Schizosaccharomyces pombe Minichromosome Maintenance-binding Protein (MCM-BP) Antagonizes MCM Helicase

The minichromosome maintenance (MCM) complex, a replicative helicase, is a heterohexamer essential for DNA duplica-tion and genome stability. We identified Schizosaccharomyces pombe mcb1 (cid:1) (Mcm-binding protein 1), an apparent orthologue of the human MCM-binding protein that associates with a subset of MCM complex proteins. mcb1 (cid:1) is an essential gene. Deletion of mcb1 (cid:1) caused cell cycle arrest after several genera-tions with a cdc phenotype and disrupted nuclear structure. Mcb1 is an abundant protein, constitutively present across the cell cycle. It is widely distributed in cytoplasm and nucleoplasm and bound to chromatin. Co-immunoprecipitation suggested that Mcb1 interacts robustly with Mcm3–7 but not Mcm2. Overproduction of Mcb1 disrupted the association of Mcm2 with other MCM proteins, resulting in inhibition of DNA replication, DNA damage, and activation of the checkpoint kinase Chk1. Thus, Mcb1 appears to antagonize the function of MCM helicase.

The minichromosome maintenance (MCM) complex, a replicative helicase, is a heterohexamer essential for DNA duplication and genome stability. We identified Schizosaccharomyces pombe mcb1 ؉ (Mcm-binding protein 1), an apparent orthologue of the human MCM-binding protein that associates with a subset of MCM complex proteins. mcb1 ؉ is an essential gene. Deletion of mcb1 ؉ caused cell cycle arrest after several generations with a cdc phenotype and disrupted nuclear structure. DNA replication requires a series of tightly coordinated events to ensure that each daughter cell receives one complete copy of genetic information (1,2). In response to damage generated by mutations in the replication machinery or by exogenous damaging agents, eukaryotic cells activate checkpoint responses that arrest S phase progression and activate DNA repair (3,4). Defects either in replication or checkpoint responses generate genome instability and increase cancer susceptibility (5,6).
The minichromosome maintenance (MCM) 3 complex is a replicative helicase conserved in eukaryotes and archaea (for reviews, see Refs. 7 and 8). The complex consists of six distinct yet structurally related subunits, Mcm2-7, assembled into a heterohexameric ring. MCM proteins are members of the AAA ATPase family sharing several distinctive protein sequences that define the family. In metazoa and plants, there are several additional MCM family members, Mcm8 and Mcm9 (9 -12), as well as developmentally specific versions of the MCMs (13), all of which contain the characteristic MCM-specific protein sequence motifs. The fission yeast genome encodes just the six core MCMs, which assemble into a complex that is constitutively located in the nucleus throughout the cell cycle (14 -17). In late M and early G 1 phases, the MCM complex is recruited onto chromatin as part of the prereplication complex and onto unreplicated DNA; this chromatin localization is dislodged as replication proceeds (18). MCM proteins are abundant and exceed the number of replication origins (19 -21). Each of the six MCM proteins is essential for viability with a similar deletion phenotype (16,(22)(23)(24)(25)(26)(27). Reduction of MCM protein levels causes genome instability in fission yeast due to replication fork collapse and DNA damage (25,28).
A novel component of the human MCM complex was discovered in human cells using tandem affinity purification (29). Human MCM-binding protein (hMCM-BP) shares no homology to MCM proteins or AAA ATPases. Biochemical analysis suggests that MCM-BP replaces Mcm2 and forms an "alternative" MCM complex with Mcm3-7. Similar to the MCM proteins, hMCM-BP localizes primarily in the nucleus and associates with chromatin in most of the cell cycle except early M phase. Recently, an hMCM-BP orthologue, ETG1, was isolated from plants (30,31). Depletion of ETG1 activates a G 2 cell cycle checkpoint, resulting in a late G 2 cell arrest, and also plays a role in establishing cohesion. Interestingly, hMCM-BP orthologues have been found in fruit flies, frogs, zebrafish, and two fission yeasts but not in budding yeast.
Here, we report the identification and characterization of the hMCM-BP orthologue in Schizosaccharomyces pombe, mcb1 ϩ (Mcm-binding protein 1). We show that mcb1 ϩ encodes an essential gene. Spores lacking mcb1 ϩ arrest after several divisions with a G 2 DNA content and a cdc phenotype similar to MCM deletion mutants. We epitope-tagged Mcb1 and showed that Mcb1 is an abundant protein constitutively expressed through the cell cycle. Mcb1 is distributed in all cellular compartments, including a substantial chromatin-bound fraction.

EXPERIMENTAL PROCEDURES
Fission Yeast Strains, Plasmids, and Manipulation-All S. pombe strains (supplemental Table 1) were constructed and maintained in yeast extract plus supplement medium or under selection in Edinburgh minimal medium (EMM) with appropriate supplements using standard techniques (32)(33)(34). Transformation was performed by electroporation. Unless noted, asynchronous cultures were grown at 32°C. In cell cycle block and release experiment, cells were grown at 25°C (permissive temperature) to early exponential phase and shifted to 36°C for 4 h (restrictive temperature). HA-tagged Mcb1 from the endogenous locus was generated by using the pFA6a series of plasmids with primers 5Ј-CGAAGAGTTTCGGTCGTCAAC-TGGTTTCAAGAATTGATTTTGAGGCTGCCCGTAGTC-TAATCAATCATTGGACTGTCAACCGGATCCCCGGGT-TAATTAA-3Ј and 5Ј-CTTGGAAATTCCAAAAAGACATG-AAAAGTAATTTCTAACATTGGTTAAATGATGTTGAT-TATAAGAAAATATGCGATCAAGAATTCGAGCTCGTT-TAAAC-3Ј (35). Doubly tagged strains were isolated by mating and from tetrad analysis. The mcb1 ϩ gene was cloned using cDNA as template and was inserted into pREP-based expression plasmids to generate pLD10 (nmt1-mcb1HA) and pLD18 (nmt1-mcb1V5), which were used for ectopic expression (36). The mcb1g gene was amplified using genomic DNA. To generate stable Mcb1HA-overproducing cells (OP-Mcb1), we made pLD14 by inserting the nmt1-mcb1HA fragment from pLD10 into pJK210. NruI-linearized pLD14 was integrated at leu1-32 locus as described (37). The strains for the mutation analysis were generated with the same approach. The nmt1 promotercontaining strains were maintained on yeast extract plus supplement agar (for integrants) or EMM with supplements and thiamine. To perform overproduction/induction experiments, liquid cultures grown in the presence of 2.5 g/ml thiamine to early exponential phase were washed twice with an equal volume of EMM before inoculating in the absence of thiamine (overproduction state) or in the presence of 5 g/ml thiamine (strong repression state) (38,39).
Construction of mcb1 ϩ Deletion-To delete the mcb1 ϩ gene (SPAC1687.04), we removed the entire coding sequence according to a gene deletion protocol using upstream prim ers 5Ј-GAGATCTAGACAGGACGATTGGACGATACT-3Ј and 5Ј-GAGACTCGAGATTATAAATATATAATTTTAT-CCTTTAAACC-3Ј and downstream primers 5Ј-GAGAG-CGGCCGCTTGATCGCATATTTTCTTATAATC-3Ј and 5Ј-GAGATCTAGAGTCGCTTTAGTACATTCTAAAC-3Ј. The resulting plasmid (pLD21) was amplified, linearized at XbaI, and transformed into a fresh mated wild-type diploid strain. Stable Uraϩ integrated diploids were selected and confirmed by PCR. The deletion was also confirmed by tetrad analysis and complementation. Bulk spore germination was performed as described previously (23).
Complementation-A heterozygous diploid strain (mcb1 ϩ / ⌬mcb1::ura4 ϩ ) was transformed with linearized leu1 ϩ integration plasmids that express HA-tagged mcb1 deletion mutants cloned from cDNA and plated on thiamine-containing selective medium. Random spore analysis (34) was used to recover haploids that were Uraϩ and Leuϩ. The resulting haploids were confirmed by PCR and Western blot.
Flow Cytometry-Flow cytometry was performed as described (40,41) with minor modifications. Briefly, cells were fixed in 70% ice-cold ethanol, rehydrated with 50 mM sodium citrate, and treated with 0.1 mg/ml RNase A. Cells were stained with 1 M SYTOX Green (Invitrogen) in 50 mM sodium citrate. Macintosh BD CellQuest TM Pro 5.2.1 software (BD Biosciences) was used to analyze and organize the data acquired by the FACScan cytometer (BD Biosciences).
Protein Extracts, Immunoprecipitation, and Immunoblotting-Total protein extracts were prepared either by glass bead lysis using trichloroacetic acid (TCA) extraction as described (44) or alkaline (NaOH) lysis protein extraction (45). The concentrations of TCA-extracted protein samples were quantified by BCA protein assay (Pierce). Twenty micrograms of total protein were separated by 8% SDS-PAGE. For alkaline lysis protein extraction, equal numbers of cells were collected, resuspended in 0.3 M NaOH, and incubated at room temperature for 10 min. Permeabilized cells were centrifuged at 1,700 ϫ g for 3 min. The cells pellets were resuspended in 30 l of 2ϫ sample buffer and boiled for 10 min. Ten microliters were loaded on an SDSpolyacrylamide gel for separation.
Immunoprecipitations were performed with 750 g of precleared soluble protein overnight at 4°C. Fifty microliters of immobilized rProtein A (IPA300, RepliGen; 1:1 in lysis buffer) were added and incubated for 2 h at 4°C. Beads were spun down and washed four times with 1 ml of cold lysis buffer. After the final wash, beads were resuspended in 2ϫ sample buffer and boiled for 5 min. An equal volume was loaded on an 8% SDSpolyacrylamide gel for separation.
For immunoblotting, samples separated by SDS-PAGE were transferred to Immobilon-P membrane (Millipore). ECL Western blotting substrate (Pierce) and Blue Ultra autoradiography film (BioExpress) were used to detect signals.
Viability Assays-Cells (FY11, FY838, FY4594, and FY4596) were grown overnight at 32°C to early exponential phase (A 595 ϳ 0.2-0.3) in EMM lacking leucine (EMM-L) with 2.5 g/ml thiamine. Cells were washed twice with EMM-L and inoculated into EMM-L. At each time point, cultures were serially diluted in yeast extract plus supplement (1:100, 1:1,000, and 1:10,000). Equal volumes of each dilution were plated on yeast extract plus supplement plates for viability testing and incubated at 30°C for 4 days. The number of colonies was averaged from different dilutions. Relative viability at time T was calculated as follows: (averaged number of colonies at time T)/(averaged number of colonies at time 0). At each time point, cells were also fixed in 70% ethanol. Cellular DNA content was analyzed by flow cytometry. Rehydrated cells were counted twice for cell growth analysis. Relative growth at time T was defined as follows: (averaged number of cells at time T)/(averaged number of cells at time 0).
In Situ Chromatin Binding Assay and Fluorescence Microscopy-An in situ chromatin binding assay was performed as described (18) with modifications as described (46). Proteins were detected using rabbit anti-V5 (1:300 volume) or rabbit anti-GFP (1:200 volume) and chicken anti-rabbit Alexa Fluor 488 (Invitrogen). Cells were mixed with poly-L-lysine (Sigma), heat-fixed on microscope slides, and mounted in 50% glycerol, phosphate-buffered saline (PBS) for visualization. DAPI staining for rehydrated cells was performed by mounting the heat-fixed cells with 1 g/ml DAPI containing antifade mounting solution (50% glycerol in water with 0.1% p-phenylenediamine dihydrochloride).
To examine Rad22-YFP foci in live cells, cells were washed twice in EMM containing 10 g/ml DAPI, air-dried on Color-Frost Plus Microscope slides (Fisher Scientific), and mounted in 50% glycerol, PBS. All pictures were taken on a Leica DMR florescence microscope using a 63ϫ oil immersion objective (Leica Plan Apochromat; numerical aperture, 1.32) and recorded with OpenLab software (Improvision).
Digital Image Manipulation-All the plates and films were electronically scanned using a ScanJet IIcx scanner (Hewlett-Packard). Digitized pictures/photos were analyzed and contrast-enhanced by NIH ImageJ software and assembled in Canvas software (ACD Systems).

RESULTS
Identification and Deletion of S. pombe mcb1 ϩ -hMCM-BP was first identified as a novel component of MCM (29). Although there is no obvious Saccharomyces cerevisiae orthologue (see yeast orthology table version 2.15 (Sanger)), we found a putative orthologue of hMCM-BP gene in S. pombe, SPAC1687.04, and named it mcb1 ϩ (supplemental Fig. S1). The mcb1 ϩ gene is highly conserved in eukaryotes. It has two exons and one intron and encodes a protein with 501 amino acids and a predicted molecular mass of 56.6 kDa. Similar to the hMCM-BP protein, Mcb1 has no obvious sequence motifs and shares no obvious homology with S. pombe MCMs or AAA ATPases (supplemental Fig. S2).
We created a heterozygous diploid mcb1 ϩ /⌬mcb1::ura4 ϩ for tetrad analysis. All tetrads displayed a 2:2 segregation of viable:inviable spores, and all viable colonies were UraϪ (supplemental Fig. S3A). The inviable colonies were ⌬mcb1::ura4 ϩ spores, which managed to complete a few cell divisions and form microcolonies of 8 -16 cells (supplemental Fig. S3B). Ectopic expression of mcb1 ϩ under its own promoter from a plasmid was able to rescue the lethality in the spores, confirming that the phenotype arises from disruption of this gene (data not shown).
We analyzed a population of ⌬mcb1 spores by a bulk spore germination assay using ⌬mcb1::ura4 ϩ /mcb1 ϩ diploid cells and wild-type ura4 ϩ /ura4-D18 diploid cells as a control. The spores were inoculated into liquid medium lacking uracil, which ensures that only the ⌬mcb1::ura4 ϩ or ura4 ϩ spores can germinate. In contrast to ⌬mcm spores (23, 24), ⌬mcb1::ura4 ϩ spores showed no obvious delays in their first S phase compared with the wild type (Fig. 1A). However, similar to ⌬mcm spores, the terminal phenotype of germinated ⌬mcb1::ura4 ϩ spores was an elongated cdc morphology and in most cases a single nucleus. About 10% of cells had an abnormally shaped nucleus, either teardrop-shaped or cut (Fig. 1B, 16 h). Thus, mcb1 ϩ is an essential gene.
Characterization of Mcb1 Protein-In fission yeast, MCM proteins are localized in the nucleus throughout the cell cycle by nuclear localization sequences on Mcm2 and Mcm3 (15).
Human MCM-BP is also a nuclear protein (29). We constructed a C-terminally HA-tagged Mcb1 (mcb1HA) to replace the wild-type copy in the genome. Mcb1HA cells showed normal growth, indicating that the tagged copy is functional. When compared with a strain expressing Mcm2HA, we observed that Mcb1HA is expressed at a higher level, indicating that it is a very abundant protein (supplemental Fig. S4).
Because human MCM-BP is chromatin-associated (29), we examined Mcb1HA localization by immunofluorescence using an mcb1HA mcm2V5 strain for an in situ chromatin binding assay (supplemental Fig. S5). Consistent with previous studies (15,47), Mcm2V5 stayed in the nucleus (supplemental Fig.  S5A) and remained chromatin-bound in binucleated (S phase) cells (supplemental Fig. S5B). However, we were unable to detect Mcb1HA cytologically (supplemental Fig. S5), suggesting that the epitope tag is inaccessible or occluded. Similar results were observed for N-terminally HA-tagged Mcb1 (HAmcb1; data not shown). We also tagged Mcb1 with GFP and mCherry at the C terminus. Although these tags were easily detected on Western blots, we were not able to detect any fluorescence in live cells (data not shown).
Therefore, we took a biochemical approach to examine localization and performed a cell fractionation assay. Asynchronous cells were treated to release different cellular compartments in subsequent fractions (supplemental Fig. S6). Equal volumes of each fraction were separated by SDS-PAGE and blotted for Mcm2V5, Mcm7, and Mcb1HA ( Fig. 2A). Fib1 (Nop1) is involved in pre-rRNA processing and is a marker for the chromatin fraction, whereas ␣-tubulin is a predominantly but not exclusively cytoplasmic (48,49). Consistent with previous studies, Mcm2V5 and Mcm7 are nuclear proteins. Mcm2V5 was mainly in the nucleoplasm, whereas the majority of Mcm7 was chromatin-bound. In contrast, Mcb1HA was present throughout the cells but strongly enriched in the nuclear fractions.
The levels of mcm2 ϩ mRNA (23, 50) and the protein level (51) are constant throughout the cell cycle. Studies suggest that all MCM subunit levels are comparable (17,47). To investigate whether levels of Mcb1 protein level fluctuate in the cell cycle, we used a cdc25-22 mcb1HA mcm2V5 strain to synchronize and release the cells by controlling the temperature. We observed no cell cycle-dependent change in total protein level or mobility (supplemental Fig. S7). Mcb1HA migrated as a doublet in this experiment, but that was not apparent in soluble lysate (Fig. 2B, lane 1).   (Fig. 2B, lane 3). Conversely, Mcm2V5 associated with Mcm4, -6, -7, and -3 but not Mcb1 (Fig. 2B, lane 4). This suggests that there are at least two MCM complexes, one with Mcm2 and one with Mcb1. We found no significant association between Mcb1HA and Mcm2V5 in this experiment (Fig. 2B, lanes 3 and 4)

Isolation of Mcb1
Mutants-To identify the functional regions of Mcb1, we constructed a series of deletions within the protein, arbitrarily defining five domains: Exon 1, amino acids 2-54; A region, amino acids 55-230; B region, amino acids 231-414; and C region, amino acids 415-501 (Fig. 3A). Constructs lacking these domains were cloned into episomal plasmids under the high strength nmt1 promoter (38). The resulting plasmids were transformed into wild-type cells. We controlled expression by the levels of thiamine. In the presence of thiamine (low amount of protein expression from the nmt1 ϩ promoter (55)), all the transformed cells were viable (Fig. 3B). However, in the absence of the thiamine (overexpression), we found that cells transformed with full-length Mcb1 (mcb1 and mcb1g) were unable to form colonies (Fig. 3B). Interestingly, cells that overexpressed mcb1-D2 and mcb1-D22 were viable but generated notably smaller colonies than the vector control. Both of these mutants lack the N-terminal exon, and D22 has an additional point mutation introduced during PCR (E423G). This suggests that D2 and D22 have residual Mcb1 activity.
We tested each mutant for complementation of mcb1⌬ by integrating the constructs into the leu1-32 locus under the nmt1 promoter in the diploid ⌬mcb1::ura4 ϩ /mcb1 ϩ . Following sporulation, we screened haploids for Uraϩ Leuϩ clones containing both the deletion and the insertion alleles in the presence of thiamine.
Full-length mcb1 ϩ and mcb1-D2 and mcb1-D22 rescued mcb1⌬ with comparable levels of expression (Figs. 3A and 4A). None of the other mutants were recovered, indicating that they are non-functional. We observed that both mcb1-D2 and mcb1-D22 cells showed elongated cells with a 2C DNA content (Fig.  4B), suggesting that they are hypomorphs of mcb1 ϩ . Their elongated cell morphology suggests that they have activated a checkpoint that delays the cell cycle. We crossed these mutants into strains lacking the S phase checkpoint (cds1⌬), the damage checkpoint (chk1⌬), or the upstream regulator of both checkpoints (rad3⌬) and determined that deletion of either chk1 or rad3 relieved the cell elongation phenotype. Thus, we conclude that the damage checkpoint is responsible for their cell cycle delay.
Overexpression of Mcb1 in Cells Inhibits S Phase and Activates DNA Damage Checkpoint-We next investigated the lethality associated with overexpression of full-length mcb1 ϩ . We expressed nmt1-mcb1HA integrated at the leu1-32 locus in an mcb1 ϩ or mcb1⌬ background. In the presence of thiamine, both strains were healthy and produced colonies with normal sizes (supplemental Fig. S8A, top left). In the absence of thiamine, the overproducing strain in mcb1⌬ was completely inviable (supplemental Fig. S8A, bottom, d). Surprisingly, the overproducing strain that also contains one copy of wild-type mcb1 ϩ had a few surviving colonies (supplemental Fig. S8A, bottom, c).
We followed the cells during promoter induction and found that the overall cell number was similar in both strains; however, the viability (plating efficiency) of the mcb1 overproducers began to drop by 9 h (supplemental Fig. S8B). Although the number of ⌬mcb1 nmt1-mcb1HA survivors continued to drop, the viability of mcb1 ϩ nmt1-mcb1HA cells plateaued at 13 h and started to increase again around 17 h. The survivor class that emerged by 26 h in mcb1 ϩ nmt1-mcb1HA showed normal 2C DNA content (Fig. 5A, panel c), whereas ⌬mcb1 nmt1-mcb1HA was lethal at the same time point (Fig. 5A, panel d).
The levels of ectopically produced Mcb1HA protein dropped significantly at 26 h (supplemental Fig. S9). We conclude that under the selective pressure of toxic overproduction, the survivors escaped by down-regulating the nmt1 promoter. Because the endogenous mcb1 ϩ gene was still intact, the cells could survive repression of the dominant negative (nmt1-mcb1HA) transgene. For the strain in which the transgene is the only source of mcb1 ϩ , there were no survivors.
The presence of elongated cells in the population suggested that some sort of checkpoint was activated in Mcb1HA overproducers (OP-Mcb1). DNA damage during replication can be visualized by the formation of repair foci containing the homologous recombination protein Rad22 (ScRad52) (e.g. Ref. 28). We observed an increase in the formation of Rad22 foci in cells overproducing Mcb1HA (Fig. 5C). Cells with Rad22-YFP foci increased to 37% by 14 h, and 13% of cells had more than one focus compared with about 10% for the cells grown in repressing (thiamine) conditions. We reasoned that these repair foci might accompany activation of the DNA damage checkpoint, so we combined the overproducer strain with cds1⌬, chk1⌬, or rad3⌬. After growing in thiamine-free medium for 16 h, only the ⌬cds1 nmt1-mcb1HA cells showed elongated cells, whereas the other strains had small cells and abnormal nuclear structure (Fig. 6A, bottom  row). This suggests that the DNA damage checkpoint is activated in some of the cells. Finally, we examined the phosphorylation status of Chk1 as a measure of Chk1 activation (57). Because detection of Chk1 relies on an HA epitope tag, we  (Fig. 6B) (Fig. 7A, compare lanes 3 and 4) (Fig. 7B).
When we precipitated Mcm2V5 in the overproducers, we were unable to detect other MCM proteins (Fig. 7A, compare  lanes 5 and 6). This suggests that the normal MCM complex is disrupted. Surprisingly, under these conditions, we detected an interaction between Mcb1HA and Mcm2V5 (Fig. 7A, lane 6 Next, we examined the effect of our deletion mutants on MCM complex formation using the same strategy. In the presence of thiamine, protein levels of most Mcb1 truncation mutants were comparable with that of full-length Mcb1 (supplemental Fig. S10). Surprisingly, we observed that Mcb1-D5, Mcb1-D6, and Mcb1-D9 associated with Mcm4 when expressed at a near normal level, although they were unable to complement the null and were not toxic when overproduced (Fig. 3A, 7C). However, when overproduced, these mutants did not dissociate the interaction between Mcm4 and the other MCMs (Fig. 7D), leading us to conclude that they do not replace Mcm2. In contrast, the two N-terminal truncation mutants (Mcb1-D2 and Mcb1-D22) did reduce association of Mcm4 with other MCMs, including Mcm2 (Fig. 7D, lanes 9 and 14). Thus, there is a clear correlation between three phenotypes: Mcb1, Mcb1-D2, and Mcb1-D22 are the only constructs that were able to complement mcb1⌬, were toxic when overproduced, and disrupted association between Mcm2 and Mcm4 upon overproduction.
In fission yeast as in most eukaryotes, MCM proteins are found predominantly in the nucleus throughout the cell cycle, but their chromatin association is cell cycle-regulated (for a review, see Ref. 7). Chromatin binding depends on activation of the prereplication complex, including the initiator protein Cdc18 (18). Previously, we showed that mutations that disrupt the MCM complex such as mcm4ts cause all the subunits to exit the nucleus (15). Thus, complex assembly is required for nuclear retention. We hypothesized that the overexpression of Mcb1, which disrupts the normal MCM complex, therefore would also disrupt chromatin binding and nuclear localization of the MCMs.
We used indirect immunofluorescence in an in situ chromatin binding assay (18,47). Proteins located on the chromatin are resistant to Triton, whereas proteins located in the nucleoplasm but not on the chromatin are removed by Triton treatment. This method allows the S phase subset of chromatinbound MCMs (on the binucleate cells) to be distinguished from the abundant unbound MCM protein in the nucleus (18,47). As a control, we used a cdc18-shutoff strain that blocks MCM binding to the chromatin. Importantly, both of these phenotypes depend on thiamine although in opposite directions. For Mcb1 overproducers (OP-Mcb1), lethality is caused by removing thiamine from the medium to induce nmt1-mcb1HA (minus thiamine condition). In the case of the cdc18-shutoff cells, lethality is caused by adding thiamine to the medium to repress nmt1-cdc18 ϩ expression (56).
As shown in Fig. 8, under permissive conditions, both Mcm4GFP and Mcm2V5 were found in the nucleus (ϪTriton) and were chromatin-bound in S phase (binucleates ϩTriton). When Mcb1HA was overexpressed in the absence of thiamine, nuclear localization of both MCM proteins was reduced, and chromatin binding (ϩTriton) was abolished. Similar results were observed if cdc18 ϩ was shut off by addition of thiamine. These data suggest that the toxic effect associated with Mcb1 expression results in delocalization of the MCM subunits to the cytoplasm and inhibition of replication.

DISCUSSION
MCM proteins are members of the AAAϩ ATPase family and share a unique motif, the MCM box, which is important for MCM complex formation and ATP hydrolysis (for reviews, see Refs. 7 and 8). Recent studies identified a novel component of the MCM complex in human as well as plants, MCM-BP (29 -31). Evidence from these systems suggests that MCM-BP is a component of the replisome and replaces the Mcm2 subunit. Recent work suggests that MCM-BP may contribute to sister chromatid cohesion and DNA repair (31). Although no orthologue has been found in budding yeast, we identified a putative orthologue of human MCM-BP that we named Mcb1. Similar to other MCM-BPs, Mcb1 shares no homology to S. pombe MCM proteins and lacks the MCM box.
Disruption of mcb1 ϩ was lethal. However, disrupted spores managed to germinate and complete several cell cycles before arresting with an elongated, cdc morphology. Most of the cells had a single nucleus of normal appearance; a few had a disordered nucleus or evidence of mitosis. This delayed lethality is likely to reflect the abundance of the maternal protein packaged in the spores. For example, mcm4⌬ cells complete S phase prior to arresting with a 2C DNA content; the mcm4⌬ cells arrest prior to S phase only if the residual maternal Mcm4 protein is inactivated with a temperature-sensitive mutation (25). Therefore, although we can conclude that Mcb1 is essential for viability and cell cycle progression, we cannot conclude at what stage(s) of the cell cycle it works.
We found Mcb1 to be amenable to epitope tagging and detection by Western blot.  in some way in its normal environment and only available upon denaturation. We used cell fractionation to examine Mcb1 localization and found that Mcb1 is ubiquitously distributed through the entire cell but enriched in the nucleus.
Similar to observations in humans (29), we found that Mcb1 normally associates with Mcm3, -4, -6, and -7, but not Mcm2, thus forming an alternative MCM complex. The canonical MCM complex consists of Mcm2-7 with 1:1:1:1:1:1 stoichiometry (52, 58 -60). Mcm4, -6, and -7 form a trimeric subcomplex known as "the MCM core." Mcm2 binds to the core and a dimer formed by Mcm3 and Mcm5. The MCMs interdigitate with one another in a ring structure in which the arginine finger of one MCM meets with the P-loop in the Walker A motif of its neighbor to form an ATP binding site (53,58). Coordinated ATP hydrolysis occurs at a subset of sites (53,58,60). Mcm2 is thought to be the gate of the ring and the site at which the ring opens to encircle chromatin (53 There is good biochemical evidence that all six canonical MCMs participate as a helicase in vivo with Cdc45 and GINS complex as cofactors (53,61,62). Mcm2 is required to recruit Mcm4, -6, and -7 into the nucleus where an intact MCM complex is necessary to retain them (15). Moreover, the phenotypes associated with mcm2 mutations are indistinguishable from mutations in other MCM subunits; if it were not a core constituent of the helicase, this would not be expected. We found no evidence that Mcb1 expression can substitute for Mcm2. Therefore, the interaction between Mcb1 and the other MCMs is likely to have some other, possibly regulatory function.
In the absence of a tight conditional allele, we used two approaches to examine Mcb1 activity. We constructed a series of deletion/truncation mutations in Mcb1 and assessed their ability to function. Most of these mutants were unable to complement an mcb1⌬ mutant. However, two mutants containing a truncation of the N terminus were apparent hypomorphs; the growing cells were elongated, and this elongation depended upon the damage checkpoint. DNA content, however, was normal, and cells formed colonies with timing similar to wild type. We conclude that attenuating Mcb1 function leads to some genome instability, which is similar to the phenotypes associated with attenuation of MCM function (e.g. Refs. 24 and 25).
We also found that overexpression of Mcb1 generates a dominant lethal phenotype. OP-Mcb1 cells showed evidence of an initiation defect characterized by an increase of cells with a sub-1C DNA content. This phenotype is reminiscent of cells with mutations in the essential replication initiation factors orc1 (63,64), cdc18 (56), and rad4/cut5 (65). The general model is that cells that do not initiate replication have no way to activate a checkpoint or register that S phase has not occurred. Thus, the cells proceed through mitosis and tear apart the unreplicated genome. We observed that OP-Mcb1 cells show a modest increase in Rad22 foci, which is indicative of DNA damage, and activation of the Chk1 damage checkpoint kinase, which has been seen in many mutants defective in DNA replication initiation (66). This may occur in the subset of cells in the population that are elongated.
This initiation-defective phenotype is likely to result from inactivation of the MCM complex. In strains overproducing Mcb1, association between the canonical MCM proteins was disrupted. This was most strikingly observed by the failure of Mcm4 to associate with Mcm2, but interaction between Mcm4 and the core MCMs was also reduced. Interestingly, the overproduced Mcb1 was able to bind to Mcm2, suggesting that when expressed at high enough levels this protein can interact with all MCM subunits.
When we examined the truncation/deletion mutants for overproduction phenotypes, we found that only the two hypomorphic mutants, which contain a short N-terminal truncation, were toxic upon overexpression. Although they were still able to form small colonies (unlike expression of the full-length protein which is lethal), they also reduced association between the canonical MCM proteins.
Several of the non-functional mutants were able to bind the MCMs but showed no evidence for complex disruption (Mcm2 and Mcm4 remained associated, for example). We conclude that there are three modes of interaction between Mcb1 and the MCM complex (supplemental Fig. S11). The first is a normal "functional" mode in which Mcb1 replaces the Mcm2 subunit. The second interaction is a "sticky" mode in which Mcb1 appears to bind nonspecifically to all the MCM subunits. This is not toxic, does not replace Mcm2, and does not disrupt the MCM complex. Finally, the third is the overproduction toxicity. Only proteins capable of functional interactions can disrupt the complex when overproduced, leading to toxicity and an apparent arrest of DNA replication initiation. Although it remains a formal possibility that the inhibitory effect Mcb1 has on the MCM complex is exacerbated by overexpression and not a representation of its true phenotype, we consider this unlikely given the correlation of functional Mcb1 with complex disassembly. We conclude that overproduction of Mcb1 causes a dramatic inhibition of replication initiation similar to that caused by mutations in the genes required for formation of the prereplication complex.
MCMs in most eukaryotes exist in three populations within the nucleus: 1) the replisome MCMs bound to chromatin at the replication fork that are detectable by ChIP, 2) "remote" MCMs bound on unreplicated DNA during S phase that can be visualized cytologically (18,47,67,68) but not by ChIP, and 3) a soluble pool not bound to chromatin. The large amount of remote MCMs creates a puzzle known as "the MCM paradox" (19,21). These "remote MCMs" are very important for distributing origins, marking unreplicated chromatin for replication, and reserving dormant replication origins to complete replication under replication stress (69 -73). Formally, Mcb1 could contribute to the formation of any of these pools possibly by disrupting the intact hexamer to change the distribution between them. This could occur by promoting removal of MCMs from the chromatin, particularly the remote MCMs. Because the Arabidopsis ETG1 protein has been linked to sister chromatid cohesion (31), another possibility is that Mcb1 changes the composition of the MCM complex to facilitate binding of cohesin assembly proteins at the fork.
It is interesting that budding yeast does not have an obvious orthologue of Mcb1, which is readily identified in other eukaryotes. In this regard, it may be worth noting that a significant difference between budding yeast behavior in S. cerevisiae compared with other systems is regulated nuclear localization. In budding yeast, MCMs cycle in and out of the nucleus during the cell cycle in a cyclin-dependent kinase-dependent pathway (74,75). Newly synthesized MCMs are preferentially transported into the budding yeast nucleus (76). By contrast, in other eukaryotes, including S. pombe, MCMs are located constitutively in the nucleus, and only their chromatin association is regulated. Dissociation of the MCM complex causes chromatin dissociation and crm1 ϩ -dependent nuclear export (15). We saw reduced MCM on chromatin in OP-Mcb1 cells, and there was a reduction in the overall nuclear signal of Mcm4 compared with Mcm2. It is possible that Mcb1 functions to regulate the MCMs at the level of chromatin association to prevent binding or activation outside of S phase. This could be a regulator function that is not needed in budding yeast. This is consistent with data from other systems. In the process of preparing this work, a new study has shown that Xenopus MCM-BP unloads MCM complex from the chromatin in late S phase to prevent rereplication by dissociating Mcm2-7 from chromatin (77). Our data support a model in which the abundant Mcb1 protein contributes to redistribution of the MCM proteins at the conclusion of S phase.