Two mcm3 mutations affect different steps in the initiation of DNA replication.

Mcm3 is a subunit of the hexameric MCM2-7 complex required for the initiation and elongation of DNA replication in eukaryotes. We have characterized two mutant alleles, mcm3-1 and mcm3-10, in Saccharomyces cerevisiae and showed that they are defective at different steps of the replication initiation process. Mcm3-10 contains a P118L substitution that compromises its interaction with Mcm5 and the recruitment of Mcm3 and Mcm7 to a replication origin. P118 is conserved between Mcm3, Mcm4, Mcm5, and Mcm7. An identical substitution of this conserved residue in Mcm5 (P83L of mcm5-bob1) strengthens the interaction between Mcm3 and Mcm5 and allows cells to enter S phase independent of Cdc7-Dbf4 kinase (Hardy, C. F., Dryga, O., Pahl, P. M. B., and Sclafani, R. A. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 3151-3155). Mcm3-1 contains a G246E mutation that diminishes the efficiency of replication initiation (Yan, H., Merchant, A. M., and Tye, B. K. (1993) Genes Dev. 7, 2149-2160) but not its interaction with Mcm5 or recruitment of the MCM2-7 complex to replication origin. These observations indicate that Mcm3-10 is defective in a step before, and Mcm3-1 is defective in a step after the recruitment of the MCM2-7 complex to replication origins.

To ensure precise duplication of the nuclear genome, the initiation of DNA synthesis from each replication origin is restricted to no more than once in a mitotic cell cycle. At the heart of this elaborate regulatory mechanism is the temporal separation of the establishment of the pre-replication complexes (pre-RC) 1 at replication origins from late M through G 1 phase and the activation of the pre-RC to initiate DNA synthesis in S phase (3,4). A key component of the pre-RC is a complex of six conserved proteins known as the MCM2-7 complex (5). The recruitment of the MCM2-7 complex to an origin, a process often referred to as "replication licensing" (6), signals the completion of the pre-RC assembly and establishes the competence of the origin for the initiation of DNA synthesis (3,7,8). The initiation of DNA synthesis is activated by a series of phosphorylation events (9 -11) that result in the conversion of the MCM2-7 complex into an active helicase (12,13). Thus, the regulation of the initiation of DNA replication can be viewed as a two-step process involving the assembly and activation of the MCM2-7 complex at replication origins (4). However, the molecular details for the assembly and activation of the MCM2-7 complex are not well understood. Although MCM2-7 hexamers have been purified from various organisms, an active heterohexameric helicase remains to be identified (14 -16).
The recruitment of the MCM2-7 complex to replication origins is dependent on the chromatin binding of the origin recognition complex and Cdc6 (17)(18)(19). Recent studies show that Cdt1 and Mcm10/Dna43 also play important roles in the chromatin binding of the complex (20 -23). An important question is whether the MCM2-7 complex is recruited to a replication origin as a pre-assembled heterohexamer or assembled on the origin de novo by the sequential recruitment of subcomplex intermediates such as the Mcm(3ϩ5), Mcm(4ϩ6ϩ7), and Mcm(2ϩ2) (15, 24 -27). In vitro replication studies using Xenopus egg extracts indicate that the establishment of the prereplication complex or "licensing" is achieved only with preassembled heterohexamers (24). However, similar studies suggest that individual MCM subunits or subcomplexes bind stepwise to chromatin during the licensing reaction to form stable hexamers (27). The characterization of mcm mutants defective in the recruitment of the MCM2-7 complex to replication origins should provide useful information on the process of establishing the pre-replication complex.
The activation of the MCM2-7 complex in the S phase requires the coordinated actions of S phase cyclin-dependent kinases (28 -31) and the Cdc7-Dbf4 kinase (32)(33)(34)(35). The cyclindependent kinases are believed to serve as global S phase promoting factors (36,37), whereas the Cdc7-Dbf4 kinase specifically phosphorylates origin-bound Mcm2 (9 -11, 25, 38, 39). The phosphorylation of Mcm2 is believed to induce a conformational change in the MCM2-7 complex essential for the melting of DNA at the origin (12) and the recruitment of Cdc45 and DNA polymerases (30,31). Interestingly, a MCM5 mutant allele, mcm5-bob1, bypasses the requirement of the Cdc7-Dbf4 kinase for mitotic DNA replication (1,40), suggesting that this mutation may induce a conformational change in the complex similar to that induced by the Cdc7-Dbf4 phosphorylation of Mcm2. Understanding the changes induced by the mcm5-bob1 mutation may provide valuable information on the activation of the MCM2-7 complex.
In this study, we report the characterization of two mcm3 mutant alleles. 1) mcm3-10 is defective in recruiting Mcm3-10 and Mcm7 to a replication origin. 2) mcm3-1 is defective in the activation of the MCM2-7 complex. Mcm3-10 contains an amino acid substitution identical to that found in Mcm5-bob1 in a conserved residue. Our results suggest that altered interactions between Mcm3 and Mcm5 caused by mutations in this conserved residue contribute to the different replication defects in the respective mutants.
Construction of Isogenic mcm3 Strains-A MCM3 integration plasmid was constructed as follows: the MCM3 gene was recovered from YCP50.MCM3 (41) and cloned into a modified version of the integration vector pRS306 (Stratagene, La Jolla, CA) at a SalI site (pRS306.MCM3). The EcoRI-SacII fragment in pRS306.MCM3 was then replaced with that from pRS314.mcm3-1 and pRS314.mcm3-10, respectively, to generate pRS306.mcm3-1 and pRS306.mcm3-10. These plasmids were digested with BamHI and integrated into the MCM3 locus of the yeast strains BJ2168 (MATa pep4 prc1prb1 ura3 leu2 trp1) and W303a (MATa ade2 ura3 leu2 trp1can1). The transformants were grown on 5Ј-fluoroorotic acid plates to select for strains containing either MCM3 or mcm3. Isogenic mcm3-10 and mcm3-1 mutants were selected by their inability to grow on YEPD-rich medium at 37°C and confirmed by sequencing the mcm3 loci in the respective mutant strains. The arrest phenotypes in these strains are similar to their respective original source strains R61-3C and D273-11A.
Chromatin Immunoprecipitation Analysis-W303a (WT) and the mcm3 mutant cells grown in YEPD medium at 25°C were first synchronized at M phase with nocodazole for 2 h and then released into YEPD medium containing ␣-factor (10 g/ml) at 37°C for 1.5 h. Formaldehyde cross-linking was then performed as described previously (23). Cross-linked cells were lysed in lysis buffer (50 mM HEPES 7.5, 140 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, and protease inhibitor mixture (Roche Molecular Biochemicals)). Immunoprecipitation was performed with anti-Orc2 (Santa Cruz Biotechnology, Santa Cruz, CA), anti-Mcm3 (2), anti-Mcm7 (47), or anti-hemagglutinin antibodies (Roche Molecular Biochemicals). Immunoprecipitates were washed twice with lysis buffer, twice with lysis buffer plus 0.5 M NaCl, twice 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 twice with TE buffer (10 mM Tris-HCl, pH 8.0, and 1 mM EDTA). Immunoprecipitates were then resuspended in TE buffer with 1% SDS and incubated at 65°C overnight to reverse protein DNA cross-link. DNA samples were purified using Qiagen PCR purification columns (Qiagen, Valencia, CA), and PCR was carried out using primers 5Ј-atggcgttattggtgttgatgta-3Ј/5Ј-ttgcggtgaaatggtaaaagtc-3Ј for ARS1 and 5Ј-aggggcagcggtttgtgag-3Ј/5Ј-tggagctgcttctacgactactga-3Ј for ATP11, a gene that encodes a chaperone for the assembly of the F 1 -ATPase at the following conditions: twice (94°C for 2 min, 96°C for 1 min, and 54°C for 4 min); 28 times (94°C for 1 min and 54°C for 1.5 min); and once (70°C for 5 min). PCR products were incubated with 1:10000 dilution of Vistra Green (Amersham Pharmacia Biotech, NJ) and separated on 3% agarose gel. Fluorescent signals were quantitated using Molecular Dynamics Storm 840 PhosphorImager and IMAGEQUANT software. The relative efficiency of origin binding of a protein in a particular strain was calculated as the ratio of ARS1/ATP11 amplified from the immunoprecipitate normalized to the ratio of ARS1/ATP11 amplified from the control immunoprecipitate.
Western Blot Analysis-Whole cell lysates were prepared and probed by anti-Mcm3, anti-LexA, and anti-actin antibodies as described previously (23,48).
Dosage Suppressor Screen-A mcm3-10 strain LRD11-5A (42) was transformed with a YEp24-based genomic library (49) and plated on 25 selective plates. One plate was incubated at 30°C to estimate the total number of transformants generated. The other 24 were incubated to select for growth at 37°C. The plasmids from these suppressors were isolated, and the identities of the clones were determined by sequencing analysis.

Mcm3-10 Blocks Entry of S Phase-MCM3
is a founding member of the MCM2-7 gene family. It was initially cloned through complementation of the ts growth defect of mcm3-1 (41,50), a mutant allele that shows a diminished efficiency of replication initiation (2). However, at the restrictive temperature, the mcm3-1 cells arrest with a largely replicated genome (41). Another ts mcm3 allele, mcm3-10 also known as dbf10, is unable to initiate DNA synthesis (43,51). To gain insights into the roles of Mcm3 at different steps of the replication initiation process, we characterized replication defects of these two alleles using isogenic strains ("Experimental Procedures"). At the restrictive temperature (37°C), both mcm3-10 and mcm3-1 cells arrest as predominantly large budded cells (85%) with undivided nuclei (Fig. 1a). In mcm3-10, the nuclei are localized near the bud neck in the mother cells (Fig. 1a). The migration of nuclei to the bud neck is normally observed before nuclear division but after the completion of DNA synthesis. In mcm3-1, the nuclei are randomly localized within the mother cells (Fig.  1a), consistent with the incomplete replication of the genome in these cells (41,44). However, FACS shows that mcm3-10 cells arrest with DNA contents close to one copy of the haploid genome at the restrictive temperature (Fig. 1b), suggesting a failure to commence DNA replication (51) even though the nuclei have assumed a bud neck position. As observed previously (41,44), mcm3-1 mutant cells replicate most if not all of the genome at the restrictive temperature (Fig. 1b), a phenotype shared by several other DNA replication initiation mutants (9,41,44,52,53). The late S phase arrest is believed to be the effect of a combination of leaky initiation defects and incomplete replication of the genome in these mutants (41,44). A slight increase in the DNA contents of both G 1 and G 2 /M phase cells was observed in all three strains after the temperature shift, probably the result of continuous amplification of mitochondrial DNA. The FACS results suggest that mcm3-10 and mcm3-1 are defective at two different steps in the replication initiation process, mcm3-10 is unable to initiate DNA synthesis and mcm3-1 is unable to complete DNA synthesis.
To determine whether recruitment or activation of the MCM2-7 complex is compromised in the mutants, we characterized their respective cell cycle block relative to START, a point in late G 1 phase when the MCM2-7 complex has already been recruited to the replication origin but not yet activated (54). Cells grown at 25°C were first synchronized at START by ␣-factor and then released into fresh medium at 37°C. Cell cycle progression during the course of the treatment was monitored by FACS analysis. The results shown in Fig. 2 indicate that mcm3-10 cells complete one round of DNA replication and then arrest in the next cell cycle with 1C DNA content after release from ␣-factor. However, under the same conditions, mcm3-1 cells arrest at the late S phase of the same cell cycle. These results suggest that inactivating Mcm3-10 after the MCM2-7 complex has been recruited to the replication origin does not prevent the initiation or completion of DNA synthesis until the next cell cycle. In contrast, the inactivation of Mcm3-1 after the recruitment of the MCM2-7 complex compromises the initiation (2) as well as the completion of DNA replication.

Recruitment of Mcm3 and Mcm7 Is Compromised in mcm3-10 Mutants-If
Mcm3-10 fails to execute an event before the recruitment of the MCM2-7 complex at replication origin, the binding of the complex to origin will probably be affected. To test this hypothesis, the association of Mcm3 and Mcm7 with ARS1 in mcm3-10 cells was analyzed by chromatin immunoprecipitation analysis (55,56). Wild type and both mcm3 mutants were first synchronized in M phase with nocodazole and then released into fresh medium containing ␣-factor at 37°C for formaldehyde cross-linking. Cell lysates were prepared for immunoprecipitation using anti-Mcm3 or Mcm7 antibodies. Immunoprecipitation was also performed using polyclonal anti-Orc2 and hemagglutinin antibodies as positive or negative controls. The results indicate that the binding of Orc2 to ARS1 is not affected by either mutant allele (Fig. 3a, lanes  1-3, and b), consistent with the observation that the MCM2-7 complex is recruited after origin recognition complex is bound to origin DNA (17,18). Mcm3-1 does not appear to significantly affect its own binding or the binding of Mcm7 to ARS1 compared with Mcm3 in wild type cells (Fig. 3a, lanes 4, 5 and 7, 8  and b), supporting the hypothesis that Mcm3-1 is fully functional during the recruitment of the MCM2-7 complex. In contrast, the association of both Mcm3-10 and Mcm7 with ARS1 is considerably reduced in mcm3-10 mutants (Fig. 3a, lanes 4, 6  and 7, 9, and b), consistent with a failure to execute an event that is required for the recruitment of the MCM2-7 complex. The wild type and mutant Mcm3 proteins were expressed at similar levels (Fig. 3c).
mcm3-10 and mcm5-bob1: Similar Mutations with Opposite Effects-To determine the structural change in Mcm3-10 that impedes the recruitment of Mcm3 and Mcm7, we identified the amino acid substitutions in the mcm3 alleles. The ts growth defect of both mcm3 alleles can be rescued by a MCM3-containing plasmid (Fig. 4a). To map the mutations in the mcm3 alleles, a series of mcm3 deletions (gapped plasmids) were analyzed for their ability to rescue the growth defect of the mutants at 37°C (Fig. 4b) (57). All but mcm3⌬SB failed to rescue the ts phenotype of mcm3-1 or mcm3-10. Thus, both mutations are localized within the 1 kilobase pair of EcoRI-SacII fragment in the open reading frame of MCM3 (Fig. 4b).
The repaired plasmids containing mcm3-1 or mcm3-10 were recovered from mcm3⌬RS transformants for sequencing analysis. mcm3-1 is a G to A mutation resulting in a glycine to glutamate substitution at residue 246 (G246E), and mcm3-10 is a C to T mutation resulting in a proline to leucine substitution at residue 118 (P118L) (Fig. 4c). Both mutations are Nterminal to the ATPase domain conserved in the Mcm2-7 proteins (5). When these mutations were reintroduced into the yeast genome at the MCM3 locus, the resulting strains exhibited phenotypes of the original mutant alleles ("Experimental Procedures"). A comparison of the amino acid sequences of MCM proteins from different species indicates that Gly 246 (mcm3-1) is conserved in the Mcm3 from yeast to human but not between members of the Mcm2-7 family (Fig. 4, c and d).
To investigate whether Mcm3-10 also bypasses the function of the Cdc7-Dbf4 kinase, a diploid strain (MCM3cdc7⌬/ mcm3-10 CDC7) was constructed and subjected to tetrad analysis. The growth phenotypes of cdc7⌬ and mcm3-10 were scored. From 50 asci, the ratio of parental ditype:nonparental ditype:tetratype was ϳ1:1:4 (Table I) teins are assembled into the hexameric complex, Mcm5 appears to be in direct contact with Mcm3 based on their strong interactions in two-hybrid and co-immunoprecipitation analyses (48,59,60). Furthermore, a stable Mcm3-Mcm5 heterodimer has been purified and shown to be an intermediate in the assembly of the heterohexamer (24,27,59). To investigate whether the phenotypes of mcm3-10, mcm3-1, and mcm5-bob1 may stem from an altered interaction between Mcm3 and Mcm5, the interactions between mutant and wild type proteins of Mcm3 and Mcm5 were examined by two-hybrid analysis (61). The results in Fig. 5a 10 and 11). The mcm5-bob1 mutation does not affect the expression level of the LexA-Mcm5-bob1 fusion protein in the two-hybrid strain (Fig. 5a, inset).
MCM5 Is an Allele-specific Dosage Suppressor of mcm3-10 -If Mcm3-10 has a specific defect in the assembly and recruitment of the MCM2-7 complex, the mcm3-10 allele may provide a valuable tool for identifying interacting factors that modulate the assembly and the recruitment of the MCM2-7 complex. One approach we took was to isolate dosage suppressors of mcm3-10 by transforming the mutant with a high copy yeast genomic library. From ϳ45,000 transformants, which represent the yeast genome ϳ30 times, 22 suppressors of the ts growth defect were isolated. Two of the suppressor plasmids contain MCM5 (data not shown). The dosage suppression of mcm3-10 by MCM5 is allele-specific, as the high copy yeast genomic library MCM5 plasmid does not complement the ts FIG. 2. mcm3-10 and mcm3-1 mutants are defective at different steps in DNA replication. WT, mcm3-10, and mcm3-1 cells grown at 25°C treated with ␣-factor 1 or 2 h after release into fresh medium at 37°C as indicated were analyzed.  4. Identification of the two mcm3 mutations. a, MCM3 complements the ts growth defect of mcm3-1 and mcm3-10. WT, mcm3-10 growth defect of the mcm3-1 mutant (Fig. 5b). Dosage suppression often results from the stabilization of a complex by an increased concentration of one of the interacting partners. The suppression of mcm3-10 by an increased dosage of Mcm5 further supports the notion that a weakened interaction between Mcm3-10 and Mcm5 is the basis for the failed recruitment of MCM proteins to the origin in the mcm3-10 mutant. For the same reason, the increased expression of Mcm5 would be expected to have no effect on the initiation of DNA replication once the MCM2-7 complex is already in place at origins as in the case of the mcm3-1 mutant.

DISCUSSION
The MCM2-7 complex is the signature component of the pre-RC that primes replication origins for the initiation of DNA synthesis. The recruitment of the complex to replication origins requires a proper assembly of the MCM2-7 complex (24,27) and the prior action of origin recognition complex, Cdc6, and Cdt1 (17, 18, 20 -22). The mechanisms involved in the regulation of this complex process are not well understood. We have shown that the mcm3-10 mutant is defective in the recruitment of the MCM2-7 complex at the restrictive temperature. We   (1,58). Although very different replication defects are observed in these mutants, the altered interaction between  (Fig. 4d), we have not been able to define an interaction domain by conventional deletion analysis (data not shown). The preservation of intact sequences appears to be critical for the structural integrity of these proteins (48). That identical amino acid substitutions (Pro to Leu) in Mcm3-10 and Mcm5-bob1 produce opposite rather than similar effects suggest asymmetric contacts between Mcm3 and Mcm5 at the interface of the heterodimer. In a recently proposed structural model, Schwacha and Bell (14) suggest that the MCM2-7 hexamer is composed of one trimer Similar to several other replication mutants (9,44,52), the mcm3-1 mutant arrests with an undivided nucleus and approximately two copies of the haploid genome DNA content (Fig. 1a) (41). This late S phase arrest suggests an incomplete replication of the genome (41,44). Consistent with this notion, we showed that mcm3-1 mutant cells released from G 1 arrest do not proceed beyond S phase at the restrictive temperature (Fig.  2). The random distribution of nuclei in the mcm3-1 mutant cells further supports this interpretation (Fig. 1a). The bud neck localization of nuclei in the mcm3-10 mutant cells seems inconsistent with a cell cycle arrest before START. However, a report of reductive cell division as a result of a complete block in DNA replication (63,64) may provide an explanation for the nuclei migration observed in mcm3-10 mutant cells. If so, checkpoint controls that obviate reductive division in mcm3-10 would have to be activated. A possible role for Mcm3 in other cell cycle events such as checkpoint controls requires further investigation.