 cyclin that is another candidate Cdc34p substrate (22).
Cdc4p, Cdc34p, and Cdc53p Are Members of a Multiprotein Complex-Gel filtration analysis shows that Cdc4p, Cdc34p, and Cdc53p are not found as monomers in yeast lysate and take part in associations resulting in the formation of high molecular mass complexes. Overlap of the three proteins' distribution suggests that the Cdc4p⅐Cdc34p⅐Cdc53p complex is approximately 400 kDa. This size is greater than that predicted for a complex containing one of each protein (218 kDa), that have a predicted molecular mass of 88, 34, and 96 kDa for Cdc4p, Cdc34p, and Cdc53p, respectively. Therefore either these proteins may act as multimers or additional proteins are part of this complex. Skp1p is required for the ubiquitination of Sic1p and has been shown to bind to Cdc4p through a Cdc4p domain called the F-box (25). However, Skp1p is only 22 kDa and perhaps other such proteins are required for Cdc34p to interact with Cdc4p and Cdc53p. Significant amounts of Cdc4p and Cdc34p co-fractionate in the absence of Cdc53p and together with our data demonstrating the stabilization of Cdc4p when co-overproduced with Cdc34p suggest that these two proteins form a heterodimer. Again, this association, of approximately 200 kDa may occur in the presence of other proteins. Therefore it appears that components of the Cdc4p⅐Cdc34p⅐Cdc53p complex may be separated and be redistributed for distinct ubiquitinating activities.
The Cdc34p Complex Is Present throughout the Cell Cycle-Finally we determined whether Cdc34p existed within a complex throughout the cell cycle. Many cell cycle regulators are themselves only present during specific times of the cell cycle. We have shown here that Cdc4p, Cdc34p, and Cdc53p are present throughout the cell cycle and that Cdc34p appears to reside as part of a complex throughout the cell cycle. Potentially, Cdc34p activity may be required throughout the cell cycle and not just at the G 1 /S transition in spite of the fact that some of its substrates are degraded in a cell cycle specific manner. Phosphorylation of Sic1p and Cln2p is a prerequisite for these substrates to be ubiquitinated by Cdc34p and Cdc34p activity may be solely regulated by recognizing phosphorylated substrates (38,39). Other candidate Cdc34p substrates include Cdc6p (40), Far1p (41), and Gcn4p (42). Both Cdc6p and Far1p are degraded in a cell cycle specific manner but outside the G 1 /S transition and therefore the Cdc34p complex must be active throughout the cell cycle.