Mcm1 binds replication origins.

Mcm1 is an essential protein required for the efficient replication of minichromosomes and the transcriptional regulation of early cell cycle genes in Saccharomyces cerevisiae. In this study, we report that Mcm1 is an abundant protein that associates globally with chromatin in a punctate pattern. We show that Mcm1 is localized at replication origins and plays an important role in the initiation of DNA synthesis at a chromosomal replication origin in vivo. Using purified Mcm1 protein, we show that Mcm1 binds cooperatively to multiple sites at autonomously replicating sequences. These results suggest that, in addition to its role as a transcription factor for the expression of replication genes, Mcm1 may influence the local structure of replication origins by direct binding.

Mcm1 is an essential protein required for the efficient replication of minichromosomes and the transcriptional regulation of early cell cycle genes in Saccharomyces cerevisiae. In this study, we report that Mcm1 is an abundant protein that associates globally with chromatin in a punctate pattern. We show that Mcm1 is localized at replication origins and plays an important role in the initiation of DNA synthesis at a chromosomal replication origin in vivo. Using purified Mcm1 protein, we show that Mcm1 binds cooperatively to multiple sites at autonomously replicating sequences. These results suggest that, in addition to its role as a transcription factor for the expression of replication genes, Mcm1 may influence the local structure of replication origins by direct binding.
The mechanism for the initiation of DNA synthesis is fundamental to all eukaryotes. The protein factors involved in this process and the sequence of events that lead to the initiation of DNA synthesis are highly conserved (1,2). However, the criteria for site selection appear to differ dramatically among eukaryotes (3). At the one extreme, Saccharomyces cerevisiae initiates DNA synthesis from defined DNA segments known as autonomously replicating sequences (ARSs) 1 of about 100 -200 bp (4). At the other extreme, any random DNA segment is a potential site for initiation in Xenopus oocytes (5). The unifying common denominators that define all replication origins are unknown. Although the entire repertoire of replication origins in the yeast genome has been mapped (6,7), researchers have yet to come up with an algorithm that independently identifies replication origins.
Replication origins in S. cerevisiae were initially identified as DNA sequences known as ARSs which support the autonomous replication of recipient plasmids (8,9). ARSs are modular in structure, containing sequence elements that are interchangeable between certain ARSs (10,11). They share an essential 11-bp AT-rich sequence known as the ARS consensus sequence or ACS, which requires the presence of other auxiliary elements to function (4). An auxiliary element B2/DUE/ATR shared by all ARSs is located 3Ј of the T-rich strand of the ACS (12). Another element frequently associated with ARSs is the B3 element. The ACS is the binding sequence for the origin recognition complex, which serves as the launching pad for the assembly of the prereplication complex (pre-RC) (13). The B2 element is an AT-rich sequence that serves as the site for the initial unwinding of origin DNA and the incorporation of the first nucleotides (14). The B3 element is a 12-bp degenerate sequence which Abf1 binds. Abf1 is a transcription factor that stimulates the activity of certain ARSs (10,15). Although ARSs have modular structures, not all ARSs have the same modules or arrangement of modules. For example, the replication enhancer sequence REN but not B3 stimulates the activity of ARS1501 independent of orientation or spacing with respect to the ACS (16). The B3 element is present in two copies in ARS121 and ARS1501, one copy in ARS1, and not at all in ARS307. Some ARSs, such as ARS1 (10) and ARS307 (17), contain a single ACS, whereas others, such as ARS121 (15), ARS110, and ASR310 (18), contain repetitive ACSs. Because only a couple of modules are obligate components, and others are auxiliary components, there would be many ways to assemble combinations of modules into a functional unit. Thus, although ARSs are modular in structure they share little sequence homology. It is becoming increasingly clear that ARSs that function in the context of a plasmid may only be part of a larger structure that must be preserved for regulation or efficient function in its native environment (18 -20). Therefore, analysis of extended sequences of an ARS is important for identifying new elements of chromosomal replication origins.
In an attempt to identify common denominators for site selection, researchers have shifted their emphasis from sequence analysis to the characterization of factors that bind replication origins (2). Among the common factors recruited to and involved in the establishment of pre-RCs (origin recognition complex, Mcm2-7 helicase, Cdc6, and Cdt1), only the origin recognition complex has been shown to bind specifically origin DNA. Thus, the identification of additional factors that can bind to and/or facilitate the assembly of pre-RCs at replication origins may pave the way toward a broader definition of replication initiation sites for all eukaryotes. For this reason, S. cerevisiae remains the best model for identifying factors that bind replication origins because of its well characterized replication origins and amenable genetics.
Mcm1 was identified initially as a factor required for the stable maintenance of minichromosomes in S. cerevisiae (21). Because of its role as a general transcription factor that regulates diverse genes including the cell cycle expression of a number of replication initiation factors, it is generally believed to play an indirect role in replication initiation (22). Among the genes regulated by Mcm1 are genes that encode critical com-ponents of the pre-RC such as CDC6, MCM3, MCM5, MCM6, and MCM7 (23,24). Mcm1 is the founding member of a family of combinatorial transcription factors known as the MADS box proteins (25). MADS proteins interact with multiple cofactors to regulate diverse genes by binding cooperatively with their cofactors to their combinatorial cognate sequences. The mechanisms by which Mcm1 regulates the expression of different genes vary widely. It may act as a positive regulator that recruits transcription activation machinery (26) or as a negative regulator that recruits repression complexes (27). It is also believed to play a role in positioning nucleosomes (28) or interfering with nucleosome assembly (29). The crystal structure of the Mcm1⅐␣2⅐DNA ternary complex has been solved (30). Like other MADS box proteins, Mcm1 dimers bind a 10-bp consensus sequence CC(AT) 6 GG interacting symmetrically with each half of the binding site to form a bent DNA structure (31,32). Direct and indirect DNA binding assays suggest that a conserved 5Ј-ATTAGG in one half of the dyad symmetry element is important for Mcm1 binding, whereas variants in the other half are tolerated, conferring versatility to Mcm1 sequence recognition (33).
In this study, we show that Mcm1 is a sequence-specific DNA-binding protein that imparts a global presence on chromatin. Chromatin immunoprecipitation experiments reveal that Mcm1 is localized at ARSs in vivo. In vitro footprinting analysis shows that Mcm1 binds and induces DNase I hypersensitivity in the immediate vicinity of the minimal ARS domain, suggesting that Mcm1 may play a direct role in organizing the local structure of chromosomal replication origins.
Minichromosome Maintenance Assays-Minichromosomes were transformed into yeast using the lithium acetate method (35). Minichromosome stability assays were performed on a minimum of three independent transformants at room temperature as described previously (36). Standard deviation from the mean value of three samples is less than 10%.
Two-dimensional Gel Electrophoresis-Procedures for two-dimensional gel electrophoresis analysis of replicating DNA have been described by Brewer and Fangman (37). For the study of ORI1, actively replicating DNA was prepared from isogenic wild type and mcm1-1 mutant strains. The probes for detecting ORI1 were prepared from ARS1 plasmid DNA digested with NcoI, gel purified, and 32 P-labeled by random oligonucleotide priming (38).
Yeast Chromatin Spreads-Yeast chromatin spreads were performed on a yeast strain containing 13Myc::Mcm1 as described previously (39) except that 1.5% paraformaldehyde rather than 3% was used in the spread. Primary antibodies used were rabbit anti-H2B and mouse anti-Myc (Covance) as described previously (40). Secondary antibodies used are anti-mouse Oregon Green and anti-rabbit Texas Red (Molecular Probes). DNA is stained with DAPI.
Chromatin Immunoprecipitation-Chromatin IP experiments were performed using formaldehyde cross-linking as described by Hecht et al. (41) with minor modifications. Affinity-purified Mcm7, Mcm1-Myc, and HA antibodies were used to immunoprecipitate chromatin cross-linked complexes. Individual PCRs were carried out with selected primer pairs and 5 Ci of [␣ 32 P]dATP in the following conditions: 2 cycles (94°C, 2 min; 96°C, 1 min; 54°C, 4 min), 24 cycles (94°C, 1 min; 54°C, 1.5 min), and 1 cycle (70°C, 5 min). PCR for the chIP was performed in a multiplex reaction with the ATP11 primers included in the CDC6, ARS1, and ARS305 reactions. All three primer pairs were added in the same reaction for the chIP of ARS121 and ARS307.
Radioactive Labeling of DNA Fragments by PCR-PCR primers were labeled by incubating 50 pmol of primer, 30 Ci of [␥-32 P]ATP, and 10 units of T4 polynucleotide kinase for 15 min at 37°C. The reactions were stopped at 65°C for 10 min, ethanol precipitated, and resuspended in water. The complementary primer, dNTPs, Taq polymerase, 10 ng of DNA template, 10ϫ PCR buffer, and MgCl 2 were added. PCR amplification was carried out for 35 cycles. ARS121, ARS1, and ARS307 primers used for mobility shift and DNase I footprinting experiments are listed above.
Mobility Shift Experiments-Purified proteins were mixed with radioactively labeled DNA (ϳ100 fmol) in buffer 1 to a final volume of 20 l. 500 ng of poly(dI⅐dC) was added to each reaction and incubated at room temperature for 20 min prior to gel electrophoresis (5-6% TBE, 5% glycerol PAGE). Gels were dried onto 3MM paper and either exposed to film or PhosphorImager screen. Specific competitor DNA used is actaatttacccagaaaggaaatttccttataaggaaaataaatgcaattcattaagtcg. Nonspecific competitor DNA used is tcgacacgcgcgcgtgattacaaagacgatgacgataagctagccgcgcgtg.
DNase I Footprinting-DNase I footprinting was performed as described for mobility shift reactions. A precalibrated amount of DNase I (Sigma) was added in a volume of 0.01 ml for 1 min at room temperature. The reaction was stopped by precipitation in a mixture of ethanol, 0.3 M sodium acetate, and 1 g of linear acrylamide. The precipitate was washed with 70% ethanol, dried, and resuspended in formamide buffer before loaded onto a 6 M urea, 6% polyacrylamide sequencing gel. DNA sequencing reactions were performed using the dsDNA cycle sequencing system (Invitrogen).

Mcm1 Is Essential for Efficient Plasmid Replication and Initiation of DNA Synthesis at a Chromosomal Replication
Origin-The first clue that the Mcm1 protein is essential for the efficient replication of ARSs comes from the ARS-specific minichromosome maintenance defect of the mcm1-1 mutant (42). Using the plasmid stability assay that measures ARS activity, representative ARSs were tested (complete list not shown). These include single copy ARSs (ARS121, ARS1), moderately repetitive ARSs (ARSTel131a) (43), early firing (ARS121, ARS1), and late firing ARSs (6) located either at the silent mating type locus (ARSHML) (19) or telomere (ARSTel131a) (44). Compared with their activity in wild type cells, these six representative ARSs replicate much less efficiently in the mcm1-1 mutant (Fig. 1A). Plasmids bearing these ARSs are lost at a rate of 1-5%/cell division in wild type strain and 14 -23%/cell division in the mcm1-1 mutant. Thus, the Mcm1 protein is required for the efficient initiation at ARSs. Because the mcm1-1 mutation affects the activity of ARSs that differ in chromosomal location, copy number, and temporal order of replication during S phase, this observation suggests that Mcm1 may operate through a common mechanism of action at different types of ARSs.
To show that the requirement of Mcm1 for replication initiation also applies to chromosomal origins, initiation bubbles generated de novo from ORI1, which corresponds to ARS1, were analyzed by two-dimensional gel (Fig. 1B). Robust initiation activity is observed at ORI1 in the wild type strain, but this activity is greatly reduced in the mcm1-1 mutant as indicated by increased passive replication through the ORI1 region (strong "Y" arc intensity).
Mcm1 Is an Abundant Chromatin-associated Protein-Because the mcm1-1 mutation was isolated from a mutant screen that also identified several members of the pre-RC (Mcm2-Mcm7, Mcm10), we investigated the possibility that Mcm1 may be directly involved in DNA replication separate from its role as a transcription factor. As a general transcription regulatory factor and a sequence-specific DNA-binding protein, Mcm1 is expected to be associated with chromatin in modest abundance.
The distribution of Mcm1 on chromatin was visualized by immunofluorescence microscopy of chromatin spreads of G 1 phase cells in which the endogenous MCM1 gene has been replaced by MCM1-Myc and probed with either anti-Myc or anti-H2B antibodies (Fig. 2). Mcm1 is localized in more than 500 foci distributed throughout the genome to give a global punctate pattern similar to that observed for histone H2B. A general comparison and overlay of these staining patterns suggest that the abundance of Mcm1, in terms of the number of intensely stained foci, may be similar to histone H2B protein. However, Mcm1 does not colocalize with histone H2B. Considering that Mcm1 is a sequence-specific DNA-binding protein whereas histones interact with DNA in a sequence-nonspecific  1. Mcm1 is required for efficient replication. Replication activity is measured by plasmid stability and two-dimensional gels in wild type (WT) and mcm1-1 mutant yeast strains. A, ARS activity assayed by plasmid stability in wild type 8534-8C and mcm1-1 mutant. A variety of ARS fragments were tested including the single copy ARS121, ARS1, ARSHML, ARSH2B, and ARSHO and the moderately repetitive subtelomeric X-ARSTel131a. For each ARS type, at least three independent transformants were grown in nonselective medium for 8 -10 generations. B, origin initiation assayed by two-dimensional gel. Autoradiograms of replicating DNA prepared from isogenic wild type and mcm1-1 mutant strains are shown, separated by two-dimensional gel electrophoresis. For ARS1, DNA was digested with NcoI and probed with the 5-kb NcoI-NcoI fragment. The direction of electrophoresis was left to right for the first dimension and top to bottom for the second dimension. The solid arrow points to "bubble" arcs; the hollow arrow points to Y arcs. manner, the similarities in abundance and genomewide distribution of these two proteins were unexpected.
Mcm1 Is Specifically Localized to Replication Origins in Vivo-Mcm1 appears to be a highly abundant protein that has a global presence on chromatin. To investigate whether Mcm1 specifically binds replication origins in vivo, we carried out chromatin immunoprecipitation experiments (Fig. 3A). Mcm1⅐DNA complexes cross-linked by formaldehyde were precipitated with anti-Mcm1 (anti-Myc) antibodies (Fig. 3A, lane 3). Specific PCR primer pairs were used to amplify four well characterized chromosomal replication origins corresponding to ARS1, ARS121, ARS305, and ARS307. Each primer pair was designed to amplify a 200 -400-bp region encompassing the essential ACS and the surrounding flanking region. Also amplified were the CDC6 promoter and the ATP11 gene to use as a positive or negative control for cross-linking. In addition, antibodies specific to Mcm7 (lane 2) and HA (lane 1) were used as positive or negative controls, respectively, for immunoprecipitation.
When Mcm1 (or Mcm7) DNA complexes were immunoprecipitated, ARS121 and CDC6 were specifically amplified to give a strong signal (Fig. 3B). Because Mcm1 and Mcm7 have been shown to bind cooperatively to the promoter of CDC6, 2 amplification of CDC6 from anti-Mcm1 and anti-Mcm7 immunoprecipitation serves as a positive control. These signals were normalized against that obtained from the immunoprecipitates of the HA antibody. ARS305, ARS307, and ARS1 were also specifically amplified though not as strongly as ARS121 or CDC6 (Fig. 3B). Within the resolution of the chIP experimental approach, both Mcm7, a known component of the pre-RC (45), and Mcm1 appear to be localized in close proximity to replication origins. Because Mcm1 has specific DNA binding properties that the Mcm2-7 complex lacks, we sought to determine which cis-acting sequences are responsible for the binding of Mcm1 to replication origins in vivo.
Mcm1 Binds Specific Sequences at ARSs-To investigate where Mcm1 binds at ARSs, we purified an E. coli-expressed GST-Mcm1 fusion protein (Fig. 4A) and carried out DNA binding assays on ARS121, ARS1, and ARS307. For each ARS, we designed PCR primer pairs that produced two or three overlapping 200 -400-bp DNA fragments that encompass a 989-bp region of ARS121 (Fig. 4B), a 420-bp region of ARS1 (Fig. 4C), and a 580-bp region of ARS307 (Fig. 4D). Increasing amounts of Mcm1 protein were incubated with ARS121 fragments I-III, ARS1 fragments I and II, or ARS307 fragments I and II in the presence of carrier DNA.
Mcm1 binds strongly to fragment II of ARS121, which contains the proximal B3, ACS, and ATR, and fragment III (domain B), which contains the ATR and adjacent 3Ј-flanking sequence (Fig. 4B). The binding of fragment II appears to be cooperative, showing an all or none binding pattern within a 2 M. Fitch, unpublished results.  lanes 10 and 11). In comparison to ARS121, Mcm1p binds weakly to both fragment I and fragment II of ARS1, requiring almost 10 times as much Mcm1p to produce a total shift of the same mole equivalent of DNA (Fig. 4C, 32.4 pmol, lanes 5 and 10 and Fig. 4B, 3.6 pmol,  lane 11). At low concentrations of Mcm1 (Fig. 4C, lanes 3 and 4,  8 and 9), a prominent species of Mcm1p⅐DNA complex is evident. But at higher concentrations (lanes 5 and 10) Mcm1p⅐DNA complexes of higher molecular weight and heterogeneous mobility are formed (indicated by a bracket). Mcm1p also binds cooperatively to ARS307 (Fig. 4D, lanes 6 and 7). Mcm1p binds fragment II, which contains the ACS, B1, and B2 elements but lacks the B3 element, to form a complex of uniform mobility at low concentrations (lanes 5 and 6) but complexes of heterogeneous mobility at higher concentrations (lanes 7 and 8). In all, the binding of Mcm1 to these ARS DNA fragments appears to be specific, concentration-dependent, and highly cooperative. Furthermore, Mcm1p binds different ARSs with different affinities.

3-fold difference in concentration (compare
Mcm1 Binds Cooperatively to Multiple Sites at ARS121-To confirm that binding of Mcm1 is sequence-specific and to determine the exact nucleotide sequence that is recognized we performed in vitro DNase I footprinting analysis on each of the ARS fragments. Purified Mcm1 was incubated with different ARS fragments from ARS121, treated with DNase I nuclease, and run on a denaturing gel to visualize the DNase I protection patterns (Fig. 5).
Protection of multiple sites in the central region of ARS121 between the B3 element and the essential ACS by Mcm1 is evident on both strands of fragment II (Fig. 5, A and B). There are at least eight protected sites on this DNA fragment, some corresponding to full Mcm1 consensus elements (MCEs) and others not (Fig. 5C). Two of these protected sites overlap the two B3 elements (Fig. 5A, a). Four adjacent Mcm1-protected sites, labeled MCE I-IV, are clearly apparent (Fig. 5A, b). Two of the MCEs, I and III, are more strongly protected than MCE sites II and IV (Fig. 5A, b; compare lanes 1 and 2). Further-more, protection of sites I and III precedes that of the other sites as the amount of Mcm1 increases by 2-fold. Upon Mcm1 binding, there is also increased DNase I hypersensitivity in the boundary sequences between each of the four MCEs, suggesting distortion of the Watson-Crick base pairing at these sequences. In the region extending from MCE sites I-IV toward  the ACS and ATR region, increased DNase I hypersensitivity occurs at a high frequency (Fig. 5, B and C). Although one or two protected MCEs appear to be located close to the ACS region, the minimal functional domain of ARS121 (ACS-ATR) is largely free of Mcm1 protection.
Nucleotide analysis of the DNase I-protected regions show that a majority of the Mcm1 binding sites resemble the 10-bp MCE, CC(AT) 6 GG, or the 5-bp half-site. For inspection, these protected sequences broken down into 15 half-sites were aligned (Fig. 6A). The consensus sequence from this alignment is C(C/T)(T/A)A(T/A), which is in general agreement with the MCE half-site. To verify specificity of binding, we compared the binding affinities of Mcm1 and the mutant Mcm1-1 protein for ARS121 DNA (Fig. 6B). Mcm1-1, which contains a P97L substitution, is the only known mcm1 mutant that causes a defect in plasmid replication (21). GST-Mcm1 binds ARS121 DNA efficiently (lanes 2 and 3), whereas GST-Mcm1-1 binds poorly (lanes 4 and 5), correlating origin binding with minichromosome maintenance.
A DNA competition experiment was also carried out to verify the binding specificity of Mcm1 for ARS121 (Fig. 6C). A 60nucleotide DNA fragment from the MCM7 promoter, which contains two MCEs, was used as a specific competitor (lanes [3][4][5], and a 55-nucleotide DNA fragment from a plasmid vector that does not contain obvious MCE half-sites was used as the nonspecific competitor (lanes 6 -8). The specific competitor competes efficiently with ARS121 DNA for the binding of Mcm1. At the same concentration as the ARS121 DNA probe, binding of Mcm1 is about 50% competed, and at 10 times the concentration of the labeled probe almost all Mcm1 binding is competed (Fig. 6D). Under the same conditions, the nonspecific competitor has little effect on the binding of Mcm1 to the ARS121 DNA probe (Fig. 6D).
Mcm1 Binds Cooperatively at Multiple Sites at ARS1-Interaction between Mcm1 and ARS1 was also investigated by DNase I footprinting analysis. In fragment I of ARS1, Mcm1 protects three closely spaced sites in the 3Ј-flanking region outside the minimal ARS (Fig. 7A, b). These three sites map to two recognizable MCEs that overlap a known B3 element. At higher Mcm1 concentrations, protection in this region occurs with the simultaneous appearance of hypersensitivity at the boundary between the MCEs and B3 element (Fig. 7A, b, lanes  1 and 2). In addition, weaker protected and hypersensitive sites are apparent near the B1 and B2 regions (Fig. 7A, a, lane 1, and  7B). Using other fragments of ARS1 (domain C), we did not observe sequence-specific binding (data not shown). Instead, we observed some degree of "white-out" along the entire fragment, suggesting that the entire region may be protected nonspecifically. These results show that in both ARS121 and ARS1, Mcm1 binds cooperatively to multiple MCEs outside the minimal functional domain (ACS-B2/ATR). In addition, Mcm1 appears to footprint MCEs that overlap the B3 elements in both ARS121 and ARS1.

DISCUSSION
Mcm1 was initially identified as a protein essential for the maintenance of minichromosomes. It was subsequently shown that Mcm1 is a general transcription factor that regulates diverse genes including replication genes such as CDC6, MCM3, MCM5, MCM6, and MCM7 (24). For this reason, the requirement for Mcm1 in minichromosome maintenance was attributed to its role in regulating the expression of proteins essential for replication initiation. However, unlike most general transcription factors, Mcm1 is essential for viability (21), suggesting that it may have other essential functions. In this study, we show that Mcm1 is required not only for plasmid replication but also for the initiation of replication at a chro-mosomal replication origin. Mcm1 is an abundant chromatinassociated protein whose global presence on chromatin can be visualized by immunofluorescence microscopy as punctate stains similar to those of histone H2B. A genomic search for the Mcm1 binding motif CCNNNWWRGG derived from known Mcm1 binding sites at promoters indicates that this motif occurs in 17% of the 6,270 intergenic regions of the S. cerevisiae genome (web.wi.mit.edu/young/). Keeping in mind that binding motif is not a predictor of protein binding in vivo and that a search algorithm does not recognize binding sites that deviate from consensus sequence (24), an estimate of ϳ1,000 genomic locations for Mcm1 is consistent with the distribution of Mcm1 in immunofluorescence microscopy. We show that Mcm1 is localized to replication origins in vivo by chIP. Gel mobility shift and DNase I protection assays show that Mcm1 binds to specific DNA sequences in all ARSs tested. A mutation in Mcm1 that is defective in minichromosome replication is also defective in binding to ARS DNA. These results together suggest that Mcm1 plays a direct role in DNA replication by binding to replication origins.
For the three ARSs studied, Mcm1 binds strongly to ARS121, less strongly to ARS307, and weakly to ARS1. A comparison of the Mcm1 protection patterns of these ARSs is shown in Fig. 8. The location and nucleotide sequences of the Mcm1 binding sites at these ARSs have a number of common features yet differ in their overall pattern and frequency. We have mapped the Mcm1 binding sites to specific nucleotide sequences. Mcm1 protects about 11 sites in ARS121 (989 bp) and fewer sites in ARS307 (580 bp) and in ARS1 (420 bp). The varied protection patterns of ARS121, ARS1, and ARS307 suggest that the degree of involvement of Mcm1 at individual ARSs may be quite different. Differential affinities of Mcm1 to promoters with a varying number of MCEs have also been observed (33,46). The affinity of Mcm1 to ARS DNA is similar to that reported for promoter DNA in the absence of a cofactor (33). Involvement of cofactors that stimulate the binding of Mcm1 to origin DNA in vivo cannot be ruled out.
Most of the Mcm1 binding sites are outside of the ACS, B1, or B2/ATR elements. Previous studies demonstrated the importance of the B3 element in ARS121 and ARS1 by site-directed mutagenesis (10,47). In this study, we show that Mcm1 binds both ARS1 and ARS121 at sites that overlap each of the B3 elements. Furthermore, the mcm1-1 mutant affects the activity of both ARSs. Although Abf1 has been shown to bind B3 in vitro, the abf1 mutant only affects the activity of ARS121 but not ARS1 (48). The apparently contradictory results for the role of Abf1 at ARS1 could be explained if the degenerate "B3" element were the collective site of action of auxiliary factors such as Mcm1 and Abf1 (49,50), which have relaxed recognition sequences. Consistent with this hypothesis, the Abf1 consensus motif TCRNNNNNNACG occurs in 24% of intergenic regions (web.wi.mit.edu/young/). Furthermore, unlike the B1 and B2 elements, which are strictly associated with the B domain, B3 may be located in either domain B or domain C of an ARS (Fig. 8).
In ARS121 and ARS1, the Mcm1 protection sites are interspersed with DNase I hypersensitive sites, suggesting that binding of Mcm1 may induce bending of DNA. A majority of the protected regions contain recognizable 10-bp CC(A/T) 6 GG MCE sites or half-sites; however, some do not. It is possible that sequence-specific binding of Mcm1 at a few high affinity sites may stimulate association of Mcm1 with neighboring weaker sites. This explanation is consistent with the observation that Mcm1 forms a distinct protein-DNA complex at low concentrations and higher molecular weight complexes of heterogeneous mobility at higher concentrations. Indeed, binding of Mcm1 to multiple sites at ARSs appears to be cooperative. An example for a nucleating sequence in the cooperative binding of a replication initiator protein has been reported for the binding of the E. coli DnaA protein to OriV of the RK2 plasmid (51).
What might be the possible function of Mcm1 that binds mostly outside of the minimal functional domain of ARSs? It is possible that Mcm1 together with other auxiliary factors, such as Abf1, may regulate the activity of an ARS by altering the accessibility of the ACS and B2/ATR element to replication initiation factors. It may do so in one of several ways. First, it may play a role in positioning nucleosomes as suggested for Abf1 (52,53) to modulate the accessibility of the ACS-B2 region. Indeed, Mcm1 together with its cofactor ␣2 have been shown to bind their cognate sequence to phase nucleosomes (29,54). Conversely, Mcm1, without the ␣2 protein, has also been implicated in interfering with nucleosomal array assembly by binding to the recombination enhancer sequence near HML in a mating type cells (29). Second, Mcm1 binding to origin DNA may recruit replication initiation factors there. Mcm1, together with its cofactors, has been shown to act as a recruiting platform for transcriptional repression or activation complexes (27,55). It is not difficult to imagine that Mcm1 may play a similar role at ARSs to recruit pre-RCs. Third, one can also envision a role for Mcm1 in bending or looping DNA to enhance accessibility of the minimal functional domain of ARSs. A classical example for looping can be found in the replication origin of OriP of Epstein-Barr virus (56,57). The Epstein-Barr viral encoded transcription factor EBNA has been shown to bind multiple sites flanking OriP to form a loop structure that facilitates the binding of the pre-RC (58 -60). Finally, it is possible that the structural changes induced by Mcm1 binding could facilitate interaction between the large number of replication factors associated at replication origins. As proposed for DnaA and integration host factor at E. coli OriC, the binding of auxiliary factors causes DNA bending that helps juxtapose and facilitate interaction between replication initiation factors bound at faraway cis-acting elements (61).
In summary, genetic and biochemical evidence from this and other studies is consistent with Mcm1 playing a direct role in replication initiation at replication origins. Further analysis is necessary to elucidate the precise functions that Mcm1 may serve at individual ARSs on plasmids and at their native chromosomal locations. The global presence of Mcm1 on chromatin suggests that Mcm1 may play a ubiquitous role regulating chromatin structure at a genomic scale. A direct role for Mcm1 in DNA replication may have important implications for other MADS box transcription factors, especially those that stimulate mitogenic growth in vertebrates.