The stability of the Cdc6 protein is regulated by cyclin-dependent kinase/cyclin B complexes in Saccharomyces cerevisiae.

The Saccharomyces cerevisiae Cdc6 protein is necessary for the formation of prereplicative complexes that are a prerequisite for firing origins during DNA replication in the S phase. In budding yeast, the presence of Cdc6 protein is normally restricted to the G(1) phase of the cell cycle, at least partly because of its proteolytic degradation in the late G(1)/early S phase. Here we show that a Cdc28-dependent mechanism targets p57(CDC6) for degradation in mitotic-arrested budding yeast cells. Consistent with this observation, Cdc6-7 and Cdc6-8 proteins, mutants lacking Cdc28 phosphorylation sites, are stabilized relative to wild-type Cdc6. Our data also suggest a correlation between the absence of Cdc28/Clb kinase activity and Cdc6 protein stabilization, because a drop in Cdc28/Clb-associated kinase activity allows mitotic-arrested cells to accumulate Cdc6 protein. Finally, we also show that cdc28 temperature-sensitive G(1) mutants accumulate Cdc6 protein because of a post-transcriptional mechanism. Our data suggest that budding yeast cells target Cdc6 for degradation through a Cdc28-dependent mechanism in each cell cycle.

Recent work aimed at elucidating the mechanisms that underlie the control of DNA replication in eukaryotes has afforded a two-step model (1)(2)(3)(4). This model first proposes the establishment of a state of competence for replication at origins, as soon as cells exit from mitosis, and then its activation for the initiation of DNA synthesis, during the G 1 -to-S phase transition. At the same time, these steps must somehow be linked to the precise control of cell cycle progression to prevent reduplication of the genome before passage through mitosis. This latter type of control appears to be integrated through modulation by cyclin-dependent kinase (CDK) 1 activity (5)(6)(7).
The state of competence means that a multisubunit protein complex must be formed at the origins of replication, the discrete sites at which DNA synthesis is initiated. In Saccharomyces cerevisiae, the sequence elements making up these origins have been well characterized (8). These origins are bound throughout the cell cycle by a complex of six proteins called the origin recognition complex (ORC), which is required for the initiation of DNA replication (9 -11). Although ORCs at origins are essential, the binding of additional factors is needed for the initiation of genome duplication, as shown by in vivo footprint pattern changes during the cell cycle (11). In S. cerevisiae a number of gene products have been identified that interact with some of the components of the ORC, participating in the assembly of the initiation complex, or prereplicative complex (pre-RC). These include Mcm proteins, Cdc45, Cdc7, and Dbf4, as well as Cdc6 (12)(13)(14)(15).
The formation of prereplicative complexes in budding yeast is dependent on Cdc6 (16). Genetic evidence has indicated that Cdc6 is rate-limiting for the initiation of DNA replication through its interaction with ORC (13). The closest homolog of CDC6 in fission yeast, cdc18ϩ, is also required for DNA replication (17). Cdc18 protein levels vary during the cell cycle, peaking at G 1 /S like Cdc6 (18,19). Additionally, the overexpression of cdc18ϩ induces continuous DNA synthesis (18,20), suggesting that this gene plays a key role in initiating DNA replication. In fission yeast a link between the cell cycle machinery and pre-RCs has emerged through the demonstration that Cdc18 interacts in vivo with Cdc2 (the fission yeast homolog of budding yeast Cdc28) and Orp2 (the fission yeast homolog of budding yeast Orc2) (21). Also, it has been shown that in budding yeast the Cdc6 protein when overexpressed interacts with Cdc28 (22,23).
The inactivation of Cdc28/B-cyclin kinase seems to be required for both the exit from mitosis and the resetting of chromatin for the subsequent S phase (6,7). This resetting involves Mcm binding to chromatin in a hypophosphorylated state (24 and references therein) and Cdc6 being recruited to origins, an essential step for the formation of pre-RCs (15,16). At the same time, reformation of the prereplicative complex is actively inhibited by Cdc28/B-cyclins (7). Because B cyclins are required for both entry into the S phase and the inhibition of nuclear resetting, another round of replication cannot occur until B-cyclin/Cdc28 kinases have been destroyed in the subsequent mitosis (6,7). Consistent with this hypothesis, it has been suggested that a Cdc2-mediated mechanism regulates Cdc18 degradation in Schizosaccharomyces pombe, preventing this fission yeast from entering a second round of DNA replication within the same cell cycle (25,26).
Our work here addresses the mechanism regulating the degradation of the key DNA replication initiator Cdc6 protein in S. cerevisiae at the G 1 -to-S phase boundary. We have previously shown that Cdc6 is ubiquitinated in vivo for degradation (27). In accordance with published data (28), our results indicate that Cdc4 and Cdc34 participate in Cdc6 proteolysis, suggesting that the initiator protein is phosphorylated for degradation through the proteasome. Consistent with this hypothesis, here we report data suggesting that the phosphorylation of Cdc6 by Clb/Cdc28 kinases stimulates Cdc6 proteolysis and therefore restricts the formation of pre-RCs to G 1 . In light of these results, we propose not only that CDC6 acts to initiate DNA replication at the onset of the S phase but also that its degradation is one of the multiple CDK-mediated events that prevent the reinitiation of DNA replication within the same cell cycle, because in the absence of Cdc6 no pre-RCs are formed.
YIpG3CDC6, YIpG3cdc6-7, and YIpG3cdc6-8 integrative transformations in S. cerevisiae were targeted into the URA3 locus and also tested by Southern blot.
Culture Conditions and General Techniques-Cells were grown in YEPD (1% yeast extract, 2% bactopeptone, 2% glucose) except when selecting for plasmids. In this case, cultures were grown in minimal media with supplemented amino acids. General molecular techniques were performed as described (34,35).
The DNA content of individual cells was measured using a Becton Dickinson FACScan. Cells were prepared for flow cytometry, using the method of Hutter and Eipel (36), by staining them with propidium iodide.
Galactose Induction Experiments-Yeast cells were grown in YEP supplemented with 2% raffinose (YEPR) at 32°C until the culture reached the midlog phase. Then, galactose was added to a final concentration of 2.5%. When temperature-sensitive strains were used, cells were grown at 25°C.
Nocodazole-induced Cell Cycle Arrest-Cultures were incubated with shaking in the presence of nocodazole (15 g/ml). Cells were further incubated and collected for fluorescence-activated cell sorting and microscopical analysis.
Protein Extract Preparation-Soluble protein extracts were prepared as described previously (37). Cells were collected, washed, and broken in 30 l of histone buffer using glass beads. The HB buffer contained 60 mM ␤-glycerophosphate, 15 mM p-nitrophenylphosphate, 25 mM 4-morpholinepropanesulfonic acid (pH 7.2), 15 mM MgCl 2 , 15 mM EGTA, 1 mM dithiothreitol, 0.1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, and 20 g/ml leupeptin and aprotinin. The glass beads were washed with 500 l of histone buffer, and the supernatant was recovered. Protein concentrations were measured using the BCA assay kit (Pierce).
Kinase Activity Assays-Total p34 CDC28 protein kinase activity was assayed after immunoprecipitation with anti-Cdc28 or with 12CA5 antibody, using histone H1 (Roche Molecular Biochemicals) as substrate (37). p34 CDC28 was immunoprecipitated from 0.8 mg of soluble protein extracts with 0.5 l of polyclonal anti-Cdc28 antibodies (39). Cdc6Ha was immunoprecipitated from 1.6 mg of soluble extract with 1.5 g of 12CA5 monoclonal antibody (Roche Molecular Biochemicals). Immunoprecipitates were incubated in a 0.1 mM ATP, 0.5 mg/ml histone H1, and 20 Ci/ml [␥-32 P]ATP reaction mix at 30°C for 30 min. Reactions were stopped with 1 ϫ final Laemmli-SDS sample buffer and denatured for 5 min at 100°C. Samples were run on a 12% SDS-polyacrylamide gel. Phosphorylated histone H1 was detected by autoradiography and quantitated using a Fujifilm BAS1200 PhosphorImager.

Cdc6
Interacts in Vivo with Cdc28 in S. cerevisiae Cells-A direct interaction between recombinant Cdc6 and ORC has been described (13). More recent biochemical evidence suggests that, when overexpressed, Cdc18, a Cdc6-like protein in fission yeast, interacts in vivo with p34 cdc2/CDC28 and Orp2, a putative component of the S. pombe ORC (21). In fact, also in budding yeast, it has been shown that Cdc6 interacts in vivo with Cdc28, at least under overexpressing conditions (22,23). To test in vivo interactions between Cdc6 and Cdc28, we constructed a series of S. cerevisiae strains carrying a single copy of Ha-tagged CDC6 (see "Experimental Procedures"). Gene replacement experiments revealed that Cdc6Ha was fully functional (data not shown). Immunoprecipitation from Cdc6Haexpressing cells with Ha monoclonal antibody coprecipitated H1 kinase activity (Fig. 1A). A cdc28-13 CDC6HA double mutant strain was also constructed. The kinase activity co-immunoprecipitated from this double mutant was thermosensitive in an in vitro assay, indicating that Cdc28 was physically interacting with Cdc6Ha (Fig. 1A). Furthermore, we detected p34 CDC28 in Cdc6Ha immunoprecipitates from those strains by Western analysis (Fig. 1A). These results suggest that Cdc6 interacts in vivo with Cdc28 in S. cerevisiae cells under physiological conditions.
Cdc6 Protein Is Degraded in Mitotic-arrested Cells by a Cdc28dependent Mechanism-The above biochemical interaction suggested that Cdc6 could be an in vivo substrate for Cdc28 kinase, indicating the possibility that Cdc6 function or stability may be regulated by Cdc28 phosphorylation. Consistent with this hypothesis, it has been shown that Cdc6 is an in vivo phosphoprotein (22) and that bacterially produced Cdc6 is an in vitro substrate for Clb-associated Cdc28 activity (Refs. 22, 23, this work, and Fig. 1B). We therefore evaluated Cdc6 stabilization in cells deficient in p34 CDC28 -associated kinase activity by Western analysis. Using a previously characterized rabbit polyclonal antibody (27) we examined the rate of Cdc6 disappearance in promoter shut off experiments in a cdc28-13 GAL1-10:CDC6 strain. CDC6 expression was induced in nocodazolearrested cells by incubating them in galactose for 60 min at 37°C. Transcription was repressed by the addition of glucose, and Cdc6 stability was estimated by Western blotting at both 25°C (permissive temperature) and 37°C (restrictive temperature) (Fig. 2). Whereas Cdc6 became rapidly unstable at the permissive temperature or in a wild-type control at both temperatures, its levels remained constant up to 45 min at the restrictive temperature for cdc28-13. Comparable results were obtained when cycloheximide was added to inhibit protein synthesis at the same time that transcription was repressed with glucose (data not shown); nevertheless, Cdc6 protein was barely detectable after releasing the cells at the permissive temperature, as shown for CDC28 wild-type cells (Fig. 3C, upper panel). These results strongly suggest that cells blocked in mitosis require Cdc28 to degrade ectopically produced Cdc6.
Cdc6-7 and Cdc6-8 Proteins Lacking Cdc28 Putative Phosphorylation Sites Are Stabilized Relative to Wild-type Cdc6 -We reasoned that if mitotic-arrested cells required a functional Cdc28 kinase to degrade Cdc6, it is conceivable that p57 CDC6 may be targeted for degradation by p34 CDC28 phosphorylation. If this were the case, it could be predicted that mutant forms of Cdc6 protein lacking its phosphorylation sites would be more stable than wild-type protein. Cdc6 contains eight Cdc28 consensus phosphorylation sites, three of them minimal ((S/T)P). The rest conform to full Cdc28 consensus sites ((S/ T)PX(K/R)). We therefore constructed a number of mutants bearing different combinations of changes involving replacement of phosphoacceptor amino acids by alanines. We analyzed two mutant forms; one of them, named cdc6-7, carries alanine substitutions of Thr-23, Thr-39, Ser-43, Thr-135, Ser-354, Thr-368, and Ser-372. In the other we changed all phosphorylatable serines or threonines at putative Cdc28 phosphorylation sites by alanines. We called this version of the gene cdc6-8.
Our additional approach to understand whether Cdc28 controls Cdc6 stability was to examine the rate of Cdc6 disappearance in promoter shut off experiments in a GAL1-10:CDC6 strain compared with Cdc6-7 and Cdc6-8 relative stability in GAL1-10:cdc6-7 or in GAL1-10:cdc6-8 strains, respectively. Similar to the experiment described in Fig. 2, CDC28 wild-type GAL1-10:CDC6, CDC28 GAL1-10:cdc6-7, or CDC28 GAL1-10: cdc6-8 cells were blocked in nocodazole for 2.5 h at 30°C. Flow cytometry and microscopic analysis demonstrated that cells were arrested in G 2 /M (data not shown). Expression of wild-FIG. 1. Cdc6 co-immunoprecipitates with Cdc28. A, Cdc6 associates with Cdc28-dependent H1 kinase activity in S. cerevisiae cells. Protein extracts were prepared from cdc28-13 CDC6HA (cdc28ts) or CDC28 CDC6HA (wt) growing exponentially at 25°C. Anti-Ha immunoprecipitates obtained from 3 mg of cdc28-13 Cdc6Ha or CDC28 Cdc6Ha cell extracts were split into four parts. Two aliquots were tested for their ability to phosphorylate H1 in vitro at 25 or 37°C, and the other two were used for the detection of p34 CDC28 by immunoblotting with ␣-Cdc28 antibody (lower bands). An untagged CDC6 control is also shown (control). B, recombinant Cdc6 protein is an in vitro substrate for Cdc28-associated kinase activity. Cdc6 was expressed in Escherichia coli, purified to homogeneity, and used as an exogenous substrate for Cdc28. The kinase was immunoprecipitated from S. cerevisiae 15Dau protein extracts with specific antibodies and assayed for its ability to phosphorylate different samples of either Cdc6 or histone H1 or both, as indicated.
FIG. 2. A Cdc28-dependent mechanism targets p57 CDC6 for degradation in mitotic-arrested budding yeast cells. Immunoblot analysis of cell extracts from cdc28-13 GAL1-10:CDC6 is shown. This double-mutant strain was blocked with nocodazole at 25°C, and galactose was added for CDC6 induction for 60 min at 37°C. Afterward, GAL1-10:CDC6 expression was repressed by the addition of glucose (0Ј). The culture was split into two aliquots, and these were further incubated at 25 or 37°C. Samples were taken at the indicated intervals, processed for Western blotting, and probed with affinity-purified ␣-Cdc6 or ␣-Cdc28 antibodies. Aliquots of the samples were subjected to fluorescence-activated cell sorting analysis, which revealed that Ͼ98% of the cells were arrested with a 2C DNA content during the experiment (data not shown). A CDC28 GAL1-10:CDC6 control is shown for comparison.  19) or rabbit antiserum to Cdc28 for loading control. C, Analysis of the relative stability of Cdc6 phosphorylation site mutants. Immunoblot analysis of cell extracts from GAL1-10:CDC6, GAL1-10:cdc6-7, and GAL1-10:cdc6-8 strains is shown. The three mutant strains were blocked with nocodazole at 30°C (NZ), and galactose was added for CDC6 induction during 60 min. Afterwards, GAL1-10 expression was repressed by addition of glucose (0Ј) as well as protein synthesis inhibited by adding cycloheximide (CH; 20 g/ml). Samples were taken at indicated intervals, processed for Western blotting, and probed with affinity-purified goat antiserum to Cdc6 or rabbit antiserum to Cdc28. type CDC6 or mutant forms cdc6-7 and cdc6-8 was induced 60 min by addition of galactose. Transcription was repressed by adding glucose. At the same time, and in addition to the GAL1-10 promoter shut off, cycloheximide was added to ensure that no new synthesis of the endogenous Cdc6 was taking place. The stability of the different forms of Cdc6 was tested by Western blot in samples taken at indicated intervals in Fig. 3, A and C. Whereas wild-type Cdc6 protein become rapidly unstable, being barely detectable 15 min after blocking transcription and translation (Fig. 3C, upper panel), both Cdc6-7 and Cdc6-8 mutant proteins remained stable for up to 2 h in these nocodazole-arrested cells (Fig. 3C, middle and lower panels). From these experiments we conclude that the elimination of putative in vivo sites for phosphorylation in Cdc6-7 and Cdc6-8 results in the stabilization of Cdc6 protein.
Cdc6-7 and Cdc6-8 Mutant Proteins Are Functional-To test whether CDC6 mutants were functional, a complementation study was carried out on a cdc6-1 strain background. All mutants alleles were placed under the control of the CDC6 wildtype promoter, which allowed us to perform a functional analysis. As a negative control, we constructed a mutant in which we changed the critical lysine at amino acid 114 to a glutamate in the unique match to the P-loop of the consensus purine nucleotide-binding motif (GXXGXGKT, at amino acids 108 -115) present in Cdc6. We called this mutant cdc6-E114. Cells bearing this temperature-sensitive allele were transformed with four different centromeric plasmids (pHR70) having no insert or containing CDC6, cdc6-7, or cdc6-8, all of them under the control of the CDC6 wild-type promoter (see "Experimental Procedures"). Transformants capable of growing on selective media at 25°C were replica-plated and further tested for growth at 37°C. All mutant alleles efficiently rescued cdc6-1 (Fig. 4). We did not observe any differences in the complementation test between mutants and wild-type CDC6. In similar experiments, cdc6-E114 was unable to rescue the temperaturesensitive defect associated with cdc6-1 (data not shown), indi-cating that the purine nucleotide-binding motif is essential for CDC6 function in vivo. Because the rescue of the cdc6-1 temperature-sensitive (ts) allele does not necessarily mean that Cdc6-7 or Cdc6-8 are functional because, for instance, they could serve to stabilize the ts protein, we set up a complementation test of a null allele of cdc6 with these cdc6 phosphorylation site mutants (Fig. 4). A cdc6 deletion strain (cdc6::KANMX4), carrying a GAL1-10:CDC6 allele at the URA3 locus that supports growth in galactose-based media, was transformed with high multicopy plasmids (YEp13) bearing wild-type CDC6, cdc6-7, or cdc6-8 under the control of the CDC6 wild-type promoter. To test the ability of these plasmids to complement the null mutation, GAL1-10:CDC6 cells were changed to glucose medium to repress GAL1-10:CDC6 expression. All mutants rescued the deletion of CDC6. Nevertheless, it is remarkable that, in consonance with a delay in G 2 , YEp13cdc6-7 transformants formed colonies of elongated cells. These results indicate that both Cdc6-7 and Cdc6-8 proteins are functional.
A Drop in Cdc28-associated Kinase Activity Allows Mitoticarrested Cells to Accumulate Cdc6 Protein-It is already known that de novo synthesis of Cdc6 protein is required for cells to enter the S phase (19,23). On the other hand, it has been shown that transient inhibition of Cdc28-Clb kinases by expression of the Cdc28 inhibitor p40 SIC1 causes mitotic-arrested cells to undergo an extra round of DNA replication (7). To test whether CDC6 function is needed for this rereplication phenotype, p40 SIC1 -conditionally expressing wild-type CDC6 and cdc6-1 mutant strains were constructed. The rereplication ability in both strains was analyzed at either 25 or 37°C. After 40 min of induction, p40 SIC1 expression was repressed, and the DNA content was measured in samples taken at regular intervals from cultures incubated at the permissive or restrictive temperatures. As shown in Fig. 5A, rereplication in nocodazolearrested cells caused by transient inhibition of Cdc28-Clb kinases was prevented in cdc6-1 mutant cells at the restrictive temperature. These data suggest that the accumulation or presence of a fully functional Cdc6 protein is required for rereplication induced by a drop in Cdc28-associated kinase activity.
To understand whether this Sic1-dependent inhibition of p34 CDC28 would cause the accumulation of p57 CDC6 in nocodazole-arrested cells, we monitored the Cdc6 protein content by Western blot using antiserum to Cdc6. After induction of p40 SIC1 , samples were taken and divided into two aliquots for Western and p34 kinase assays. As shown in Fig. 5B, Cdc6 protein accumulated as soon as the kinase activity dropped, suggesting a correlation between the lack of p34 CDC28 activity and p57 CDC6 stabilization in mitotic-arrested cells.

Rereplication in Nocodazole-arrested Cdc6-7-expressing Cells Is Dependent on a Fall in Cdc28
Kinase Activity-The evidence presented above suggest that it might be important for mitotic cells to degrade the Cdc6 protein to prevent and extra round of DNA replication before cell division. If degradation of Cdc6 during mitosis is the only condition for preventing rereplication, it could be predicted that a stable form of the initiator protein, Cdc6-7 or Cdc6-8, would induce an extra round of DNA synthesis independently of a Sic1-dependent drop in Cdc28/ Clb-associated kinase activity in nocodazole-arrested cells. To address this possibility we repeated the experiment described in Fig. 5 in a cdc6-1 GAL1-10:SIC1 strain transformed with a plasmid expressing cdc6-7 from the wild type CDC6 promoter (pHR70 cdc6-7 ), with the idea of testing DNA re-replication ability of the cdc6-7 allele at the restrictive temperature for cdc6-1. SIC1 expression was induced in cells previously blocked in nocodazole, as checked by fluorescence-activated cell sorting and morphological analysis. After 40 min of induction, GAL1-10-regulated expression was repressed with glucose, and cell DNA content was measured in samples taken at regular intervals from cultures incubated at 25 or 37°C, permissive and restrictive temperatures for cdc6-1, respectively. As shown in Fig. 6, rereplication was dependent upon the SIC1-induced drop in Cdc28-kinase activity, suggesting that stabilization of Cdc6 does not interfere with regular controls limiting DNA re-replication during mitosis. A similar result was obtained with cdc6-1 GAL1-10:SIC1 pHR70CDC6 cells used as wild type control.
Cdc6 Protein Becomes Stabilized on G 1 CDK-deficient Mutant Yeast Strains Because of a Post-transcriptional Mechanism-Based on nocodazole-arrested cells, to here, our results clearly suggest that there is an inverse correlation between Cdc6 protein levels and CDK kinase activity during mitosis. We were interested also in understanding whether G 1 CDKdeficient mutant cells accumulate the Cdc6 protein by means of a post-transcriptional mechanism. To evaluate Cdc6 stabiliza-tion in cells deficient in p34 CDC28 -associated kinase activity, we took advantage of cdc28-4 and cdc28-13 thermosensitive alleles. Cells bearing the temperature-sensitive cdc28-13 allele were incubated at the permissive (25°C) or the restrictive (38°C) temperature for 60 or 90 min, and protein extracts were analyzed with antiserum to Cdc6 (Fig. 7A, right panel). Cdc6 wild-type protein accumulated at the restrictive temperature as a single band of 57 kDa. We also analyzed the accumulation of Cdc6 in a cdc28-4 mutant background (Fig. 7A, left panel). It has previously been described that this particular mutation of the kinase leads to an efficient pre-START block of S. cerevisiae cells (40). In fact, p34 immunoprecipitates from these mutant cells have low kinase activity (Fig. 8A), and we therefore expected to find a significant stabilization of p57 CDC6 in this background. An exponentially growing culture at the permissive temperature (25°C) was shifted to the restrictive temperature (37°C), and samples were taken at 1-h intervals for fluorescence-activated cell sorting analysis and protein detection. Flow cytometry and microscopical analysis demonstrated that cells were arrested in G 1 (data not shown). Western analysis of the cell extracts revealed that Cdc6 accumulated as a single band that was not present either in cells deleted for CDC6 or in asynchronous cells (Fig. 7A, left panel). The accumulation of the Cdc6 protein in Cdk-deficient yeast strains cdc28-4 and cdc28-13 was attributable to a post-transcriptional mechanism, given that CDC6 mRNA did not accumulate when both mutant strains were blocked at the restrictive temperature (Fig. 7B), whereas the CLN1 mRNA already accumulated as described previously (41). These findings suggest that Cdc6 protein is stabilized in S. cerevisiae cells devoid of a functional p34 CDC28 kinase.
A Transient Loss of Cdc28 Function in S. cerevisiae cdc28-4 or cdc28-13 Mutant Cells Correlates the Absence of Kinase Activity with Cdc6 Protein Stabilization-An advantage of using these two G 1 -deficient alleles of CDC28 is that they show different levels of kinase activity under permissive or restrictive conditions (Fig. 8A). In fact, on measuring the in vitro kinase activity immunoprecipitable from cdc28-4 or cdc28-13 FIG. 5. A p40 SIC1 -induced drop in CDK-associated Clb-kinase activity causes mitotic-arrested cells to accumulate p57 CDC6 . A, cultures in raffinose-based media of wild-type and congenic GAL1-10: SIC1 or cdc6-1 GAL1-10:SIC1 cells in log phase at 25°C were arrested in mitosis, incubated for 2.5 h with nocodazole, split into two aliquots, and further incubated in the presence of the drug at the permissive (25°C) or the restrictive (37°C) temperature. After 40 min of induction in galactose, the expression of the CDK inhibitor p40 SIC1 was repressed. Samples were taken at the indicated intervals and processed for DNA content analysis. Asynchronous (asynch.) cultures are shown in order to indicate the relative positions of 1C and 2C DNA content peaks. B, left panels, immunoblot analysis of cell extracts from the GAL1-10:SIC1 strain arrested in mitosis with nocodazole. After 2.5 h of incubation in the presence of the drug, p40 SIC1 was induced by adding galactose to the medium. After galactose addition, samples were taken at the indicated intervals and split into two aliquots for immunoprecipitation and Western analysis. 100 g of total protein from cell lysates were loaded in each lane and resolved by SDS-polyacrylamide gel electrophoresis, transferred to nitrocellulose membranes, and probed with rabbit ␣-Cdc6 or ␣-Cdc28 antibodies, the latter for a loading control. A protein extract from cdc28-4 cells blocked for 4 h at 37°C is also shown for comparison (C). Right panels, anti-Cdc28 immunoprecipitates obtained from half of the above protein samples were split into two aliquots for H1 kinase activity assays and p34 CDC28 immunoblot analysis for reference to the total amount of Cdc28p protein immunoprecipitated.
FIG. 6. Mitotic-arrested Cdc6-7-expressing cells rereplicate, comparably with wild-type cells, only after a drop in Cdc28 kinase activity. A culture in raffinose-based media of cdc6-1 GAL1-10:SIC1 pHR70cdc6-7 cells in log phase at 25°C was arrested in mitosis (nocodazole, 2.5 h), split into four parts, and further incubated in the presence of the drug at the permissive (25°C) or the restrictive (37°C) temperature (two aliquots at 25°C and the other two at 37°C). In one of the cultures at 25 and 37°C, SIC1 expression was induced for 40 min, and then glucose was added to repress the GAL1-10 promoter. Samples were taken at indicated intervals and processed for DNA content analysis. An asynchronous (Asyn.) culture is shown to indicate the relative positions of 1C and 2C DNA content peaks. cell extracts at 25 and 37°C, we found significant differences between the two strains for the same amount of immunoprecipitated Cdc28 protein (Fig. 8A), temperature conditions at which the wild-type Cdc28 shows no difference (Fig. 1A). If Cdc28 does regulate the stability of Cdc6, it could be predicted that alleles retaining less kinase activity should accumulate more Cdc6 protein. As shown above, both cdc28-4 and cdc28-13 accumulated significant levels of Cdc6 at the restrictive temperature, but to detect similar amounts in both mutant backgrounds, cdc28-13 had to be incubated at 38°C (Fig. 7A). We therefore sought to determine whether a pulse of limited duration at 37°C (a generation time) would cause a different degree of accumulation of Cdc6 protein in cdc28-4 compared with cdc28-13 S. cerevisiae mutant cells and whether Cdc6 would be more stable on a return to 25°C in cdc28-4 mutant cells, given that this allele shows lower p34 CDC28 -associated kinase activity. At this temperature cdc28-13 mutant cells accumulated less Cdc6 protein than did cdc28-4 (Fig. 8B), consistent with the leakiness of the former mutant allele (42). After release at 25°C, most p57 CDC6 had disappeared by ϳ120 min in cdc28-13 cells (Fig. 8B, upper panel). The behavior of the strain bearing the cdc28-4 temperature-sensitive mutant allele was quite different. At the restrictive temperature Cdc6 accumulated at higher levels, and after release the protein became stabilized for up to 180 min and was still, although barely, detectable after 260 min (Fig. 8B, lower panel). Again the accumulation of the Cdc6 protein was attributable to a posttranscritional mechanism of stabilization, because CDC6 mRNA did not accumulate during the experiment, as checked by Northern analysis (data not shown). These experiments show that cdc28-4 cells accumulate more p57 CDC6 protein than cdc28-13 cells and suggest a correlation between low kinase activity and Cdc6 protein stabilization in G 1 .

DISCUSSION
Control Regulating Cdc6 Proteolysis in S. cerevisiae-Biochemical and genetic studies have strongly suggested that the Cdc6 DNA replication protein is ubiquitinated for degradation in each cell cycle at the beginning of the S phase by a Cdc4/ Cdc34/Cdc53-dependent mechanism (23,27,28). Although Cdc28-mediated phosphorylation of key regulatory proteins is known to play a role in regulating their Cdc4/Cdc34/Cdc53 ubiquitination, as is the case for Sic1 and Cln2 (39,43,44), to our knowledge until now the role of Cdc28/Clb-mediated Cdc6 phosphorylation in S. cerevisiae remains to be elucidated.
In this report we have described a regulatory mechanism controlling Cdc6 protein degradation. Several of the lines of evidence presented here suggest that one consequence of the rise in Cdc28 activity during the G 1 -to-S phase transition is the phosphorylation of Cdc6 protein for targeting it for proteolysis. In this regard, we have shown that under restrictive conditions, in which the Cdc28 kinase is thermolabile, p57 CDC6 accumulates in cdc28-4 and cdc28-13 ts alleles (Figs. 7A and 8), mutants defective in Cdc28-kinase activity (40,42). Nevertheless, the accumulation of a protein could be achieved by regulating its turnover or its synthesis. Northern analysis of the CDC6 mRNA abundance (Fig. 7B) rules out a role for the accumulation of the CDC6 transcript in the regulation of the Cdc6 protein abundance. These lines of evidence suggest that Cdc6 is stabilized in S. cerevisiae cells devoid of a functional p34 CDC28 kinase because of a post-transcriptional mechanism. Because cdc28-4 accumulates more Cdc6 protein than cdc28-13 (Fig. 8), our experiments with these two cdc28 ts alleles also FIG. 7. Cdc6 accumulates in cdc28-4 and cdc28-13 pre-STARTarrested cells. A, immunoblot analysis using affinity-purified rabbit ␣-Cdc6 antibody with cell extracts from cdc28-4 (left) or cdc28-13 (right) arrested at their restrictive temperature. Samples were taken at the indicated intervals from midlog phase-growing cultures after being changed to the restrictive temperature. A cdc6::URA3 control sample was obtained by repressing a strain deleted for CDC6 and carrying GAL1::CDC6 in glucose at 37°C for 2 h. A cross-reacting yeast protein appears in the Western analysis that is also detectable in a cdc6-deleted strain. B, analysis of CDC6 transcript in cdc28-4 and cdc28-13 strains arrested at 37°C. Total RNA was prepared from asynchronous wild type cells (Asyn.), cells overexpressing CDC6 from the GAL1-10 promoter (opCDC6), or the indicated temperature-sensitive cdc mutants growing either asynchronously at 25°C (0H) or after arrest by shift to the restrictive temperature at the indicated intervals. Ten micrograms of total RNA from each sample were separated on 1% formaldehydeagarose gels and analyzed by Northern blot using CDC6, CLN1, or LEU2 as probes (the latter for a loading control). Note the accumulation of the CLN1 mRNA during the cell cycle arrest as described previously for cdc28-4 (41).
FIG. 8. cdc28ts mutants correlate the levels of CDK-associated kinase activity with the degree of Cdc6 protein stabilization. A, H1 kinase activity in cdc28-4 and cdc28-13 mutant yeast strains. Anti-Cdc28 immunoprecipitates obtained from exponentially growing cdc28ts cells at 25°C were split into two aliquots, and H1 kinase activity was assayed at 25 or 37°C. An immunoblot of p34 CDC28 of aliquots of these samples is included for reference to the total Cdc28p protein immunoprecipitated. B, Western analysis of cdc28-13 and cdc28-4 cells blocked at 37°C and released at 25°C. Midlog phase cells growing on YEPD at the permissive temperature were shifted to 37°C for 90 min and then released at 25°C. Samples from indicated intervals (minutes) were processed for Western analysis, so that lysates from these samples were separated on SDS-polyacrylamide gels, transferred to Hybond-nitrocellulose, and probed with either rabbit ␣-Cdc6 or ␣-Cdc28 antibodies.
suggest a direct correlation between the lack of Cdc28 kinase and Cdc6 protein stabilization (Figs. 7 and 8). Nevertheless, the accumulation of a given protein could be an indirect consequence of the cell cycle defect of the mutant being analyzed. In fact, wild-type Cdc6 protein accumulates in cdc28-13 and cdc28-4 mutant strains of S. cerevisiae, cdc28 ts alleles that block cell cycle progression in G 1 (45,46), but not in cdc28-1N (data not shown), a G 2 -defective allele (40,47). Nonetheless, our data regarding Cdc6 accumulation in mitotic-arrested cells by inducing a reduction in Cdc28 kinase activity (Fig. 5) suggest that the initiator protein stabilizes because of a genuine defect in p34 kinase. Moreover, our approach to examining whether Cdc28 affects or controls Cdc6 stability by analyzing the rate of its disappearance in promoter shut off experiments showed that nocodazole-arrested cells require Cdc28 to degrade Cdc6 protein (Fig. 2). Taken together, these data and the biochemical evidence described above suggest that a Cdc28-dependent mechanism targets Cdc6 for degradation. Consistent with this hypothesis, the mutation of seven of eight or all Cdc28 consensus phosphorylation sites results in Cdc6 protein stabilization (Fig. 3). In fact, certain data not included in this report suggest that Cdc6 has two different "hot" Cdc28 consensus phosphorylation locations, each involving two phosphorylatable residues and hence playing a role in protein stability, because double mutations in residues T39A and S43A or T368A and S372A result in mutant forms of Cdc6 protein that are considerably more stable than wild-type Cdc6 (to be described elsewhere).
Finally, it is worth mentioning that p57 CDC6 is ubiquitinated for degradation through the Cdc4/34/53 pathway (27,28), because it has been argued that this complex may act to specifically target the degradation of proteins involved in cell cycle control in a Cdc28-dependent manner of phosphorylation (48). Our work, together with the above observations regarding Cdc6 proteolysis (27,28), indicates that Cdc28 phosphorylates Cdc6 and, on doing so, targets it for degradation through the ubiquitin-mediated proteolysis pathway. Yet some important details regarding Cdc6 protein phosphorylation remain to be elucidated. For example, at present it is unknown whether the CDK consensus posphorylation sites present in Cdc6 are in vivo substrates for Cdc28. However, given the interactions between Cdc6p and Cdc28 (Ref. 22 and Fig. 1A) and the results presented here with Cdc6-7 and Cdc6-8 mutant proteins (Fig.  3), we think this is most likely to be the case.
Does Cdc6 Proteolysis Play Any Role in Genome Ploidy Maintenance?-The experiments presented in this paper strongly suggest that Cdc6 is degraded by a Cdc28-dependent mechanism, given that elimination of the CDK consensus phosphorylation sites of Cdc6 results in the stabilization of the initiator protein (Fig. 3). On top of that, the mutants lacking these phosphorylation sites are fully functional (Figs. 4 and 6), because they complement the deletion of CDC6 (Fig. 4), and fully stable. Our results are consistent with those presented previously regarding a truncated form of Cdc6 in which the Cdc4 interaction domain has been deleted (28), resulting in a stable protein also able to promote DNA synthesis. Thus, it is reasonable to ask why the Cdc6 protein is unstable. Cyclin B-CDK complexes prevent rereplication during G 2 and M phases by inhibiting reformation of pre-RCs at replication origins (7). At the same time, Cdc6 is an absolute requirement for DNA replication, because in the absence of this initiator protein no pre-RCs are formed (16). Furthermore, normal S. cerevisiae cells degrade Cdc6 every cell cycle at the beginning of S phase (19,28), most probably in a Cdc28-dependent manner (this work). Taken together these lines of evidence suggest that budding yeast Cdc28-controlled Cdc4/34/53-mediated Cdc6 proteolysis is part of a redundant CDK-mediated mechanism that prevents the reinitiation of DNA replication within the same cell cycle.
Role of CDK-mediated Phosphorylation of the Cdc6/Cdc18 Class of Proteins in Eukaryotes-The closest homolog of Cdc6 in fission yeast is the cdc18ϩ gene-encoded product. Both in fission and budding yeast, gene expression of cdc18ϩ and CDC6 are transcriptionally regulated, and both Cdc18 and Cdc6 proteins levels vary during the cell cycle, peaking at the G 1 -to-S phase transition (18,19,49,50). In these unicellular eukaryotes, de novo Cdc18 and Cdc6 protein synthesis is critical for the formation of pre-RCs required for the initiation of DNA replication. Cdc18 phosphorylation by CDKs has recently been shown to play a role in targeting the initiator protein for degradation at the G 1 -to-S phase transition (25,26). Thus, from the biochemical evidence presented here, and given that Cdc18 and Cdc6 are ubiquitinated for proteolysis in fission and budding yeast (27,51), the Cdc28/2-mediated mechanism by which these simple eukaryotes control Cdc18/Cdc6 stability appears to be conserved. Although human Cdc6 is also a substrate for CDKs, the consequences of such phosphorylation seem to be quite different. It has recently been reported that CDK phosphorylation of human Cdc6 regulates its subcellular localization throughout the cell cycle (52). Unphosphorylated protein is nuclear during G 1 , whereas CDK-phosphorylated Cdc6 localizes at the cytoplasm along the rest of the cell cycle.