Mcm3 Is Polyubiquitinated during Mitosis before Establishment of the Pre-replication Complex*

To ensure fidelity in genome duplication, eukaryotes restrict DNA synthesis to once every cell division by a cascade of regulated steps. Central to this cascade is the periodic assembly of the hexameric MCM2–7 complex at replication origins. However, in Saccharomyces cerevisiae, only a fraction of each MCM protein is able to assemble into hexamers and associate with replication origins during M phase, suggesting that MCM complex assembly and recruitment may be regulated post-translationally. Here we show that a small fraction of Mcm3p is polyubiquitinated at the onset of MCM complex assembly. Reducing the rate of ubiquitination byuba1-165, a suppressor of mcm3-10, restored the interaction of Mcm3-10p with subunits of the MCM complex and its recruitment to the replication origin. Possible roles for ubiquitinated Mcm3p in the assembly of the MCM complex at replication origins are discussed.

The accurate and timely replication of DNA is essential to the faithful propagation of cell identity in eukaryotes. Crucial to this regulation is the MCM (minichromosome maintenance) complex containing six highly conserved subunits, . Dynamic changes in the assembly and disassembly of the MCM complex at replication origins are critical for restricting DNA replication to once per cell division cycle (2)(3)(4). Upon recruitment to replication origins, a series of events must take place before the hexameric MCM complex is converted to an active helicase (5)(6)(7)(8)(9). Achieving the final state of replication competence involves multiple factors, including cyclin-dependent kinases (10,11) and the Cdc7-Dbf4 kinase (12,13).
Regulating the activity of the MCM proteins is complicated further by the fact that all six MCM proteins are estimated in tens of thousands of copies/cell in Saccharomyces cerevisiae (14,15), ϳ100 -500 times more than the number of replication origins (16). However, only a fraction of each MCM protein associates with chromatin from late M phase to the beginning of S phase (6,(17)(18)(19)(20)(21). The mechanism for selecting this particular fraction is unknown. It is possible that only a small fraction of MCM proteins acquire the ability to initiate DNA replication after being modified during M phase. Indeed, posttranslational modifications of the MCM proteins have been reported. MCM proteins in different subcellular locations and at different cell cycle stages exist in multiple phosphorylated states (19,22,23). The cyclin B-dependent Cdc28 kinase appears to regulate the nuclear accumulation of the MCM proteins during G 2 /M phase (21). The Cdc7-Dbf4 kinase phosphorylates Mcm2p to activate the MCM complex after it has been recruited to replication origins (12,24,25). A recent report identified a novel acetyltransferase in human that specifically acetylates chromatin-bound Mcm3p in vivo and may inhibit DNA replication (26).
Mcm3p is probably the most abundant of all the MCM proteins, present in ϳ200,000 copies/cell (14,27). In this study, we show that a small fraction of Mcm3p is polyubiquitinated during M phase in S. cerevisiae. Ubiquitin is a highly conserved 76-residue protein that covalently modifies other proteins in both the cytoplasm and nucleus of eukaryotic cells. Its complex structure bears information that specifies various functions to different protein targets (28). A series of enzymatic reactions catalyzes ubiquitination of target proteins. Free ubiquitin is first activated by the ubiquitin-activating enzyme (E1) 1 in an ATP-dependent step. The activated ubiquitin is then transferred to one of many ubiquitin-carrier proteins (E2) and ubiquitin-protein isopeptide ligases (E3) that confer substrate specificity. Upon transfer of the first ubiquitin to the substrate, formation of a polyubiquitin chain frequently follows via an isopeptide linkage between the ⑀-amino group of one of the lysine residues in one ubiquitin and the carboxyl-terminal glycine of the next ubiquitin (28,29).
The Mcm3-10 mutant protein is defective in its interactions with Mcm5p as well as in the assembly of the MCM2-7 complex at replication origins (30). Through suppressor analysis, we isolated uba1-165, a mutation in the E1 enzyme that induces DNA replication defects and suppresses the growth and replication initiation defect of the mcm3-10 mutant. Furthermore, Mcm3p polyubiquitination is reduced in this uba1-165 mutant. Polyubiquitination of proteins is known to target selected proteins for proteolysis by the 26 S proteasome (29,31 (14), and an excess amount of Mcm3p results in DNA replication defects (32). Targeted proteolysis of a select pool of Mcm3p by ubiquitination degradation during M phase may be one of several regulatory steps that modulate the periodic assembly of pre-replication complexes in the initiation of DNA synthesis.
Protein extracts and immunoprecipitates were electrophoresed using 6% SDS-polyacrylamide gel. For Western blotting, 40 g of total protein extracts from each sample were transferred to Immobilon-P membranes (Millipore Corp.) and incubated with a specific antibody. Horseradish peroxidase-conjugated anti-rabbit or anti-mouse secondary antibodies (Bio-Rad) and a chemiluminescent detection kit (Pierce) were used for protein detection.
Mcm3p Ubiquitination Analysis-To analyze Mcm3p ubiquitination in vivo, 50 ml of SUB280 and SUB280-221 cells were grown to A 600 nm ϭ 0.8. Protein extracts with the proteasome inhibitor MG132 were incubated with 1 l of anti-Mcm3p antibody or 10 l of anti-Myc antibody (9E10, Santa Cruz Biotechnology) at 4°C. After incubation overnight with antibodies, protein A-Sepharose beads (Sigma) were added for another 2 h at 4°C. The resin was then washed five times with radioimmune precipitation assay buffer (1% sodium deoxycholate, 1% Nonidet P-40, 0.1% SDS, 150 mM NaCl, and 50 mM Tris, pH 8.0), and proteins were eluted with 40 l of SDS-PAGE sample buffer.
Chromatin Immunoprecipitation-Formaldehyde cross-linking was performed as described by Hecht et al. (34) with modifications. Strains were grown in complete medium to A 600 nm ϭ 0.6. 40 ml of cells were cross-linked and immunoprecipitated with 0.5-10 l of specific antibodies. The immunocomplexes were washed two times with lysis buffer (50 mM HEPES, pH 7.5, 140 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, and Roche Complete protease inhibitor mixture), two times with lysis buffer plus 0.5 M NaCl, two times with wash buffer (10 mM Tris, pH 8.0, 0.25 M LiCl, 0.75% Nonidet P-40, 0.5% sodium deoxycholate, and 1 mM EDTA), and two times with Tris/EDTA. Protein-DNA cross-links were reversed overnight at 65°C in 150 l of elution buffer (50 mM Tris, pH 8.0, 10 mM EDTA, and 1% SDS). DNA samples were purified on PCR purification columns (QIAGEN Inc.) and eluted with 40 l of Tris/EDTA. PCR was carried out under the following conditions: two cycles at 94°C for 2 min, 96°C for 1 min, and 54°C for 4 min; 28 cycles at 94°C for 1 min and 54°C for 1.5 min; and one cycle at 70°C for 5 min. PCR products were incubated with a 1:10,000 dilution of Vistra Green (Amersham Biosciences) and separated on 3% agarose gel. Quantitation of fluorescence signals was performed on a Molecular Dynamics Storm 840 PhosphorImager, and quantitative analysis was performed using ImageQuant software. -Fold enrichment of origin binding was calculated as the ratio of ARS1 to ACT1 (or ATP11) precipitated by a specific antibody and was normalized to the ratio of ARS1 to ACT1 (or ATP11) precipitated by a control antibody.
␤-Galactosidase Stability Assay-Cells carrying pUB23L were grown in glucose medium to A 600 nm ϭ 0.3 before shifting to galactose medium for 1 h to induce expression from the GAL promoter. Cells were resuspended in glucose medium pre-equilibrated at 13°C. ␤-Galactosidase assays were performed using a published procedure (35). Triplicate 0.2-ml samples were taken from each culture and placed in 0.4 ml of Z buffer (Na 2 HPO 4 ⅐7H 2 O 16 g/liter, NaH 2 PO 4 ⅐H 2 O 5.5 g/liter, KCl 0.75 g/liter, MgSO 4 ⅐7H 2 O 0.2 g/liter, ␤-mercaptoethanol 2.7 ml/liter) at 13°C. After a 0.5-h incubation, 0.15 ml of 4 mg/ml o-nitrophenyl-␤-Dgalactopyranoside were added. Reactions were stopped with 0.2 ml of 2 M Na 2 CO 3 upon detection of a visible yellow color.
Mcm3-td Degron Experiment-YKL44 cells were arrested at 24°C in YP medium ϩ raffinose ϩ 0.1 mM CuSO 4 by ␣-factor or nocodazole before shifting to prewarmed YP (yeast extract peptone) medium ϩ galactose (no CuSO 4 ) at 37°C to induce degradation of the Mcm3-td protein. The cultures were then returned to 24°C in YPD (YP ϩ dextrose) medium ϩ 0.1 mM CuSO 4 for 30 min to allow resynthesis of Mcm3-td under the same arrest conditions.
Other Methods-Plasmid stability assay was performed as described by Lei et al. (14). Flow cytometric analysis was performed as described (36).

A Mutation in the E1 Enzyme Suppresses Growth and Replication Defects in the mcm3-10 Mutant-Mcm2p-Mcm7p
interact to form the presumed replicative helicase that melts origin DNA and unwinds growing forks when cells transition from G 1 to S phase. Mcm3p is believed to be a regulatory subunit of this MCM2-7 helicase (7,9,37,38). One of the MCM3 mutant alleles, mcm3-10, had a growth defect at 37°C (Fig. 1A) and a minichromosome maintenance (mcm) defect at 30°C (Table I)  To better understand the nature of the mcm3-10 defect at a molecular level, spontaneous suppressors of mcm3-10 were selected. One of the suppressors was identified as uba1-165, which rescued the mcm3-10 growth defect at 37°C (Fig. 1A) and the mcm defect at 30°C (Table I), but acquired a coldsensitive growth phenotype in the single and double mutants at 13°C (Fig. 1A). Suppression of mcm3-10 by uba1-165 was allele-specific. The mcm3-1 mutant, which is defective in a step subsequent to the assembly of the MCM complex at replication origins (30,39) and whose defective gene product interacts well with Mcm5p (30), was not suppressed by uba1-165 (data not shown).
The Uba1p is involved in the first step of the ubiquitination pathway by activating ubiquitin. A defect in this enzyme is pleiotropic, affecting the turnover rates of a large number of proteins at different stages of the cell cycle (31,40). Surprisingly, at 13°C, the majority (ϳ85%) of the uba1-165 mutant cells arrested growth at G 1 phase, showing large buds with unreplicated DNA (Fig. 2A). The uba1-165 mutant also exhibited phenotypes of plasmid instability (Table I), consistent with defects in DNA replication. Because the E1 enzyme affects the turnover rates of a large number of proteins, including those that regulate DNA replication and cell cycle progression, replication defects associated with uba1-165 are to be expected.
Reducing E1 Activity or the Rate of Polyubiquitin Chain Formation Corrects the Growth and Minichromosome Maintenance Defects of mcm3-10 -To characterize the defect of uba1-165, the mutant allele was sequenced, and the mutation was identified as R603P in the highly conserved catalytic center of the E1 enzyme (Fig. 2B). Arg 603 is proximal to the only cysteine (Fig. 2B, black boxes) that forms a thiol ester linkage with ubiquitin (41). The efficiency of the ubiquitination pathway was examined in vivo by measuring the stability of the Ub-L-lacZ reporter, in which a single ubiquitin is fused N-terminal to lacZ (42). At 13°C, a nonpermissive condition for uba1-165, the reporter was more stable in the uba1-165 mutant than in the wild-type or the control cold-sensitive DNA replication mutant cdc45-1 (Fig. 2C) (43). Therefore, the uba1-165 mutation induces a deficiency in the ubiquitination pathway.
To investigate whether inefficient polyubiquitination is the basis for the suppression of mcm3-10 by uba1-165, various lysine-to-arginine substitution mutants and wild-type ubiquitin were overexpressed in the mcm3-10 strain. Ubiquitin has seven lysine residues, all of which are capable of mediating linkages to form polyubiquitin chains in vitro. The functions of each of these different linkages of polyubiquitin chains have been investigated in budding yeast (44 -46). We found that overexpression of Ub(K48R) suppressed mcm3-10 at 37°C (Fig. 2D). Furthermore, minichromosomes were stabilized in mcm3-10 expressing Ub(K48R), a phenotype also observed in the mcm3-10 uba1-165 mutant (Table I). However, expression of the other six lysine-to-arginine ubiquitin mutants did suppress the mcm3-10 growth defect or the mcm defect (data not shown). Ub(K48R) blocked the attachment of additional ubiquitin to a growing Lys 48 -linked chain in a reversible reaction. Expressing Ub(K48R) in addition to the endogenous ubiquitinencoding genes (UBI1-4) reduced the rate of ubiquitin polymerization through Lys 48 . Lys 48 -linked polyubiquitin chain has been implicated in signaling protein degradation (45). Therefore, defects of the mcm3-10 mutant can be suppressed by reducing either the rate of Lys 48 -linked polyubiquitin chain growth or the pool size of activated ubiquitin.
Mcm3p Is Ubiquitinated in Vivo-There are a number of  possible explanations for how defects in the ubiquitination pathway may suppress the growth defect of mcm3-10 and destabilize minichromosomes. One possible scenario is that Mcm3p is modified by ubiquitination and that reducing the rate of Mcm3p ubiquitination corrects the replication defects of mcm3-10. Mcm3p in a crude lysate of a wild-type strain appeared predominantly as a 120-kDa protein with a few minor higher molecular mass species when analyzed by Western blotting (Fig. 3A, lane 1) (lower panel). B, shown is the sequence alignment of the E1 enzyme in S. cerevisiae and other organisms. The uba1-165 mutation is an R603P substitution located in a highly conserved region of the E1 enzyme. The essential cysteine residue that forms the thiol ester linkage with ubiquitin is highlighted by black boxes. C, uba1-165 was defective in the catalytic activity of the E1 enzyme. The stability of the GAL-controlled N-terminal end rule reporter substrate Ub-L-LacZ was measured in the uba1-165 (f), cdc45-1 (q), and wild-type (Wt; OE) strains. Reporter protein expression was induced in galactose medium and then stopped by transferring to glucose medium at 13°C. The decay curve after shifting to glucose medium was determined by ␤-galactosidase (␤-Gal) assay. D, wild-type and mcm3-10 cells were transformed with plasmids containing either wild-type ubiquitin (Ubwt) or the chain terminator mutant Ub(K48R). The plates were incubated first at 37°C and then at 30°C for 2 days. inhibits 26 S proteasome activity, was used. An asynchronous culture of a MG132-permeable strain (ise1) (47) was treated with this compound for 2 h. There was no significant change in the overall level of Mcm3p in the presence or absence of the proteasome inhibitor MG132 in these cells (Fig. 3C, (Fig. 4A, lane 3), but not in G 1 or S phasearrested cells (lanes 1 and 2) accumulation of Mcm3p at M phase when proteasome activity is inhibited was investigated. The pre1-1 pre4-1 strain is defective in two of the subunits of the 26 S proteasome, resulting in reduced ubiquitin/proteasome degradation at the restrictive temperature of 37°C (48). Wild-type (PRE ϩ ) and proteasome mutant (pre Ϫ ) strains were first arrested at M phase by nocodazole at 30°C and then shifted to 37°C for 2 h. Protein extracts were analyzed by SDS-PAGE and Western blotting. An increase in the level of Mcm3p of ϳ1.6-fold was detected in the M phase-arrested pre1-1 pre4-1 mutant (pre Ϫ ) compared with the wild-type strain (PRE ϩ ) (Fig. 4B) M phase in S. cerevisiae can be viewed as either the beginning or the end of a replication cycle. Early M phase is when replication has been completed, and negative signals that prevent over-replication must be removed (52). Late M phase is when the pre-replication complex is assembled for the next round of DNA synthesis ( Cells were arrested at metaphase by depletion of Cdc20p, an activator of the anaphase-promoting complex (APC) (53), and then allowed to resume progression of the cell cycle by induction of Cdc20p synthesis. Cell samples were collected at 20-min intervals. FACS analysis showed that cells were in M phase until 60 min after resuming cell cycle progression (Fig. 5A). Western blot analysis showed that ubiquitination of Mcm3p was apparent at the metaphase-anaphase transition and was most prominent at ϳ20 min after release from arrest (Fig. 5B). ChIP indicated that most of the Mcm3p was recruited to ARS1 between 20 and 40 min after release from arrest and remained associated with ARS1 throughout M phase and G 1 phase (Fig.  5C). The timing of events is consistent with the ubiquitination of Mcm3p occurring during M phase just before or as Mcm3p is recruited to ARS1.

Abolishing Mcm3p Ubiquitination Has No Effect on Recruitment of Subunits of the MCM Complex to ARS1-To distinguish whether ubiquitination positively or negatively regulates
Mcm3p-origin binding, we examined the binding of the MCM proteins to ARS1 in uba1-165. We have shown that the ubiquitinated form of Mcm3p was significantly reduced under this condition in uba1-165 (Fig. 3B, lane 7). As a control, we in-

DISCUSSION
The initiation of DNA synthesis at replication origins involves two temporally separable steps: the assembly and recruitment of the hexameric MCM complex and the activation of the MCM complex to form a helicase (54). Mcm3p is believed to be a regulatory subunit of the MCM helicase. A mutant allele (mcm3-10) arrests growth in late G 1 phase at 37°C and exhibits a minichromosome maintenance defect at 30°C. Mcm3-10p has reduced affinity for other Mcm proteins and prevents recruitment of the MCM2-7 complex to replication origins (30).
To understand in greater detail the process of assembly of the MCM complex, we isolated cold-sensitive suppressors of the mcm3-10 mutant. An R603P mutation in the E1 enzyme of the ubiquitin pathway suppressed the growth and mcm defect of mcm3-10. Furthermore, expression of Ub(K48R), which reduced the efficiency of proteolysis by the 26 S proteasome, also suppressed the growth and minichromosome maintenance defects of the mcm3-10 strain. These results prompted us to investigate whether Mcm3p is modified by ubiquitin. We showed that a small population of Mcm3p was polyubiquitinated transiently during M phase. The ubiquitination of nascent Mcm3p synthesized during M phase suggests that if Ub-Mcm3p plays a role in regulating DNA replication, it is more likely to play a role in establishing the next replication cycle than in removing expended Mcm3p from the previous replication cycle.
A primary determinant of substrate specificity in the ubiquitination pathway is the E2-E3 complex. One of the E3 complexes, the APC, activated only from M phase to G 1 phase, eliminates anaphase inhibitors, mitotic cyclins, and components of the mitotic spindle (55). The ubiquitination of Mcm3p occurs at approximately the same time that the APC is activated. However, Mcm3p was ubiquitinated in the absence of Cdc20p (Fig. 4), the first activator associated with the APC. Unless the APC has other unknown activators that act before Cdc20p, ubiquitination of Mcm3p is most likely not catalyzed by the APC. The only E3 enzyme in the literature that may relate to the MCM proteins is the human E6-associated protein, which associates with human papilloma virus protein 18E6 and acts as an E3 enzyme that targets human MCM7 for ubiquitination. Both Mcm7p and Mcm3p are co-immunoprecipitated with anti-human papilloma virus protein 18E6 antibody (56). Therefore, potential E6-associated protein homologs in yeast are good candidates for Mcm3p-specific ubiquitin ligases. K7428 cells (GAL-CDC20 ⌬cdc20) were synchronized in metaphase in glucose medium for 3 h before release into galactose medium. Samples were collected at 20-min intervals after shifting to galactose medium and subjected to the following analyses. A, cells were subjected to FACS analysis. B, the Mcm3p blot was exposed for 2 min (upper panel) or 10 s (middle panel). Actin was used as a loading control (lower panel). C, Mcm3p cross-linked to ARS1 DNA was analyzed by ChIP. Anti-hemagglutinin (HA) antibody was used as a negative control for immunoprecipitation (IP), and ACT1 was used as a control for non-origin DNA. -Fold enrichment of Mcm3p association was calculated as the ratio of ARS1 to ACT1 and was normalized against that of the hemagglutinin control.
The molecular mechanisms to explain the suppression of mcm3-10 by uba1-165 and the replication defect associated with uba1-165 require further study. A trivial explanation is that Mcm3-10p might become misfolded and quickly degraded by the ubiquitin/proteasome pathway (57). A lesion in the ubiquitination pathway would slow down the degradation process and effectively stabilize Mcm3-10p. However, a previous study showed that the stability of Mcm3-10p is comparable to that of Mcm3p (30). Another explanation is that a compromised ubiquitin degradation system may result in an accumulation of targeted substrates that play a direct or indirect role in replication because all ubiquitinated proteins are dependent on Uba1p function. Delaying ubiquitin-mediated degradation results in the accumulation of cyclins, cyclin-dependent kinase inhibitors, and anaphase inhibitors, thus extending the cell cycle and allowing adequate time to assemble Mcm3-10p into the MCM2-7 complex. An additional explanation, which warrants elaboration, is that ubiquitination degradation modulates the size of a select pool of Mcm3p destined for complex assembly. In this hypothesis, the reduced affinity of Mcm3-10p for other subunits of the MCM complex is compensated for by an increase in this select pool of Mcm3-10p. A prime example of pool regulation by targeted degradation is demonstrated by the selective degradation of differentially modified Sic1, an inhibitor of cyclin-dependent kinase, by the ubiquitination pathway in response to the phosphorylation states of Sic1p molecules (58).
In S. cerevisiae, only a small fraction of Mcm3p is periodically assembled into the hexamer (14,16,19) at the beginning of late M phase. This study suggests that Mcm3p is ubiquitinated during M phase, just before or as the MCM complex is assembled at replication origins. Because the ubiquitination pathway is not required for the recruitment of the MCM complex to replication origins, ubiquitination of Mcm3p may regulate the assembly of the MCM complex. Previous studies have indicated that the level of Mcm3p is 2-5 times more abundant than that of other MCM proteins (14). Furthermore, overexpression of Mcm3p, although not Mcm2p, destabilizes minichromosomes (39,59). On the other hand, a reduced dosage of other MCM subunits (Mcm2p, Mcm6p, and Mcm7p), but not Mcm3p, destabilizes minichromosomes (14). 2 These results suggest that a stoichiometric imbalance in the MCM subunits (and in particular, an excess of Mcm3p) may lead to the formation of nonproductive MCM complexes. It is possible that the ubiquitin degradation pathway down-regulates Mcm3p at M phase to maintain a stoichiometric balance during MCM complex assembly. We favor the idea that a pool of Mcm3p may be pre-selected for assembly into the MCM2-7 complex by an unknown mechanism and that ubiquitin/proteasome-mediated degradation may restrict the size of this select pool of Mcm3p during MCM complex assembly (Fig. 7). An excess amount of Mcm3p due to either overexpression of Mcm3p or a defect in the ubiquitination pathway may cause the formation of nonproductive MCM complexes, which may assemble at an origin, but fail to initiate DNA synthesis. We envision that ubiquitin/ proteasome-directed proteolysis ensures a particular pool size of Mcm3p during M phase for the next round of DNA synthesis.
The model elaborated here is not the only plausible model consistent with a role for Ub-Mcm3p in MCM complex assembly. Other models such as targeted recruitment and/or stabilization of the MCM hexamer by Ub-Mcm3p cannot be ruled out. Post-translational modifications of the MCM proteins such as the phosphorylation and polyubiquitination of Mcm3p may hold the key for unraveling the secret of assembling an active hexameric helicase, a feat that has yet to be reconstituted in 2 Y. Kawasaki and M. Fitch, unpublished data. vitro (7,9,37,38). Further studies on how this small fraction of Mcm proteins may be selected and the consequence of a mutation in MCM3 that prevents ubiquitination will be central to our understanding of the many functions of the Mcm proteins.